Stabilization of delayed teleoperation using predictive time-domain passivity control

Stability is an important issue in bilateral teleoperation due to different factors like varying operating conditions, existence of time-delays, etc. Recently, time-domain passivity control (see [1]) has been proposed to stabilize unstable telerobotic systems and has shown good performance in situations where one or more components of the system behave actively. However, this scheme requires the computation and updating of passivity controllers in real-time in order to dissipate the surplus energy thus making it difficult to be used in many telerobotic scenarios where time-delays are inevitable. In this contribution, we present a unique approach to solve this problem by using a Kalman filter based energy predictor which allows us to design master and slave passivity controllers even in the presence of time-delays in forward and backward channels. In order to accurately predict the slave-side energy levels, the predictor makes use of a recursively updated model of the remote system. Provision of slave or environment models is not required because an adaptive mechanism constructs them from the command velocity and delayed force information. Simulation results are provided showing the validity of proposed approach.