北京航空航天大学学报 ›› 2018, Vol. 44 ›› Issue (7): 1547-1553.doi: 10.13700/j.bh.1001-5965.2017.0556

• 论文 • 上一篇    下一篇

基于支持向量机的机场智能驱鸟决策

陈唯实, 闫军, 张洁, 李敬   

  1. 中国民航科学技术研究院 机场研究所, 北京 100028
  • 收稿日期:2017-08-31 出版日期:2018-07-20 发布日期:2018-07-25
  • 通讯作者: 陈唯实.E-mail:chenwsh@mail.castc.org.cn E-mail:chenwsh@mail.castc.org.cn
  • 作者简介:陈唯实 男,高级工程师。主要研究方向:低空空域安全监视、雷达目标检测与跟踪、机场安全运行技术。
  • 基金资助:
    国家自然科学基金委员会-中国民航局民航联合研究基金(U1633122);国家重点研发计划(2016YFC0800406)

Intelligent decision making for airport bird-repelling with support vector machine

CHEN Weishi, YAN Jun, ZHANG Jie, LI Jing   

  1. Airport Research Institute, China Academy of Civil Aviation Science and Technology, Beijing 100028, China
  • Received:2017-08-31 Online:2018-07-20 Published:2018-07-25

摘要: 为提高机场鸟击防范管理水平,实现探鸟雷达与多种驱鸟设备联动,提出一种基于支持向量机(SVM)的机场智能驱鸟决策方法。该方法包括训练和测试两部分。训练部分利用机场鸟类探测预警与驱赶联动系统获取的大量历史鸟情信息,结合专家知识,通过数据预处理与支持向量机训练,建立驱鸟策略分类模型;测试部分根据驱鸟实时智能决策结果,对驱鸟策略分类模型进行持续修正与优化。通过某机场的实测鸟情信息数据与若干驱鸟实例,证明驱鸟策略分类模型具有较高的决策正确率,并能够通过自身修正与优化应对各种新问题。本文方法针对实时鸟情信息,实现了多种驱鸟设备的优化组合,克服了驱鸟设备长期重复运行造成的鸟类对驱鸟设备的耐受性问题,极大改善了驱鸟效果。

关键词: 支持向量机(SVM), 机场, 驱鸟, 分类, 决策

Abstract: To impove the management of bird-strike avoidance at airport and realize the linkage of avian radar with multiple bird-repelling devices, an intelligent decision making method was proposed for airport bird-repelling based on support vector machine (SVM). The method includes two steps of training and testing. In the training step, the bird-repelling strategy classification model was established by data pretreatment and SVM training, which are combined with expert knowledge and large amount of historical bird information collected by the airport linkage system for bird detection, surveillance and repelling. In the testing step, the bird-repelling strategy classification model was continuously corrected and optimized according to the real-time intelligent bird-repelling strategy results. Through the real bird information data and several bird-repelling examples of a certain airport, it is demonstrated that the decision accuracy of bird-repelling strategy classification model is relatively high, and it can solve new problems by self correction and optimization. The proposed method achieves the optimized combination of multiple bird-repelling devices against real-time bird information with great improvement of bird-repelling effect, overcoming the tolerance of birds to the bird-repelling devices due to their long-term repeated operation.

Key words: support vector machine (SVM), airport, bird-repelling, classification, decision making

中图分类号: 


版权所有 © 《北京航空航天大学学报》编辑部
通讯地址:北京市海淀区学院路37号 北京航空航天大学学报编辑部 邮编:100191 E-mail:jbuaa@buaa.edu.cn
本系统由北京玛格泰克科技发展有限公司设计开发