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

Learn More →

Nonlinear Distortion Mitigation in Multi-IF over Fiber Transmission Using Modulation-Based Adaptive Power Allocation

Nonlinear Distortion Mitigation in Multi-IF over Fiber Transmission Using Modulation-Based... hv photonics Article Nonlinear Distortion Mitigation in Multi-IF over Fiber Transmission Using Modulation-Based Adaptive Power Allocation Inho Ha, Hyoung-Joon Park, Soo-Min Kang and Sang-Kook Han * Department of Electrical and Electronics Engineering, Yonsei University, Seoul 03722, Korea; 91hainho@yonsei.ac.kr (I.H.); phjokokok@yonsei.ac.kr (H.-J.P.); roemee817@yonsei.ac.kr (S.-M.K.) * Correspondence: skhan@yonsei.ac.kr Abstract: We propose a modulation-based adaptive power allocation (MBAPA) technique for nonlin- ear distortion mitigation in intermediate frequency over fiber (IFoF) systems. The technique allocates the spectral power of each IF band according to the required signal-to-noise power ratio (SNR) of the modulation format. To demonstrate the performance of the technique, transmission experiments were performed in 10 km and 20 km with 24-IF bands using OFDM signals. The feasibility of the proposed MBAPA technique was experimentally verified by reducing inter-modulation distortion (IMD) power and enhancing channel linearity. Keywords: intermodulation distortion; multi-IFoF transmission; mobile fronthaul network; power allocation 1. Introduction With the advent of diverse applications and devices, the number of devices per single user and the required data capacity has been increasing significantly, leading to an expo- Citation: Ha, I.; Park, H.-J.; nential growth in mobile data traffic [1]. As a result, due to the limitation of data traffic per Kang, S.-M.; Han, S.-K. Nonlinear Dis- cell and the increase in the amount of data required per device, cell coverage is reduced to tortion Mitigation in Multi-IF over Fiber small cells. Many cells, including small cells, were configured to support the required data Transmission Using Modulation-Based traffic. To reduce capital expenditure (CAPEX) and operating expenditures (OPEX), the Adaptive Power Allocation. Photon- radio access network (RAN) structure changed from a distributed-RAN (D-RAN) structure ics 2021, 8, 2. https://dx.doi.org/ to a centralized RAN (C-RAN) structure in which digital units (DU) are separated from 10.3390/photonics8010002 the distributed cell sites. Because communication was required between the DU and the cell site where there are only radio units (RU) available, the common public radio inter- Received: 13 November 2020 face (CPRI) has been standardized and used as the digital optical transmission protocol Accepted: 19 December 2020 between the DU and the RU [1,2]. However, when the signal is converted for digital Published: 22 December 2020 transmission, the amount of transmission data is dynamically increased. For example, long term evolution (LTE, 4th generation) signal with 20 MHz bandwidth requires 2.457 Publisher’s Note: MDPI stays neu- Gbps speed of CPRI signal in digital transmission [2]. Likewise, the bandwidth required tral with regard to jurisdictional claims in 5G (5th generation) increases further, and the CPRI transmission capacity increases in published maps and institutional significantly. Thus, there is a limitation in digital transmission systems as the data re- affiliations. quired by the user increases in the current mobile fronthaul network (MFN). In order to prevent data capacity enlargement and support the required data traffic, Radio over fiber (RoF) transmission as an analog transmission has been proposed as an alternative technol- Copyright: © 2020 by the authors. Li- ogy [1–5]. The RoF transmission has higher spectral efficiency than the CPRI. The enhanced censee MDPI, Basel, Switzerland. This CPRI (eCPRI) technique that developed CPRI has been proposed. In the case of eCPRI, article is an open access article distributed data enlargement is reduced due to ethernet transmission through packet-based optical under the terms and conditions of the transmission. However, since eCPRI technology is a packet-based ethernet transmission, it Creative Commons Attribution (CC BY) is necessary to convert the data for mobile communication. It makes the system complexity license (https://creativecommons.org/ at the RU increase. In the RoF transmission, there is no process of digitizing the signal, licenses/by/4.0/). Photonics 2021, 8, 2. https://dx.doi.org/10.3390/photonics8010002 https://www.mdpi.com/journal/photonics Photonics 2020, 7, x FOR PEER REVIEW 2 of 9 CPRI has been proposed. In the case of eCPRI, data enlargement is reduced due to ethernet transmission through packet-based optical transmission. However, since eCPRI technology is a packet-based ethernet transmission, it is necessary to convert the data for Photonics 2021, 8, 2 2 of 9 mobile communication. It makes the system complexity at the RU increase. In the RoF transmission, there is no process of digitizing the signal, so the data capacity enlargement has not occurred. Accordingly, the major drawback of CPRI, data capacity enlargement, so the data capacity enlargement has not occurred. Accordingly, the major drawback of does not occur in the RoF transmission. Moreover, to use frequency resources efficiently, CPRI, data capacity enlargement, does not occur in the RoF transmission. Moreover, to use multiband intermediate frequency over fiber (multi-IFoF) transmission has been frequency resources efficiently, multiband intermediate frequency over fiber (multi-IFoF) proposed and studied [4,6]. The RoF transmission, utilizing an intermediate carrier transmission has been proposed and studied [4,6]. The RoF transmission, utilizing an frequency, is the most bandwidth-efficient technique in the MFN because it preserves intermediate carrier frequency, is the most bandwidth-efficient technique in the MFN spectral bandwidths of the signals. For transmitting high data capacity, CPRI requires because it preserves spectral bandwidths of the signals. For transmitting high data capacity, transceivers and WDM components because of data capacity enlargement, but only a CPRI requires transceivers and WDM components because of data capacity enlargement, single analog optical transceiver is required in the multi-IFoF transmission. Therefore, the but only a single analog optical transceiver is required in the multi-IFoF transmission. IFoF transmission could significantly reduce the required transmission capacity [7]. Therefore, the IFoF transmission could significantly reduce the required transmission ca- Though the multi-IFoF transmission has a high-efficiency of frequency resource and has pacity [7]. Though the multi-IFoF transmission has a high-efficiency of frequency resource no need for an additional device to convert digital form, IFoF transmission is vulnerable and has no need for an additional device to convert digital form, IFoF transmission is to noise, especially intermodulation distortion (IMD) noise, compared to the digital vulnerable to noise, especially intermodulation distortion (IMD) noise, compared to the transmission. Because of the difference in the receiving technique of the analog signal and digital transmission. Because of the difference in the receiving technique of the analog the digital signal. Several techniques have been used in previous research to reduce IMD signal and the digital signal. Several techniques have been used in previous research to noise effects [6–14]. One technique is the equalizer using many memory taps, another reduce IMD noise effects [6–14]. One technique is the equalizer using many memory taps, technique is to pre-distort the signal to mitigate the nonlinear channel effect, and the other another technique is to pre-distort the signal to mitigate the nonlinear channel effect, and is employing the additional external modulators into the system [8–11]. However, as the the other is employing the additional external modulators into the system [8–11]. However, number of devices in the cell increases and the required bandwidth of each device gets as the number of devices in the cell increases and the required bandwidth of each device broader, the total bandwidth and data capacity increases. The signal is affected by IMD gets broader, the total bandwidth and data capacity increases. The signal is affected by noises, including IMD2 and IMD3, which are generated by nonlinear devices and through IMD noises, including IMD2 and IMD3, which are generated by nonlinear devices and the fiber. Because of these complex IMD noises, it will be very difficult to estimate the through the fiber. Because of these complex IMD noises, it will be very difficult to estimate channel. Therefore, it is hard to reduce signal distortions by equalization and signal pre- the channel. Therefore, it is hard to reduce signal distortions by equalization and signal distortion requiring accurate channel information. These techniques are not suitable to pre-distortion requiring accurate channel information. These techniques are not suitable mitigate IMD noise in MFN. Existing studies have conducted just tone experiments for to mitigate IMD noise in MFN. Existing studies have conducted just tone experiments for mitigating nonlinear components [6–9,11] or have assumed that all IF bands have the same mitigating nonlinear components [6–9,11] or have assumed that all IF bands have the same modulation format. Furthermore, in conventional techniques, all IF bands have the same modulation format. Furthermore, in conventional techniques, all IF bands have the same power regardless of modulation format, as shown in Figure 1. However, because each IF power regardless of modulation format, as shown in Figure 1. However, because each IF band supports different devices and applications, each IF band passes through different band supports different devices and applications, each IF band passes through different channels and has different modulation formats. After receiving the signal at the receiver, channels and has different modulation formats. After receiving the signal at the receiver, the SNR the SNR is m is e measur asured a ednand d infinformed ormed the the cha channel nnel to the tra to the transmitter nsmitter. . (a) (b) Figure 1. Conventional signal spectrum of (a) transmitted signal and (b) received signal. Figure 1. Conventional signal spectrum of (a) transmitted signal and (b) received signal. Accordingly, the transmitter sets the modulation format, and power loading is per- Accordingly, the transmitter sets the modulation format, and power loading is formed with the water-pouring algorithm [15]. In the case of the CPRI, the generated performed with the water-pouring algorithm [15]. In the case of the CPRI, the generated orthogonal frequency-division multiplexing (OFDM) signal is converted to the time do- orthogonal frequency-division multiplexing (OFDM) signal is converted to the time main through the inverse Fourier transform (IFT), and the converted signal is digitized domain through the inverse Fourier transform (IFT), and the converted signal is digitized again to the baseband digital signal, then the signal is transmitted to the RU. However, again to the baseband digital signal, then the signal is transmitted to the RU. However, since this conversion leads to explosive growth in data traffic if the OFDM signal is directly since this conversion leads to explosive growth in data traffic if the OFDM signal is transmitted to the RU using the IFoF transmission, IMD noise is generated, and this IMD directly transmitted to the RU using the IFoF transmission, IMD noise is generated, and noise distorts the signal. At the RU, the distorted optical signal is downconverted to the RF this IMD noise distorts the signal. At the RU, the distorted optical signal is downconverted frequency, and the RF signal is amplified again and transmitted to each user through the wireless channel. As shown in Figure 2a, at the RU, the distorted optical signal is downcon- verted to the RF frequency, and the RF signal is amplified again and transmitted to each user through the wireless channel. Since the CPRI signal is converted to the OFDM signal at Photonics 2020, 7, x FOR PEER REVIEW 3 of 9 to the RF frequency, and the RF signal is amplified again and transmitted to each user through the wireless channel. As shown in Figure 2a, at the RU, the distorted optical signal is downconverted to the RF frequency, and the RF signal is amplified again and transmitted to each user through the wireless channel. Since the CPRI signal is converted to the OFDM signal at the RU, the IMD noise is not considered in the MFN using CPRI, and also, the transmitted signal power, which affects IMD noise power, was not a major consideration in the MFN. However, in the case of IFoF transmission, as shown in Figure 2b, after downconversion of the received signal, the signal is amplified without removing IMD noise at the RU, so the noise is also amplified and transmitted with the signal. Because of using an analog signal in IFoF transmission, the power of the nonlinear component affected by the signal power can be considered as the main distortion, and the generated IMD is amplified without being removed from the RU stage, which can continuously affect the signal distortion. In this paper, we propose a spectral power allocation technique dependent on the Photonics 2021, 8, 2 3 of 9 modulation format and verify the feasibility of system linearity enhancement of multi- IFoF-based MFN. The signal spectral power was allocated for each IF band according to modulation order, which is dependent on the required signal-to-noise power ratio (SNR), the RU, the IMD noise is not considered in the MFN using CPRI, and also, the transmitted which can be converted from the error vector magnitude (EVM) requirement. Since the signal power, which affects IMD noise power, was not a major consideration in the MFN. IMD noise power is affected by the transmitted signal power, when the total transmitted However, in the case of IFoF transmission, as shown in Figure 2b, after downconversion of signal power is changed, the IMD noise power would be significantly changed compared the received signal, the signal is amplified without removing IMD noise at the RU, so the to the signal power variation. We have experimentally demonstrated the reduction of noise is also amplified and transmitted with the signal. Because of using an analog signal IMD noise using the proposed IF power allocation. Transmission performance in IFoF transmission, the power of the nonlinear component affected by the signal power enhancement using the technique was verified experimentally by the EVM improvement can be considered as the main distortion, and the generated IMD is amplified without of the received signal. This technique was briefly proposed at the Asian communication being removed from the RU stage, which can continuously affect the signal distortion. and photonics (ACP) conference 2019 and is described in detail in this paper [16]. (a) (b) Figure 2. Figure 2. The schematic The schematic diagram diagram of ( of (a a)) common common pu public blic rad radio io interfac interfacee (CP (CPRI), RI), ( (bb ) IF ) IF over over fiber fiber (IFoF) (IFoF) transmission. transmission. In this paper, we propose a spectral power allocation technique dependent on the 2. Modulation-Based Adaptive Power Allocation modulation format and verify the feasibility of system linearity enhancement of multi- When the transmitted signal passes through the nonlinear devices or channel, the IFoF-based MFN. The signal spectral power was allocated for each IF band according to nonlinear noise is generated according to the channel transfer function, as shown in modulation order, which is dependent on the required signal-to-noise power ratio (SNR), Equation (1). The transfer function has different coefficients for each different order as a which can be converted from the error vector magnitude (EVM) requirement. Since the function of the frequency [11,12]. IMD noise power is affected by the transmitted signal power, when the total transmitted signal power is changed, 𝑦 𝑡, 𝑓 the𝑎 IMD𝑎 noise 𝑓 𝑥 power 𝑡 𝑎 wou 𝑓 𝑥 ld 𝑡 be𝑎significantly 𝑓 𝑥 𝑡 ∙∙∙ changed compar(1 ed ) to the signal power variation. We have experimentally demonstrated the reduction of IMD In MFN using CPRI, the signal modulation format is determined by the signal noise using the proposed IF power allocation. Transmission performance enhancement degradation by the wireless channel between RU-Device and the fixed additive white using the technique was verified experimentally by the EVM improvement of the received gaussian noise (AWGN) in the entire channel, and the signal power is allocated by the signal. This technique was briefly proposed at the Asian communication and photonics water-pouring algorithm as shown in Figure 3 [17]. (ACP) conference 2019 and is described in detail in this paper [16]. 