Model-Based Predictive Control for Improving Stability and Transparency in Time-Delay Teleopertation

Abstract

<p>Prior research on time-delay bilateral teleoperation has mainly resulted in methods that favor the robust stability of the system at the expense of its transparency. In contrast, it is demonstrated in this thesis that carefully designed model-based predictive controllers are able to achieve high levels of transparency while maintaining teleoperation stability. This is accomplished by utilizing available information on system model and time delay within a predictive control framework. The performance objectives are delay-free position tracking between the master and slave and the establishment of a virtual mass-damper tool impedance between the user and environment. Three model-based predictive controllers are proposed in this thesis which utilize both force and position measurements at the master and slave sites. The first two controllers can only handle known and fixed time delays whereas the third controller is applicable to systems with fixed or variable delay. </p> <p>First, building upon our recent work in [1, 2], a decentralized model-based predictive controller is introduced that can enhance the time-delay teleoperation transparency and robust stability by allowing the use of local delay-free measurements in local master I slave controllers. The stability of the system is analyzed using a frequency sweeping test. Numerical analysis demonstrates improved performance and robustness compared with the centralized controller in [2].</p><p>Next, a robust predictive controller is proposed to deal with modeling uncertainty, particularly in the operator and environment dynamics. In a two-step control approach, first local adaptive/nonlinear controllers are applied to linearize the system dynamics and to eliminate dependency on the master and slave parameters. Teleoperation coordination is then achieved by formulating an input/ output (I/O) time-delay H00 robust control synthesis. The transparency and robust stability properties of the proposed method are examined via numerical analysis.</p> <p>Although less sensitive to modeling uncertainty, the robust controller can still sacrifice teleoperation transparency in favor of its stability since it utilizes a fixed controller for the entire range of teleoperation. In an attempt to avoid such a tradeoff, a stable adaptive predictive controller for teleoperation systems with constant and varying communication delay is proposed. The controller utilizes a model of the system dynamics and the time delay within a predictive control framework to achieve the desired transparency objectives while maintaining the system stability. The controller adapts to uncertainties in the system dynamics by estimating the model parameters in real time. A Lyapunov analysis of the performance and stability of the resulting system is presented.</p> <p>The proposed controllers for time-delay bilateral teleoperation are implemented and experimentally evaluated. The results demonstrate the effectiveness of the proposed methods in providing a stable transparent interface for teleoperation under time delay.</p>