Decomposition method for the constrained path planning of articulated systems

This paper describes a general approach to the optimal path planning of industrial robots. The proposed method is based on nonlinear programming, decomposition-coordination and feedback decoupling and linearization. The optimal path is obtained through the optimization of a time-energy criterion for the linearized model by static feedback. The proposed method does not suffer from curse of dimensionality. Decomposition of the original optimization problem at the joints levels makes possible parallel processing and reduces substantially the CPU time needed. The application of this control scheme to an n-link articulated system is developed in details. It is programmed on VAX-station and is used as a CACSD tool for controlling the PAMIR robot developed at the Control Laboratory of ESIEE.<>