2. Modulation-Based Adaptive Power Allocation When the transmitted signal passes through the nonlinear devices or channel, the nonlinear noise is generated according to the channel transfer function, as shown in Equation (1). The transfer function has different coefficients for each different order as a function of the frequency [11,12]. 2 3 y(t, f ) = a + a ( f )x(t) + a ( f )x(t) + a ( f )x(t) +    (1) 0 1 2 3 In MFN using CPRI, the signal modulation format is determined by the signal degra- dation by the wireless channel between RU-Device and the fixed additive white gaussian noise (AWGN) in the entire channel, and the signal power is allocated by the water-pouring algorithm as shown in Figure 3 [17]. Photonics 2021, 8, 2 4 of 9 Photonics 2020, 7, x FOR PEER REVIEW 4 of 9 Photonics 2020, 7, x FOR PEER REVIEW 4 of 9 Photonics 2020, 7, x FOR PEER REVIEW 4 of 9 Figure 3. Power loading with the water-pouring algorithm. Figure 3. Power loading with the water-pouring algorithm. Figure 3. Power loading with the water-pouring algorithm. Figure 3. Power loading with the water-pouring algorithm. However, in the IFoF transmission, the OFDM signals are upconverted to each IF and However, in the IFoF transmission, the OFDM signals are upconverted to each IF and However, in the IFoF transmission, the OFDM signals are upconverted to each IF are transmitted to the RU, so the IMD noise and AWGN are combined and distort the are transmitted to the RU, so the IMD noise and AWGN are combined and distort the and ar However, e transmitted in thto e IFoF tr the RU, ansmission, the so the IMD noise OFDM sign and AWGN als are are upconverted to each combined and distort IF an thed signal. Since the amplifier amplifies not only the signal power also noise power, the signal signal. signSince al. Since the th amplifier e amplifier amplifies amplifies not not only on the ly tsignal he sign power al power a also lnoise so noise powe power, the r, tsignal he signal are transmitted to the RU, so the IMD noise and AWGN are combined and distort the is transmitted to the device with the amplified noise. is transmitted is transmitted to the devi to the device with ce withe th the a amplified mplifinoise. ed noise. signal. Since the amplifier amplifies not only the signal power also noise power, the signal In other words, the MFN using the RoF system should transmit the signal to the RU In other In other wor words, d the s, the MFN MF using N usin the g the RoF Ro system F system should should tra transmit nsmi the t the si signal gna tolthe to the RU RU is transmitted to the device with the amplified noise. without distortion as much as possible. In the proposed technique, the signal is generated without without di distortion stortias on a much s muas ch a possible. s possibIn le. In th the pr e proposed technique, t oposed technique, the signal he sign is al generated is generated In other words, the MFN using the RoF system should transmit the signal to the RU without without su surplus rplus power. It power. It is pos is possible sible t to o reduce reduce t the he IMD IMD noise noise power gener power generated ated by t by the he without surplus power. It is possible to reduce the IMD noise power generated by the without distortion as much as possible. In the proposed technique, the signal is generated surplus surplus powe power r of the of thesign signal. al. As As shown shown in Fig in Figur ures 4 an es 4d 5 and show the 5 showspectrum of the signal the spectrum of the surplus power of the signal. As shown in Figures 4 and 5 show the spectrum of the signal without surplus power. It is possible to reduce the IMD noise power generated by the signal passing t passing hrough t through he non thelnonlinear inear chann channel el using using the p theo power wer lo loading ading and and us using ing p prroposed oposed passing through the nonlinear channel using the power loading and using proposed surplus power of the signal. As shown in Figures 4 and 5 show the spectrum of the signal techniques. techniques. techniques. passing through the nonlinear channel using the power loading and using proposed techniques. Figure 4. Power allocation with the proposed technique. Figure 4. Power allocation with the proposed technique. Figure 4. Power allocation with the proposed technique. Figure 4. Power allocation with the proposed technique. Figure 5. (a) Multiband orthogonal frequency-division multiplexing (OFDM) signal spectrum with- Figure 5. (a) Multiband orthogonal frequency-division multiplexing (OFDM) signal spectrum Figure 5. (a) Multiband orthogonal frequency-division multiplexing (OFDM) signal spectrum out power allocation, (b) signal spectrum with power loading, (c) signal spectrum with proposed without power allocation, (b) signal spectrum with power loading, (c) signal spectrum with without power allocation, (b) signal spectrum with power loading, (c) signal spectrum with Figure 5. (a) Multiband orthogonal frequency-division multiplexing (OFDM) signal spectrum technique, (d) signal spectrum comparison of power loading and proposed technique. proposed technique, (d) signal spectrum comparison of power loading and proposed technique. proposed technique, (d) signal spectrum comparison of power loading and proposed technique. without power allocation, (b) signal spectrum with power loading, (c) signal spectrum with proposed technique, (d) signal spectrum comparison of power loading and proposed technique. The IMD components are generated by device nonlinearity and fiber nonlinearity. The IMD components are generated by device nonlinearity and fiber nonlinearity. In The IMD components are generated by device nonlinearity and fiber nonlinearity. In In the case of fiber nonlinearity, the IMD2 noise is mainly the nonlinear component gen- the case of fiber nonlinearity, the IMD2 noise is mainly the nonlinear component the case of fiber nonlinearity, the IMD2 noise is mainly the nonlinear component The IMD components are generated by device nonlinearity and fiber nonlinearity. In erated by the mixing laser chirp and chromatic dispersion when the laser is directly generated by the mixing laser chirp and chromatic dispersion when the laser is directly generated by the mixing laser chirp and chromatic dispersion when the laser is directly the case of fiber nonlinearity, the IMD2 noise is mainly the nonlinear component modulated [18,19]. The effect of dispersion is increased by increasing the transmission modulated [18,19]. The effect of dispersion is increased by increasing the transmission modulated [18,19]. The effect of dispersion is increased by increasing the transmission generated by the mixing laser chirp and chromatic dispersion when the laser is directly distance. When the directly modulated signal suffered dispersion, the IMD2 component distance. When the directly modulated signal suffered dispersion, the IMD2 component distance. When the directly modulated signal suffered dispersion, the IMD2 component modulated [18,19]. The effect of dispersion is increased by increasing the transmission is generated greatly [3]. Therefore, in the short distance, the IMD components caused by is generated greatly [3]. Therefore, in the short distance, the IMD components caused by is generated greatly [3]. Therefore, in the short distance, the IMD components caused by distance. When the directly modulated signal suffered dispersion, the IMD2 component device nonlinearity, which are generated in RF amplifiers, a laser diode (LD), photodetector device nonlinearity, which are generated in RF amplifiers, a laser diode (LD), device nonlinearity, which are generated in RF amplifiers, a laser diode (LD), is generated greatly [3]. Therefore, in the short distance, the IMD components caused by (PD), and, etc. are mainly generated as IMD3. The IMD noise power is determined by the photodetector (PD), and, etc. are mainly generated as IMD3. The IMD noise power is photodetector (PD), and, etc. are mainly generated as IMD3. The IMD noise power is device nonlinearity, which are generated in RF amplifiers, a laser diode (LD), signal power. For example, the IMD3 noise power is expressed as the product of coefficient, determined by the signal power. For example, the IMD3 noise power is expressed as the determined by the signal power. For example, the IMD3 noise power is expressed as the photodetector (PD), and, etc. are mainly generated as IMD3. The IMD noise power is and the IMD3 noise power is three times of signal power [17]. That is, if the signal power product of coefficient, and the IMD3 noise power is three times of signal power [17]. That product of coefficient, and the IMD3 noise power is three times of signal power [17]. That determined by the signal power. For example, the IMD3 noise power is expressed as the decreases, the IMD3 noise power is reduced by three times the amount of the signal de- is, if the signal power decreases, the IMD3 noise power is reduced by three times the is, if the signal power decreases, the IMD3 noise power is reduced by three times the product of coefficient, and the IMD3 noise power is three times of signal power [17]. That crease. Therefore, a dramatic reduction in IMD noise power can be achieved by controlling amount of the signal decrease. Therefore, a dramatic reduction in IMD noise power can amount of the signal decrease. Therefore, a dramatic reduction in IMD noise power can is, if the signal power decreases, the IMD3 noise power is reduced by three times the the signal power [14]. In multi-IFoF transmission, there are two different IMD components. be achieved by controlling the signal power [14]. In multi-IFoF transmission, there are two be achieved by controlling the signal power [14]. In multi-IFoF transmission, there are two amount of the signal decrease. Therefore, a dramatic reduction in IMD noise power can The first is the intra-band distortion generated by beating between subcarriers within a different IMD components. The first is the intra-band distortion generated by beating different IMD components. The first is the intra-band distortion generated by beating be achieved by controlling the signal power [14]. In multi-IFoF transmission, there are two single IF band, and the other is inter-band distortion due to beating between subcarriers between subcarriers within a single IF band, and the other is inter-band distortion due to between subcarriers within a single IF band, and the other is inter-band distortion due to different IMD components. The first is the intra-band distortion generated by beating from other bands. In other words, in the IFoF transmission, the intra-band and inter-band beating between subcarriers from other bands. In other words, in the IFoF transmission, beating between subcarriers from other bands. In other words, in the IFoF transmission, between subcarriers within a single IF band, and the other is inter-band distortion due to IMD components are combined to generate a complex IMD component. In multi-IFoF the intra-band and inter-band IMD components are combined to generate a complex IMD the intra-band and inter-band IMD components are combined to generate a complex IMD beating between subcarriers from other bands. In other words, in the IFoF transmission, transmission, each IF band passes through different channels and supports different de- component. In multi-IFoF transmission, each IF band passes through different channels component. In multi-IFoF transmission, each IF band passes through different channels the intra-band and inter-band IMD components are combined to generate a complex IMD vices, where each IF band signal is modulated in a different format. Because different and supports different devices, where each IF band signal is modulated in a different and supports different devices, where each IF band signal is modulated in a different component. In multi-IFoF transmission, each IF band passes through different channels and supports different devices, where each IF band signal is modulated in a different Photonics 2021, 8, 2 5 of 9 Photonics 2020, 7, x FOR PEER REVIEW 5 of 9 format. Because different modulation formats have different BERs in the same SNR environment, the modulation formats require different EVM values, as shown in below modulation formats have different BERs in the same SNR environment, the modulation Table 1 [20]. formats require different EVM values, as shown in below Table 1 [20]. Table 1. Modulation format dependent error vector magnitude (EVM) and signal-to-noise power Table 1. Modulation format dependent error vector magnitude (EVM) and signal-to-noise power