利用少量体压传感器和支持向量机算法的坐姿识别方法

针对传统坐姿识别系统中传感器数量多和系统较复杂导致成本过高等问题,设计一种基于少量体压传感器和支持向量机(SVM)算法的坐姿识别方法.首先,设计一种由少量薄膜压力传感器构成的体压传感阵列,将其置于坐垫内部;然后,利用该传感阵列采集不同坐姿的体压数据,并绘制相应的体压分布等高线图;最后,以体压数据作为特征向量,结合支持向量机算法建模,以实现坐姿分类自动识别.测试结果表明:少量体压传感器也能获取不同坐姿的体压分布特征;SVM坐姿分类模型在熟悉样本下的坐姿识别准确率达98.3%,在陌生样本下的坐姿识别准确率达92.5%.

Sitting Posture Recognition Method Using Small Number of Pressure Sensors and Support Vector Machine Algorithm

In view of the high cost of the sitting posture recognition system due to the large number of sensors and complex system, a sitting posture recognition method based on a small number of body pressure sensors and support vector machine(SVM)algorithm is designed. Firstly, a body pressure sensor array based on a small number of thin film pressure sensors is designed, which is placed in the cushion. Then, the sensor array is used to collect the body pressure data of different sitting posture, and the corresponding contour maps of body pressure distribution are drawn. Finally, taking the body pressure data as the feature vector, the support vector machine model is combined to realize the automatic recognition of sitting posture. The test results show that a small number of body pressure sensors can also obtain the characteristics of body pressure distribution in different sitting positions. The accuracy of sitting posture recognition of SVM sitting posture classification model is 98.3% in familiar samples and 92.5% in unfamiliar samples.