遗传算法在质量矩导弹姿态控制中的应用

研究了三轴稳定控制系统的姿态控制问题.建立了内部带有3个可移动滑块的数学模型.利用RBF(Radial Basis Function)网络的快速自学习以及自适应特性,将其作为控制器不断调整滑块的位置.同时利用遗传算法(GA)的全局、并行寻优及增强式学习能力对神经网络控制器的权系进行优化.通过对滑块位置的协调控制,改变导弹的飞行姿态.最后讨论了滑块运动特性对系统性能产生的影响并对系统稳定性进行了分析.仿真结果表明该控制律能在保证系统稳定的情况下有效地实现导弹的姿态调整,并且提高了系统的动态响应品质. 

Moving-mass attitude control law based on genetic algorithm

The ability of the moving-mass attitude control system was investigated to control a vehicle with three-axial stabilization in intra-atmospheric space.And the general nonlinear motion equations with three internal moving masses were provided.For the fast self-learning and adaptive capacities of radial basis function(RBF) networks,used it as the controller to calculate the moving masses' position.With the coordination of the control,the masses were positioned independently to generate the modest attitude corrections for the vehicle.At the same time,the weights of the networks were optimized by genetic algorithm(GA).At last,the influence to the system performance caused by the moving masses' motion was discussed and the system stability was analyzed. Simulation results show that the system stabilization can be satisfied to realize efficiently attitude adjustment for the vehicle,and the dynamic performances of the system are improved.