U-Shaped Obstacle Avoidance for a Bionic Robotic Fish: A Virtual Sentinel Obstacle Strategy Based on the Artificial Potential Field Method

Reliable obstacle avoidance is essential for bionic robotic fish operating in complex underwater environments. However, when a robotic fish performs depth-keeping cruising near U-shaped obstacles, the traditional artificial potential field (APF) method is prone to local minima, which can cause the vehicle to become trapped and lead to obstacle avoidance failure. To address this problem, this paper proposes a virtual sentinel obstacle strategy based on the APF method. Virtual sentinel obstacles are deployed near the entrance of U-shaped obstacles, and corresponding deployment rules are formulated to prevent the robotic fish from entering the local minimum region. To further improve path planning performance, a two-stage fuzzy controller is developed to adjust heading rotation and cruising step size. The proposed method is evaluated through numerical simulations and physical experiments using a self-developed bionic robotic fish prototype. The results show that the virtual sentinel obstacle strategy prevents entrapment around the tested U-shaped obstacles, while fuzzy control shortens the path and improves smoothness. The physical experiments further verify the feasibility of the proposed strategy in a two-dimensional underwater obstacle avoidance scenario. These results indicate that combining virtual sentinel obstacles with APF-based planning provides a feasible approach for U-shaped obstacle avoidance by bionic robotic fish.