A gain scheduling approach to the feedback control of the knee-joint movement in paraplegic patients, who have recovered partial functionality of muscles through Functional Electrical Stimulation (FES), is studied. Since it was not possible to perform stimulation sessions on the patient during the work development, collected experimental data have been used to tune a known physiological model of the musculo-skeletal system involved FES, which is here adopted as a virtual patient, i.e. as the system to control. So, a nonlinear black box model is developed, by using I/O data set obtained from the physiological model simulator. With reference to such a black box model a nonlinear gain scheduling controller is designed, by using the knee-joint position as a scheduling variable and by properly interpolating different Linear Quadratic regulators. It is proven that the Linearization Property holds for the proposed controller. Furthermore, the performance of such a controller are analysed through closed-loop simulations, where the physiological model simulator is used to represent the knee-joint dynamics. The proposed controller shows good tracking and robustness properties on the full range of extension of the knee joint angle and simulation show that the presented strategy could perform better than any proposed linear approach to this problem. The results achieved for the control of the considered muscle group, encourage further research to exploit the use of the proposed technique with reference to more complex control problems, where more muscle groups at a time are involved, like FES-standing and FES-cycling.
Design of a Gain Scheduling Controller for Knee-joint Angle Control by Functional Electrical Stimulation
Carpanzano E;
2003
Abstract
A gain scheduling approach to the feedback control of the knee-joint movement in paraplegic patients, who have recovered partial functionality of muscles through Functional Electrical Stimulation (FES), is studied. Since it was not possible to perform stimulation sessions on the patient during the work development, collected experimental data have been used to tune a known physiological model of the musculo-skeletal system involved FES, which is here adopted as a virtual patient, i.e. as the system to control. So, a nonlinear black box model is developed, by using I/O data set obtained from the physiological model simulator. With reference to such a black box model a nonlinear gain scheduling controller is designed, by using the knee-joint position as a scheduling variable and by properly interpolating different Linear Quadratic regulators. It is proven that the Linearization Property holds for the proposed controller. Furthermore, the performance of such a controller are analysed through closed-loop simulations, where the physiological model simulator is used to represent the knee-joint dynamics. The proposed controller shows good tracking and robustness properties on the full range of extension of the knee joint angle and simulation show that the presented strategy could perform better than any proposed linear approach to this problem. The results achieved for the control of the considered muscle group, encourage further research to exploit the use of the proposed technique with reference to more complex control problems, where more muscle groups at a time are involved, like FES-standing and FES-cycling.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
