北京航空航天大学学报 ›› 2016, Vol. 42 ›› Issue (4): 780-788.doi: 10.13700/j.bh.1001-5965.2015.0296

• 论文 • 上一篇    下一篇

面向多威胁的无人机智能目标跟随策略设计

祁晓明, 魏瑞轩, 周凯   

  1. 空军工程大学航空航天工程学院, 西安 710038
  • 收稿日期:2015-05-11 修回日期:2015-08-06 出版日期:2016-04-20 发布日期:2016-04-29
  • 通讯作者: 魏瑞轩, Tel.: 17082452676 E-mail: 2856402009@qq.com E-mail:2856402009@qq.com
  • 作者简介:祁晓明 男,博士研究生。主要研究方向:多无人机协同搜索决策方法。 E-mail: fancyxiaoming@163.com;魏瑞轩 男,博士,教授,博士生导师。主要研究方向:多无人机自主协同控制。 Tel.: 17082452676 E-mail: 2856402009@qq.com;周凯 男,博士研究生。主要研究方向:多无人机协同控制方法。 E-mail: kaigemima@163.com
  • 基金资助:
    国家自然科学基金(61573373);航空科学基金(20135896027)

Intelligent target following strategy design for UAV against multi-threats

QI Xiaoming, WEI Ruixuan, ZHOU Kai   

  1. Aerospace Engineering College, Air Force Engineering University, Xi'an 710038, China
  • Received:2015-05-11 Revised:2015-08-06 Online:2016-04-20 Published:2016-04-29
  • Supported by:
    National Natural Science Foundation of China(61573373);Aeronautical Science Foundation of China (20135896027)

摘要: 随着无人机(UAV)一体化作战的不断发展,无人机在搜索到运动目标之后,需要立即转入跟随模式,考虑到战场环境的复杂性,研究了在多个威胁源条件下无人机跟随运动目标的问题,为了保证无人机的安全性以及跟随目标的精确性,提出了一种基于决策树的无人机智能目标跟随策略。首先对威胁概率图(TPM)进行建模;然后采用几何图法及任务优先级生成不同的规则,建立相应的决策树,并设计了不同规则下无人机飞行航向及速度指令;最后通过仿真验证所提出的智能目标跟随策略的有效性。

关键词: 威胁建模, 无人机(UAV), 梯度下降, 规则, 决策树, 目标跟随, 智能导航

Abstract: With the development of integrative combat, the task will be converted into target following mode when unmanned aerial vehicle (UAV) finishes searching ground moving target. Complex combat environment is considered, and the problem of UAV following moving target in the condition of multi-threats source are studied. An intelligent target following strategy based on decision trees is proposed in order to ensure the security of UAV and the precision of target following. Firstly, threat probability map (TPM) model was established. Secondly, the problem how to solve the minimum of TPM was researched. Then, the different rules were generated on the basis of the different priorities for object tasks by geometric method; the complete decision trees were established; the heading and speed commands of UAV were generated for different rules. Finally, simulation results demonstrate the validity of the proposed method.

Key words: threat modeling, unmanned aerial vehicle (UAV), gradient descent, rules, decision trees, target following, intelligent navigation

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