Robot model validation

Abstract

Some initial results arising from a project involving robot validation are described. A theme of the project is that control, whether adaptive or non-adaptive, should be model-based; that is, the controller should contain a model of the dynamic system describing the robotic manipulator. Good control requires a good model, and obtaining such a model is the aim of the validation exercise. The paper focuses on validating models of three components of the robot system: link flexibility, DC motors, and manipulator dynamics. In each case, algorithms, simulation and experimental results are presented. The use of symbolic algebra techniques is emphasised. Relationships with other area of control and systems theory are explored. In particular, the relationships with parameter identification and fault detection are noted.