Detecting Change between Urban Road Environments along a Route Based on Static Road Object Occurrences

Zoltán Fazekas; László Gerencsér; Péter Gáspár

doi:10.3390/app11083666

Fazekas, Zoltán;Gerencsér, László;Gáspár, Péter

2021-04-19 00:00:00

applied sciences Article Detecting Change between Urban Road Environments along a Route Based on Static Road Object Occurrences 1 , 2 , 1 1 Zoltán Fazekas * , László Gerencsér and Péter Gáspár Institute for Computer Science and Control (SZTAKI), Eötvös Loránd Research Network (ELKH), 13-17. Kende Utca, H-1111 Budapest, Hungary; laszlo.gerencser@sztaki.hu (L.G.); peter.gaspar@sztaki.hu (P.G.) Department of Control for Transportation and Vehicle Systems, Faculty of Transportation Engineering and Vehicle Engineering, Budapest University of Technology and Economics (BME), 2 Stoczek Utca, H-1111 Budapest, Hungary * Correspondence: zoltan.fazekas@sztaki.hu Featured Application: A road environment-type (RET) detection function could improve the road awareness of inexperienced car drivers, especially in urban areas, and by doing so, it could slightly raise the urban trafﬁc safety. A pragmatic implementation could make use of static road object data, e.g., trafﬁc sign (TS) data, that is already collected and available on-board. It could rely on the TS recognition function offered by advanced driver assistance systems (ADAS). Fur- thermore, apart from its primary function, the RET detection system could provide reciprocal information—with respect to the current RET—for various ADAS and autonomous driving (AD) computations and subsystems. Making use of such reciprocal information could speed up the ADAS/AD computations, and render their results more accurate and more reliable, e.g., via intro- ducing parameter constraints and marking regions-of-interest. Abstract: For over a decade, urban road environment detection has been a target of intensive Citation: Fazekas, Z.; Gerencsér, L.; research. The topic is relevant for the design and implementation of advanced driver assistance Gáspár, P. Detecting Change between systems. Typically, embedded systems are deployed in these for the operation. The environments can Urban Road Environments along a be categorized into road environment-types. Abrupt transitions between these pose a trafﬁc safety Route Based on Static Road Object risk. Road environment-type transitions along a route manifest themselves also in changes in the Occurrences. Appl. Sci. 2021, 11, 3666. distribution of trafﬁc signs and other road objects. Can the placement and the detection of trafﬁc https://doi.org/10.3390/app11083666 signs be modelled jointly with an easy-to-handle stochastic point process, e.g., an inhomogeneous Academic Editor: Luís Picado Santos marked Poisson process? Does this model lend itself for real-time application, e.g., via analysis of a log generated by a trafﬁc sign detection and recognition system? How can the chosen change detector Received: 24 March 2021 help in mitigating the trafﬁc safety risk? A change detection method frequently used for Poisson Accepted: 14 April 2021 processes is the cumulative sum (CUSUM) method. Herein, this method is tailored to the speciﬁc Published: 19 April 2021 stochastic model and tested on realistic logs. The use of several change detectors is also considered. Results indicate that a trafﬁc sign-based road environment-type change detection is feasible, though Publisher’s Note: MDPI stays neutral it is not suitable for an immediate intervention. with regard to jurisdictional claims in published maps and institutional afﬁl- Keywords: marked Poisson processes; change detection methods; urban road environment detection; iations. trafﬁc sign detection and recognition; advanced driver assistance systems Copyright: © 2021 by the authors. 1. Introduction Licensee MDPI, Basel, Switzerland. Despite of the on-going research on self-explaining road layouts and designs [1,2], and This article is an open access article on the computerized recognition methods of such designs and layouts, e.g., on methods distributed under the terms and that apply artiﬁcial intelligence methodology [3], setting up trafﬁc signs (TSs) along the conditions of the Creative Commons roads and trafﬁc lights in road junctions and near pedestrian crossings by the transport Attribution (CC BY) license (https:// authorities still remains a customary measure for reducing trafﬁc safety risks in urban creativecommons.org/licenses/by/ areas [4]. Clearly, there are other viable alternative measures, as well as supplementary ones 4.0/). Appl. Sci. 2021, 11, 3666. https://doi.org/10.3390/app11083666 https://www.mdpi.com/journal/applsci Appl. Sci. 2021, 11, 3666 2 of 17 for the purpose. These include—among many others—the installation of speed reduction markings onto the road-surface [5] and the installation of vehicle- to-infrastructure (V2I) communication facilities, e.g., to succor the TS recognition (TSR) function offered by advanced driver assistance systems (ADAS) [6] and self-driving cars [7]. In a wider sense, V2I communication succors the road, trafﬁc and vehicle data gathering, fusion, and dissemination, and through these data processes, it is expected to have a signiﬁcant beneﬁcial impact on trafﬁc safety [8]. More speciﬁcally, V2I communication can be used for raising the road-awareness of car-drivers, as well as that of the intelligent and the self-driving road vehicles. Furthermore, it can be used for providing the human drivers and the smart vehicular systems with current trafﬁc information with respect to the region, town, and area, on the one hand, and with some very speciﬁc dynamic information on individual vehicles in the vicinity, on the other [9]. When speaking about raising road awareness of drivers, one is obliged to speak about the Global Navigation Satellite System (GNSS), a system that is used by masses of people around the world. According to [10], the GNSS devices per capita averaged out at 0.8 across the countries of world in 2019. The GNSS is used with wide variety of devices running map-based applications, a signiﬁcant percentage of these devices are installed on-board cars. The brief history of the navigational systems and their respective precisions are presented in [11]. The paper provides a fresh outlook on the navigational needs of and the available navigational solutions for autonomous vehicles and systems. As it often happens to popular services, devices, and applications, threats against these surface from time to time. Such threats have surfaced also against the GNSS service [12]. Although the number of successful navigational spooﬁng attacks is still negligible, the navigational signal deteriorations due to other—i.e., non-hostile—factors are clearly not. For instance, the signal reception is often brought down, or even blocked by the high-rise buildings in densely built urban areas. Some examples in this context are presented in [13]. The speed reduction measures implemented in urban areas are motivated by the trafﬁc safety concerns associated with the intense road trafﬁc and the limited space available there for the driving maneuvers [14]. While driving, and particularly while driving in urban areas, drivers need to perform numerous mental and control tasks—ranging from those associated with limb-movement to those required for complex driving maneuver planning and execution—within stringent time and spatial constraints and with high reliability [15]. Furthermore, these tasks must be performed in presence of disturbances, such as unfavorable lighting, adverse weather, and trafﬁc conditions [16]. In addition, the older age of the driver may contribute to the perceived difﬁculty of these tasks [17]. A system, which pays attention to the driver ’s activity within the car and also to aspects of the urban road environment, was developed as part of the Urban Intelligent Assist Research Initiative some years ago [18], and since then, other systems with similar, or enhanced capabilities followed suit [19,20]. The effect of driving experience on drivers’ adaptation to road environment complexity—a notion closely related to that of the road environment type (RET) used herein—in urban areas was investigated in a simulation study [21]. The ﬁndings of the study underline the need for an automatic RET detection function, and indicate that such a function is particularly useful for car-drivers lacking prolonged driving experience, and also for older drivers. Several algorithmic approaches and sensor arrangements were devised, applied, and tested for detecting, characterizing, and categorizing urban road environments based on image and/or point cloud data [22–24]. In the application considered herein, the urban road environment appears around and sweeps past an ego-car while it is driven in an urban area. The data streams used for the purpose of road environment detection and analysis originate—among others—from one or more camera and one or more light detection and ranging (LiDAR) sensor. In a viable implementation of a road environment detection and classiﬁcation system that is capable of assisting a car driver while driving, either a comprehensive real-time on-board processing of the respective raw data streams is required (direct processing) or a timely access to and further processing of the data—rendered by Appl. Sci. 2021, 11, 3666 3 of 17 some other real-time application/subsystem on-board—on certain distinguishing road objects (ROs) are necessary (indirect processing). In the above cited papers, the real-time requirements were limited to data synchro- nization and data collection issues, while the bulk of the processing, e.g., simultaneous localization and mapping (SLAM), and object segmentation, were carried out in a post- processing manner. Nevertheless, a large portion of the processing presented in these papers is real time capable and could be used in direct implementations. The approach presented in [22] builds and then segments the point cloud originating from a ground-level LiDAR device moving along a given trajectory. The aim of the authors was to produce editable—simpliﬁed, but visually still pleasing—object-models that lend themselves for fast visualization. The target areas were the residential urban areas in the United States. These areas are characterized by their low-rise buildings without strong and extensive repetitive patterns. The semantical labeling and various analysis steps follow the mentioned preprocessing steps. Simple models of the individual houses in the area are then created. The basic building blocks of the models are simple, symmetric, and convex geometric blocks. These blocks—together with their spatial arrangement and their connection graph—form an easy-to-handle geometric model of the individual buildings. By aggregating the certain features of the individual buildings for an area (e.g., by computing the average building dimensions and the average distance between nearest buildings), the residential urban road environment can be adequately characterized. The system presented in [23] extracts the characteristics of individual buildings rather than those of more extensive road environments. Nonetheless, the building characteristics, such as building height and building complexity—again aggregated for a given area, or along a route—together with the spatial densities of the buildings there, are deﬁnitive in the respect of the RET. A multi-sensor and multi-precision data collection campaign is described in [24]. It was a car-based campaign that made use of an array of different environment perception, navigational, and motion sensors. These included four LiDARs, a pair of stereo cameras, a ﬁber optics gyroscope and encoder sensors for the tires. The data collection trips covered diverse complex urban environments in Korea, with a clear emphasis on those environ- ments, where GPS reception is highly unreliable. The collected data were organized into a publicly accessible dataset that includes the measured ego-car trajectories, the raw and processed point cloud data from the LiDAR sensors, as well as the ego-car trajectories with improved precision computed via SLAM. Other approaches, e.g., the ones presented in [25,26], rely object-level data as inputs to the urban RET detection function, i.e., they follow an indirect processing approach. In a feasible realization, the raw data streams originate from the very same sensors as in the direct case, but the respective data streams reach the RET detection subsystem only after having been processed and considerably compressed by one or more ADAS subsystem. The resulting data are an object-level description of the road environment, i.e., an RO log. This log serves as an input to the indirect RET detection function. The on-board data processing described above, as well as other road, trafﬁc, and vehicular data processing carried out in various ADAS subsystems (e.g., lane detection, TSR, detection of nearby vehicles) can also be looked at from, analyzed with respect to, and formulated using a static reference point. Setting up and using a local dynamic map (LDM) [27] could serve these purposes, and provide additional conceptual support for the developers of ADAS functions. LDM is a widely used model for representing, and a standardized technology for integrating static, temporary, and dynamic road, trafﬁc, and vehicular information into a static geographical context by means of a common coordinate reference. Customarily, it has four object layers describing and managing ROs that are subject to change and exhibit dynamics at different time scales. More concretely, these layers store and handle data on permanent static, transient static, transient dynamic, and highly dynamic ROs, respectively. For instance, when framing the static RO-based urban RET detection task in LDM, the ego-car is seen as a highly dynamic RO. A crossroads Appl. Sci. 2021, 11, 3666 4 of 17 (CRs) intersection of streets is a permanent static RO in this model. The intersection can be associated with other ROs (e.g., with ﬁxed trafﬁc lights located there, which are transient static ROs). The lanes, lane markings, pedestrian crossings, and the conventional TSs are transient static ROs, while the TSs displayed by variable message sign boards can be classiﬁed as transient dynamic ROs. The RETs can be treated as permanent static features of an area or of a sub-network of roads. One could look at the ROs in an urban settlement and collect and compile their location and categorical data into a map layer, e.g., in the way data contributors of OpenStreetMap maps do with roads, railways, rivers, and various locations of importance [28]. By selecting appropriate subsets of TSs—i.e., TS subsets that are characteristic to certain RETs—various sublayers of the RO layer can be created, displayed, and analyzed. The analysis could include a Delaunay triangulation of the TS locations within a sublayer, and then one could look for dense clusters of triangles in the generated structure. By carrying out similar processing for a number of sublayers, a TS-based RET categorization of the urban area can be created. By further processing the map-based representation of TSs and other ROs, one could derive other interesting sublayers that relate to seasonal, weekly, or daily validity of the TSs and could derive a sublayer representing weather-related TSs (e.g., TSs applicable for wet, snowy and icy road conditions). For instance, the sublayer representing the within-the-day validity of TSs—indicated by auxiliary signs or time intervals attached to the TSs—should reﬂect the daily dynamics of trafﬁc source and sink structure of the area [29]. Clearly, the mentioned dynamics are closely related to the RET categorization used herein. In our view, such sublayers—compiled, e.g., from data gathered in car-based data collection campaigns—could give useful hints to road authorities and administration as to where to place additional TSs and auxiliary signs or remove unnecessary existing ones. Herein, however, we stick to the route-based sampling of the TSs of the urban area, the map-based processing touched upon above will be addressed in further research. In [25,26], the urban road environments were categorized into three RETs, namely, into downtown (Dt), residential (Res), and industrial/commercial (Ind) areas. The ROs represented in the object-log were the TSs and CRs encountered along the route. In an advantageous implementation foreseen, both the TS and the CR data originate from their respective dedicated ADAS subsystems. While in case of the TS data, the corresponding subsystem, i.e., the TSR ADAS subsystem, is quite common in recent production cars, the CR detection ADAS function is fairly uncommon at this point of time. It is expected though that in the coming years, the LiDAR sensors developed for automotive applications will pave the way for the spread of such an ADAS subsystem. A good insight in ADAS system architectures, various ADAS subsystems and func- tions, as well as the respective methods and computations involved is given in [30]. A survey on TSR methods and systems is given in [31], while in [32], a mapping and naviga- tion system developed for large-scale global positioning-denied sites is introduced. The system is capable of detecting CRs, intersections, and other road infrastructure. The static RO-based urban RET detection approach proposed in [25], and some further approaches make use of a variety of classiﬁcation and change detection (CD) methods known from the statistical inference literature. In the cited paper, it is presumed that the static ROs in general, and the considered TSs and the CRs, in particular, occur along the route according to an inhomogeneous discrete-variable binomial process. The minimum description length (MDL) methodology is then applied to detect and locate change in the character of the road environment sweeping past the ego-car. The lane-keep assist ADAS and the lane following autonomous driving (AD) subsys- tems, which perforce continually identify the current and neighboring lanes, and estimate their widths, as well as the TSR ADAS and AD subsystems, which locate, identify, and track the TSs encountered by the ego-vehicle, are of particular interest in the context of RET detection. First, such ADAS subsystems are already available on-board many production cars, second, the categorical and spatial distribution of TSs, as well as, the lane-widths and Appl. Sci. 2021, 11, 3666 5 of 17 the number of lanes—in the current cross-section of the road or in an aggregated form (e.g., average lane-width, average number of lanes)—carry information that can be useful in determining the RET of the given urban area. It should be emphasized that a timely feedback of the RET information to the above ADAS and AD subsystems could increase their effective processing speed and lower the rate of misclassiﬁcations via setting practical parameter constraints for the computations involved. Such constraints could be of geometrical nature and could take the forms of Boolean, probabilistic, and fuzzy regions-of-interest (ROIs), respectively, e.g., within image frames of video sequences [33]. While in case of point clouds, volumes-of-interest, again meant in a Boolean, in a probabilistic, and in a fuzzy way, respectively, could be marked and used [34]. As a further application of such reciprocal information, the characteristic size range of TSs—for a given RET—could be used for validating the detected TS candidates [35]. Similar processing beneﬁt could be gained from the above outlined information feedback in case of other presently not so wide-spread driver assistance functions, such as the CR detection. Furthermore, information on the current RET is also important for suggesting/choosing appropriate vehicle speed and acceleration/deceleration for the ego-car. An embedded testbed architecture for testing functions of self-driving cars was proposed in [36]. It could also facilitate the seamless integration of the static RO-based RET detection function into the intelligent vehicle control systems. In the following, it will be assumed that TS occurrences are reliably detected and logged by the on-board TSR ADAS subsystem, moreover, this log is passed on to the RO-based—in the following practically TS-based—RET detection system in real-time. It was our aim to choose, adapt, and validate a mathematically sound CD method that makes provision for and relies on a simple, but realistic stochastic model of the static RO placement and occurrences, in general, and of the TS placement and occurrences, in particular, for the purpose and in the context of detecting transitions between road environments of different character—or more concretely, between road environments of different RETs—in order to assist car drivers, human, and robotic drivers alike, in their driving tasks and activities. The continuous-time inhomogeneous marked Poisson process (IMPP) was identiﬁed as a possible stochastic model to work with. It should be noted, however, that in real life, the static RO placements—including those of TSs and trafﬁc lights—are governed by technical and administrative guidelines [37], from time to time they are subjects of potentially lengthy conciliatory procedures between locals and road administration. The ﬁnal decisions are therefore taken at different administrative levels. Some aspects of this occasionally complicated process are outlined in [38]. As in [25,26] also herein, the occurrences are considered along routes. These routes are assumed to be random, but they are, in fact, based on intelligent choices made by the drivers. Results gained via simulation implementing the IMPP model and making use of realistic data indicate that a TS-based RET CD is feasible and can be used for driver assistance, though it is not suitable for initiating an immediate intervention in critical situations. A more varied selection of static ROs—including, e.g., CRs, trafﬁc lights, and pedestrian crossings—would further improve the feasibility of the RET CD. Similar utility and feasibility are expected for the RET detection and identiﬁcation function computed with several RET change detectors and an artiﬁcial neural network (ANN) that merges and mushes together the detected RET transitions. 2. Materials and Methods 2.1. Car-Based Collection of Static Road Object Data from Various Urban Road Environments A series of car-based static RO data collection trips was carried out in Hungary in 2017. The data were collected from a number of urban areas. Data concerning a richer set—than presented here—of TSs and of some more characteristic ROs was gathered. The TSs and other ROs were recorded manually along the routes—together with the RETs of AppA l. p Sp cli.. S 2c 0i2 . 1 2,0 1 21 1,, x 1 1 F,O xR F O PEE R P REE RE RV R IE EW VI EW 6 of6 1o 8f 18 AppA l. p S pcli.. S 2c 0i2 . 