Functional electrical stimulation is an assistive technique used to produce functional movements in patients suffering from neurological impairments. However, existing open-loop clinical FES systems are not adequately equipped to compensate for the nonlinear, time-varying behaviour of the muscles. On the other hand, closed-loop FES systems can compensate for the aforementioned effects by regulating the stimulation to induce desired contractions. Therefore, this work aims to present an approach to implement a closed-loop FES system to enable angular positional control during wrist flexion. First, a Wiener model describing the response of the wrist flexor to pulse width modulated stimulation was identified for two healthy volunteers. Second, a nonlinear PID controller (subject-specific) was designed based on the identified models to enable angular positional control during wrist flexion. Subsequently, the controller was implemented in real-time and was tested against two reference angles on healthy volunteers. This study shows promise that the presented closed-loop FES approach can be implemented to control the angular position during wrist flexion or a novelty of the work when compared with the existing work.

Functional electrical stimulation is an assistive technique that utilizes electrical discharges to produce functional movements in patients suffering from neurological impairments. In this work, a biphasic, programmable current- controlled functional electrical stimulator system is designed to enable hand grasping facilitated by wrist flexion. The developed system utilizes an operational amplifier based current source and is supported by a user interface to adjust stimulation parameters. The device is integrated with an accelerometer to measure the degree of stimulated movement. The system is validated, firstly, on two passive electrical loads and subsequently on four healthy volunteers. The device is designed to deliver currents between 0-30mA, and the error between the measured current and simulated current for two loads were -0.967±0.676mA and -0.995±0.97mA. The angular data from the accelerometer provided information regarding variations in movement between the subjects. The architecture of the proposed system is such that it can, in principle, automatically adjust the parameters of simulation to induce the desired movement optimally by measuring a stimulated movement artifact (e.g., angular position) in real time.