Autonomous systems from the Faculty of Electrical Engineering at CTU have already proven in many cases that they are capable of being deployed in demanding conditions in both the defense and civilian sectors. “Dual use is a reality. Most of the technologies we use today, and the vast majority of robots in Ukraine, are of civilian origin,” says Jan Faigl from the Laboratory of Computational Robotics. “Our ambition is to develop technologies that have a real impact – whether in the field of defense, security or critical infrastructure,” describes Petr Páta, Dean of the Faculty of Electrical Engineering at CTU. Companies from the listed areas are interested in cooperating with the projects, and according to Páta, such research strengthens the technological sovereignty of the Czech Republic. “It turns out that every crown you put into such activities will come back to you many times over. Research shows that two to four times, depending on the nature of the technology,” he says.
CTU showed its research projects at the Defence Research Day event, where TAROS 6×6 was on display, among others. The autonomous vehicle is being developed in cooperation with VOP CZ, a state-owned enterprise that verifies the functionality of advanced unmanned platforms in demanding conditions and their portability into practice. The vehicle has advanced control and navigation, a waterproof design, four electric motors, a 48V architecture, 4 turning modes for different types of terrain, and hydraulic attachment points for integrating equipment. Its payload is up to 1,000 kg and towing capacity is 750 kg, and it travels at a speed of 10 to 15 km/h, depending on the terrain. It fits into the rapidly developing European trend of unmanned systems. It is designed for a range of tasks, from reconnaissance and monitoring to logistical support to deployment in crisis situations. CTU is the lead partner in the European IMAS 2 project, which brings together 14 countries and is the European Commission’s flagship project in the field of robotics. “The dialogue with the end user is absolutely crucial. I am very pleased that the Ministry of Defense and the General Staff are involved, because this dialogue with them is the most important thing for us to be able to define the scenarios that are relevant,” explains Faigl. CTU says that researchers from the Department of Computer Graphics and Interaction are developing technologies that fundamentally transform the preparation and management of operations in security and defense scenarios. They are working on using virtual reality for training bomb disposal officers, where complex situations related to the identification and disposal of unexploded ordnance can be safely simulated. “It’s not just about training in cutting blue and red wires, as you know from movies, but also about disarming unexploded ordnance. It’s a very dangerous job that needs to be supported in some way when training new specialists,” explains David Sedláček from the department. The virtual environment allows for training in decision-making processes, working with detection tools, and controlling specialized robots, to a degree that would be difficult to implement in real conditions in terms of time, money, and safety. “It’s not just about manual skills, but primarily about following processes, the correct approach to the ordnance, its identification, and solving the problem so that nothing is destroyed, if possible, and most importantly, the operator is not injured,” adds Sedláček.
Another project, HOLO-Swarm, created in collaboration with the Multirobotic Systems Group and the company QuaternAR, brings the concept of a holographic control room and allows you to display the situation in the field as an interactive 3D model, into which data on the movement of units, for example drone swarms, their status or planned routes are projected. Users can share this data, analyze it and directly influence the course of the operation in real time. “Imagine that you have a large table in front of you – but it is not real here. Thanks to this, we can place it in any room. We can call several people at the same time to consult, for example, the plan of a rescue operation in the context of a given environment,” describes Sedláček. The system is already connected to the army’s command and control systems and allows you to display and change the trajectories of drones in real time. “Augmented reality glasses are much lighter than physical tables, so we can deploy such an overview basically anywhere,” he adds. The AR Rescue system consists of a helmet with a visor that shows augmented reality to the user - special warnings are displayed on it. The system is intended for firefighters, military units and other units responding to CBRN environments, i.e. in situations associated with chemical, biological, radiological or nuclear threats. It connects vital signs sensors, detection of hazardous substances and augmented reality to support response units. Its aim is to provide responders with key information in real time about their own condition, the condition of their colleagues and the dangers in the surroundings, without unnecessary congestion or distraction. "The visor is a transparent glass that folds down. It allows unit members to see, for example, through a wall if someone is there - which can be advantageous if a wounded person needs to be located quickly," explains Miroslav Bureš from the Laboratory of Intelligent Systems Testing. The response commander can monitor more detailed data in a specialized application. The system was designed to work even with poor connectivity and to distract its users as little as possible.
