GBAS based autoland system: A bottom up approach for GAST-D requirements

• Published in: IEEE/ION Position, Location and Navigation Symposium pp. 566 - 574
• Main Authors: Belabbas, Boubeker, Dautermann, Thomas, Looye, Gertjan, Kladetzke, Jan
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.05.2010
• Subjects: Aircraft navigation; Airports; Automatic testing; Availability; Certification; Instruments; Mechatronics; Multilevel systems; Robotics and automation; System testing
• ISBN: 1424450365; 9781424450367
• ISSN: 2153-358X; 2153-3598
• DOI: 10.1109/PLANS.2010.5507312

Ground Based Augmentation Systems (GBAS) for the precision approach of aircraft have never been so close to support automatic approach and landings under CAT IIIc conditions as they are today. However, as one key requirement for certification, it has to be demonstrated that the required integrity, continuity and availability allocations for GBAS are met, even under the most unfavorable circumstances. As for ILS/MLS, the performance of the total system (GBAS, the aircraft and its automatic landing system) is the key to eventual certification. Contrary to ILS/MLS-based guidance, at this moment no standard models and parameters for navigation errors and disturbances have been defined yet for assessing automatic landing performance based on GBAS. This raises the questions (1) how well does the integrated system perform with state of the art GBAS error models, and (2) for a given performance level reached with ILS/MLS, how much error could be tolerated when using GBAS instead? In this paper we have made a preliminary step to find an answer to the first question. To this end we have taken an existing, flight tested automatic landing system, tuned for use with DLRs fly-by-wire test bed ATTAS (Advanced Technologies Testing Aircraft System), which is representative for EASA Part 25-class aircraft. The landing system was adapted for use with GBAS and an advanced GBAS error model was incorporated in the simulation model. With this set-up, we performed the same Monte-Carlo simulations as were originally done with ILS. Initial results indicate that, based in the adopted error models, CS-AWO criteria can be met with more ease than with ILS/MLS-based guidance. It is expected that performance of GBAS-autoland systems can be further improved by exploiting new features that GBAS has on offer, tolerating larger NSE errors or possibly allowing operational autoland limitations, like maximum crosswind, to be relaxed.