ratio (SNR) requirements in long term evolution (LTE) standard [20]. ratio (SNR) requirements in long term evolution (LTE) standard [20]. Format QPSK 16QAM 64QAM 256QAM Format QPSK 16QAM 64QAM 256QAM EVM (%) 17.5 12.5 8 4 SNR ( EVM dB) (%) 15 17.5 18 12.5 22 8 28 4 SNR (dB) 15 18 22 28 The high order modulation format requires high received SNR. On the contrary, the lower-order modulation format requires a relatively low received SNR. Because of the The high order modulation format requires high received SNR. On the contrary, the relationship between the EVM and the SNR, required SNRs are about a minimum of 15 lower-order modulation format requires a relatively low received SNR. Because of the dB for quadrature phase-shift keying (QPSK) and 28 dB for 256 quadrature amplitude relationship between the EVM and the SNR, required SNRs are about a minimum of modulation (QAM) in the transmission system. The higher the modulation order, the 15 dB for quadrature phase-shift keying (QPSK) and 28 dB for 256 quadrature amplitude more signal power is required. In MFN, the signal is transmitted through the fiber from modulation (QAM) in the transmission system. The higher the modulation order, the more DU to RU, amplified at the RU, and then sent to each device or user over the wireless signal power is required. In MFN, the signal is transmitted through the fiber from DU to channel. The dominant noise which degrades SNR occurs mainly in the nonlinear device RU, amplified at the RU, and then sent to each device or user over the wireless channel. or wireless channel. When the device receives a signal, the device periodically measures The dominant noise which degrades SNR occurs mainly in the nonlinear device or wireless the channel. Then, the device sends a channel quality indicator (CQI) containing the channel. When the device receives a signal, the device periodically measures the channel. measured SNR from the de Then, the device sends avice. Since the modula channel quality indicator tion format of the (CQI) containing signthe al is dete measur rmined ed SNR by the SNR from the device. from CQI [2 Since 1], thethat is, the t modulationrformat ansmitted signal of the signal from the is determined DU to the RU needs by the SNR from CQI [21], that is, the transmitted signal from the DU to the RU needs only an SNR that is only an SNR that is enough to satisfy the EVM requirements until the antenna where the sign enough al is amp to lif satisfy ied. The therEVM efore, rin t equir heements IFoF tran until smission, the the antenna spectr wher al powers e the signal of t is he band amplified. s Therefore, in the IFoF transmission, the spectral powers of the bands are allocated based are allocated based on the required SNRs according to the modulation format of each on the required SNRs according to the modulation format of each band, and the allocated band, and the allocated signal is transmitted. If the signal power per IF is allocated by the signal is transmitted. If the signal power per IF is allocated by the SNR required by each SNR required by each modulation format, the total power of the signal can be reduced, modulation format, the total power of the signal can be reduced, and the power of the and the power of the nonlinear component is also reduced. When the SNR required by nonlinear component is also reduced. When the SNR required by each modulation format each modulation format is allocated to each IF band, the power of the signal is adjusted, is allocated to each IF band, the power of the signal is adjusted, and the IMD power can and the IMD power can also be decreased, as shown in Figure 6. also be decreased, as shown in Figure 6. (a) (b) Figure 6. Signal spectrum with the proposed technique of (a) transmitted signal and (b) received Figure 6. Signal spectrum with the proposed technique of (a) transmitted signal and (b) received signal. signal. The operating principle of the proposed modulation based adaptive power allocation (MBAPA) technique is as follows: When the initial power of all bands has the same power, The operating principle of the proposed modulation based adaptive power allocation the power of the signal band of 256QAM, the highest order, is maintained. Based on the (MBAPA) technique is as follows: When the initial power of all bands has the same power, the power of 256QAM signal the signa power l band of , the excess 256QAM, th power e highest order, in each band is is m reduced aintaine by d the . Br aequir sed on the ed SNR difference for each order. In other words, the power of the QPSK signal is reduced by 13 dB, 256QAM signal power, the excess power in each band is reduced by the required SNR which is the difference from the required SNR of 256QAM. Even if the surplus power of difference for each order. In other words, the power of the QPSK signal is reduced by 13 the signal is reduced, the required EVM can be satisfied so that the signal to which the dB, which is the difference from the required SNR of 256QAM. Even if the surplus power MBAPA technology to control surplus power is applied can be received. of the signal is reduced, the required EVM can be satisfied so that the signal to which the MBAPA technology to control surplus power is applied can be received. Photonics 2021, 8, 2 6 of 9 Photonics 2020, 7, x FOR PEER REVIEW 6 of 9 Photonics 2020, 7, x FOR PEER REVIEW 6 of 9 3. Experiments and Results 3. Experiments and Results The feasibility of the proposed MBAPA technique was verified experimentally in The feasibility of the proposed MBAPA technique was verified experimentally in 3. Experiments and Results multi-IFoF transmission by intensity-modulated and direct detection (IM/DD). In the multi-IFoF transmission by intensity-modulated and direct detection (IM/DD). In the The feasibility of the proposed MBAPA technique was verified experimentally in multi- experiment, as shown in Figure 7, randomly generated bits were mapped to symbols experiment, as shown in Figure 7, randomly generated bits were mapped to symbols IFoF transmission by intensity-modulated and direct detection (IM/DD). In the experiment, depending on each modulation format of subcarriers (QPSK, 16QAM, 64QAM, and depending on each modulation format of subcarriers (QPSK, 16QAM, 64QAM, and as shown in Figure 7, randomly generated bits were mapped to symbols depending on each 256QAM). Bandwidth and spacing of subcarriers were set for generating OFDM signal. 256QAM). Bandwidth and spacing of subcarriers were set for generating OFDM signal. modulation format of subcarriers (QPSK, 16QAM, 64QAM, and 256QAM). Bandwidth and To have 20 MHz bandwidth per single IF, the number of subcarriers per band was 1200. To have 20 MHz bandwidth per single IF, the number of subcarriers per band was 1200. spacing of subcarriers were set for generating OFDM signal. To have 20 MHz bandwidth The subcarrier spacing of the OFDM signal was set to be 15 kHz. The spacing between The subcarrier spacing of the OFDM signal was set to be 15 kHz. The spacing between per single IF, the number of subcarriers per band was 1200. The subcarrier spacing of the bands was set as 10 MHz, and 24 and 48 multi-IF bands were generated. The optical signal bands was set as 10 MHz, and 24 and 48 multi-IF bands were generated. The optical signal OFDM signal was set to be 15 kHz. The spacing between bands was set as 10 MHz, and 24 was transmitted through 10 km and 20 km optical single-mode-fiber. The received optical was transmitted through 10 km and 20 km optical single-mode-fiber. The received optical and 48 multi-IF bands were generated. The optical signal was transmitted through 10 km power at the photodetector (PD) was about −1.5 dBm, and the RF spectra of the signal, power at the photodetector (PD) was about −1.5 dBm, and the RF spectra of the signal, and 20 km optical single-mode-fiber. The received optical power at the photodetector (PD) which was converted from optical to electrical signal at PD, were measured. The proposed which was converted from optical to electrical signal at PD, were measured. The proposed was about 1.5 dBm, and the RF spectra of the signal, which was converted from optical MBAPA signal can suppress the nonlinearity of the system. It was assumed that the IMD MBAPA signal can suppress the nonlinearity of the system. It was assumed that the IMD to electrical signal at PD, were measured. The proposed MBAPA signal can suppress the components were generated by transfer functions of RF amplifier and LD and by power components were generated by transfer functions of RF amplifier and LD and by power nonlinearity of the system. It was assumed that the IMD components were generated by saturation at PD. saturation at PD. transfer functions of RF amplifier and LD and by power saturation at PD. Figure 7. Experimental setup. Figure 7. Experimental setup. Figure 7. Experimental setup. Figure 8a,b shows the transmitted signal spectrum with the conventional technique Figure 8a,b shows the transmitted signal spectrum with the conventional technique Figure 8a,b shows the transmitted signal spectrum with the conventional technique and with the proposed technique, respectively. After 10 km transmission, two nonlinear and with the proposed technique, respectively. After 10 km transmission, two nonlinear and with the proposed technique, respectively. After 10 km transmission, two nonlinear distortions were generated and interfered with each other in a spectrum, as shown in distortions were generated and interfered with each other in a spectrum, as shown in distortions were generated and interfered with each other in a spectrum, as shown in Figure 9. Figure 9. Figure 9. (a) (b) (a) (b) Figure 8. The power spectral density of transmitted signal (a) without the technique and (b) with the Figure 8. The power spectral density of transmitted signal (a) without the technique and (b) with Figure 8. The power spectral density of transmitted signal (a) without the technique and (b) with the technique l technique l. . the technique l. Photonics 2021, 8, 2 7 of 9 Photonics 2020, 7, x FOR PEER REVIEW 7 of 9 Photonics 2020, 7, x FOR PEER REVIEW 7 of 9 Figure 9. Received 24-IF bands signal spectrum after 10 km transmission. Figure 9. Figure Recei 9. Received ved 24-IF24-IF bandbands s signal spe signal ctr spectr um after 10 km transmission. um after 10 km transmission. In Figure 9, using the proposed technique, the IMD power was reduced by about 10 In Figure 9, using the proposed technique, the IMD power was reduced by about In Figure 9, using the proposed technique, the IMD power was reduced by about 10 dB. Figure 10 shows the constellations of the first IF band, which was modulated by QPSK, 10 dB. Figure 10 shows the constellations of the first IF band, which was modulated by dB. Figure 10 shows the constellations of the first IF band, which was modulated by QPSK, and the third IF band modulated by 256QAM after a 10 km transmission with 24-IFs. QPSK, and the third IF band modulated by 256QAM after a 10 km transmission with 24-IFs. and the third IF band modulated by 256QAM after a 10 km transmission with 24-IFs. (a) (b) (c) (d) (a) (b) (c) (d) Figure 10. Constellation of the received signal after 10 km transmission with 24-IFs (a) quadrature phase-shift keying Figure 10. Constellation of the received signal after 10 km transmission with 24-IFs (a) quadrature phase-shift keying Figure 10. Constellation of the received signal after 10 km transmission with 24-IFs (a) quadrature phase-shift keying (QPSK) without the technique, (b) QPSK with the technique, (c) 256QAM without the technique and (d) 256QAM with (QPSK) without the technique, (b) QPSK with the technique, (c) 256QAM without the technique and (d) 256QAM with (QPSK) without the technique, (b) QPSK with the technique, (c) 256QAM without the technique and (d) 256QAM with the technique. the technique. the technique. Figure 10a,b shows QPSK constellation diagrams, respectively, without the proposed Figure 10a,b shows QPSK constellation diagrams, respectively, without the proposed Figure 10a,b shows QPSK constellation diagrams, respectively, without the proposed technique and with the proposed technique. In the case of QPSK, the EVM value without technique and with the proposed technique. In the case of QPSK, the EVM value with- technique and with the proposed technique. In the case of QPSK, the EVM value without the technique was 7%, and The EVM value with the technique was 11%. QPSK is a low- out the technique was 7%, and The EVM value with the technique was 11%. QPSK is a the technique was 7%, and The EVM value with the technique was 11%. QPSK is a low- order modulation that requires less signal power. Therefore, the proposed technique low-order modulation that requires less signal power. Therefore, the proposed technique order modulation that requires less signal power. Therefore, the proposed technique reduces the surplus of the signal power, improves the SNR, which causes deterioration of reduces the surplus of the signal power, improves the SNR, which causes deterioration reduces the surplus of the signal power, improves the SNR, which causes deterioration of the EVM. The EVM values of both signals satisfy the EVM requirement of the QPSK of the EVM. The EVM values of both signals satisfy the EVM requirement of the QPSK the EVM. The EVM values of both signals satisfy the EVM requirement of the QPSK format. Figure 10c,d is the 256QAM constellation; (c) is constellation without the proposed format. Figure 10c,d is the 256QAM constellation; (c) is constellation without the proposed format. Figure 10c,d is the 256QAM constellation; (c) is constellation without the proposed technique, and (d) is with the proposed technique. In the case of 256QAM, unlike QPSK, technique, and (d) is with the proposed technique. In the case of 256QAM, unlike QPSK, technique, and (d) is with the proposed technique. In the case of 256QAM, unlike QPSK, the proposed technique improved EVM performance from 10.8% to 3.5%. The received the proposed technique improved EVM performance from 10.8% to 3.5%. The received the proposed technique improved EVM performance from 10.8% to 3.5%. The received signal without the technique could not satisfy the EVM requirement to receive by signal without the technique could not satisfy the EVM requirement to receive by nonlinear signal without the technique could not satisfy the EVM requirement to receive by distortion nonlinear distortion noise. On the other ha noise. On the other hand, with the pr nd, wi oposed th technique, the proposed techniq the EVM requir ue, tement he EVis M nonlinear distortion noise. On the other hand, with the proposed technique, the EVM satisfied, requirement so the is signal satisfied with , so the thtechnique e signal wit is h suitable the tec for hniq multi-IFoF