AGENTFLY is a multi-agent system enabling large-scale simulation of civilian and unmanned air traffic. The system integrates advanced flight path planning, decentralized collision avoidance with highly detailed models of the airplanes and the environment.

Subsequent projects developed on top of the underlying AGENTFLY framework focus on large-scale civilian air traffic simulation or tactical missions of UAVs in urban areas.

Project videos:

(0:43, 43MB, audio)(4:11, 260MB, audio)(1:06, 66MB)(1:09, 256MB)(4:05, 209MB)(3:30, 242MB)


The project develops and integrated framework with a rich set of tools and techniques for the analysis and classification of maritime domain data, for creation of behavioral models of typical classes of vessels and prediction of their routes and actions, for planning of optimal routes of long range transport vessels and for risk assessment of potentially dangerous areas.

The project explores multi-agent systems can be used to improve maritime security, with particular focus on fighting maritime piracy. The ultimate objective of the project is to develop an integrated set of algorithmic techniques for maximizing transit security given the limited protection resources available. This is achieved by improving the coordination of the movement of merchant vessels and naval patrols, while taking into account the behavior of pirates. To this end, the project develops agent-based simulation of maritime traffic, piracy risk modeling and assessment algorithms and game-theoretic techniques for coordinated, risk-minimizing routing of transit traffic and navy patrols.

Vessel trajectories data Vessel trajectories data
Maritime traffic simulation Maritime traffic simulation
Piracy risk estimation Piracy risk estimation
Risk-minimizing transit and patrol routing Risk-minimizing transit and patrol routing


AgentPolis is a fully agent-based platform for modeling multi-modal transportation systems. It comprises a high-performance discrete-event simulation core, a cohesive set of high-level abstractions for building extensible agent-based models and a library of predefined components frequently used in transportation and mobility models. Together with a suite of supporting tools, AgentPolis enables rapid prototyping and execution of data-driven simulations of a wide range of mobility and transportation phenomena. We illustrate the capabilities of the platform on a model of fare inspection in public transportation networks.

London mobility simulation
London mobility simulation
Simulation events
Simulation events
In-station statistics
In-station statistics
Zone-based statistics
Zone-based statistics


Agent-based Computing for Intelligent Transport Systems

We explore a wide range of agent-based simulation, planning, coordination and game theoretic techniques that will underpin the future intelligent reflective multi-modal transport systems.

Agent-based transport simulation
Agent-based transport simulation
Group travel planner
Group travel planner
Fare inspection optimization
Fare inspection optimization
SuperHub project
SuperHub project


Nowadays, the state of the art in robotics reached a point when troops in military operations theatre routinely deploy unmanned robotic assets, be it unmanned aerial and ground vehicles, or unattended ground sensors supporting them in carrying our their tactical missions. Tele-operation of large numbers of such assets drastically raises costs of such operations in terms of costly and scarce expert human resources.

Tactical Agents is an ongoing cluster of U.S. Army/CERDEC-sponsored research projects aimed at development and evaluation of multi-agent coordination techniques supporting information-collection missions in tactical urban warfare. The core objective of the project cluster is development and evaluation of techniques allowing deployment of teams of autonomous assets which are capable of autonomous coordination and teamwork with the goal to carry out the high-level mission assigned to the team. In particular, we are developing agent-based coordination, planning and game-theoretic techniques for heterogeneous multi-agent teams allowing them to carry out various information collection tasks, such as reconnaissance, convoy and perimeter protection, pursuit-evasion of smart targets, surveillance, safe area traversal, etc.


Ground and aerial teams towards tactical missions.

Short Horizon Trajecotry Planning for UGVsGround Common Operational PictureUAV tracking of ground convoy


AgentDrive is a consolidated simulation framework for realistic vehicles simulation enabling testing of agent-based algorithms for route planning, navigation, cooperative driving, traffic optimization, and others vehicle coordination and cooperation methods.


Collective robotics

  • Contact Person: Milan Rollo
  • Key Partner / Sponsorship: CAK
  • Years: 2006 -

This work was done in cooperation with the Intelligent and Mobile Robotics Group.

Collective robotics becomes a very important area of research nowdays. In the past robots were usually teleoperated or only a single autonomous robot was operating in the environment. At present as robots become more cost-effective, easier to maintain and the communication infrastructure has improved, situation drives towards wider use of robotic teams.

These teams may consist of number of heterogenous mobile robots, static sensors and even a humans. As humans more and more rely on hardware robots, it is necessary to develop robust and reliable control algorithms.

Software agents and multiagent systems are often used to simulate the behavior of the distributed systems. These simulations allow to adjust parameters of the simulated environments, control the communication accessibility, simulate high number of robots (scalability tests) and are faster and cheaper then real world experiments.

Within the collective robotics research effort we work on following projects:

  • NAIMT (Naval Automation and Information Management Technology) project provides simulation environment for the underwater surveillance and propose architecture of control part of autonomous robot capable of efficient operation in such environment. Besides this the algorithms for decentralized coordination within a group of such robots and video stream transmission path planning were developed. Development of these algorithms was necessary due to the nature of environment, where individual robots can become temporarily inaccessible.
  • Center of Applied Cybernetics

Software simulation however can not be so complex to cover all situations and problems that can occure in real world environment. In cooperation with the Intelligent Mobile Robotics group (part of the Gerstner lab) we are working on the development of a framwork that will allow us to verify the functionallity of algorithms that control the team of autonomous robots. These algorithms are originally developed for multiagent systems and their deployment on hardware robots should allow us to further improve their stability and robustness. In this project we are currently working on problems of collision avoidance and multirobot exploration.

