Development of a European Ionosphere Threat Model in support of GBAS Deployment
Emilien Robert, Petr Jonas, EUROCONTROL, Belgium; Jonathan Vuillaume, Daniel Salos, Egis Avia; Philippe Yaya, Louis Hecker, CLS
GBAS is a Ground Based Augmentation System for GNSS that provides precision approach operations to aircraft. Besides the GNSS space segment the GBAS system consists of a GBAS ground subsystem and a GBAS aircraft subsystem. The GBAS ground system precisely determines its position using the GNSS satellites signal and calculates pseudo range corrections for these satellites. These pseudo range corrections are broadcasted to the aircraft via a VHF link. Together with integrity parameters, these pseudo-corrections allow the aircraft to compute a position with the required accuracy and integrity to support precision approach operation.
One of the main source of ranging error is the propagation delay induces by the ionosphere. This propagation delay is taken into account in the pseudo range correction broadcasted by the GBAS ground system and is fully corrected under nominal ionosphere condition. However, under anomalous ionosphere condition, the ionosphere can have a significant spatial variation and the aircraft can experience a different ionosphere than the GBAS ground system. The pseudo range corrections broadcasted by the ground station become inappropriate and the aircraft will experience an unusual position error. To mitigate this risk, local ionosphere threat model that bounds the maximum ionosphere spatial variation needs to be developed. Anomalous ionosphere structure are modelled as a travelling ionosphere front characterized by its gradient, speed, direction and width.
Based on analysis of historic GNSS data recorded over most of the previous solar cycle, a threat model covering the contiguous United State that mitigate anomalous ionosphere has been developed by Stanford University. The methodology was based on the use of pairs of GNSS ground receiver, one receiver representing the GBAS ground subsystem, the other one the GBAS aircraft subsystem. Using dual frequency measurement, ionosphere delays between each ground station and a GPS satellite have been evaluated. Then, the difference of ionosphere delay between the stations divided by the geographical distance provided the ionosphere spatial variation. In order to limit the computation needs, days of interest have been pre-selected using the Kp and Dst geomagnetic indexes. The largest ionosphere gradient was measured on November, 20th, 2003 and reached 425 mm/Km. A similar study performed by the DLR Institute of Communications and Navigation analyzed eleven years of data from 1998 to 2008 over Germany. The observed ionosphere fronts over Germany were significantly lower than the ones measured over the CONUS area and the study proposed a threat model with a maximum ionosphere gradient of 150 mm/Km.
EUROCONTROL launch a project in 2012 to develop a European ionosphere threat model in support of current and future European GBAS implementations.
First, the appropriate GNSS data have been selected and retrieved. Based on existing networks of ground receivers, several clusters of GNSS ground receivers have been selected. This selection process was a trade-off between a reasonable numbers of ground stations on one side, and the need to covers the European latitude on the other side. In addition, within a cluster, the smallest distance between ground station and the largest station density were targeted. As a result, 14 clusters were selected, located in the Canary Island (Spain, 28°N), Madrid (Spain, 40°N), Corsica (France, 42°N), Toulouse (France, 43°N), Friuli region (Italy, 46°N), Paris (France, 48°N), Prague (Czech Republic, 50°N), London (United Kingdom, 51°N), Hamburg (Germany, 53°N), Malmö (Sweden, 55°N), Göteborg (Sweden, 57°N), Stockholm (Sweden, 59°N), Lulea (Sweden, 65°N), and Gällivare (Sweden, 67°N). Data were retrieved from a total of 215 ground receivers and from October 8th, 2012 to March, 31th, 2016.
Second, a dedicated MATLAB-based tool initially developed by Stanford to define the CONUS threat model and provided to EUROCONTROL by the FAA has been used to process the GNSS data. This tool has been adapted and complemented with additional functions to match the European specificities. The GNSS raw data processing has been divided into different steps. The first step processes GNSS raw measurement stored in RINEX files and performs various corrections to compute “true” ionosphere delay per line of sight. In order to improve the raw measurement processing, an innovative technique based on a statistical analysis has been developed to correct for the cycle slips. The second step estimates the ionosphere gradient for every pairs of ground stations within a cluster and performs various consistency checks to detect and select the potential relevant ionosphere gradient. For each potential gradient detected during the second step, the third step computes the additional characteristic: the ionosphere front speed, direction and width. The last step is the manual validation of the processing. Although various and efficient correction and consistency checks are applied to the raw measurement, the processing needs to be manually validated to separate the gradients coming from measurement artefact (“fake” gradients) with the true gradients coming from ionosphere spatial variation. The various raw measurement corrections and consistency checks have been gradually improved through the development of several version of the tool in order to reduce the number of measurement artefact detected. Additionally, specific tools have been developed to reduce the required effort to manually validate a gradient. The combination of these developments allowed the manual validation of a very large number of ionosphere gradients.
All data from October 8th, 2012 to December, 31th, 2016 have been processed, and a total of 60716 potential gradients have been manually validated out of which, 1457 were true ionosphere gradients. The largest measured ionosphere gradient has a slope of 403 mm/Km, which remains lower than the boundary of the threat model developed for the Contiguous United State. However, this event has been measured at very low latitude from the Canary Island clusters, and GBAS users at mid-latitude may use a different boundary. The ionosphere gradient location analysis has shown that most of them are occurring either at low latitude (below 35°N) or high latitude (above 60°N). In addition, the analysis of the gradient time of occurrence has shown that most of the gradients occurred around the equinoxes and during local night hours. These characteristics are very consistent with typical ionosphere behaviour and reinforce the confidence in the overall data processing methodology.
By the time of the ION PLANS conference, all data until September 30th, 2017 will have been processed and potential gradients will have been manually validated.