Research on Efficient Data Collection Strategy of Wireless Sensor Network

Recently, compressed sensing (CS) has been widely studied to collect data in wireless sensor network (WSN), which usually adopts greedy algorithm or convex optimization method to reconstruct the original signal, but noise and data packet loss are not considered in the compressed sensing (CS) model. Therefore, this paper proposes an efficient data collection strategy based on Bayesian compressed sensing (BCS-DCS) to improve the performance of data acquisition performance and network lifetime. Firstly, adopt a random packet loss matrix to update the compressed sensing (CS) model and employ Bayesian method to reconstruct the original signal. Then, according to the covariance of the reconstructed signal and the energy constraint of the sensor nodes, construct the active node selection optimization framework to optimize the node selection matrix quickly and effectively. Simulation results show that the efficient data collection strategy based on Bayesian compressed sensing (BCS-DCS) proposed in this paper can improve the signal acquisition performance and lifetime of wireless sensor network (WSN).