ue is suit transmission. able for multIn i-IFoF ad- requirement is satisfied, so the signal with the technique is suitable for multi-IFoF transmission. In addition, in order to verify the effect of the technique on the fiber dition, in order to verify the effect of the technique on the fiber nonlinearity, experiments transmission. In addition, in order to verify the effect of the technique on the fiber wer nonlinearity, e conducted experiments by increasing were cond the transmission ucted by incre distance. asing the transmission distance. nonlinearity, experiments were conducted by increasing the transmission distance. The The dispersion effect is dispersion effect is mormore sever e severe as incr e as easing increa the sing transmission the transmis distance. sion dist Asance already . As The dispersion effect is more severe as increasing the transmission distance. As written above, the dispersion mixed with the chirp effect of laser-generated the IMD2 more. already written above, the dispersion mixed with the chirp effect of laser-generated the already written above, the dispersion mixed with the chirp effect of laser-generated the As Figure 11 shows, the constellation of the 20 km transmitted signal with the same IMD2 more. As Figure 11 shows, the constellation of the 20 km transmitted signal with IMD2 more. As Figure 11 shows, the constellation of the 20 km transmitted signal with received optical power about 1.5 dBm. Similar to Figure 10, the EVM from (a) without the the same received optical power about −1.5 dBm. Similar to Figure 10, the EVM from (a) the same received optical power about −1.5 dBm. Similar to Figure 10, the EVM from (a) technique to (b) with the technique deteriorated from 6.5% to 13.4%, and both EVM values without the technique to (b) with the technique deteriorated from 6.5% to 13.4%, and both without the technique to (b) with the technique deteriorated from 6.5% to 13.4%, and both satisfied the EVM requirement. In addition, when using the proposed technique from EVM values satisfied the EVM requirement. In addition, when using the proposed EVM values satisfied the EVM requirement. In addition, when using the proposed (c) to (d), the EVM improved from 9.2% to 4%. It was proven that the IMD2 component technique from (c) to (d), the EVM improved from 9.2% to 4%. It was proven that the technique from (c) to (d), the EVM improved from 9.2% to 4%. It was proven that the generated by the dispersion could be effectively reduced by the proposed technique. IMD2 component generated by the dispersion could be effectively reduced by the IMD2 component generated by the dispersion could be effectively reduced by the proposed technique. proposed technique. Photonics 2021, 8, 2 8 of 9 Photonics 2020, 7, x FOR PEER REVIEW 8 of 9 Photonics 2020, 7, x FOR PEER REVIEW 8 of 9 (a) (b) (c) (d) (a) (b) (c) (d) Figure 11. Constellation of the received signal after 20 km transmission with 24-IFs (a) QPSK without the technique, (b) Figure 11. Constellation of the received signal after 20 km transmission with 24-IFs (a) QPSK without the technique, Figure 11. Constellation of the received signal after 20 km transmission with 24-IFs (a) QPSK without the technique, (b) QPSK with the technique, (c) 256QAM without the technique and (d) 256QAM with the technique. (b) QPSK with the technique, (c) 256QAM without the technique and (d) 256QAM with the technique. QPSK with the technique, (c) 256QAM without the technique and (d) 256QAM with the technique. Furthermore, an experiment with increasing the total bandwidth of the transmission Furthermore, an experiment with increasing the total bandwidth of the transmission Furthermore, an experiment with increasing the total bandwidth of the transmission signal was conducted. The transmission signal was changed from a signal using 24-IF signal was conducted. The transmission signal was changed from a signal using 24-IF signal was conducted. The transmission signal was changed from a signal using 24-IF bands to a signal using 48 IF bands. As the number of IF bands used increases, the total bands to a signal using 48 IF bands. As the number of IF bands used increases, the total bands to a signal using 48 IF bands. As the number of IF bands used increases, the total bandwidth becomes wider, and the effects of the inter-band IMD2 and IMD3 on the signal bandwidth becomes wider, and the effects of the inter-band IMD2 and IMD3 on the signal bandwidth becomes wider, and the effects of the inter-band IMD2 and IMD3 on the signal are significantly increased. The constellation of the 48-IF signal is shown in Figure 12 after are significantly increased. The constellation of the 48-IF signal is shown in Figure 12 after are significantly increased. The constellation of the 48-IF signal is shown in Figure 12 after 10 km transmission. As shown in Figure 12a,b, EVM became worse from 7.1% to 10.8% 10 km transmission. As shown in Figure 12a,b, EVM became worse from 7.1% to 10.8% 10 km transmission. As shown in Figure 12a,b, EVM became worse from 7.1% to 10.8% using the proposed technique, but the signals satisfied the EVM requirement for QPSK. using the proposed technique, but the signals satisfied the EVM requirement for QPSK. In using the proposed technique, but the signals satisfied the EVM requirement for QPSK. In addition, when using the proposed technique for 256QAM from (c) to (d), the EVM was addition, when using the proposed technique for 256QAM from (c) to (d), the EVM was In addition, when using the proposed technique for 256QAM from (c) to (d), the EVM was improved from 10.2% to 3.4%. It was demonstrated that the proposed technique mitigates improved from 10.2% to 3.4%. It was demonstrated that the proposed technique mitigates improved from 10.2% to 3.4%. It was demonstrated that the proposed technique mitigates the effects of complex inter-band IMD. the effects of complex inter-band IMD. the effects of complex inter-band IMD. (b) (c) (d) (a) (b) (c) (d) (a) Figure 12. Constellation of the received signal after 10 km transmission with 48-IFs (a) QPSK without the technique, Figure 12. Constellation of the received signal after 10 km transmission with 48-IFs (a) QPSK without the technique, (b) Figure 12. Constellation of the received signal after 10 km transmission with 48-IFs (a) QPSK without the technique, (b) (b) QPSK with the technique, (c) 256QAM without the technique and (d) 256QAM with the technique. QPSK with the technique, (c) 256QAM without the technique and (d) 256QAM with the technique. QPSK with the technique, (c) 256QAM without the technique and (d) 256QAM with the technique. 4. Conclusions 4. Conclusions 4. Conclusions We proposed the modulation format based spectral power allocation technique to We proposed the modulation format based spectral power allocation technique to We proposed the modulation format based spectral power allocation technique to mitigate the IMD noise and improve channel linearity in multi-IFoF transmission. For the mitigate the IMD noise and improve channel linearity in multi-IFoF transmission. For the mitigate the IMD noise and improve channel linearity in multi-IFoF transmission. For the 24-IFoF 10 km transmission system, the proposed technique obtained 7% of EVM 24-IFoF 10 km transmission system, the proposed technique obtained 7% of EVM im- 24-IFoF 10 km transmission system, the proposed technique obtained 7% of EVM improvement when 256QAM was used, and also, the MBAPA technique suppressed IMD provement when 256QAM was used, and also, the MBAPA technique suppressed IMD improvement when 256QAM was used, and also, the MBAPA technique suppressed IMD power around 10 dB. In addition, the system linearity was improved for a 48-IFoF 10 km power around 10 dB. In addition, the system linearity was improved for a 48-IFoF 10 km power around 10 dB. In addition, the system linearity was improved for a 48-IFoF 10 km transmission system and a 24-IFoF 10 km transmission system. The proposed technique transmission system and a 24-IFoF 10 km transmission system. The proposed technique in transmission system and a 24-IFoF 10 km transmission system. The proposed technique in the reduction of IMD power effectively will be the key solution for the 5G MFN the reduction of IMD power effectively will be the key solution for the 5G MFN supporting in the reduction of IMD power effectively will be the key solution for the 5G MFN supporting multi-user. multi-user. supporting multi-user. Author Contributions: Conceptualization, I.H.; Data curation, I.H.; Formal analysis, I.H. and H.- Author Contributions: Conceptualization, I.H.; Data curation, I.H.; Formal analysis, I.H. and H.-J.P.; Author Contributions: Conceptualization, I.H.; Data curation, I.H.; Formal analysis, I.H. and H.- J.P.; Project administration, S.-K.H.; Writing—original draft, I.H.; Writing—review & editing, S.- Project administration, S.-K.H.; Writing—original draft, I.H.; Writing—review & editing, S.-M.K. J.P.; Project administration, S.-K.H.; Writing—original draft, I.H.; Writing—review & editing, S.- M.K. All authors have read and agreed to the published version of the manuscript. All authors have read and agreed to the published version of the manuscript. M.K. All authors have read and agreed to the published version of the manuscript. Funding: This work was supported by the National Research Foundation of Korea (NRF) grant Funding: This work was supported by the National Research Foundation of Korea (NRF) grant Funding: This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korean government (MSIT; Ministry of Science and ICT) (No. NRF- funded by the Korean government (MSIT; Ministry of Science and ICT) (No. NRF-2019R1A2C3007934). funded by the Korean government (MSIT; Ministry of Science and ICT) (No. NRF- 2019R1A2C3007934). 2019R1A2C3007934). Conflicts of Interest: The authors declare no conflict of interest. Conflicts of Interest: The authors declare no conflicts of interest. Conflicts of Interest: The authors declare no conflicts of interest. Photonics 2021, 8, 2 9 of 9 References 1. Tanaka, K.; Agata, A. Next-generation optical access networks for C-RAN. In Proceedings of the Optical Fiber Communication Conference, Los Angeles, CA, USA, 22–26 March 2015. Paper Tu2E.1. 2. Common Public Radio Interface. CPRI Specification V7.0. Stand. Doc. Specif. 2015. Available online: http://www.cpri.info/ downloads/CPRI_v_7_0_2015-10-09.pdf (accessed on 2 November 2020). 3. Radio-over-fibre (RoF) technologies and their applications. In ITU-T, Suppl. 55, G-Series Recommendations; ITU: Geneva, Switzer- land, 2015. 4. Ishimura, S.; Bekkali, A.; Tanaka, K.; Nishimura, K.; Suzuki, M. 1.032-Tb/s CPRI-equivalent data rate transmission using IF-over-Fiber system for high capacity mobile fronthaul links. In Proceedings of the 2017 European Conference on Optical Communication (ECOC), Gothenburg, Sweden, 17–21 September 2017. Paper Th.PDP.B.6. 5. Gomes, N.J.; Morant-Perez, M.; Alphones, A.; Cabon, B.; Mitchell, J.E.; Lethien, C.; Csornyei, M.; Stoehr, A.; Iezekiel, S. Radio over fibre transport for the support of broadband services. Invit. Pap. OSA J. Opt. Net. 2009, 8, 156–178. [CrossRef] 6. Han, C.; Sung, M.; Cho, S.-H.; Chung, H.S.; Kim, S.M.; Lee, J.H. Performance improvement of multi-IFoF-based mobile fronthaul using dispersion-induced distortion mitigation with IF optimization. J. Lightw. Technol. 2016, 34, 4772–4778. [CrossRef] 7. Sung, M.; Cho, S.-H.; Kim, J.; Lee, J.K.; Lee, J.H.; Chung, H.S. Demonstration of IFoF-based mobile fronthaul in 5G prototype with 28-GHz millimeter wave. J. Lightw. Technol. 2018, 36, 601–609. [CrossRef] 8. Han, C.; Cho, S.; Chung, H.S.; Lee, J.H. Linearity improvement of directly-modulated multi-IF-over-fibre LTE-A mobile fronthaul link using shunt diode predistorter. In Proceedings of the 2015 European Conference on Optical Communication (ECOC), Valencia, Spain, 27 September–1 October 2015; p. We.4.4.4. 9. Li, J.; Zhang, Y.C.; Yu, S.; Jiang, T.; Xie, Q.; Gu, W. Third-order intermodulation distortion elimination of microwave photonics link based on integrated dual-drive dual-parallel Mach–Zehnder modulator. Opt. Exp. 2013, 38, 4285–4287. [CrossRef] [PubMed] 10. Peccarelli, N.; Fulton, C. Mitigation of nonlinear spurious products using least mean-square (LMS). In Proceedings of the GOMACTech, Miami, FL, USA, 12–15 March 2018. 11. Jiang, W.; Tan, Q.; Qin, W.; Liang, D.; Li, X.; Ma, H.; Zhu, Z. A linearization analog photonic link with high third-order intermodulation distortion suppression based on dual-parallel Mach–Zehnder modulator. IEEE Photon. J. 2015, 7, 1–8. [CrossRef] 12. Park, H.J.; Ha, I.H.; Han, S.K. Distortion mitigation in IFoF system employing frequency-separated DPD. In Proceedings of the MWP, Toulouse France, 22–25 October 2018. 13. Najarro, A.C.; Kim, S.M. Predistortion for Frequency-Dependent Nonlinearity of a Laser in RoF Systems. J. Inf. Commun. Converg. Eng. 2016, 14, 147–152. [CrossRef] 14. Ha, I.H.; Park, H.J.; Han, S.K. Nonlinearity mitigation in multi-IFoF-based mobile fronthaul transmission using spectral power shaping. Opt. Comm. 2019, 437, 382–387. [CrossRef] 15. Sonalkar, R.; Shively, R. An efficient bit-loading algorithms for DMT applications. In Proceedings of the GLOBECOM, Sydney, NSW, Australia, 8–12 November 1998; pp. 2683–2688. 16. Ha, I.H.; Park, H.J.; Han, S.K. Nonlinear Distortion Mitigation Technique using Modulation Format Dependent Spectral Power Al- location in Multi-IFoF system. In Asia Communications and Photonics Conference; Optical Society of America: Washington, DC, USA, 17. Gallager, R.G. Information Theory and Reliable Communication; Wiley: New York, NY, USA, 1968. 18. Brendel, F.; Poëtte, J.; Cabon, B.; Zwick, T.; van Dijk, F.; Lelarge, F.; Accard, A. Chromatic dispersion in 60 GHz radio-over-fiber networks based on mode-locked laser. J. Lightw. Technol. 2011, 29, 3810–3816. [CrossRef] 19. Lim, C.; Lee, K.-L.; Nirmalathas, A.; Novak, D.; Waterhouse, R. Impact of chromatic dispersion on 60 GHz radio-over-fiber transmission. In Proceedings of the IEEE 21st LEOS Annual Meeting, Newport Beach, CA, USA, 9–13 November 2008; pp. 89–90. 20. 3GPP TS 36.104 Version 15.3.0 Release 14, 3GPP. 2018. 21. Hamed, M.M.; Shukry, S.; El-Mahallawy, M.S.; El-Ramly, S.H. Modified earliest deadline first scheduling with channel quality indicator for downlink real-time traffic in LTE networks. In Proceedings of the Third International Conference on e-Technologies and Networks for Development (ICeND2014), Beirut, Lebanon, 29 April–1 May 2014. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Photonics Multidisciplinary Digital Publishing Institute

Nonlinear Distortion Mitigation in Multi-IF over Fiber Transmission Using Modulation-Based Adaptive Power Allocation

Photonics , Volume 8 (1) – Dec 22, 2020

Loading next page...
