Data-based training and artificial intelligence




FODA data from flights have been analyzed for a long time in airline operations and are incorporated into optimization concepts. But emergency situations are extremely rare in this context – and that is a good thing. We enrich this database with borderline situations that are practiced, recorded, and evaluated during training, so that new insights can be gained to enable more targeted and individualized training.
In various research projects conducted in cooperation with training equipment manufacturers and universities, we collect training data from the full-flight simulator and subsequently analyze it.
Together with CAE and Brussels Airline, for example, we initiated a testing procedure on the topic of data-based training. During the training on an A320 full-flight simulator, data are collected that will be analyzed and evaluated in the second phase. These data will serve as the basis for training analyses, can be integrated into competency-based training & assessment, and can, in analogous application of flight data monitoring, lead to optimization.



Artificial intelligence (AI) is also of great importance in the training sector and promises to revolutionize various aspects of flight training. We are integrating AI into our training programs in the context of various pilot projects. To give but one example: With the help of a number of our trainees we are testing the use of artificial intelligence in SEP training. In addition to the usual criteria, other visual dimensions such as the trainees' field of vision are recorded by video in order to optimize training results, Finally, the artificial intelligence also analyzes the processes in dimensions that heretofore had proven elusive, thus enabling us to identify potential for improvement. 
In all pilot projects, data protection is a key concern for us. That is why, for purposes of data evaluation, individuals are rendered unrecognizable. 
The following is to serve as an example our AI-based research models: We are extending the existing Large Language Model (LLM) from GPT by adding a context layer. In this context layer, data relevant to training and flight operations from manuals such as OM-A, OM-B, and FSM are added for the Airbus A320 use case. With the support of Microsoft, our data are hosted in a protected cloud environment. In a trial run, pilots can then access the AI support via a chat and voice interface in a training scenario. 


WE're Here for you

Oliver Hofmann

Business Development

To the contact form