Research On Millimeter-Wave Massive MIMO Hybrid Precoding Techniques With Low-Resolution ADC/DAC

With the booming development of the mobile Internet and Internet of Things,the current fourth-generation mobile communication system(4G)seems unable to meet the surge in traffic demand.In order to meet the requirement of the growing traffic data,the the fifth-generation mobile communication system(5G)is developed quickly,and the 5G first-phase standard is now finished.As one of the key technologies of the 5G,millimeter wave communication plays an important role due to the rich spectrum resources available.At the same time,the spacing between the antennas can be greatly reduced,and thereby the massive multiple input multiple output(mMIMO)equipped with hundreds or thousands of antennas at the Base Station(BS)becomes possilbe.In future mobile communication systems,the combination of millimeter wave and mMIMO technology will be inevitable.In millimeter-wave mMIMO systems,in order to reduce receiver complexity,hardware cost and power consumption,the hybrid precoding techniques are usually employed.Namely,the precoding is divided into digital precoding and analog precoding,which is generally implemented with low-cost analog phase shifters.The number of radio frequency(RF)chains is also reduced.Further,by reducing the resolution of the Digital to Analog Conventers(DACs)and the Analog to Digital Conventers(ADCs),the base station and the user can reduce the power consumption of the system.Under the Additive Quantization Noise Model(AQNM),the thesis developed two low-complexity HP schemes for low-resolution ADCs and mixed-resolution ADCs for multi-user millimeter-wave mMIMO systems,respectively.The main contents are summerized as follows.Three low-complexity HP schemes based on optimal beam indexing criteria are developed with low resolution ADCs: Full Connected Hybrid Precoding(FC-HP),Fixed Sub-Connected Hybrid Precoding(FSC-HP)and Dynamically Sub-Connected Hybrid Precoding(DSC-HP).For the FC-HP and FSC-HP schemes,the best beam is achieved based on the maximizing the equivalent channel gain between the BS transmitting RF chain and the user receiving RF chain.By searching optimal receiving beam,the optimal target vector in the candidate analog precoding vector set and the candidate analog receiving vector set are designed.For DSC-HP scheme,the transmitting antenna dynamic grouping is implemented according to the user equivalent channel gain maximization criterion firstly,and then utilizing the optimal receive beam index to design analog precoding matrix and analog receive combination matrix.Lastly,the three schemes employ the ZF criterion to design a corresponding digital precoding matrix to eliminate inter-user interference.Considering the severe nonlinear distortion of low-resolution ADC,two HP schemes are developed based on fixed and dynamic DACs,respectively.Under AQNM,the theoretical expressions of the sum rate of the millimeter-wave mMIMO downlink based on mixed-resolution DACs is derived firstly.The digital precoding is further obtained based on the analog precoding matrix and analog combination matrix.The digital precoding matrix are designed by ZF method to eliminate the inter-user interference.Secondly,the HP is desiged based on the dynamic antenna grouping,which loose the diagonalization limitation in the analog precoding of the traditional FSC-HP scheme and the huge array gain brought by mMIMO technology can be more rationally utilized.Due to the fragmentation limitation of the analog precoding matrix in the HP scheme based on fixed mixture DACs,the analog beamforming gain and the limited high resolution DACs resources cannot be fully utilized.Therefore,a dynamic mixture DACs scheme based on user feedback is proposed.In the mixed DACs scheme,the feedback coefficient of the users is utilized to make the reasonable DACs matched with the desired RF chain,hence the limited high-resolution DACs resources are fully used.