Access the full text.
Sign up today, get DeepDyve free for 14 days.
Hindawi Wireless Power Transfer Volume 2022, Article ID 7941978, 10 pages https://doi.org/10.1155/2022/7941978 Research Article Stage Spectrum Sensing Technique for Cognitive Radio Network Using Energy and Entropy Detection Mustefa Badri Usman , Ram Sewak Singh, and S Rajkumar Department of Electronics and Communication Engineering, School of Electrical Engineering and Computing, Adama Science and Technology University, P.O.Box:1888, Adama, Ethiopia Correspondence should be addressed to Mustefa Badri Usman; firstname.lastname@example.org Received 6 April 2022; Accepted 27 July 2022; Published 24 August 2022 Academic Editor: Jiafeng Zhou Copyright © 2022 Mustefa Badri Usman et al. )is is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. )e radio spectrum is one of the world’s most highly regulated and limited natural resources. )e number of wireless devices has increased dramatically in recent years, resulting in a scarcity of available radio spectrum due to static spectrum allocation. However, many studies on static allocation show that the licensed spectrum bands are underutilized. Cognitive radio has been considered as a viable solution to the issues of spectrum scarcity and underutilization. Spectrum sensing is an important part in cognitive radio for detecting spectrum holes. To detect the availability or unavailability of primary user signals, many spectrum sensing techniques such as matched ﬁlter detection, cyclostationary feature detection, and energy detection have been developed. Energy detection has gained signiﬁcant attention from researchers because of its ease of implementation, fast sensing time, and low computational complexity. Conventional detectors’ performance degrades rapidly at low SNR due to their sensitivity to the uncertainty of noise. To mitigate noise uncertainty, Shannon, Tsallis, Kapur, and Renyi entropy-based detection has been used in this study, and their performances are compared to choose the best performer. According to the comparison results, the Renyi entropy outperforms other entropy methods. In this study, two-stage spectrum sensing is proposed using energy detection as the coarse stage and Renyi entropy-based detection as the ﬁne stage to improve the performance of single-stage detection techniques. Furthermore, the performance comparison among conventional energy detection, entropy-based detection, and the proposed two-stage techniques over AWGN channel are performed. )e parameters such as probability of detection, false alarm probability, miss-detection probability, and receiver operating characteristics curve are used to evaluate the performance of spectrum sensing techniques. It has been shown that the proposed two-stage sensing technique outperforms single-stage energy detection and Renyi entropy-based detection by 11dB and 1dB, respectively. Cognitive radio (CR) is a critical technology that allows for 1. Introduction more eﬃcient use of restricted and ineﬃciently utilized fre- Nowadays, wireless communication technologies are rapidly quency bands using an opportunistic manner . CR has four evolving to accommodate people’s demands and require- main tasks/functions: spectrum sensing (SS), spectrum deci- sion/management, spectrum sharing/allocation, and spectrum ments, which are changing exponentially . As wireless technology advances, the need for spectrum resources is mobility/handoﬀ. SS is used to determine the portion of the increasing, which has resulted in a scarcity of spectrum due spectrum that are vacant and senses the presence of licensed to a static spectrum allocation policy. On the contrary, recent primary users (PUs). Spectrum management selects the most studies on current spectrum allocation show the underuti- suitable vacant spectrum holes among the detected ones. )e lization of the allocated spectrum to the licensed user at any goal of spectrum sharing is to evenly or fairly distribute the speciﬁc location and time. Cognitive radio has been iden- spectrum holes among the secondary users (SUs). Spectrum tiﬁed as a possible technology for addressing the issues of mobility aims to maintain communication while transitioning spectral scarcity and underutilization . to better spectrum holes [3, 4]. 2 Wireless Power Transfer )e two most important parameters used to evaluate the Among all functions of CR, SS is regarded as the most critical component in its establishment. A number of SS performance of any SS method are probability of detection (P methods have been suggested in recent decades, which can ) and probability of false alarm (P ). )e probability of d f be classiﬁed into two categories based on the bandwidth of correctly detecting the presence of a primary (licensed) user interest for spectrum sensing: wideband and narrowband. is denoted by P . As a result, a high P is preferable because d d Narrowband spectrum sensing examines a single frequency it ensures less interference to primary users. P is the band at a time, whereas wideband spectrum sensing analyzes probability of incorrectly declaring the primary user multiple frequencies simultaneously . Another way to presence. categorize sensing techniques is by the requirement for prior )e structure of this