基于概率运动基元的移动机器人轨迹学习与避障算法研究

Research on Trajectory Learning and Obstacle Avoidance Algorithms for Mobile Robots Based on Probabilistic Motion Primitives

示教学习在移动机器人的路径规划中展现出潜力,但将其直接应用于三维空间时,常面临效率低下、障碍物碰撞等挑战。提出了一种基于概率运动基元建模的移动机器人三维路径规划方法。通过简化速度信息,对示教路径点的时域坐标建模,实现了高效在线规划,并借助条件高斯计算提升路径准确性。设计了避障算法,利用障碍物信息赋予路径偏置值,结合一阶系统吸引点模型,确保路径平滑避障。实验验证表明,该模型在三维空间中规划效果良好,时间成本低,且避障算法有效,为移动机器人在复杂环境中的自主导航提供了新思路。

Demonstration learning has shown potential in path planning for mobile robots,but when it is directly applied to three-dimensional space,it often faces challenges such as low efficiency and obstacle collisions.A threedimensional path planning method for mobile robots based on probabilistic motion element modeling is proposed.By simplifying speed information and modeling the time-domain coordinates of teaching path points,efficient online planning has been achieved,and path accuracy has been improved through conditional Gaussian calculation.An obstacle avoidance algorithm that utilizes obstacle information to assign path bias values,combined with a first-order system attraction point model is designed to ensure smooth obstacle avoidance along the path.Experimental verification shows that the model has good planning effect in three-dimensional space,low time cost,and effective obstacle avoidance algorithm,providing a new idea for autonomous navigation of mobile robots in complex environments.