Research on multi-sensor pedestrian dead reckoning method with UKF algorithm

In order to solve the problem that the positioning error of pedestrian dead reckoning (PDR) with inertial measurement unit will accumulate over time, an indoor pedestrian space estimation method based on multi-sensor fusion is proposed. Based on the low-cost multi-sensor of smart mobile device, the initial alignment based on the unscented Kalman filter (UKF) is designed and four threshold conditions are set for carrying out the step state detection. In the process of walking, aiming at solving the cumulative problem of the step error and the heading angle error, the method of combining zero velocity update (ZUPT), zero angle rate update (ZARU), and magnetometer based on UKF is used to correct the error, where ZUPT corrects the speed error, the combination of ZARU and magnetometer corrects the gyro error and the heading angle error, which effectively reduce the position error and improve the pedestrian positioning accuracy. The experimental results show that the method proposed in this paper can effectively reduce the average position deviation of PDR, and the positioning error is about 1.5% of the total walking distance. Compared with the original PDR system algorithms, the proposed method uses low-cost multi-sensor to obtain better positioning accuracy.

Automated summary

The proposed method utilizes multi-sensor fusion, correction with UKF, and calibration methods such as ZUPT and ZARU to effectively reduce position deviation and improve positioning accuracy of PDR.