基于遗传算法优化LSSVM的着靶速度建模与预测

Modeling and Prediction of Target Velocity Based on Genetic Algorithm Optimized LSSVM

测试弹丸的着靶速度是靶场试验的重要科目,但是当弹丸飞行状态异常时,雷达就无法准确测试着靶速度,所以利用已测数据对未能准确测试的数据进行预测就很必要。常用的预测模型是GM(1,1)灰色模型,预测精度不理想。为了提升预测精度,选择采用BP神经网络和支持向量回归机进行预测,但这两个模型的参数是随机选取的,预测精度不是最高,所以选择利用遗传算法优化最小二乘支持向量机预测最优参数。实验结果表明,遗传算法优化最小二乘支持向量机的预测精度最高,误差小于2‰,是预测着靶速度的最佳模型。

Testing the target velocity of projectile is an important subject of rang test,but when the projectile flight state is abnormal,the radar can not accurately test the target velocity,for it is necessary to use the measured data to predict the data that can not be accurately tested.The commonly used prediction model is GM(1,1)Grey model,but there are some shortcomings in this model,resulting in poor prediction accuracy.In order to improve the prediction accuracy,BP neural network and support vector regression machine were selected for prediction,but the parameters of these two models were randomly selected,and the prediction accuracy was not the highest.Therefore,the least squares support vector machine was optimized by genetic algorithm for prediction,and the genetic algorithm can find the optimal parameters of support vector machine.The experimental results show that the prediction accuracy of least squares support vector machine optimized by genetic algorithm is the highest,and the error is less than 2‰,and it is the best model to predict the target velocity.