An AUV collision avoidance algorithm in unknown environment with multiple constraints

Ensuring the safety of Autonomous Underwater Vehicles (AUVs) is paramount for successfully executing their marine operations. Grounded in partial observation, motion characteristics, and safety distance constraints of AUVs navigating in uncharted environments, this study proposes a collision avoidance algorithm to guarantee AUV safety. The algorithm integrates the Dubins Improved Hybrid A* (DIHA*) algorithm and the Fuzzy Heading Avoidance (FHA) algorithm. The DIHA*-FHA algorithm dissects collision avoidance paths into target-oriented global paths and collision avoidance-oriented local paths. Subsequently, algorithmic functions are invoked based on different types of obstacle information. The components are delineated as follows: (1) We are employing the DIHA* algorithm to compute the optimal path solution aligning with the actual motion characteristics of the AUV, leveraging known prior obstacle information in the environment to guide the AUV toward the target. (2) Utilizing the FHA algorithm, grounded in a fuzzy controller, to swiftly respond to obstacle information detected by the AUV during its operation in an unknown environment. This involves calculating a local smooth path that maintains a safe distance from obstacles, facilitating effective collision avoidance. Ultimately, this study substantiates the superior performance of the DIHA*-FHA algorithm in real-time, path quality, and mitigation of local minima issues under multiple constraints in unknown environments, as compared to other algorithms. Experimental results underscore that this algorithm enhances the collision avoidance capability of AUVs in unknown environments with multiple constraints, playing a pivotal role in ensuring AUV safety in practical maritime operations.