Energy Aware Opportunistic Routing for Energy Harvesting Wireless Sensor Networks

摘要

In this paper, the joint optimization of information transmission and energy utilization in energy harvesting wireless sensor networks (EH-WSNs) is studied. In EH-WSNs, the imbalance of energy harvesting and consumption affects the sensing and communication ability of the network. To solve this problem, a novel energy aware opportunistic routing protocol with long-short term memory (LSTM) based solar energy prediction is proposed. The protocol innovatively takes account of the nodes’ current residual energy and the harvesting solar energy in a short term which is predicted by a LSTM neural network as key factors in forwarding candidates election process of the opportunistic routing. Additionally, in order to jointly optimize the energy consumption and information transmission, a new metric with combined energy factor and relay efficiency for each node is proposed to assist candidate node selection where the relay priority considers the residual energy and working history of nodes. Simulation results show that the proposed protocol increases the network throughput by 12% and reduces retransmission rate by 15% compared to opportunistic routing based on geography. Simultaneously, it has stronger ability to balance energy consumption among sensor nodes.