基于RBF网络机载火控系统动态精度建模  

Dynamic Precision Model of Airborne Fire Control System Based on Radial Basis Function Neuron Network

作　　者：李阳[1] 周德云[1] 张堃[1] 

机构地区：[1]西北工业大学电子信息学院,陕西西安710072

出　　处：《指挥控制与仿真》2015年第4期45-48,共4页Command Control & Simulation

基　　金：国家自然科学基金(61401363)

摘　　要：针对传统机载火控系统动态精度分析方法中建立模型形式固定、预测精度较差等问题,提出一种基于径向基(RBF)网络的机载火控系统动态精度建模方法。首先进行误差序列自回归(AR)模型辨识,根据AR模型阶数识别准则以及最小信息准则确定RBF网络输入神经元个数,然后利用RBF网络对动态误差序列进行建模,最后通过预测精度评定模型的优劣。仿真结果表明,相对于传统方法建立AR模型,选用RBF网络建立的模型,预测误差更小,精度更高。For the problem of the fitting model is fixed and the precision is low of the traditional airborne fire control system dynamic accuracy analysis, research on dynamic precision model of airborne fire control system on radial basis function neuron network is proposed. Firstly, it needs a sequence error auto regressive （AR） model identification, to determined RBF network number of input neurons based on AR model identification criteria and Akaike information criteria. Secondly, the model for dynamic error sequence is established using RBF neuron network. Finally, it evaluates the merits of the model through predicting accuracy. The simulation results show the smaller prediction error and higher precision using the radial basis function neuron network model compared with the traditional method.

关 键 词：动态精度 AR模型 最小信息准则 RBF网络 

分 类 号：E920.2[军事—军事装备学]