study is as follows: Section 2 information of PU signals, which are classiﬁed as coherent presents a review of various works that are closely related to and noncoherent detection. However, spectrum sensing the proposed system, and Section 3 discusses a system model techniques are popularly classiﬁed into three diﬀerent of conventional energy detection, entropy detection, and the techniques for sensing spectrum holes: transmitter-based, proposed two-stage spectrum sensing. In Section 4, the interference-based, and receiver-based detection. In this simulation results using MATLAB 2020a are presented with study, we only look at transmitter-based detection. Under their discussion. Finally, Section 5 brings the study to a close transmitter-based detection, many techniques are presented and makes suggestions for future research. such as eigenvalue detection, energy detection (ED), cyclostationary feature detection (CFD), and matched ﬁlter 2. Related Works detection (MFD). In this study, ED is generally adopted for SS due to its In , the performance of SS (spectrum sensing) for CR was noncoherent method, in which the SU receiver requires no examined under noise ﬂoor and low SNR values were prior information about PU signals. It has a low compu- presented. )is study combines dynamic or adaptive tational complexity and a fast sensing time. However, noise threshold ED with the kernel principal component analysis uncertainty, which is the random or unpredictable and (KPCA) to overcome the vulnerability of energy detection unavoidable variation of noise in any wireless communi- (ED) to noise uncertainty at low SNR values. A novel cation system, severely reduces ED performance, especially mathematical approach is proposed to obtain the optimal when SNR is low . Many methods have been proposed to sensing duration for energy detection and to analyze the address the challenges of ED. Among those methods, en- eﬀect of sample numbers on the existence of noise uncer- tropy detection becomes the most robust one due to the tainty in . In , an investigation of the use of ED for noise uncertainty at low SNR values. Furthermore, because it spectrum sensing technique in CR is presented. In this study, does not require prior information about primary signals, its the concept of diﬀerent types of spectrum sensing techniques implementation complexity is comparable to that of ED. with their theoretical and mathematical formulas is dis- )ere are many types of entropy detection, such as Shannon, cussed clearly. In this study, the ED method is one of the SS Renyi, Kapur, and Tsallis entropy. So, by comparing these techniques that is analyzed in detail. An energy detection types of entropy with each other, the one that outperforms method is used to detect the unused portions of the spectrum better is examined in this study for developing two-stage SS and make them available for reuse. Using the energy de- with conventional ED (CED). tection method, we can identify and allocate gaps in the In this study, to address the challenges of ED and to spectrum to secondary users. Also, the eﬀects of fading, further enhance the performance of entropy-based detec- shadowing, and hidden terminal problems on detection tion, two-stage SS techniques are proposed, using conven- performance are discussed. )is study analyzes energy de- tional energy detection (CED) and best performer entropy tection techniques well, but it fails to detect PU signals at low types among the entropies listed above. Because it accom- SNR levels. plishes SS within the shortest time and provides accurate In , a numerical analysis of histogram-based esti- detection at high SNR values, ED is used as the coarse stage mation techniques for entropy-based spectrum sensing is (ﬁrst stage) for the proposed technique. However, due to its proposed. In this study, spectrum detection based on en- robustness to noise uncertainty in low SNR values, the best tropy had been proposed to sense primary transmission in performer entropy detection technique is used in the ﬁne cognitive radio network (CRN). To estimate entropy, the stage (second stage). histogram method was used. )e performance of the en- )e main contributions of this study are as follows. tropy-based detection with respect to several rules for cal- We applied diﬀerent types of entropy-based detection to culating the number of bins in the histogram is evaluated. It alleviate the performance degradation of CED, which occurs demonstrated that the performance of detection is diﬀerent at low SNR condition due to uncertainty of noise. Moreover, for each of the aforesaid rules due to the probability dis- the performance comparisons among various types of en- tribution of the PU signal. )e main focus of this study was tropy detection are performed to choose the best performer on the optimal determination of number of bins. However, entropy method. the current hot research topics on the area of CRs are Depending on the results found from the comparison, a improving the performance of spectrum sensing at low SNR. novel two-stage spectrum sensing techniques are developed In addition, the Shannon entropy is the only subject of this using CED as a coarse stage and the best performer entropy work. Other types of entropy, such as