Optimal robot plant planning using the minimum-time criterion

Abstract

A path planning technique is presented which produces time-optimal manipulator motions in a workspace containing obstacles. The full nonlinear equations of motion are used in conjunction with the actuator limitations to produce optimal trajectories. The Cartesian path of the manipulator is represented with B-spline polynomials, and the shape of this path is varied in a manner that minimizes the traversal time. Obstacle avoidance constraints are included in the problem through the use of distance functions. In addition to computing the optimal path, the time-optimal open-loop joint forces and corresponding joint displacements are obtained as functions of time. The examples presented show a reduction in the time required for typical motions. >