A robust decoupled adaptive control of robots
A decoupled adaptive scheme is proposed for the trajectory control of robots. Under moderate constraints, the scheme is stable and robust in the presence of feedback delay and signal hold due to digital computation in the control loop. The design procedure involves a search for compensators with the aim of minimizing a quadratic performance index in order to minimize the maximum tracking error. Stability analysis of the proposed control law (the selected compensators) then provides stability bounds of disturbances, control and adaptation gains, and desired trajectories and their time-derivatives. The region of attraction is local due to unmodelled dynamics.
© 1995 Taylor & Francis Ltd. Received 1 February 1992. Revised 25 July 1994. Communicated by Professor P. A. Ioannou. The authors would like to thank Professors Joel Burdick, Thomas Caughey, and Athanasios Sideris at Caltech, and Dr Homayoun Seraji at the Jet Propulsion Laboratory, for valuable discussions, comments and suggestions. This work was supported partly by Caltech funds and partly by the Caltech/JPL Director's Discretionary Fund.