Explicit Model-Following Design of Propulsion Control Aircraft Using Genetic Algorithms

Abstract

This paper presents the design of flight control system for crippled aircraft with engine thrust only. The longitudinal attitudes of propulsion control aircraft are guided through the explicit model-following tracker. The technique of command generator tracker is applied to convert the model-matching problem into a regulator problem. The model mismatch error is minimized by means of linear quadratic optimal control. Furthermore, the genetic algorithms (GAs) are employed to search the optimal parameters of the robust controller. Computer simulation demonstrates the responses of crippled aircraft behave like those of nominal model and the controller possesses the robustness against elevator with partial effective, engine delay, and sensor noise.

Optimal linear quadratic model following with an application to a flexible aircraft

“…Using an analogy with LQR, in Section 4, the LTV results are approximated by LTI. This approach is similar to the developments of Sato (2009) and Yu (2006). However, the results determined here present the conditions at which the approximation is valid and clarify some stabilization and robustness properties.…”

“…Such input is found to be given by a backward integration scheme involving the RM input. This LTV result has more generality than the LTI controls of Asseo (1970), Kreindler and Rothschild (1976), Sato (2009), and Yu (2006).…”

“…LQMF has also been developed by Sato (2009) and Yu (2006). The control law is based on previous results that do not consider optimality, which is composed by an LTI state feedback and LTI feed-forward terms involving RM state and input command.…”

“…The approaches of Sato (2009) are applied to an experimental aircraft and evaluated via flight tests. A damaged aircraft is studied by Yu (2006) via simulation, which is required to follow a standard behaviour.…”

A Robust Adaptive Chattering-Free Sliding Mode Control Strategy for Automotive Electronic Throttle System via Genetic Algorithm

“…In actual experiments, too large or too small values of the parameter will cause instability of the system even accidents, so we need to find the approximate range of parameters before experiments. Since GA can evaluate multiple points parallel in the parameter space, it is more likely to converge toward a global solution [28]. Because of such advantages, this underlying GA-based global optimization technique has been embedded or integrated into other control methodology [29]- [35], such as PID, H infinity control and fuzzy control.…”