Section: Application Domains

Non invasive stimulation (external FES)

Both triggered open-loop and closed-loop FES controllers that we are developing for movement involve several sensors and stimulators whose activities must be precisely coordinated by the controller. For instance, the stimulation controller is fed back by various sensors, such as limb joint angles, IMUs providing accelerations, and electrophysiological signals like EMG. These signals are then used by feedback controllers to accurately control the artificially actuated limbs by means of stimulators. This distributed architecture is often deployed on a wireless network since it distinctively complies with mobility constraints, leading to good acceptance from human users. The quality of service (QoS) of this network influences the controlled system properties and the quality of control (QoC). The control performance and robustness of this system can be very far from expectations if implementation-induced disturbances are not taken into account. Thus, the overall performance of a real-time control system must be assessed not only with respect to deadlines (as in classical scheduling analysis) but also by considering other criteria such as time-varying delays and jitter. Hence, research on the joint design of control, computation and communication has to be carried out and applied [49] to the particular case of FES control loops distributed over imperfect links and low power nodes. In addition to the elaboration or adaptation of algorithms, specific tools must be further developed to assess the effectiveness of the new control algorithms and to support their implementation. In particular, realistic simulations remain a precious tool ahead of real experiments to ensure that the implementation meets the functional and safety requirements without danger. This is, for example, the case of the hybrid simulation framework of our distributed FES system currently under development [6]. Understanding and modeling the influence of an implementation (support system) on QoC is a challenging objective in a distributed control design process, but it is mandatory to guarantee the system’s safety and effectiveness.