Adaptive control of robot manipulators with controller/update law modularity (Q1304032)

The authors develop an adaptive link position tracking controller for robotic manipulators which achieves modularity of the controller and update laws. They design a torque input control law to achieve input-to-state stability. The parameter estimation error is treated as a disturbance input to the link position tracking error dynamics. Consequently, the torque controller ensures closed-loop stability for any adaptive update law design that guarantees that the parameter estimates remain bounded. The authors then show how the design of the torque input control law provides global asymptotic link position tracking for any adaptive update law design which also satisfies the following conditions: the prediction error is square integrable, and the estimated inertia matrix is positive definite. To illustrate the advantages of their modular adaptive control design, the authors compare their proposed law using a standard least square estimator to the well-known controller of Slotine and Li using both simulation and experiments. Experimental results show improved tracking performance probably due to limited, indirect improvement in the control of transient performance.