Robust Kalman Filter applied to GNSS positioning in harsh environment

GNSS-based navigation, nowadays the most spread system, is very challenging in scenario where the presence of obstacles could block, diffract or reflect the signals coming from satellites. In this harsh environment the navigation algorithm has to face out with low redundancy of measurements or the presence of blunders among them. A possible approach to overcome this problem is the use of robust estimation that is applied, in this study, to single point positioning, implemented with an Extended Kalman Filter (EKF). In the update phase of the filter, the measurement covariance matrix is computed considering the robust Huber method. The satisfactory performance of the proposed method are investigated processing real data, collected by a low cost GNSS receiver in an urban scenario.