Research on a fusion gait real-time recognition algorithm

Abstract

Based on the current market prospects of wearable devices, gait recognition accuracy and real-time market demand, a fusion gait real-time recognition algorithm is designed. The paper introduces the method of limiting the original step signal by using the method of limiting filtering and moving smoothing filtering. At the same time, using dynamic threshold detection and similarity algorithm fusion method, using dynamic window method, the step data is sequentially subjected to dynamic threshold. Only if the two algorithms are simultaneously established, the decision step is established. The fusion gait recognition algorithm is designed to achieve real-time gait recognition. and finally obtain more accurate detection results.