Kalman smoothing with soft inequality constraints for space robot teleoperation

Abstract

Space robot teleoperation system with haptic device has two operation modes: offline teleoperation mode and online teleoperation mode. High acceleration and jerk produced by the teleoperation system may degrade the performance of trajectory tracking and cause large residual vibration. In this article, human hand is treated as a kind of sensor, and two smoothers based on the standard Kalman smoother have been proposed to limit the acceleration and jerk in joint space in a certain range. On the one hand, the offline smoother for offline teleoperation mode makes use of a smoothing algorithm proposed in the literature, which is based on the interior point techniques. On the other hand, the online smoother for online teleoperation mode is proposed and named as reduced-Q method, which is a fixed-lag smoother with adjustable measurement noise. Experimental results have shown that the methods can constrain the acceleration and jerk of the trajectories in joint space within the specified range approximately and can improve the tracking accuracy without causing too many deformations. Thus, the psychology burden of operator will be alleviated.