Renyi, Kapur, and detector as a ﬁne stage. Tsallis entropy, are not included. Wireless Power Transfer 3 predicting PU activity without accomplishing any SS task in )e authors in  present a comparison of Tsallis and Kapur entropy in communication systems. )is work fo- system operation. After that, a deep Q-network (DQN)- based energy and spectrum eﬃcient routing strategy is cuses on analyzing the best entropy in communication system among the Kapur, Tsallis, and generalized entropy. suggested to maximize the CRN throughput. In , an )e authors of  considered a cyclostationary feature for intelligent reﬂecting surface (IRS)-enhanced ED for SS in cognitive radio based on entropy estimation to enhance the CRNs is proposed. )is study clearly discusses both situa- SS. In this study, the proposed algorithms are veriﬁed under tions where the PU is directly connected to SU and those single-node and multi-node/cooperative situations. In where they are not. Furthermore, IRS-enhanced ED for [13, 14], entropy-based spectrum sensing for cognitive radio cooperative SS is also proposed to improve the detection performance of SS. )e authors in  present a kernel fuzzy is proposed. A numerical study is performed in  to analyze histogram estimation methods for entropy-based SS. C-means clustering (KFCM) on ED-based cooperative SS. )is study focuses on enhancing the performance of de- )e performance of the entropy-based detection is evaluated in this study in relation to numerous rules for calculating the tection using KFCM rather than fuzzy C-means clustering (FCM). )ere are two types of KFCM, which are KFCM-K number of bins in the histogram. In , two-stage SS techniques for CR with adaptive and KFCM-F. )e KFCM-F forms are considered in this thresholds are proposed by combining two-well techniques: study since it does not require an inverse mapping from wavelet denoising and energy detection. In this study, an ED kernel space. )e summary of recent research paper is listed technique is used to determine the availability or presence of in Table 1. PU signal in the presence of a high SNR value by comparing the energy of the received signal with values of the threshold. 3. System Model However, in the presence of low SNR values, a wavelet In this section, the spectrum sensing (SS) methods such as denoising stage is used prior to energy detection (ED) to energy detection, entropy-based detection, and the proposed decrease the noise inﬂuence and to sense the PU signal in two-stage SS are discussed. Each method of SS has its own set noisy environments. )e authors of  present a two-step of beneﬁts and drawbacks. SS technique for cognitive radio networks (CRNs). In this )e main task of SS in CRNs is to determine whether an work, two-step techniques are implemented using energy authorized PU (primary user) is present or not on a given detection (ED) as the ﬁrst stage and maximum eigenvalue channel, so that secondary users can eﬃciently access and detection (MED) as the second stage. According to the exploit the unoccupied or unused spectrum. SS depends on comparison results of this study, the proposed method takes a well-known method known as signal detection. In a noisy less sensing time than MED and ED while achieving ap- environment, signal detection methods are used to detect proximately the same sensing performance as MED. In , the availability of signal. Signal detection could be sim- an improved two-stage SS for CRNs is proposed. )e author pliﬁed to a basic identiﬁcation problem that can be for- focuses on resolving the non-optimality issues occurring in mulated as hypothesis test analytically . Generally, the spectrum sensing performance of two-stage and im- spectrum sensing can be modeled as a binary hypothesis proves the probability of detection. problem in the detection theory and can be given as  In , new parallel fully blind multistage detectors are follows: proposed. Appropriate stages are assumed based on the estimated SNR value that is achieved from the SNR esti- w : H mator. Energy detection is used in the ﬁrst stage for its y(n) � , (1) h s(n) + w(n) : H simplicity and sensing accuracy at high SNR. For low SNR, a 1 maximum eigenvalue detector techniques with diﬀerent where n �1, 2, 3, . . ., N is the sample number of the sampled smoothing factors are adopted for higher stages. )e sensing signal that has been received, y (n) is the sampled signal that accuracy for maximum eigenvalue detector increases with an has been received by secondary users, w (n) is the noise increase in smoothing factor. Also, they analyzed the per- introduced by AWGN (additive white Gaussian noise) formance of two cases of the proposed detector: two-stage channel with zero mean and variance σ , s (n) is the signal and three-stage schemes. However, the computational from PU with variance σ and zero mean, h is the impulse complexity at the higher stages becomes increased due to the response of the channel or the channel amplitude gain use of eigenvalue detector and an increase in smoothing between PU transmitter and secondary user (SU) receiver factor. )e authors of  presented the SS in cognitive since we use AWGN channel h �1. H andH represent 0 1 vehicular networks for uniform mobility model. In this absence (null hypothesis) and