No CrossRef data available.
Published online by Cambridge University Press: 13 June 2025
In this paper, an intricate autopilot architecture is built for a highly nonlinear aircraft. The mode control of the autopilot is illustrated with state transition diagram. The fundamental concept of this architecture is a model-free design, which means that aerodynamic derivatives and model linearisation are not required. However, only the dynamic control allocation (CA) technique involves the use of derivatives for aerodynamic control. With the use of the Incremental Nonlinear Dynamic Inversion (INDI) in this framework, reduced dependency on models can be achieved. The algorithm addresses system nonlinearities. Our study simulates the approach, and the typical terminal arrival phase of the flight. For lateral navigation, a line-of-sight (LOS) based guidance approach is employed. The principle of the LOS is look ahead-based steering and, mainly the cross-track error is primarily used in the literature. The cross-track error is not the only error for establishing the predefined path. The error between the aircraft heading and the path course is required to describe how fast heading change occurs to re-establish the path again. This proposed modified LOS improves not only the path following ability, but also fault tolerance during aileron or rudder malfunctions. In addition, control surface lock-in-place failures, right/left engine flames-out, and turbulence/wind cases are studied. The objective is a safe approach without reconfiguration. The effectiveness of the proposed control architecture is demonstrated with numerical simulations for the NASA Generic Transport Model (GTM) T-2.