On-Line Kinematical Optimal Trajectory Planning for Manipulator

There are several obstacles to apply trajectory planning on-line to Haptic Feedback Manipulator (HFM), one of which is to quickly find a plan to optimize the cost desired, while avoiding local minima. In this paper, we proposed and tested an on-line trajectory planning method based on kinematical optimization. The particle swarm optimization (PSO)-based trajectory planner of the polynomial interpolation method optimized the minimum time and acceleration for velocity constraint. To enable the real-time of the planner, machine learning was employed for building a regression model able to find an efficient initial solution. Simulation results indicated that trajectories with the compromise between safety and rapidity can be optimized in real time the by presented method.