presence (alternative hy- study, cooperative spectrum sensing (CSS) based on ED pothesis) of the PU, respectively. scheme is enabled for CR vehicular networks by considering the velocity of PU and SU nodes as a uniform motion. Initially, a distance-dependent distribution function is 3.1. Energy Detection. Energy detection (ED) is one of the proposed to ﬁnd the probability of a secondary user resides most frequently used approaches, and because of it, it re- in particular coverage of the PU transmission zone. quires no prior information about the PU signal and has In , a machine learning for spectrum information small computational complexity. Figure 1 depicts the ED and routing in multi-hop green CRNs is presented. In this block diagram, which is used to identify whether or not PU is study, a support vector machine (SVM) is used for present in a given channel. Initially, an ideal band pass ﬁlter 4 Wireless Power Transfer Table 1: A summary of some related work. Authors Year Title Techniques and concept covered Gaps or concept not covered Entropy-based detection is proposed Numerical analysis of for spectrum sensing. Several rules for )is study focuses only on the Shannon G. Prieto histogram-based estimation determining the number of bins in entropy. It does not include other types et al.  techniques for entropy-based histogram are evaluated. )ose rules of entropy such as Renyi, Kapur, and spectrum sensing are as follows: square root rule, Scott Tsallis entropy rule, and Sturges rule Even though it improves the performance of SS, the proposed Adaptive two-stage spectrum Two-stage spectrum sensing techniques technique is not robust to noise A. Fawzi sensing model using energy 2020 are proposed by combining two well uncertainty at low SNR. Moreover, since et al.  detection and wavelet denoising methods: ED and WD WD is integrated with ED at low SNR, for CR systems sensing time and computational complexity are very high Cooperative spectrum sensing using Analysis of energy detection A. D. Sahithi ED is used to overcome the problem of )is study does not focus on overcoming 2020 spectrum sensing technique in et al.  uncertainty occurred due to fading and the problem of noise uncertainty cognitive radio hidden primary terminal Two-stage and three-stage detector for Even though three-stage and two-stage SS are discussed in detail spectrum sensing are performed, the An integrated parallel overall performance of SS is not good at F. Mashta 2021 multistage spectrum sensing for ED and maximum eigenvalue detector low SNR due to its sensitivity to noise et al.  cognitive radio with diﬀerent smoothing factors are uncertainty. Moreover, since the used eigenvalue-based detector is used, the complexity of overall system increases Energy Detection Summation y (t) BPF ADC (.) T≥λ ED or Integration T<λ ED Energy Measurement, Figure 1: Block diagram of energy detection. with a bandwidth of interest is used to preﬁlter the received 2 2 λ − Nσ + σ T≥ λ ED w s ED ⎛ ⎜ ⎞ ⎟ ⎜ ⎟ signal in an ED. )en, the ﬁltered signal is fed into ADC ⎜ ⎟ ⎝ ����������� � ⎠ P � P � Q , (3) 2 2 (analog-to-digital converter) converter. After that, the ADC 2Nσ + σ w s output signal is squared and integrated over the pre- determined time interval. )e obtained signal from the T< λ λ − Nσ integrator is used to develop a test statistics . Finally, the ED ⎛ ⎜ ED w⎞ ⎟ ⎜ ⎟ ⎜ ⎟ ⎝ ����� ⎠ P � P � Q , (4) deﬁned or developed test statistics are compared with a H 2Nσ predeﬁned threshold “λ ,” to identify whether or not the ED licensed user is present. Test statistics of ED is given as  where T denotes the test statistics, λ is the predeﬁned ED follows: threshold of ED, N is the sample number, σ is the variance of AWGN, and σ is the variance of PU signal. Here, T(y) � |y(n)| , (2) N ∞ n�1 −y √�� � Q(x) � exp dy, (5) 2π 2 where T(y) represents the test statistics of ED. )e detection probability (P ) and false alarm prob- where Q (x) is Q function. ability (P ) are two parameters used to evaluate the de- )e detection threshold can be given as follows: tection performance of any SS method. P and P are d f associated with a speciﬁc threshold λ that test decision √��� ED 2 − 1 λ � σ 2N Q P + N. (6) statistics : ED f w Wireless Power Transfer 5 Entropy Detection T ≤ λ Calculate L Histogram EnD y (n) DFT T > λ EnD Entropy Estimation, T (y) reshold, EnD Figure 2: Block diagram of entropy-based detector. 3.2. Entropy-Based Detection. Figure 2 shows the basic block where E is Tsallis entropy. diagram of entropy-based detector. After applying DFT to a Kapur’s entropy formula is given by the following binary hypothesis given in equation (1), we get the following: equation : → → α n 1/α y (k) � w (k). (7) 1 − P k�1 k (12) E � , 1 − α → → → y (k) � s (k) + w (k), (8) where E is Kapur’s entropy, α is the order of entropy, and → → → p is the frequency of occurrences in kth bins and is given as where s (k), y (k), and w (k) are frequency spectrum [10, 14]follows: representation of the primary user (PU) signal, received signal, and noise signal, respectively. p � . (13) Numerous methods for predicting the entropy (ran- domness) of random variables depending on a limited set of By substituting equation (13) in equations (9), (10), (11), observations have been