Task-specific optimal simultaneous kinematic, dynamic, and control design of high-performance robotic systems

Abstract

A task-specific optimal simultaneous kinematic, dynamic and control design approach is proposed for high-performance computer-controlled machines, such as robots. This mechatronics design approach is based on the trajectory pattern method and a fundamentally new design philosophy that such machines, in general, and ultrahigh-performance machines, in particular, must only be designed to perform a class or classes of motions effectively. In the proposed approach, given the structure of the manipulator, its kinematic, dynamic, and control parameters are optimized simultaneously with the parameters that describe a selected trajectory pattern with which the desired class(es) of task(s) can best be performed. In one example, a weighted sum of the norms of the higher harmonics appearing in the actuating torques and the integral of the position and velocity tracking errors are used to form the optimality criterion. The selected optimality criterion should yield a system that is optimally designed to accurately follow the specified trajectory at high speed. Other objective functions can be readily formulated to synthesize systems for optimal performance. Based on the developed design methodology, a two-degrees-of-freedom robot manipulator with a closed-loop chain is optimally designed and constructed for point-to-point motions. The preliminary results of experiments indicate that the robot can, in fact, execute point-to-point motions rapidly and with minimal residual vibration. The potentials of the developed method and its implementation for generally defined motion patterns are discussed.