Parallel Processing of Robot Dynamics Simulation Using Optimal Multiprocessor Scheduling Algorithms

This paper discusses the parallel processing of real‐time robot dynamics simulation. Real‐time robot dynamics simulation is an indispensable technique for a robot with a high‐level function. It is a computation procedure to determine the robot motion (joint position, speed, acceleration, etc.) for the given torque and driving force at each joint. In the proposed scheme, the optimal multiprocessor scheduling algorithms developed in this paper are employed, and the dynamics of the robot arm with arbitrary shape can be simulated with the minimum processing time on a multiprocessor system composed of an arbitrary number of parallel processors. The effectiveness and the practical usefulness of the proposed parallel processing scheme ire demonstrated on the robot motion simulator using a prototype multiprocessor system. This is the first report of the robot dynamics simulation being realized efficiently by the parallel processing on the real‐time multiprocessor system. Thus it is verified that the multiprocessor robot dynamics simulator with an excellent cost‐performance ratio can be realized.