A Study On Student Learning State Recognition Based On Electrical Skin Signals

In recent years,the use of physiological signals to identify students’ learning status has potential advantages and application prospects in the field of education.Through reasonable analysis of students’ physiological signals(e.g.,EEG signals,electrodermal signals,heart rate,EMG,etc.),teachers can obtain students’ learning and cognitive status in a timely manner.However,the acquisition of physiological signals often requires specialized sensor equipment,which limits its development in large-scale applications.Compared with traditional electronic devices,wearable devices have better human-computer interaction functions,and have the advantages of convenient and easy to carry and simple operation.The use of wearable devices to collect and analyze physiological signals has broad development space.Based on the above advantages,this paper proposes a research method for student learning state recognition based on electrical skin signals,and develops a student learning state recognition system based on MATLAB GUI.Firstly,a portable electrodermal acquisition system was designed based on STM32 microcomputer,and the system was used to collect the electrodermal data of students in different states.After acquisition,the data is denoised by a Butterworth low-pass filter.Subsequently,the median filter is used to extract the galvanic skin response component in the electrodermal signal for feature extraction.Secondly,a total of 35 features were extracted in the time domain,frequency domain and time frequency domain for classification training.The discrete binary particle swarm optimization algorithm is used to optimize the feature combination,and the 18 optimal features obtained are used to build a learning state recognition model.By analyzing the advantages and disadvantages of the three classifiers deployed in embedded hardware systems,the model deployment in embedded hardware devices is realized.Thirdly,a student learning status recognition system is designed based on the MATLAB GUI.The system can update the prediction results of the embedded device on the student’s learning state in real time,and realize the visualization of the prediction results.Finally,the collected data is identified and predicted in MATLAB,and the classifycation accuracy rate exceeds 80%.In addition,the one-to-one method of SVM realizes triclassification prediction in embedded hardware devices,and the final recognition accuracy is more than 60%.This result verifies that the learning status of students can be identified using embedded systems.