A Stride Detection Algorithm Based on Triaxial Acceleration Characteristics of Pedestrians

Stride detection plays a vital role in the Pedestrian Dead Reckoning (PDR) algorithm, and the accuracy of stride detection will directly affect the accuracy of PDR algorithm. In particular, missed detection and false detection are critical problems in stride detection. For traditional stride detection algorithms, such as peak detection and zero-crossing detection, the thresholds need to be set in advance. However, due to the peak acceleration and stride frequency are not fixed during walking, it is difficult to determine the threshold of peak detection and the zero position of zero-crossing detection. Aiming at the problems existing in the traditional stride detection algorithms, a new stride detection algorithm based on triaxial acceleration characteristics of pedestrians is proposed. The algorithm uses smartphones to obtain acceleration data based on hand posture. After data processing, the algorithm uses peak detection as the first detection condition, and uses dynamic time warping (DTW) to perform similarity matching on the triaxial acceleration of pedestrians as the second detection condition for stride detection. Finally, the stride detection experiments are carried out on experimenters of different heights, and the results show that the algorithm can effectively filter out abnormal strides and improve the accuracy of stride detection.