消防炮伺服系统控制策略

Fire extinguishing cannon servo system control strategy

针对远距离消防炮随动发射装置位置伺服系统中存在的强耦合、非线性以及参数时变等不确定因素,提出一种基于遗传算法优化双隐层神经网络的控制策略。为提高控制精度,将遗传算法和组合神经网络结合,对RBF-BP组合神经网络的初始阀值和权值进行优化。数值仿真结果表明:该种控制策略与单一的RBF、BP神经网络控制策略相比较,具有控制精度高、响应速度快、迭代次数少的优点,改善了位置伺服系统的静态响应和动态性能,提高了系统的鲁棒性。

Against the strong coupling, nonlinearity and uncertainty in position servo system of long-distance distinguishing cannon's launching device, a system control strategy based on Genetic Algo?rithms optimization RBF-BP Neural Network was proposed. The genetic algorithms was combined with neural network to improve the control precision, optimizing the initial weights and threshold. In comparison with the single RBF and BP neural network control?ler, the numerical simulation results showed that the proposed method has the advantages of high control precision, fast response speed and less iteration times. It can improve static, dynamic per?formances and robustness of the position servo system.