1 2,0 1 21 1,, x 1 1 F , O x R F O PEE R P R EE RR EV RIE EW VIEW 6 o6 f 1 o8 f 18 2. Ma 2. Ma terita elrs i a aln s d a n Me d Me thotd hs o ds 2. Ma 2. Ma teri ta el rs ia a ln s d a n Me d Me tho td hs o ds 2.1.2 C .1a . rC -B ar a- sB ed a sC ed o lle Co clle tioc n t io of nS otf aS tic ta R tic o aR do O ad b je Ocb tje D ct a tD a a fr ta o m fr o V m a rV ioau rs io U us r b U an rb R an o aR do E ad n v E ir no vn ir m on en m ts e nts 2.12 . .C 1a . rC -B ar a-sB ed as C ed o lle Co clle tiocn tio on f S otf aS tic ta t R ic o a R do O adb je Oc btje D cta t D a afr ta o m fr oV m a rV io au rio s U usr b U ar nb a Rn o a R do E ad n v Eir no vn irm on en m te sn ts Appl. Sci. 2021, 11, 3666 6 of 17 A sA er i se es r io ef s c oa fr c -b aa r- sb ea ds e sd ta ti stca ti Rc O R d O a ta d ac ta o ll ce oclti le o cn ti o tr n i p tr s ip w sa s w c aa sr c ri aerd ri e od ut oiut n Hun in Hun garg ya i rn y in A s A er s ie es r io es f c oa f rc -b ar a-sb ea d s e sd ta s titc a ti RcO R d O a ta da c ta o lc le oc lti leo cn ti o tr ni p tr si p w sa w s a ca s rc ra ie rd ri e od ut out in Hun in Hun garg ya r in y in 2017. The data were collected from a number of urban areas. Data concerning a richer 2017. The data were collected from a number of urban areas. Data concerning a richer 201 27 0.1 T 7h . e T h de a ta da w tae w ree c re o lc le oc lte led cte fd ro f m ro a m n a u m nu bm erb o er f o ur f b ur an b a an r ea ar se . aD s.a D taa c ta o n cc oe n rc ne ir nn gi n ag r a ic h rie crh er set— seth t— ath n p an re p se re ns te en dte hd er h ee — re o— f T oS f sT a Sn sd a n od f so ofm so em m eo m reo crh ea c rh aa cr te ar cite sti ri cs R tic O R s O wsa w s g aa st g ha et rh ee dr.e T dh . e T he set— set th — ath n a pn re p sr ee nste en dte hd e rh ee — reo — f T oS f s T a Sn s d a n od f s o o fm so em m eo m reo c rh e a crh aa cr te ac rite sti rics ti Rc O R s O wsa w s g aa s tg ha etrh ee d r.e T dh . e T he TSsT a Sn s d a n od th o eth r R er O R s O wse r w ee rre ec o re rc do erd d e m da m nua anlua ly la ll y o a nlg o n th ge th ro eute route s—s to — gto eth ge eth r w eri th w ith the th Re E T Rs E o Tf s of TSs T a Sn s d a n od th o eth r R er O R s O w se w ree r re ec r oerc d oerd d e m da m nua anlua ly la ly lo a nlg o n th ge th ro eute route s—s to — gto eth ge eth r w eri th w ith the th Re E R Ts E T os f of the given areas—with the help of a dedicated tablet-based Android application, while the Appl. Sci. 2021, 11, x FOR PEER REVIEW theth giev g ei n v a e 6r n e o a f a r s 1e — 8a s w — ith w ith the th he el h pe o lp f a o d f a ed die cd ate icd ate ta d b ta let b-lb ea t- sb ea ds A ed n A dr n od id ro a id pp alp ic p alti ico an ti,o w nh , w ile h th ilee th e theth ge iv g ein v e an re a ar se — asw — ith w ith the th he e lh pe o lp f a o f d a e d die cd ate ica dte ta db ta leb t-lb ea t-sb ea d s e A dn A dr no d ir do a id p p alp ic pa lti ico an ti,o w n,h w ile h ith lee th e trajectory data of the trip was collected automatically by the app in every few seconds, and trajtr ec ato jer cy to d ry a ta d ao ta f th ofe th tre i p tr w ipa s w c ao s ll ce oclte led cte ad uto auto mam tica ati llcy a lb ly y b th ye th ap e p a p in p e in v ee rv ye f re yw f es w ec s oe n cd os n , ds, trajtr ec ato jec rto y r d y a ta da at o ta fthe th ofe times th tr ei p tr w iof pa w the s a co sTS lc le oc and lte lec dte other ad uto auto m RO am tic a entries a ti ll cy a lb ly y [ b 25 th y]. e th ae p p a p in p e in v e erv ye r fe yw fe s w ec s oen cd os n,d s, Appl. Sci. 2021, 11, x FOR PEER REVIEW 6 of 18 Appl. Sci. 2021, 11, x FOR PEER REVIEW 6 of 18 and a n ad t th at eth tie m ti em s o efs th ofe th Te S T an S d a n od th o eth r R eO r R eO ntr ei n etr s i[e 2s5 [ ]2 . 5]. and a n ad t th at eth tie m ti em s o ef s th o The fe th Te data S T aS n d a collection n od th o eth r R er O R personnel e O n tr en ie tr s i[e2 s5 [] consisted 2 . 5]. of two persons: a driver and a data entry Appl. Sci. 2021, 11, x FOR PEER REVIEW 6 of 18 The T d he a td a a cto al lc eo cltlie ocn ti o pn er p so en rsn oe n l n ce oln c so is n ts eid st e od f to w f o tw po er p so en rs so : n as d : a ri v de rr iv a en r d a n a d d a a td a a etn at e ry n ta rs ys ia ss -sis- assistant. The manual data entry was made easy by the array-like screen design with TSs The T h de a td aa c to al l ce o cltlie o cn ti o pn e rp so er n sn oe nln ce o ln cs o is nts eid st e od f to w f o tw po e rp so er n ss o : n as :d a ri d ve ri rv a en r d a n ad d a a td aa e tn a te rn y ta ry ss a is s-sis- tant ta . n Tt h . e T m hea m nu aa n lu d aa l td a a etn at r ey n tw ry a s w m asa m dea e da es e ya b sy y t b h ye ta hr er a ay rr -l aiy k-el is kce r e se cn re d en es d ig en si g w n i tw h iT th S sT a Sn sd a n Rd O R O and RO symbols. In case of parametrized TSs, e.g., speed limits, the standard options (i.e., tantta . n Tth . e T h m ea m nu aa nlu d aa l td aa e tn a te rn y tw rya w s m asa m de a d ea es e ya b sy y b th ye t h ar er a ay rr -a liy k-e li s kc er e sc er n e e dn e sd ig en si g w ni tw h iT th S s T a Sn s d a n R d O R O 2. Materials and Methods sym sy bm olb s.o Iln 10 s. c Ia km/h, ns e ca o sfe p oa 20 fr p akm/h, a m ra et m rie zte r .d i.z .T ed ,S70 s T , S ekm/h) .sg , .e , .s g p .,e s e wer p de le id m e l of iitm sfer , itth s ed—also e , t s h te a n st d aa n rd in da o r pictorial d p to io pn tiso (n i.s form—after e .(,i .1 e0 ., k 1m 0 k /h m , 2 /h the 0, 20 general sym sy bm olb s.o Iln s. c Ia ns c ea o sf e p oa f rp aa m ra em trie zte rd iz e Td S s T , S es .g , .e , .s gp .,e se p d e e lid m liitm s, itth s,e t s hte a n std aa nrd da o rd p to io pn tis o ( n i.se ( .,i .1 e0 ., k 10 m k /h m , / 2 h 0, 20 2.1. Car-Based Collection of Static Road2 O . Ma bjec tt e r D ia al ts a a frn od m Me Vartio hu os d U s rban Road Environments 2. Materials and Methods kmk /h m , /… h,, … 7 TS 0, k 7 type. 0 m k /h m ) Speciﬁc /w h)e r w e eo rf efsymbols, e orfe fd er — ed a— lsi.e., o a lisn o touch-scr p in ic t p oir cita olr een i fa olr m fo keys, — rm a— fte wer a rf tte h e re of t g hf e eer n g e ed e rn al e for r T aS l entering T ty Sp te y . p Se p . e the S cp ife iconsider c c ific ed kmk /h m , /… h, , … 70 , 7 k0 m k /h m ) /w h)e w ree o re ff e orfe fe dr— ed a— lso a li sn o p in ic p to ic rita olr ifa olr m for — m a— fte arf tte h re t h ge e n ge ern ae l rT aS l T ty Sp te y.p S ep . e Scp ife ic ci fic 2. Materials and Methods A series of car-based static RO2 d .1 a.ta C acro -B lla es ce ti do C no tr lle ic pts io w n a of s S cta artr ic ie R do o ad ut Oib nje Hun ct Da gta ar fr y oim n Various Urban Road Environments 2.1. Car-Based Collection of Static Road Object Data from Various Urban Road Environments RETs, the repeated entries, for the cancelations of the last entry, and for entering verbal sym sy bm olb s,o il.s e,. ,i .te o .,u tco h u -s cc hr-e se cn re k en ey k se , y w se , r w e e o rf ef e orfe fd er f eo dr fe on rt e en ritn eg ri n th ge tc ho en c so id ne sr id ed er e Rd E T Rs E , T th s,e tr he ep r ee ap te ed at ed sym sy bm olb s,o il.s e,. ,i .t eo .,u tc o h u -c sh cr -e sc er n e e kn e y k se , y w s,e w ree o re ff e orfe fe dr e fo dr fe on r te en ritn eg ri n th ge t h co en cs o id ns eird ee dr e R d E T Rs E , T th s,e t h re ep re ea p te ed at ed 2017. T 2h .1e . C da ar ta -B w ase ed re C c oo lle llc etc io te nd o f fr S o tm atic a R nou am d O be br je o ctf D ur ab taa fr n oa m re V aa sr . io D u asta U c rb oa n n c e R ro n aid n g E n av ir rio cn hm ere nts A series of car-based static RO data en c tr o il els e,c comments. fti oo r n th tr e icp as n c w The ela ast ic o location, a nrsr io efd t h o eut time, las in t eHun n TS/RO, try,g a an rd yand fio nr e RET ntericategorical ng verbal co data mme wer ntse . T stor he led oca in tio antext-ﬁle , entries, for the cancelations of the last entry, and for entering verbal comments. The location, ente ri n ets r,i e fo s,r fto h re t h ca en cca en la ct eilo an tis o o nf s to hfe t h lae s tl a es n t te rn y,t r a y n , d a n fo dr fe on r te en ritn eg ri n vg e rv be ar lb ca o lm co m m em nte sn . T tsh . e T h lo ec a lo tico an ti,o n, A series of car-based static RO data collection trips was carried out in Hungary in set—than presented here—of TSs and of some more characteristic ROs was gathered. The 2017. The data were collected from a number of urban areas. Data concerning a richer A series of car-based static RO data collection trips was carrite im d te i o ,m ut Te S ,i /T n RS O in Hun /R , a a O ncomma ,g d a a n R rd y E T R in E c a T separated t e cg ao te rg ic o arli c d a values a lt d a a w ta e r w (cvs) e e srte o r se format. to dr i en d a i n te a After x t t- efx ilte -f i the in le a i n trips, c o am co m m the am se ap csv s ae rp aﬁles a te rd at v ewer d al u ve ae s lu stor es ed as 2017. The data were collected from a number of urban areas. Data concerning a richer timte im , T eS , /T RS O /R , a O n , d a n R d E T R E cT at e cg at o erg ic oa rli c d aa l td aa w tae w ree s rte o r se to dr e in d a in t e a x t t- efx ilte -f i il n e a in c o am co m m am se ap sa erp aa te ra dt e vd a lv ua elsu es TSs and other ROs were recorded manually along the routes—together with the RETs of set—than presented here—of TSs and of some m spr oreadsheets e character and isticwer ROe s converted was gatherto edvarious . The formats (e.g., kml) for post-processing and 2017. The data were collected from a number of urban areas. Data (cv c s (o c ) v n fo s c)r e m f ro n a rim tn . g A a tf a . t A errif c tte h hre e t rt h re ip ts r,i p th s,e tc hs ev c fsiv le f si lw es e r w e esrte o r se to dr e ad s s a p sr s ep ad re sa h d es eh ts e e atn sd a n w d e r w e ecro en cv oen rv te ed rt e to d to (cv( sc )v fs o)r m for am t. a A t.f tA erf tte h re t h tre ip ts ri , p th s,e t h cs ev c f siv le f si lw ese w ree s rte o r se to dr e ad s s ap s rs ep ar d esa h d es eh ts e e atn s d a n w de w ree c ro en cv oe n rv te ed rt e to d to set—than presented here—of TSs and of some more characteristic ROs was gathered. The the given areas—with the help of a dedicated tablet-based Android application, while the TSs and other ROs were recorded manually alo visualization. ng the routes—together with the RETs of set—than presented here—of TSs and of some more characteristic R vO ar sv i o w au ra iso s fu o gs r a m fto ha re m tr se a (d e ts ..g T (.e ,h .k g em . , k l)m fo l) r fp oo r sp t-o p sr to -p ce ro ss cie n sg si n an gd a n vd is u va is lu iza altiiz o an ti . on. varv io au ris o f u osr m for am ts a (t es .g (.e , .k gm ., k l)m fo l)r fp oo r s p t- o p sr t- o p cr eo sc se in sg si n ag n d a n vd is u vi aslu iz aa lt iz io an ti.o n. TSs and other ROs were recorded manually along the routes—together with the RETs of trajectory data of the trip was collected automatically by the app in every few seconds, From the above data collection, the relevant TS rates—along routes in the considered TSs and other ROs were recorth de ed g m iva en nua arle la ys a — lo w ni g th th th e e r o hute elps o — f to a d ge ed th ic ea r te w d ith ta b th le et -R bE aT se sd o A f ndroid application, while the FroF m ro th me th ab eo av be o d va e td a a cto al l ce oclti le o cn ti,o th n,e th re el e rv el ae n v t aT nS t T ra Ste rs a— tes a— loa nlg o n ro gute route s ins th ine th co en cso id ns eird ee dr ed FroF m ro th me th ab e o av be o v de a td aa c to al c le oc lti leo cn ti,o th n,e th re el e re vla en vt aT nS t T ra Ste ra s— tesa — loa nlg o n ro gute route s in s th ine th co en cs oin ds eird ee dr ed the given areas—with the help of a dedicated tablet-based Android application, while the and at the times of the TS and other RO entries [25]. the given areas—with the help tr o afj e ac d to erd yi c d aa te ta d o ta f b th lee t -tr ba ip se w d a A sn c d orlo le id cte ad p p aluto icati m oa nti , c w ah lliy l eb th y e th e app in every few seconds, RET R Ea T r ea ar s RET e — as w — ar er w eas—wer ee r ken o kw no ne w . known. n T.h e T h ee m The e pm iri p empirical cia rli cr aa l te rs a te rates fo s r fo tfor h r et h the e (“No (“No (“No sto stopping” p sto pip np gi” n g ((NS)), ” NS (NS )), )), ,, (“Parking , RER T Ea Tr ea ar se — asw — ew ree rk en o kw no nw . n T.h e T heem e pm irp icia rl i ca ra l te ra s te fs o rf o tr h et he (“ No (“No sto s pto pip np gi” n g(” NS (NS )), )), , , trajectory data of the trip was collected automatically by the app in every few seconds, The data collection personnel consisted of two persons: a driver and a data entry assis- and at the times of the TS and other RO entries [25]. trajectory data of the trip was collected automatically by the app (i“ nP ( e a “ v rP k ea r in y rk g f i e n lw o gt ” lso e(tc P ” o L n ()P d ),L s ,) ), (“G (i“ v G e iv w ea w y”a y (G ” W) (GW) ), a)n , d a nd (“M (“ aM x a sx p esep de e 3d 0 3 k0 m k /h m ”/ h (S ”L ()S ) LT )) S sT Ss and at the times of the TS and other RO entries [25]. (“P(a “rP ka ir nk gi n lg o t” lo lot” ( t” P L ((PL)), ) P )L , )), (“ G (“Give (“ iv G ei v w eway” aw y” a y (” G (GW)), W) (GW) ), a) and n , d a nd (“Max (“ M (“ aM xspeed a sx p e se p d e 30 e3 d km/h” 0 3 k0 m k /h m ” (SL)) / h (” S L(TSs )S )L T )) wer S s T Se s used herein tant. The manual data entry was made easy by the array-like screen design with TSs and RO The data collection personnel consisted of two persons: a driver and a data entry assis- and at the times of the TS and other RO entries [25]. wer w ee us ree us d e hd e rh ee in re a in s a a s p a ri o pr rii o er siti e m sti am tes a te of s th ofe th re ef e rr ee fe nr ce en r ca ete rs a te fo sr fth ore th m ea m rka er d k e Pd o iP ss oo is ns o pn ro p -ro- as a priori estimates of the reference rates for the marked Poisson processes. These rates The data collection personnel consisted of two persons: a driver and a data entry assis- wew ree us re e us d e hd e rh ee in re a in s a a sp a r ip or ri io ersiti em sti am tea s te os f th ofe th re ef e re re fe nrc ee n r ca ete ra s te fo sr f th ore th m ea m rka erd k e P d o iP so so is n s o pn r o p -ro- symbols. In case of parametrized TSs, e.g., speed limits, the standard options (i.e., 10 km/h, 20 tant. The manual data entry was made easy by the array-like screen design with TSs and RO The data collection personnel consisted of two persons: a driverc a en sc d se e a s s .s d e Ta sh .t a e T se h en e r t sa re y te r a s as te a sr is s e -a g ri ev g ei n v i en n T in a b T la e b 1l e f o 1r fr o o rute route s wsi th wiin th D in t D an t d a n w di th wiin th R in e s R a erse a ar se . as. cesc se es ss . e T sh . e Tsh ee r sa ar ete r ea sgiven te ar se a g re iin v g ein T vable e in n T in a 1 b T for la eb 1 l r e outes f o 1r f r oo rwithin ute route s w si Dt th wiin th and D int within D an t d a n w di Res th wiin th ar R in eas. e s R a er se a ar se . as. tant. The manual data entry was made easy by the array-like screen design with TSs and RO km/h, …, 70 km/h) were offered—also in pictorial form—after the general TS type. Specific tant. The manual data entry wassy m ma bd oe ls e . a In sy c a bsy e to h fe p a arrra am y-e li tk rie z e sc dr e Te Sn s ,d ee .g si .,g s n p w eeid th l iT m Sist s a , n th de R sO ta ndard options (i.e., 10 km/h, 20 symbols. In case of parametrized TSs, e.g., speed limits, the standard options (i.e., 10 km/h, 20 symbols, i.e., touch-screen keys, were offered for entering the considered RETs, the repeated Tab Ta le b 1le . T 1 h.e Tem he p em irip ca irli c tr aa l ffi trc a ffi sig c n s i( g TS n ( ) TS ra) tes ra t aes lo n ag lo a n g ra a n r da o n m d o rm ou r te ou fo te r f to hr e th ho e m ho og m en og eo en ueo s m us a r m ked ark ed symbols. In case of parametrizek dm T/S hs,, … e.g , .7 , 0 sp k em ed /h li)m w ites T r,e able t h oe f fs e 1. tr a e The n d d — Ta aempirical ra d b lTa s le o op b 1 it le n .i o T p 1 n h .traf is e c T t ( em h o iﬁc .e r ei.p em a ,si il1 r gn 0 if p c o i a k r r (TS) lm i m c tr a / — a lh ffi rates t,a r 2 a f c0 t ffi e s ir along c g t n sh i g (e TS n g a ()e TS random r n ae )t r es ra al ta es T lo S r n a oute l g to y n a p g r e f a or .a n S r d the p ao n e m d c homogeneous io f ri m o cu rto e u ft oe r fto hre tmarked h ho e m ho om gen o Poisson g eo en ueo s m upr sa m r ocesses ka ed rk eddescribing km/h, …, 70 km/h) were offered—also in pictorial form—after the general TS type. Specific entries, for the cancelations of the last entry, and for entering verbal comments.P T oh ise s o lo nc p ar to io cn es , ses describing (downtown) Dt and residential (Res) urban road environments. Poisson processes describing (downtown) Dt and residential (Res) urban road environments. symbols, i.e., touch-scre (downtown) en keys, w Pe o Dt r ie P ss o and o o if s n fs e p o r rn r esidential e o d p c es rfo o s cres es e s n d es t es (Res) e r d cir es n ib g ci urban rn tih b ge i n ( d c go o r(w n oad ds n oiw t d o envir e n w rte n od w ) onments. Dt R n) E a Dt T nsd ,a t n rh es de i rd r es e en p id e ten ia at le t ( id R al es (R ) es ur)b u ar n b r ao na r d o en adv en iro vn ir m on en m ts en . ts. km/h, …, 70 km/h) were offered—also in pictorial form—after the general TS type. Specific symbols, i.e., touch-screen keys, were offered for entering the considered RETs, the repeated time, TS/RO, and RET categorical data were stored in a text-file in a comma separated values entries, for the cancelations of the last entry, and for entering verbal comments. The location, symbols, i.e., touch-screen keys, were offered for entering the considered RETs, the repeated entries, for the cancelations of the last entry, and for entering verbal comments. The location, Exp Ee xcp te ed ct e nd u m nu bm erb e E r x p Ee xcp te ed ct e nd u m nu bm erb e N ra N tua rt au l rla olg la o -ga Ch - Ch ar- ar- Exp Ee xc p te ec dt e nd u m nu b m erb e r E x p Ee xc p te ec dt e nd u m nu b m erb e N r aN tua rtau l rla olg la o-gaCh - Ch ar-ar- Expected Number Expected Number of Natural (cvs) format. After the trips, the csv files were stored as spreadsheets and were converted to TS TSA bA be bv b ie - vi- time, TS/RO, and RET categorical da T ta S T wS e A reb s A b to e b rv b eid e - v in i- a text-file in a comma separated values entries, for the cancelations of the last entry, and for entering verbal comments. The location, Characteristic time, TS/RO, and RET categorical data were stored in a text-file in a comma separated values TS Type Abbeviation Index Ind Ie nx of d e Occurrences x of o oc fc o uc rc ru en rrce en sc e Occurrences s o f o oc fc o uc rc ru eper n rrce en km sc p ee s rp e rr iLogarithm t h rm ith o m f to h fe of t ha ec te ar cits e-ris- Ind Ie nx d ex of o oc f co u crcru er n rc ee n sc e s o f o oc f co u crcru er n rc ee n sc p es e r p e rr i th ri m th o m f t o h fe t ha ec ta ec ri ts e- ris- various formats (e.g., kml) for post-processing and visualization. to (cvs) format. After the trips, the csv fileT s y w p T ee r y e p s et a ot rieo a d n t ia o sn s preadsheets and were converted to time, TS/RO, and RET categorical data were stored in a text-file in T ay cp T oe m y p m e a a t sie o a p n ta i o ra nt ed values (cvs) format. After the trips, the csv files were stored as spreadsheets and were converted to per km in Dt in Res Areas the Rate-Ratio perp k em r k im n Dt in Dt kmk im n R in e s R a er se a ar se as rate ra -rta et-iro a tio tic t ti o c to per p k er m k im n Dt in Dt km k im n R in e s R a er se a arse as rate ra -r ta et -i ro a tio tic ttio c to From the above data collection, the relevant TS rates—along routes in the considered (cvs) format. After the trips, the v a cr siv o u fis le fs o r w m ea re ts s (te o.r ge .,d k a m s ls ) p fo re ra p d o ss h te -p er ts o c ae n sd si n w g e r ae n d co v niv su er atle iz da t to io n. various formats (e.g., kml) for post-processing and visualization. NS 1 2.00 0.35 1.74 Dt NS 1 2.00 0.35 1.74 Dt NS NSNS 1 1 1 2.00 2.02 0. 00 0.30.35 0 5. 35 1.71 4. 74 1.74 Dt Dt Dt RET areas—were known. The empirical rates for the (“No stopping” (NS)), , From the above data collection, the relevant TS rates—along routes in the considered various formats (e.g., kml) for post-processing and visualization. From the above data collection, the relevant TS rates—along routes in the considered PL PL 2 PL 2 2 1.70 1.70 1 .70 0.25 0.25 0 .25 1.1.92 92 1 .92 Dt Dt Dt PL PL 2 2 1.71 0. 70 0.20 5. 25 1.91 2. 92 Dt Dt (“Parking lot” (PL)), (“Give way” (GW)), and (“Max speed 30 km/h” (SL)) TSs RET areas—were known. The empirical rates for the (“No stopping” (NS)), , From the above data collection, the relevant TS rates—along routes in the considered RET areas—were known. The empirical rates for the (“No stopping” (NS)), , GW 3 0.70 0.80 −0.13 Res GW 3 0.70 0.80 −0.13 Res were used herein as a priori estimates of the reference rates for the marke d PoG isW sG o n W p ro-3 3 0.70 0. 70 0.80 0. 80 −0.− 10 3. 13 Res R es GW 3 0.70 0.80 0.13 Res (“Parking lot” (PL)), (“Give way” (GW)), and (“Max speed 30 km/h” (SL)) TSs RET areas—were known. The empirical rates for the (“No stopping” (NS)), , (“Parking lot” (PL)), (“Give way” (GW)), and (“Max speed 30 km/h” (SL)) TSs SL SL 4 4 0.20 0 .20 0.40 0 .40 −0.6 −9 0 .69 ResR es cesses. These rates are given in Table 1 for routes within Dt and within Re s a reaS sL . SL 4 4 0.20 0. 20 0.40 0. 40 −0.− 60 9. 69 Res R es SL 4 0.20 0.40 0.69 Res were used herein as a priori estimates of the reference rates for the marked Poisson pro- (“Parking lot” (PL)), (“Give way” (GW)), and (“Max speed 30 km/h” (SL)) TSs were used herein as a priori estimates of the reference rates for the marked Poisson pro- AnA y ny 4.60 4 .60 1.80 1 .80 Dt Dt AnA y ny 4.64 0. 60 1.81 0. 80 Dt Dt were used herein as a priori ec se tism sea ste . T s h oe fs th e r ea r te es fe a rr ee n g ce iv re an te is n fT oa r b th lee 1 m fo ar r k re od ute Po s iw ssio th ni n p r D ot - and within Res areas. cesses. These rates are given in Table 1 for routes within Dt and within Res areas. Any 4.60 1.80 Dt Table 1. The empirical traffic sign (TS) rates along a random route for the homogeneous marked cesses. These rates are given in Table 1 for routes within Dt and within Res areas. Poisson processes describing (downtown) Dt and residential (Res) urban road environments. Table 1. The empirical traffic sign (TS) ra2 tes .2 .2 a M .l2 o. n a M g th a e a m t rh aa e n tm d ico a am tl ic M r ao l ou M dte e o ls f d o a erls n td h a e M nd he o t M m ho e o d tg h sen o deo s us marked 2.22 . .M 2.a M tha em tha et m icaatl ic M al oM deo ls d e als n da n M de M tho et dh so ds Table 1. The empirical traffic sign (TS) rates along a random route for the homogeneous marked Poisson processes describing (downtown) Dt and residential (Res) urban road environments. 