PPO-based charging path planning in WRSN for wide-area lake with ellipsoidal MC
Year: 2026
Type: article
Abstract: The deployment of underwater wireless rechargeable sensor networks (UWRSN) in wide-area lake environments for underwater resources monitoring represents an emerging approach to support effective water resource management and aquatic ecosystems preservation. A key challenge in such networks lies in maintaining the operation of underwater sensors, as their batteries are difficult to replace and their limited battery capacity constrains long-term data acquisition. Employing mobile chargers (MC) to recharge these sensors through optimized path planning offers a viable solution. However, underwater MCs are subject to complex physical influences, such as resistance, buoyancy, and gravity, which considerably impact their mobility and consequently, their charging efficiency. Existing studies on charging path planning, largely focused on two-dimensional and aerial scenarios, are not directly applicable to underwater environments due to their insufficient modeling of MC morphology and mobility-related energy consumption. To address these issues, this paper proposes a novel charging path planning for WRSN in wide-area lakes using ellipsoidal MCs. The proposed approach is specifically designed for three-dimensional underwater settings and incorporates detailed modeling of the ellipsoidal MC. By leveraging deep reinforcement learning techniques, it optimizes the charging path of MCs. Extensive simulation results show that the proposed method achieves significantly better charging performance compared to conventional strategies such as NJNP, tabular Q-learning, and DQN. Evaluated under identical hydrodynamic conditions with consistent depth-dependent water density and drag coefficients, the proposed method offers notable advantages in terms of charging efficiency, sensor node mortality rate, and total energy consumption to ensure the long-term integrity of environmental data.
Source: IEEE Sensors Journal
Language: English
Subfield: Ocean Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: hybrid
Estimated APC paid: $2,645
OpenAlex ID: W7220740913