北京航空航天大学学报 ›› 2019, Vol. 45 ›› Issue (2): 347-356.doi: 10.13700/j.bh.1001-5965.2018.0230

• 论文 • 上一篇    下一篇

未知环境下无人机集群协同区域搜索算法

侯岳奇1,2, 梁晓龙1,2, 何吕龙1,2, 刘流1,2   

  1. 1. 空军工程大学 国家空管防相撞技术重点实验室, 西安 710051;
    2. 空军工程大学 陕西省电子信息系统综合集成重点实验室, 西安 710051
  • 收稿日期:2018-04-25 出版日期:2019-02-20 发布日期:2019-03-04
  • 通讯作者: 梁晓龙 E-mail:afeu_lxl@sina.com
  • 作者简介:侯岳奇男,硕士研究生。主要研究方向:航空集群智能决策;梁晓龙男,博士,教授,硕士生导师。主要研究方向:航空集群指挥与控制、智能系统、空管智能化;何吕龙男,博士研究生。主要研究方向:航空集群编队控制;刘流男,硕士研究生。主要研究方向:航空集群编队控制。
  • 基金资助:
    国家自然科学基金(61472443,61703427);陕西省自然科学基础研究计划(2017JQ6035)

Cooperative area search algorithm for UAV swarm in unknown environment

HOU Yueqi1,2, LIANG Xiaolong1,2, HE Lyulong1,2, LIU Liu1,2   

  1. 1. National Key Laboratory of Air Traffic Collision Prevention, Air Force Engineering University, Xi'an 710051, China;
    2. Shaanxi Province Lab. of Meta-synthesis for Electronic & Information System, Air Force Engineering University, Xi'an 710051, China
  • Received:2018-04-25 Online:2019-02-20 Published:2019-03-04
  • Supported by:
    National Natural Science Foundation of China (61472443,61703427); Natrual Science Basic Research Plan in Shaanxi Province of China(2017JQ6035)

摘要: 针对无人机集群在无先验信息的未知环境中协同搜索的问题,提出了一种以覆盖率为实时搜索奖励的无人机集群协同区域搜索算法。首先建立覆盖分布地图(CDM)来描述任务环境,并采用Hadamard积实现CDM的快速更新,继而基于CDM计算覆盖率来定量描述实时搜索效果。将无人机集群视为一个控制系统,基于分布式模型预测控制理论建立系统的预测模型,并将预测周期内最大覆盖率增量设为奖励函数,采用差分进化算法进行求解,得到最优解作为系统的最优输入。仿真结果表明,所提算法能够对区域进行覆盖搜索,在出现突发情况时,覆盖率远高于平行搜索方法。

关键词: 未知环境, 无人机集群, 协同搜索, Hadamard积, 覆盖率, 分布式模型预测控制

Abstract: Aimed at the problem of cooperative search for UAV swarm in an unknown environment without prior information, a cooperative area search algorithm for UAV swarm with coverage rate as real-time search rewards is proposed. First, coverage distribution map (CDM) is established to describe the mission area, and the rapid update of CDM is realized by using Hadamard product. Then, the coverage rate is calculated based on CDM to describe the search results quantitatively. Considering UAV swarm as a control system, a predictive model of the system is established based on the distributed model predictive control theory, and the maximum increment of coverage rate in the predictive period is determined as a reward function. The optimal solution, as the optimal input of system, is obtained by differential evolution algorithm. Simulation results demonstrate that the proposed algorithm can complete the coverage and search of region effectively. In the event of emergencies, its area coverage rate is much higher than that of the parallel search method.

Key words: unknown environment, UAV swarm, cooperative search, Hadamard product, coverage rate, distributed model predictive control

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