敏捷性导弹的神经网络控制与制导

为了使导弹获得更高的机动性、敏捷性和导引精度,大多采用推力矢量控制方案.本文讨论了新型推力矢量控制导弹的动力学建模,在此基础上提出了适用于推力矢量导弹的神经网络控制与制导方案,其中的导弹控制方案包括双网络逆动态学习控制和混合神经网络控制两种.神经网络制导采用基于线性二次型的最优制导的有教师学习方案.神经网络学习算法中因引入了基于经验的模糊规则,具有较好的学习性能.仿真表明神经网络控制与制导方案具有较好的末制导导引精度.

NEURAL NETWORK CONTROL AND GUIDANCE FOR AGILE MISSILES

Advanced agile missiles possess maneuverability, agility and precise guidance performance by adopting thruster vector control. System dynamics and modelling for the advanced missiles with thrust vectoring is investigated, and neural network control and guidance for advanced missiles is proposed. Two neural network control methods are presented: the twin dynamic inversion networks learning control and the integrated neural network control.The neural network guidance uses the teacher learning method according to the linear quadratic optimal guidance law. Fuzzy learning rules are introduced to improve the learning ability of the neural network system. Numerical simulation results show that the proposed neural network control and guidance has good performance to hit maneuvering targets.