Research On Multi-user Signal Detection Technology Of Time Reversal Division Multiple Access System

Time Reversal(TR)technology is a signal processing technology,which has caused extensive research by scholars due to its space-time focusing characteristics.There have been studies using TR technology as multiple access for wireless systems.Time Reversal Division Multiple Access(TRDMA)can be regarded as a special type of spatial multiple access.According to the characteristics of TR focusing,it can focus on a specific spatial location,so as to distinguish different users and achieve multiple access.TRDMA enables the signal to be focused at a specific position according to the temporal and spatial focusing characteristics.In this way,the transmitted signal is more accurately focused on the receiving end,which has less interference and higher system capacity than traditional multiple access methods.However,when there is a strong correlation between channels,TR cannot distinguish users,and pseudo-focus phenomenon will occur.Aiming at the problem that channel correlation in TRDMA systems will cause multi-user interference,this article focuses on eliminating multi-user interference caused by strong channel correlation in TRDMA.The main work isFirstly,the Minimum Mean Square Error(MMSE)algorithm is introduced into the TRDMA system to eliminate interference between users,but the calculation complexity of the MMSE detection matrix is too large.Low complexity approximation algorithm for matrix factorization about Three diagonal lines based on MMSE in the TRDMA system is proposed.First,the three diagonal matrices containing the main diagonal in the MMSE detection matrix is been extracted,and then the three diagonal matrices are divided into two two diagonal matrices.Then get the inverses of these double two-diagonal matrices by the inverse rule of two-diagonal matrix respectively,and the inverse of the three-diagonal matrices is obtained.Finally,based on the Neumann series approximation,the inverse of the MMSE detection matrix is approximated by the inverse of the sparse matrix of the tridiagonal matrix.Simulation results show that the algorithm has obvious performance advantages in terms of bit error rate and spectral efficiency,and has lower complexity.It can obtain nearly optimal performance gain when the complexity is lowSecondly,the Barzilai-Borwein(B-B)iterative algorithm is introduced into TRDMA to solve the problem of interference between users when the channel is highly correlated,and the method of solving linear equations by B-B iteration is used to avoid the inverse process of the complex MMSE detection matrix.Conjugate gradient(CG)algorithm is introduced to accelerate the convergence speed of B-B algorithm,and a CG accelerated iterative detection algorithm based on B-B algorithm in TRDMA system is proposed.The CG algorithm first iterates twice to find the fastest descent direction,and then performs B-B iteration.B-B continues to iterate along the CG gradient descent direction.Simulations show that the improved algorithm converges faster than the B-B and CG algorithms,and the complexity is lower than that of the CG and MMSE algorithms,remains at O(N~2).