Comparison of Step Length Estimation Models Using Inertial Sensor on Pelvis

In this paper, step length models using an inertial sensor on pelvis are compared by experiment. Step length estimation is an important process of pedestrian dead reckoning (PDR). However, the step length is difficult to measure directly by an inertial sensor, so it should be derived indirectly by measuring other gait features. Biomechanical and statistical regression models are widely used models, and those are commonly applied to data from pelvis because it is near to the center of mass of the body. In this paper, by comparing step length estimation models from several previous research, the most accurate model is suggested in individual and general case respectively. The models are compared with RMSE calculated based on experimental data.