Indoor Pedestrian Tracking Using Consumer-Grade Inertial Sensors With PZTD Heading Correction

Abstract

Foot-mounted navigation based on inertial navigation sensors struggles with unobservable heading error. Existing solutions for heading error rely on acquiring geomagnetic information or prior knowledge, such as building maps, which are susceptible to magnetic disturbance or lack of maps, respectively. A prominent approach to mitigate the heading error is by applying heuristic drift elimination (HDE), where orthogonal indoor corridors are used as reference for headings. Any heading discrepancy in walking straight along the orthogonal corridors or so-called dominant directions will be attenuated. State of the art emphasizes that HDE should not correct the heading when pedestrians walk along curved paths or straight paths in non-dominant directions; if it is not a case, over-correction will occur and lead to wrong trajectories. Also, the techniques to determine the type of walks heavily rely on thresholding, which is dependent on the person or operating conditions. In this paper, a threshold-less turn detection method known as pelvic rotation-ZUPT turn detector has been applied to correct the heading in the Kalman filter-based foot-mounted navigation. The turning information proves to be a significant aid to the pedestrian tracking even though consumer-grade inertial sensors are used. The pedestrian tracking system achieves return position error and traveled distance error about 2.3% and 2.9% in average, respectively.