Force-reflecting teleoperation over wide-area networks

Bilateral (force-reflecting) teleoperation in the presence of network induced communication constraints presents significant challenges in terms of stability and transparency of the teleoperator system. It is well-known that the stability of the force reflecting teleoperators is compromised in the presence of even small communication delays. Also, stability and transparency are conflicting goals; in particular, high force reflection gain provides a better kinesthetic as well as tactile feedback, however, it also destabilizes the overall system due to increasing the closed-loop gain. In this thesis, a set of results is presented towards stable and transparent force-reflecting teleoperation in the presence of communication constraints typical for serial communication networks. A small gain approach to network-based bilateral teleoperation is systematically developed. The approach is built upon a new version of the input-to-output stability small gain theorem for systems which communicate over multiple networked channels. Based on this theorem, schemes for bilateral teleoperation over networks are developed that guarantee the stability/tracking properties in the presence of network induced communication constraints. Projection-based force reflection algorithms are introduced that solve the contradiction between stability, maneuvrability, and high force reflection gain; in particular, these algorithms allow for achieving the stability for arbitrarily low damping of the master manipulator and arbitrarily high force reflection gain Next, the problem of design of the network-based teleoperators enhanced by means of virtual environment is addressed. The approach proposed is based on a nonlinear sampled-data design framework, and uses ideas from model-based control as well as multi-rate sampled data systems. The control algorithms and the communication protocols are presented that guarantee stability of the overall system under mild assumptions imposed on communication process.