Back to the main AGENTFLY page

Tactical AGENTFLY is an ongoing U.S. Army-sponsored research project developing agent-based coordination and planning techniques for multi-UAV information collection missions with special emphasis on complex urban environments.

Mission schemeUrban area modeled after real-world data

Problem Specification

Unmanned aerial vehicles (UAVs) have been increasingly deployed for aerial surveillance. The present UAVs are able to perform most of their flight control level tasks autonomously and there is a strong demand for intelligent systems providing an interface for high-level commands. The project aims to provide such high-level control, allowing the operator to communicate with a fleet of UAVs on a high level in terms of collection tasks and their results, with detailed allocation of the tasks to individuals UAVs and planning of their optimum trajectories performed automatically. The following basic information collection tasks are supported:

  • persistent surveillance - provide and maintain the information about the target area, keeping the average information age minimum
  • tracking - provide uninterrupted information about the selected mobile ground target (or a group of targets)

In both cases, the control algorithms have to respect UAV’s flight model constraints, sensory constraints and maintain collision-free trajectories. The development of such robust and effective multi-agent control mechanisms for the individual information tasks is an important objective of the project. The ultimate objective, however, is the development of methods that can efficiently and autonomously address heterogeneous combinations of the basic tasks (termed mixed information collection). The primary performance metric is the difference between the true state of the tracked targets/area and the state reported by the UAVs employing the proposed control mechanisms. All metrics are evaluated both on synthetic experiments and on a realistic simulation of a search-and-capture operation in a complex urban environment.

Realistic model of the UAV's on-board sensorInternal visibility model used for calculating surveillance trajectories

Navigation of ground entitiesOptimized surveillance trajectory


The project has been producing a number of results of formal, algorithmic and software nature. Specifically, we have developed and are continuously extending the following:

  • a detailed model of the urban environment based on real-world data
  • a realistic model of UAV’s on-board sensors including the effects of limited sensor range and sensor occlusions
  • decentralized robust control algorithms for persistent surveillance and tracking of mobile ground targets
  • autonomous integrated information collection mechanisms for efficiently allocating a mix of heterogeneous information collection tasks between a fleet of UAVs
  • graphical C2 interface for specifying information collection tasks and inspecting their results
  • a detailed simulation of a search-and-capture mission in an urban environment, including executable behavioral models of several classes of ground entities (armed forces, civilians, militants etc.)

All developed component are integrated with the Core AGENTFLY technology providing support for the base UAV functionality (detailed simulation of UAV in-flight operation, optimized flight path planner respecting aircraft’s physical constraints, robust decentralized utility-based collision avoidance mechanisms etc.)

C2 PanelOperational Picture recency map

Project Videos

(4:05, 209MB)


Dr. Michal Pechoucek (principal investigator), Dr. Michal Jakob (project leader), Eduard Semsch, Dusan Pavlicek and Vojtech Elias (team members)

Back to the main AGENTFLY page



  • David Sislak and Premysl Volf and Michal Pechoucek and Christopher T. Cannon and Duc N. Nguyen and William C. Regli: Multi-Agent Simulation of En-Route Human Air-Traffic Controller. In Proceedings of the Twenty-Fourth Innovative Appications of Artificial Intelligence Conference. Toronto, Canada: AAAI Press, 2012, p. 2323-2328. ISBN 978-1-57735-568-7.
    BiBTeX | PDF (382)
  • David Sislak and Premysl Volf and Dusan Pavlicek and Michal Pechoucek: AGENTFLY: Multi-Agent Simulation of Air-Traffic Management. In 20th European Conference on Artificial Intelligence. Montpellier, France: IOS Press, 2012. ISBN 978-1-61499-097-0.
    BiBTeX (383)


  • David Sislak, Premysl Volf, Stepan Kopriva and Michal Pechoucek: AgentFly: Scalable, High-Fidelity Framework for Simulation, Planning and Collision Avoidance of Multiple UAVs. In Sense and Avoid in UAS: Research and Applications. Wiley: John Wiley&Sons, Inc., 2012, p. 235-264.
    BiBTeX (315)