The Research Of BTT Missile Control System Based On Dynamic Inversion System And Neural Network Theory

To improve the maneuverability of missile, many countries are doing research on all kinds of control system with great concentration. The Bank-to-Turn control is an advanced technology in the area of today's world missile control, and it has great significance in developing the long distance, high speed and high maneuverability missile system both in theory and engineering application. The features of BTT control have determined that BTT missile is a strong coupling and time varying system, which is hard to design a control system. The traditional design method of three independent channels is not suitable. So to develop an efficient method for BTT missile's control become a hotspot for researchers in the area of missile control. With the development of nonlinear theory and artificial intelligence, the dynamic inverse system theorem and neural network are becoming strong tools in analysing and designing uncertain nonlinear system.In the aspects of theory and engineering application, this paper has done deep research on the application of dynamic inverse system and neural network control theory in the control system of BTT missile. The main work includes:First, the model of missile's control system is presented, and the features of BTT missile's control system are displayed. In order to design BTT missile's strong-coupled and time-varying control system, this paper develops an efficient approach to design nonlinear control system-dynamic inverse system method. For the method cannot apply to Non-minimum phase system directly, two schemes are provided to realize the application: 1) decompound the system dynamics into slowness and swiftness sections, 2) redefine the outputs of control system of BTT missile. In this way, the dynamic inverse system controller is designed.For that the nonlinear dynamic inverse system requires accurate mathematic model but the mathematic model of missile is uncertain, this paper designs the robust dynamic inverse controller for missile integrating dynamic inverse system method with adaptive neural network control. To quicken the speed of neural network algorithm which is to realize real time control for missile, this paper has done research on the BP neural network algorithm and its ameliorated algorithm.In order to comprehensively validate the specifications and robustness of the controller which is guiding the flying missile to track objects, this paper sets up the mathematic model of the nonlinear time varying control system of BTT missile and does the simulation.The main innovation of technology includes:1. A new mathematic model including all coupling variables in which all state variables are observable is proposed. This model utilizes the information of BTT missile has known to simplify the design, realize the robust control of BTT missile and make them available. In this way, the dynamic inverse controller of one BTT missile has been designed successfully. It has been validated that the outputs of track gradually impended over zero in the theory.2. This paper brings up an adaptive neural network to compensate inverse error and designs the adaptive control architecture of study online of neural network. When the aerodynamic parameters have stirred or the system has been disturbed, the feedback error is able to adjust network weights online to impend over inverse errors, so that the affection on the whole control system is killed dynamically.3. Based the inverse system theory and neural network adaptive theory, this paper gives a control scheme for missile's BTT control and its application method in detail, designs its robust dynamic inverse controller, and settles the problem of the robust of missile's BTT control sysem.4. Based on Marquardt algorithm, this paper proposes a new ameliorated BP algorithm to quicken the speed to converge., which is to realize real time control for missile.The work of this paper has indicated that the robust controller based on dynamic inverse system and adaptive neural network theorem is available, and meets the specification of the system. The approach is an efficient engineering method for the coupled nonlinear time-varying system synthesis, and is a good reference for engineers of control theory, its engineering applications, aircraft design and so on.