Pedestrian Inertial Positioning Method Based on Foot Quasi-Zero Velocity Observation Under Multiple Motion Modes

Abstract

As the traditional zero velocity update (ZUPT) algorithm based on fixed threshold is difficult to solve the problem of effective positioning for pedestrians under multiple motion modes, a simpler, flexible and stable inertial positioning algorithm is proposed in this paper. A foot kinematic model based on rotating “L" shaped rigid body is proposed to describe the process from heel landing to heel off the ground. The quasi-zero velocity interval (QZVI) of foot motion is defined, and a method to determine the QZVI based on peak detection and gait cycle constraint is proposed to avoid the dependence of system on predetermined threshold. The proposed foot kinematic model is used to calculate the “near zero" velocity in the QZVI. Then difference between the calculated “near zero" velocity and the inertial navigation solution velocity is used as the observation of Kalman filter to estimate the position error, velocity error and attitude error of inertial navigation system (INS). Experiments show that the proposed algorithm has positioning error relative to the mileage of less than 2% under the mileage of 500m or less, effectively adapting to different motion modes such as fast walking, slow walking and running, so as to realize the accurate autonomous positioning under the condition of multiple motion modes.