Towards DAE methodology for the control of cable-driven parallel robots

Models obtained from the laws of Physics are naturally written as differential-algebraic equations (DAE) also known as descriptor models. However, the usual methodology in control consists in reducing the model to a set of ordinary equations. Remaining in the DAE word has several advantages: the reduction step is not necessary any more and the parameter dependance might be simplified. This paper is a step forward towards a control methodology that would be entirely based on DAE models. Its goal is twofold. First, the tools and methods available for DAE models are reviewed. Second, a case taken form robotics is considered. The simulation being done with Maple, a linear model is derived and transferred to Matlab for synthesis. A H_∞ controller is then synthesized and evaluated. The controlled system is then simulated with Maple.