Research On Compressed Sensing-Based Wireless Channel Estimation For Massive MIMO

Massive MIMO(multiple input multiple output)is further researched and applied for 5G communication,as it can achieve fast data transmission,low delay propagation,enhanced channel capacity and high efficiency in frequency domain.As antenna number gradually grows at base station nowadays,traditional channel estimation algorithm performances poorly as channel estimation becomes more complicated.When CS(compressed sensing)theory is applied to channel estimation,better performance will be realized because of low pilot cost and low complexity algorithm.The main point of this thesis is to overcome the problems,e.g.,insufficient estimation accuracy and large calculation,appeared in CS based algorithm for sparse channel estimation.(1)The CS theory is introduced,and several matching pursuit algorithms are described in detail.Then the sparse channel estimation theory is introduced,which paves the way for the thesis.(2)In this thesis,a variable metric method gradient pursuit(VMMGP)algorithm is proposed for improving the performance in channel estimation of massive MIMO,which aims to re-duce the pilot cost and algorithm complexity,and also benefits to estimation accuracy and signal reconstruction efficiency.(3)A weighted orthogonal matching pursuit(WOMP)algorithm is proposed.The simula-tion results shows that WOMP(weighted orthogonal matching pursuit)can improve the low channel estimation accuracy of OMP(orthogonal matching pursuit)algorithm at low SNR(signal-to-noise ratio),and achieve higher estimation accuracy and lower pilot cost.