suggested. Entropy estimation based and (12), we get the test statistics [10, 14]. on histogram is evaluated in this study due to its lower complexity. T(y) � E. (14) By splitting the ranges of (y − y ) of values in y into max min L bins of constant width A, the data set for histogram, which )e detection threshold is determined as [10, 14]follows: is y � y , y , y , . . . , y , is obtained. Let n denotes the 0 1 2 N−1 k − 1 λ � H + Q 1 − P σ , (15) EnD L f n number of items in y that fall within the kth bins such that n � N [6, 10, 14]. After constructing the histogram, k�1 k where the entropy, which is denoted by E, is calculated as follows. )e Shannon entropy formula is given by the following L c √� H � ln + + 1. (16) equation [10, 14]: 2 2 ( ) In Gaussian noise entropy, the number of bins is denoted E � − p log , (9) S k 2 by L, c represents Euler–Mascheroni constants, P is false k�1 f alarm probability, and σ is the standard deviation of H where E is Shannon entropy, L is the number of bins, and P S k under H . is the frequency of occurrences in kth bins. )e Renyi entropy formula is given by the following equation : 3.3. Proposed Two-Stage Spectrum Sensing. In this study, two-stage SS techniques are developed to enhance the ⎝ ⎠ ⎛ ⎞ performance of detection as shown in Figure 3. In the ﬁrst E � log P , (10) 1 − α stage (coarse stage), ED is used because it has faster sensing k�1 time than other methods, and then, the average power of the where E is Renyi entropy, α is the order of entropy, L is the received signal is compared with a threshold λ ; if it ex- ED number of bins, and P is the frequency of occurrences in kth ceeds λ , then the channel is declared to be occupied by PU; ED bins. else, we proceed to the second stage. In the second stage (ﬁne )e Tsallis entropy formula is given by the following stage), one of the entropy-based detection methods that equation [11, 14]: outperform other methods of entropy is used, and then, statistical tests are executed using entropy formulas and ⎝ ⎠ ⎛ ⎞ compared it with a threshold λ to decide whether channel E � 1 − P , (11) EnD T k α − 1 k�1 is idle or occupied. PU Absent 6 Wireless Power Transfer Coarse Stage Fine Stage Energy No Entropy No T ≥ λ Y (n) T ≤ λ ED EnD Detection Detection Yes Yes PU Present Figure 3: System model for proposed two-stage spectrum sensing. )e general mathematical equations for false alarm probability and detection probability in two-stage SS are 0.9 given as follows, respectively: (coarse) (fine) (coarse) 0.8 P � P + P ∗ 1 − P , (17) f f f f 0.7 (coarse) (fine) (coarse) 0.6 (18) P � P + P ∗ 1 − P , d d d d 0.5 where P is false alarm probability of overall system, P is fa d (coarse) 0.4 detection probability of overall system, P is false alarm (fine) probability of coarse stage, P is false alarm probability of 0.3 (coarse) ﬁne stage, P is detection probability of coarse stage, 0.2 (fine) and P is detection probability of ﬁne stage. 0.1 4. Simulation Results –30 –25 –20 –15 –10 –5 0 SNR in dB In this section, the simulation results of energy detection, entropy-based detection, and two-stage detection are dis- Shannon Entropy kapurs Entropy cussed using MATLAB version 2020a. Simulations are Renyi Entropy Energy Detection provided to evaluate the performance of the proposed two- tsallis Entropy stage SS with respect to conventional energy detection and Figure 4: P vs. SNR at Pf �0.1 and α �4 for CED and diﬀerent entropy-based detection under AWGN channels. 10,000 types of entropy. Monte Carlo simulations and 1000 sample number were used to generate the simulation results. According to IEEE wall compared with Tsallis, Shannon, and Kapur entropy, 802.22 standard, all simulations must take into account the requisite detection probability (≥90%), the probability of respectively. Depending on this result, the Renyi entropy can detect a weaker PU signal than other entropy methods since false alarm (≤10%), and the probability of miss-detection (<10%) for cognitive radio. SNR wall is a key factor eval- it has a lower SNR wall than others. )e detection technique with a lower SNR wall value has better sensitivity. Hence, it uated in all graphs of the result to compare the proposed can be concluded that the Renyi entropy detection out- techniques with the existing ones. Moreover, SNR wall is performs both CED and all other types of entropy. used to compare various types of entropy-based detection Figure 5 depicts the receiver operating characteristic with each other. )e minimal SNR below which detection is (ROC) curve for both CED and various types of entropy- not possible is referred to as the SNR wall. based detection at SNR= −18dB. )is simulation is done by Figure 4 depicts the relationship between detection setting the number of bins to 15 and the order of entropy to probability (Pd) and SNR for conventional energy detection 4. As shown in Figure 5, an increase in the probability of false (CED) and various types of entropy-based detection at a probability of false alarm (P alarm enhances the detection probability of entropy. Renyi �0.1), number of bins (L �15), entropy-based detection achieves the desired probability of and order of entropy (α �4). Any spectrum sensing tech- nique with a detection probability greater than or