Massive MIMO-OFDM Channel Estimation via Distributed Compressed Sensing

Massive multiple-input multiple-output orthogonal frequency division multiplexing (mMIMO-OFDM) channel estimation is considered recently utilizing compressed sensing (CS) based methods. Here, we proposed to use the joint sparsity of mMIMO-OFDM channels using stage-wise forward backward pursuit (StFBP) algorithm. In order to increase the speed of convergence and accuracy of estimation, we proposed to gather multiple good atoms in each step and to exploit common sparsity in the system model, respectively. Furthermore, the backward steps improve the accuracy by omitting bad previously gathered atoms. Simulation results represent the superiority of the proposed StFBP approach rather than the conventional CS-based and non-CS-based approaches.