Sensor fusion of INS, odometer and GPS for robot localization

Abstract

This paper presents data fusion of three sensors Inertial Navigation System (INS), Global positioning systems (GPS) and odometer for determining the correct location of a differential drive mobile robot. The data from INS and odometer is combined using Kalman Filter (KF) based sensor fusion technique. The KF filtered signal and the GPS signal is fused by assigning weights. The proposed technique is tested in simulation. Mathematical models of the three sensors as well as the robot model is developed in MATLAB/Simulink environment to generate the data for simulation purpose. It has been demonstrated that with the proposed sensor fusion architecture exact geo-location of a differential drive wheeled robot can be determined to a greater degree of accuracy in an indoor or outdoor environment.