Threshold-less Zero-Velocity Detection Algorithm for Pedestrian Dead Reckoning

摘要

GPS (Global Positioning System)/GNSS (Global Navigation Satellite System) and wireless positioning systems, which provide reliable location information only in a fully constructed infrastructure environment, are not suitable navigation methods for multimodal pedestrians or those who are moving through gait where the infrastructure is not built. To solve this problem, research on PDR (Pedestrian Dead Reckoning) using inertial sensors has been going on for the last two decades. Recently, a method of calculating the position through the INS algorithm using the inertial sensors mounted on the shoe and correcting the error through ZUPT (Zero-velocity UPdaTe) algorithm by detecting the zero-velocity at the moment when the foot is stepped on the ground is widely used. Most of the zero-velocity detection algorithms use threshold values after manipulating the sensor outputs. However, these threshold values should be set differently depending on the pedestrian, the type of walking, etc. In this paper, we propose a zero-velocity detection algorithm that does not require any threshold value. After processing the gyro and accelerometer outputs based on an appropriate algorithm, it detects the moment when the foot touches the ground without any threshold value. The performance of the proposed algorithm is verified experimentally.