Successive hypothesis testing based sparse signal recovery and its application to MUD in random access

Based on successive hypothesis testing, we propose an approach for sparse signal recovery and apply it to random access to detect multiple block-sparse signals over frequency-selective fading channels. By introducing the sparsity variable, the proposed approach decides the presence or absence of the signal in each stage. To mitigate the error propagation, adaptive ordering is also employed as a greedy algorithm. From simulation results, it is shown that the proposed approach performs better than the block orthogonal matching pursuit algorithm, which is a well-known greedy compressive sensing algorithm for compressive random access.