Vision-Based Task-Level Control of a Flexible-Link Manipulator

Abstract

This paper discusses a vision-based approach to implement task-level control in flexible-link manipulators. The proposed approach emphasizes the advantage of using vision in the control of flexible manipulators. It is pointed out that taking advantage of the inherent robustness, implementation of an image-based visual servo can be regarded as a synthetic solution to precise task-level control of flexible manipulators. This approach is implemented in a three-dimensional flexible-link manipulator. The implementation makes good use of filters in decoupling task-level control and vibration suppression control. Moreover, we point out that although the robustness of the approach can help to overcome the difficulty in control resulting from the complex measurement of the link's elastic deformation, it lacks in capability of tip trajectory specification. This problem is analyzed in this research and it leads to the proposal for the integration of the image interpolation technique. This technique makes the proposed approach adequate for tasks involved with complex tip trajectories. For flexible-link manipulators, the proposed approach with the remedy is the first vision-based synthetic solution that attempts to make a flexible manipulator usable for a practical task.