Collective Perception Using UAVs: Autonomous Aerial Reconnaissance in a Complex Urban Environment

Sensors

卷:20

Dalibor Procházka¹  Karel Šilinger²  Petr Stodola³  Jan Drozd⁴  Jan Hodický⁵ 

[1] Centre for Security and Military Strategic Studies, University of Defence, 66210 Brno, Czech Republic;

[2] Department of Fire Support, University of Defence, 66210 Brno, Czech Republic;

[3] Department of Intelligence Support, University of Defence, 66210 Brno, Czech Republic;

[4] Department of Tactics, University of Defence, 66210 Brno, Czech Republic;

[5] NATO Headquarters Supreme Allied Commander Transformation, Norfolk, VA 23551, USA;

关键词: autonomous aerial reconnaissance;    collective perception;    metaheuristic algorithm;    simulated annealing;    experiments;   

DOI  :  10.3390/s20102926

来源: DOAJ

【 摘 要 】

This article examines autonomous reconnaissance in a complex urban environment using unmanned aerial vehicles (UAVs). Environments with many buildings and other types of obstacles and/or an uneven terrain are harder to be explored as occlusion of objects of interest may often occur. First, in this article, the problem of autonomous reconnaissance in a complex urban environment via a swarm of UAVs is formulated. Then, the algorithm based on the metaheuristic approach is proposed for a solution. This solution lies in deploying a number of waypoints in the area of interest to be explored, from which the monitoring is performed, and planning the routes for available UAVs among these waypoints so that the monitored area is as large as possible and the operation as short as possible. In the last part of this article, two types of main experiments based on computer simulations are designed to verify the proposed algorithms. The first type focuses on comparing the results achieved on the benchmark instances with the optimal solutions. The second one presents and discusses the results obtained from a number of scenarios, which are based on typical reconnaissance operations in real environments.

【 授权许可】

Unknown