Research On Key Technologies In Improving Multiuser Detection Performance Of MUSA

Due to the vigorous development of the Internet of Things and the increasing demand of people,the future wireless communication system will have more users accessing,which will bring huge system capacity requirements.Therefore,Multi-user shared access technology has been studied.It uses a complex sequence as an spreading sequence to improve spectrum efficiency,which can provide services for the huge business needs of the Internet of Things in the future.In order to improve the detection performance of MUSA system,this paper mainly studies and optimizes the complex spreading sequence,power allocation and multi-user detection algorithms in MUSA technology.Firstly,this paper introduces the basic principle of MUSA technology and the related theory of MUSA technology,including complex spreading code sequence and interference cancellation receiving algorithm.At the same time,the overload performance of MUSA technology is analyzed by MATLAB simulation.It is verified that MUSA technology can achieve a certain user overload rate through non-orthogonal multiplexing of complex sequence,thus improving system capacity.Then we optimize the complex sequence used by MUSA,eliminating some complex sequences which are not suitable for spreading sequence.The optimized sequence used as spreading sequence of MUSA can effectively improve the detection performance of MUSA system.In addition,the influence of sequence collision in MUSA on detection performance is compared and analyzed,which enlightens us to avoid the occurrence of sequence collision as much as possible in the use of MUSA technology.Then,aiming at the downstream scenario of MUSA,the influence of power allocation on the performance of MUSA system is studied from the perspective of channel capacity and detection BER.The analysis of channel capacity is extended from simple scenarios of two users to scenarios of multiple users.It is verified that different user power allocation will affect the channel capacity of MUSA.The BER curve of MUSA is obtained by MATLAB simulation.The simulation results show that the average BER of MUSA users is affected by power allocation.Through the analysis of two aspects,it is concluded that the MUSA system can achieve better performance capacity and lower detection BER through appropriate power allocation scheme.Finally,this paper applies the deep neural network method to the receiver of MUSA,and designs a DNN multi-user detection scheme which is different from the traditional MUSA detection method.The performance of the trained DNN receiver in the channel environment of Gauss white noise and the channel environment with Rayleigh fading and Gauss white noise can reach or even surpass that of the traditional MUSA detection method.In addition,changing the training sample data of DNN receiver and the number of layers and neurons in the hidden layer of DNN model can change the fitting expression ability of DNN model,thus changing the detection performance of DNN receiver.