A Hybrid APF Approach for Obstacle Avoidance in a 3D Environment

This study's main objective is to look at strategies, for avoiding underwater collisions and planning trajectories while navigating around static obstacles. Traditional Artificial Potential Fields (APF) face challenges such as getting stuck in local minima, which diminishes the path's optimality. Additionally, conventional APF often results in trajectories characterized by sharp turns or sudden changes in direction, particularly in environments populated with numerous obstacles. These obstacles can significantly hinder Autonomous Underwater Vehicles (AUVs) that depend on smooth trajectories for maintaining stability and control. To address these drawbacks, we present the Hybrid A* - Artificial Potential Field (HA-APF) algorithm. When using AUVs for collision avoidance duties, adherence to safety procedures is crucial. We propose an enhanced HA-APF algorithm to address this issue for efficient path planning and collision avoidance. Additionally, the A* algorithm, which has been altered to mitigate the innate problems with local optimality in the conventional APF approach, is used in conjunction with the improved APF technique. Finally, validations are conducted in three different regions with different beginning and finishing positions. These tests show how feasible and successful the suggested algorithm is in assisting AUVs in avoiding collisions while also carrying out path-planning duties. Through this work, we provide a robust tool for advancing AUV capabilities in real-world applications, ultimately allowing for more effective underwater exploration and intervention.