基于ARIMA-SVR组合模型预测弹丸初速的方法

Method for Predicting Muzzle Velocity of Projectile Based on ARIMA-SVR Combined Model

靶场试验中有时会出现雷达死机或红外未启等现象,导致雷达丢失弹丸信号无法准确测试弹丸初速。针对这种情况,选择利用ARIMA模型和SVR模型建立ARIMA-SVR组合模型预测缺失弹丸的初速数据,ARIMA模型负责预测弹丸初速的线性特征,SVR模型负责预测弹丸初速的非线性特征,ARIMA-SVR组合模型就可以同时预测弹丸初速中的线性特征和非线性特征。三个模型预测值与实测值的误差表明,ARIMA-SVR组合模型的预测精度最高,是预测弹丸初速的最佳模型。

In the range test,it is a common test subject to accurately test the initial velocity of the projectile by using the initial velocity radar,because the initial velocity of the projectile is related to whether the weapon system can accurately hit the enemy target.However,in the test,the radar will crash or the infrared will not start,which will lead to the loss of the signal of the projectile and the failure to accurately test the initial velocity of the projectile.Aiming at the situation of missing data,ARIMA model and SVR model were selected to establish ARIMA-SVR combined model to predict the muzzle velocity data of missing projectile.ARIMA model was responsible for predicting the linear characteristics of muzzle velocity,and SVR model was responsible for predicting the nonlinear characteristics of muzzle velocity.ARIMA-SVR combined model can simultaneously predict the linear characteristics and nonlinear characteristics of muzzle velocity.The results show that ARIMA-SVR combined model is the best model to predict the muzzle velocity.