In this study the concept of using GBAS corrections and integrity parameters for UAV operations is explored. The GBAS messages are received and decoded at the airport to overcome the issue of VDB coverage for alternative applications that do not have line-ofsight to the GBAS. A flight test at a distance of about 15 km from an operational GBAS ground station was conducted using a quadrotor UAV as test vehicle. The recorded GBAS messages and recorded pseudoranges and carrier phases were subsequently used to evaluate the performance that can be obtained. One main issue that was identified is the number of satellites lost when maneuvering the drone. Loss of tracking requires a re-initialization of the smoothing filter and thus entails a wait time before a satellite can be used for positioning again. The GBAS corrections are generated based on 100 s smoothing, requiring a user to use the same smoothing time constant do avoid potential differential error buildups due to different filters. Hence, the time until a satellite can be used again after reacquisition may be substantial. When losing several satellites simultaneously or shortly after another this decreases the number of available satellites significantly and negatively impacts accuracy and protection level performance of this proposed mode of operation. The paper also studies the multipath effect at different altitudes and for different smoothing time constants compared to the standardized error models that were defined for large transport aircraft. When using 100 s smoothing the results indicate that the models are suitable also for the use case of UAV operations.

In this paper, we investigate the potential performance of the envisioned Differentially Corrected Positioning Service (DCPS) for the Ground Based Augmentation System (GBAS). For this purpose, we receive and decode the GBAS messages transmitted by the operational GBAS station providing GBAS Approach Service Type (GAST) C (i.e. a CAT-I service) at Zurich airport. Based on these parameters we calculate the respective protection levels for the GBAS DCPS and compare them with the SBAS protection levels. Due to the fact, that the GBAS corrections are locally generated, the achievable performance is usually better than with SBAS. However, the GNSS navigation errors decorrelate with distance, leading to an increase in the protection level. Both, the GBAS and SBAS protection levels, are dependent on a k-factor. This factor is derived from the allocated integrity risk to an integrity risk and equals 6.18 in the case of SBAS. For GBAS, the k-factor is 10. This is a very conservative value, which would be adapted in the future as the DCPS, its use-cases and the required level of integrity are developed further. To show a less conservative and more realistic scenario for comparing SBAS and GBAS protection levels, a GBAS protection level with a reduced k-factor of 6.18 was additionally calculated. The resulting performance of the regular GBAS and the GBAS with a reduced k-factor is then compared to the protection levels of a purely SBAS-based navigation to identify areas of potential use of the proposed system. Based on these evaluations we show that the GBAS DCPS yields better performance in an area of (on average) more than 50 km around the GBAS station.