Self-adaptive pedestrian mobile phone posture recognition method based on Gaussian mixture clustering algorithm

The invention discloses a self-adaptive pedestrian mobile phone posture recognition method based on a Gaussian mixture clustering algorithm, and the method comprises the steps: carrying out collectionand quantification of three-axis data of an internal acceleration sensor, a gyroscope and a close-range sensor under each posture of a mobile phone of a plurality of pedestrians, solving a gravity component of the acceleration sensor, and virtualizing the gravity component into gravimeter data; matching the characteristics of each attitude signal by adopting a Gaussian mixture clustering algorithm; setting the size of the sliding window to be M, storing cluster division results of a series of sampling points at the moment, and detecting a series of continuous mobile phone pose conversion in the sliding window; and outputting a posture conversion detection judgment result. The method provided by the invention has the advantages of strong universality, high reliability, simple calculation and good real-time perform