SA-SWS: A Self-Adaptive Sleep/Wake-Up Scheduling Approach for Linear Sensor Networks

Linear sensor networks (LSNs) have attracted the increasing attention from many researchers due to a wide range of application scenarios with linear areas, such as bridges, tunnels, pipelines, etc. However, with constrained energy provision, LSNs are always facing the problems of limited energy and serious network holes. In this paper, SA-SWS, an adaptive sleep/wake-up scheduling approach, is proposed. This method can guide the nodes to decide by themselves to select actions in each time slot based on reinforcement learning. In addition, SA-SWSR, a derivative protocol of SA-SWS suitable for LSNs with redundant nodes deployed, is introduced to alleviate the energy hole problem. Extensive simulation results in comparison with the state-of-the-art work show the superiority of the proposed approach in terms of packet delivery latency, energy consumption, packet delivery ratio, and network throughput.