Comparative analysis of pedestrian stride length estimation methods

Abstract

The stride length estimation is a crucial step used in indoor and outdoor pedestrian navigation. The accuracy of the navigation process depends on how accurately the average stride length is calculated. The stride length estimation methods use some parameters such as step frequency, acceleration, and pedestrian height. These parameters are applied to different datasets containing various movements of pedestrians, such as running and walking. However, the limited datasets used in academic studies preclude scientific comparability in the literature. This study compares ten stride length estimation methods using the open-source datasets created by Wang and analyzes their accuracy. The case results show that Weinberg's approach was quite successful for navigation dynamics.