Ionosphere Threat Space Model Assessment for GBAS

Kurzfassung

Ground Based Augmentation Systems (GBAS) can correct the majority of the GNSS pseudo range errors experienced by an aircraft in the vicinity of an airport. Not corrected (spatially uncorrelated) errors between ground and airborne subsystems must be overbounded and kept as small as possible in order to reach the required level of integrity defined by ICAO. Ionosphere gradients remain in general very small (can be bounded by 4mm/km in the CONUS region). This "normal" behavior of the ionosphere has a very limited impact on the position error. The confidence interval of the user position is fully acceptable for precision approach. Unfortunately, the ionosphere medium is sometimes subject to perturbations due to the strong temporal and spatial variability of the ionospheric plasma. When GNSS signals received by the aircraft are delayed in a different way than the GNSS signals received by the GBAS ground facility (GGF), the corrections provided by the GGF can cause unacceptably large position errors at aircraft level. While most of the time this environmental effect is behaving in a normal way, there has been found an anomalous ionospheric behavior which occurs rarely (few occurences in 10 years) but can be a serious threat to GBAS integrity. The CAT I GBAS architecture can principally not fully mitigate these effects by monitoring. According to the agreed approach in the GBAS CAT I community, the remaining risk is therefore treated as follows: At each epoch the worst-case ionospheric threat is assumed to occur in 100% of the time. The threat is mitigated, e.g., by preventing the aircraft from using unsafe combinations of GNSS satellites. To permit the analysis, it is essential to first define the ionosphere threat space. Since the anomalous ionosphere threat consists of moving ionospheric fronts, the ionosphere threat space is spanned by the slope, velocity and width of such an ionospheric front. The anomalous ionosphere threat model is defined by specifying a domain inside the three-dimensional threat space. For the CONUS (conterminous US) region this domain in threat space has been determined by using empirical data collected during the last solar cycle. In order to use the proposed mitigation algorithms for the ionosphere threat in a different geographical region, the anomalous ionosphere threat model has to be established for that region as well. For the certification of a GBAS ground facility in Germany, both the anomalous ionosphere threat space and the nominal ionospheric de-correlation for a region including Germany were determined. This work has been done within the ITMA (ionosphere threat model assessment) project which is a joint project between the German Air Navigation Service Provider DFS and the German Aerospace Center DLR, funded by DFS. In the first phase of the project, an automated data-screening has been performed using all publicly available dual-frequency RINEX data of the entire last solar cycle in the region considered in this study. This first phase aims to identify the days of most extreme ionospheric events. When dual-frequency data is available, there are a number of ways of extracting the ionospheric delays. While for the data-screening the difference between the phase measurements on GPS L1 and L2 were used, we also consider code minus carrier (CMC) and code-difference observable in order to validate extreme events. In the data-screening we computed the number of (formal) spatial gradients above certain limits derived from time-differences of phase-difference-derived ionospheric delays. We call these gradients formal, since they are not corrected for the movement of the ionospheric front and independent temporal changes. Nevertheless periods of extreme ionospheric activity can be identified using these formal gradients as an indicator. By analyzing all available data from a complete 11 years solar activity cycle period (1998-2008) using an automatic screening process, we have determined 16 time periods of relevant ionospheric activity. All these events have been examined manually in order to exclude false alarms caused, e.g., by cycle slips or corrupted data. In the second phase of the project, algorithms have been developed for estimating the parameters in the anomalous ionosphere threat model, i.e. the slope, speed and width of a ionospheric front. In order to estimate the key threat model parameters, the slope of the ionospheric front, the velocity and direction of the ionospheric front have to be determined. For determining the front velocity, we used both a least-squares-based estimation technique and a more direct way of computing the velocity and slope directly from (calibrated) ionospheric delays. For each period of anomalous ionospheric activity we have manually identified the time and location of extreme events. Additional data from the German geodetic reference network (SAPOS) for the determined locations has been used to increase the spatial resolution. Then, for each event threat model parameters have been estimated. Again, CMC and code-difference observables were used in addition to phase-difference-derived ionosphere delays for validation purposes. As a result, a threat space valid for the mid Europe especially Germany was derived. In addition to the extreme ionosphere behavior, which is captured by the anomalous ionosphere threat model, we have also determined the nominal ionospheric gradients in the considered region using calibrated vertical ionospheric delays from periods of quiet to moderate ionospheric activity. The paper will give a detailed insight in the derived methods for ionospheric threat space determination and the achieved results for the geographical region "Germany". The results act as an essential input for further parameterization and certification of GBAS systems in Europe.