2.2. Mathematical Models and Methods Table 1. The empirical traffic sign (TS) rates along a random route for the homogeneous marked Poisson processes describing (downtown) Dt and residential (Res) urban road environments. Expected number Expected number Natural loga A - s A Ch it sw a it a r- s w m ase m nti eo nn tie o d n i en d th ine th In etr In od truc od ti uc on ti,o th n,e th co en cti on nuo tinu uo s-u tis m -ti em IM e P IM P P sto P c sh to ac sh tia cs m tico m de old el As A it sw it aw s m ase m nti eo nn tie od n e in d th ine th In etr In otr duc odti uc on ti,o th n,e th co en cti on nuo tinu uo s-u tis m -ti em IM e IM PP P sP to s cto ha csh ti acs ti m co m de old el TS Abbevi- Poisson processes describing (downtown) Dt and residential (Res) urban road environments. As it was mentioned in the Introduction, the continuous-time IMPP stochastic model Index of occurrences of occurrences per rithm o hfa t d h h a b ed ee a b n ce te c eh n ri o s cs - h eo ns e fo nr fc oh ra c rh aa cr te ar cite zirn iz gi n th ge th aleo a nlg o-n th ge -th -ro eute -route pla p ce la m ce em nt ea nn t d a n od cc o ur cc rur enr ce en o cf e T oS f sT Ss had h a bd e e b n e e cn h o cs h eo ns e fo nr f c oh r a crh aa cr te ac rite zirn iz gi n th ge th ae lo a nlg o-n th ge -th -ro e- ute route pla p ce la m ce em nt ea nn t d a n od c co ur ccrur enrc ee n o ce f T oS f s T Ss Expected number Expected number Natural loga- Char- Expected number Expected number Natural loga- Char- Type ation TS Abbevi- had been chosen for characterizing the along-the-route placement and occurrence of TSs TS Abbevi- join jo tliy n tl fo yr fth ore th pe ur p p ur os p eo o se f R oE f T R E C T D C iD n th ine th pe re p se re ns t es n tud t stud y. F yo . rF a o rp a ro p fo ro un fod un tr d e a tr tm eae tm nt eo nft th ofe th e per km in Dt km in Res areas rja otie n jo -tl ria y nt tl ifo o y r f o th re tth i c p e ur t o p p ur os p eo o se f R of E R T E C TD C iD n th ine th pe r ep sr ee nst es ntud t stud y. F yo . r F o a rp a r o p fr oo un fod un tr de a tr tm eae tm nt eo nf t th ofe th e Expected number EI xn pd ee ct xe d nu of m o b ce cr u rN re an tu ce ra s l loo gf a -ocCh cura rr e-nces per rithm of the acteris- Index of occurrences of occurrences per rithm of the acteris- TS Abbevi- Type ation jointly for the purpose of RET CD in the present study. For a profound treatment of the Type ation mam tha et m he am tica ati l c th ale th ore yo o ry f P oo f iP ss oo is ns o pn ro p cr eo sc se es s,s e ss e,e s [e 3e9 [ ]3 . 9]. mam tha etm he am tic aa ti l cth ale th ore y o o ry f P oo f iP so so is n s o pn r o p creo sc se es ss , e ss e,e s [ e3 e9 []3 . 9]. NS 1 2.00 0.35 1.74 Dt Index of occurrences of occurrencp es e rp k em r riin th Dt m of thk e m a c in te R rie s- s areas rate-ratio tic to per km in Dt km in Res areas rate-ratio tic to Type ation mathematical theory of Poisson processes, see [39]. In th Ine th ch eo csh eo ns s etn o c sh to ac sh tia cs a tip cp ar p op ar co ha , c th h,e th TS e T dS ata d a lo ta g l so a grse a sre ee s ne a es n ra esa r lie za alt ii zo an tiso o nfs a o n f I aM n P IM P.P P. In th Ine th ch eo cs h eo ns e st n o s ct h oa csh ti acs t aip c p ar p o p arco ha ,c th h,e th Te S T dS a ta da lta og ls o a gr se a s re ee s n e e an s r ae sa r li ez aa li tz io an tis o o nf s a on f a IM n I P M P.P P. PL 2 1.70 0.25 1.92 Dt per km in Dt NS km1 i n Res areas2 .00 rate-ratio tic0 t.o 3 5 1.74 Dt NS 1 2.00 0.35 1.74 Dt In the chosen stochastic approach, the TS data logs are seen as realizations of an IMPP. The T h Ce D C m De m the oth d o us de us d e cd om co m m om nlo yn iln y c io nn cjo un njc un tio cn ti o w ni tw h iP th o iP ss oo is ns o pn ro p cr eo ss ce es s sie ss th ise th cum e cum ulau tilv ae t ive The T h CeD C m De m the oth d o us d eus d e cd o m co m m om nlo yn iln y c in o n cjo un njc un tio cn ti o w ni tw h iP th o iP ss oo is n s o pn r o p cr eo sc se es ss iess th ise th cum e cum ulau tilv ae ti ve GW 3 0.70 0.80 −0.13 Res NS 1 2 .00 PL 2 0.35 1.70 1.74 D 0t. 25 1.92 Dt PL 2 1.70 0.25 1.92 Dt The CD method used commonly in conjunction with Poisson processes is the cumulative sum sum (C U (C SU UM SU ) M m)e m the oth d.o D de . tD aielt ea d il e ed xp e o xsp it o io sin tiso o nf s s o u fc s h u m che m the oth dso c da sn c a bn e f bo eu fn od u n in d [i4 n0 ,[4 41 0],.4 1 B]y . By sum sum (C U (C SU US M U )M m )e m the oth d.o D d.e t D ae ilte ad il e ed x p eo xs p io tis oin tis o n os f s ou f csh u c m he m the oth ds o d ca sn c a bn e b fo eu fn od u n in d [i4 n0 [ ,4 41 0] ,4 . 1 B]y . By SL 4 0.20 0.40 −0.69 Res PL 2 1. 70 GW 3 0.25 0.70 1.92 D 0t. 80 −0.13 Res GW 3 0.70 0.80 −0.13 Res sum (CUSUM) method. Detailed expositions of such methods can be found in [40,41]. By assum assum ingi n th ge th va el i v d ailtiy d io tf y t o hfe t I h M e P IM P P m Po m de o ld — ea l— t le aa t s le t a w st i th w irth es p re es cp t e to ct th toe th co en cso id ne si rd ed er R ed E T Rs E a Tn sd a nd assa um ssum ingi n th ge th ve a lv id ailt iy d io ty f to hfe t h IM e I P M P P m Po m de old — el a— t la et a l se t a w sti th w ir th es r pe esc p t e tc ot th toe th co en cs o in ds eird ee dr e R d E R Ts E T an s d a nd Any 4.60 1.80 Dt GW 3 0. 70 SL 4 0.80 0.20 −0.13 Re 0s .4 0 −0.69 Res assuming the validity of the IMPP model—at least with respect to the considered RETs SL 4 0.20 0.40 −0.69 Res concso id ne si rd ed er e Td S sT — Ss fo — r fd or e sd cr eisb cirn ib gi n ag n d a n cd h ac rh aa ct re ar citz eir n iz gi n Tg S T pS la p ce la m ce em nts e na ts n d a n od c co uc rc ru en rr ce en s ca elso n alg o ng con cs o in ds eird ee dr e T d S s T — Ss f— orf o dr e sd ce ri sb cirn ib gi n ag n d a n cd h a crh aa cr te ar citz eirn iz gi n T g S T pS la p ce la m ce em nts en a ts n d a n o d c co ucrc ru er nrceen sc a es lo a nlg o ng SL 4 0.20 0.40 −0.69 Res Any 4.60 1.80 Dt and considered TSs—for describing and characterizing TS placements and occurrences Any 4.60 1.80 Dt rourto eu s tw esi tw hiin th a in n d a n bd etw bee tw ene e un rb u ar n b a en n v ein ro vn irm on em nte s,n o ts ur , o ta urs k ta n sk a rn ro aw rro ed w e dd o w do nw to n a td o a ap d t aa p t su ai t s-uit- rou rto eu s tw esi tw hiin th a in n d a n bd e tw bee tw en e e un r b u arn b a en n v ein ro vn irm on em nte sn , t osur , o ur tas k ta s nk a rn ra or w ro ew d e d d o w do nw to n a to d a ap dt aa p t su a is t- uit- 2.2. Mathematical Models and Methods Any 4.60 1.80 along D routes t within and between urban environments, our task narrowed down to adapt a able a b C le U C SU UM SU m Me m the oth d o fo dr fto hre t p hu e rp p u orsp eo , s ae n,d a n vd al i v d aa li td e a it t ew it i tw h irte ha r lie sa ti lc is T tiS c T dS a ta d.a ta. abla eb C le U C SU US M U m Me m the oth d o fo dr f to h re t h pe u r p p u o rs p eo , s ae n , d a n vd a lv id aa litd ea it te w iti tw h ir te ha r li esa tliics t T ic S T dS a ta da . ta. As it was mentioned in the Intro 2d .2 uc . M tio at n h,e th me at c ic oan l ti M no uo deu lss a -ti nm d M e IM ethP od Ps stochastic m suitable odel CUSUM method for the purpose, and validate it with realistic TS data. 2.2. Mathematical Models and Methods It w It aw s o aur s o iur nte in nte tio nn ti o to n a to d o ap dt oa pn t d a n vd a lv id aa lite d aa te c o an cti on nuo tinus uo -us tim -ti em va e rv ia an rit ao nft th ofe th Ce U C SU UM SU M It w It aw s o as ur o ur inte in nte tio nn ti o to n a to d o ap dt oa pn t d a n vd a lv id aa lite da a te c o a n cti on nuo tinus uo-us tim -ti em ve a rv ia an rit ao nft th ofe th Ce U C SU US M U M It was our intention to adopt and validate a continuous-time variant of the CUSUM had be 2e.n 2. c M ho at sh ee n m fa otr ic c ah l a M ra oc dte els r ia zn in dg M th ete h o ad lo s ng-the-route placement and occurrence of TSs As it was mentioned in the Introduction, the continuous-time IMPP stochastic model mem the oth d o fo dr fC orD C . D Fur . F tur het rh m eo rm reo , rtr e,a d tre a-d oe ff- o if sf sio s ug soh ug t b he t tb w ee te w ne e th ne t h fa el sfe a la se la r am la rr m ate r aa te n d a n th de t he mem the oth d o fd o rf o CrD C . D Fur . F tur hetrh m er om reo , rtr e,a d tre a-d oe ff - oifsf s io s ug soh ug t h bt etb w ee tw en e e tn h et h fa el s fe a ls ae la a rm lar m ra te ra a te n d a n td h et he As it was mentioned in the Introduction, the continuous-time IMPP stochastic model method for CD. Furthermore, trade-off is sought between the false alarm rate and the jointly for the purpose of RET CD in the present study. For a profound treatment of the As it was mentioned in thh ea In d tr be oe dn uc cti ho os ne , n th fe o r c o cn hti ar na uo cte ursi- zti in m ge th IM e P a eP lx o p s ne to g c-c te th hd a e s -d r tio e cte ute m cti o p d oln e al c le am g e an ss t oa cn ia dte od c cw ur ith re n tc he e o TfS T sS esq uences and to provide hints for choos- expected detection lag associated with the TS sequences and to provide hints for choos- exp ee xc p te ec dte d d e te de cte tio cn ti o la ng l a ag ss a os cs io ate cia dte w di th w it th he t h Te S T sS eq sue eqn ue ce n sc a en s d a n to d p to r o p v rio dv ei d he in h ts in fto sr f o ch r o co hs o -os- had been chosen for characterizing the along-the-route placement and occurrence of TSs expected detection lag associated with the TS sequences and to provide hints for choosing mathematical theory of Poisson processes, see [39]. jointly for the purpose of RET CD in the present study. For a profound treatment of the had been chosen for characterizing the along-the-route placement ia n n g d i n a o p gc p a cr ur p op p rre ro in a p c te r ei a o th te f r T e th S sh r se oslh do sl f d osr fth ore th ch ea c n h g ae n d ge e te dc eto ter cs to . rs. jointly for the purpose of RET CD in the present study. For a profound treatment of the ingi n ag p p ar p o p p rr oip ate ria th ter th esr heo slh do sl d fo sr f th ore th ch e a cn hg ae n g de e te de cte toc rto s. rs. appropriate thresholds for the change detectors. In the chosen stochastic approach, the TS data logs are seen as realizations of an IMPP. mathematical theory of Poisson processes, see [39]. jointly for the purpose of RET CD in the present study. For a profound treatment of the mathematical theory of Poisson processes, see [39]. The continuous-variable variant of the CUSUM method for CD in IMPP realizations The CD method used commonly in conjunction with Poisson processes is the cumulative In the chosen stochastic approach, the TS data logs are seen as realizations of an IMPP. mathematical theory of Poisson processes, see [39]. In the chosen stochastic approach, the TS data logs are seen as realizations of an IMPP. is derived in this subsection. The working of the RET change detectors implementing this sum (CUSUM) method. Detailed expositions of such methods can be found in [40,41]. By In the chosen stochastic ap Tp hreo a CcD h , m the eth To Sd d us ata e d lo g co s m arm e o se ne ln y a in s r ce o a n lijun zatc io tin os n o w f a itn h IP M oP isP so . n processes is the cumulative The CD method used commonly in conjunction with Poisson processes is the cumulative method is demonstrated in Sections 3.1 and 3.2. The change detectors presented therein assuming the validity of the IMPP model—at least with respect to the considered RETs and The CD method used common sum ly i n ( C cU on SjU un M ct )i o m ne th wio td h . P D oe is ta so iln ed p r eo xc pe o ss sie ti so in ss th oe f s cu um ch u m lae tth ivo e ds can be found in [40,41]. By sum (CUSUM) method. Detailed expositions of such methods can be found in [40,41]. By have been tuned to detect two different RET transitions, namely, Dt to Res and Res to Dt considered TSs—for describing and characterizing TS placements and occurrences along assuming the validity of the IMPP model—at least with respect to the considered RETs and sum (CUSUM) method. Detailed expositions of such methods can be found in [40,41]. By assuming the validity of the IMPP model—at least with respect to the considered RETs and transitions, and are tested with synthetic input sequences. routes within and between urban environments, our task narrowed down to adapt a suit- considered TSs—for describing and characterizing TS placements and occurrences along assuming the validity of the IMPP model—at least with respect to the considered RETs and considered TSs—for describing and characterizing TS placements and occurrences along able CUSUM method for the purpose, and validate it with realistic TS data. routes within and between urban environments, our task narrowed down to adapt a suit- considered TSs—for describing and characterizing TS placements and occurrences along routes within and between urban environments, our task narrowed down to adapt a suit- It was our intention to adopt and validate a continuous-time variant of the CUSUM routes within and between urb aa bn le e C nU viS rU on M m m ene tth s, o od ur f o ta r s tk h e n p ar urro p w os ee d, d an od w v na tlo id a ad te a p itt w a is th u ir te - alistic TS data. able CUSUM method for the purpose, and validate it with realistic TS data. method for CD. Furthermore, trade-off is sought between the false alarm rate and the able CUSUM method for the purpoIt se,w aa nsd o v ur al iid na te te n i ti t o w ni tto h r ae d ao lip st t ic a n Td S v da alta id . ate a continuous-time variant of the CUSUM It was our intention to adopt and validate a continuous-time variant of the CUSUM expected detection lag associated with the TS sequences and to provide hints for choos- It was our intention to ad m op eth t a o n d d fv oa rl iC dD ate . F aur co th ne ti rn m uo orus e, -tr tim ad ee - vo afrfi a is n t so oug f th he t C bU etS w U eM en the false alarm rate and the method for CD. Furthermore, trade-off is sought between the false alarm rate and the ing appropriate thresholds for the change detectors. expected detection lag associated with the TS sequences and to provide hints for choos- method for CD. Furthermore, trade-off is sought between the false alarm rate and the expected detection lag associated with the TS sequences and to provide hints for choos- ing appropriate thresholds for the change detectors. expected detection lag associated with the TS sequences and to provide hints for choos- ing appropriate thresholds for the change detectors. ing appropriate thresholds for the change detectors. Appl. Sci. 2021, 11, 3666 7 of 17 2.2.1. Modelling TS Occurrences within a Given Urban Road Environment As a ﬁrst step, we are going to model the TS placements, occurrences, and detections— along a random route and within a certain urban road environment—jointly as events of a continuous-variable homogeneous marked Poisson process (HMPP). In the literature dealing with Poisson processes usually the mentioned continuous variable is the time. Although, with the notations used herein, as well as with the verbal expressions describing relations between values, we are going to comply with this “tempo- ral” convention, and it should be emphasized that the path-length—that has been covered by the ego-car—is chosen to be the continuous spatial variable. Let fT , k g, where k 2 f1, 2, . . . , Kg and K 2 N, be a marked Poisson process n n n k k with counting measures N (). These are deﬁned as N ( A) = #fn : T 2 Ag, where A is typically an interval. Let the rates associated with the marked Poisson process be l , assuming spatial—and particularly along-the-route—homogeneity, and let the corresponding reference rates of the Poisson process be l . Then, the negative logarithm of the likelihood-ratio of an observation sequence (fT , k g), T < T is given by n n n N 0 k k D , T l l N T log (1) ( ) T å k å k k where N (T) is the number of events of type k prior to T. 0 0 Assuming now that l = l is the true set of process parameters, furthermore N N assuming that l = fl g is a set of tentative parameter values, and writing D = D (l), T T we have the following inequality: n o E D (l) 0. (2) The left-hand side is simply the Kullback-Leibler (KL) divergence of the true distri- bution from the estimated one. Using the common notation of the KL divergence, the left-hand-side of the above inequality can be rewritten as n o N 0 E D (l) = D mPois l T k mPois(lT) , (3) K L where mPois represents the distribution corresponding to the marked Poisson process, ( ) while D mPois l T k mPois(lT) is the expected number of extra nats—NB: not bits, K L but nats, as the natural logarithm is used in Equation (1), not log —required to encode the observation sequence from the distribution mPois l T using a code optimized for the distribution mPois(lT) rather than using the code optimized for mPois l T . Associated with D (l) is the computable quantity L (l) , T l N (T) log(l T) (4) å k å k k k L (l) can be interpreted as the approximate length of an optimal code encoding the observation sequence, were l the true set of the process parameters. Considering, however, that L (l) is dependent on l —i.e., on the “real” true set of process parameters—via k 0 N , we can write L l = L l , l , and similarly, we can indicate the same kind of ( ) T T N N N N 0 dependency for D , i.e., D = D (l) = D l , l . T T T T 2.2.2. Modelling TS Occurrences in a Neighboring Urban Road Environment In conjunction with a second road environment that borders the one looked at in the previous subsection, let us now consider another marked Poisson process with parameters 0 0 0 m = m . The probability distribution corresponding to this process is mPois m T . The counting measures M () are used for counting the events of different types, i.e., for counting the occurrences of the various TSs, separately. Appl. Sci. 2021, 11, 3666 8 of 17 Let the a priori estimate of m be m. Then, similarly to our comments with respect to L (l) deﬁned in Equation (4), the approximate length of an optimal code encoding the observation sequence observed within this second road environment—were m the true set of the process parameters—is given in Equation (5) J (m) , T m M (T) log(m T). (5) T å k å k k k Considering that J (m) is dependent on m —i.e., on the “real” true set of process k 0 parameters—through event counts M , we can write J (m) = J m , m . Furthermore, the T T negative logarithm of the likelihood-ratio of an observation sequence—observed within M M M 0 this second road environment—can be written as D = D (m) = D m , m . T T T 2.2.3. Modelling TS Occurrences over Two Neighboring Urban Road Environments 0 0 N 0 If l and m are reasonable estimates of l and m , respectively, then D l , l and M 0 D m , m are going to remain relatively small. We shall consider the case, when the 0 0 parameter-sets of the two marked Poisson process differ considerably, i.e., l and m differ 0 0 considerably with l and l still being close to each other, and with m and m still being N 0 M 0 close to each other. Then D l , m and D m , l are going to be large. Furthermore, using the encoding argument outlined above, L (l) L (m) will have a T T tendency to decrease in time (i.e., with T), and similarly, J (m) J (l) will also have such T T a tendency. The negated version of the latter, i.e., J (l) J (m), on the other hand, will T T have a tendency to increase in time. More clearly, using the deﬁning formulae in Equations (4) and (5), respectively, L (l) L (m) = T (l m ) N (T) log (6) T T å k k å k k tends to decrease with T, while J (l) J (m) = T (l m ) M (T) log (7) T T å k k å k k tends to increase with T. 2.2.4. Detecting Change between Urban Road Environments and Locating the Change Point Assume now a switch from the ﬁrst marked Poisson process to the second, i.e., from 0 0 mPois l T to mPois m T , and accordingly a switch from respective counting measures k k N () to M () at time t. Then for T t g , g (l, m) , L (l) L (m) (8) T T T T tends to decrease with T. Let us now introduce the notation T = T t. For T > t, i.e., for T > 0, g is deﬁned as follows t t g , g (l, m) , g + J (l) J (m), (9) T T t T T t t where J (l) J (m) is to be computed according to a modiﬁed version of Equation (7), T T as shown below t t k J (l) J (m) = T (l m ) M (T + t) log (10) å k k å t T T k k k k here the counting measures M () count events in the same way as M () with the only difference that they now count events only from t onwards. Appl. Sci. 2021, 11, 3666 9 of 17 For T > t, g tends to increase with T. To estimate the location t, one should monitor function g . Then, in order to determine t, we need to wait for an increasing trend in g T T to appear. In order to ﬁnd out whether a change in the stochastic model has occurred, or not, and if it has, when/where, one should compute the minimum of g on-the-ﬂy with a Page-Hinkley change detector (PHCD), see [42,43] for the detailed derivation of the change detector. Using g , the PHCD, h is deﬁned as T T h , g inf g . (11) T T sT A change is thought to have occurred if h exceeds a threshold d > 0. As it will be clear from the examples presented in Sections 3.1 and 3.2, the choice of d is crucial for the proper working of the detector. It can be shown—following the line of thoughts presented in [44]—that under the hypothesis of no-change, h is L-mixing, and the false alarm probability is exponentially ad decaying in d: P(h d) Ce with some a > 0. Hence the false alarm rate itself is exponentially decaying in d. As a consequence, the false alarm rate can be effectively reduced by choosing larger d. On the other hand, if d is chosen to be too large, then the detection lag can be too long or even transitions can be missed. 