Time Domain Passivity Control for Position-Position Teleoperation Architectures

Abstract

This article presents a method for passivating the communication channel of a symmetric position-position teleoperation architecture on the time domain. The time domain passivity control approach has recently gained appeal in the context of time-delayed teleoperation because passivity is not established as a design constraint, which often forces conservative rules, but rather as a property which the system must preserve during operation. Since passivity is a network property, the first design rule within this framework is to represent consistent and comprehensible circuit (i.e., network) representations of the mechanical teleoperation system. In particular, the energetic behavior of these networks is interesting because it allows straightforward conclusions about system stability. By means of so-called passivity observers (PO) and passivity controllers (PC) (Hannaford & Ryu, 2001), the energetic response of a delayed communication channel is captured and modulated over time so that the network in question never becomes nonpassive. The case analyzed in this paper tackles a communication channel that conveys position data back and forth. This type of channel does not offer intuitive network representation since only flows are actually being transmitted. Although energy clearly travels from one side to the other, port power identification, as defined by the correlated pair flow and effort, is not evident. This work first investigates how this kind of channel can be represented by means of circuit networks even with the lack of physical effort being transmitted through the channel, and identifies which networks are susceptible to become nonpassive due to the channel characteristics (i.e., time delay, discretization or package loss). Once achieved, a distributed control structure is presented based on a PC series that keeps the system at the verge of passivity (and therefore stability) independent from the channel properties. The results obtained by the simulation and by experiment sustain the presented approach.