旋转弹的神经网络变质心姿态控制

研究了旋转弹变质心控制系统的姿态控制问题.建立了内部带有n个可移动滑块的导弹系统仿真数学模型,分析了滑块与弹壳间的相对运动对系统运行产生的耦合影响.利用神经网络的自学习性以及自适应特性,设计了基于神经网络控制的姿态控制律来计算系统质心的期望位置.利用最优原理确定了各滑块的期望偏移以实现系统质心位置的改变,从而达到改变导弹飞行姿态的目的,提高了系统的动态响应品质.以带有2个滑块的旋转弹姿态控制系统为例进行非线性仿真,证明了所设计控制律的有效性.

Moving-mass attitude control system for spinning vehicles based on neural networks

The ability of a moving-mass attitude control system to control a spinning vehicle was investigated.The general nonlinear equations of motion with n internal moving elements were provided.The coupling influence to the system caused by the relative movement of the moving masses to the vehicle shell was described.For the self-learning and adaptive capacities of the neural network control system,the attitude control scheme based on the neural networks was produced to improve the dynamic performances,which calculated the desired center position of the system mass.In order to generate the modest attitude corrections of the vehicle,the mass position algorithm was presented by using the optimal principle to get the commanded position of each mass to realize the desired change of the system mass center.A nonlinear simulation of the spinning vehicle with two internal moving mass actuators demonstrated the ability of the controller to effectively control the vehicle attitude.