The department controller assistance develops prototypic support systems for air traffic controllers. Controllers issue commands to aircraft pilots via voice. These commands also need to be entered into the air traffic control system to enable support functionalities for controllers. Entering such commands via keyboard or mouse creates additional effort for controllers. Furthermore, the relevant information already exists within the verbal command. Hence, automatic speech recognition is used to extract the contents of controller commands to feed the air traffic control system.
DLR developed an assistant based speech recognition (ABSR) system with command recognition error rates below 1 % for the approach area. The ABSR consists of a speech recognition engine and a command hypotheses generator forecasting the most probable commands in the current situation. These commands reduce the search space for the speech recognizer and lead to lower command recognition error rates. In a first version, hypotheses were rely on radar data, flight plan data and meteorological data as well as expert knowledge about the airspace and air traffic in a specific environment. An enhanced version of the hypotheses generator used machine learning algorithms to forecast commands for all relevant aircraft depending on their position. For the first time, hypotheses have also been generated for the tower area automatically learned from annotated speech and radar data. For even more accurate forecasts, a state machine approach should be used. Particularly aircraft on the ground follow a sequence of actions respectively states such as engine start-up, pushback, taxi, line-up, and take-off, etc. If it is known, which commands the aircraft has already received respectively executed, the number of predicted commands can be further reduced. This knowledge can be generated from radar as well as speech data with some lower accuracy (probability) from the speech recognition system itself. Hence, algorithms should be developed to forecast controller commands based on a state machine approach assigned with a probability. Those hypotheses should then be evaluated against recorded data from human-in-the-loop trials with tower controllers from Hungary and Lithuania.
In the course of your master’s/bachelor’s thesis, the following tasks need to be handled:
Look forward to a fulfilling job with an employer who appreciates your commitment and supports your personal and professional development.
Our unique infrastructure offers you a working environment in which you have unparalled scope to develop your creative ideas and accomplish your professional objectives.
Our human resources policy places great value on a healthy family and work-life-balance as well as equal opportunities for persons of all genders (m/f/non-binary).
Individuals with disabilities will be given preferential consideration in the event their qualifications are equivalent to those of other candidates.
You can send this job advertisement via e-mail and complete your application on a personal computer or laptop.
We need your digital application documents (PDF). The document upload function is not supported by all mobile devices. Please complete your application on a PC/laptop.
Dr Oliver Ohneiser
Institute of Flight Guidance
Phone: +49 531 295-2566