Compressed Sensing-Based Sparsity Adaptive Doubly Selective Channel Estimation for Massive MIMO Systems

Algorithm 1

The B-SAMP Algorithm.

Input: sensing matrix , observed matrix , step size ;

Output: -sparse block approximation

Initialization: the iterative index , the support set

, the residual matrix , the initial channel

sparsity level and the stage index

Repeat

(1) (correlation)

(2) (support obtain)

(3) (support pruning)

(4) (block estimate)

(5) (residue update)

If  

(6) (Continue loop iteration)

Else

(7) (stage index update)

(8) (sparsity level update)

Until halting condition ture

Output: and obtain the estimation of channel