Dynamic restructuring mechanisms for robot manipulator controls

The characteristic behavior of three robot manipulator control schemes are analyzed. The controller designs were based on three different control theories, i.e. linear multivariable control, variable structure system, and model-reference adaptive control. The strength and weakness of each controller design are first identified. Then, the three controllers are combined to form a bank of controllers. Finally, two restructuring control mechanisms and three restructuring laws are derived and used to coordinate the operational behavior of the controllers and to dynamically select the proper control action for a prespecified sub-task. The performance of the restructuring control schemes and laws is illustrated by presenting simulation results of the PUMA-560 robot arm.<>