Path Planning Algorithm Based on A_star Algorithm and Q-Learning Algorithm

The path planning algorithm is one of the most important algorithms in indoor mobile robot applications. As an integral part of ground mobile robot research, the path planning problem has greater research and application value. Based on machine learning, the mobile robot is continuously tried and trained in the simulation environment to eventually achieve the optimal path planning requirements for real-time obstacle avoidance, resulting in a new path planning algorithm. To make the planning goal smoother, after optimizing the global path planning A_star algorithm, it is necessary to combine the Q-learning algorithm, so this paper proposes the HA-Q algorithm. Under the HA-Q algorithm, the mobile robot can smoothly move from the specified starting point to the target point where the specified function is designated, to realize the functions of obstacle avoidance and path selection. After some simulation experiments, the HA-Q algorithm is more consistent with the ground mobile robot movement in the actual scene compared to the traditional algorithm. At the same time, these experimental results also show that the algorithm can be used to obtain a short and smooth path, avoid obstacles in real time, and effectively avoid the problem of falling into a locally optimal solution.