Research On Optimization Algorithm Of Channel Decoding And Detector For 5G High Speed Data Transmission

Since the fifth-generation mobile communication technology was officially put into commercial use,it has become one of the key technologies of China’s technological independent innovation and information technology.The realization of the high-speed,high-reliability,and large-capacity goals of the 5G system is inseparable from the support of low-complexity channel decoding and efficient MIMO detection technology.Low-density parity-check(LDPC)codes have become the channel coding of 5G data services due to their excellent performance,and the realization of their low-complexity decoding algorithms has become a research focus.MIMO technology can significantly improve channel capacity and spectrum efficiency,and is one of the key technologies of 5G.The joint optimization of decoder and MIMO detector is an important way for high-speed data rate transmission.The work of this paper is mainly reflected in the following two aspects:(1)The optimization of low complexity decoding of LDPC code.As far as LDPC codes are concerned,the computational complexity of decoding mainly lies in the check nodes.In order to reduce the computation times of the check nodes,the addition correction factor and the multiplication correction factor are introduced into the implementation of the decoder.The selection of these factors needs to be optimized by using extrinsic information transfer(EXIT)chart.Considering the existence of variable nodes with degree 1 in 5G LDPC code,this paper uses EXIT chart with 3D structure to represent the transfer relationship of extrinsic information.The simulation results show that when the EXIT chart is used to optimize the correction factors,the simulation performance of bit error rate(BER)of multiplicative factor is in good agreement with the relationship of extrinsic information measurement of EXIT chart;but that of additive factor is not significant.(2)The concatenation of DetNet detector based neural network and decoder.Based on the current situation that the channel decoder needs to input bit soft information,and the output of DetNet is symbol soft information,this paper uses the mapping method from symbols to bits to realize the calculation of soft information,and introduces the scaling factor of soft information to improve the reliability of concatenated system,and finally realizes the improvement of concatenation performance of decoder and detector.