Numerical complexity of decentralized dynamic control laws for manipulator systems

Abstract

SUMMARY The numerical complexity of robot dynamic models, inverse control structures and decentralized control laws is analyzed in a systematic fashion. Dynamic models of actuators (DC motors and hydraulic cylinders) are taken into account, too. The exact number of floating-point multiplications and additions/subtractions for an n −link manipulator is determined. The results are applied to several typical industrial manipulators. Measuring of computation time for various control structures is realized on an up-to-date 16-bit microcomputer, widely used in robot controllers. Thereby, the real-time applicability and effectively the “price” of different control laws are estimated.