Adaptive Control for Cooperative Dual Robot Arms

NASA 's Jet Propulsion Laboratory, Pasadena, California

Three strategies have been proposed for the adaptive control of two cooperative robot arms. In comparison with other digital control systems that have been considered for multiple robot arms, controllers to implement these strategies would have simple structures and would be computationally fast enough for operation at high sampling rates. Dual-arm control strategies of the new type have yielded promising results in numerical simulations, and the general approach can be extended to greater numbers of arms. The control architecture is based on the tri-level hierarchical control of two ro bot arms (see figure). The high level plans the task to be performed and decomposes the task into appropriate subtasks for the right and left arms. In the intermedia te level, each subtask is transformed into a sequence of synchronous desired trajectories of end-effector motions and applied forces. The low level is concerned with the execution of the desired trajectories and employs feedback from the current status of the arms. In this hierarchy, the low level "achieves" the desired motion and operates in "millisecond" time scale, the intermediate level "determines" the motion desired for the subtask in "second" time scale, and the high level "plans" the sub-task sequences in "minute" time scale. Single-arm control strategies developed previously are applied at the low level of the hierarchy. In each dual-arm strategy, a suitable task-related coordinate fram e of reference is chosen for both arms, and the desired motions and applied forces for each arm are expressed in this frame. Then each arm moves as though it were carrying out the commanded motion by itself in this frame. The adaptive controllers ensure that the controlled variables track the desired reference commands and reject the unwanted disturbances caused by interaction forces and torques exerted by each arm on the other arm through the load. In the position/position control strategy, the adaptive controllers ensure that t he end-effector positions of both arms track desired trajectories in Cartesian space despite unknown time-varying interaction forces exerted through the load. In the position/hybrid control strategy, the adaptive controller of one arm controls end - effector motions in the free directions and applied forces in the constraint directions, while the adaptive controller of the other arm ensures that the end effector tracks desired position trajectories. In the hybrid/hybrid control strat egy, the adaptive controllers ensure that both end effectors track reference position trajectories while simultaneously exerting desired forces on the load. In the generalization to N cooperative arms (N > 2), an important feature of this approach is that the overall control system is reduced to N decentralized independent single-arm controllers. Available techniques for single-arm control can be utilized directly in multiple-arm environments. The control schemes require no communication among individual controllersQan appealing feature in terms of both computation and reliability. The new control strategies do not require the knowledge of such load parameters as mass and stiffness or such dynamic parameters of the robot as the masses and inertias of links. Therefore, these strategies can cope with uncertainties or variations in the parameters. Furthermore, the complex dynamic models of the arms are not used in generating the control actions.

Point of Contact:
Homayoun Seraji,
Mail Stop 198-219
Jet Propulsion Laboratory
4800 Oak Grove Drive
Pasadena, CA 91109
seraji@telerobotics.jpl.nasa.gov

Telerobotics Program page

Please email the site webmaster with any comments, criticisms or corrections for this page.
Maintained by: Dave Lavery
Last updated: May 10, 1996