Adaptive Inverse Control Based On Neural Networks And Its Application In An Automatic Aircraft Landing

This thesis combines artificial neural networks(NN) with adaptive inverse control(NNAI) on the basis of analyzing their principle and characteristics, and applies them in the control of nonlinear system. An application in the design of automatic landing is shown.The main works are:1. Evolutionary neural networks. Present the models, algorithms and the existing problems of the Back Propagation(BP) network, a typical multilayer neural network. Introduce the genetic algorithm to train neural networks in order to improve the learning efficiency.2. Adaptive inverse control. First give the elementary knowledge of adaptive inverse, then give the construction of filter which is proper to adaptive inverse, and LMS adaptive algorithm. The policy is used in the system control, and an achieving way in the design of nonlinear plant is proposed. Last, provide the related simulation results.3. Automatic landing system of airplane using adaptive inverse control. Traffic accidents occur during landing frequently. Adaptive inverse control is applied in the design of automatic landing system and robustness is verified. Good simulation results are obtained. At last, NNAI is contrasted with the system adopting fuzzy control after linearization.The study shows that NNAI is effective and practicable.