Building a modular robot control system using passivity and scattering theory

This paper analyses the problems and presents solutions for building a modular robot control system. The approach requires modeling the entire robot system using multi-dimensional passive networks, breaking the system into subnetwork "modules" and then discretizing the subnetworks, or n-ports, in a passivity preserving fashion. The main difficulty is the existence of "algebraic loops" in the discretized system. This problem is overcome by the use of scattering theory, whereby the inputs and outputs of the n-ports are mapped into wave variables before being discretized. By first segmenting the n-ports into nonlinear memoryless subnetworks and linear dynamic subnetworks and then discretizing using passivity preserving techniques such as Tustin's method, a complete modular robot control solution is obtained.

[1]  Robert J. Anderson,et al.  Using virtual objects to aid underground storage tank teleoperation , 1994, Proceedings of the 1994 IEEE International Conference on Robotics and Automation.

[2]  Robert J. Anderson Teleoperation with virtual force feedback , 1993, Other Conferences.

[3]  Mark W. Spong,et al.  Asymptotic Stability for Force Reflecting Teleoperators with Time Delay , 1992 .

[4]  M. V. Valkenburg Network Analysis , 1964 .

[5]  K. W. Lilly,et al.  A modular approach to multi-robot control , 1995, Proceedings of 1995 34th IEEE Conference on Decision and Control.

[6]  Robert J. Anderson,et al.  Autonomous, teleoperated, and shared control of robot systems , 1996, Proceedings of IEEE International Conference on Robotics and Automation.

[7]  Dean Karnopp,et al.  Introduction to physical system dynamics , 1983 .

[8]  Robert J. Anderson,et al.  SMART: a modular control architecture for telerobotics , 1995, IEEE Robotics Autom. Mag..

[9]  George C. Verghese,et al.  Design issues in 2-port network models of bilateral remote manipulation , 1989, Proceedings, 1989 International Conference on Robotics and Automation.