Research On Signal Detection Algorithm In Massive MIMO Low Precision ADC System

Massive multiple-input and multiple-output(massive MIMO)technology and millimeter wave(mm Wave)technology have been extensively studied because they can improve the capacity of mobile communication systems.However,the application of these technologies brings high cost and power consumption problems to the system.Thus,low-precision quantification technology with low power consumption characteristics has been widely concerned.The signal detection algorithm suitable for high-precision quantization system is not suitable for low-precision quantization system.Therefore,this thesis studies the signal detection problem in low-precision quantization system.The following is the main work and problems to be solved in this thesis.The soft detection algorithm in the single bit quantization system are studied.Soft output detection(SOD)algorithm has good performance,but the computational complexity is very high.Therefore,based on SOD algorithm and successive interference cancellation algorithm,ordered successive interference cancellation soft output detection(OSIC-SOD)algorithm is proposed.The computational complexity of the proposed algorithm increases as the number of signals detected each time increases.When only one signal is detected each time,and the transmit antenna and modulation order are both 4,the computational complexity of the proposed algorithm is only 6.25%of the SOD algorithm.When all signals are detected at one time,the computational complexity of the proposed algorithm is the same as that of SOD algorithm.In 32×4 MU-MIMO system,when the polarization code rate is 0.5 and the frame error rate(FER)is 10^3-,compared to the SOD algorithm,the performance loss of the proposed algorithm is about 0.5d B when detecting one transmission signal at a time.As the number of signals detected in a single time increases,the FER of the proposed algorithm gradually close to the SOD algorithm.Research and analysis of the existing signal detection algorithms in multi-bit quantization systems,mainly analyze the generalized expectation consistent signal recovery(GEC-SR)algorithm and the generalized turbo signal recovery(GTurbo-SR)algorithm.GTurbo-SR algorithm has low computational complexity,but the algorithm is only suitable for single-input single-output OFDM systems.Therefore,this thesis proposes a signal detection algorithm suitable for MIMO-OFDM based on the GTurbo-SR algorithm:generalized turbo signal recovery algorithm aided by vector approximation message passing(GTurbo-SR-VAMP)algorithm.The proposed algorithm uses the VAMP algorithm to iterate in the frequency domain,which makes the computational complexity of the algorithm much lower than that of the GEC-SR algorithm.When the number of transmitting antennas,receiving antennas,number of subcarriers,and modulation order are 8,64,64,and 4,respectively,the computational complexity of the proposed algorithm is only 10^5-of the GEC-SR algorithm.In the optimal iteration,the bit error rate(BER)of the GTurbo-SR-VAMP algorithm is slightly higher than that of the GEC-SR algorithm.