Experiments in modelling and end-point control of two-link flexible manipulators.

摘要

Current space manipulators employ control schemes that are severely limited by structural rigidity requirements. As a result, tasks are executed slowly and tediously. The goal of this research is to increase the performance of flexible manipulators by developing modelling techniques and control strategies that overcome these limitations.;Toward this goal, nonlinear equations of motion for two-link structurally-flexible manipulators were developed using the assumed-modes method. A novel approach for representing the link deflections, called the System Modes Representation, was developed, to realize fully the advantages of the assumed-modes method. It produced an accurate, low-order model. Accuracy gained with this modelling technique improved controller performance.;The manipulator model was used to design a variety of controllers--two are fully contrasted. The baseline controller is a conventional proportional-plus-integral-plus-derivative (PID) joint controller based on joint angle and joint rate measurements. The performance of this type of control is severely limited because the sensors are located coincident with the actuators.;Direct measurement of the manipulator end point can be used to overcome the limitations of joint control. End-point measurements were integrated into a controller design based on linear-quadratic Gaussian (LQG) theory. The end-point controller improved, in some cases by an order of magnitude, the trajectory tracking and disturbance rejection capabilities of the baseline controller.;An experimental two-link flexible manipulator, supported by air cushions on a smooth horizontal surface, was designed and constructed to verify and refine the modelling and control strategies.