An integrated rapidly exploring random tree-based path planning method with initial and terminal attitude constraints for underactuated AUV

For underactuated autonomous underwater vehicles (AUVs), homing path planning with initial and terminal attitude constraints is essential for feasible docking. Conventional methods generally use position reachability as the terminal condition and may produce unsuitable approach directions. This paper proposes IAPRRT, an integrated attitude-constrained homing planner based on a modified rapidly exploring random tree (RRT). First, an improved A* algorithm (IA*) constructs a global guiding path through midpoint substitution, and the resulting waypoints guide RRT sampling and termination. Then, the RRT incorporates hybrid sampling, attitude-constrained node generation, parent-node expansion limitation, and terminal-attitude termination conditions. Finally, multi-population particle swarm optimization (MPPSO) refines the path to satisfy attitude-variation constraints. Comparative simulations demonstrate that IAPRRT generates shorter paths with higher planning efficiency than conventional methods while satisfying position and attitude constraints. The method provides an effective path planning solution for AUV homing to support autonomous docking.