Synergistic Characterization Of Human Body Based On Rhythmic Motion Generation

In nature,human beings have excellent cooperative movement ability,and can adjust cooperative strategy in time to adapt to the target of sports task according to the actual situation.However,due to the influence of the complex physiological structure and exercise habits of human body,the characteristics of movement coordination are personalized and differentiated,which makes the extraction and characterization of the characteristics of human rhythmic movement coordination become one of the most challenging subjects in the study of human movement.Therefore,how to effectively extract the synergistic characteristics of human rhythmic movement and accurately depict the coupling and synergistic law between various joints in synergistic movement has become an urgent problem to be solved.In view of the difficulty in extracting and describing the collaborative features of human rhythmic movement,this paper proposes a method of generating rhythmic movement trajectory based on conditional generation antagonistic network to generate the rhythmic movement of human joints in an orderly manner,and introduces an adaptive Hopf oscillator based on Hebbian theory to realize the explicit expression of the implicit features of joint synergy.The main work of this paper is as follows:Firstly,a simplified kinematics model and a rhythmic movement database should be established based on the structure of human body segments and the distribution of mass and center of mass.Then,according to the rhythmic characteristics of walking and rope skipping and the characteristics of joint rotation,the different stages of rhythmic movement cycle should be divided,which lays a foundation for the research of joint cooperative phase of human rhythmic movement in the next step.Secondly,based on the periodic characteristics of human rhythmic movement,the analysis matrix of rhythmic movement should be constructed.In addition,singular value decomposition(SVD)and principal component analysis(PCA)were used to extract and analyze the joint numerical characteristics of different rhythmic motion matrices of different subjects.Then,by setting the joint phase reference point,the cooperative phase calculation criterion describing the joint coupling rotation timing sequence should be established,and the characteristics of convergence and difference coexistence of cooperative features among different individuals with different rhythmic movements should be analyzed and verified.Then,based on the idea of generating against,the rhythmic motion of joints data as sample data generated by the study,based on conditions generated against network method to generate the rhythm of movement,by setting the data labels,rhythmic movement features different individuals into multi-layer neural network weighting parameters,rhythmic movement data accurately generate reiteration.In addition,the hidden space dimension reduction method is used to realize the visualization of joint cooperative features in network parameters to verify the effectiveness of the proposed method.Finally,in order to avoid generating against neural network problem of opaque black box model,further clarify the physical meaning of parameters in the model,realize the explicit expression of movement coordination body rhythm characteristics,this article introduced based on dynamic adaptive CPG Hebbian theory,the establishment and the corresponding human body joint unit parameter identification model,through the joint trajectory of adaptive identification process,realize the coordinated characteristics of joints rhythm of parametric representation.On this basis,combined with the joint cooperative phase calculation criteria previously proposed,the coupling rotational timing between the joints was integrated into the CPG phase coupling network,and the reconstructed data and real data were analyzed and compared through kinematics simulation experiment to verify the effectiveness and accuracy of the proposed method.