A new historical data based GNSS fault detection and exclusion algorithm for UAV positioning

Recently, unmanned aerial vehicle (UAV) development is increasing dramatically. To reach the full potential of UAV, the requirement of Global Navigation Satellite Systems (GNSS) performance is increasing for UAV applications. However, the accuracy and reliability of GNSS are always affected by interference such as multipath effects in some special environments, which may deteriorate the safety of general and civil aviation operations. It is therefore essential to enable an effectively GNSS fault detection scheme to make sure the high accuracy of UAV positioning calculation. In this paper, we have proposed a new historical data based GNSS fault detection and exclusion algorithm for UAV positioning. The proposed algorithm enables a real-time online dataset based fault detection and exclusion scheme, which allows accurate detection of simultaneous multiple faults of different satellites for real-time GNSS measurement quality control in dynamic positioning applications. The UAV flight tests with fault satellites of various pseudorange errors have demonstrated the effectiveness of the proposed algorithm with a fault detection rate of 100% for the multi-satellite faults scenario. With fault detection and exclusion, the horizontal and 3D positioning accuracy RMSE is greatly improved to 1. 01m and 6. 73m, respectively, while the accuracy could be tens of meters (horizontal) and more than 100 meters (3D) before fault satellite exclusion.