基于人工势场法的机械臂动态运动基元局部平滑避障方法OA

Aiming at the low modeling efficiency and oscillation problem of obstacle avoidance trajectory planning for manipulators in space,a local smooth obstacle avoidance method based on artificial potential field method for dynamic motion primitives of manipulators is proposed.This method uses dynamic motion primitives(DMP)to learn the charac-teristic parameters of a single teaching trajectory,realizing fast generalised modeling of similar trajectories for any target point in space;by setting obstacle distance constraints and introducing the artificial potential field method(APF)to plan local obstacle avoidance for the generalised trajectory of DMP,a virtual target point strategy is proposed to improve the planning stability of the trajectory;using cubic Bezier interpolation to generate transition curves,achieving smooth obstacle avoidance and eliminating trajectory oscillation problems.To verify the effectiveness of the method,obstacle avoidance experiments for manipulators in static and dynamic obstacle scenarios are conducted against the background of non-ferrous metal de-slagging processes.Experimental results show that compared to traditional obstacle avoidance planning methods,the method proposed in this paper has improved average planning time efficiency and trajectory smoothness,with promising application prospects.