equal to detection (P ≥0.9) with the lowest probability of false alarm when compared to other methods. As illustrated in the 0.9 (≥90%) can distinguish PU signal from noise signals, according to IEEE 802.22 standards. As observed, Tsallis ﬁgure, the Kapur entropy has better performance than the Shannon and Tsallis entropy. Also, the Shannon entropy has entropy, Shannon entropy, energy detection, and Kapur’s a better probability of detection than the Tsallis entropy. As entropy detect primary user (PU) signal at SNR wall of observed, the Renyi entropy achieves the desired probability −7dB, −9dB, −10dB, and −14dB, respectively. However, of detection with 0.05 probability of false alarm, while the the Renyi entropy detects the PU signal at an SNR wall of Kapur entropy, Shannon entropy, Tsallis entropy, and CED −20dB. Renyi entropy detection has a signiﬁcant im- achieve it with 0.22, 0.33, 0.36, and 0.9, respectively. )e provement of about 13dB, 11dB, 10dB, and 6dB in SNR Probability of Detection (Pd) Wireless Power Transfer 7 1 0 0.9 0.8 0.7 0.6 –1 0.5 0.4 0.3 0.2 0.1 –2 –2 –1 0 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 10 10 10 Probability of False Alarm (Pf ) Probability of False Alarm (Pf ) Shannon Entropy kapurs Entropy Shannon Entropy kapurs Entropy Renyi Entropy Energy Detection Renyi Entropy Energy Detection tsallis Entropy tsallis Entropy Figure 5: ROC curve for CED and various types of entropy de- Figure 6: CROC curve for CED and various types of entropy tection at SNR � −18dB. detection at SNR � −18dB. detector with a higher probability of false alarm leads to poor spectrum utilization. Hence, it can be deduced that the Renyi 0.9 entropy outperforms all other methods since it requires a 0.8 lower probability of false alarm to obtain the required de- 0.7 tection probability than the others. 0.6 Figure 6 depicts the complementary receiver operating 0.5 characteristics (CROC) curve for both CED and various 0.4 types of entropy-based detection at SNR � −18dB. As it can 0.3 be seen, the Renyi entropy-based detection has the lowest 0.2 probability of miss-detection when compared to other 0.1 methods. )e method with the lowest miss-detection probability can make eﬃcient use of the spectrum holes. –30 –25 –20 –15 –10 –5 0 Figure 7 illustrates the impact of the number of bins on SNR in dB the Renyi entropy-based detection. As observed from the ﬁgure, the detection probability decreases as the number of Renyi entropy at L=15 bins increases. When the number of bins is 15 (L �15), the Renyi entropy at L=17 Renyi entropy detection can detect the primary signal up to Renyi entropy at L=20 −20dB SNR values, whereas at L �17 and L �20, it can Figure 7: P vs SNR curves for Renyi entropy detection at various detect the PU signal up to −19dB and −18dB, respectively. numbers of bins. As a result, the sensing performance of entropy detection increases as the number of bins decreases. Figure 8 depicts the comparison among the proposed proposed two-stage methods have a lower probability of false alarm than Renyi entropy and CED for detecting PU two-stage SS, Renyi entropy-based detection and conven- tional energy detection (CED), at a particular probability of signals within the desired probability of detection. It is clear false alarm of 0.1, number of bins of 15, and order of entropy from the ﬁgure that the proposed detector performs better in of 4. )e comparison results show that the proposed method terms of detection. For instance, at a given probability of outperforms both Renyi entropy and CED by a signiﬁcant false alarm of 0.1, the detection performance of the proposed technique is 0.631, while the detection probability of Renyi performance improvement. For instance, at a given SNR of −23dB, the detection probability of the proposed technique entropy and CED is 0.5298 and 0.2153, respectively. In other words, the proposed two-stage SS technique achieves the is 0.6336, while the detection probability of Renyi entropy and CED is 0.5354 and 0.2113, respectively. In other words, desired probability of detection with 0.16 probability of false alarm, while the Renyi entropy and CED achieve it with 0.18 the proposed two-stage technique has a signiﬁcant im- provement of about 11dB and 1dB in SNR wall when and 0.96, respectively. As a result, it is possible to deduce that compared to CED and Renyi entropy, respectively. the proposed techniques have better detection performance Figure 9 shows the ROC curve that compares the pro- than both Renyi entropy and CED since it requires a lower posed two-stage SS with Renyi entropy and CED at probability of false alarm to obtain the desired probability of detection than others. SNR � −23dB, Pf �0.1, and L �15. As it can be seen, the Probability of Detection (Pd) Probability of Detection (Pd) Probability of Miss Detection (Pm) 8 Wireless Power Transfer 1 0 0.9 0.8 0.7 0.6 –1 0.5 0.4 0.3 0.2 –2 –2 –1 0 0.1 10 10 10 Probability of False Alarm (Pf ) –30 –25 –20 –15 –10 –5 0 Energy Detection SNR in dB Renyi Entropy Renyi Entropy Proposed Two-Stage Energy Detection Figure 10: CROC curve for proposed two-stage SS technique at Proposed Two-Stage SNR �-23dB. Figure 8: P vs SNR at Pf �0.1 and α �4 for CED and diﬀerent types of entropy. 