摘要
火炮炮口振动数据准确预测后,可以对火炮随机振动误差进行补偿,对高机动条件下提高装备火控系统的精度具有重要意义。机器学习方法的进步为提高数据预测精度提供了有效途径。提出一种基于深度置信网络的火炮炮口振动预测模型,通过定义阶段、预处理阶段、训练阶段、测试阶段和评价阶段5个阶段,对火炮炮口振动数据进行了预测。结果表明,深度置信网络对炮口振动数据预测可以根据历史数据提取更丰富的数据特征,相较于传统BP神经网络,可以达到更好的预测精度。
The accurate data prediction of muzzle vibration can compensate for the random vibration eror of the muzzle,which is of great significance for improving the precision of the equipped fire control system under high maneuvering conditions.In recent years,advances in machine learning methods have provided an effective way to improve the precision of data prediction.A muzzle vibration prediction model based on deep belief network is proposed,and the muzzle vibration data is predicted by five stages,definition stage,pre-processing stage,training stage,testing stage and evaluation stage.The results show that the prediction of muzzle vibration data by the deep belief network(DBN)can extract richer data features based on historical data,and the achieved prediction accuracy can be better than that of BP neural network.
作者
高星
高飞
高原
何泽源
程春阳
韩志贺
GAO Xing;GAO Fei;GAO Yuan;HE Zeyuan;CHENG Chunyang;HAN Zhihe(North Automatic Control Technology Instute,Taiyuan 030006,China;Unit 31669 of PLA,Harbin 150036,China)
出处
《火力与指挥控制》
CSCD
北大核心
2023年第11期164-168,共5页
Fire Control & Command Control
关键词
深度置信网络
炮口振动
时间序列预测
预测模型
火控系统
deep belief network
muzzle vibration
time series prediction
prediction model
fire control system