AcListant Exponat auf dem ATC World Congress

Kurzfassung

Nowadays Arrival Managers (AMANs) are implemented to give the controller guidance advisories for optimized approaches. Information about deviations from the planned sequence is exchanged between controller and pilot via radio communication. The AMAN gets this information only by changes in the track of the aircraft with a delay of about 30 seconds. Therefore, the planning algorithm of the AMAN sometimes is not up to date. The demonstration at the AT-One booth 470 shows how to avoid this delay by analysing the controller-pilot-communication by an automatic speech recognition system (ASR), being developed by the Institute of Spoken Language Systems at the Universität des Saarlandes (UdS). The radar data of the first validation trial for Düsseldorf Approach are input into two different versions of DLR’s Arrival Manager 4D-CARMA. The left screen shows the output without using the information of a speech recognizer. The Arrival Manager on the right screen, however, uses the input of the second input sensor. Controllers giving tactical advisories outside the training environment do not always stick to the book. Nevertheless a speech recognizer with a high recognition rate is necessary. This is achieved by communication in the opposite direction: The AMAN does not the controllers’ intent, but a set of possible intents. Therefore, the AMAN can provide the ASR system with a set of expected controller commands which reduces the possible hypotheses of the ASR system. Again we show in an interactive environment the results of two speech recognizers: The first one, which uses the input from AMAN, and the other one, which has to rely on the controller’s utterance alone. This report describes the setup and the responsibilities before and on the ATC World Congress.