Research On The Nonlinear Parameter Identification And Robust Control Of Aircraft

The parameter identification and robust control of the aircraft have great theoretic and practical significance to the design and realization of modern aircraft. Directly based on nonlinear model of the aircraft, the thesis has mainly studied the two correlative contents: the aerodynamic parameter identification and the robust control of aircraft. And the thesis makes efforts in theoretically and technically supporting the aircraft parameter identification and robust control. The thesis pays attention to the utilization of transcendental information, such as engineering experience, so as to make the results more easily accepted by engineers. The thesis also makes efforts in the improved efficiency and easy application of methods.The main contributions are listed as follows:1. A maximum likelihood method based on the interior point method is proposed, which combines the interior point method, an effective optimization tool, with the maximum likelihood method, and has improved the efficiency and convergence ability of maximum likelihood method. When converged, the new algorithm reserves the nice properties of maximum likelihood method. And the new algorithm has the ability to always keep the identified results stay in a reasonable range. The thesis studied offline parameter identification of the aircraft mainly based on the algorithm, which verified the validity of the algorithm.2. A recursive maximum likelihood method based on the interior point method is proposed, which combines the interior point method with the maximum likelihood method. The new algorithm converges faster than generic recursive maximum likelihood method, and always keeps the identified results stay in a reasonable range. So the new algorithm is more appropriate to online applications. When converged, the new algorithm reserves the nice properties of recursive maximum likelihood method. Some instances of aircraft's online parameter identification in the thesis indicate that the algorithm has goodapplication foreground.3. An approach to combine the maximum likelihood method with the genetic algorithm is proposed. For aircraft parameter identification, the maximum likelihood method is used as a local optimization operator to make use of the global optimization ability of genetic algorithm and the convergence rate and accuracy of maximum likelihood method. The algorithm was applied to the area of aircraft parameter identification, which indicated the validity of it.4. We propose a scheme of nonlinear aircraft robust control. In this scheme the model control error of inverse system method is compensated by a neural network with a rational structure and a formally simple learning rule of the weights. The scheme was applied to some typical and important fields of flight control, such as the autolanding and the high angle-of-attack control, which verified the validity of the scheme.