Autonomous Flight Envelope Estimation for Loss-of-Control Prevention

Abstract

A nonlinear aircraft dynamics modeling and system identification approach is formulated and used to derive a rapid, convex optimization-based algorithm for the estimation of aerodynamic force coefficients from noisy sensor data, with uncertainty quantification. Furthermore, with the aerodynamic force coefficients identified, a computationally fast method for generating the corresponding aircraft trim envelope given safe limits on achievable thrust and angle of attack is presented. This unified approach achieves the online Bayesian inference of the aircraft aerodynamic performance capability while running in the background as the aircraft is controlled independently by the pilot or automation system. The resulting information is readily incorporated into the determination of the extended safe maneuvering envelope, pilot displays, and automation algorithms for flight planning, trajectory generation, and guidance: all to help maintain safe aircraft operations under both nominal and off-nominal flight conditions.