LNPP: Logical Neural Path Planning of Mobile Beacon for Ocean Sensor Networks in Uncertain Environments Using Hierarchical Reinforcement Learning

Localization of sensor nodes is a critical issue in Ocean Sensor Networks (OSNs). A key feature of these methods is the use of GPS-enabled mobile beacons that periodically broadcast their locations through path planning to achieve sensor localization. However, mobile beacons in OSNs are vulnerable to wind, waves, and currents, which prevent them from following pre-planned paths. Additionally, existing path planning algorithms are not suitable for handling uncertain obstacles and disturbances in OSN environments. To address these challenges, this article introduces a novel Logic Neural Path Planning (LNPP) scheme for mobile beacons. LNPP incorporates logical rule knowledge into its neural network in order to enhance learning efficiency and reduce unnecessary explorations. Furthermore, to effectively learn strategies across different time scales, a hierarchical reinforcement learning algorithm is proposed. The experimental results demonstrate that, compared to other methods, the proposed path planning scheme reduces the average localization time and the trajectory length by over 21 percent in obstacle-free OSN environments. Moreover, our scheme effectively operates in OSNs with uncertain obstacles.