Massive MIMO Channel Estimation Based on Improved Variable Step Size Regular Backtracking SAMP Algorithms

Abstract

The traditional channel estimation methods LS and MMSE are used in MIMO channel estimation. Because of the large number of pilots required, the computation of inverse covariance matrix is required, which results in high computational complexity. Considering the sparsity of wireless channel in time domain, compressed sensing theory can be used to estimate the channel. The common greedy algorithms of compressed sensing include OMP algorithm and CoSaMP algorithm, which need to take sparsity as a known condition, so their use is limited. In this paper, an improved variable step-size regular backtracking SAMP algorithm based on compressed sensing theory is proposed to estimate the channel of Massive MIMO system. This algorithm improves the reconstruction accuracy of traditional SAMP algorithm in channel estimation and avoids the need for known sparsity, so it has good application value. In addition, good estimation results are also obtained under the noise environment, which proves the advantages of the algorithm.