RL-EAR: reinforcement learning-based energy-aware routing for software-defined wireless sensor network

Efficient energy utilization in wireless sensor networks is paramount for the success of Internet of things (IoT) applications. Traditional energy-saving routing protocols typically compute isolated routes, which may only partially optimize energy consumption globally. Software-defined wireless sensor networking (SDWSN) is emerging as a crucial architecture for enabling IoT, leveraging software-defined network principles to facilitate routing operations. However, existing routing algorithms often need more capability to identify the most optimized paths. In this paper, we propose a reinforcement learning-based energy-aware routing (RL-EAR) algorithm to optimize the routing paths of SDWSN. Our method utilizes a reward function considering factors such as remaining energy, hop count, and congestion to optimize routes from any node to the controller node, aiming to enhance network lifetime. The SDWSN controller, acting as the agent, receives rewards based on its actions and refines routes over time through interactions with the data plane. The simulation results of the RL-EAR algorithm demonstrate its prominence over existing routing algorithms. Metrics such as network lifetime, packet delivery ratio, and the number of dead nodes over time are utilized to evaluate the effectiveness of our proposed algorithm. As a result, RL-EAR increases network lifetime by 42% compared to existing routing algorithms.