Random Access Compressed Sensing with Unequal Probabilities in Wireless Sensor Networks

This paper considers the problem of selecting nodes in distributed compressed sensing, which is one of the best ways to make the measurement matrix. We propose a novel method based on Random Access Compressed Sensing (RACS) denoted by Unequal Probabilities Random Access Compressed Sensing (UPRACS). In the proposed scheme, nodes do not use the same probability when they decide whether to participate in the channel access process. Each node chooses a higher probability or a lower one according to the comparison between its measurement and a threshold. Simulation results demonstrate that UPRACS can lead lower recovery errors than RACS with the same energy consumption both in the whole fields and the interesting fields.