Novel Data Driven Approach to Enhance Pilot Training on Situational Awareness

Abstract

Situational Awareness is an important aspect of flight performance and safety. The concept describes the perception, integration and projection of elements of the environment that are relevant to the mission. In many cases a breakdown of situational awareness leads to accidents or serious incidents. Yet, the training and measurement for the concept still miss important pieces of information that could help improve flight safety.

I proposed a system to close these gaps, collect information on SA performance and introduce methods through direct feedback to improve flight training. The methods designed include an automated scenario generation using evolutionary algorithms and a structured toolbox incorporating elements of past flight performance. Furthermore, I designed a composite measurement method that combines subjective proven measures with data driven machine learning methods and topological analysis. I evaluated the systems through a set of experiments in the flight simulator and through surveys.