Design And Research Of Intelligent Upper Limb Based On Surface EMG Signal

The intelligent upper limb prosthesis can effectively ameliorate the living situation of the amputated patient and improve their quality of life,which is a bio-electronic device used to replace the missing or damaged part of the human upper limb.However,most of the commercial upper limbs on the market are decorative prostheses or expensive functional prostheses,which can not meet the amputated patients demand for economical,lightweight and high-performance of artificial limbs.Therefore,an economical and practical intelligent upper limb prosthesis have a good market prospect.Traditional myoelectric control prostheses use traditional switch control,threshold control,etc.,which control system can only realize some simple functions such as prosthetic opening and closing control.The performance of Multi-degree-of-freedom myoelectric control prosthesis based on traditional machine learning pattern recognition algorithm depends on the characteristics of artificial design,which robustness is poor.In addition,the sensitivity of the acquisition device to the electrode position results in the system stability is poor.And it is inconvenience to the user that the patient has to carry out tedious calibration experiments before use.Most existing prostheses cannot meet the needs of the amputation patients.Aiming at the above problems,the paper proposes an upper limb motion pattern recognition algorithm based on multi-channel deep convolutional neural network for the myoelectric control of intelligent upper limb prosthesis.This algorithm converts the 8-channel surface EMG signal into an EMG image and decomposes it into four partial images based on skeletal muscle motion theory.The whole image and the four partial images are used as the input of the 5-channel deep convolutional neural network respectively.The algorithm network can learn overall feature information and key local feature information of the EMG image,which improves the ability of the network model to express the features of the EMG image.Improves the accuracy of upper limb motion pattern recognition.At the same time,because the convolutional neural network has a good ability to express deep abstract features,the upper limb motion pattern recognition algorithm based on multi-channel deep convolutional neural network is more robust and overcomes electrode displacement and impedance changes.As well as the influence of factors such as muscle fatigue on the recognition accuracy,the system's robustness is improved,and the control performance of the myoelectric control prosthesis and the comfort of the patient are improved.The paper proposed a myoelectric control prosthetic system based on sliding tactile perception feedback.This system used PVDF piezoelectric film as the sensor of sliding tactile information to collect the EMG signal,then input the EMG signal into the signal preprocessing circuit,Through the sliding tactile perception and discrimination,the fine prosthetic grip control is performed,so that the intelligent upper limb prosthesis has better bionic,dexterity and intelligence.The intelligent upper limb prosthesis adopts the electromyography control method based on the sliding tactile sense feedback to realize the autonomy and flexibility control of the intelligent upper limb prosthesis and faces the use of the forearm long residual limb amputation patient.At the end of this paper,the research results are summarized and the future research is prospected.