Performance Analysis of Machine Learning Algorithms to Human-Pilot-Model Parameter Estimation
Rocío Jado-Puente, Stephen Brutch, Hever Moncayo · 2025
The behavior of a pilot is highly complex and nonlinear, making their responses difficult to predict. Pilot errors often lead to safety risks, especially when entering flight conditions considered as Loss-of-Control regimes. Predicting a pilot's inputs to the flight control system can help avoid dangerous flight conditions. The work presented in this paper compares different machine learning pilot parameter estimators through numerical simulations and pilot-in-the-loop experiments to evaluate the accuracy of these techniques in estimating pilot characteristics and parameters using a McRuer pilot model.