In situ heading drift correction for human position tracking using foot-mounted inertial/magnetic sensors

Abstract

-This paper presents a heading drift correction method and experimental results for position tracking of human movement based on the use of foot-mounted inertial/magnetic sensor modules.A position tracking algorithm was previously developed, which incorporated a zero velocity update technique for correcting accelerometer drift. Previous experiments indicated the presence of a persistent heading drift in the estimated position. In this paper, a simple method for correcting this drift is presented. The method requires the user to walk over a closed loop path with the footmounted sensor module. Assuming a constant sensor bias for this initial walk, the resulting position error is then used to accomplish an in situ correction for position estimates during future walks. Experimental results validate the effectiveness of the drift correction method and show a significant improvement in position tracking accuracy. Accuracy is determined based on the final position estimates following walks of 100 and 400 meters. Estimated distance traveled averages within 0.2% of actual distance traveled and distance from the actual position averages within 0.28% of actual distance traveled.

Estimation of Walking Direction Estimation using a Shoe-mounted Acceleration Sensor

“…Estimation of walking direction has been attracting interest in the field of context recognition [1][2][3][4], dead-reckoning navigation [5][6][7], and biomechanics [8][9][10][11]. In the field of context recognition, especially in cases where a mobile phone is used, an acceleration signal is acquired from the on-body mobile phone and processed through principle component analysis (PCA) for estimating the user's walking direction [12].…”

Virtual reality in the wild: A self-contained and wearable simulation system

“…Each participant attached the SCIPT sensors to their shoes and donned the backpack rendering system. A quick calibration routine was run to adjust the inertial tracking algorithm to that user to reduce angular drift, which required the user to walk in a circuitous path of their choosing and return to the starting point [2].…”

“…IEEE Virtual Reality 2012 4-8 March, Orange County, CA, USA 978-1-4673-1246-2/12/$31.00 ©2012 IEEE precision and dynamic accuracy within 1%[2]…”

A Pedestrian Navigation System by Low-cost Dual Foot-Mounted IMUs and Inter-foot Ranging

“…A number of methods have been proposed to solve this problem. These methods can be divided into the constraintbased methods [13][14][15][16][17][18][19] and extra sensor-assisted methods [20][21][22][23]. The works presented in [13,16] attempt to correct the yaw gyro bias by assuming a constant heading if a person walks on a straight line.…”

“…The works presented in [13,16] attempt to correct the yaw gyro bias by assuming a constant heading if a person walks on a straight line. When the users walk back to the starting position, a zero position update is used in [15]. Some authors utilize an inequality constraint in the Kalman filter in view of the fact that there exits an upper bound on the maximum spatial separation between two feet in a system of dual foot-mounted IMUs [18,24].…”