But cyber security is also an important component of the defense industry. “Today, it’s not just about who can penetrate the system, but who can influence your thinking,” comments Brigadier General Václav Žid, Deputy Director of Military Intelligence and a graduate of the Czech Technical University’s Faculty of Engineering and Technology, who was inducted into the Cybersecurity Hall of Fame in 2023. The ALICE system uses AI to map the spread of information on social networks, identify their sources, and detect coordinated or automated campaigns, thereby allowing people to systematically “read” the space. It uses a combination of machine learning, language models, and statistical methods to track how information is created, how it spreads between accounts, and how it changes over time. The map then allows, among other things, to identify influential actors. “We are trying to find out how information flows through the information space – who creates it, who forwards it, who amplifies it, and who the key consumers are,” explains Associate Professor Tomáš Pevný from the Center for Artificial Intelligence. The system can predict whether the message will be amplified. “We need to know if the message will clutter the information space, and if necessary, react – for example, to release a counter-message that will prepare the population for disinformation,” he adds. The project is being created in cooperation with the company Gerulata, which integrates the developed methods into tools for security institutions and significantly speeds up the analysis of the information space, which would otherwise take weeks or months. Autonomous drones are strongly represented at CTU, the technology allows them to fly in an environment without GNSS, other systems ensure the coordination of swarms. “The development of drones at this time requires autonomy,” says Petr Štěpán from the Multi-Robot Systems Group. According to CTU, the conditions of modern conflicts are leading to development towards autonomy, faster response to changes in the environment and securing robotic platforms against takeover or misuse. Drones build a map of the environment using a camera and localize themselves based on it. Thanks to this, they can also operate indoors. Researchers are working on drones inspired by bird movements and those that can capture an enemy drone without the machine endangering the space below it. They use their own optical communication to coordinate drone swarms. “It is very resistant to interference and at the same time resistant to detection,” adds Štěpán. The project is being created in cooperation with BUT Brno, the University of Defense and several private companies. Research in this area focuses not only on the flight itself and the coordination of multiple machines, but also on zero trust principles, secure on-board computers, flight control units and secure communication. Experts also showed a system for acoustic detection and localization of gunshots, which uses autonomous sensor units equipped with microphones and a central evaluation unit with a software application. The measured signal is processed by advanced algorithms and classifiers using AI so that it is possible to distinguish gunshots from ambient noise and minimize false alarms. “We can detect an event and locate the source of the shooting, but we can also use artificial intelligence to determine what weapon the shot came from,” says Jan Holub from the Department of Measurement. The database for AI training is being created on its own so that the system is not dependent on unreliable data from elsewhere. On the other hand, experts have also come up with a handheld GPS interference detector that allows detecting local interference in the L1 GPS and Galileo E1 bands, both in omnidirectional and directional modes. The device is used by the Czech Telecommunications Office for measuring and assessing complaints about signal interference. The technology can also locate members of the IZS and security forces in complex spaces and in places without GNSS, while the systems can be worn in shoes.
Robots designed for complex terrain are receiving special attention at CTU. Researchers focus on situations where the robot must estimate what it can afford, how to move safely, and how to react to terrain or flow that cannot be easily described in advance. “For a ground robot, this complex environment is represented by uneven terrain, vegetation, sand, mud,” explains Associate Professor Karel Zimmermann from the Data for Robots and Autonomous Systems group. For drones, the problem is predicting the trajectory and wind around obstacles. Helhest was the first robotic spin-off from CTU, which can function even after tipping over, move without GPS based solely on passive sensors, and aims to be used in both civilian and military scenarios. “It is a three-wheeled all-terrain robot that overcomes obstacles up to half a meter high, climbs 47 degrees, reaches speeds of 20 km/h even on terrain and can carry a load of around 100 kilograms. It offers capabilities that are better than what is currently available on the global market – and is relatively inexpensive,” describes Zimmermann. The researchers focus on situations where complex systems used in critical infrastructure, IZS or defense face outages, interference, partial damage or cyberattack. The methods they presented rely on intentionally simple, explainable models and transparent AI algorithms so that risky situations can not only be detected, but also understood and explained to the operator. Some problems are detected by artificial intelligence agents before they happen. "We used to think that the life cycle of technologies was years, maybe even decades. But now it changes in a time frame of months, maybe more like weeks," describes Brigadier General Žid.