直升机飞行／推进综合系统的神经网络控制器设计

提出一种H∞控制理论和人工神经网络相结合的直升机飞机/推进综合控制系统设计方法。仿真结果表明，所设计的控制器使直升机在整个飞行包线内部都具有较强的稳定鲁棒性、性能鲁棒性和参数鲁棒性。

Design of Helicopter IFPC System for the Full Envelope Flight

We study a synthesis design method that combines H∞ theory with artificial neural network (ANN) for helicopter IFPC (Integrated Flight/Propulsion Control) syst em. The control system required is nonlinear. Like previous researchers, we divide the full envelope into five flight conditions. We use H∞ theory to design linear controllers for the five flight conditions. We use ANN to combine the five H∞ controllers into a nonlinear IFPC system. Ref.［4］ used ANN to combine controllers for different flight conditions into a nonlinear system, but our method of using ANN is different from and is perhaps better than his. The simulation results, as shown in Figs. 4 and 5, show preliminarily that the IFPC system possesses a strong robustness and can achieve the desired system performance.