 
/lp/multidisciplinary-digital-publishing-institute/nonlinear-distortion-mitigation-in-multi-if-over-fiber-transmission-BDqIh0lLMx
Publisher
Multidisciplinary Digital Publishing Institute
Copyright
© 1996-2020 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
2304-6732
DOI
10.3390/photonics8010002
Publisher site
See Article on Publisher Site

Abstract

hv photonics Article Nonlinear Distortion Mitigation in Multi-IF over Fiber Transmission Using Modulation-Based Adaptive Power Allocation Inho Ha, Hyoung-Joon Park, Soo-Min Kang and Sang-Kook Han * Department of Electrical and Electronics Engineering, Yonsei University, Seoul 03722, Korea; 91hainho@yonsei.ac.kr (I.H.); phjokokok@yonsei.ac.kr (H.-J.P.); roemee817@yonsei.ac.kr (S.-M.K.) * Correspondence: skhan@yonsei.ac.kr Abstract: We propose a modulation-based adaptive power allocation (MBAPA) technique for nonlin- ear distortion mitigation in intermediate frequency over fiber (IFoF) systems. The technique allocates the spectral power of each IF band according to the required signal-to-noise power ratio (SNR) of the modulation format. To demonstrate the performance of the technique, transmission experiments were performed in 10 km and 20 km with 24-IF bands using OFDM signals. The feasibility of the proposed MBAPA technique was experimentally verified by reducing inter-modulation distortion (IMD) power and enhancing channel linearity. Keywords: intermodulation distortion; multi-IFoF transmission; mobile fronthaul network; power allocation 1. Introduction With the advent of diverse applications and devices, the number of devices per single user and the required data capacity has been increasing significantly, leading to an expo- Citation: Ha, I.; Park, H.-J.; nential growth in mobile data traffic [1]. As a result, due to the limitation of data traffic per Kang, S.-M.; Han, S.-K. Nonlinear Dis- cell and the increase in the amount of data required per device, cell coverage is reduced to tortion Mitigation in Multi-IF over Fiber small cells. Many cells, including small cells, were configured to support the required data Transmission Using Modulation-Based traffic. To reduce capital expenditure (CAPEX) and operating expenditures (OPEX), the Adaptive Power Allocation. Photon- radio access network (RAN) structure changed from a distributed-RAN (D-RAN) structure ics 2021, 8, 2. https://dx.doi.org/ to a centralized RAN (C-RAN) structure in which digital units (DU) are separated from 10.3390/photonics8010002 the distributed cell sites. Because communication was required between the DU and the cell site where there are only radio units (RU) available, the common public radio inter- Received: 13 November 2020 face (CPRI) has been standardized and used as the digital optical transmission protocol Accepted: 19 December 2020 between the DU and the RU [1,2]. However, when the signal is converted for digital Published: 22 December 2020 transmission, the amount of transmission data is dynamically increased. For example, long term evolution (LTE, 4th generation) signal with 20 MHz bandwidth requires 2.457 Publisher’s Note: MDPI stays neu- Gbps speed of CPRI signal in digital transmission [2]. Likewise, the bandwidth required tral with regard to jurisdictional claims in 5G (5th generation) increases further, and the CPRI transmission capacity increases in published maps and institutional significantly. Thus, there is a limitation in digital transmission systems as the data re- affiliations. quired by the user increases in the current mobile fronthaul network (MFN). In order to prevent data capacity enlargement and support the required data traffic, Radio over fiber (RoF) transmission as an analog transmission has been proposed as an alternative technol- Copyright: © 2020 by the authors. Li- ogy [1–5]. The RoF transmission has higher spectral efficiency than the CPRI. The enhanced censee MDPI, Basel, Switzerland. This CPRI (eCPRI) technique that developed CPRI has been proposed. In the case of eCPRI, article is an open access article distributed data enlargement is reduced due to ethernet transmission through packet-based optical under the terms and conditions of the transmission. However, since eCPRI technology is a packet-based ethernet transmission, it Creative Commons Attribution (CC BY) is necessary to convert the data for mobile communication. It makes the system complexity license (https://creativecommons.org/ at the RU increase. In the RoF transmission, there is no process of digitizing the signal, licenses/by/4.0/). Photonics 2021, 8, 2. https://dx.doi.org/10.3390/photonics8010002 https://www.mdpi.com/journal/photonics Photonics 2020, 7, x FOR PEER REVIEW 2 of 9 CPRI has been proposed. In the case of eCPRI, data enlargement is reduced due to ethernet transmission through packet-based optical transmission. However, since eCPRI technology is a packet-based ethernet transmission, it is necessary to convert the data for Photonics 2021, 8, 2 2 of 9 mobile communication. It makes the system complexity at the RU increase. In the RoF transmission, there is no process of digitizing the signal, so the data capacity enlargement has not occurred. Accordingly, the major drawback of CPRI, data capacity enlargement, so the data capacity enlargement has not occurred. Accordingly, the major drawback of does not occur in the RoF transmission. Moreover, to use frequency resources efficiently, CPRI, data capacity enlargement, does not occur in the RoF transmission. Moreover, to use multiband intermediate frequency over fiber (multi-IFoF) transmission has been frequency resources efficiently, multiband intermediate frequency over fiber (multi-IFoF) proposed and studied [4,6]. The RoF transmission, utilizing an intermediate carrier transmission has been proposed and studied [4,6]. The RoF transmission, utilizing an frequency, is the most bandwidth-efficient technique in the MFN because it preserves intermediate carrier frequency, is the most bandwidth-efficient technique in the MFN spectral bandwidths of the signals. For transmitting high data capacity, CPRI requires because it preserves spectral bandwidths of the signals. For transmitting high data capacity, transceivers and WDM components because of data capacity enlargement, but only a CPRI requires transceivers and WDM components because of data capacity enlargement, single analog optical transceiver is required in the multi-IFoF transmission. Therefore, the but only a single analog optical transceiver is required in the multi-IFoF transmission. IFoF transmission could significantly reduce the required transmission capacity [7]. Therefore, the IFoF transmission could significantly reduce the required transmission ca- Though the multi-IFoF transmission has a high-efficiency of frequency resource and has pacity [7]. Though the multi-IFoF transmission has a high-efficiency of frequency resource no need for an additional device to convert digital form, IFoF transmission is vulnerable and has no need for an additional device to convert digital form, IFoF transmission is to noise, especially intermodulation distortion (IMD) noise, compared to the digital vulnerable to noise, especially intermodulation distortion (IMD) noise, compared to the transmission. Because of the difference in the receiving technique of the analog signal and digital transmission. Because of the difference in the receiving technique of the analog the digital signal. Several techniques have been used in previous research to reduce IMD signal and the digital signal. Several techniques have been used in previous research to noise effects [6–14]. One technique is the equalizer using many memory taps, another reduce IMD noise effects [6–14]. One technique is the equalizer using many memory taps, technique is to pre-distort the signal to mitigate the nonlinear channel effect, and the other another technique is to pre-distort the signal to mitigate the nonlinear channel effect, and is employing the additional external modulators into the system [8–11]. However, as the the other is employing the additional external modulators into the system [8–11]. However, number of devices in the cell increases and the required bandwidth of each device gets as the number of devices in the cell increases and the required bandwidth of each device broader, the total bandwidth and data capacity increases. The signal is affected by IMD gets broader, the total bandwidth and data capacity increases. The signal is affected by noises, including IMD2 and IMD3, which are generated by nonlinear devices and through IMD noises, including IMD2 and IMD3, which are generated by nonlinear devices and the fiber. Because of these complex IMD noises, it will be very difficult to estimate the through the fiber. Because of these complex IMD noises, it will be very difficult to estimate channel. Therefore, it is hard to reduce signal distortions by equalization and signal pre- the channel. Therefore, it is hard to reduce signal distortions by equalization and signal distortion requiring accurate channel information. These techniques are not suitable to pre-distortion requiring accurate channel information. These techniques are not suitable mitigate IMD noise in MFN. Existing studies have conducted just tone experiments for to mitigate IMD noise in MFN. Existing studies have conducted just tone experiments for mitigating nonlinear components [6–9,11] or have assumed that all IF bands have the same mitigating nonlinear components [6–9,11] or have assumed that all IF bands have the same modulation format. Furthermore, in conventional techniques, all IF bands have the same modulation format. Furthermore, in conventional techniques, all IF bands have the same power regardless of modulation format, as shown in Figure 1. However, because each IF power regardless of modulation format, as shown in Figure 1. However, because each IF band supports different devices and applications, each IF band passes through different band supports different devices and applications, each IF band passes through different channels and has different modulation formats. After receiving the signal at the receiver, channels and has different modulation formats. After receiving the signal at the receiver, the SNR the SNR is m is e measur asured a ednand d infinformed ormed the the cha channel nnel to the tra to the transmitter nsmitter. . (a) (b) Figure 1. Conventional signal spectrum of (a) transmitted signal and (b) received signal. Figure 1. Conventional signal spectrum of (a) transmitted signal and (b) received signal. Accordingly, the transmitter sets the modulation format, and power loading is per- Accordingly, the transmitter sets the modulation format, and power loading is formed with the water-pouring algorithm [15]. In the case of the CPRI, the generated performed with the water-pouring algorithm [15]. In the case of the CPRI, the generated orthogonal frequency-division multiplexing (OFDM) signal is converted to the time do- orthogonal frequency-division multiplexing (OFDM) signal is converted to the time main through the inverse Fourier transform (IFT), and the converted signal is digitized domain through the inverse Fourier transform (IFT), and the converted signal is digitized again to the baseband digital signal, then the signal is transmitted to the RU. However, again to the baseband digital signal, then the signal is transmitted to the RU. However, since this conversion leads to explosive growth in data traffic if the OFDM signal is directly since this conversion leads to explosive growth in data traffic if the OFDM signal is transmitted to the RU using the IFoF transmission, IMD noise is generated, and this IMD directly transmitted to the RU using the IFoF transmission, IMD noise is generated, and noise distorts the signal. At the RU, the distorted optical signal is downconverted to the RF this IMD noise distorts the signal. At the RU, the distorted optical signal is downconverted frequency, and the RF signal is amplified again and transmitted to each user through the wireless channel. As shown in Figure 2a, at the RU, the distorted optical signal is downcon- verted to the RF frequency, and the RF signal is amplified again and transmitted to each user through the wireless channel. Since the CPRI signal is converted to the OFDM signal at Photonics 2020, 7, x FOR PEER REVIEW 3 of 9 to the RF frequency, and the RF signal is amplified again and transmitted to each user through the wireless channel. As shown in Figure 2a, at the RU, the distorted optical signal is downconverted to the RF frequency, and the RF signal is amplified again and transmitted to each user through the wireless channel. Since the CPRI signal is converted to the OFDM signal at the RU, the IMD noise is not considered in the MFN using CPRI, and also, the transmitted signal power, which affects IMD noise power, was not a major consideration in the MFN. However, in the case of IFoF transmission, as shown in Figure 2b, after downconversion of the received signal, the signal is amplified without removing IMD noise at the RU, so the noise is also amplified and transmitted with the signal. Because of using an analog signal in IFoF transmission, the power of the nonlinear component affected by the signal power can be considered as the main distortion, and the generated IMD is amplified without being removed from the RU stage, which can continuously affect the signal distortion. In this paper, we propose a spectral power allocation technique dependent on the Photonics 2021, 8, 2 3 of 9 modulation format and verify the feasibility of system linearity enhancement of multi- IFoF-based MFN. The signal spectral power was allocated for each IF band according to modulation order, which is dependent on the required signal-to-noise power ratio (SNR), the RU, the IMD noise is not considered in the MFN using CPRI, and also, the transmitted which can be converted from the error vector magnitude (EVM) requirement. Since the signal power, which affects IMD noise power, was not a major consideration in the MFN. IMD noise power is affected by the transmitted signal power, when the total transmitted However, in the case of IFoF transmission, as shown in Figure 2b, after downconversion of signal power is changed, the IMD noise power would be significantly changed compared the received signal, the signal is amplified without removing IMD noise at the RU, so the to the signal power variation. We have experimentally demonstrated the reduction of noise is also amplified and transmitted with the signal. Because of using an analog signal IMD noise using the proposed IF power allocation. Transmission performance in IFoF transmission, the power of the nonlinear component affected by the signal power enhancement using the technique was verified