Collective Robotics Screenshots

Collective roboticsCollective roboticsCollective roboticsCollective robotics


  • Contact Person: Martin Rehák
  • Key Partners / Sponsorship: European Research Office of the US Army and Czech Ministry of Education
  • Years: 2006 -
  • CAMNEP Project Page

Joint research effort with Institute of Computer Science, Masaryk University. under Contract No. N62558-07-C-0001 and Czech Ministry of Education grants 1M0567, 6840770038 (CTU) and 6383917201 (CESNET)

Project CAMNEP consists of two activities:

  • The collaborative agent-based Network Behavior Analysis system which has been deployed on real networks and is able to perform on-line surveillance of gigabit-speed links and
  • The prototype of Intrusion protection System running inside a network simulator. Use of the simulation allows us to investigate the problems related to autonomous response to self-propagating malware and Denial of Service attacks in local networks. The essential system components are:
    • multi-agent simulation, used to implement the simulated network hosts and create traffic with natural characteristics;
    • trust, which is used to classify the traffic as either legitimate/malicious using the feedback from host-based IDS/IPS systems;
    • reflective agents, located on network nodes (routers) used to implement the response and enforce automatically created filtering policies;
    • distributed learning, which has been applied to distribute the intrusion detection process into between several peer nodes and
    • Extended Contract Net Protocol used to efficiently distribute the filters between the nodes with filtering capability.




  • Contact Person: Viliam Lisý
  • Key Partners / Sponsorship: US Air Force
  • Years: 2007 -

U.S. Air Force Research Lab Project Deep Adversary (DeepA) investigates planning and adversarial reasoning in realistic games involving multiple players of different types, strength and goals. As a part of the project, a sound formal model of adversarial behavior is being developed. The model forms a basis for advanced techniques for modelling, detecting and predicting adversarial behavior.

In the project, we focus on three basic applications of the opponent model: execution, utilization, and generation. The first application is crucial for the other ones. We develop intelligent agents that can be parameterized by the opponent model and play based on the easily adjustable characteristics.

When playing against these players, we investigate how knowing the opponent model the other players execute can help to improve performance of the agents in the game. We utilize the information about the opponent model in the planning process of the agents as well as in prediction the opponents’ actions.

In order to complete the framework, we are using machine learning and regression techniques to generate models of the opponents based on the observations made in past games or earlier in the history of the game, where we want to utilize the model.

The techniques are being evaluated using a sophisticated A-globe-based game testbed, which allows simulating complex games in realistic environments.

Some of the results of this projects were developed in collaboration with CMU in the project of the Czech Ministry of Education, Youth and Sports under grant ME09053.


DeepA DeepA DeepA DeepA


AGLOBE is an agent platform designed for testing experimental scenarios featuring agents’ position and communication inaccessibility, but it can be also used without these extended functions. The platform provides functions for the residing agents, such as communication infrastructure, store, directory services, migration function, deploy service, etc. Communication in A-globe is very fast and the platform is relatively lightweight.

AGLOBE is suitable for real-world simulations including both static (e.g. towns, ports, etc.) and mobile units (e.g. vehicles). In such case the platform can be started in extended version with Geographical Information System (GIS) services and Environment Simulator (ES) agent.


Alite - a multipurpose toolkit (not only) for Agent Oriented Prototyping.

Alite is a software toolkit helping with particular implementation steps during construction of multi-agent simulations and multi-agent systems in general. The goals of the toolkit are to provide highly modular, variable, and open set of functionalities defined by clear and simple API. The toolkit does not serve as a pre-designed framework for one complex purpose, it rather associates number of highly refined functional elements, which can be variably combined and extended into a wide spectrum of possible systems.

ATG projects using Alite:

Malware Capture Facility

This is a research project to capture, monitor, analyze and publish long-lived real malware network traffic. The malware is executed with only two restrictions on the output traffic: a limit on the bandwith and the interception of spam. The most important characteristic of this project is the execution of malware during long periods of time, that can go up to several months. The traffic is stored in pcap files, pre-process, analyzed, labeled and made public for the research comunity. The preprocessing includes RRD files with the history of traffic shape, bidirectional Argus flows (both the binary file and the text file), web logs for all the web traffic and a dns report among others. The labels are manually generated by a group of security experts and added to both Argus files and to the weblogs.


The datasets created in this facility are used in the research projects of botnet behavior analysis and anomaly detection. 

If you use these dataset for your own research please reference it accordingly. Also consider a colaboration with the project to make the dataset better.


The researcher in charge of this project is PhD student Sebastian Garcia.

sebastian.garcia at



  • Contact Person: Milan Rollo
  • Key Partner / Sponsorship: Technology Agency of the Czech Republic
  • Years: 2011 - 2013
  • SafeFly project page

Goal of the SafeFly project is to develop a telemetric system which will increase operation safety in a light sport aircraft domain. This system will be capable of autonomous detection of other aircraft within its proximity based on (i) communication with other airplanes equipped with the same system and (ii) detection of other airplanes using the radio-telemetric device. On-board "glass cocpit" control system will be extended with collision detection algorithms predicting these situations in 5-10 minutes time horizon. This detection will be based on mutual exchange of flight trajectories between airplanes in case of airplanes equipped with the same telemetric systems or based on trajectory prediction in other cases. Possible solution of the collision situation will be presented to the pilot using a graphical user interface. Features of the proposed system are being verified by a series of flight tests. System is prepared for integration of an autopilot which will allow to follow modified collision-free trajectories autonomously.

Project is a joint effort of CTU, F AIR flight school and TL Elektronic.

Test flight with Zaon XRX deviceTesting airplane on groundTest flight