基于高斯混合模型的机器人曲面打磨路径规划

A Path Planning Method for Robotic Polishing of Curved Surfaces Based onGaussian Mixture Models

针对自由曲面上不规则区域的机器人打磨路径规划问题,提出一种基于高斯混合模型(GMM)的自适应路径规划方法。首先借助测地线的数值求解实现多边形网格上的指数映射与对数映射,然后基于黎曼流形上的GMM实现曲面上不规则区域的统计建模,并基于平行样式完成对不规则区域的打磨路径规划。所提方法简练描述了曲面上不规则区域的分布情况,并高效实现了局部覆盖路径规划。最后使用工业机器人在NURBS曲面工件的实验平台上进行实验,实验结果验证了该算法的可行性和有效性。

Aiming to solve the path planning problem for robotic polishing of randomly distributed regions on curved surfaces,an adaptive path planning method based on the Gaussian mixture model is proposed.Firstly,the approximate exponential map and the logarithmic map on polyhedral surfaces are achieved with the help of numerical computation of geodesics.Then the Gaussian mixture model on Riemannian manifolds is deployed to statistically model the shape of randomly distributed regions on curved surfaces.Besides,coverage path planning among these regions is performed by selecting the path pattern as direction-parallel.The GMM can concisely describe the distribution of randomly distributed regions on the surface,and the path planning method can efficiently plan the local coverage path.Finally,robotic polishing experiments on NURBS-designed curved-surface workpieces are completed.The experimental result verifies the feasibility and effectiveness of the method.