Improved Hybrid A* Algorithm Based on Iterative Acceleration for Autonomous Vehicles

Efficient path planning for autonomous vehicles in complex indoor environments remains challenging due to narrow passages, dense obstacles, large search spaces, and nonholonomic constraints. To balance planning efficiency and path quality, this paper proposes an improved Hybrid A* path planning algorithm based on multi-objective optimization and iterative acceleration. First, a four-dimensional cost function is designed to jointly consider path length, curvature variation, obstacle potential, and reversing cost. Second, a dynamic acceleration factor is introduced to generate an initial feasible path rapidly and then progressively refine the path as ε decreases from 5 to 1. Third, a mean-filtering-based key-point extraction method is used to reduce the computational burden of subsequent path smoothing while preserving trajectory continuity. Real-vehicle experiments on unmanned transporters in a construction site environment show that the proposed algorithm reduces the initial planning time from 27,682.25 ms to 218.35 ms, shortens the final path length by 30.3%, and improves the initial planning efficiency by a factor of 126 while maintaining feasible and smooth trajectories in constrained spaces.