Multi-objective three-dimensional path planning algorithm for autonomous underwater vehicles in marine environments

This paper addresses AUV path planning for multi-location detection in complex underwater environments. A two-stage hybrid algorithm, RRT* Swarm Synergy for Multi-Location Path (RSS-MLP), is proposed, integrating improved RRT* with enhanced particle swarm optimization (PSO). It decomposes the multi-location problem into single-location subproblems via segmented planning: improved RRT* generates initial feasible subpaths, and enhanced PSO optimizes them locally. Experimental results in simulated 3D underwater environments show that RSS-MLP produces paths 6–15% shorter than those from particle swarm, ant colony, and genetic algorithms, with significantly better planning efficiency, convergence speed, and accuracy. RSS-MLP offers superior comprehensive performance and can plan efficient, smooth, and robust 3D navigation paths for AUVs.