Fast AM-RRT*: A Fast and Efficient Path Planning Algorithm Based on AM-RRT* in Complex Environments

As the application of robotic technology in complex environments continues to expand, efficient path planning algorithms have become crucial. AM-RRT*, a variant of the rapidly exploring random tree algorithm, is widely applied in robot planning problems due to its efficiency. This paper proposes an extended algorithm based on AM-RRT*, namely Fast AM-RRT*, aimed at improving the efficiency and quality of path planning for robots in dynamic environments. The Fast AM-RRT* algorithm introduces a bidirectional search framework, allowing the path search to be conducted simultaneously from both the start and goal points, significantly reducing search time. Additionally, this algorithm employs a novel node selection strategy that considers both the distance and directionality between nodes, resulting in smoother and shorter paths. Through extensive simulation experiments, we validate the performance of Fast AM-RRT* in various scenarios. The results show that compared with the existing AM-RRT*, Fast AM-RRT* reduces the path length and search time by an average of 4.8% and 64.7%, respectively. These improvements make the Fast AM-RRT* algorithm advantageous in terms of real-time response and dynamic obstacle avoidance, providing an effective solution for the field of robotic motion planning.