摘要
战场态势感知和辅助决策是智能空战的重要组成部分,飞行动作识别是多项关键技术的基础。文中提出了一种基于微分思想和卷积神经网路的高效飞行动作识别方法。首先,对训练数据按照微分思想进行分割;然后,通过卷积神经网络对飞参数据的训练实现飞行动作特征的自动提取、对飞参数据进行飞行动作识别,该方法扩展性好,有利于快速处理大量飞参数据;最后,对四种基本飞行动作单元进行仿真实验,仿真结果表明,该方法具有良好的有效性和准确性。
Battlefield situational awareness and aided decision making are important components of intelligent air combat,and flight action recognition is the basis of many key technologies. an efficient flight action recognition method based on convolutional neural network is proposed. Firstly,the automatic extraction of flight action features is realized through the training of flight parameters data by network. The network training updates the weights through the back propagation method,and finally determines the weights to complete the network training,and then uses the weights to recognize flight movements from flight parameters. This method has good expansibility and is beneficial to processing a large number of flight data quickly. Finally,four basic flight movements are simulated,and simulation results show that the proposed method has good effectiveness and accuracy.
作者
方伟
王玉
闫文君
宫跃
FANG Wei;WANG Yu;YAN Wen-jun;GONG Yue(Naval Aviation University,Yantai 264001,China)
出处
《中国电子科学研究院学报》
北大核心
2021年第4期347-353,共7页
Journal of China Academy of Electronics and Information Technology
基金
国家自然基金资助项目(91538201)
泰山学者工程专项经费基金资助项目(ts201511020)
信息系统安全技术重点实验室基金资助项目(6142111190404)。
关键词
飞参数据
动作识别
基本飞行动作单元
卷积神经网路
微分思想
flight data
action recognition
basic flight action unit
convolutional neural network
differential thought
