Partition Heuristic RRT Algorithm of Path Planning Based on Q-learning

To solve the problem of high randomness in traditional rapid exploration random tree (RRT) path planning, a partition heuristic RRT planning algorithm based on Q-learning (Q-PRRT) is proposed. Partition heuristic rules are established by designing a sampling strategy of target bias and obstacle-avoided guidance. Markov modeling is carried out for partition heuristic RRT algorithm based on Q-learning, and actions are constructed based on partition heuristic rules. Each node is evaluated by designing the global optimal path reward function with Q-learning method, retain path nodes and eliminate redundant nodes based on greedy strategies. The simulation results show that Q-PRRT algorithm guarantees the optimality of the global planning path, obtains a smoother planning path, also improves path search efficiency and obstacle avoidance ability. It has better adaptability in different obstacle environments.