0.9 0.8 0.9 0.7 0.8 0.6 0.7 0.5 0.6 0.4 0.5 0.3 0.4 0.2 0.1 0.3 0.2 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0.1 Probability of False Alarm (Pf ) 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 N=1000 N=2000 Probability of False Alarm (Pf ) N=3000 Energy Detection Figure 11: ROC curve of proposed two-stage SS technique with the Renyi Entropy diﬀerent number of samples. Proposed Two-Stage Figure 9: ROC curve for proposed two-stage SS technique at SNR �-23dB. SU. )erefore, from this result, it can be concluded that the proposed technique is better at distinguishing PU signals from noise signals than Renyi entropy and CED. Figure 10 shows the CROC curve that compares the Figure 11 illustrates the ROC curve for the proposed proposed two-stage SS with Renyi entropy and CED at two-stage SS technique at various sample sizes. In this SNR � −23dB, Pf �0.1, and L �15. As it can be observed simulation, the SNR value is set to −27dB. )e detection from the ﬁgure, the proposed technique has a lower missed performance improves as the number of samples increases detection probability over all ranges of P compared with for a particular probability of false alarm. At a particular or Renyi entropy and CED. For instance, at a given probability given SNR and false alarm probability, only a sample size of of false alarm of 0.07, the missed detection probability of the 3000 may attain IEEE 802.22 standards, as seen in the graph. proposed technique is 0.5885, while the missed detection )e ROC curve plotted for the proposed two-stage SS probability of Renyi entropy and CED is 0.7127 and 0.8258, technique with 3000 sample sizes achieves the desired respectively. )e detection technique with a lower proba- probability of detection (i.e., P ≥90%) with 0.09 probability bility of missed detection leads to less interference to PU by of false alarm, while the ROC curve plotted with 2500 and Probability of Detection (Pd) Probability of Detection (Pd) Probability of Detection (Pd) Probability of Miss Detection (Pm) Wireless Power Transfer 9 1 improved by 2.9986 times compared with CED. In other words, the proposed two-stage technique has a signiﬁcant 0.9 improvement of about 11dB and 1dB in SNR wall when 0.8 compared to CED and Renyi entropy, respectively. 0.7 According to the simulation results, increasing the number of samples improves the detection probability of the 0.6 spectrum sensing scheme. Moreover, it has been observed 0.5 that the detection probability of the SS scheme increases as 0.4 both SNR and the probability of false alarm increase. In addition, this study investigates the performance analysis 0.3 among four types of entropy detections, and the simulation 0.2 result shows that the Renyi entropy is the best one at the 0.1 entropy order of 4 (α �4). In the future, it is recommended to analyze the impact of cooperative sensing on the proposed –30 –25 –20 –15 –10 –5 0 two-stage SS scheme. Another possibility for future research is to adapt the proposed two-stage SS scheme to analyze its SNR in dB behavior within MIMO detecting circumstances and in- Pf=0.05 Pf=0.15 vestigate its eﬀect on the performance of sensing. In this Pf=0.1 Pf=0.2 study, the AWGN channel was employed for detection, and the other channel like fading channel (Rayleigh and Rician) Figure 12: Performance comparison of the proposed two-stage SS can be applied for detection. technique at various values of false alarm probability. Data Availability 1000 sample size achieves it with 0.15 and 0.26, respectively. As a result, it is possible to deduce that the performance of )e data are available from the corresponding author upon proposed SS technique is enhanced by increasing the sample request. sizes (number of samples) since an increasing sample size decreases the false alarm probability of attaining the re- Conflicts of Interest quired probability of detection. Figure 12 shows the detection performance of the )e authors declare that they have no conﬂicts of interest. proposed two-stage SS scheme at the sample number of 1000 (N �1000) for various values of the probability of false References alarm. As illustrated in the graph, the detection probability is  I. Develi, “Spectrum sensing in cognitive radio networks: increased when the probability of false alarm increases. As threshold optimization and analysis,” EURASIP Journal on shown from the ﬁgure, the curve that is plotted with P �0.2 Wireless Communications and Networking, vol. 255, 2020. achieves the desired detection probability at an SNR of  S. Srinu, S. L. Sabat, and S. K. Udgata, “Spectrum sensing −24dB, whereas the curves that are plotted with P �0.15, using frequency domain entropy estimation and its FPGA P �0.1, and P �0.05 achieve it at SNRs of −22dB, −20dB, f f implementation for cognitive radio,” Procedia Engineering, and −18dB, respectively. In other words, the curve that is vol. 30, pp. 289–296, 2012. plotted using P �0.2 has the highest probability of de- f  A. Fawzi, W. El-Shafai, M. Abd-Elnaby, A. Zekry, and tection when compared to others, while Pf �0.05 has the F. E. Abd El-Samie, “Adaptive two-stage spectrum sensing lowest detection performance. However, the maximum model using energy detection and wavelet denoising for cognitive radio systems,” International Journal of Commu- acceptable P for cognitive radio is 0.1, which cannot be nication Systems, vol. 33, no. 16, pp. e4400–25, 2020. surpassed according to IEEE 802.22 standards.  