The identification studie and the model obtained are really satisfactory regarding the objectives we define at the beginning of the studie. In term of definition of the prothesis we have to do a lot of work in term of ``what the system will be''. The sensor for example has to be completly defined, maybe the pressure is not the best information to work with. Probably a deplacement sensor will be more accurate. Regarding the comportment of the entire system we can ask another very important question. Has the system (like the muscular system) to be continuously active? And what kind of system can we use to reproduce (if it's necessary!) the non linearity of the response? It's necessary that all this questions will be considere not only from a ``system point of vue'' but essentially from a medical and a efficiency point of vue.