GPS ephemeris error screening and results for 2006-2009

For the Global Positioning System (GPS), real-time satellite orbits and clock biases are derived from predicted ephemeris and clock parameters in broadcast navigation messages. The performance of broadcast ephemerides is critical to billions of GPS users in terms of position accuracy and integrity. A typical way to evaluate ephemeris errors is comparing broadcast ephemerides with precise ones. At times, broadcast ephemerides data obtained from a tracking network include errors caused by receivers. Besides, the receivers at different locations may not receive the same broadcast ephemeris message as a satellite rises and sets. In this paper, a powerful systematic screening methodology is presented to cope with all above problems and the screening results for year 2006-2009 is provided. The broadcast ephemerides are retrieved from all active International GNSS Service (IGS) stations. The following types of data defects are observed: losses, duplications, inconsistencies, discrepancies, and errors. A data purification algorithm based on error-correction and majority-vote is devised and implemented to remove all erroneous ephemerides and to generate validated daily global combined broadcast ephemerides. The validated broadcast ephemerides are employed to propagate broadcast satellite positions and clocks at 15-minute intervals that coincide with the precise ephemerides from National Geospatial-Intelligence Agency. Then an analytic method is utilized to calculate the worst-case signal-in-space range error (SISRE). Finally, ephemeris anomalies are identified by comparing the worst-case SISRE with the signal-in-space not-to-exceed tolerance, 4.42 times of the user range accuracy upper bound. All GPS ephemerides from 2006 to 2009 are screened, and all potential anomalies and Issue of Data, Clock (IODC) reuse problems are documented. In comparison with the daily global combined broadcast ephemerides provided by IGS, our validated ephemerides include far fewer errors and greatly reduce the number of false anomalies.