2.2.5. Basic Properties of Functions g and h T T Before getting on to elaborate concrete TS-based RET CD examples in Sections 3.1 and 3.2, it is worthwhile to look more closely at the functions involved in the CD computations, namely, to g and h . The diagrams of these functions are composed of linear segments. T T In case of g , each of these linear segments has the same slope, and at either end of a segment, there can be a “jump”. The jump can be either an upward jump, or a downward jump depending on the particular event and on the process parameters. In case of h , the situation is slightly more complicated. Apart from the linear segments with the same slope, if any such segment remains in the diagram, there can be broken lines reaching the horizontal axis, and a number of linear segments along this axis. Furthermore, all the constituting linear segments and broken lines of h are located in the upper half- plane that includes also the horizontal T axis. 3. Results 3.1. Examples Let us now see two CD examples from the given application ﬁeld. In these examples, we intend to detect change in the RET based on TS occurrences along a route. The full length of the trip represented in the table is 4.6 km. The TSs are assumed to be detected and located by an on-board TSR system. The TS locations are marked with the corresponding TSs in the top band of Table 2. The sequence given there is synthetic, and has been compiled for the purpose of demonstrating a RET transition from a Dt to a Res area (denoted by Dt ! Res), if the virtual journey is taken from the left, and a RET transition from a Res to a Dt area (denoted by Res ! Dt), if the journey is taken from the right. In the middle and the bottom bands of Table 2, the actual counts of the NS, PL, GW, and SL TSs for the virtual trips starting from the left, and from the right, respectively, are given. These counts have been produced in a uniﬁed and generic manner, i.e., without demarking the validity intervals of the respective counting measures. NS PL GW SL As these counts are denoted simply by N , N , N , and N in the present and in the subsequent subsections, care should be taken to use them properly, i.e., according to the direction of the virtual trip. Appl. Sci. 2021, 11, x FOR PEER REV AI pEW pl. S ci. 2021, 11, x FOR PEER REVIEW Appl. Sci. 2021, 11, x FOR PEER REVIEW Appl. Sci. 2021 6, 1 o1 f, x 1 8F OR PEER REVIEW 6 of 18 6 of 18 6 of 18 Appl. Sci. 2021, 11, x FOR PEER R AE pV plI.EW Sci. 202A 1p , p 1l 1., S xc iF . O 20 R 2 1 P,EE 11R , x R F EO VR IEW PEE R REVIEW 6 of 18 6 of 18 6 of 18 Appl. Sci. 2021, 11, x FO A R p A p Pp lEE .p S l.R c S i .R c2 iE .0 2 2 V 0 1 I2 ,EW 1 1,1 1 , 1 x, F xO FR O P REE PEE R R RE R V EIV EW IEW Appl. Sci. 2021, 11, x FOR PEER REVIEW 6 of 18 6 6 o f o1 f 81 8 6 of 18 2. Materials and Method 2s. Ma 2. Ma terita elrsi a aln sd a n Me d Me thotd hs 2 o .d Ma s teria2 ls . Ma and te Me r2 ia . lt Ma s h a on d te d sr iMe als ta h n od d s Me thods 2. Materials and Methods 2. Materials and Methods 2. Materials and Methods 2. Materials and Me 2. Ma thot d esr ials and Methods 2.1. Car-Based Collection o2 f .S 1t.2 a C .t1 ic a.r C R -B a oa r as - dB e d O a s C b eje d olle cC t c o D t lle io at c n a t io o fr f n o 2 S m .o t 1 f a .V t S C ic t aa a rR t r io ic -oB u a R s a d 2 s o U .e O 1 a d r .d b b C C je O ao a n cb lle r t - je R D B 2 c c o .a ta t 1 a io s td .D e a n C d E a fr o a t C n o a f rv o m - S fr B lle ir t o a V a oc m s t n a t e ic io m r d V io R n e C a u n o r o os ta io lle f s d U S u c O t s rta b io U b ta ic je n n r c b R o R t af o n D oS a a R d a td a t o O a t E a ic fr n b d je v R o E ir m c 2 o n t.a o 1 v V d D n .ir m a O a C o r te n a io b a n r m je u fr - tB s c s eo t n a m U D s ts e rV a d bt a a a C n r o io fr R lle o uo m s c a t U d io V r E a n b r n a o io v n f ir u S R s o t o n a U a t m ic r db e E a n R n n tos v a R ir do o O an db m E jeen cn t v t ir D s o an ta m fr en om ts Various Urban Road Environments 2.1. Car-Based Collection of Static Roa 2d .1 O . C bje ar c-tB D as ae td a C fro olle m c V tio arn io o uf sS U ta rtbic an R R oa od a d O E bn jev cir t o D nam tae n frto sm Various Urban Road Environments 2.1. Car-Based Colle2 c.t1 io . n C a or f -S Bta as te ic d R Co oa lle d c O tio bje n co tf D St aa tt aic fr R om oa d V a O rio bje uc st U Dra bta an fr R oo m a d V E ar nio vu irso n U m rb ea nn ts R oad Environments A series of car-based staA tic s A R er O ise es d r ia o eta fs c o c afo r - l cl b a ea r cs -ti b eo a dn s e s tr t d a iti s p A t c s a ti R s w e c O a r i R se d O s c a a ta o r d A fr a ic c e ta s o a d e l r lr c -e o i b o e c ut a ls ti ls A e o e o i cn d n fti s c e o tr s Hun a r t n ii a rp e - ti tr b s sc g a io w p s a R fs e r a O d y c s w a i s c r d a n - t a a s b a r ta ti a r cis a c e e c rd R d o r il O e o lse d ut t c a d ti o ti a io ut c ta n n R Hun i c tr O n o il Hun p ld e g s A a c a ta ti w rs o g y a e c n a s ro ir i n tr e l c y l s a e i ir p c o n rti s f i e o w c d n a a r o tr s -ut b ica p as s ire n rw d i e Hun a d ss t a o c ti ut g ac r a rr R iiy n e O d iHun n d o a ut ta gi a n cro y Hun l li en c ti g oa n r y tr iin p s was carried out in Hungary in A series of car-bas A ed A s e s s r te iae rti i se c s o R fo O c f a c r d a -a b rta -a b s a e cs d o e ld l se t 2c a s0 t ti ti 1 ao c 7 ti n .c R T tr O R h iO e p d s a d d ta w a ata ta A a cs o w c sl c e o la e e r lr r l i ce r e e ti ic s e o c ti d o n o o f l n o l tr 2 c e ut 0 a i c tr p 1 te r2 i- 7 s ip b d 0 n . s w a 1 T Hun f s 7 w a r h e .o s d e a T m c sh d s a g t c e a r a a a a r ta r ti d r in y e r c a u i d w ta e iR m n d e o O r b w ut o e e d ut e r c r i a o n e o ta il f n lHun c e o 2 ur c c Hun 0 o lte lb 1 leld 7 g a c e.te n a c g fr ti T r d a y a o o h r r m y n e f e ir n a 2 o d i tr s 0 a n m a .i 1 p ta n D 7 s a u . a w m T w ta nh e u b a 2 e r c s m e 0 e o rd 1 c b n c 7 a o a e c o .r ta fr e r l T lr i ur o e e n h w c f d i e b te n e ur a o r d g d n e ut b a a f ta a a cr o n r ir o n e i l w m c la a e Hun h s r e ce .e a r te a r e D n s d c a .g u o ta f a D m r lrl o a y e b cm ta c o e ite r n n c a c d o o e f n n r f u n r c ur o e i 2 m n m r 0 bn g b 1 a i e 7 n a n a r . g n a T r o r u ia f h c e m h e a ur r e s ib d c .r b e h a D a rta e n a r o ta f a w r ur e e cr a o b e s n a .c c n D e o r la a n lr e ta ie c n a te g c so d . a n D f c r ra e io c ta rm h n e ic n r a o g n n c a u e m r ri n c b ih n ee r gr o af rur ich be ar n areas. Data concerning a richer set—than presented here—o sfe T t— Ssth an an d p or f e ss o em nte e d m h oe re re c— ha o rfa T cte Ss r ia sn tid c R ofO s ss o e t m w — a eth s m g ao a n r t h e p e r cr e h e s a d e rn .a T te cte h d e r h ise ti rc e — RO of s T w Sa ss s a e g n ta — d t h o th e fr a s en o d m p . T r ee h m se e o nr te e d c h ha erra ec— teo rifs T tiS c sR a O ns d w of a s s o gm ate h e m re od re . T ch ha er acteristic ROs was gathered. The 2017. The data we2 r0 e 21 0 c 71 o . 7 lT l . e h T ce te h e d d a d f ta a ro ta w m w e r ae e r n e cu oc m lo le b ll c e e te r c te d o f d f r ur o frm b oa m 2 n a0 a 1 n a 7 r u n e .m u a Ts m b h . e e b D r e d a o ra ta fo ta ur fc ur o w s bn e a e b t c n r — a e e n r a th n c ra o ie n a r l an l e g s e a .c p a ste D r . e rd D a isc ta e a h fn ta r e o c te r o m c d n o c h n a ee c r n r e n e u ri— n n m sig e n o b t g f — e a r T a r th S o i c r s f a h i a c n ur e h n r s p e d b e r r t a e o — n s fe s th a n o rte a m en a d e s p .h m re D e o rs a e r e ta — e n c te o c ho f d a n T r h a c Se c e sr te r e a nr — n iin d so g ti o f c a fT R sS r o O i sm c s a h e n w e d r m a s o o f g r e s ao tch m hea e rr e m a d co .te T rr e h i s c eti h c a rR aO cte s r w isa ti sc g R aO thse w rea ds. g Ta hte h ered. The TSs and other ROs were recT or Sd s ea d n d m o an th ua erl lR yO as lo w ng er th e r ee r co oute rdes d — m to ag ne ua th T le l Sy rs w a aln io th d n g o th th th ee e R r r E R oT O ute ss o s w f— er to e g re eth coe rr d w ed T i th m Ss a th a nn e ua d R lo E ly th T a se lr o o fR n g O th s w e e rr oe ute res c— orto deg d e th me ar n w ua il th ly th ale o n Rg E T th se o rfo utes—together with the RETs of set—than presentesd es th e — e t— r th e— th an a o n p f r T p eS r se s e s n a en te nd te d o d h f e h sro e er m — e— e o m fo T f o S T rs e S a s cn h ad a s ne r d o a t— f c o te sfth o r sm io a sm n ti e c m p e R rm o e O r s o e e sr n c e w T te h c S a a h d ss r a a a h g r c n a e a te d r c th e te r o — ies r th rti ie o sc e d ti f r R .c T R T O S R h O sO s e a s w s n w w a ds e a o rg s e f a g s r to a T h em t c S e h o r se e e r a r d d m e n e .d d o d T .r h o T e m e th T h c a S h e e n s a r ua r a R a n l c O ld te y s o r a w ith lso ti ee n r c r e g R R r th O e O c e s s o w r r w o d a ute e e s r d e g s m a r— t eh a cto n e or ua rg e d e d e lth l.d y T e m a h rl e o w a n n iua g th th lth ly ee r a R o lo ute E n T gs s th — ofe to rg oe ute the sr — w to ith ge th the e rR w Ei T th s o th f e RETs of TSs and other ROs were re th co er g die vd e n m T a S a r s n e a ua an s— d ll y o w th ailth o er n th g R e O th th h se e e th w g r lp o ie e v ute r o g e efn i v r a s e a e — d c r n o e e to a a d rd s r g ie — c e e a a d th te s w — m e d ir th a w ta w n ith b ua th ith le e l th th t l h - th y b e e e l a a e p h g s l R o e i e o v n l d E f p e g T a A n o s th d n f a th o e r d a ef e d e r d r a o ig o s c e iiute — a d d v te i e a c w n d s a p th — i te p a th ta er ld to ib e g c th a l ta ia g e v sti e e tb — e - th o h l n b e n w e a e t a ,ls - r i r p b w e th e w a d o a h s i s f th A ie th — la d e n e d w th th d A h e ri e n e e d o th l d ip iR c d r th a E o o a te T f ie p d a s d p h a d o ta l e p if e lc p p b d all i ti o e ic c t o f a a - th n a te b ti ,a d e o d w s e n g e ta d h ,d i v i w ib c lA e l e a h e n n te th t i l - a d e d b e rr a e th o ta s a ie e s d b d — l a e A p t w -p n b ith l d a is c r e o a th d ti id e o A n a h n p ,e d w p lp r lh o i o c iia lf d e ti a a th o d p n e e p , d lw iic ch a ai ti te le od n th ,ta e w b h le it le -b th ase e d Android application, while the TSs and other ROsT w Sse r ae n d re o co th re dr e d R O m sa w nua ere ll y re a clo orn d g e d th m e r ao nute uals l— y a to lo gn eg th th ere w ro ith ute th s— e R to E gT es th oefr with the RETs of the given areas—with the h tre aljp e co to f a ry th d d e ed a g ta ii cv a e o te n f d th a ta re eb atr l se — it p tr - b w a w a jitr e s a th c e s ato d jth e c r A o cy e to ln l h e d d rc e y a rte l o ta p d i d d a o o ta a fa f uto a p th o d p fe l e m i th d c tr a a ie c i ti ti p a c tr o te tr a w n ila d p ,la jy w e ta s w c b h to c b a y i o l s lr e l e th y lt c e - th o tr b c e d lte a e l a a a e s jta d p e c ete c p d a to o d uto A i fn r th n y tr a e uto d m a e d v r j e a a e o tr c ti m ta r ito i y d c p a a o a r f ti lw e y f p lc w y p th a a d l s lb a l i e s y y c ta c e a tr o cb th ti o lo iy lo p n e e f n d c th th w a te ,s p w e ,a e d p s h a tr a p c iil uto i n o p e p l e l th iw e v n m c e e a te e a r sv ti y d c e c f o r a a e tr y ll uto w ll e a y fc je e s te m b w e cy to c d a o s ti th r e a n y c c uto d e a o d s ln a l,a y m p d ta p s b a , y ti o in c fth ath e le v ly e e a rp b tr y y p i p f th e i n w w e e a s a v se p ec cp ro o y n li ln f d ee c s w e te , v d e sr e a y cuto ofn ed w m s ,a s e tic co an lld ys b , y the app in every few seconds, the given areas—w th ith e g th iv ee h ne a lr pe o as f — a d w eid th ic th ate ed h ta elp bl o et f- a b a dse ed di c A ate nd dr o ta id bl a ep t-p bl ais ce ad ti o A nn , d w rh oi il d e a th pe p lication, while the trajectory data of the trip a w na d s a ct oth lltr e ec a ti te je m d c to ea suto r o yf d m th aa e ta ti T c o S a a f ln a lth y d n a d e n b a t d y o tr th th i a th p t e e e th r w ti a R m e a p O sti e p c s m e o in o n e lf tr l s e e th ic o v ete fe e s r th d [ T y 2 S e a 5 f a uto ] e a T .n w n Sd d m a s a n o e a t d th cti th o o c e n e a r th d a lti R ls n e y ,m O r d b R e e a y s O n t o th tr th e fa i n e e e th n tr s ti a d e i p [m e 2 a T p s 5 t e S ] [ s th i .2 n a o 5 n e f e ] d .ti v th m e oe r th e y T s eS fr o e a f w R n th O d se e e o c T n th oS tr n e ia d r en s s R ,d [O 2 o5 th e ]n .a etr n rd iR e s a O t [ 2 th e5 ne ]tr . ti ie m s e [2 s 5 o]f. the TS and other RO entries [25]. trajectory data of th tra ej etr cto ipr w y a dsa ta co lo le f cth tee d tr ai uto p w m aa sti cco alllle yc te by d th auto e am pp a ti in ca e ll v y e r b y y fth ew e s ae pcp o n in d s e,v ery few seconds, and at the times ofa th n ad n e d a Tt S ath t ath e n d ti e m o tith m ese e o r s fR o th f O th e e T e n S tr T S i ae n a sd n [ 2 d o5 th o ]T .th e h re e R a rd n O R a d t O e a a n e c t tr o n th li tr le e ei sc e ti t [s i2 m o [5 n 2 e ]5 .s p ] .e o r fs th on T en h T e eT l S d c h a o a e n n t a d d s i a c s o to ta th e l ld c e eo c r o l t li f R e o tc O n w t i p o o en n e p r tr p s eo ir e e s n rs o s n T o n [e h 2 n s le 5 : n c ] a d e o . ln d a c t s ra o i is v n c te e T s o r id l h s la t e e o e n c f d d t d i t a o o w a tn f a o d T tp c w a h p o e te e o l r al s r e d e o p sc a o n n e tti n t r n a o r s s e y n o c :l o n a a p c ls s o l e d :s e n rr ia c s s i s t o v -i i d s o n e r tn r n i e v d e a p e ln e r o c d r f a o s n a t n o w d s n d i o n s a at e t p e a d ld e c a e r o to n s a n o f t r s e n tT y iw n s s h :t ta o r e e a s y d s p d d ia e o s a rs - r f i ts s v ai to e s w c n - ro o s a l: l n p e ad c e d tr ia s o r o in d v n a e p st r :a e a a re s n d n o d r t n r ia n v y e e d a r la sc a t so a n is n d e -sn ia s t r td e ya d a t a o s s fe i s n tw -tro y p ae ss ri ss o -ns: a driver and a data entry assis- tant. The manual data entry w taa n st .m Ta hd ee m ea as ny u a bly d ta hte a a ern rt ary y- lw ika es s m cra ed en e d ea es st y ia g n b nt y .w T th ih te h e a T m rS ra a sn y a u -n la id k l e d R s a O c tr a e e en nt r d y e s w ig an s m wti a a td n he t T . e T Sa s h s e y a n m b d y a R n th u O e a la d rra aty a- l ein kt er y sc w rea es n m de as die g n e a w sy it h b y T t S h se a a nrd ra R yO -li ke screen design with TSs and RO The data collectionT p h T ee h r s d eo a d n ta n at e c a lo c c lo o le ln c le s ti i co s tt in e o d n p e o p rfe s r o tw sn on o n e n p l e e clr o s c n o os n n is s s:t i e s a T t d e h d d o e r f i o v d tfe w a r ttw o a a o n c po d e p l t r la e a s en r o c d s t n to .i a s o T n t:n a s h a :e e p d a n e m r td ri rs a v r y o i n e v n a r u e n sa a r s e l n i a ls d d n - ca o d a tn ad a s e i a d s n tta a te r t d e a y ta n e o w n tn f r t a .y tt r s T w y a m h s o a e s a s i p m d s se t - ie a s r a n s -n eo t a u .n sa T s y l:h d b a ey a d m tta rh a i e v e n n e u a r tr a r r a y la n d y w d a - a lta ia s k d e e m n a sa t tc a r d ry e e e e n w e n ta a r s d y sy e m a s b s ia y g sd i n s te h -w e e i a a ts h ry r T a b y S y -sl i ta k h n e ed s a c R rr rO e ae y n -l i d ke es i sg cn re w en i td he T si S g sn a w nd it h R O TS s and RO symbols. In case of params ey tr m is zy b eo m dl s b T .o S Iln ss, . c e Ia .n g s .e c , a o ss p fe e p o e ad fr p a li m a m re a is tm tr y si,m e z tt e h r b d i e o z T l s es td S .a s I n T ,n s d e S y c .a s m g a ,r .s d e ,b e .s o g o p o l.p s ,f e .t s s e p iI p y d o n am e n r le c a i sb m a d m (o s il ie .l e i t e s m s to ..r ,, f I iit 1 z n th p 0 s e e a ,c d k a t r s h m a s T te e a m S / n o h s s e d t ,f ,t a e a r 2 p n i .r 0 g z a d d e r .a , a d o s rm p d p Tt e e S o ie t o s p r d n ,i tz e s ilo e . ig ( m d n i.., s e iT t s .(,s s S p i ,y 1 .s e e t0 m ,e h . , e d k e b 1 .g m 0 o s l.it ,lm / k a s h s.m n p i,I t d e n 2 s /h a e ,0 c r d t , a h d 2 s le 0 io e m s p o tia tfti n s o p ,d n a ta s h r a r e (d m i .s e o te .a ,p t n r 1 tid i 0 z o a e k n r d m s d T ( /o ih S .p e ,s .t ,2 , i e 0 o 1. 0 n g s .k , ( m sip .e /e h .,e ,1 d 2 0 0 l i k m m it /s h,, t2 h0 e standard options (i.e., 10 km/h, 20 tant. The manual dt aa tt n a a tn e . n t T .t h T re y h m e w m a an sa u m na u a l a d d le a d te a aa t e s ay n e tb n ry y tr y w th w a es a a m s r rm a ad y a - e t d la ie e k n a e t e s . a s y T s c y h b re y e b e m y t n h t d a e h n e a e u sr ia a r g r a ln r y d a -w a y lit -k ia ltie h k e s e n T c ts S r rc e s y r e e a n w e n n d a ds e d R s m e iO g sa in d g n w e e w ita h ist y h T S b T s y S a s tn h ad e n a d R r O r R aO y- like screen design with TSs and RO symbols. In case of parametk ri m ze /h d, T … Ss s , ,y 7 e m 0 .g b k .,o m slp s /h .e I) e n d w c le a im r se ei t o o k sf f m , f p t e/ k h a r h m e e r, d a s / … m — t h a,,e n a … t 7 d lrs 0 i a o ,z r k e 7 d im d 0 n o / T k p p h m S it )c i s o / t w ,h o n e) r e .s i g r w a ( e .i l , . e s o e fr p o k .fe ,f r e m 1 e m o e 0 r/ d f e h — fk d e l ,m ir — a m … e f/d th a i,e t— l,s r 7 s k 2 ,o 0 t m a 0 th h li k s e / n e h o m g s ,p i / t e … n i h a c n n )tk p e ,o d w r m i7 rc a a i 0 e t l / r a o r h d lk T e r , f m i S o o … a o p l r / f t h y f m t ,fe i o ) p o 7 r — r e 0 n w em . d s a k e S — — f (r m p t ie .e e a e a /ro .h l c f ,s t i t f )1 o e f h f ie 0 r w c e i r n k te e g h d m re p e e — n / i g c h o et a e r ,fo f ln a 2 s e rl0 e o ir a r T e a l id S n lf— o T tp y r k S a im p m c lt s t e — y o o /.h r p S a i i,e a n p f… .l t e e p S f c r ,o p i i c f 7 r t e ith m 0 c o c e ir k — f iig m a ca e l /n ffh to e e )r r r m a w tlh — eT e rS e a g fo te ty e f n f p re e e t r rh .a e d l e S p — T ge S ec a n i t lf y s e io r p ca e iln . T S p S p i e c ty c to ip frie ica . lS f p oe rc m if— ic after the general TS type. Specific symbols. In case of s p ya m ra b m ole st.r I in ze c d a s T eS o s,f e p.a gr .,a s m pe et erd iz le im d i T ts S,s t,h ee .g s.t , a sn pd ee ad rd l io m pittis o,n th s e (i .se t.a , n 1d 0 a k rm d / o h p , t2 io 0 ns (i.e., 10 km/h, 20 Appl. Sci. 2021, 11, 3666 10 of 17 km/h, …, 70 km/h) were ofs fy em red bo — lsa ,k lism .o e. / ,i h n t,o p u … icc,h t o 7 -s 0 rc i a r k e lm e fn o /s h ry k m )m ew s — y y b s e o m ,a r lf w e s b t ,e o e o r ir l f . e s e fte ,.h ,o r i e .t e fe o f d g .e u ,— e rt c e n o h d a e u -l r s c s fa c o o h lrr -e i T sn e e c S n n r p e t ts k e i e y y c r e n p t i m y o n e k s r .g b ,e i S o a y w tp lh ls s e e f ,e , o r c w ie c r i.s f e o m e y i o .n c r ,m f — e s f t e o ib o d a r u o f e f e f c t l d r e e s h e r ,r s - f d e s i y o t d . c e r m h R r .f e ,e e E o b e n tT g r o o n t e s le u e s ,n k r n c ,i t e e h t n i h r e .y - e g e a r s s . l i c ,,r t n r h T t e w e g o p e S e u e e t n c r h t c a o e y h e t k n p e - o e c s d s e fy o c if . d r s e n S e ,r e s p e e w r in d d e ee d c e k f r ir o f e R e e i r c y d o E e s f T R ,f n e s w t E ,r e e T e t s rh d y i r sn e e , m f g to o r h b e r f te o f h p e e le r e n r se a e ,c t p t d e io e .e r e n d f i a .n o s , t ie r g td o d e e tu n h rc e e th e d c r - s i o R n c n E r g s e T ie td s h n ,e e r t k h e ce o d e y n r s R s e , i E p w d T e e e a s rr ,t e e e t d h d o e R ff E r ee r T p es d e , a tft h o ee r d r e e np te er aitn eg d the considered RETs, the repeated km/h, …, 70 km/h)k w me /h re , … off , e7 r0 e d k— ma /h ls)o w in e rp e ic otfo fe ri ra eld f— ora m ls— o i an ft e pri ctth oe r ig al e n fo er rm al — TS a ft ty er p e th . e S p ge ec n if eir ca l TS type. Specific symbols, i.e., touchs -y ss c m y re m b eo n b l o sk ,l s e i,.y eis ..,e , t .w ,o t u e o r c u e h c - o h sfc -fs r ec e rr e ee n d e n k fo e k r e y e n s e y ,t n r s w t i,e e w e r sr i,e n e fr g o s o ey r f t o f m h te h fe f r b e e e c o r d c o e la s n d fn ,o s ic f ir .d o e e e l .r e ,a n re t te ti o n e o d u t r n e ic e R n s r h n i E g o n - t T s r e f g ti c n h s e t r,h tt e s e h r t ,e e i h c e f e n o lo e s a c n r ,k o s r s f t e t e n o i h p y e d s r e n s e i e d t ,a t c r h r e w t e a y e e r d n e d ,e c c r d a R a e e n n l E R a d o c T t E f e if f s o T lo e ,a n r s r t te , h s i e e o d te n o n h n f f t t r e s r o e e t ir r r h o p e i e f e s n e e p , a t n g le h fa tt o a e v e e sr t d t r e e e l it r a e n n d h b s n tg e t a r t i r le t c e y h n c a s,o e ,t n a r m fc c y o n e e o ,n m r d ln a a t tr e s f n t h o i iin d e o e d rt s n e s c f ,e r .s o a n fe T r o n o t d r h e c e f e r n e t R til h h t n a E le e e o t g T r i c c o l is v a a n a n ,t s e g n is t t r o h c b v e o n e e a e n f l, a l r r tt b r t e h ciy p a o o e,lm e n l a c a a sn m o t se o d m td e f e n f m tn o h tr t s e e r .n e y T ln t e a ,s h n t s a .e e t t n T r r e d li ih o n e n e s f c g to ,r a l y r v f o to i ,e c e o r a a r n n b n t tt ,i h a e d oe l r n i f c c n o ,o a g rm n e v cn m e etl re e a b n r ta i itn o ls n g c . o T sv m h o ee f r m b tlh o a el e n c a c tls t o ai.s m o T tn m h e , n ee tl n ro y tc s , a .a t T n io h d n e f , l o o rc e an tit o en ri , ng verbal comments. The location, time, TS/RO, and RET categot riim cae l ,d T a S t/ aR w Oe , ra en s d to R re Ed T ic n a t ae t g eo xrti-c fa ille d in at a a c w oe m rte i m m sa te o ,sr e T ep d Sa / i R rn a O ta e , d ta e n v xd t a- lfR u ile E es T i n c a at c eo gm orm ica al sd e ti a p m ta ae r a w , t T e eS d r/ e R v sO a to l,u ra e en d s d i n R a E T te c xa t- tf eig le o i rn ic a a lc d oa m ta m w a e sre ep s atr o arte ed d iv na a lu te es x t-file in a comma separated values entries, for the canc ee n le a tn r tii te o rs in e , s s f ,o o frf o tt rh h te e h c e la a c s n a tc n e ec n la e tl t ra iy o t,in o asn n o s d f o fto fh r te h ele n a l s ta e t s r e e t in n n e tt g n r r iy t v e r ,s e y a ,r ,n b fa o a d n rl d ft c o h o fr e o m e r cn m a e t tn n ie e m c t r n e e it e n r ls a ,ig .n t T T ig v o S h e n /v e R rs e b O lr o a o b ,l fc a a a c tlh n o t i c e d m o o n lm R m a , s E m e tT n e e t n c n sa t .t r t s T y e . h g ,t T ia e o m h n rle ie o d c ,lc a o f T a lo c tS d r a io / t a R e in to n t O a ,i n t m e w ,, r a e i e n n ,r d T g e S s R v / te E R orr T O b e a d ,c la a ic t n n e o d g m a o R tm re E ix c e T a tn -lc f ti d a s l.t e a e T ti g a n h o w e a r i le c c o r a o c e lm a d s tt i m a o otn r a ae , s w d e p e in r ae r a a s tt te e ox d rt e -v d fia lil e n u i e a n s t a e x co t-m film e ia n s a e p co ar m at m ed a v se ap lu ae ra s ted values Table 2. A synthetic sequence of TSs used in the examples (top band), and the uniﬁed counts for the No stopping (NS), (cvs) format. After the trips, t(h cv e sc )s f v o r fim lea st .w A ef rt ee r s tto h re e d tr ia p ss s , p th re ea c d ss vh f eie le ts s a w n (e c d r v e s w )s e t fo o re r r e m cd o a n a t.v s A e sr p ftte r ee d ra td th ose h t ere ip tss ,a n th d e w cse v (r c e v f is c le ) o sf n o w v re m errta e et d .s t A t o o r f t ee d r a th s es p trrie pa sd , s th he e e cts sv a f n il d e s w w er ee r ec o st n ov re er d t ea d s tso p readsheets and were converted to time, TS/RO, and RtE im T ti m e c,a e T t,e S T g /S R o/r O R ic,O a a l, n d ad n a t d R a E R w T E e T c ra ec t a e st t g o eo g rr e o id c ra i i c ln a d la a d t ta a ett x iw a m t -w e fe ir,le e e T r s e iS n t / o s R t a ro e O c r d o e , d m a in n im d n a( a c t a R e v s E t x s e e t T )p x - f f t a c io -lr a fe r a itm le t ie e n g a d i o n t a .r v c i a A c a o a c lfm u t o le e d m m rs a t m a th a s e a w e s tp r e ei a p r p r e a s a ( r c ,s t a e v t to h t d se r ) e d e v fc d a o v s l rv i u a m n l e u f s a a i (e l c t e t s .v e s A s x ) w t f - f te f o eir r r le e m t h s ia n te o t .a r t A r e c id p o ft s m a e,r sm t h t sh p a ee rs c e e t sa r p v id p af s r s ih l a ,e e tts e e h d tw e s v c e as a r n v e ld u s f e it w l s o e r e se r w d e c e ao r se n s v s p t e r o r e r ta e ed d ds t a h o s e e st p sr e aa nd ds w he eerte s c ao nn dv w ere te rd e c to on verted to Parking lot (PL), the Give way (GW), and Max speed 30 km/h (SL) TSs in case of trips starting from the left and from the (cvs) format. After the tripsv , a th rie o u cs sv f o (fc irlv m es s) a w t fs oe r (r e m e .g a s.t t ,. o k A rm efd t le ) v a r f a o st r h r i v so p e p au r r o ts i e r s o a if tp u - o d p s s r s, r m h fo t o e h a c r e e e t m t s s s c s a ( s i e a tv n .s n g g d f (. i e ,a l .w k e n gs m .d e ,w r l k v e )m e i v fs c r o a u o l er )r n a s i f p lo v o tio o z u e rs r a r s p e tt t -e f d i o p o o d s r n r a t v o m - t s .a p o c r s e a rip o s to s s r c u ie e (n s e a s g .d s fg v o is a n .a r ,h n g m r k e d i m o e a a t v n u ts ls id ) s s a ( f u fe v o n o a .i r d r g s l m i p u .z ,w o a a k al s e t t m it i s r z o - e p l a (n ) e tc r .i .fo o g o oc n .n r ,e v .k p s e s m o r is n tle t) g -d p f a o r tn r o o d c p e o v ss s it s i- v n u p a g a rr l o i a io z c n a e u d s ts i s o v f in o n is r g .u m a aa n litd z sa ( v te iio .sg u n .,.a k lim zalt)i o fo nr. post-processing and visualization. (cvs) format. After (tc h v es ) t rfio pr sm , ta hte . A csfv te f ri lte h se w tre ir pe s ,s t th or ee c d s v a sf is le ps r e w ad er se h e se to ts r ea d n d as w sp er re e a cd os n h v ee ertts e d an td o were converted to right, respectively (middle and bottom bands). various formats (e.g., kml) for poF sr t- o p m v ra o th r cie o e su sa i sn b fg o o v r am e n d d at a v st ia (s e u c .g o a.ll,il z k e Fa c m rtti o ilo F o m ) n r n fo o .th , m r th e p e th o a s r b e t e o - la p v eb r v eo o a d c v n e a e t s t s d a T i a n S ct g o r a la a lc te e n o c F d s lti r l— e o v oc m in a s ti ,l u o o th th a n n le i e ,g z th a r a r F e b t o e r ilo ute o o e rv n v m ee .a l s e n th d v it n a a e F T t n th a r a S t o b c e r T m o o a S c v lte l o th e e rn s a c d e — s te ti i a a o d s a tb n a — e lo o r ,c e v n a th o d e l g lo e l d e r n r o c a g e ti ute t la r e o o v c n ute s o a , i n lth n lt e s e c th T i ti n r S e e o th r ln c e ao ,v te e n th a c s s n F o — e ir t d n r o T e a s e m r i l S ld o e e r d n v e th a r g a te e e n r d s t a o — b T ute o S a v l r s e o a in te d n g a sth t — ra o e c a ute c o lo o lln n s e g s i ci n ti d r o o th eute n re e , d th cs o e in n s r e ith d le e ev r c e ao d nn t sT id S e rr a ete ds —along routes in the considered various formats (e.g v.a , r k im ou ls ) f fo or r m po as ts t- p (er.o gc .,e k ss m in l)g f o an r d p o vs it s-u pa rlo iz ce as tiso in ng . and visualization. From the above daF ta rF o c rm o olm lth e c th e ti a e ob n ao b , v th oe v e d e r a d e tl a a e t v c ao a cR l n o le t E ll c T T e ti S c o ti a rn r o ae ,te n th a,s s th — e — r ea w e F r llr o e e eo l r n v e m e a g v n a r k th t o n n T ute e t o S T a w r S b s R n a o r ite .E n v a T te e s R T th — h E d sa e — e T a a r t c e la a o e o a a l m n n c s r oo e — g s n p a i l g d l r is w e r o — e i r c ute e c r o ti e a r w ute o e d l s n e r r k ,is a e n th n te i o n th k e s w n th e rn o f e c e o R lw .o e r E cv n T n o T a t s .h n h in e d s e T a t ie r d h T e r ee e m e S a r d s R e r p e — a d (m E i“ te r T w No ip c s i e — aa rr l ir e c a s e r a R lto a a o k ls E te n n p — r T g o s p a w w te ira n fo o n e r s g ute e r r .” e a f o T s ts ( h — r k h NS e in e n t w o h ) th e w e e )m ,r e n e ( p . c “ o i k No T r ,n n i(h c s “ o a e i No w d ls e e n to rr m .a e p s te d p T to p is h ir p n e if c p g o a i ” e R r n l m g E (r tNS ” p T h a te ie r ( a i ) NS sc ) r ,a e fla o ) ) s r (r ,— “ a , No tte h w s e ,e rf se o to r k p(n tp “ h o i No e n w gn ”.s to (T NS (“ p hNo e p ) i)n ,e g m s” to p ,ip ( r NS p icia n )l) g , ” ra te (NS ,s f )o ),r th,e (“No stopping” (NS)), , —— —— — —— — — —— ———— ——— —— ——— —— RET areas—were R k E R n T E o T w a r n a e .r ae s T a — h s— ew w e em re er p ek i rn ik c o n a w o l w n r.a n te .T ( “ s h T P e h a f o r ee k r m ien tm p h g R ie p E rl iio T c rt a i ” c la a r ( ( lre P “ aa r No L te s a)— te s ) , s w f s o to f e ro r p (( e r “ tp “ h P G it k e ( a n h “ n ir g e v P k o ” e a iw n r w (k g n ( NS i “ .a n l (No y o “ g T )” No t) h ” , l e o (s G ( tto P ” e s W) L ,m to p ()P p ) p p ),L ,i i p n r )a iig )n n c ,” a g d (l( ” “ ( “ NS G r P ( a (NS i a “ te v r ) G (e k )“ s, ) i i M ) v w n f ,e o g a a r x (y ,w l “ o ” t P ,s a h tp a ” y (e G e r ” k ( e W) P i d (n G L (g 3 ) “ ) W) ) ( ,0 P ,“ l a o a No k )n t r,m ” k d a i / ( ( n n s h “ P g d to G ” L p ) i l( ( v o )S “ p ,e t M L i ” n ( ) w “ a ) g ( P M x a ” T ( L y “ S a s” ( ) G p s x NS ) ,i e ( v s G ep e ) d W) ) e , w e 3 (“ d 0 ) aG , y ,3 k a ” i 0 v m n e ( d k / G h m w (W) ” “a / P h ( y ) a S ( ” ” ,r “ L k a M ( ) (S i n ) G n a L d T W) g x ) S ) ls so ) T p ,t S e ” ( a s e “ n d ( M d P L a 3x 0 )) , s k p (m “ eM / e h ( d a “ ” x G 3 (0 i S sv p L e k e ) m ) e w d T /a h S y 3 ” s0 ” ( S k (G L m ) W) ) / hT ” )S , ( s a S n Ld )) TS s ( “Max speed 30 km/h” (SL)) TSs were used herein as a prioriw ee sr ti em us ate ed s h oe f rth eie n ra es f e ar e pn rc io er r ia e te sti s m foa rte th s e o m fw th a ere rk e r e e us df e P erd o e i n h sc s e e o r n e ra i n p te ra s os -f o a rp th rio e rm i e as rti km ew d a e te P ro e s i us o ss fo e th n d e p h r r ee o rf - ee ir ne n as c e a rp ar te io sr f io e rs ti th m e a m tea sr k oe f d th P eo rie sf se orn e n pc re o - rates for the marked Poisson pro- (“Parking lot” (PL () “)(P ,“ a P ra kr ik ( n “ ig n G g l io v lt e o ” t w ” (P a (L y P) ” L ) ,) ()G , W) ( “ )(G ,“ a G iv nie d v e w w ay ( a (” “ y “ P M ” (a G r ( aG W) k xi W) n s)g p , ) e a l ,e o n a d td ” n 3 d (0 w P L k e() m r “ )e ( ,M / “h us M a ”x e a ( d (x S s “ p L h G se ) e p i) e r v e d e e T e i d n S 3 w s 0 a 3 a s 0 y k a ” m k p m / (h G r/ i” o h W) w r ” (e iS ) r ( e L ,e S s) a L ti ) us n )m T )d e S a T dte s w S h s s e e r o ( re “ e f M ius th na e e a x d s r s e a h p f e e p er rr e e e id i o n n r c3 ie a 0 e s r s a k a ti te m m ps/ r a h ifo te o ”r rs i ( th S e oL s fe ti )th )m m e T a a S r r te s e k s fee d o r e fP n th o cie e s s r ro a en te fe s p r e r fn o oc - re th ra ete m sa fr o k re th d e P o m is asro kn e d p r P oo -isson pro- ! 1 2 cesses. These rates are giv ce es ns c e i en s s. s T e Ta s h .b e T ls eh e 1 er s a fe o te r r s a r te o ar ute s e a g r sie v w c g ee in ith v s s ie i n e n n s T .D i n a Tt b h T a le c e a n s e b d 1 e sl s e f r w e o a 1 sr te i. th fr T s o o c i h r a n ute e e r r sR s e o se e ute s e g s s r w i. a v a T s te ie r th h n w e se a ii ia n s s n th r e . D e T ir n g t a a te i b a D v n ls e e t d n a a 1 n r w if e n d o i g th r T w ir a v ii o n b e th ute l n R e i n e i1 n s s R fw T a o er a r s ic e th b e r a a lo s s r e is n ute e . e 1 a s D s f .s . o t T r w a h n re io th d sute e i w n ra i sD th te w t isn i a th a n R rd ie e n s w g D i aiv r t th e ea n a in n s i d . n R w e Ts i a th a br lie e n a 1 R s .fe osr a rr o eute as. s within Dt and within Res areas. were used herein a w sw e a re e p r e us ri o us erd ie d e hs e h ti re m er ie n aite n a s sa o a s fa p th rp ie o ri ro rie re f ie se ti rs e m ti nm c ae te a w r te sa eo s te r f e o s th f us f o th ee r e r d th e r f h e e ee f rm e r ee r n a e ic n n re k c a e r e sa d r a te a P te p so rf s iio s o f s ro r o ith r n eth e s p ti r e m m o m a -a ra te kre k sd e o d P f o th Pio s es i s o rs e n o f n e pr rp e o n r-o ce - rates for the marked Poisson pro- ! 1 2 cesses. These rates are given in Tabc le es 1 s e fs o.r T rh oe ute se s r a w te ith s a in re D gt iv ae nn d iw n iT th ab in le R 1 e s fo a rr r eo aute s. s within Dt and within Res areas. cesses. These ratesc a er se s e g si.v T eh ne is n e T ra ate bls e a 1r e fo g r ir vo eute n in s w Ta ith ble in 1 D fo t ra r no dute wisth w in ith Rie ns D ar t e aa n sd . within Res areas. ! 1 2 3 4 Table 1. The empirical traffi Ta c s b iTa g le n b 1 (le .TS T 1 h ) .e r T aem h tes e p em a il ro ip c n a ig r l i t a cra a rla ffi tn rc d a Ta ffi o sim b g cn le s r io ( g 1 TS u n . tT e )( h TS r fTa a e ot ) rem es b r ta h le a t p e es li 1 o h r. n io a c T g Ta m la h o la n o e b t g g r rem le a a en a n ffi 1 d p r eo .c a o iT r n s u m ih i d c s g e a o r n m lo m em t u ( a r TS r t r ap e k o ffi ) i ed u fr o r c t ia e c r s a tt i f es lg h o tn r e r a a th ( lh ffi o TS o e n m c g ) h o s r o a ia g m g t r en n es a o n g ( eo TS a d en lo u Ta o)m eo s n r b g m a u r le ta o s a es u r r m 1 k t a a .e ed n a lT o r f d h o k n o e r ed g m t em h a r e r o p a h u n io t rd e im c o f a o m o lg rt en r r ta h offi u e eo th c e u o s f sm o i g m ro n t a g h ( ren TS e ked h eo )o r m u as to es m g en a al ro eo kn ed u g s a m ra an rk ded om route for the homogeneous marked ! 1 2 3 Poisson processes describinP go (id s P s o o o w in sn s p t oo r n o w c p n es r) os Dt ces es a s des n es d c d r res es ibc i id n rien g b P i(n o td ii g a o s ls w ( o d (R n n ot es w o pw ) r n o u t P n c o r o ) es w b iDt s a s n s n es ) o a r n Dt n o d p a d es P a d r r o o n c es en i r c d s ies b i s r v d o is es n ien n es rg o i p d t n d ( ir en d a m o es lo c t en ( w c es iR a rn i t es lsb s es t (.i o R ) n w u es d gr n es ) b () d a u cDt o n r rw i b b r a a o n in n n a t o g d d rw o (r en a d es n d o ) v i w en d Dt irn en ov tn a o P itm n r iw o a o d ien l n n s r ( s m ) R es t o Dt sen es n .i d p )t a en s r u n .o r d tc b i a es r ales n s ( es R ird o es en a d) d es tu ien c a rr b l i v a ( b R in irn es o rg n o ) m a (u d d en r o b en w a ts n n v . t ir o ro o w a nd n m ) en en Dt v t i s a r.n o d n m res en id tsen . tial (Res) urban road environments. Table 1. The empirical traffic sign (TS) Ta rat b es le a 1 l.o T nh ge aem ran pd ir o im ca lr o tru atffi e c fo sri g th ne (TS ho) m ro ag tes en a eo lou ns g m a a rr ak ned do m route for the homogeneous marked Table 1. The empiric Ta al b trle affi 1.c T sh ig e nem (TS p) ir r ia ca tes l t ra alffi oncg s a ig rn a n (TS do)m ra rto es u ta e lo fo nr g t a h e ra h n o d m oo m g en rou eo te uf so m r t ah re ked ho mogeneous marked 2 1 Poisson processes des Po c P i rs o is b is o is n n o g p n ( rd p oo r co w es cn es ses to ses w dn es d ) es cDt ri cb r a i in b ni g d n ( g rd es (o d i w d on w en tn o tt i w a oln w P () R n oDt i ) es sDt s )a o u n n a r d n p b d r r aes o n r c es ir es d oien a s d es d en t i en a d tlies a v (l R ic r(r es o R in b es )i m n u )g en ru b ( r a t d b s n o .a w r no n r ao t d o a w d en n en v ) iv r Dt o ir n o a m n nm en d en r tes s.t s id . ential (Res) urban road environments. Expected number ExpeE cx te p E d ex c n p tu e ed m ct n e bd u e r m n u N bm e ar tb u e rE r a x l p lE o ex g cE p t ae x e -d p ct e Ch n ecd u te a m n d ru - b n E m eu x rb p m N e eb r c a t e t N e r u d E a r a x tn E u l p u x r l e p m a o cl e g t b e c l ao e t d -e g r d n a Ch u -n Em u x a Ch p m r b -e e b a c rr t e - e r d EN x np a ut e m u cr tb e ae d l r l n oN u ga m a E - tx u bp r Ch ea e rl c a N t le o ra - d gt a u n -r u am Ch l lb o a e g r ra - - ECh xpe ac rt -ed number Natural loga- Char- 2 1 TS Abbevi- TS TSA bb Ae b v b i- evi- TS AbbT ev Si - Ab T bS e vi-Abbevi- TS Abbevi- Expected number E x E E p x xe p p ce e tc e ct t d e e d d n u n nm u um m be b b re e r r I E n N x d E p a x ee tp x u c e t rc e atd le l o d n o f u g n o a m u c -m c b u e Ch E b rrr x e e p r N an r e N a - c ct e ta u e stI r d u n a r l n d a I o l u e ln o f x m ld g o o a e b g c- x e c a u r -o Ch rf r Ch E e o a o x n c r fp a c c - o u r e e-c s c rc r tp e e ue d n rr r c n e en r u si c m tI e h n s b m d o e e f ro x o f o N cf tc a h o u te u c r I o c r n rf e a u a d n c o l re t r c c le e x o c e r n u s g i s c r a p - I e r-n e e so n r d Ch p fc e e r e o x i rs a c t h c rr - u m it rh o r o o e f m ff n o o t c o c c h e c c fe s u u t r h r a r re e c e o t n n e f a c c r c e o e its s c s e- c r p u ie s rr -r o e fr n i o c tc h ec Is m u n p r d o r e e e f r x n t c h re ie ts h o p a m f c e t o r o e c f rr c it i u s h t-h re rm en a o c ct f ee s trh i s e- o afc o te crciu s- rrences per rithm of the acteris- TS Abbevi- TS Abbe4 vi- 3 2 1 TS Abbevi- TS Abbevi- Type ation Typ Te y pe atio an ti on Type atT io yn p e T at y ip oe n ation Type ation Index of occurrences of occu Ir n rd ee nx c es po er f orc ic th um rre o n fc t eh se ao ct fe o rc is c-urrences per rithm of the acteris- Index of occurI rn en dc ee xs o of f o oc cc cu ur rr re en nc ce es s p er o fr o itc h cm ur r oe fn p tc h ee e rs k p a m c et r e i n rri iDt s t- hm of t k h m e ia nc p R te ee r rp s ik s e a - m rr e k i a m n s Dt in Dt rat e-k ra m ti k o im n R p in e e t s r i R ca k e t r m s o e a a is rn e p a Dt e s rr k at m e r- a r ip tk n a e e t m -i Dt r ro a k i t n m io R ite n is c k Dt a t tm o irc e ia tn o s Rk es m r a a r i tn e ea - R rsa e ts io a r ra et a p e s t- e i r c ra t k to im o r a tie n- rDt ta ic ti o to k ti m c tio n Res areas rate-ratio tic to Type ation Ty Tp ye p e ata io tin o n Type ation 3 2 1 per km in Dt p ke p m re k r i n m k m R ie n i s n Dt a Dt r ea s km k ra m i tn e i- n R ra e R ts ie o a s r a er aes ta i s cp te or r a k r ta e m t-e r i a -n r ta i o Dt ti o tic ti k t co m t o i n Res areas rate-ratio tic to NS 1 2.0 NS 0 NS 1 1 0. 35 2.0NS 0 2 .0 0 1 1 NS . 74 0NS .31 5 0 D .32 t 5 . 00 1 2.1 0.0 7 4 1 .742 .0 00 .3 5 D t DtNS 0. 35 1.0 1 7. 4 3 5 1.7 D 4t 2.001 .74D t Dt 0.35 1.74 Dt NS 1 NS NS 2 .00 1 1 2.0 2 00 ..0 3 0 5 P L NS 2 1.0 7.4 3 0 5 .1 3 5 1.7 PD 0 L t P L 2 1..0 7 10 2 4 .7 4 2 0D . 2D t5 t1 .70 P 10 .L 7 .3 05 1 2 P . 9 L 2 01 .P 2 .7 L 2 5 0 4 D . 21 t 5 . 70 2 Dt 1.1 7.0 9 2 1 .921 .7 00 .2 5 D t DtP 0 L. 25 1.0 2 9. 2 2 5 1.9 D 2t 1.701 .92D t Dt 0.25 1.92 Dt In the sequence given in the top band, each “—” signiﬁes a 0.2 km path-length along PL 2 1.70 PL 0.2 2 5 1.7 10 .9 2 Dt 0.25 1.92 Dt PL 2 PL 1.70 2 1.7 00 .2 G 5 W 3 1.0 9.2 2 5 0.G 7D 0 W G t W 1.93 2 3 0D .8 t0 0.7G 0 0 W .70 −G 3 0 .W 1 3 0 G .8 W 3 0 0 R . 8 e 0 0 s. 70 3 0− .7 00 .1 − 3 0 .13 0 .7 00 .8 R 0 esR eG s W 0.8 0 −00 3 .1 . 8 30 −0R .1e 3s 0 .70 − 0.1R 3 es Res 0.80 −0.13 Res the route without any relevant TSs—facing the ego-car—installed/detected along the GW 3 GG WW 0 .70 3 3 0.7 0 00 ..7 8 0 0S L GW 4 −00 .1 .8 0 30 . 3 8 0 0.2 S R 0 L e s S L − 00 .− 7 .1 0 0 43 . 1 34 0 R . 4 e R 0 se 0 s .20 S 00 L .2 .8 00 −4 0 S .L 6 9 0 −.S 0 4.L 4 0 1 0 R 3 .4 e 0 0 s. 20 4 Res0 − .2 00 .6 − 9 0 .69 0 .2 00 .4 R 0 esR esS 0 L. 40 −00 4 .6 . 4 90 −0R .6e 9s 0 .20 − 0.6R 9 es Res 0.40 −0.69 Res corresponding patch of road. It should be underlined that these ﬁxed path-length route- SL 4 0.20 SL 0.4 4 0 0.− 20 0. 69 Res 0.40 −0.69 Res SL 4 SL 0.20 4 Any0 .2 00 .4 0 −00 .6 .4 90 A ny4 .6R 0e s −0.69 1 R .8 e0 s 4.60 An y 1D .8t 0 4.A 60 n y Dt 1 .80 4.60 Dt 1.80 Dt Any Any 4.60A ny 1.80 4.60 4.60 1.80 D t 1.80 D t Dt segments are used in the examples simply for convenience, i.e., to make diagrammatic Any 4.6A 0 ny 1.8 0 4.60 Dt 1.80 Dt Any A ny 4.60 4.6 10 .8 0 1 .80 Dt Dt representation of the TS sequence, as well as the virtual trip easier to follow, and to make 2.2. Mathematical Models a 2n .2 d. 2 M M .2e .a tt M h ho e a d m ts h aetm ica at l ic M al od M els od a en ls 2d . 2 a M .n M de t M a hto h ed t eh s m 2 o .a 2 dt .s ic M al at M h 2e o .m 2 d.e a M ls tic a aa t n l h d M e m M oa d ett e ic h ls a o l d an s M do M dee ls t h ao n dd s Meth2 od .2 s. Mathematical Models and Methods the calculations and diagrams easier to verify. 2.2. Mathematical Models and Method 2s. 2. Mathematical Models and Methods 2.2. Mathematical M 2o .2 d.e M ls a at n hd e m M aettic ha ol dM s odels and Methods As it was mentioned in th A es In A it tr s w o ia t d s w uc m ati s e o n m n ti e ,o n th n ti e e o d c n o ie n n d ti th in n e A uo th In s e u itr t s In o w -ti d tr a m uc s o e d m A ti uc IM o e sn n i ti t P ti , o w th P o n n a e s , e s to A th c d m o s c e i n h ie n c t ti a n o w th s n ti n ti uo a o e ti cs n n In m u m e uo s d tr o e - d ti o u n in d e m s ti - luc th o ti en m e IM ti e In o d e n P IM tr i,n P o th th P s dto e P uc e c c s In o ti h to n a o tr c ti s n h o ti n ,d a c uo th s A uc m ti e s u c ti o c i s m t d o o -ti w n n eo m l,ti a d th n s e e uo l m e IM c e u o n P s n ti - P ti ti o s n m n to uo e ec d IM h u ia s n- s P ti ti th P m ce s m e to In IM oc tr d h o e P ad ls P ti uc scto ti m o ch o nd a , s th eti l e c c m oo nd tie nluo us-time IMPP stochastic model In each of the two examples that are presented in this subsection, the speciﬁc PHCD is As it was mentioned ih na th d e b In eetr n o cd huc oA se ti s n o i n t fo ,w r th acs e h m c ao r h e a n a n c ti d ti te h n o a b r uo n id e z e e i u b d n n s e g - ic e n ti h n th m th o c e s e h e e a o IM n In ls o f e tr n o P n g r o P - f d c th o sh uc to ra e c r - ti ch r a h h o o c a a a n ute te r s d ,a ti r th ci b c te z p e e m ir e l n c a in o g z o c d i e n c h n th e m h ti a g lo e d n e th s uo n a e b l t e n e o u a e n a fn s n o l h g -o d ti r - a c n th d m h c o g h o c e - b e c a s - th e r ur e r IM e o a n en ute c r - P r f e te c o o n P h r r ute c i p o s e z cto ls i h a o e n a c p c n f g e r h l T a a m f a th c o c S se te e r s ti en m c r c a t i h e l m z a o a n in n r n o t a d g g d a c -e n o te th th ld c re c e io ur z -a r c il o n c r o h ute e ur g n a n d g th r c - e e p b th e n e o l c a a e e f e l c - n o T r e o o n m S c fute g h sT e - o n th S s t s e p e a n l-a n r f c o d o e ute r m o c ce h cp n ur al t r aa r a ce c n en te d m c re o e i z n c o ic t n fur a g T n r S th d e sn e o c c a ec l ur o of n rg T e-S n th s c e e -o ro f ute TSs placement and occurrence of TSs As it was mentione A ds iin t w tha es In m tr eo nd tiuc onti eo dn i,n th th ee c In on tr tio n d uo uc u ti so -ti nm , th e e IM co P nP ti s nto uo ch ua s- sti tim c m e IM ode PlP applied stochasto tic a m RET odeltransition that is homologous to the one, the PHCD has been tuned to. had been chosen fo hr a h d c ah d b a e b re a e n c e te n ch rc io h zs io e n s n g e n fth o fr e o c r a h lco a h r n a a g r ca -te th jc ote rie in z - rtl r i iz o n y iute g n f g o th h rth p a e th d le a a e lc b a o e e p ln m o eur g n n e - g th n p c-h t o th eo s a -e e n s r o e -d o r n ute o jfo o ute fR i c o n c E jr p o tl ur T lc i p y a n h r l C ctl e a f a e o D n c r y m r a e c m c f e i e th o n te n o e r e t th r n f th i a p t z T e n i ur e a S n d p n s p g p r d o ur e o th c s o s c e p e c e ur n o c o a t ur s r lf e s e o j r tud R o n n o eic E g f n n e T - c R tl y th e o .E y C fe o F T D f T -f o o r C S r o T r i s j n D ute S a o th s i th p n ien r tl p e o p y th lf ur p a o f c r e un o p e e j o p r s m o ie d r th n se n e e tl n s tr t e e y o t s e n p ftud a a f t ur o R n tm s r d E p tud y th T e o o .n s c e F C t y e c o D p ur o .o rur f F f r a io th n e R p r p n E e o th r a c s T o e e e p f o C o r p o f o un D r f f T e o R s d i S un n e E s n tr T th d t eC s e a tr tud D tm jp e o r a iie e n n tm y n s tl .e th t y F e n o o n e t ff r t o s p th r tud o a r f e th e p s th r e y eo n .e p f t F o ur o s un tud rp d a o s y p tr e.r e o o F a f f o tm o r R un a E e T n d pt rC tr o oD f e fo a th un i tm n e d th e n tr e t e p o ar ftm eth se e e n n t t s otud f th y e. For a profound treatment of the First, let the ego-car start its virtual trip from the left. This case is described in detail mathematical theory of Poism soa n t h pe rm oca eti sc sa es l ,th se ee o r [y 39 o ].f Poisson processes m , a se th ee [m 39a ]ti . cal theory of Poisson pm roa cte h se sm esa , ti se ce a l[ 3 th 9e ].o ry of Poisson processes, see [39]. jointly for the purp joo jio n si e tl n y o tl fy f o R frE o th r T th e C e p D ur p iur n p o th pso e es e p o f ro e R fs E e Rn T E t T C s tud D CD in y jio .n th iF n th o e tl re p y a rp f e o p r se r r e s o n th e ft o n es un t tud p sur tud d m y p a tr .ty o e h F .s a e o e F tm m r o o a a r f e ti n a p R c t r E p a o o T lrf fo th o C th fun o e D e un od rin iy n d tr o Example th e tr fa e e P tm a o p m tm ie r sa e n se s to t h n en o No. n e t f m t p o th r s fa o tud th ti 1. e c m c e e a a s y ls t .h e th F s e,o e m s o re a ra e ti y p c [o 3 a rf9 l o P ]th f.o oe un is osrd o yn tr o p f e r a P o tm o cie se s sn s o e t n so ,p f s r e th o ec e [ e 3 s9 s] e.s , see [39]. In the chosen stochastic apprIo na th che , th che o T se Sn d sa to ta c h lo ag ss ti c a ra ep s p er eo na a ch s ,r th eae li z Ta St id I o n a n ta th s o l eo f c g a h s n o a I sr M e en P s s e P te .o n c h aa s s rte ic a la iz pa p tr io on ac s h o,f I th n a n th e I T e M S c P h dP o as .ta e n lo sg to s c ah re a sste ic e n a p ap s r ro ea ac lih z,a th tio en T sS o d f a ata n IlM og P sP a.r e seen as realizations of an IMPP. mathematical theom ry m a o ta h ft e h P m e om a isti s ac o ti a n c l a p th l rth e oo cee ro y sr s y o e f so ,P fs o P eie o s s i[s o 3s9 n o ] n p . rp orco m ec se a ss t eh s se e , s m s,e s a ee ti [ ec 3 a [93 l] 9 .th ]. eory Io nf th Po ei c ss ho on se p nr o stc o ecsh sa es st , is ce a ep [ Ip 3 nr 9 o th ].a e c h c,h th os I e e n n T th S s t e d o c a ch ta ho a s lso e tig n cs s a a tp o rp e c h rso a ea s ec tn ih c a , a s th p re p e a r T o lS ia z c d ah t a i,ta o th n ls e o o g Tf sS a a d n re a I ta M se l P e o n P g .a s sa r re ea sle iz ea nt ia o sn rse o al fi z aa nt i Io M nP s P o.f an IMPP. 3.1.1. Example No. 1: A Dt! Res Change Detector Applied to a Homologous RET Transition In the chosen stochasti T ch a ep C pr D o a m ch e,th th I o n e d th T us S e e d cd h a ta o cs o e lm o T ng h m s s e t T o o aC h r n ce e h D l y a s C e s m iD t e n ie n c c th m a a op s o e n p th d r jun e r o a us ol d a c ie t z c ius d h a ot , n ic e th o o d w n m e s c i T T t o m o h h S m f e o P d a m n C o n a li y o ta D I s n M s il o l n m o y P n T g e c P ih s o th p n . e n a r o c o r jC un o d e cn e D s us s c je un s t T e m i e e o n h s d c e n e t a th i ic s s o C w o o r n th D m id e ta w e h m m us l ic iP z t o um e e h o a n th d t ilP i s y o o u s c o d n o o liia n n s s m us t s o ic o p m v f o e n r e a n o d o n p n c j un c e r lIo y o s M s m c ci e t e P n is m s o P s c in o .e s o s n n w th lijy s un ie t T h th ic n c h um P e te c i o o o c C in um n u sD s jlw un a o t u n m iitc v l h a p e te i t th r o P io v n o o c e i d e w ss sus s io t eh n s e d P p is o r c o th io s cm s e eo sm c n sum e o p sn rio u lsy c lth a e it n si e s v e c c e s o um n is jun u th lc a ett iic o v um n e w uilta h t iv Pe o isson processes is the cumulative In the chosen stochIa ns t th ice a c p h p ors o ea n c h st,o th ch ea T sS ti c d a ap ta p lro og as c h a,r e th se e e Tn S a d sa r ta ea llo iz ga st a io re n s s e oe fn a n as I M rea PlP iz . ations of an IMPP. The CD method used coms m um on l(y C U in T S h c U o eM n C jun )D m c m e tith o eth n o d o w .d iD s tus h um e tP e s ad o um il (i e C s cd s o U o ( m e C S nx U m U p pM S o o ro U n s)i c l t M y e m io s ) is e n n e th m s s c o o e o ith s d fn .s th j o In un u D d e c s e .the h um c c tD t a um im i o e l TS n e t e (a d C u th i w l lU sequence e e a o ix d t S t d ih p s U v se um o e P M x c s p a o i) t n i o i ( so m s C s b i n o t U given e e s in s th o S f um o n o p U o fs u r d M s o n o .u (in c d C f )D e c s h m U s ithe e u n s t S e e m c a [ U s th h i4 e l top i M 0 e th o m sd , d 4 ) th o e 1 .e band th m d ]D e x .s p o e e B cc th o d t um y a a s s o n ii of lt c d e i u b a o d .T l n e n a D able e t s f b i e x o v e o t p u a e fo f n i 2 o l s s d e ,u u id t the c n i i o n h s d en um x [ m intended s i4 p n 0 e o o ,th [ s ( f 4 4 C i1 s t 0 o i U ] u ,o d .4 c S n B s 1 h U s ] y c . RET m M a o B n fe y ) s th b m u e transition o c e h d fth o sm u o c n d e ad th n . D io b n e d e (i.e., t [ s a 4 f o i0 c lu a ,e 4 n n d 1 d ]b e . x e iB n p f y o o [ 4 s ui 0 n t,i4 d o1 n i ]n s . B o [4 y f 0 s,u 41 ch ]. m By e thods can be found in [40,41]. By The CD method us Te h d e c C oD m m mo en th ly o d in us co ed nj un com ctim on o n w ly it h in P co oin ss jun onc p tir oo nc e w ss it eh s P iso th iss eo c num pro uc la et sis v ee s is the cumulative sum (CUSUM) method. De at sa sium ledi n eg x sp um th os e i ( t v C ia oU ln id s S i U o ty f M s o) u f c m ta h hs e e s m th um I aM e s os th d P um in .o P D g d m i s e th n to g c a e d a i th lv n e el a e d — b l iv e e d a a x i f t the lt o p iy ld u e o o a in st Dt s i fy d t tt i o h o w i! a n f n es i t s th s I [h Res) M 4 um o e 0 r f P ,e I 4 M s i s P 1 n u p occurs ]P m g c e . h P c B a th o t s y m m d t e s o e um e o v lth th — at d ae li e o i a about n ld d — t c g a io s lt s e a th y n s c a t um s a o s e lin t e f the d v a w t e ib a h s n ri e l te e g th i path-length w d d f Ith o M i i r t R u th e y e P E n s o v p P T r d f e a e s m s t lc i i h a n p t d o n e e ti [ d o d c t 4 Iy e t M of 0 th lt ,o — 4 o P e 2.2 f1 P th a t c ]h .t o m e km e B ln e c o y I s a o M i d s d n (i.e., e tP e s lw r — i P d e a i d m th e a s having r t sR o e um l rd e d E ea s e T R s p l is — n t E e a w g T c cover a n t s th itd th t a o l e e n th a rv d ed e sa s e t lp w ic eleven d e o ic ith n tt y st io r o d e fth e s dashes t r p h e e e e d c c t o IR M t n E o s P T ith d P se e a m r n c eo o d dd n R e si lE — dT es a r e ta d ln e d R as E t T w s ia th n d re spect to the considered RETs and sum (CUSUM) mes th um od .