Abstract
Ultra-wideband (UWB) and inertial measurement unit (IMU) fusion is an efficient method to resolve the uncertainties of UWB in non-line-of-sight (NLOS) situations because of signals refraction, the effect of multipath and inertial positioning error accumulation in indoor environments. Existing systems, however, are focused only on foot-mounted IMUs that restrict the system's implementation to particular real situations. In this research, using foot-mounted IMU, we suggest combining UWB ranging and IMU pedestrian dead reckoning (PDR), which can provide a generic indoor positioning solution. The issues such as position and orientation drift, interferences and divergence in strap-down inertial navigation system (SINS) based orientation estimates could be addressed by a UWB ranging sensor fusing with an IMU using the extended Kalman filter (EKF). The main goal of this research is to investigate and compare two different sensor data fusion techniques. For instance, adaptive Kalman filter (AKF) and least-squares (LSs) incorporate a foot-mounted IMU tightly coupled to a 2D pedestrian positioning solution derived from UWB signals. Moreover, we consider the UWB NLOS and IMU error identification. A real-time ranging error compensation model based on the LS method and AKF positioning algorithm are used for fixing such problems. We propose a new tightly coupled inertial navigation system (INS) with a two-way ranging (TWR) fusion positioning algorithm to improve accuracy, integrating UWB and IMU sensors based on the EKF in pedestrian navigation. Experiments in dynamic indoor environment validate the effectiveness of the proposed approach that uses EKF to combine AKF and LS for error minimization.