Teleoperation from movable bases: Modeling, analysis, design and experiments with a hydraulic motion platform.

摘要

There are manual control and teleoperation systems in which the operator is subject to base motion. Examples are aircraft piloting and joystick control of heavy hydraulic machines such as excavators. This thesis presents a new framework for the modeling, analysis, design and evaluation of controllers for teleoperation/manual control from movable bases.; First, a general model for teleoperation/manual control from movable bases is presented. Second, models for the operator dynamics are introduced and identified experimentally. Third, a novel robust μ-synthesis based control system design approach is proposed that addresses robust stability and performance issues. The approach is illustrated by a prototype single-degree-of freedom manual control task in which the operator positions his/her base, a hydraulic motion simulator, using a force-feedback joystick. The designed controllers are robustly stable with respect to parametric uncertainties in the arm/joystick as well as the feedthrough dynamics, and achieve a desired level of performance based on relevant measures. The proposed methods are compared with controllers that ignore the base motion through analysis and a set of experiments. The robust controllers suppress the operator-induced oscillations and produce well-damped responses, whereas the fixed base controllers become unstable.; Motivated by the need for high performance motion simulation in the evaluation of controllers for feedthrough cancellation, the position control of mechanical systems driven by hydraulic actuators is studied. This includes single-cylinder hydraulic servo-systems and multiple-degree-of-freedom hydraulic robots. Nonlinear dynamics of the actuators and nonlinear rigid body dynamics are considered in the model of hydraulic robots. For single-degree-of-freedom systems, valve dynamics are also included. Parametric uncertainties are allowed in these model.; Using the backstepping technique, novel nonlinear and adaptive/nonlinear control laws as proposed that incorporate the system dynamics into their design. It is proven via Lyapunov analysis that the closed-loop systems are stable and that the tracking errors converge to zero under the proposed feedback laws. The effectiveness of these approaches is demonstrated through a series of simulations and experiments carried out with the UBC hydraulic motion simulator.; Finally, a numerical approach is developed to optimize the feedback gains for a simplifies version of the proposed nonlinear controllers for hydraulic servo-systems.