基于SVR的火控机解算性能测试  

Testing the Calculation Performance of Fire Control Computer Based on SVR

作　　者：张子宁[1] 单甘霖[1] 段修生[1] 张岐龙[1] 

机构地区：[1]军械工程学院光学与电子工程系,石家庄050003

出　　处：《科学技术与工程》2009年第16期4799-4802,共4页Science Technology and Engineering

摘　　要：针对火控机解算性能的测试问题,提出使用支持向量机回归模型来拟合射击诸元曲线。通过使其与火控机解算出的实际射击诸元进行互相关计算来统一时序关系,以便于比对误差,从而检测火控机解算的正确性。为了验证该拟合方法的可行性,对支持向量机回归的拟合过程进行了计算机仿真,并与用人工神经网络等方法拟合的效果进行了比较。According to the test problem of fire control computer calculation performance,put forward that using support vector machine regression model to fit shooting parameters curve. And by calculating the cross-correlation between the curve fitted by SVR and actual shooting parameters calculated by fire control computer to make their time-sequential relation same in order to be easy to get the deviation between them , the correctness of calculation of fire control computer could be detected. Computer emulation is used for verifying feasibility of the fitting approach by SVR, and which result is compared with the fitting results by other approaches, such as BP networks and so on.

关 键 词：支持向量机回归 射击诸元 互相关 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程]