Search

Functional electrical stimulation (FES) can be used to initiate lower limb muscle contractions and has been widely applied in gait rehabilitation. Establishing the correct timing of FES activation during each phase of the gait (walking) cycle remains challenging as most FES systems rely on open-loop control, whereby the controller receives no feedback about joint kinematics and instead relies on predetermined/timed muscle stimulation. The objective of this study was to develop and validate a closed-loop FES-based control solution for gait rehabilitation using a finite state machine (FSM) model. A two-phased study approach was taken: (1) Experimentally-Informed Study: A neuromuscular-derived FSM model was developed to drive closed-loop FES-based control for gait rehabilitation. The finite states were determined using electromyography and joint kinematics data of 12 non-disabled adults, collected during treadmill walking. The gait cycles were divided into four states, namely: swing-to-stance, push off, pre-swing, and toe up. (2) Simulation Study: A closed-loop FES-based control solution that employed the resulting FSM model, was validated through comparisons of neuro-musculo-skeletal computer simulations of impaired versus healthy gait. This closed-loop controller yielded steadier simulated impaired gait, in comparison to an open-loop alternative. The simulation results confirmed that accurate timing of FES activation during the gait cycle, as informed by kinematics data, is important to natural gait retraining. The closed-loop FES-based solution, introduced in this study, contributes to the repository of gait rehabilitation control options and offers the advantage of being simplistic to implement. Furthermore, this control solution is expected to integrate well with powered exoskeleton technologies. [ABSTRACT FROM AUTHOR]

We have recently demonstrated in simulations and experiments that a proportional and derivative (PD) feedback controller can regulate the active ankle torque during quiet stance and stabilize the body despite a long sensory-motor time delay. The purpose of the present study was to: 1) model the active and passive ankle torque mechanisms and identify their contributions to the total ankle torque during standing and 2) investigate whether a neural-mechanical control scheme that implements the PD controller as the neural controller can successfully generate the total ankle torque as observed in healthy individuals during quiet stance. Fourteen young subjects were asked to stand still on a force platform to acquire data for model optimization and validation. During two trials of 30 s each, the fluctuation of the body angle, the electromyogram of the right soleus muscle, and the ankle torque were recorded. Using these data, the parameters of: 1) the active and passive torque mechanisms (Model I) and 2) the PD controller within the neural-mechanical control scheme (Model II) were optimized to achieve potential matching between the measured and predicted ankle torque. The performance of the two models was finally validated with a new set of data. Our results indicate that not only the passive, but also the active ankle torque mechanism contributes significantly to the total ankle torque and, hence, to body stabilization during quiet stance. In addition, we conclude that the proposed neural-mechanical control scheme successfully mimics the physiological control strategy during quiet stance and that a PD controller is a legitimate model for the strategy that the central nervous system applies to regulate the active ankle torque in spite of a long sensory-motor time delay. [ABSTRACT FROM AUTHOR]

Our studies have recently demonstrated that a proportional and derivative (PD) feedback controller, which takes advantage of the body's position and velocity information to regulate balance during quiet standing, can compensate for long neurological time delays and generate a control command that precedes body sway by 100-200 ms. Furthermore, PD gain pairs were identified that ensure a robust system behavior and at the same time generate dynamic responses as observed in quiet standing experiments with able-bodied subjects. The purpose of the present study was to experimentally verify that the PD con- troller identified in our previous study can: 1) regulate the active ankle torque to stabilize the body during quiet standing in spite of long neurological time delays and 2) generate system dynamics, i.e., a motor command and body sway fluctuation, that success- fully mimic those of the physiologic system of quiet standing. Our real-time closed-loop feedback circuit consisted of a center of mass position sensor and a functional electrical stimulator that elicited contractions of the plantar flexors as determined by the aforementioned PD controller. The control system regulated upright stance of a subject who was partially de-efferented and de-efferented due to a neurological disorder called von Hippel-Lindau Syndrome (McCormick Grade III). While the subject was able to generate a motor command for the ankle joints, he could not regulate the resulting torque sufficiently due to a lack of sensory feedback and motor control. It is important to mention that a time delay was included in the closed-loop circuit of the PD controller to mimic the actual neurological time delay observed in able-bodied individuals. The experimental results of this case study suggest that the proposed PD controller in combination with a functional electrical stimulation system can regulate the active ankle torque during quiet stance and generate the same system dynamics as observed in healthy individuals. While these findings do not imply that the CNS actually applies a PD-like control strategy to regulate balance, they suggest that it is at least theoretically possible. [ABSTRACT FROM AUTHOR]