Infrastructure-free indoor pedestrian tracking based on foot mounted UWB/IMU sensor fusion

Abstract

Accurate indoor human localization without requiring any pre-installed infrastructure is essential for many applications, such as search and rescue in fire disaster areas or human social interaction. Ultra-wideband (UWB) is a very promising technology for accurate indoor positioning with pre-installed receivers. An infrastructure-free methodology, called Pedestrian Dead-Reckoning (PDR), which uses an inertial measurement unit (IMU), can also be used for position estimation. In this approach, the drift errors of IMU in each step length estimation are compensated based on zero-velocity update (ZUPT), zero angular rate update (ZARU) and heuristic heading drift reduction (HDR) algorithms. An accurate step detection can be achieved by relying on the data provided by accelerometers and gyroscopes. In order to further improve the accuracy, a novel approach, which combines IMU PDR and UWB ranging measurements by Extended Kalman filter (EKF) without any pre-installed infrastructure, is proposed. All the components in this approach, the IMU, the mobile station (MS) and the receiver of the UWB are mounted on the feet. The biases in the IMU measurements, which cause inaccurate step length estimation, can be compensated by range measurements provided by UWB. The performance of the normal PDR with EKF is evaluated as comparison to the proposed approach. The real test results show that the proposed approach with EKF is the most effective way to reduce the error.