Quantitative Assessment of Drone Pilot Performance

This paper introduces a quantitative methodology for assessing drone pilot performance, aiming to reduce drone-related incidents by understanding the human factors influencing performance. The challenge lies in balancing evaluations in operationally relevant environments with those in a standardized test environment for statistical relevance. The proposed methodology employs a novel virtual test environment that records not only basic flight metrics but also complex mission performance metrics, such as the video quality from a target. A group of Belgian Defence drone pilots were trained using this simulator system, yielding several practical results. These include a human-performance model linking human factors to pilot performance, an AI co-pilot providing real-time flight performance guidance, a tool for generating optimal flight trajectories, a mission planning tool for ideal pilot assignment, and a method for iterative training improvement based on quantitative input. The training results with real pilots demonstrate the methodology’s effectiveness in evaluating pilot performance for complex military missions, suggesting its potential as a valuable addition to new pilot training programs.