Real-time Trajectory Optimization of Impaired Aircraft based on Steady State Manoeuvres

TL;DR

This paper introduces a real-time trajectory optimization method for impaired aircraft using advanced mathematical techniques, enabling quick, terrain-aware flight path planning after failures to enhance safety.

Contribution

It combines differential flatness, pseudospectral methods, and inverse dynamics to generate near-optimal, real-time trajectories for impaired aircraft in complex terrains.

Findings

Real-time control inputs computed in less than a second.

Near-optimal trajectories retain up to 80% of the optimal solution.

Method successfully applied to NASA GTM with rudder failure.

Abstract

Aircraft failures alter dynamics, diminishing manoeuvrability. Such manoeuvring flight envelope variations, governed by the aircraft's complex nonlinear dynamics, are unpredictable by pilots and existing flight management systems. To prevent in-flight Loss of Control, post-failure trajectories must be optimal, planned in real-time, avoid terrain, and adhere to the impaired aircraft's reduced manoeuvrability and dynamic constraints. This paper presents a novel real-time trajectory optimization method for impaired aircraft based on a combination of differential flatness theory, the pseudospectral method, nonlinear programming, and inverse dynamics. In the proposed method, which utilizes a high-fidelity nonlinear six degree-of-freedom model, to conform to aircraft's altered dynamics a sequence of trim points is selected from the impaired aircraft's manoeuvring flight envelope based on the chosen optimization criteria, ensuring that the resulting three-dimensional trajectory observes terrain avoidance. Then, the required control inputs are obtained for each manoeuvre in less than a second. The method is applied to the NASA Generic Transport Model with rudder failure near a complex mountainous terrain. Both an optimal one-piece trajectory and a near-optimal piecewise path consisting of several optimal trajectories, are generated in non-real-time and real-time, respectively, and compared. Results show that the near-optimal real-time trajectory retains up to 80% of the optimality.