Artificially Intelligent Assistance for Pilot Performance Assessment

Current teaching and evaluation practice for flight training lessons are carried out by humans. The instructor is responsible for evaluation of the student pilot skills. Pilot performance is typically assessed by the comparison of printed flight profiles with expected trajectories. This manual process means a significant workload for instructors who must deal with several evaluations of pilots’ critical tasks. Artificial Intelligence (AI) is an attractive ad hoc solution for highly skilled and sophisticated activities traditionally performed by human beings. This paper presents an approach to assess airplane’s pilot performance based on AI algorithms. This paper describes a setup to provide a capability for data collection from a certified flight simulator and employ the above algorithms capable of evaluating airplane’s pilots’ behavior as well as pilot’s progress during a training. Initial simulation results are encouraging as to produce satisfactory comparisons of pilots’ performance and suggest a potential AI-assisted support for pilot performance assessment. Concluding remarks and future work are also presented.