( C D U et Sa U ilM ed) e m xp eth oso itd io . n D se o ta fi lse u d c h e x m po eth sito io dn s sc o an f sb u ec h fo m un ed th io nd [s4 0 ca ,4 n 1 ]b . eB f y o und in [40,41]. By assuming the valida is ta y ss um o sum f tih ni e g n Ig th M th e P v e P a v m la id l o iid d ty e it ly o — fo a tfh t t e lh e I e c a M o s In M tP s w P i P d iP m th e r m o e a rd d e s oe s s d um p l T — ee S lc — s a i t — n tt a g l o t f e o l a th th e rs a e t e d s w v c t e o s a w ith c n lcir io s d th i ib r n id e t c i s y r e n s o ie r p d g n o s e e e s p f d a c r it e d e t n h R c d e t d e t E o r te T I T c o th M d h S s th e s a P a T — rn c e P S a o d c f s c m n o — t o fr e s rn o r if om d s i d d o z ie e e r d ir n s le — e d the c g r d r e c e i a s T o b d R t c n left i S r E ln R s e iT g b p iE a d i s l s T in n a a e ta s c n g rw n e e the d ad d m a in c th n c o d e T h d n band). n r S a s ts e c s r is h — d a p a c a ee t f n r re o c c a e d r t o r The d ci t t n z o d o e T is c r e n ith S c id s g pr z u s c e i e — r r n T r esence i r c e g b S e f o d o in n n T p r c T s g S le i d a S s d a c p e s of e e n a — s lr m a c l d e o c r such f d e i n e o c b n m r g h R i ts n a d e E g r n e a T a a s ts c a s n RET c t n d a r e a d in rb n o i d z id c c n transition ih c n g o u a g c r r a a c rT n c e u co d S n tre n r cp r c e s e ih in ls z d a a in c ic e a n r e e r la g s o the m e c n d a t T e e g l S o n r T synthetic i n ts S z p g i s ln — a ag n ce f d o T m r S o e c d n p sequence c e ts l u s ar c cr a r e e i n m b n d ic e n e n o g sts c c a a u n la o r d n r n d e g cn h o ca c er c sa u c a rtlr e o e rn n iz g cie n sg a T lo Sn g p lacements and occurrences along routes within and between u rr ob ua tn es e w nv itih ro in n m an e was d n tb se , secur tw our e eta ed n su k by r b na a having n rr o en w v eid rinserted o d n ro m ow u et n n e ts ts o into ,w o a iur d th a it i p n ta ﬁve t s a a k n s d n u TSs a b ir te - rtw o that w ee ed n ar d u eo rmor w ban n r e te o on characteristic u a v td e ira so p w n t m ia th esn iu n ti sta ,- n of od ur the b ta etw s Res ke n en a ar r u r eas o rb w ae n d e d no vw iro nn tm o e an dta sp , t oa ur s u ta it s-k narrowed down to adapt a suit- considered TSs—fo co rc n o d sn e id s sc ie d rrie e b r d ie n d T g S T a sS n — s d — f o cfr h o a d rr e d asc e cts r ec ir b riii z b n ii n g n g g a n T ad S n d c p co h lc a n a h c s re a a id m r ca e te e c rt n r ee id ts z r i i z n T aig S n n s d g T — r S o T o f c S o p u c rl t u p a ed r l c sr a e ee w c s m n e cim c r e ti h e n b e s its in n n a ts g a l ao n n a n a d d n n g d d b o e co c tw ch c u c a r e u r r e a r e n r r c n e o tc u e n u e r r c t s ib e e z a s s a i n n lw a o g e ln i o n tT g r n h v S o g iin u r p o ta e ln a n sm c d w e e m b in te e h ttw s n i,n ts o e a ur e a n n n d ta d u b r so b e kc tw a c n n u a ee r r e r n rn e o v n w u ic rr o e eb d s n a m a n dl o e o en w n ntg v n s i , rto o our n a m d ta e an s pk t t s a ,n o a sr ur u rio t - ta ws ek d n da orw ro n w te od a d do ap wt na t s ou a it d -apt a suit- able CUSUM method for a th bl ee a p b C u le U r p C So U U sM e S,U a m M ne d th m vo e a than th d li d f oo a d t r of e ft o a h ithe tb r e w l te p hi u Dt e C th rU p p r ar u o S er U s a eas—see p a e lM io ,b s a s lt e i e n m c ,C d T a e U n v th Sd a S a d the o lU b iv d a d lM ta a e a flr o t i .C ightmost e d m r U a ite tt h S e th w e U i o p i tM t d w u h r f m i rp o column te h o r e a s th lt rie h e s,o a e t i a d l c p in s f u T d t of o ir S c r p v T d T t a o h able a l S sie ta e d d ,p a .a a t u ta 1 e n r—in d p i . ta o v w b sa le i e lconsecut ,t i h d C a a n r U te d e S a v U i lt ia s M w l tion i id c i m t a T hte and S e rth e d it a o a lw i d ta rs easonably it.f ti o h cr T r t e S h a e d li p s atta u ic r . p To Ss d e,a a ta n.d validate it with realistic TS data. routes within and r bo e ru tw otu ee t se e w s n w iu th i rti b h n a i n n a n e ad n n v d bi e r b tw o en tw e m ee e n e n n u ts ru ,b r o a b ur n a n e ta n ev s r n o k iv ru o in r tn a e om r s n r m w o en w i et tn e h s,t d i s n o ,d ur o a o n ur w ta d n ta s b k t e s o tw k n a a n d e ra r e a r o n p rw o t u w e a rd b e s a d u d n io d t - e w o n w n v i n tr o o t o n a d m aa d e p a n t p ts a t , a s o u ur siu t- ita t- sk narrowed down to adapt a suit- close to each other starting from this path-length along the virtual route. able CUSUM method for the puIt rpw oa s ae b s,l e o a n ur C d U iv n Sa U te liM n dti a m t o en e i t th to w o a id td It h fo o r w p r e It t a a th l a s w in e so t a d p iur s c u v o T r a iur p S n li o te d d s ia a n n eta te ,ti te a o . n n a n ti d c o to o v n n a a ti lto i d n d o It uo a a p td w e t us o i a a p tn s -t w ti d o a m i ur n t v h d e a It l i rv i n v e w d a te a aa r l l a i i i te n s a s d ti t n a o ia o c t te ur It n c o T o f a w S to i n n th c d a ti te o a s a e n n d ta n o uo C ti o ti ur .p n U o us t uo n S ia n U -to n ti us te M d m n a - ti v e ti d a m o o v ln p i a e d t r to v i a a a te a n n a rd t id a a o o v n c fp a t o th l t n o id a ti fe n a th n C d te uo It e U v a C us S w a c U lU i a o -d M ti S s n a U m ti o te ur n M e uo av ic a n us o r te in a -n ti ti nti n m t o uo o e n f v us th to a- r ea ti i a C d m n o U t e p S o t v U fa a th M n rid a e n v C t a U o lifS d th U ate M e C a U co Sn U ti M nuo us-time variant of the CUSUM able CUSUM metha ob d l ef o C rU th Se U p M u r m po eth seo , a dn fd o rv ta h li ed p au ter p ito w se i,t h a n re da v lia sltiid c a T tS e i d t a w ta i.t h realistic TS data. According to the CUSUM-based RET CD method derived in Section 2.2, functions It was our intention to m a ed th oo pd t a fn or d C v It D alw .i d F a a ur ste o t ur h a e c ro im m n nte e o tith m r n n e ti uo o e , o th d tr n us o a f to o d d -ti r e f a m -C o o dr e f D o f C v p .i a s t F D ra ur s i .n a o F t n d ug h ur t e v o h r ta f h m t l m i e th b d o re e m a r et th e te w C o , o r U tr e a e d e a S ,c n d o U ftr o n e m tM a r h -ti d o e e C n f e th f uo D f - a o o i.l s f d us s f F e s m ur if o -s o ti a ug e r l ts m a th h o C r h e e ug m o r t D v d m b .h a r e o f r t a F o t i r te ur w a b e rn e , e C ta t t tr h e w n D o n e a d e f r .d e m tth F h e tn h - ur o e e o e t r f h f C e tf a h ,e U li e s tr s f r S e a m a U s ld a o so M le e ug a r -r e o am ,h lf a ft tr rr ib m a m s ad e te s t e e r w o th - aa o ug te eo n fe f d d h n a i t n s f tt o h b d h sr e e e o t t C ug h w fa e D e lh s e . t e n F b ur a te lh a ttw e r h m e e fa r e m ln rsa e o te t h ra e e la ,a n f r tr a d m a l s d te r h e a e - a te o la f fr a m n isd r s a o th te ug e a hn t d b e th tw e een the false alarm rate and the It was our intentioIt n w toa a s d oo ur pt ia nn te d n v tia oln id to ate a d ao cp ot na tin n d uo vus ali-d tia m tee a v c ao ri n ati nn t uo of us th- eti C m U eS v U aM ria nt of the CUSUM g and h are to be used for the purpose of CD. These functions are now denoted by method for CD. Fm ur m e tth h ee th o rd m o d o fo rfr e o , r C tr D C aD .d F e . ur -F our ftfh te ih s r e m s ro m oug e ro e xr ,h p etr t e , c b a tr te d eatd e d w m - e o d e - e fe e o f th te n fif o s c t iti d h s s o e of s n ug o o f a r ug l l a h s C g t e h D b a t a .s e e l b s t a F x e w o r ur p tm c e w e e ix t e a c e h p n te te re e a e n d d r te t cm h te t d w e h a o d ei n e r fte th a e d d f l,c e a sti tt tr te e lh h s o a e e c e a n d ti lT a a e o ll S r T a -n a m o g r s f l m e f a a r q s g i a ue s s r e te o a a x s s c te n p T o s i a c a e ug o n e a c te c d s te n ih d a d a d t t te n h w b td d e d e h i e e x th e t to w te p w e ic te p th h c ti e r te e o n o td n T h v e t i S e h x l d d a p e T e s e g e e te S f h c q a a c te s i ue s lti n e s s d e t q o o n s ue n c d a cf i e l e a o la n s a te te rr c g a m c c d e n ti h a s d w s o o r a s a o n n ito o te th s d c l- a i p a a to t gh r te n o a e d p d v s T r s itw o d o S hv e c e is th i i a h d ete q iet n ue h d h te e s in w x n f T p c o tie S s e th r s c fs c te o a e t h r h n q d o e d ue c o h d T s to o n e -S te o c p e s sc - s e rti o q a o v ue n n id d n l e a c to g e h si p a n a s rts n o so v d f c i o i d to a re te c p h h d r io o n w o v ts s i i - th d fo e r th h c ie n h t o T so S f s o s -e r q cue hon oc se -s and to provide hints for choos- Dt!Res Dt!Res g and h , respectively. The lower and upper indices indicate the function’s ing appropriate thresholds fio nrg th ap e p cr ho ap nr g ie a te d e th terce to sh ro s.l ds for the change in d ge a te p cp to ro rs p .r iate thresholds for the in cg h a an pg pe r o dp er te ia cte to th rs.r esholds for the change detectors. expected detection e x le a p x g e p c a e te s cs te d o d c di e a d te te ete c dti cw o tin i o th n la l g ta h g a e s a T ss o S sc o s ia c ei te q aue te d d n w cw ie e th x s ith p a te h n c te d h te e T to d S T d p S se e rs te q o eue v c qi ti ue d n io n e c n n g e h c s li a e a n a p s g t n p s a a d r n f s o o d s to p r o to r c c i p ia h a rp te o te or o v o d th s iv d -w r ie e d is e h th h ih o n tilth n d se t isf n s o T f g f o r S o r a c r s p th h e c p o q h e r o ue o o c so p h i -n n s r a - c g in a e g a te se p a th p d nr e d r o te e p to sc r h to i o a pl r te r d so . s th v fio d re r es th h ho ie n l d c ts h s a ffo n or g r e c th h d e o e o c te s h- c ato ng rs e. detectors. l l dependence on path-length l and the RET transition actually monitored by the change ing appropriate thirn eig s n h g a o p a ld p pr s p o f rp o or r p i th a ri te a ete th ch th ra en r se g hs e o h ld o de ls d te fso c fto ro th r r s th e . c eh c ia n h n g ag n a e g p d e p e d rte o ep te cto rc ia to rte sr . s th . resholds for the change detectors. detector, respectively. These functions are given in Equations (12) and (13), respectively. In the former, the coefﬁcients are given with the same numerical precision as was used in Table 1. Dt!Res 1 NS g 2.80 km l 1.74 N (l) (12) PL GW SL 1.92 N (l) + 0.13 N (l) + 0.69 N (l) Dt!Res Dt!Res Dt!Res h = g infg . (13) l l sl For the generation of these functions, the actual counts of the four considered TS types—at a particular path-length—are required. These counts are given in the middle band of Table 2 for virtual trips starting from the left. Dt!Res Dt!Res In Figure 1, functions g and h have been plotted, respectively, for the TS !l !l sequence given in Table 2. The virtual trip in this case had started from the left, as indicated by “!” in the lower indices. Appl. Sci. 2021, 11, x FOR PEER REVIEW 12 of 18 Appl. Sci. 2021, 11, 3666 11 of 17 (a) (b) Figur Figure e 1. A1. do A w downtown ntown (Dt(Dt) ) → ! res residential idential (R (Res) es) cchange hange detector detectorapplied applied to to a homologous a homologor u oad s ro envir ad en onment vironm type ent (RET) type (RET) Dt!Res DtRes transition showing up in the trafﬁc sign (TS) sequence given in the top band of Table 2: (a) Function g for the transition showing up in the traffic sign (TS) sequence given in the top band of Table 2: (a) Function l g for the Dt!Res considered TS sequence and (b) Function h . !l DtRes considered TS sequence and (b) Function h . l Dt!Res According to the diagram of h in Figure 1, a threshold, say, d = 3.0 could be !l selected to detect the RET change. With this threshold, the change is detected at about the DtRes According to the diagram of in Figure 1, a threshold, say,   3.0 could be path-length 2.7 km, i.e., with a detection lag of 0.5 km. If smaller thresholds, e.g., d = 1.0 or l d = 2.0 are used instead, then 3 and 1 false alarms will occur, respectively, along the 2.2 km selected to detect the RET change. With this threshold, the change is detected at about the of the Dt route. If, on the other hand, larger thresholds are used, such as d = 4.0 or d = 5.0, path-length 2.7 km, i.e., with a detection lag of 0.5 km. If smaller thresholds, e.g.,   1.0 then unnecessary extra detection lags of 0.3 and 0.6 km will occur. or   2.0 are used instead, then 3 and 1 false alarms will occur, respectively, along the 2.2 km of the Dt route. If, on the other hand, larger thresholds are used, such as   4.0 or 3.1.2. Example No. 2: A Res! Dt Change Detector Applied to a Homologous RET Transition   5.0 , then unnecessary extra detection lags of 0.3 and 0.6 km will occur. To make full use of the synthetic TS sequence given in the top band of Table 2 and to provide a deeper insight into the proposed CD method, let now the ego-car start its virtual 3.1journey .2. Exam frp om le the Noright. . 2: A Res → Dt Change Detector Applied to a Homologous RET Tran- As a side note, we can reverse such a virtual journey fairly easily in a table. In real life sition and in real trafﬁc, however, it would be much more problematic as one might encounter To make full use of the synthetic TS sequence given in the top band of Table 2 and to very different TSs on the way back (if at all the reversed route is permitted by the TSs provide a deeper insight into the proposed CD method, let now the ego-car start its vir- installed). Furthermore, the TSs that are facing us now are located on the opposite side of tual journey from the right. the road as in the ﬁrst journey. However, even in case of this wieldy virtual journey, some As a side note, we can reverse such a virtual journey fairly easily in a table. In real modiﬁcations in the table, namely, with respect to the TS counts, are necessary. life and These in re new al trTS affi counts, c, howi.e ev., er the , it counts would for beright-to-left much mor virtual e probtrips, lemati ar ce a given s onein m the ight en- bottom band of the table. In this reversed case, the ego-car is driven from an intended Res counter very different TSs on the way back (if at all the reversed route is permitted by the area to an intended Dt area, i.e., a Res ! Dt transition is expected. TSs installed). Furthermore, the TSs that are facing us now are located on the opposite Again, as in the ﬁrst example, we wish to form functions g and h that signal the T T side of the road as in the first journey. However, even in case of this wieldy virtual jour- RET transition that is actually expected. Accordingly, these functions are now denoted by ney, some modifications in the table, namely, with respect to the TS counts, are necessary. Res!Dt Res!Dt g and h , respectively. l l These new TS counts, i.e., the counts for right-to-left virtual trips, are given in the In order to detect such a transition, the roles of the process parameters l and m —given in k k bottom band of the table. In this reversed case, the ego-car is driven from an intended Res the fourth and the ﬁfth column of Table 1, respectively—need to be swapped in Equations area to an intended Dt area, i.e., a Res → Dt transition is expected. (8) and (9), as now, it is supposed that the ﬁrst road environment is a Res area, rather than a Dt area. Again, the uniﬁed counts are to be used, namely, the TS counts given in Again, as in the first example, we wish to form functions g and h that signal T T bottom band. the RET transition that is actually expected. Accordingly, these functions are now de- Res!Dt Res!Dt Functions g and h are given in Equations (14) and (15), respectively. l l ResDt ResDt noted by g and h , respectively. l l Res!Dt NS g 2.80 km l + 1.74 N (l)+ In order to detect such a transition, the roles of the process parameters  and (14) PL GW SL +1.92 N (l) 0.13 N (l) 0.69 N (l)  —given in the fourth and the fifth column of Table 1, respectively—need to be Res!Dt Res!Dt Res!Dt h = g infg . (15) l l s swapped in Equations (8) and (9), as now, it is supposed that the first road environment sl is a Res area, rather than a Dt area. Again, the unified counts are to be used, namely, the TS counts given in bottom band. ResDt ResDt Functions g and h are given in Equations (14) and (15), respectively. l l ResDt 1 NS g 2.80 km l 1.74N l    l Appl. Sci. 2021, 11, x FOR PEER REVIEW 13 of 18 PL GW SL 1.92N l 0.13N l 0.69N l (14) Appl. Sci. 2021, 11, 3666 12 of 17 ResDt ResDt ResDt h  g  inf g (15) l l s sl . ResDt ResDt In Figure 2, the functions g and h have been plotted. The corre- l l Res!Dt Res!Dt In Figure 2, the functions g and h have been plotted. The corresponding l l sponding virtual journey had started from the right in the band, as indicated by the “ ” virtual journey had started from the right in the band, as indicated by the “ ” in the lower in the lower indices. The diagrams presented herein corresponding to virtual journeys indices. The diagrams presented herein corresponding to virtual journeys from the right from the right are presented in green. are presented in green. (a) (b) Figure 2. A Res → Dt change detector applied to a homologous RET transition appearing in the TS sequence given in the Figure 2. A Res ! Dt change detector applied to a homologous RET transition appearing in the TS sequence given in the ResDt ResDt Res!Dt Res!Dt top topband bandof of TTa able ble 22 : :( a (a ) )Function Functiong g for fo the r tconsider he consied derTS ed sequence TS sequen and ce a(n bd ) Function (b) Functih on h . . l l l l Again, threshold d = 3.0 would be an appropriate choice based on the diagram in Again, threshold   3.0 would be an appropriate choice based on the diagram in Figure 2. With this threshold, the RET change is detected at about the path-length 3 km. Figure 2. With this threshold, the RET change is detected at about the path-length 3 km. Since the ﬁrst signs of the Dt area begin to appear at the path-length 2.8 km, the detection Since the first signs of the Dt area begin to appear at the path-length 2.8 km, the detection lag is 0.2 km long. lag is 0.2 km long. Dt!Res Res!Dt Comparing the diagrams of g and g shown in Figures 1a and 2a, respec- !l l DtRes ResDt g g Comparing the diagrams of and shown in Figure 1.a and Figure tively, one notices the symmetry between these. This symmetry can be traced back to two l l facts: 2.a, re ﬁrst, specti frv om elyEquations , one notic(8) es th and e sy (9)—as mmetrused y betw ineExample en these. 