Adaptive Control Of Joint Movement Induced By Functional Electrical Stimulation

Functional electrical stimulation(FES)is a kind of neuromuscular rehabilitation technique to improve the functional movement ability for spinal cord injured(SCI)or stroke patients.It has some advantages such as portability,convenience,blood circulation promotion and muscle atrophy prevention.Because of complex,highly nonlinear and timevarying properties of human musculoskeletal system,external disturbance,and muscle fatigue,FES systems with open-loop control cannot guarantee desired,accurate and persistent movements.Thus,this thesis proposes a close-loop control method in order to adapt to the effects from the musculoskeletal modeling,external disturbance and muscle fatigue,and the following aspects are researched:(1)The dynamic properties of the skeletal muscle and the knee joint responding to the electrical stimulation are modelled.A neuro sliding mode control algorithm is proposed through boundary layer method to achieve equivalent sliding mode control.In this control algorithm,the radial basis function network is applied to estimate the system uncertainties.(2)Based on the above theoretical method,the FES experimental platform is built.The parameters of the human musculoskeletal models are determined and validated.The performance of the control method is evaluated in inducing knee joint movement only through agonist muscle stimulation in both simulation and experiments.(3)In order to achieve more natural movement performance,the proposed control method is extended to induce knee joint movement through stimulating a pair of agonistantagonistic muscles.The performance of the control method is also validated in simulation and human FES experimental study.