Analysis on the Parameter Estimation Performance in an UWB/IMU Tightly-Coupled System

Abstract

Combining ultra-wideband (UWB) and low-cost inertial measurement unit (IMU) by a tight coupling fashion is an effective way to provide an integrated indoor localization system with high accuracy and continuity. However, the performance of such an integrated system may be poor in the presence of undesirable geometry between an UWB target node and a number of UWB anchors. In addition, some parameters in the state vector of Kalman filter (used for data fusion) are unobservable from UWB ranges. These problems have not been discussed in detail in the literature related to UWB localization. In this paper, we analyze the performance of integrated localization under different geometries between the UWB node and anchors based on Fisher information matrix. Also, the parameter observability with respect to UWB ranges is analyzed.