LAUSR

Due to the disadvantages of traditional posture recognition technologies in cameras and sensors, such as expensive, blind spots and complex deployment, a new method of posture recognition based on WiFi signals is proposed. In this method, channel state information (CSI) is obtained by the wireless network card. Phase calibration is adopted in advantage of the phase difference between different antennas. Besides, both the amplitude and phase information of the channel state are considered in the feature extraction, the improved linear discriminant analysis and Softmax regression algorithm are used to generate the human activity model. Experimental results show that the method has high accuracy and robustness, and the average recognition accuracy of 95.87% can be reached.