The Research On Key Technologies For Both Analog And Digital Radio-over-fiber Based 5G Mobile Fronthaul

According to the latest data of ministry of industry and information technology,the number of 4G subscribers and base stations have reached up to 1.03 billion and 3.34 million in China,respectively.Monthly mobile internet access traffic has also reached up to 2.7GB.China,the United States of America,the European countries,Japan,and South Korea have claimed that they will realize the commercialization of 5G by 2020.Besides this,they will realese the 5G spectrum planning in November 2017,July 2016,November 2016,July 2016 and 2018,respectively.With the flourish of high-speed internet access,high resolution multi-media entertainment with virtual reality and machine-to-machine communications,the global mobile data traffic in the 5G era will grow exponentially.As the key component of centralized radio access network(C-RAN),mobile fronthaul(MFH)faces great challenges such as demands of higher capacity,higher bandwidth efficiency,lower latency,and lower cost.To address these issues,both digital and analog radio-over-fiber based MFH architectures have been widely investigated,due to the excellent immunity to nonlinear distortions and high bandwidth efficiency,respectively.However,the digital MFH has the drawback of relatively low bandwidth efficiency;analog MFH suffers from high sensitivity to nonlinear distortions and the relatively high implementation complexity,which are the key issues in 5G MFH.To resolve the key problems above,systematical and profound studies are carried out in this dissertation.We propose and experimentally demonstrate the digital MFH architecture employing high order delta-sigma modulator with PAM-4 format,differential pulse coding modulation with employing noise shaping,enhanced noise shaping based pulse code modulation,and discrete cosine transform combined with multi-band quantization,respectively.Besides this,the analog MFH architecture employing digital code-division multiplexing(CDM)channel aggregation is firstly proposed.Synchronous transmission of both the I/Q waveforms of wireless signals and the control words(CWs)used for the purpose of control and management using the CDM approach is also presented and proof-of concept experiments are carried out.The main several innovations of my dissertation can be summarized as follows:First of all,in order to improve the bandwidth efficiency of digital MFH and the EVM performance of conventional delta-sigma modulation,a digital MFH architecture employing high order delta-sigma modulator with PAM-4 format is proposed in this dissertation.The experimental results indicate that,significant improvement up to 68% is achieved in the average EVM performance compared to the previous delta-sigma modulation-based digital MFH.Compared to common public radio interface(CPRI),the optical bandwidth efficiency can be improved by 3~4 times(Refer to H.Li et al.,Opt.Express,2017,25(1): 1-9.).Secondly,to circumvent the low bandwidth efficiency of digital MFH and achieve better EVM performance compared to DPCM,we propose a bandwidth efficient digital MFH employing differential pulse coding modulation with noise shaping(NS-DPCM).By reshaping the spectrum of quantization noise,the noise power is suppressed in the data carrying subcarriers of OFDM signals,which reduces the data distortion caused by quantization noise.The experimental results show that,significant improvement up to 40% can be observed compared to DPCM.Meanwhile,the optical bandwidth efficiency can be improved by 3~10 times(according to the modulation format)compared to CPRI(Refer to H.Li et al.,Opt.Express,2018,26(9): 11407-11417.).Thirdly,two issues existed in NS-DPCM based digital MFH: One is that DPCM with multi-tap linear predictor needs online training to configure the predictor coefficients.Besides this,the noise power enhancement out of signal band will make the mobile signals not be applicable for asynchronous transmission with diverse applications in 5G.To address these two issues,we propose an enhanced noise shaping scheme based pulse code modulation to further improve the bandwidth efficiency of digital MFH.Compared to NS-DPCM,the noise shaping function is enhanced to obtain better suppression of quantization noise at the transmitter.At the receiver,the well-designed filter is utilized to effectively suppress the undesired side lobes of signals sent to RF frontend.Up to 5.7dB signal-to-quantization-noise-ratio(SQNR)improvement can be achieved compared with NS-DPCM.Meanwhile,the side lobes of decoded OFDM signal could be reduced up to ~50dB(Refer to H.Li et al.,Technical Digest of ECOC'2018,2018.paper Tu3 B.4.).Fourthly,for the application of DSP based frequency division multiplexing(FDM)channel aggregation in analog MFH,the computational complexity of the digital FDM channel aggregation will be huge due to the required large size fast Fourier transform/ inverse fast Fourier transform(FFT/IFFT).To address this issue,a low-complexity analog mobile fronthaul architecture via digital CDM channel aggregation is proposed in this dissertation.The FFT/IFFT operations are replaced by simple sign selection and addition,leading to the significant reduction of computational complexity due to nonuse of real multiplication.The experimental results indicate that almost the same performance could be maintained via CDM based aggregation,in comparison with the FDM based aggregation(Refer to H.Li et al.,Technical digest of OFC'2017,2017.paper Th3 A.5).Fifthly,to facilitate the control and management of the fronthaul equipment in 5G,it is necessary to transmit the control words(CWs)with high quality and bandwidth efficiency.Synchronous transmission of both the I/Q waveforms of wireless signals and the control words used for the purpose of control and management using the CDM approach is proposed.As shown in the experimental results,when the received power is more than-22 dBm,CWs could be error-free recovered while the OFDM signals are recovered with an EVM of ?3.6% after transmission over 5 km SSMF(Refer to H.Li et al.,IEEE Photon.Journal,2018,10(2): Article#: 7902710.).Sixthly,to further imporove the bandwidth efficiency of digital MFH,a spectrally efficient digital mobile fronthaul with discrete cosine transform(DCT)and partitioned quantization is firstly proposed and experimentally demonstrated.Different from digitizing each OFDM sample with the same number of quantization bits in previous digital mobile fronthaul,partitioned quantization(PQ)is introduced to separately digitize different parts of the DCT coefficients with different number of quantization bits according to the significance of these coefficients.Furthermore,the part of transform coefficients with very little significance is discarded.Thus,the total number of required QBs for digitizing an OFDM symbol can be greatly reduced,thereby high bandwidth efficiency could be achieved(Refer to H.Li et al.,Opt.Letters,2018,43(20): 5130-5133.).