Improving the Position Accuracy of the ZUPT-Aided Pedestrian Inertial Navigation by Using a Differential Layout MIMU Array

The pedestrian inertial navigation system (PINS) based on a miniature inertial measurement unit (MIMU) and zero-velocity update (ZUPT) can achieve pedestrian positioning without any external equipment. However, the ZUPT algorithm cannot correct the drift error of MIMU due to its low observability of heading angle, leading to the divergence of pedestrian tracks. Using multiple MIMUs can improve the measurement accuracy of inertial sensors. Therefore, a differential layout MIMU array composed of multiple low-cost MIMUs is designed for the foot-mounted PINS in this article. The system consists of two-stage data fusion. In the first stage, multiple sensors are arranged in differential structures on the axes of the carrier, and a Kalman filter (KF) is presented to fuse measurements of multiple MIMUs. In the second stage, a ZUPT KF is designed to correct attitude and location errors. Finally, the device-level test results showed that the noise performance of the differential MIMU array was reduced by 82.5%. The system-level test results showed that the positioning error of pedestrian inertial navigation based on the MIMU array accounted for about 0.36% of the total mileage of 1.4 km.