experimentally by the EVM improvement can be considered as the main distortion, and the generated IMD is amplified without of the received signal. This technique was briefly proposed at the Asian communication being removed from the RU stage, which can continuously affect the signal distortion. and photonics (ACP) conference 2019 and is described in detail in this paper [16]. (a) (b) Figure 2. Figure 2. The schematic The schematic diagram diagram of ( of (a a)) common common pu public blic rad radio io interfac interfacee (CP (CPRI), RI), ( (bb ) IF ) IF over over fiber fiber (IFoF) (IFoF) transmission. transmission. In this paper, we propose a spectral power allocation technique dependent on the 2. Modulation-Based Adaptive Power Allocation modulation format and verify the feasibility of system linearity enhancement of multi- When the transmitted signal passes through the nonlinear devices or channel, the IFoF-based MFN. The signal spectral power was allocated for each IF band according to nonlinear noise is generated according to the channel transfer function, as shown in modulation order, which is dependent on the required signal-to-noise power ratio (SNR), Equation (1). The transfer function has different coefficients for each different order as a which can be converted from the error vector magnitude (EVM) requirement. Since the function of the frequency [11,12]. IMD noise power is affected by the transmitted signal power, when the total transmitted signal power is changed, 𝑦 𝑡, 𝑓 the𝑎 IMD𝑎 noise 𝑓 𝑥 power 𝑡 𝑎 wou 𝑓 𝑥 ld 𝑡 be𝑎significantly 𝑓 𝑥 𝑡 ∙∙∙ changed compar(1 ed ) to the signal power variation. We have experimentally demonstrated the reduction of IMD In MFN using CPRI, the signal modulation format is determined by the signal noise using the proposed IF power allocation. Transmission performance enhancement degradation by the wireless channel between RU-Device and the fixed additive white using the technique was verified experimentally by the EVM improvement of the received gaussian noise (AWGN) in the entire channel, and the signal power is allocated by the signal. This technique was briefly proposed at the Asian communication and photonics water-pouring algorithm as shown in Figure 3 [17]. (ACP) conference 2019 and is described in detail in this paper [16]. 2. Modulation-Based Adaptive Power Allocation When the transmitted signal passes through the nonlinear devices or channel, the nonlinear noise is generated according to the channel transfer function, as shown in Equation (1). The transfer function has different coefficients for each different order as a function of the frequency [11,12]. 2 3 y(t, f ) = a + a ( f )x(t) + a ( f )x(t) + a ( f )x(t) +    (1) 0 1 2 3 In MFN using CPRI, the signal modulation format is determined by the signal degra- dation by the wireless channel between RU-Device and the fixed additive white gaussian noise (AWGN) in the entire channel, and the signal power is allocated by the water-pouring algorithm as shown in Figure 3 [17]. Photonics 2021, 8, 2 4 of 9 Photonics 2020, 7, x FOR PEER REVIEW 4 of 9 Photonics 2020, 7, x FOR PEER REVIEW 4 of 9 Photonics 2020, 7, x FOR PEER REVIEW 4 of 9 Figure 3. Power loading with the water-pouring algorithm. Figure 3. Power loading with the water-pouring algorithm. Figure 3. Power loading with the water-pouring algorithm. Figure 3. Power loading with the water-pouring algorithm. However, in the IFoF transmission, the OFDM signals are upconverted to each IF and However, in the IFoF transmission, the OFDM signals are upconverted to each IF and However, in the IFoF transmission, the OFDM signals are upconverted to each IF are transmitted to the RU, so the IMD noise and AWGN are combined and distort the are transmitted to the RU, so the IMD noise and AWGN are combined and distort the and ar However, e transmitted in thto e IFoF tr the RU, ansmission, the so the IMD noise OFDM sign and AWGN als are are upconverted to each combined and distort IF an thed signal. Since the amplifier amplifies not only the signal power also noise power, the signal signal. signSince al. Since the th amplifier e amplifier amplifies amplifies not not only on the ly tsignal he sign power al power a also lnoise so noise powe power, the r, tsignal he signal are transmitted to the RU, so the IMD noise and AWGN are combined and distort the is transmitted to the device with the amplified noise. is transmitted is transmitted to the devi to the device with ce withe th the a amplified mplifinoise. ed noise. signal. Since the amplifier amplifies not only the signal power also noise power, the signal In other words, the MFN using the RoF system should transmit the signal to the RU In other In other wor words, d the s, the MFN MF using N usin the g the RoF Ro system F system should should tra transmit nsmi the t the si signal gna tolthe to the RU RU is transmitted to the device with the amplified noise. without distortion as much as possible. In the proposed technique, the signal is generated without without di distortion stortias on a much s muas ch a possible. s possibIn le. In th the pr e proposed technique, t oposed technique, the signal he sign is al generated is generated In other words, the MFN using the RoF system should transmit the signal to the RU without without su surplus rplus power. It power. It is pos is possible sible t to o reduce reduce t the he IMD IMD noise noise power gener power generated ated by t by the he without surplus power. It is possible to reduce the IMD noise power generated by the without distortion as much as possible. In the proposed technique, the signal is generated surplus surplus powe power r of the of thesign signal. al. As As shown shown in Fig in Figur ures 4 an es 4d 5 and show the 5 showspectrum of the signal the spectrum of the surplus power of the signal. As shown in Figures 4 and 5 show the spectrum of the signal without surplus power. It is possible to reduce the IMD noise power generated by the signal passing t passing hrough t through he non thelnonlinear inear chann channel el using using the p theo power wer lo loading ading and and us using ing p prroposed oposed passing through the nonlinear channel using the power loading and using proposed surplus power of the signal. As shown in Figures 4 and 5 show the spectrum of the signal techniques. techniques. techniques. passing through the nonlinear channel using the power loading and using proposed techniques. Figure 4. Power allocation with the proposed technique. Figure 4. Power allocation with the proposed technique. Figure 4. Power allocation with the proposed technique. Figure 4. Power allocation with the proposed technique. Figure 5. (a) Multiband orthogonal frequency-division multiplexing (OFDM) signal spectrum with- Figure 5. (a) Multiband orthogonal frequency-division multiplexing (OFDM) signal spectrum Figure 5. (a) Multiband orthogonal frequency-division multiplexing (OFDM) signal spectrum out power allocation, (b) signal spectrum with power loading, (c) signal spectrum with proposed without power allocation, (b) signal spectrum with power loading, (c) signal spectrum with without power allocation, (b) signal spectrum with power loading, (c) signal spectrum with Figure 5. (a) Multiband orthogonal frequency-division multiplexing (OFDM) signal spectrum technique, (d) signal spectrum comparison of power loading and proposed technique. proposed technique, (d) signal spectrum comparison of power loading and proposed technique. proposed technique, (d) signal spectrum comparison of power loading and proposed technique. without power allocation, (b) signal spectrum with power loading, (c) signal spectrum with proposed technique, (d) signal spectrum comparison of power loading and proposed technique. The IMD components are generated by device nonlinearity and fiber nonlinearity. The IMD components are generated by device nonlinearity and fiber nonlinearity. In The IMD components are generated by device nonlinearity and fiber nonlinearity. In In the case of fiber nonlinearity, the IMD2 noise is mainly the nonlinear component gen- the case of fiber nonlinearity, the IMD2 noise is mainly the nonlinear component the case of fiber nonlinearity, the IMD2 noise is mainly the nonlinear component The IMD components are generated by device nonlinearity and fiber nonlinearity. In erated by the mixing laser chirp and chromatic dispersion when the laser is directly generated by the mixing laser chirp and chromatic dispersion when the laser is directly generated by the mixing laser chirp and chromatic dispersion when the laser is directly the case of fiber nonlinearity, the IMD2 noise is mainly the nonlinear component modulated [18,19]. The effect of dispersion is increased by increasing the transmission modulated [18,19]. The effect of dispersion is increased by increasing the transmission modulated [18,19]. The effect of dispersion is increased by increasing the transmission generated by the mixing laser chirp and chromatic dispersion when the laser is directly distance. When the directly modulated signal suffered dispersion, the IMD2 component distance. When the directly modulated signal suffered dispersion, the IMD2 component distance. When the directly modulated signal suffered dispersion, the IMD2 component modulated [18,19]. The effect of dispersion is increased by increasing the transmission is generated greatly [3]. Therefore, in the short distance, the IMD components caused by is generated greatly [3]. Therefore, in the short distance, the IMD components caused by is generated greatly [3]. Therefore, in the short distance, the IMD components caused by distance. When the directly modulated signal suffered dispersion, the IMD2 component device nonlinearity, which are generated in RF amplifiers, a laser diode (LD), photodetector device nonlinearity, which are generated in RF amplifiers, a laser diode (LD), device nonlinearity, which are generated in RF amplifiers, a laser diode (LD), is generated greatly [3]. Therefore, in the short distance, the IMD components caused by (PD), and, etc. are mainly generated as IMD3. The IMD noise power is determined by the photodetector (PD), and, etc. are mainly generated as IMD3. The IMD noise power is photodetector (PD), and, etc. are mainly generated as IMD3. The IMD noise power is device nonlinearity, which are generated in RF amplifiers, a laser diode (LD), signal power. For example, the IMD3 noise power is expressed as the product of coefficient, determined by the signal power. For example, the IMD3 noise power is expressed as the determined by the signal power. For example, the IMD3 noise power is expressed as the photodetector (PD), and, etc. are mainly generated as IMD3. The IMD noise power is and the IMD3 noise power is three times of signal power [17]. That is, if the signal power product of coefficient, and the IMD3 noise power is three times of signal power [17]. That product of coefficient, and the IMD3 noise power is three times of signal power [17]. That determined by the signal power. For example, the IMD3 noise power is expressed as the decreases, the IMD3 noise power is reduced by three times the amount of the signal de- is, if the signal power decreases, the IMD3 noise power is reduced by three times the is, if the signal power decreases, the IMD3 noise power is reduced by three times the product of coefficient, and the IMD3 noise power is three times of signal power [17]. That crease. Therefore, a dramatic reduction in IMD noise power can be achieved by controlling amount of the signal decrease. Therefore, a dramatic reduction in IMD noise power can amount of the signal decrease. Therefore, a dramatic reduction in IMD noise power can is, if the signal power decreases, the IMD3 noise power is reduced by three times the the signal power [14]. In multi-IFoF transmission, there are two different IMD components. be achieved by controlling the signal power [14]. In multi-IFoF transmission, there are two be achieved by controlling the signal power [14]. In multi-IFoF transmission, there are two amount of the signal decrease. Therefore, a dramatic reduction in IMD noise power can The first is the intra-band distortion generated by beating between subcarriers within a different IMD components. The first is the intra-band distortion generated by beating different IMD components. The first is the intra-band distortion generated by beating be achieved by controlling the signal power [14]. In multi-IFoF transmission, there are two single IF band, and the other is inter-band distortion due to beating between subcarriers between subcarriers within a single IF band, and the other is inter-band distortion due to between subcarriers within a single IF band, and the other is inter-band distortion due to different IMD components. The first is the intra-band distortion generated by beating from other bands. In other words, in the IFoF transmission, the intra-band and inter-band beating between subcarriers from other bands. In other words, in the IFoF transmission, beating between subcarriers from other bands. In other words, in the IFoF transmission, between subcarriers within a single IF band, and the other is inter-band distortion due to IMD components are combined to generate a complex IMD component. In multi-IFoF the intra-band and inter-band IMD components are combined to generate a complex IMD the intra-band and inter-band IMD components are combined to generate a complex IMD beating between subcarriers from other bands. In other words, in the IFoF transmission, transmission, each IF band passes through different channels and supports different de- component. In multi-IFoF transmission, each IF band passes through different channels component. In multi-IFoF transmission, each IF band passes through different channels the intra-band and inter-band IMD components are combined to generate a complex IMD vices, where each IF band signal is modulated in a different format. Because different and supports different devices, where each IF band signal is modulated in a different and supports different devices, where each IF band signal is modulated in a different component. In multi-IFoF transmission, each IF band passes through different channels and supports different devices, where each IF band signal is modulated in a different Photonics 2021, 8, 2 5 of 9 Photonics 2020, 7, x FOR PEER REVIEW 5 of 9 format. Because different modulation formats have different BERs in the same SNR environment, the modulation formats require different EVM values, as shown in below modulation formats have different BERs in the same SNR environment, the modulation Table 1 [20]. formats require different EVM values, as shown in below Table 1 [20]. Table 1. Modulation format dependent error vector magnitude (EVM) and