Adaptive prediction of aircraft loss-of-control margins and safe control envelopes with longitudinally and laterally coupled higher-order dynamics

Abstract

Predictive systems with the ability to warn pilots of impending entry into loss-of-control have the potential to improve safety in flight. Along these lines, research at the General Aviation Flight Lab at Wichita State University has sought to develop an early warning mechanism to predict an aircraft’s “receding-horizon” safe control envelope and warn pilots of future impending excursion of its safe flight envelope. The intent of the concept is to mitigate entry into loss-of-control by continually keeping the aircraft at a certain time-distance from the edge of the flight envelope. Our prior work has analyzed the decoupled 2nd and 4th order longitudinal and lateral/directional modes independently. While this approach was sufficient for adverse flight configurations occurring exclusively in either the pitch or roll/yaw axes, lossof-control often involves flight configurations with high angles-of-attack and large sideslip angles occurring simultaneously, leading to cross-coupling in the aircraft dynamics. In addressing this, that prior architecture has been refined and extended to the 8th order aircraft with cross-coupled dynamics between the longitudinal and lateral/directional modes. The adaptive prediction architecture estimates the aircraft's proximity to its control loss boundaries through calculations of critical control deflection, rate, acceleration, and higher order input limits that would drive the aircraft to control loss at some point several seconds in the future. Adaptive parameter estimation is implemented to enable real-time modeling error identification, accounting for uncertainties such as failures or modeling error in the aircraft. A pilot advisory display presents the predicted control margins to the pilot through head-up augmented-reality technologies, providing pre-emptive warning of impending entry into control loss within the look-ahead prediction window. The architectures are applied to an 8th order light business jet model in simulation, and results demonstrate successful prediction of the critical control limits that should not be exceeded in order to avoid near-term future entry into control loss.