Research On Distribution Of Pressure Field In Contact Surface Of Seat And Recognition Of Sitting Posture

Seat has become an indispensable tool for people's daily life. With the increase of people's working hours, the time for people to keep the sitting position is more and more long. Therefore, for the study of the classification of sitting posture has become the focus of people's attention. In this paper, based on the sensor technology to collect the sitting position information, the method of data processing is used to recognize the sitting position.This paper introduces the measurement of the pressure in the contact surface of the human body, and describes the layout of the sensors. Data are stored so that can be applied to analyze. On the basis of this, the information can be analyzed in time domain and frequency domain. The time domain analysis mainly includes the mean, variance and zero crossing of the data which is used to compare the pressure and the the sitting position information. The time domain features are extracted to distinguish different sitting postures through the experiments. But sometimes the time domain features of different sitting postures still overlap. Therefore, on the basis of the time domain analysis, frequency domain analysis is performed on the pressure of the contact surface. The main method use the FFT to process the pressure of the contact surface for distinguishing the difference of sitting postures.The time domain analysis and frequency domain analysis can distinguish sitting posture. But there is still large error in the recognition rate. In this paper,we use the SVM to train the experimental samples, and extract the time domain characteristics and frequency domain characteristics to distinguish different sitting postures to improve the recognition rate and reduce the harm caused by the error identification. Through the data acquisition of different experimenters,the corresponding time domain characteristics and frequency domain characteristics are extracted to train the sample of the support vector machine.Using the time domain and frequency domain features of the samples train the classifier, and compare the results of different training samples. Analyze the different reasons of the classification in results of different training samples.Finally, experiments are conducted to verify the correctness of the classification results.