基于卷积神经网络的反无人机蜂群智能决策算法
Intelligent decision-making algorithm for anti-UAV swarm based on convolutional neural networks
孙学章 1时晨光 1周建江1
作者信息
- 1. 南京航空航天大学雷达成像与微波光子技术教育部重点实验室,南京 210016
- 折叠
摘要
针对无人机蜂群目标突发来袭时拦截方式的选择问题,提出基于卷积神经网络的反无人机蜂群智能决策算法.采用模糊数学理论对无人机蜂群目标行为因素进行威胁度评估;建立无人机蜂群行为因素到无人机蜂群拦截方式的触发逻辑,基于此生成样本数据;针对无人机蜂群行为因素与无人机拦截方式之间的非线性关系,构建卷积神经网络结构;在此基础上,对卷积神经网络进行训练和测试,并将所提算法和反向传播(Back Propagation,BP)神经网络算法、支持向量机(Support Vector Machine,SVM)分类算法和极限学习机(Extreme Learning Machine,ELM)分类算法进行性能比较.仿真结果表明,卷积神经网络能够实现对无人机蜂群拦截方式的智能决策,且具有较高的决策精度.
Abstract
An intelligent decision-making algorithm based on convolutional neural networks is proposed to select the interception methods when facing a sudden attack from unmanned aerial vehicle(UAV)swarms.The theory of fuzzy mathematics is applied to assess the threat level of the UAV swarm's behavior factors.The trigger logic from the UAV swarm's behavior factors to the interception methods is established in order to generate the sample data.The structure of convolutional neural networks is designed to model the nonlinear relationship between the UAV swarm's behavior factors and the corresponding interception methods.On this basis,the model is trained and tested by using the sample data,and the performance of the proposed algorithm is compared with the Back Propagation(BP)neural network algorithm and the Support Vector Machine(SVM)classification algorithm.The simulation results show that the convolutional neural network can achieve the intelligent decision-making of the interception methods for UAV swarm and have a high accuracy.
关键词
反无人机蜂群/卷积神经网络/威胁评估/智能决策/模糊数学/隶属度函数/拦截触发逻辑Key words
anti-UAV swarm/convolutional neural network/threat assessment/intelligent decision-making/fuzzy mathematics/membership function/interception trigger logic引用本文复制引用
基金项目
国家自然科学基金(62271247)
国防基础科研计划资助项目(JCKY2021210B004)
航空科学基金(20220055052001)
南京航空航天大学前瞻布局科研专项资金()
出版年
2024
NETL
NSTL
万方数据