Back to the main AGENTFLY page

AGENTFLY is a software prototype of multi-agent technology deployment in distributed air traffic control of heterogeneous groups of aerial vehicles supporting the free flight concept.

All aerial vehicles are modeled by autonomous software agents. Each vehicle/agent is responsible for its own flight operation. The high-level mission of each vehicle is specified by an arbitrary number of waypoints. The operation is tentatively planned before take-off without consideration of possible collisions with other flying objects. During the flight, the agents detect mutual future conflicts in their flight plans and engage in peer-to-peer negotiation aimed at sophisticated re-planning in order to avoid the conflicts and maintain collision-free trajectories.

AGENTFLY is built on top of the AGLOBE multi-agent platform.

AGENTFLY provides:

  • distributed model of flight simulation and control,
  • free-flight concept
  • flight path planning
  • flexible collision avoidance architecture
  • scalability
  • advanced 2D/3D visualization,
  • realtime measurement and data collection for subsequent evaluation and analysis


On the conceptual level, AGENTFLY simulation consists of two main components: a synchronous simulation of the virtual environment and asynchronous simulated aerial assets. Each asset is simulated as a separate computational entity which can run in a distributed manner on individual computers. The simulation of the whole virtual environment is conceptually a single computational unit but it can also be distributed over multiple computers which enables to reduce to overall computational load.

Flight Path Planning

Each aerial vehicle (software agent) can invoke its own flight path planner. The generated trajectory is smooth and continuous, consisting essentially of a sequence of straight elements and turns. The trajectory has to respect the physical restrictions of the particular asset, such as the minimal turning radius, maximal climbing angle etc. The algorithm used for the flight path planning is Accelerated A* which was developed specifically for this purpose. During the expansion of the search-space states, the algorithm determines the length of the expansion step dynamically, depending on the distance of the expanded state from any obstacles.

Collision Avoidance

Core AGENTFLY provides a number of collision avoidance (CA) methods, each implemented as a separate plug-in module, which can be activated and deactivated in runtime, depending on the current situation. AGENTFLY supports both cooperative and non-cooperative mode of CA. The non-cooperative mode can be employed when the other vehicle is unable or unwilling to communicate. A particular vehicle can use different CA modes towards other aerial assets (the multi-layer architecture).

Three cooperative CA modules are currently available:

  • rule-based (RBCA): based on visual flight rules specified by FAA
  • iterative peer-to-peer (IPPCA): pair-wise negotiation
  • multi-party (MPCA): group-wise, semi-centralized negotiation

IPPCA and MPCA take advantage of negotiation in order to optimize the resulting flight plan with respect to a pre-defined criterion, such as minimal length, minimal fuel consumption etc.

Visualization and Data Analysis

AGENTFLY also contains an advanced 2D/3D visualization module which provides realtime visual feedback and allows interactive monitoring of the currently running simulation.

In-built data loggers enable automatic data collection for subsequent analysis and evaluation.


(3:35, 140MB, audio)
(1:27, 10MB, audio)
(0:47, 38MB, audio)
(1:10, 65MB, audio)
(1:14, 29MB, audio)
(0:43, 43MB, audio)
(1:48, 86MB)
(1:39, 14MB)
(1:11, 14MB)
(0:46, 13MB)
(0:25, 6MB, audio)
(1:59, 32MB)
(1:13, 13MB)
(1:02, 25MB, audio)
(0:43, 26MB, audio)
(0:36, 6MB)
(1:01, 12MB, audio)
(2:17, 87MB, audio)
(4:11, 260MB, audio)
(1:56, 32MB)


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)