Sparse Multipath Channel Estimation For Massive MIMO-OFDM

For Massive MIMO systems,accurate channel estimation information plays an important role in precoding,resource allocation at the base station and signal detection,equalization at the user side.Sparse channel estimation technology in frequency selective multipath fading channels is one of the most important and challenging research directions in Massive MIMO systems.However,most of the existing channel estimation techniques are only suitable for 4G small scale MIMO systems(such as the LTE-A system of 8 antennas).Therefore,this paper explores the 5G large-scale MIMO multipath sparse channel estimation technology.Based on the characteristics of space-time common sparsity of Massive MIMO channels,this paper uses advanced compressed sensing theory to study the channel estimation algorithm that reduces the pilot overhead without reducing the estimation accuracy.Firstly,based on the space-time common sparsity of Massive MIMO channels and joint difference theory,a joint differential scheme based sparse channel estimation is proposed.The characteristics of the improved algorithm are as follows: First,the improved algorithm updates multiple vectors at the same time during each iteration,enhances the sparsity structure,and improves the performance of the algorithm.Second,in the process of reconstruction,the improved algorithm does not only deal with the pilot signals received at the present time,but the difference of the pilot signal of the previous frame to further enhance the sparsity,thus improving the precision of the algorithm reconfiguration.The simulation results show that the proposed algorithm can achieve better parameter estimation performance while reducing the pilot cost.It also proves that the compressed sensing theory is robust in the Massive MIMO-OFDM system environment.Secondly,in consideration of sparsity unknown in real environment,a sparsity adaptive channel estimation scheme is proposed.Most of the compressed sensing theory reconstruction algorithms,such as CoSaMP,need the sparsity of the known signal as a priori condition.However,in the actual environment,the sparsity of the wireless channel is unknown.Therefore,the proposed algorithm uses the space-time common sparsity of the Massive MIMO channel to reasonably set the stopping iteration parameters under different SNR to obtain accurate dynamic sparsity.The experimental results show that,compared with the traditional CoSaMP and S-CoSaMP algorithms,the SSA-CoSaMP algorithm has better channel estimation performance under the same SNR and can obtain the sparsity adaptively.