Multi-Condition Constraint Adaptive Step Detection Method Based on the Characteristics of Gait

摘要

The Pedestrian Dead Reckoning (PDR) system plays an increasingly important role in the field of seamless localization of indoor and outdoor space. However, the accuracy of PDR system is affected by many factors. Step, stride and heading are three main factors. Therefore, improving the step, stride and heading measurement precision can enhance performance of the PDR system In this paper, we propose an original multi-condition adaptive step detection algorithm This algorithm based on gait characteristics analysis, then using the periodicity of the gait to realize gait detection. In this way we can solve the problem of step detection inaccuracy. The proposed algorithm is verified with the real experiment. Also, the experiment results indicate the proposed step detection algorithms can meet the 99% detection accuracy better than the traditional algorithm.