Principles and design of model-based robot controllers

Abstract

Model-based control algorithms for industrial manipulators require the on-line evaluation of robot dynamics and are particularly sensitive to modelling errors. The development of a unifying framework for the analysis and design of model-based robot control strategies is the theme of this paper. In this framework, the practical problems associated with real-time implementation are highlighted and methods to improve the robustness of the closed-loop system are suggested.