In this paper, we propose a motion recognition based step length estimation algorithm using a smart phone. In smart phone usage environment, there are a lot of motions and output patterns of sensors are differentiated according to each motion. In order to reduce the effect of motion on the accuracy of step length estimation, we identified the motion of pedestrian using the sensory functions of a smart phone. By adapting the parameters on step length estimation algorithm, proper process of step length estimation for each motion is conducted. The gait characteristic of pedestrian is analyzed using the map data based in-flight calibration process for accurate step length estimation. In order to verify the proposed step length estimation algorithm, we performed experiments in indoor corridors to demonstrate its improved accuracy.

Keywords: motion recognition, step length estimation, map data, sensor calibration