Neuro-Muscular Engineering
Hamid Reza Kobravi; Abbas Erfanian Omidvar
Volume 2, Issue 4 , June 2008, , Pages 335-349
Abstract
In this paper an adaptive robust fuzzy controller based on sliding mode control (SMC) approach is proposed to control the knee joint position using quadriceps electrical stimulation and it has been tested on three subjects. The proposed method is based on SMC. The main advantage of SMC derives from the ...
Read More
In this paper an adaptive robust fuzzy controller based on sliding mode control (SMC) approach is proposed to control the knee joint position using quadriceps electrical stimulation and it has been tested on three subjects. The proposed method is based on SMC. The main advantage of SMC derives from the property of robustness to system uncertainties and external disturbances. However, a large value has to be applied to the control gain when the boundary of uncertainties is unknown. Unfortunately, this large control gain may cause chattering on the sliding surface and therefore deteriorate the system performance. In this paper a robust control strategy proposed which is based on the combination of sliding mode, fuzzy logic systems, and an adaptive compensator to reduce the system uncertainties while alleviating the effects of chattering. The fuzzy logic system is used to identify the muscle-joint dynamics. The parameters of this fuzzy system were estimated using another fuzzy system. The controller is evaluated through the simulation studies on a virtual patient and experimental studies on intact subjects. The results show that the adaptive robust controller provides an accurate tracking of desired knee-joint angle for different subjects and different days and can generate control signals to compensate the muscle fatigue and reject the external disturbance.
Neuro-Muscular Engineering
Amir Homayoun Jafari; Seyed Mohammad Reza Hashemi Golpayegani; Farzad Towhidkhah; Ali Fallah
Volume -2, Issue 1 , July 2005, , Pages 57-70
Abstract
A hierarchical structure model with three levels is presented for modeling motor control in skill movements. At each level, based on accuracy and quality of control, a specific controller is activated. At first level, control concepts are qualitative. The duty of the first level is to provide stability ...
Read More
A hierarchical structure model with three levels is presented for modeling motor control in skill movements. At each level, based on accuracy and quality of control, a specific controller is activated. At first level, control concepts are qualitative. The duty of the first level is to provide stability of system, based on the received qualitative information from second level such as the decrement or increment of error. A self-organized controller at first level is used to generate qualitative control commands, and it plays an encouragement-punishment role to keep the stability of system by sending discrete commands to the second level. This controller only contributes at control action when the controller of second level can not preserve stability individually. At second level, control concepts are quantitative. The duty of the second level is adaptation and control of system accurately. The received information at this level generally comes from sensory and visual feedbacks, and it includes more accurate concepts of control action - like the amount of movement error. A model based on the predictive controller at second level generates quantitative control commands and indeed, determines trajectory of movement accurately. A fuzzy switch combines the control commands of first and second levels, based on the sliding mode strategy, to provide a robust control. At third level, this command is interpreted and then is applied to the involved muscles in movement. The received information at this level is generally the contribution of muscles in performing movement and the effects of environment on the movement, which comes from sensory feedbacks. The presented model with this hierarchical structure has a proper ability to control and keep the stability of system. The simulation results confirm this subject.