1, Th and is sy by mm swapping etry can the be tr roles aced of parameters l and m for equations corresponding to Example 2—furthermore, from back to two facts: first, from Equations (8) and (9)—as used in Example 1, and by swap- k k Equations (12) and (14), it follows that ping the roles of parameters  and  for equations corresponding to Example k k 2—furthermore, from Equations (12) and (14), it follows that Res!Dt Dt!Res g 1.0 g . l l ResDt DtRes g 1.0g . l l and second, to the fact that the same route—with its TSs—was covered from opposite and second, to the fact that the same route—with its TSs—was covered from opposite directions. ...!... directions. There will be further symmetries perceptible between the respective g diagrams. ...l... ...... ...!... However, similar symmetries do not show up amongst the respective h functions. There will be further symmetries perceptible between the respective g dia- ...l... ...l ... ...... grams. However, similar symmetries do not show up amongst the respective h 3.2. Further Examples ...l ... functions. In each of the two examples below, an “off-the-tune” RET change detector is consid- ered. These change detectors are applied to the TS sequence used above, furthermore, 3.2. Further Examples the same marked Poisson process reference parameters are used. The examples below In each of the two examples below, an “off-the-tune” RET change detector is consid- demonstrate that “off-the-tune” change detectors may behave fairly haphazardly. In [17], appr eredoaches . These ar che an pr goposed e detecto to rs deal are a with ppliesuch d to tbehavior he TS seqof uethe nce change used abdetectors. ove, furtheThese rmore, ap- the pr saoaches me mar ar ke ed essential Poisson if pr several ocess re dif fer fer enently ce par tuned ametechange rs are udetectors sed. The ear xa em to plbe es used below within demon a- compound system, e.g., for the purpose of RET detection and identiﬁcation, rather than CD. strate that “off-the-tune” change detectors may behave fairly haphazardly. In [17], ap- proaches are proposed to deal with such behavior of the change detectors. These approaches 3.2.1. Example No. 3: A Res ! Dt Change Detector Applied to a Dt ! Res Transition Res!Dt Res!Dt Functions g and h to be used for signaling a RET transition Res ! Dt l l have already been presented—in conjunction with the Example No. 2—in Equations (14) and (15), respectively. The only difference is that now we need to use these functions with the TS counts given in the middle band of Table 2, rather than with the TS counts given Appl. Sci. 2021, 11, x FOR PEER REVIEW 14 of 18 are essential if several differently tuned change detectors are to be used within a compound system, e.g., for the purpose of RET detection and identification, rather than CD. 3.2.1. Example No. 3: A Res → Dt Change Detector Applied to a Dt → Res Transition ResDt ResDt Functions g and h to be used for signaling a RET transition Res → Dt l l have already been presented—in conjunction with the Example No. 2—in Equations (14) and (15), respectively. The only difference is that now we need to use these functions with the TS counts given in the middle band of Table 2, rather than with the TS counts given in the bottom band, as the car is now driven from the left to the right (i.e., a Dt → Res transition is expected). ResDt ResDt In Figure 3, the diagrams of g and h are shown for an intended RET l l transition Dt → Res, respectively. The detector takes the first 0.8 km starting from the left for a Res area and then detects a change Res → Dt at that point (e.g., with a threshold   1.0 ). If threshold   3.0 is used instead, then the RET transition will be detected at 1.0 km. The corresponding detection lags for these two thresholds are 0.8 and 1.0 km, Appl. Sci. 2021, 11, 3666 13 of 17 respectively. If we use larger thresholds, say,   4.0 , or   5.0 , then for the former, an extra detection lag of 0.9 km will be introduced, while the latter will completely miss the Dt segment of the route. in the bottom band, as the car is now driven from the left to the right (i.e., a Dt ! Res The intended RET change of transition Dt → Res remains undetected by this detec- transition is expected). tor no matter what   0 is used. Res!Dt Res!Dt In Figure 3, the diagrams of g and h are shown for an intended RET !l !l transition Dt ! Res, respectively. The detector takes the ﬁrst 0.8 km starting from the left 3.2.2. Example No. 4: A Dt → Res Change Detector Applied to a Res → Dt Transition for a Res area and then detects a change Res ! Dt at that point (e.g., with a threshold DtRes In Figure 4, the diagram of function g is shown for the intended RET transi- l d = 1.0). If threshold d = 3.0 is used instead, then the RET transition will be detected at tio 1.0 n R km. es → The Dt. corr Thi espondi s transiti ng on detection is presenlags t whfor en these drivintwo g fro thr m esholds the righar t a elo 0.8 ngand the 1.0 con km, sid- ered TS sequence. Note that for the given TS sequence and for the given process param- respectively. If we use larger thresholds, say, d = 4.0, or d = 5.0, then for the former, an DtRes DtRes extra detection lag of 0.9 km will be introduced, while the latter will completely miss the eters, functions g and h happen to be identical. Using threshold   3.0 , l l Dt segment of the route. the detector signals a change Dt → Res at about the path-length of 1 km. (a) (b) Figure 3. A Res → Dt change detector applied to a Dt → Res transition showing up in the TS sequence given in the top Figure 3. A Res ! Dt change detector applied to a Dt ! Res transition showing up in the TS sequence given in the top ResDt ResDt Res!Dt Res!Dt band bandof of TTa able ble 22 : :( a (a ) ) Function Functiong g for fo the r tconsider he consid ed erTS ed sequence TS sequen and ce a( n b d ) ( Function b) Functih on h . . !ll !l l The intended RET change of transition Dt ! Res remains undetected by this detector no matter what d > 0 is used. 3.2.2. Example No. 4: A Dt ! Res Change Detector Applied to a Res ! Dt Transition Dt!Res In Figure 4, the diagram of function g is shown for the intended RET transition Res ! Dt. This transition is present when driving from the right along the considered TS sequence. Note that for the given TS sequence and for the given process parameters, Appl. Sci. 2021, 11, x FOR PEER REVIEW 15 of 18 Dt!Res Dt!Res functions g and h happen to be identical. Using threshold d = 3.0, the detector l l signals a change Dt ! Res at about the path-length of 1 km. DtRes Dt!Res Figure 4. Function g for the considered TS sequence. For the given sequence and for the given Figure 4. Function g for the considered TS sequence. For the given sequence and for the l Dt!Res Dt!Res process parameters functions g and h happen to be identical. DtRes DtRes l l given process parameters functions g and h happen to be identical. l l 4. Discussion According to the approach derived in Section 2.2, the RET changes can be detected with CUSUM change detectors, which rely on the on-the-fly minimization effected by PHCDs. In order to detect all kinds of the RET transitions between the three RETs considered herein, the simultaneous use of six differently tuned PHCDs is necessary. In Section 3.2, ...... ...... we have demonstrated what happens to the functions g and h when the ac- ...l ... ...l ... tual RET change is not what the detector is tuned to detect. In fact, in the examples given there, we have applied change detectors that were tuned to the inverse transitions. If one wanted to use the aforementioned PHCDs for the purpose of detecting not only the changes between different RETs, but also the actual RETs themselves, further- more, wished to overcome the haphazard behavior of the “off-the-tune” PHCDs, there ...... are promising possibilities; for instance, the respective functions h can be generated and considered within a sliding window, furthermore, several overlapping sliding win- dows can be used at the same time. In addition, these could be multi-scale windows. An artificial neural network (ANN) proposed for TS-based RET detection was pre- sented in [26]. The ANN-based method made use of sliding multi-scale windows, and for these windows, TS histograms were calculated. The network proposed there could well ...... be extended to input and make good use of the “summaries” of functions h , rather than the TS histograms. These summaries could be of syntactic nature. A tool capable of exploring time series data for pattern and query search tasks, as well as for generating syntactic descriptions of the time series was proposed and demonstrated in [45]. The ...... syntactic descriptions of the functions h should preferably be computed for sliding multi-scale windows. The TS-based RET change, the inferred actual RET within, and the complete inferred RET structure—i.e., a map layer, or sublayer—of an urban area could be utilized in var- ious manners in automotive applications. First, the TS-based RET change, or the TS-based actual RET could initiate warnings to novice drivers, e.g., “You are now driving in a downtown area.” What actually is meant by this warning is as follows: “The area might be uncrowded now, but in half an hour, or so it could turn very busy and could be loaded with intense car traffic. Therefore, find a parking place now, if want to stay in this area.” It also hints at reducing speed to, say, 40 km/h. In a Res area, the respective warning could, for instance, instruct the novice driver to watch out for groups of children playing on the streets. Appl. Sci. 2021, 11, 3666 14 of 17 4. Discussion According to the approach derived in Section 2.2, the RET changes can be detected with CUSUM change detectors, which rely on the on-the-ﬂy minimization effected by PHCDs. In order to detect all kinds of the RET transitions between the three RETs considered herein, the simultaneous use of six differently tuned PHCDs is necessary. In Section 3.2, we ...!... ...!... have demonstrated what happens to the functions g and h when the actual RET ...l... ...l... change is not what the detector is tuned to detect. In fact, in the examples given there, we have applied change detectors that were tuned to the inverse transitions. If one wanted to use the aforementioned PHCDs for the purpose of detecting not only the changes between different RETs, but also the actual RETs themselves, further- more, wished to overcome the haphazard behavior of the “off-the-tune” PHCDs, there are ...!... promising possibilities; for instance, the respective functions h can be generated and considered within a sliding window, furthermore, several overlapping sliding windows can be used at the same time. In addition, these could be multi-scale windows. An artiﬁcial neural network (ANN) proposed for TS-based RET detection was pre- sented in [26]. The ANN-based method made use of sliding multi-scale windows, and for these windows, TS histograms were calculated. The network proposed there could well ...!... be extended to input and make good use of the “summaries” of functions h , rather than the TS histograms. These summaries could be of syntactic nature. A tool capable of exploring time series data for pattern and query search tasks, as well as for generating syntactic descriptions of the time series was proposed and demonstrated in [45]. The ...!... syntactic descriptions of the functions h should preferably be computed for sliding multi-scale windows. The TS-based RET change, the inferred actual RET within, and the complete inferred RET structure—i.e., a map layer, or sublayer—of an urban area could be utilized in various manners in automotive applications. First, the TS-based RET change, or the TS-based actual RET could initiate warnings to novice drivers, e.g., “You are now driving in a downtown area.” What actually is meant by this warning is as follows: “The area might be uncrowded now, but in half an hour, or so it could turn very busy and could be loaded with intense car trafﬁc. Therefore, ﬁnd a parking place now, if want to stay in this area.” It also hints at reducing speed to, say, 40 km/h. In a Res area, the respective warning could, for instance, instruct the novice driver to watch out for groups of children playing on the streets. In relation to the control of smart cars, the preferred speed could be set to some lower than 50 km/h speed in the Dt area, especially during and close to the usual peak hours. The maximum acceleration and deceleration values could be set to safer values. In relation to the ADAS/AD computations carried out on-board smart cars, particu- larly to the computations related to TS detection and recognition, a speciﬁc geometrical size range for TSs can be used. In narrow streets of historical districts, often smaller TSs are installed by the road authorities, and that size should be allowed in the TS veriﬁcation phase of the computing. The detection of trafﬁc lanes and the estimation of the distances to the TSs from the ego-car are examples for computations that implicitly make use of some spatial models of the road and its environment. In Res areas—at least in our country—multi-lane roads are infrequent, therefore simpler road structures/models should be matched against the camera images of the road scenery. Concerning the road administration and management, the TS-based RET map layer compiled from data gathered through car-based data collection trips could be used to improve the match between seasonal, weekly, and daily trafﬁc patterns and the inferred RETs, thereby creating a more perceivable and more self-explaining urban environment that is hopefully also safer. Appl. Sci. 2021, 11, 3666 15 of 17 5. Conclusions The road environment appears around and sweeps past an ego-car, while it is being driven. The character of the urban road environments can be categorized into urban RETs. Abrupt changes in the character of the road environment, i.e., transitions between areas of different RETs, pose trafﬁc safety risk, especially, for drivers lacking prolonged driving experience and also for drivers of old age. The urban RET transitions per se manifest themselves in changes in trafﬁc density and in the composition of the trafﬁc. These are transient dynamic features describing an urban area, i.e., a subnetwork of an urban road network. Nonetheless, urban RET transitions manifest themselves also in changes that concern static ROs, e.g., CRs (permanent static) and conventional TSs (transient static). So, e.g., the density and the “mixture” of TSs are expected to change between areas of different RETs. As a consequence, the RET change could also be detected via monitoring static RO occurrences along the route. Herein, TS occurrences were considered only. These are noted in TS data logs. These logs can be interpreted as realizations of a continuous-variable IMPP, and the RET change can be detected—relying on this assumption—from them. CD methods, e.g., the CUSUM method, are known and widely used for “simpler” inhomogeneous Poisson processes. The mentioned method was adopted and modiﬁed for detecting change between RETs based on a TS log. The behavior of the change detector was tested on a synthetic TS sequence. Nonetheless, the sequence was used in four different ways in Examples Nos. 1–4, and some observations and conclusions were drawn from these. The presented simulation results indicate that a TS-based RET CD is feasible, and can be adopted for driver assistance, though it is not suitable for initiating an immediate intervention in critical situations. The continuous-time approach presented herein serves as a clariﬁcation of the discrete- time model and method proposed in [25], and it was not meant and it was not expected to improve for the processing and detection characteristics achieved therein. This is due to the underlying similarity between the two stochastic models, i.e., between the marked binomial and the marked Poisson models. For this reason, the precision and the delay of the RET change detection are expected to be in the same range, respectively, for both approaches for any realistic parameter-choices in the given context. Further research and development have been suggested in Section 4 and have been motivated with regard to the integration of the RET change detector into an ANN-based detector solution proposed earlier. Author Contributions: Conceptualization, Z.F. and L.G.; data curation, Z.F.; formal analysis, L.G.; funding acquisition, P.G.; investigation, Z.F.; methodology, Z.F. and L.G.; project administration, P.G.; resources, P.G.; software, Z.F.; supervision, P.G.; validation, Z.F., L.G. and P.G.; visualization, Z.F.; writing—original draft, Z.F. and L.G.; writing—review and editing, P.G. All authors have read and agreed to the published version of the manuscript. Funding: The research was supported by the Ministry of Innovation and Technology NRDI Ofﬁce within the framework of the Autonomous Systems National Laboratory Program. Institutional Review Board Statement: Not applicable. Informed Consent Statement: Not applicable. Data Availability Statement: Data used in the article are presented herein. Conﬂicts of Interest: The authors declare no conﬂict of interest. Appl. Sci. 2021, 11, 3666 16 of 17 References 1. Theeuwes, J. Self-explaining roads: Subjective categorization of road environments. In Vision in Vehicles VI; Gale, A., Brown, I.D., Taylor, S.P., Haslegrave, C.M., Eds.; Elsevier: Amsterdam, The Netherlands, 1998; pp. 279–287. 2. Charlton, S.G.; Mackie, H.W.; Baas, P.H.; Hay, K.; Menezes, M.; Dixon, C. Using endemic road features to create self-explaining roads and reduce vehicle speeds. Accid. Anal. Prev. 2010, 42, 1989–1998. [CrossRef] 3. Qin, Y.; Chen, Y.; Lin, K. Quantifying the effects of visual road information on drivers’ speed choices to promote self-explaining plaining roads. Int. J. Environ. Res. Public Health 2020, 17, 2437. [CrossRef] [PubMed] 4. Saha, D.; Dumbaugh, E.; Merlin, L.A. A conceptual framework to understand the role of built environment on trafﬁc safety. J. Saf. Res. 2020, 75, 41–50. [CrossRef] [PubMed] 5. Ding, H.; Zhao, X.; Rong, J.; Ma, J. Experimental research on the effectiveness of speed reduction markings based on driving simulation: A case study. Accid. Anal. Prev. 2013, 60, 211–218. [CrossRef] 6. Banach, M.; Długosz, R. Techniques to facilitate the use of V2I communication system as support for trafﬁc sign recognition algo- rithms. In Proceedings of the 24th International Conference on Methods and Models in Automation and Robotics, Miedzyzdroje, Poland, 26–29 August 2019; pp. 308–313. 7. Ahangar, M.N.; Ahmed, Q.Z.; Khan, F.A.; Hafeez, M. A survey of autonomous vehicles: Enabling communication technologies and challenges. Sensors 2021, 21, 706. [CrossRef] [PubMed] 8. Popescu, O.; Sha-Mohammad, S.; Abdel-Wahab, H.; Popescu, D.C.; El-Tawab, S. Automatic incident detection in intelligent transportation systems using aggregation of trafﬁc parameters collected through V2I communications. IEEE Intell. Transp. Syst. Mag. 2017, 9, 64–75. [CrossRef] 9. Sattar, F.; Karray, F.; Kamel, M.; Nassar, L.; Golestan, K. Recent advances on context-awareness and data/information fusion in ITS. Int. J. Intell. Transp. Syst. Res. 2016, 14, 1–19. [CrossRef] 10. Global Navigation Satellite System (GNSS) Devices per Capita Worldwide 2019–2029. Available online: https://www.statista. com/statistics/1174568/gnss-devices-per-capita-worldwide/#statisticContainer (accessed on 21 March 2021). 11. Reid, T.G.; Chan, B.; Goel, A.; Gunning, K.; Manning, B.; Martin, J.; Tarantino, P. Satellite navigation for the age of autonomy. In Proceedings of the 2020 IEEE/ION Position, Location and Navigation Symposium, Portland, OR, USA, 20–23 April 2020; pp. 342–352. 12. Schmidt, D.; Radke, K.; Camtepe, S.; Foo, E.; Ren, M. A survey and analysis of the GNSS spooﬁng threat and countermeasures. ACM Comput. Surv. 2016, 48, 64. [CrossRef] 13. Welzel, A.; Reisdorf, P.; Wanielik, G. Improving urban vehicle localization with trafﬁc sign recognition. In Proceedings of the 18th IEEE International Conference on Intelligent Transportation Systems, Gran Canaria, Spain, 15–18 September 2015; pp. 2728–2732. 14. The Trafﬁc Safety Problem in Urban Areas. Available online: http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.572.9 092&rep=rep1&type=pdf (accessed on 9 December 2020). 15. Wechsler, K.; Drescher, U.; Janouch, C.; Haeger, M.; Voelcker-Rehage, C.; Bock, O. Multitasking during simulated car driving: A comparison of young and older persons. Front. Psychol. 2018, 9, 910–921. [CrossRef] 16. Faure, V.; Lobjois, R.; Benguigui, N. The effects of driving environment complexity and dual tasking on drivers’ mental workload and eye blink behavior. Transp. Res. Part F Trafﬁc Psychol. Behav. 2016, 40, 78–90. [CrossRef] 17. Payyanadan, R.P.; Lee, J.D.; Grepo, L.C. Challenges for older drivers in urban, suburban, and rural settings. Geriatrics 2018, 3, 14. [CrossRef] 18. Tawari, A.; Sivaraman, S.; Trivedi, M.M.; Shannon, T.; Tippelhofer, M. Looking-in and looking-out vision for urban intelligent assistance: Estimation of driver attentive state and dynamic surround for safe merging and braking. In Proceedings of the IEEE Intelligent Vehicles Symposium, Dearborn, MI, USA, 8–11 June 2014; pp. 115–120. 19. Stapel, J.; El Hassnaoui, M.; Happee, R. Measuring driver perception: Combining eye-tracking and automated road scene perception. Hum. Factors. 2020. [CrossRef] 20. Tang, R.; Jiang, Z. Driver ’s perception model in driving assist. In Proceedings of the IEEE 20th International Conference on Software Quality, Reliability and Security Companion, Macau, China, 11–14 December 2020; pp. 237–240. 21. Rudin-Brown, C.M.; Edquist, J.; Lenne, M.G. Effects of driving experience and sensation-seeking on drivers’ adaptation to road environment complexity. Saf. Sci. 2014, 62, 121–129. [CrossRef] 22. Lin, H.; Gao, J.; Zhou, Y.; Lu, G.; Ye, M.; Zhang, C.; Liu, L.; Yang, R. Semantic decomposition and reconstruction of residential scenes from LiDAR data. ACM Trans. Graph. 2013, 32, 1–10. [CrossRef] 23. Wang, Y.; Cheng, L.; Chen, Y.; Wu, Y.; Li, M. Building point detection from vehicle-borne LiDAR data based on voxel group and horizontal hollow analysis. Remote Sens. 2016, 8, 419. [CrossRef] 24. Jeong, J.; Cho, Y.; Shin, Y.-S.; Roh, H.; Kim, A. Complex urban dataset with multi-level sensors from highly diverse urban environments. Int. J. Rob. Res. 2019, 38, 642–657. [CrossRef] 25. Fazekas, Z.; Balázs, G.; Gerencsér, L.; Gáspár, P. Inferring the actual urban road environment from trafﬁc sign data using a minimum description length approach. Transp. Res. Proc. 2017, 27, 516–523. [CrossRef] 26. Fazekas, Z.; Balázs, G.; Gáspár, P. ANN-based classiﬁcation of urban road environments from trafﬁc sign and crossroad data. Acta Polytech. Hung. 2018, 15, 83–100. 27. SAFESPOT Final Report 2010. Available online: http://www.safespot-eu.org/documents/D8.1.1_Final_Report_-_Public_v1.0 .pdf (accessed on 22 March 2021). Appl. Sci. 2021, 11, 3666 17 of 17 28. Map Features. Available online: https://wiki.openstreetmap.org/wiki/Map_features (accessed on 8 April 2021). 29. Liu, X.; Gong, L.; Gong, Y.; Liu, Y. Revealing travel patterns and city structure with taxi trip data. J. Transp. Geogr. 2015, 43, 78–90. [CrossRef] 30. Kastner, R.; Michalke, T.; Adamy, J.; Fritsch, J.; Goerick, C. Task-based environment interpretation and system architecture for next generation ADAS. IEEE Trans. Intell. Transp. Syst. 2011, 3, 20–33. [CrossRef] 31. Sindhu, O.S.S.V.; Victer, P.P. Computer vision model for trafﬁc sign recognition and detection: A survey. In Proceedings of the International Conference on Communications and Cyber Physical Engineering, Hyderabad, India, 24–25 January 2018; pp. 679–690. 32. Veronese, L.P.; Badue, C.; Cheein, F.A.; Guivant, J.; De Souza, A.F. A single sensor system for mapping in GNSS-denied environments. Cogn. Syst. Res. 2019, 56, 246–261. [CrossRef] 33. Lin, C.C.; Wang, M.S. Road sign recognition with fuzzy adaptive pre-processing models. Sensors 2012, 12, 6415–6433. [CrossRef] 34. Kim, T.; Park, T.-H. Extended Kalman Filter (EKF) design for vehicle position tracking using reliability function of Radar and Lidar. Sensors 2020, 20, 4126. [CrossRef] 35. Akhlaq, M.; Sheltami, T.R.; Helgeson, B.; Shakshuki, E.M. Designing an integrated driver assistance system using image sensors. J. Intell. Manuf. 2012, 23, 2109–2132. [CrossRef] 36. Belbachir, A. An embedded testbed architecture to evaluate autonomous car driving. Intell. Serv. Robot. 2017, 10, 109–119. [CrossRef] 37. Manual for Streets, Department for Transport and Ministry of Housing, Communities & Local Government, UK, 29 March 2007. Available online: https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/ﬁle/34 1513/pdfmanforstreets.pdf (accessed on 22 March 2021). 38. Roads 2018. Available online: https://ec.europa.eu/transport/road_safety/sites/roadsafety/ﬁles/pdf/ersosynthesis2018-roads. pdf (accessed on 22 March 2021). 39. Kingman, J.F.C. Poisson processes. In Encyclopedia of Biostatistics; Armitage, P., Colton, T., Eds.; Wiley and Sons: Hoboken, NJ, USA, 2005. 40. Poor, H.V.; Hadjiliadis, O. Quickest Detection; Cambridge University Press: Cambridge, UK, 2008. 41. Tartakovsky, A.; Nikiforov, I.; Basseville, M. Sequential Analysis: Hypothesis Testing and Change Detection; Chapman and Hall/CRC: New York, NY, USA, 2014. 42. Page, E.S. Continuous inspection schemes. Biometrika 1954, 41, 100–115. [CrossRef] 43. Hinkley, D.V. Inference about the change-point in a sequence of random variables. Biometrika 1970, 57, 1–17. [CrossRef] 44. Gerencsér, L.; Prosdocimi, C. Input-output properties of the Page-Hinkley detector. Syst. Control Lett. 2011, 60, 486–491. [CrossRef] 45. Rodrigues, J.; Folgado, D.; Belo, D.; Gamboa, H. SSTS: A syntactic tool for pattern search on time series. Inf. Process. Manag. 2019, 56, 61–76. [CrossRef]

