The Research Of Gesture Recognition Based On Surface Electromyography Signal And Inertial Measurement Unit

Gesture recognition is currently a hot topic in the field of artificial intelligence(AI)technology,which potentially transform all aspects of human life and society,including but not limited to games,medical care,education,shopping and so on.Gesture recognition technology involves a wide range of disciplines,such as pattern recognition,deep learning,mathematical analysis.It translates hand shape and posture to the digital form representation,allowing the computer to understand the meaning of gestures.In recent years,great progress has been made in gesture recognition technology as an unique branch of applied science.However,to achieve real-time responsiveness such that the technology can benefit various aspects of our daily life remains to be very challenging so far.There are two reasons account for this phenomenon.Firstly,due to the complexity and speed of the presently available gesture recognition algorithm,it is difficult to achieve the purpose of rapid identification.Secondly,for the specific gesture recognition algorithm,the number of recognized gestures is limited.Even the gesture recognition techniques based on multi-sensor fusionstill faces challenges in the lack of adaptivity and flexibility to the diversity of human gestures.Therefore,increasing the computational speed of the algorithm and increasing the recognizable types of gestures become the focus of experimental and theoretical research activities.The research of this paper focuses on the algorithm of gesture recognition.Taking myoelectricity,or electromyography(EMG)information and inertial measurement unit information collected by MYO sensor(Gesture control arm band)as the main target,and the existing gesture recognition algorithms are improved.The gesture recognition based on EMG signals and acceleration signals are studied respectively.The main work is as follows:1.This paper introduces the overseas and domestic research status and related progress of gesture recognition technology in detail,and provides a theoretical basis for the study of gesture recognition.It also introduces the basic technique of gesture recognition,and a new HMM(Hidden Markov Model)method is proposed for experimental research objects.2.The hand gesture recognition based on complex EMG signals was researched and a special treatment method is used..Firstly,the raw EMG data is segmented by the overlapped sliding averaged energy to detect active actions.This method extracts the main part of the gesture phase.Then,the feature data of segmentation is extracted by using mean absolute value(MAV).Finally,the EMG signal of eight dimensions is fused and the gesture recognition is obtained by matching the test samples with the template in DTW method.The templatemaking is devised by finding the warping paths.The recognition accuracy of the EMG signals for the hand gestures can reach as high as 96.09%.The experiment verifies the feasibility of dynamic time warping algorithm to identify EMG signals.3.The dynamic gesture recognition of the fusion myoelectric signal and inertial measurement unit is studied,this research proposes a collaborative sparse representation classifier to recognize the ACC pose signal for gesture recognition.Based on acceleration,the optimal number of samples and the dimensions of the dimensionality reduction were studied to reduce the complexity of gesture recognition.It is demonstrated that the ACC signal for four postures can achieve 96.88% and the method has fast calculation speed.The recognition accuracy for the 12 dynamic gestures can reach 96.11%.4.To simplify the representation of gesture recognition results,the user interface of gesture recognition based on myoelectric signal and inertial measurement unit is designed and coded using MATLAB software.It is demonstrated that the speed and recognition accuracy of dynamic gesture recognition have been greatly improved by researching and improving the dynamic gesture recognition algorithm.The effectiveness and superiority of the method proposed in this study have been verified by experimental results.