signal-to-noise power Table 1. Modulation format dependent error vector magnitude (EVM) and signal-to-noise power ratio (SNR) requirements in long term evolution (LTE) standard [20]. ratio (SNR) requirements in long term evolution (LTE) standard [20]. Format QPSK 16QAM 64QAM 256QAM Format QPSK 16QAM 64QAM 256QAM EVM (%) 17.5 12.5 8 4 SNR ( EVM dB) (%) 15 17.5 18 12.5 22 8 28 4 SNR (dB) 15 18 22 28 The high order modulation format requires high received SNR. On the contrary, the lower-order modulation format requires a relatively low received SNR. Because of the The high order modulation format requires high received SNR. On the contrary, the relationship between the EVM and the SNR, required SNRs are about a minimum of 15 lower-order modulation format requires a relatively low received SNR. Because of the dB for quadrature phase-shift keying (QPSK) and 28 dB for 256 quadrature amplitude relationship between the EVM and the SNR, required SNRs are about a minimum of modulation (QAM) in the transmission system. The higher the modulation order, the 15 dB for quadrature phase-shift keying (QPSK) and 28 dB for 256 quadrature amplitude more signal power is required. In MFN, the signal is transmitted through the fiber from modulation (QAM) in the transmission system. The higher the modulation order, the more DU to RU, amplified at the RU, and then sent to each device or user over the wireless signal power is required. In MFN, the signal is transmitted through the fiber from DU to channel. The dominant noise which degrades SNR occurs mainly in the nonlinear device RU, amplified at the RU, and then sent to each device or user over the wireless channel. or wireless channel. When the device receives a signal, the device periodically measures The dominant noise which degrades SNR occurs mainly in the nonlinear device or wireless the channel. Then, the device sends a channel quality indicator (CQI) containing the channel. When the device receives a signal, the device periodically measures the channel. measured SNR from the de Then, the device sends avice. Since the modula channel quality indicator tion format of the (CQI) containing signthe al is dete measur rmined ed SNR by the SNR from the device. from CQI [2 Since 1], thethat is, the t modulationrformat ansmitted signal of the signal from the is determined DU to the RU needs by the SNR from CQI [21], that is, the transmitted signal from the DU to the RU needs only an SNR that is only an SNR that is enough to satisfy the EVM requirements until the antenna where the sign enough al is amp to lif satisfy ied. The therEVM efore, rin t equir heements IFoF tran until smission, the the antenna spectr wher al powers e the signal of t is he band amplified. s Therefore, in the IFoF transmission, the spectral powers of the bands are allocated based are allocated based on the required SNRs according to the modulation format of each on the required SNRs according to the modulation format of each band, and the allocated band, and the allocated signal is transmitted. If the signal power per IF is allocated by the signal is transmitted. If the signal power per IF is allocated by the SNR required by each SNR required by each modulation format, the total power of the signal can be reduced, modulation format, the total power of the signal can be reduced, and the power of the and the power of the nonlinear component is also reduced. When the SNR required by nonlinear component is also reduced. When the SNR required by each modulation format each modulation format is allocated to each IF band, the power of the signal is adjusted, is allocated to each IF band, the power of the signal is adjusted, and the IMD power can and the IMD power can also be decreased, as shown in Figure 6. also be decreased, as shown in Figure 6. (a) (b) Figure 6. Signal spectrum with the proposed technique of (a) transmitted signal and (b) received Figure 6. Signal spectrum with the proposed technique of (a) transmitted signal and (b) received signal. signal. The operating principle of the proposed modulation based adaptive power allocation (MBAPA) technique is as follows: When the initial power of all bands has the same power, The operating principle of the proposed modulation based adaptive power allocation the power of the signal band of 256QAM, the highest order, is maintained. Based on the (MBAPA) technique is as follows: When the initial power of all bands has the same power, the power of 256QAM signal the signa power l band of , the excess 256QAM, th power e highest order, in each band is is m reduced aintaine by d the . Br aequir sed on the ed SNR difference for each order. In other words, the power of the QPSK signal is reduced by 13 dB, 256QAM signal power, the excess power in each band is reduced by the required SNR which is the difference from the required SNR of 256QAM. Even if the surplus power of difference for each order. In other words, the power of the QPSK signal is reduced by 13 the signal is reduced, the required EVM can be satisfied so that the signal to which the dB, which is the difference from the required SNR of 256QAM. Even if the surplus power MBAPA technology to control surplus power is applied can be received. of the signal is reduced, the required EVM can be satisfied so that the signal to which the MBAPA technology to control surplus power is applied can be received. Photonics 2021, 8, 2 6 of 9 Photonics 2020, 7, x FOR PEER REVIEW 6 of 9 Photonics 2020, 7, x FOR PEER REVIEW 6 of 9 3. Experiments and Results 3. Experiments and Results The feasibility of the proposed MBAPA technique was verified experimentally in The feasibility of the proposed MBAPA technique was verified experimentally in 3. Experiments and Results multi-IFoF transmission by intensity-modulated and direct detection (IM/DD). In the multi-IFoF transmission by intensity-modulated and direct detection (IM/DD). In the The feasibility of the proposed MBAPA technique was verified experimentally in multi- experiment, as shown in Figure 7, randomly generated bits were mapped to symbols experiment, as shown in Figure 7, randomly generated bits were mapped to symbols IFoF transmission by intensity-modulated and direct detection (IM/DD). In the experiment, depending on each modulation format of subcarriers (QPSK, 16QAM, 64QAM, and depending on each modulation format of subcarriers (QPSK, 16QAM, 64QAM, and as shown in Figure 7, randomly generated bits were mapped to symbols depending on each 256QAM). Bandwidth and spacing of subcarriers were set for generating OFDM signal. 256QAM). Bandwidth and spacing of subcarriers were set for generating OFDM signal. modulation format of subcarriers (QPSK, 16QAM, 64QAM, and 256QAM). Bandwidth and To have 20 MHz bandwidth per single IF, the number of subcarriers per band was 1200. To have 20 MHz bandwidth per single IF, the number of subcarriers per band was 1200. spacing of subcarriers were set for generating OFDM signal. To have 20 MHz bandwidth The subcarrier spacing of the OFDM signal was set to be 15 kHz. The spacing between The subcarrier spacing of the OFDM signal was set to be 15 kHz. The spacing between per single IF, the number of subcarriers per band was 1200. The subcarrier spacing of the bands was set as 10 MHz, and 24 and 48 multi-IF bands were generated. The optical signal bands was set as 10 MHz, and 24 and 48 multi-IF bands were generated. The optical signal OFDM signal was set to be 15 kHz. The spacing between bands was set as 10 MHz, and 24 was transmitted through 10 km and 20 km optical single-mode-fiber. The received optical was transmitted through 10 km and 20 km optical single-mode-fiber. The received optical and 48 multi-IF bands were generated. The optical signal was transmitted through 10 km power at the photodetector (PD) was about −1.5 dBm, and the RF spectra of the signal, power at the photodetector (PD) was about −1.5 dBm, and the RF spectra of the signal, and 20 km optical single-mode-fiber. The received optical power at the photodetector (PD) which was converted from optical to electrical signal at PD, were measured. The proposed which was converted from optical to electrical signal at PD, were measured. The proposed was about 1.5 dBm, and the RF spectra of the signal, which was converted from optical MBAPA signal can suppress the nonlinearity of the system. It was assumed that the IMD MBAPA signal can suppress the nonlinearity of the system. It was assumed that the IMD to electrical signal at PD, were measured. The proposed MBAPA signal can suppress the components were generated by transfer functions of RF amplifier and LD and by power components were generated by transfer functions of RF amplifier and LD and by power nonlinearity of the system. It was assumed that the IMD components were generated by saturation at PD. saturation at PD. transfer functions of RF amplifier and LD and by power saturation at PD. Figure 7. Experimental setup. Figure 7. Experimental setup. Figure 7. Experimental setup. Figure 8a,b shows the transmitted signal spectrum with the conventional technique Figure 8a,b shows the transmitted signal spectrum with the conventional technique Figure 8a,b shows the transmitted signal spectrum with the conventional technique and with the proposed technique, respectively. After 10 km transmission, two nonlinear and with the proposed technique, respectively. After 10 km transmission, two nonlinear and with the proposed technique, respectively. After 10 km transmission, two nonlinear distortions were generated and interfered with each other in a spectrum, as shown in distortions were generated and interfered with each other in a spectrum, as shown in distortions were generated and interfered with each other in a spectrum, as shown in Figure 9. Figure 9. Figure 9. (a) (b) (a) (b) Figure 8. The power spectral density of transmitted signal (a) without the technique and (b) with the Figure 8. The power spectral density of transmitted signal (a) without the technique and (b) with Figure 8. The power spectral density of transmitted signal (a) without the technique and (b) with the technique l technique l. . the technique l. Photonics 2021, 8, 2 7 of 9 Photonics 2020, 7, x FOR PEER REVIEW 7 of 9 Photonics 2020, 7, x FOR PEER REVIEW 7 of 9 Figure 9. Received 24-IF bands signal spectrum after 10 km transmission. Figure 9. Figure Recei 9. Received ved 24-IF24-IF bandbands s signal spe signal ctr spectr um after 10 km transmission. um after 10 km transmission. In Figure 9, using the proposed technique, the IMD power was reduced by about 10 In Figure 9, using the proposed technique, the IMD power was reduced by about In Figure 9, using the proposed technique, the IMD power was reduced by about 10 dB. Figure 10 shows the constellations of the first IF band, which was modulated by QPSK, 10 dB. Figure 10 shows the constellations of the first IF band, which was modulated by dB. Figure 10 shows the constellations of the first IF band, which was modulated by QPSK, and the third IF band modulated by 256QAM after a 10 km transmission with 24-IFs. QPSK, and the third IF band modulated by 256QAM after a 10 km transmission with 24-IFs. and the third IF band modulated by 256QAM after a 10 km transmission with 24-IFs. (a) (b) (c) (d) (a) (b) (c) (d) Figure 10. Constellation of the received signal after 10 km transmission with 24-IFs (a) quadrature phase-shift keying Figure 10. Constellation of the received signal after 10 km transmission with 24-IFs (a) quadrature phase-shift keying Figure 10. Constellation of the received signal after 10 km transmission with 24-IFs (a) quadrature phase-shift keying (QPSK) without the technique, (b) QPSK with the technique, (c) 256QAM without the technique and (d) 256QAM with (QPSK) without the technique, (b) QPSK with the technique, (c) 256QAM without the technique and (d) 256QAM with (QPSK) without the technique, (b) QPSK with the technique, (c) 256QAM without the technique and (d) 256QAM with the technique. the technique. the technique. Figure 10a,b shows QPSK constellation diagrams, respectively, without the proposed Figure 10a,b shows QPSK constellation diagrams, respectively, without the proposed Figure 10a,b shows QPSK constellation diagrams, respectively, without the proposed technique and with the proposed technique. In the case of QPSK, the EVM value without technique and with the proposed technique. In the case of QPSK, the EVM value with- technique and with the proposed technique. In the case of QPSK, the EVM value without the technique was 7%, and The EVM value with the technique was 11%. QPSK is a low- out the technique was 7%, and The EVM value with the technique was 11%. QPSK is a the technique was 7%, and The EVM value with the technique was 11%. QPSK is a low- order modulation that requires less signal power. Therefore, the proposed technique low-order modulation that requires less signal power. Therefore, the proposed technique order modulation that requires less signal power. Therefore, the proposed technique reduces the surplus of the signal power, improves the SNR, which causes deterioration of reduces the surplus of the signal power, improves the SNR, which causes deterioration reduces the surplus of the signal power, improves the SNR, which causes deterioration of the EVM. The EVM values of both signals satisfy the EVM requirement of the QPSK of the EVM. The EVM values of both signals satisfy the EVM requirement of the QPSK the EVM. The EVM values of both signals satisfy the EVM requirement of the QPSK format. Figure 10c,d is the 256QAM constellation; (c) is constellation without the proposed format. Figure 10c,d is the 256QAM constellation; (c) is constellation without the proposed format. Figure 10c,d is the 256QAM constellation; (c) is constellation without the proposed technique, and (d) is with the proposed technique. In the case of 256QAM, unlike QPSK, technique, and (d) is with the proposed technique. In the case of 256QAM, unlike QPSK, technique, and (d) is with the proposed technique. In the case of 256QAM, unlike QPSK, the proposed technique improved EVM performance from 10.8% to 3.5%. The received the proposed technique improved EVM performance from 10.8% to 3.5%. The received the proposed technique improved EVM performance from 10.8% to 3.5%. The received signal without the technique could not satisfy the EVM requirement to receive by signal without the technique could not satisfy the EVM requirement to receive by nonlinear signal without the technique could not satisfy the EVM requirement to receive by distortion nonlinear distortion noise. On the other ha noise. On the other hand, with the pr nd, wi oposed th technique, the proposed techniq the EVM requir ue, tement he EVis M nonlinear distortion noise. On the other hand, with the proposed technique, the