Robot path guiding method and device based on reinforcement learning optimization and medium

The invention belongs to the field of robot path planning, particularly provides a robot path guiding method and device based on reinforcement learning optimization, computer equipment and a storage medium, and adopts a reinforcement learning algorithm to optimize a DMP algorithm so as to improve the trajectory planning capability and the obstacle avoidance capability at the same time. The problem that an original algorithm cannot give consideration to the obstacle avoidance performance and the trajectory simulation performance at the same time is effectively solved, the trajectory fitting degree is improved on the premise that obstacle avoidance is guaranteed through the algorithm, trajectory planning of fixed path points can be achieved, and the capability of customizing a specific trajectory according to needs is achieved. 本发明属于机器人路径规划领域，具体提供了一种基于强化学习优化的机器人路径引导方法、装置、计算机设备及存储介质，采用强化学习算法优化DMP算法，以同时提高轨迹规划能力和避障能力。有效解决了原有算法不能同时兼顾避障性能和轨迹模仿性能的问题，使得算法在保证避障的前提下，提升了轨迹贴合程度，可以实现固定路径点的轨迹规划，具备按需求定制特定轨迹的能力。