Research On Emg Based Control And Force Tactile Perception Feedback Of The Prosthetic Hand

Prosthetic hand is a typical human-machine interaction device.It is important for upper limb amputee recovering hand function.In the early time,prosthetic hand is for the purpose of decoration.At present,electric-driven prosthetic hands are the main prostheticses.It integrates many new technologies including biomedical science,computer science,electronics,control,robotics,etc.Among numerous electric-driven prosthetic hands,the electromyographic(EMG)based control prosthetic hands are the most popular.This research is supported by the National Science Foundation Program and Jiangsu Province Science and Technology Plan Project.This paper has carried on the research of EMG based control and force tactile perception feedback of the prosthetic hand for the purpose of improving the controllability and perceptibility of the prosthetic hand.Firstly,the research status on the aspects of mechanism design,control strategy,information perception and etc are analyzed.Based on the analysis of the research status,a prosthetic hand with two degree of freedom(DOF)is designed.The designed hand has the function of hand open,hand close,wrist rotation.The mechanism design of this prosthetic hand is described in detail.According to the need of information perception of the prosthetic hand,a machinery-sensor integrated prosthetic hand finger is designed.The grasp force and the contact position of the prosthetic hand finger and the object can be detected by the designed finger.Based on the mechanism design of the prosthetic hand,the measurement and control system is designed including hardware measurement and control circuit and software.In terms of hand motion recognition,the support vector machine(SVM)is used for the recognition of the eight motions of hand and wrist.The generalized regression neural network(GRNN)is used for the estimation of the grasp force and three-dimension push and pull force.In one DOF commercial prosthetic hand,a pair of surface EMG sensors is attached on the surface skin of a pair of antagonistic muscles of the arm stump to capture the EMG signals.The captured EMG signals are used to control the prosthetic hand after signal shapping and filtering.However,different people have different EMG signal strengths,and the EMG signals are also different when EMG sensors are placed in different positions.Due to these factors,the control parameters need to be adjusted to different individuals when they begin using a prosthetic hand.Additionally,a long period is usually required to fully adapt to their prosthetic hands.In order to overcome these drawbacks,an EMG signal decoupling adaptive learning motion recognition method is proposed in this paper.A two channel EMG signals decoupling model is proposed to eliminate the overlap interference between the two channel signals.An EMG adaptive learner is designed to adjust the normalized scale factor.With the help of signal decoupling model and the adaptive learner,the individual differences of the EMG can be eliminated,and the hand motion adaptive recognition based on surface EMG can be realized.On the basis of the hand motion recognition,a fuzzy neural network PID controller is designed to realize the proportional control of the prosthetic hand.In the process of grasping an object,the kinetic parameters of the object will change.It is bad for the stable grasp control of the prosthetic hand.In this paper,a control method based on the object's stiffness fuzzy observation is proposed to realize the stable grasp.A backstepping controller based on Lyapunov theory is designed.A fuzzy observer is designed to detect the stiffness of the object in real time.And the stiffness is used to adjust the parameter of the backstepping controller.Information perception is an important feature of the intelligence prosthetic hand.At present,most of the prosthetic hands are lack of perception.Some prosthetic hands have information perception,but the perceived information is only feedback to the controller of the hand and the user cannot perceive the state of the hand.In response to this issue,the research of force tactile perception feedback is carried on in the paper.A tactile representation device is designed.And the designed device is used to provide the hand's user with the statues information of the prosthetic hand.Tactile coding of three force tactile information(grasp force,wrist three-dimensional force and sliding information of object in the prosthetic hand)are designed based on the vibration tactile technology.