EVM satisfied, requirement so the is signal satisfied with , so the thtechnique e signal wit is h suitable the tec for hniq multi-IFoF ue is suit transmission. able for multIn i-IFoF ad- requirement is satisfied, so the signal with the technique is suitable for multi-IFoF transmission. In addition, in order to verify the effect of the technique on the fiber dition, in order to verify the effect of the technique on the fiber nonlinearity, experiments transmission. In addition, in order to verify the effect of the technique on the fiber wer nonlinearity, e conducted experiments by increasing were cond the transmission ucted by incre distance. asing the transmission distance. nonlinearity, experiments were conducted by increasing the transmission distance. The The dispersion effect is dispersion effect is mormore sever e severe as incr e as easing increa the sing transmission the transmis distance. sion dist Asance already . As The dispersion effect is more severe as increasing the transmission distance. As written above, the dispersion mixed with the chirp effect of laser-generated the IMD2 more. already written above, the dispersion mixed with the chirp effect of laser-generated the already written above, the dispersion mixed with the chirp effect of laser-generated the As Figure 11 shows, the constellation of the 20 km transmitted signal with the same IMD2 more. As Figure 11 shows, the constellation of the 20 km transmitted signal with IMD2 more. As Figure 11 shows, the constellation of the 20 km transmitted signal with received optical power about 1.5 dBm. Similar to Figure 10, the EVM from (a) without the the same received optical power about −1.5 dBm. Similar to Figure 10, the EVM from (a) the same received optical power about −1.5 dBm. Similar to Figure 10, the EVM from (a) technique to (b) with the technique deteriorated from 6.5% to 13.4%, and both EVM values without the technique to (b) with the technique deteriorated from 6.5% to 13.4%, and both without the technique to (b) with the technique deteriorated from 6.5% to 13.4%, and both satisfied the EVM requirement. In addition, when using the proposed technique from EVM values satisfied the EVM requirement. In addition, when using the proposed EVM values satisfied the EVM requirement. In addition, when using the proposed (c) to (d), the EVM improved from 9.2% to 4%. It was proven that the IMD2 component technique from (c) to (d), the EVM improved from 9.2% to 4%. It was proven that the technique from (c) to (d), the EVM improved from 9.2% to 4%. It was proven that the generated by the dispersion could be effectively reduced by the proposed technique. IMD2 component generated by the dispersion could be effectively reduced by the IMD2 component generated by the dispersion could be effectively reduced by the proposed technique. proposed technique. Photonics 2021, 8, 2 8 of 9 Photonics 2020, 7, x FOR PEER REVIEW 8 of 9 Photonics 2020, 7, x FOR PEER REVIEW 8 of 9 (a) (b) (c) (d) (a) (b) (c) (d) Figure 11. Constellation of the received signal after 20 km transmission with 24-IFs (a) QPSK without the technique, (b) Figure 11. Constellation of the received signal after 20 km transmission with 24-IFs (a) QPSK without the technique, Figure 11. Constellation of the received signal after 20 km transmission with 24-IFs (a) QPSK without the technique, (b) QPSK with the technique, (c) 256QAM without the technique and (d) 256QAM with the technique. (b) QPSK with the technique, (c) 256QAM without the technique and (d) 256QAM with the technique. QPSK with the technique, (c) 256QAM without the technique and (d) 256QAM with the technique. Furthermore, an experiment with increasing the total bandwidth of the transmission Furthermore, an experiment with increasing the total bandwidth of the transmission Furthermore, an experiment with increasing the total bandwidth of the transmission signal was conducted. The transmission signal was changed from a signal using 24-IF signal was conducted. The transmission signal was changed from a signal using 24-IF signal was conducted. The transmission signal was changed from a signal using 24-IF bands to a signal using 48 IF bands. As the number of IF bands used increases, the total bands to a signal using 48 IF bands. As the number of IF bands used increases, the total bands to a signal using 48 IF bands. As the number of IF bands used increases, the total bandwidth becomes wider, and the effects of the inter-band IMD2 and IMD3 on the signal bandwidth becomes wider, and the effects of the inter-band IMD2 and IMD3 on the signal bandwidth becomes wider, and the effects of the inter-band IMD2 and IMD3 on the signal are significantly increased. The constellation of the 48-IF signal is shown in Figure 12 after are significantly increased. The constellation of the 48-IF signal is shown in Figure 12 after are significantly increased. The constellation of the 48-IF signal is shown in Figure 12 after 10 km transmission. As shown in Figure 12a,b, EVM became worse from 7.1% to 10.8% 10 km transmission. As shown in Figure 12a,b, EVM became worse from 7.1% to 10.8% 10 km transmission. As shown in Figure 12a,b, EVM became worse from 7.1% to 10.8% using the proposed technique, but the signals satisfied the EVM requirement for QPSK. using the proposed technique, but the signals satisfied the EVM requirement for QPSK. In using the proposed technique, but the signals satisfied the EVM requirement for QPSK. In addition, when using the proposed technique for 256QAM from (c) to (d), the EVM was addition, when using the proposed technique for 256QAM from (c) to (d), the EVM was In addition, when using the proposed technique for 256QAM from (c) to (d), the EVM was improved from 10.2% to 3.4%. It was demonstrated that the proposed technique mitigates improved from 10.2% to 3.4%. It was demonstrated that the proposed technique mitigates improved from 10.2% to 3.4%. It was demonstrated that the proposed technique mitigates the effects of complex inter-band IMD. the effects of complex inter-band IMD. the effects of complex inter-band IMD. (b) (c) (d) (a) (b) (c) (d) (a) Figure 12. Constellation of the received signal after 10 km transmission with 48-IFs (a) QPSK without the technique, Figure 12. Constellation of the received signal after 10 km transmission with 48-IFs (a) QPSK without the technique, (b) Figure 12. Constellation of the received signal after 10 km transmission with 48-IFs (a) QPSK without the technique, (b) (b) QPSK with the technique, (c) 256QAM without the technique and (d) 256QAM with the technique. QPSK with the technique, (c) 256QAM without the technique and (d) 256QAM with the technique. QPSK with the technique, (c) 256QAM without the technique and (d) 256QAM with the technique. 4. Conclusions 4. Conclusions 4. Conclusions We proposed the modulation format based spectral power allocation technique to We proposed the modulation format based spectral power allocation technique to We proposed the modulation format based spectral power allocation technique to mitigate the IMD noise and improve channel linearity in multi-IFoF transmission. For the mitigate the IMD noise and improve channel linearity in multi-IFoF transmission. For the mitigate the IMD noise and improve channel linearity in multi-IFoF transmission. For the 24-IFoF 10 km transmission system, the proposed technique obtained 7% of EVM 24-IFoF 10 km transmission system, the proposed technique obtained 7% of EVM im- 24-IFoF 10 km transmission system, the proposed technique obtained 7% of EVM improvement when 256QAM was used, and also, the MBAPA technique suppressed IMD provement when 256QAM was used, and also, the MBAPA technique suppressed IMD improvement when 256QAM was used, and also, the MBAPA technique suppressed IMD power around 10 dB. In addition, the system linearity was improved for a 48-IFoF 10 km power around 10 dB. In addition, the system linearity was improved for a 48-IFoF 10 km power around 10 dB. In addition, the system linearity was improved for a 48-IFoF 10 km transmission system and a 24-IFoF 10 km transmission system. The proposed technique transmission system and a 24-IFoF 10 km transmission system. The proposed technique in transmission system and a 24-IFoF 10 km transmission system. The proposed technique in the reduction of IMD power effectively will be the key solution for the 5G MFN the reduction of IMD power effectively will be the key solution for the 5G MFN supporting in the reduction of IMD power effectively will be the key solution for the 5G MFN supporting multi-user. multi-user. supporting multi-user. Author Contributions: Conceptualization, I.H.; Data curation, I.H.; Formal analysis, I.H. and H.- Author Contributions: Conceptualization, I.H.; Data curation, I.H.; Formal analysis, I.H. and H.-J.P.; Author Contributions: Conceptualization, I.H.; Data curation, I.H.; Formal analysis, I.H. and H.- J.P.; Project administration, S.-K.H.; Writing—original draft, I.H.; Writing—review & editing, S.- Project administration, S.-K.H.; Writing—original draft, I.H.; Writing—review & editing, S.-M.K. J.P.; Project administration, S.-K.H.; Writing—original draft, I.H.; Writing—review & editing, S.- M.K. All authors have read and agreed to the published version of the manuscript. All authors have read and agreed to the published version of the manuscript. M.K. All authors have read and agreed to the published version of the manuscript. Funding: This work was supported by the National Research Foundation of Korea (NRF) grant Funding: This work was supported by the National Research Foundation of Korea (NRF) grant Funding: This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korean government (MSIT; Ministry of Science and ICT) (No. NRF- funded by the Korean government (MSIT; Ministry of Science and ICT) (No. NRF-2019R1A2C3007934). funded by the Korean government (MSIT; Ministry of Science and ICT) (No. NRF- 2019R1A2C3007934). 2019R1A2C3007934). Conflicts of Interest: The authors declare no conflict of interest. Conflicts of Interest: The authors declare no conflicts of interest. Conflicts of Interest: The authors declare no conflicts of interest. Photonics 2021, 8, 2 9 of 9 References 1. Tanaka, K.; Agata, A. Next-generation optical access networks for C-RAN. In Proceedings of the Optical Fiber Communication Conference, Los Angeles, CA, USA, 22–26 March 2015. Paper Tu2E.1. 2. Common Public Radio Interface. CPRI Specification V7.0. Stand. Doc. Specif. 2015. Available online: http://www.cpri.info/ downloads/CPRI_v_7_0_2015-10-09.pdf (accessed on 2 November 2020). 3. Radio-over-fibre (RoF) technologies and their applications. In ITU-T, Suppl. 55, G-Series Recommendations; ITU: Geneva, Switzer- land, 2015. 4. Ishimura, S.; Bekkali, A.; Tanaka, K.; Nishimura, K.; Suzuki, M. 1.032-Tb/s CPRI-equivalent data rate transmission using IF-over-Fiber system for high capacity mobile fronthaul links. In Proceedings of the 2017 European Conference on Optical Communication (ECOC), Gothenburg, Sweden, 17–21 September 2017. Paper Th.PDP.B.6. 5. Gomes, N.J.; Morant-Perez, M.; Alphones, A.; Cabon, B.; Mitchell, J.E.; Lethien, C.; Csornyei, M.; Stoehr, A.; Iezekiel, S. Radio over fibre transport for the support of broadband services. Invit. Pap. OSA J. Opt. Net. 2009, 8, 156–178. [CrossRef] 6. Han, C.; Sung, M.; Cho, S.-H.; Chung, H.S.; Kim, S.M.; Lee, J.H. Performance improvement of multi-IFoF-based mobile fronthaul using dispersion-induced distortion mitigation with IF optimization. J. Lightw. Technol. 2016, 34, 4772–4778. [CrossRef] 7. Sung, M.; Cho, S.-H.; Kim, J.; Lee, J.K.; Lee, J.H.; Chung, H.S. Demonstration of IFoF-based mobile fronthaul in 5G prototype with 28-GHz millimeter wave. J. Lightw. Technol. 2018, 36, 601–609. [CrossRef] 8. Han, C.; Cho, S.; Chung, H.S.; Lee, J.H. Linearity improvement of directly-modulated multi-IF-over-fibre LTE-A mobile fronthaul link using shunt diode predistorter. In Proceedings of the 2015 European Conference on Optical Communication (ECOC), Valencia, Spain, 27 September–1 October 2015; p. We.4.4.4. 9. Li, J.; Zhang, Y.C.; Yu, S.; Jiang, T.; Xie, Q.; Gu, W. Third-order intermodulation distortion elimination of microwave photonics link based on integrated dual-drive dual-parallel Mach–Zehnder modulator. Opt. Exp. 2013, 38, 4285–4287. [CrossRef] [PubMed] 10. Peccarelli, N.; Fulton, C. Mitigation of nonlinear spurious products using least mean-square (LMS). In Proceedings of the GOMACTech, Miami, FL, USA, 12–15 March 2018. 11. Jiang, W.; Tan, Q.; Qin, W.; Liang, D.; Li, X.; Ma, H.; Zhu, Z. A linearization analog photonic link with high third-order intermodulation distortion suppression based on dual-parallel Mach–Zehnder modulator. IEEE Photon. J. 2015, 7, 1–8. [CrossRef] 12. Park, H.J.; Ha, I.H.; Han, S.K. Distortion mitigation in IFoF system employing frequency-separated DPD. In Proceedings of the MWP, Toulouse France, 22–25 October 2018. 13. Najarro, A.C.; Kim, S.M. Predistortion for Frequency-Dependent Nonlinearity of a Laser in RoF Systems. J. Inf. Commun. Converg. Eng. 2016, 14, 147–152. [CrossRef] 14. Ha, I.H.; Park, H.J.; Han, S.K. Nonlinearity mitigation in multi-IFoF-based mobile fronthaul transmission using spectral power shaping. Opt. Comm. 2019, 437, 382–387. [CrossRef] 15. Sonalkar, R.; Shively, R. An efficient bit-loading algorithms for DMT applications. In Proceedings of the GLOBECOM, Sydney, NSW, Australia, 8–12 November 1998; pp. 2683–2688. 16. Ha, I.H.; Park, H.J.; Han, S.K. Nonlinear Distortion Mitigation Technique using Modulation Format Dependent Spectral Power Al- location in Multi-IFoF system. In Asia Communications and Photonics Conference; Optical Society of America: Washington, DC, USA, 17. Gallager, R.G. Information Theory and Reliable Communication; Wiley: New York, NY, USA, 1968. 18. Brendel, F.; Poëtte, J.; Cabon, B.; Zwick, T.; van Dijk, F.; Lelarge, F.; Accard, A. Chromatic dispersion in 60 GHz radio-over-fiber networks based on mode-locked laser. J. Lightw. Technol. 2011, 29, 3810–3816. [CrossRef] 19. Lim, C.; Lee, K.-L.; Nirmalathas, A.; Novak, D.; Waterhouse, R. Impact of chromatic dispersion on 60 GHz radio-over-fiber transmission. In Proceedings of the IEEE 21st LEOS Annual Meeting, Newport Beach, CA, USA, 9–13 November 2008; pp. 89–90. 20. 3GPP TS 36.104 Version 15.3.0 Release 14, 3GPP. 2018. 21. Hamed, M.M.; Shukry, S.; El-Mahallawy, M.S.; El-Ramly, S.H. Modified earliest deadline first scheduling with channel quality indicator for downlink real-time traffic in LTE networks. In Proceedings of the Third International Conference on e-Technologies and Networks for Development (ICeND2014), Beirut, Lebanon, 29 April–1 May 2014.

Journal

PhotonicsMultidisciplinary Digital Publishing Institute

Published: Dec 22, 2020

There are no references for this article.