Remote Robot Control With High Force-Feedback Gain

NASA s Jet Propulsion Laboratory, Pasadena, California

An improved scheme for force-reflecting hand control of a remote robotic manipulator provides unprecedently high force-reflection gainQas much as 2 or 3, even when dissimilar master and slave arms are used. Until now, a gain of 1/10 has been the maximum attainable without loss of stability for a very stiff typical industrial robot arm; that is, the maximum force that could be applied to an operator's hand via a control handle was only one-tenth the force of contact (as measured by force and torque sensors on the robot) between the associated remote manipulator and the manipulated object. With such a low gain, small differences between the actual and commanded positions of the manipulator could create undesirably large contact forces. For safety and reliability, larger gains are needed. Part of the reflected force in the improved scheme is proportional to the positio n error, which is the difference between the actual position of the robot arm and t he position commanded by the operator. This position-error force-reflection concept is combined with compliance control (described below) to enable the required high gain.

Compliance control is implemented at the robot (which can be remote from the operatorUs control station) by low-pass-filtering the outputs of force and/or tor que sensors on the robot and using the filtered signals to alter the operatorUs posit ion and/or orientation command. The filtered force and/or torque feedback makes the robot hand behave as through, in each degree-of-freedom, a damped spring were in series with the otherwise stiff, controlled position manipulator. The improved control scheme was compared with seven other control schemes in experiments in which the operators controlled robots so that they inserted a peg in a hole and inserted and removed a screw in a tapped hole. In both tasks, position-error-based force feedback with compliance control proved to be the best control scheme, yielding the shortest completion times with the smallest contact forces.

Point of Contact:
Won S. Kim
Mail Stop 198-219
Jet Propulsion Laboratory
4800 Oak Grove Drive
Pasadena, CA 91109
818-354-5047
wonsoo@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