Development Of A Haptic Intelligent Prosthetic Hand Control System Based On Myoelectric Pattern Recognition

Prosthetic hands can help people with physical disabilities to modify their physical defects,and most of them have basic functions such as gripping.Most of the existing commercial prosthetic hands do not have a haptic perception,which makes it difficult for people with physical disabilities to effectively acquire information about their surroundings.To improve the grasping ability of the prosthetic hand for people with physical disabilities,researchers have proposed the idea of an intelligent prosthetic hand with a haptic sensing function.To address these needs,an intelligent hand prosthesis integrating pressure,slip,temperature and myoelectricity sensing is designed,and a pattern recognition algorithm based on deep learning is introduced to analyze the myoelectricity signals to recognize the operator’s action intention,and then realize the accurate manipulation of the hand prosthesis.The work in this paper is of great value to the physically handicapped people to improve the manipulation ability of the prosthetic hand,avoid the dangerous environment and improve their quality of life.The main research of this paper includes:1.Overall scheme.By analyzing the overall requirements of the haptic intelligent prosthetic hand,the system architecture of the intelligent prosthetic hand is built,and the bionic hand scheme,myoelectric control scheme,haptic perception scheme,and prosthetic hand control scheme are developed based on this architecture.2.Myoelectric pattern recognition algorithm.A myoelectric pattern recognition algorithm is designed based on deep learning,and the core of the algorithm is a two-stream twodimensional convolutional neural network model with ring filling.Based on this algorithm,the performance of the model is optimized by adjusting the key hyperparameters.Finally,it is experimentally demonstrated that the algorithm can achieve 95.3% accuracy of action intention recognition while ensuring real-time performance.3.System hardware.Based on the hardware architecture of the prosthetic hand,a haptic sensory system including temperature and pressure information collection and corresponding visual and vibration feedback devices are designed.Among them,the vibration feedback device is also responsible for the warning of danger information.4.System software.Based on the software architecture of the prosthetic hand,the program of the initialization mode and normal working mode of the prosthetic hand,as well as the software of the three parts of the myoelectric recognition module,the haptic sensory module,and the control module of the prosthetic hand in the normal working mode are designed.5.Experiment and test.According to the function of the intelligent prosthesis,the manipulation experiment of the prosthesis and the temperature,pressure,and slip perception experiment were carried out,and the results showed that the prosthesis has good operation performance and haptic perception ability.In this paper,deep learning models and haptic perception are introduced into the intelligent prosthetic hand system to assist the user in accurately controlling the prosthetic hand and obtaining information about the surrounding environment to meet the needs of people with physical disabilities to use the prosthetic hand.