Gesture And Arm Motion Recognition Based On Myo

Surface Electromyography(s EMG)recognition is one of the more mature technologies in the field of human intention recognition.Through the collection and processing of EMG signals,it can identify the movement category of human limbs.It is widely used in clinical medicine and rehabilitation engineering.And the control field of mechanical prosthetic hands.By processing the inertial measurement unit(IMU)information worn by the limb,the intention of the human body movement can be obtained and the movement of the limb can be recognized.Compared with traditional EMG acquisition equipment and inertial measurement unit equipment,the Myo armband combines the functions of the two.It can also collect the movement posture information of the arm when collecting EMG signals,and is convenient to wear and use.Therefore,the purpose of this article is based on the surface EMG signal and inertial measurement unit information collected by the Myo armband.Through the pattern classification method of machine learning,a series of processing of the s EMG signal is carried out to obtain the gesture action information;by analyzing the joint structure of the human arm To model,use the IMU signal of the arm to calculate the motion angle of the arm joint,and combine with Matlab simulation to verify the feasibility.The main research contents of this paper are as follows:(1)Classification and recognition of gesture action patterns based on s EMG.Use IIR digital filter and fourth-order wavelet denoising method to preprocess the original EMG signal,extract 8 time domain features and 3 frequency domain features,and use LDA,DT,SVM,KNN,RF and BP neural network to pair Ninapro DB3 data set is trained to verify the accuracy of algorithm recognition in the EMG recognition scene.Compared with the experimental results,the SVM algorithm is selected as the classification algorithm for the manipulator grasping experiment.(2)Build a robotic grasping experiment platform.A robotic grasping experiment platform based on EMG signal classification was built,and a visual-assisted grasping strategy was added to the traditional EMG signal recognition and grasping.Experiments have proved that,compared with traditional EMG signal capture,adding a visual aid strategy to the robotic grasp can improve the accuracy of the grasp.(3)Research on arm posture calculation based on IMU.By agreeing on the coordinate system of the inertial measurement unit and the arm wearing mode in Myo,the robot arm model is established based on the joint freedom of the real arm,and the method for calculating the rotation angle of the arm joint is proposed.The experiment selects five arm motion postures for calculation,and The calculated joint angle is used as the Matlab simulation input,which verifies the feasibility of the method for calculating the arm posture.The study of the arm's electromyographic signal and IMU signal can obtain the movement intention of the human arm,expand production scenarios in the industrial field,and improve production efficiency;in life,it can bring more convenient life and services to humans,and target physical disabilities It is especially important for the patient to return to a normal life.