Manipulator Motion Planning in the Presence of Obstacles and Dynamic Constraints

Abstract

Motion planning algorithms are traditionally restricted to finding spatial or "physical" paths only, while manipulator dynamics is ignored. The attendant low-level problem of timing and velocity along the path is left to the trajectory planner. In this study, these two aspects have been combined. The overall problem has been cast as an optimal control problem. The obstacle avoidance conditions have been for mulated as state-variable inequality constraints. These con straints are shown to be continuously differentiable in our formulation. The continuous-time problem is converted to a discrete-time problem to permit the use of nonlinear pro gramming algorithms for the heavily constrained problem. Computational issues such as discretization effects and al gorithmic efficiency have been addressed. Simulations are performed for the proposed method.