Improved GNSS velocity estimation using sensor fusion

This work presents a generic method for improving the velocity estimation of GNSS devices. The suggested algorithm is based on the famous Kalman Filter which utilizes both GNSS and IMU measurements. The suggested sensor-fusion method contains a built-in classifier for identifying measurements with low confidence. The algorithm was implemented and tested using both simulation and real-world data. The results show significant improvement in the velocity estimation. This improvement can be used for further positioning accuracy improvement of a wide range of COTS GNSS devices.