Total System Performance of GBAS-based Automatic Landings

Kurzfassung

In this work, automatic landings of aircraft based on signals from Global Navigation Satellite Systems (GNSS), augmented by a Ground Based Augmentation System (GBAS), are investigated. By taking into account available knowledge that is currently not used and a more realistic modelling of the autoland performance several suggestions for improving GBAS are made. After a short discussion of the Instrument Landing System (ILS) and GBAS a motivation for the use of GBAS is given by comparing the performance of both guidance systems in flight trials. The results show that GBAS is much less susceptible to disturbances and provides more precise and smoother guidance. The next chapter continues with a description and a discussion of the derivation of navigation requirements for GBAS from the definition of a safe landing. The total error budged has to be split between the autopilot and the navigation system. A critical review of the derivation process proposes adjustments by taking into account available knowledge about the satellite geometry in order to reduce the monitoring requirements and thus increase system availability. This discussion is followed by an investigation of the autoland performance based on one example autopilot implementation. The results show that the currently used way to model the touchdown performance by a Gaussian distribution is not well suited. Either very conservative inflation is necessary or the tail probability of landing outside the required touchdown zone may be significantly underestimated. It is therefore suggested to model the touchdown performance by a Johnson distribution, better fitting to the obtained results. Finally, the contribution of the navigation system to the total system error is discussed. As the main concern in differential navigation techniques, such as GBAS, results from ionospheric disturbances these phenomena are discussed more in detail. For the GBAS ground system an improved ionospheric monitor is proposed, based on adding an additional reference receiver for monitoring purposes. Furthermore, an ionospheric monitor for future dual-frequency GBAS modes is developed. Such a monitor is necessary if positioning is to be done based on single frequency modes, which seems to be a very likely way forward. Shifting this monitoring task to the airborne system has the advantage that all knowledge about current navigation performance and autopilot performance for that specific aircraft type can be exploited, facilitating the ionospheric monitoring task significantly.