Multi-UAV Cooperative Search Using an Opportunistic Learning Method

[+] Author and Article Information
Yanli Yang1

Large Power Systems Division, Caterpillar Inc., Peoria, IL 61656-1875Yangyanl@gmail.com

Marios M. Polycarpou

Department of Electrical and Computer Engineering, University of Cyprus, Nicosia 1678, Cyprus

Ali A. Minai

Department of Electrical and Computer Engineering and Computer Science, University of Cincinnati, Cincinnati, OH 45221-0030


Corresponding author.

J. Dyn. Sys., Meas., Control 129(5), 716-728 (Jan 10, 2007) (13 pages) doi:10.1115/1.2764515 History: Received April 16, 2006; Revised January 10, 2007

The control of networked multivehicle systems designed to perform complex coordinated tasks is currently an important and challenging field of research. This paper addresses a cooperative search problem where a team of uninhabited aerial vehicles (UAVs) seeks to find targets of interest in an uncertain environment. We present a practical framework for online planning and control of a group of UAVs for cooperative search based on two interdependent tasks: (i) incrementally updating “cognitive maps” used as the representation of the environment through new sensor readings; (ii) continuously planning the path for each vehicle based on the information obtained through the search. We formulate the cooperative search problem and develop a decentralized strategy based on an opportunistic cooperative learning method, where the emergent coordination among vehicles is enabled by letting each vehicle consider other vehicles’ actions in its path planning procedure. By using the developed strategy, physically feasible paths for the vehicles to follow are generated, where constraints on aerial vehicles, including physical maneuverabilities, are considered and the dynamic nature of the environment is taken into account. We also present some mathematical analysis of the developed search strategy. Our analysis shows that this strategy guarantees a complete search of the environment and is robust to a partial loss of UAVs. A lower bound on the search time for any strategy and a relaxed upper bound for the proposed strategy are given. Simulation results are used to illustrate the effectiveness of the proposed strategy.

Copyright © 2007 by American Society of Mechanical Engineers
Your Session has timed out. Please sign back in to continue.



Grahic Jump Location
Figure 4

Illustration of the three orientations related to pv (as arrows) and five target cells

Grahic Jump Location
Figure 5

Search performance for 15 UAVs. The CL and DL algorithms use 100 steps of learning.

Grahic Jump Location
Figure 6

Search efficiency: The increments on each bar indicate the number of search steps needed to reduce uncertainty by 50%, 75%, 90%, and 98%. The system has 15 UAVs searching a 20×20 environment.

Grahic Jump Location
Figure 7

Search performance for 15 UAVs. The CL and DL algorithms use 100 steps of learning.

Grahic Jump Location
Figure 8

Search paths for UAVs in a five UAV system using greedy search. Note that many paths overlap, reducing search efficiency.

Grahic Jump Location
Figure 9

Search paths for UAVs in a five UAV system using OCL with π=0.5

Grahic Jump Location
Figure 3

Illustration of an unconstrained path, a constrained path, and a path obtained by the turn-straight-turn procedure from pu to pv

Grahic Jump Location
Figure 2

Example orientation transition choices for UAVs

Grahic Jump Location
Figure 1

A general cooperative search framework




Some tools below are only available to our subscribers or users with an online account.

Related Content

Customize your page view by dragging and repositioning the boxes below.

Related Journal Articles
Related eBook Content
Topic Collections

Sorry! You do not have access to this content. For assistance or to subscribe, please contact us:

  • TELEPHONE: 1-800-843-2763 (Toll-free in the USA)
  • EMAIL: asmedigitalcollection@asme.org
Sign In