http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png

Applied Sciences Multidisciplinary Digital Publishing Institute

http://www.deepdyve.com/lp/multidisciplinary-digital-publishing-institute/detecting-change-between-urban-road-environments-along-a-route-based-gFPnyMP1eI

Detecting Change between Urban Road Environments along a Route Based on Static Road Object Occurrences

Loading next page...

References

References for this paper are not available at this time. We will be adding them shortly, thank you for your patience.

Publisher: Multidisciplinary Digital Publishing Institute
Copyright: © 1996-2021 MDPI (Basel, Switzerland) unless otherwise stated Disclaimer The statements, opinions and data contained in the journals are solely those of the individual authors and contributors and not of the publisher and the editor(s). MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. Terms and Conditions Privacy Policy
ISSN: 2076-3417
DOI: 10.3390/app11083666
Publisher site: See Article on Publisher Site

Abstract

Journal

Applied Sciences – Multidisciplinary Digital Publishing Institute

Published: Apr 19, 2021

Keywords: marked Poisson processes; change detection methods; urban road environment detection; traffic sign detection and recognition; advanced driver assistance systems

Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Detecting Change between Urban Road Environments along a Route Based on Static Road Object Occurrences

Detecting Change between Urban Road Environments along a Route Based on Static Road Object Occurrences

Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Detecting Change between Urban Road Environments along a Route Based on Static Road Object Occurrences

Detecting Change between Urban Road Environments along a Route Based on Static Road Object Occurrences

References

Abstract

Journal

Recommended Articles

There are no references for this article.

Our policy towards the use of cookies