A New Model for Precise Point Positioning to Improve Fault Detection

Abstract

The availability of real-time corrections to broadcast satellite orbits and clock offset enables implementation of real-time Precise Point Positioning (RT-PPP) in important applications such as natural hazard warning systems and intelligent transport systems. However, current RT-PPP models combine the observations with the orbit and clock corrections in one term. Accordingly, faults in these corrections, for instance due to spoofing, will result in exclusion of their related satellite measurements, which would lead to degradation of the positioning quality, with a total disabling of RT-PPP if errors are introduced in all corrections. In this contribution, a new PPP model that treats the corrections to broadcast orbit and clock offset as quasi-observations is presented. This model enables fault detection and exclusion of these corrections separate from the observations. The excluded faulty corrections can be replaced by predicted values, using for instance the predicted IGS ultra-rapid orbits, and a linear polynomial with sinusoid terms for the clock corrections. Thus, the method preserves positioning by keeping the measurements that have faulty corrections, and using them along with the predicted corrections. The proposed method is validated at three IGS stations, where its results was compared to results of the traditional PPP methodology. Artificial faults were inserted at random events and the test was repeated with a varying number of faults. Results show that using the proposed method, positioning was maintained during the faulted periods, whereas the traditional PPP accuracy degraded sharply with the increase of number of faults.