A Bayesian Compressive Data Gathering Scheme in Wireless Sensor Networks With One Mobile Sink

This paper studies the compressive data gathering problem in terms of Bayesian theory for wireless sensor networks (WSNs) with a mobile sink. In such WSNs, designing an optimal tour path for sink is a challenge task, because topology is time-varying and needs to consider latency. In this paper, we formulate the above-mentioned issue as an optimal problem. To resolve the NP-hard problem, a modified shuffled frog leaping algorithm with a delay constraint is provided, where chaos techniques are utilized to obtain a diversified population and an adaptive step update strategy is given to accelerate convergence ratio. In particular, a novel Bayesian compressive sensing-data gathering strategy is introduced, where constraint selection schedule of gathering nodes (GNs) is developed. It jointly considers nodes' residual energy and distance to the center of deployment area, thereby balancing network load. Mobile sink only visits those GNs rather than all sensor nodes (SNs) along the controlled path. In addition, SNs transmit data to its own GN through the shortest path, and thus, they are included within the same region called cell. More importantly, the capacity of our proposed scheme is analyzed, and we derive that it can achieve \Theta (W/(M\times (2+\Delta)^{2})) capacity per node. Finally, extensive simulations are implemented to demonstrate the efficiency of the proposed algorithm.