Hierarchical Approach to Adaptive Control for Improved Flight Safety

Following failures of primary aerodynamic actuators, safe flight can be maintained by introducing alternative actuation systems, such as analytically redundant secondary aerodynamic surfaces and propulsion, for higher-priority stability and control augmentation tasks. An intelligent hierarchical flight control system architecture is presented that is designed using nonlinear adaptive synthesis techniques and online learning neural networks to enhance flight safety. Pseudocontrol hedging is used for proper adaptation in the presence of actuator saturation, rate limits, and failure. The hierarchical structure incorporates nonactive secondary actuation channels that are engaged after a failure of a primary control surface is encountered. The methodology requires only the knowledge that a failure in a specific actuator has occurred. A model of the failed aircraft, the failure type, and the failure size need not to be known: The neural network element of the secondary channel will adapt to the failed actuator effect. The secondary control channels are designed to account for the typically lower authority and degraded performance that can be expected with secondary actuation systems. The proposed hierarchical flight control architecture is attractive, in particular, as a retrofit to existing certified flight control systems for enhanced flight safety. The proposed flight control architecture is evaluated in a nonlinear flight simulation environment, demonstrating its retrofit features. (Author)