An integrated IMU and UWB sensor based indoor positioning system

Abstract

This paper describes an Indoor Positioning System (IPS) that fuses an Ultra-Wideband (UWB) sensor-based positioning solution with an Inertial Measurement Unit (IMU) sensor-based positioning solution to obtain a robust, yet, optimal positioning performance. Sensor fusion is accomplished via an Extended Kalman Filter (EKF) design which simultaneously estimates the IMU sensors' systematic errors and corrects the positioning errors. Fault detection, identification, and isolation are built into the EKF design to prevent the corrupted UWB sensor measurement data due to obstructions, multi-path and other interferences from degrading the positioning performance. General formulation of an IPS using IMU for both pure inertial and Kalman filter aided modes of operation using UWB sensor data is given in the paper for tracking a six degree-of-freedom (DOF) platform motion. The derivations of an 8-state EKF design are detailed in the paper for a 3 DOF (two translational and one angular motions) platform planar motion, where data from a 9-axis Motion Tracking device, MPU-9250, and four UWB radio sensor devices, DWM1000, are used. A Matlab-based simulation model is developed and built to assess the proposed IPS performance along with their performance sensitivities to platform motion profiles, UWB/inertial sensor errors, and filter update rates. With specific motion profiles, computer simulation results indicate that more than 100% positioning performance improvement over the UWB sensor-based positioning solution along can be obtained through the proposed sensor fusion solution. A laboratory test bed for a 3 DOF motion platform is designed, built, and tested to validate the proposed IPS performance. The IPS performance obtained from actual laboratory tests correlated very well with the simulation results.