Practical design of independent joint controllers for industrial robot manipulators

Compensation of manipulator nonlinear dynamics, which is typical of model-based control schemes, has been shown to give tracking performance improvements not only for direct-drive robots but also for gear-drive robots. This paper addresses the problem of robust design of linear independent joint feedback controllers in order to contrast imperfect dynamic modeling and disturbance effects. A new scheme that makes use of additional acceleration feedback is proposed and compared with the classical position + velocity feedback scheme, so as to provide robustness to disturbances at very low expense. Further, it is shown how linear feedforward compensation confers enhanced tracking capabilities to the schemes in the case of good model accuracy. Simulation results are provided.