G. Tomar, A. Bagwari, and J. Kanti, Introduction to Cognitive Radio Networks and Applications, CRC Presss, Boca Raton, 5. Conclusion and Future Research FL, USA, 2016.  Y. Arjoune and N. Kaabouch, “A comprehensive survey on In this work, a two-stage SS scheme for CR has been de- spectrum sensing in cognitive radio networks: recent ad- veloped to improve the detection performance. )e pro- vances, new challenges, and future research directions,” posed detector consists of conventional energy detection as Sensors, vol. 19, no. 1, 2019.  G. Prieto, A. G. Andrade, D. M. Mart´ınez, and G. Galaviz, “On coarse stage and Renyi entropy-based detection as ﬁne stage. the evaluation of an entropy-based spectrum sensing strategy )e comparison results show that the proposed method applied to cognitive radio networks,” IEEE Access, vol. 6, outperforms both Renyi entropy and CED by a signiﬁcant pp. 64828–64835, 2018. performance improvement. For instance, at a given SNR of  P. Venkatapathi, H. Khan, and S. Rao, “Performance analysis −23dB, the detection probability of the proposed technique of spectrum sensing in cognitive radio under low SNR and is 0.6336, while the detection probability of Renyi entropy noise ﬂoor,” International Journal of Engineering and Ad- and CED is 0.5354 and 0.2113, respectively. )is indicates vanced Technology, vol. 9, no. 2, pp. 2655–2661, 2019. that the detection of the proposed technique is improved by  G. Mahendru, A. Shukla, and P. Banerjee, “A novel mathe- 1.1834 times compared with Renyi entropy, while it is matical model for energy detection based spectrum sensing in Probability of Detection (Pd) 10 Wireless Power Transfer cognitive radio networks,” Wireless Personal Communica- tions, vol. 110, no. 3, pp. 1237–1249, 2020.  A. D. Sahithi, E. L. Priya, and N. L. Pratap, “Analysis of energy detection spectrum sensing technique in cognitive radio,” International Journal of Scientiﬁc and Technology Research, vol. 9, no. 1, pp. 1772–1778, 2020.  G. Prieto, A. G. Andrade, and D. M. Mart´ınez, “Numerical analysis of histogram-based estimation techniques for en- tropy-based spectrum sensing numerical analysis of histo- gram-based estimation techniques for entropy-based spectrum sensing,” IETE Technical Review, vol. 4602, 2019.  V. Kumar and P. Goyel, “A comparative study of tsalli ’ s and kapur ’ s entropy in communication systems,” International Journal of Computer Application, vol. 62, p. 7, 2013.  S. L. Sabat, S. Srinu, A. Raveendranadh, and S. K. Udgata, “Spectrum sensing based on entropy estimation using cyclostationary features for cognitive radio,” in Proceedings of the Fourth International Conference on Communication Sys- tems and Networks (COMSNETS 2012), 2012.  X. Chen and S. Nagaraj, “Entropy based spectrum sensing in cognitive radio,” in Proceedings of the Wireless Telecommu- nications Symposium, 2008.  G. Vaidehi, N. Swetha, and P. N. Sastry, “Entropy based spectrum sensing in cognitive radio networks,” International Journal of Advanced Research Computer and Communication Engineering, vol. 4, no. 11, pp. 39–43, 2015.  Z. Li, H. Wang, and J. Kuang, “A two-step spectrum sensing scheme for cognitive radio networks,” in Proceedings of the International Conference on Information Science and Tech- nology, IEEE, Nanjing, China, 2011.  F. Wasonga, T. O. Olwal, and A. Abu-Mahfouz, “Improved two-stage spectrum sensing for cognitive radio networks,” Journal of Advanced Computational Intelligence and Intelli- gent Informatics, vol. 23, no. 6, pp. 1052–1062, 2019.  F. Mashta, M. Wainakh, and W. Altabban, “An integrated parallel multistage spectrum sensing for cognitive radio,” International Journal of Embedded and Real-Time Commu- nication Systems, vol. 12, no. 2, pp. 1–20, 2021.  A. Paul, P. Kunarapu, A. Banerjee, and S. P. Maity, “Spectrum sensing in cognitive vehicular networks for uniform mobility model,” IET Communications, vol. 13, no. 19, pp. 3127–3134,  A. Paul and S. P. Maity, “Machine learning for spectrum information and routing in multihop green cognitive radio networks,” IEEE Transactions on Green Communications and Networking, vol. 6, no. 2, pp. 825–835, 2022.  W. Wu, Z. Wang, L. Yuan et al., “IRS-enhanced energy de- tection for spectrum sensing in cognitive radio networks,” IEEE Wireless Communications Letters, vol. 10, no. 10, pp. 2254–2258, 2021.  A. Paul and S. P. Maity, “Kernel fuzzy c-means clustering on energy detection based cooperative spectrum sensing,” Digital Communications and Networks, vol. 2, no. 4, pp. 196–205,  M. Abdo-tuko, “Performance evaluation and comparison of diﬀerent transmitter detection techniques for application in cognitive radio,” International Journal of Networks and Communications, vol. 5, no. 5, pp. 83–96, 2015.  A. Bagwari and G. S. Tomar, “Two-stage detectors with multiple energy detectors and adaptive double threshold in cognitive radio networks,” International Journal of Distrib- uted Sensor Networks, vol. 9, Article ID 656495, 2013.
Wireless Power Transfer – Hindawi Publishing Corporation
Published: Aug 24, 2022
Access the full text.
Sign up today, get DeepDyve free for 14 days.