Research On Control And Evaluation System Of Intelligent Movement Of Massage Chair

At present the massage chair on the market realizes massage manipulation through the mechanism.And the novel massage chair can contain three independent degrees of freedom,theoretically can achieve the space of arbitrary massage manipulation.In addition,the current method of evaluating the effect of massage chair mainly relies on the subjective evaluation of a few users,lacking objectivity.Electroencephalogram(EEG)signals belong to objective physiological signals of human body and contain abundant human body information,including fatigue state information of human body.Relieving fatigue is also a major function of the massage chair,so it is feasible to evaluate the massage effect with EEG signal as the signal source and fatigue as the evaluation index.Firstly,the control system of novel massage chair is built.According to the characteristics,a special massage manipulation trajectory control algorithm is designed,and the optimized curve is interpolated by the quantic polynomial to make the massage manipulation smoother.For the pressure control,a fuzzy PD controller is designed to control the pressure feedback,and the fuzzy rule base is established through experiments.Secondly,band-pass filtering is used to remove power frequency interference and amplifier drift noise in EEG signals.Hilbert-Huang Transform(HHT)and Independent Component Analysis(ICA)are used to remove Electrocardio(ECG)and Electro-oculogram(EOG)in EEG signals,and the effects of the two methods are compared.Then wavelet packet decomposition is used to extract the energy characteristics of EEG signals for subsequent models.Then three basic models Light Gradient Boosting Machine(Light GBM),Support Vector Machine(SVM)and Neural Network(NN)are selected to classify the EEG signals and judge the resting state and fatigue state of the human body,taking this as the main body of the evaluation system.Logistic regression model is selected as the second layer model,and the results of the three models are integrated with the two-layer Stacking framework to further improve the overall effect of the model.Finally,fatigue experiment is designed to obtain resting and fatigue EEG signals to verify the effectiveness of this method.T test is used to preliminarily test the validity of features and optimize the combination of features.The final experimental results show that it is feasible and effective to use EEG signals to judge whether the body is in a state of fatigue so as to evaluate the massage effect.