Trajectory planning and control of parallel manipulators

摘要

Kinematics based trajectory planning for parallel manipulators has been studied extensively, e.g., to determine a singularity-free path. Dynamics based trajectory planning, however, gains little concern. In order to fully exploit capacity of parallel robots, it is inevitable to conduct motion planning considering both kinematic and dynamic constraints. In this paper, three dynamics based trajectory planning methods, the time-optimal, the jerk-bounded, and the cubic polynomial planning, are investigated and implemented. The Orthopod, a 3-DoF purely translational parallel manipulator, is used as the experimental plant. The computed-torque control scheme is applied to follow the computed trajectory. In stead of manually tuning, control parameters are determined by formulating an optimization problem. Optimal parameters are thus obtained systematically. This study provides a quantitative comparison of the three planning methods and an engineering method to obtain optimal control parameters. The experimental results provide a guideline to choose an appropriate planning method for practical industrial applications.