基于神经网络优化直 / 气复合控制导弹迅捷转弯

The control of the quick turning of air-to-air missile is to complete the entire maneuvering process in the shortest possible time under the premise of ensuring the stable flight of the missile. In order to achieve the above tactical indicators, a lateral thrust / aerodynamics composite controller based on sliding mode variable structure control and fuzzy logic is designed considering the fast time denaturation, strong nonlinearity and large interference of the missile quick turning. In view of the uncertain initial value and excessive external interference that may occur in actual flight, the fixed-time convergence theory and the extended state observer are introduced to effectively avoid the control input jitter caused by interference and ensure that the attitude angle of missile can converge quickly and stably within a fixed time. The neural network is used to optimize the design parameters of sliding mode variable structure control, which not only ensures the quick turning of missile, but also reduces the mass of missileś lateral engine. The effectiveness of the proposed control and distribution scheme is verified by simulation. The designed controller has certain robustness, stable and smooth control effect, strong anti-interference ability and excellent attitude tracking performance in complex situations such as model uncertainty and external interference, and the tracking effect of the control algorithm after parameter optimization is better, which reduces the original convergence time by 12. 6% and the original maximum control force by 9. 6% .