The solution handles BVLOS (Beyond Visual Line of Sight) flights, automatic 3D LiDAR scanning, detection of key infrastructure elements, and multi-robot mission planning in real terrain.
Safer work, unified data, and a lower carbon footprint
"From the approximate GPS positions of the pylons, our drones independently map the environment with LiDAR, recognize insulators and the geometry of the high-voltage pylon, and automatically plan the imaging locations and the entire mission for multiple drones. The result is faster, safer, and scalable inspection," says Dr. Robert Pěnička from the Multi-robot Systems (MRS) Group at the Department of Cybernetics, FEE CTU. The technology thus reduces the risks for field workers who have to climb pylons in current inspections. This significantly reduces the time needed to inspect long sections of the network, and the energy distributor obtains uniform data suitable for predictive maintenance, which helps prevent outages. Another benefit is a lower carbon footprint—there is no need to use helicopters, and there are fewer trips.
From research to real-world deployment, AI helps
HIVE represents a qualitative shift from previous academic experiments (e.g., within the European Aerial-Core project) towards practical deployment. The system was designed to automatically evaluate what and where to scan after on-site 3D scanning – typically insulators from two angles, the base of the pylon, and the corridor along the line. It then plans the inspection for two or more drones and can adjust the plan during the flight according to the situation. Compatibility is also important: the MRS planner prepares missions both for research drones equipped with LiDAR and cameras that can map the environment, and for commercial drones operated by operators that scan according to a pre-prepared map. E.ON has developed an operator application that allows simultaneous control of multiple drones and bulk uploading of prepared missions, allowing one operator to oversee the entire fleet.
"The most challenging part was integration – connecting scanning, processing point clouds from LiDAR, detection, and safe planning so that the system would be robust and collision-free even in real-world conditions," adds Dr. Pěnička. During testing, drones took off and landed from several locations; the typical flight time was about 30 minutes per battery, and depending on the mission profile, it was possible to cover a distance of approximately two kilometers in a single flight. The images were immediately sent to a server for further processing. Artificial intelligence also helps, which is mainly used to understand LiDAR data—in detecting and orienting pylons and locating insulators—and is followed by automatic mission planning; at the level of defect recognition in images, the project plans to connect to the partners' operational tools.
New opportunities for power engineering and industry
E.ON and EG.D are now evaluating the further development of the solution towards product deployment, not only for long-distance lines, but also for substations and other types of critical infrastructure. HIVE technology is designed to be scalable from pilot projects to European coverage of large networks. The goal is to help improve the way EG.D maintains its electricity network. "Its maintenance is essential for reliability and safety. Our infrastructure stretches for thousands of kilometers in rural and urban areas. That is why we carry out inspections to ensure a stable and secure supply of electricity to our customers," explains Josef Basík, head of EG.D's extra-high voltage network operations. To prevent various outages or blackouts, EG.D employees carry out inspections. “This is done either on foot, when they visually inspect the overhead lines to check that everything is in order and that there are no signs of wear or damage. Or by climbing, when our technicians climb the pylon every four years and check the individual components of the pylon and the lines,” adds Josef Basík.
Drones at EG.D can help inspect insulators in high resolution, which can also be more efficient than current inspections. Their main advantage is safety, but they are also efficient and provide structured data. "Thanks to advanced sensors and algorithms, drones can automatically detect insulators and other key components on the line. The use of two drones controlled by one pilot is unique in the world and significantly increases the efficiency of inspection activities. The result is up to twice the inspection volume per drone operator and, as a result, a potential reduction in overall operating costs," calculates Jan Čech, project manager at EG.D.
However, unmanned aircraft are not intended to replace human labor. "They can help in situations where it is difficult to find workers, but they are not intended to replace human employees. Advances in drone technology are significant, but their purpose is primarily to increase efficiency and facilitate work," explains Jan Čech.
MRS group leader Martin Saska adds that the use of multi-robotic systems in power line inspection has proven successful and also has great potential for the development of other services in other industries. "Drones can be used in agriculture, among other things, to scan fields along power lines or to inspect photovoltaic power plants. E.ON approached the CTU spinout Fly4Future with a request to develop a robotic system that could handle these tasks. Product development is in its final stages, and if the technology is successfully verified, it is expected to be deployed in Germany first and then in other European countries," adds doc. Martin Saska.
Students involved in research that will be put into practice
Master's and doctoral students from the Faculty of Electrical Engineering at CTU also participated in the project. The mast detector in LIDAR data was developed during a short student internship and proved to be more accurate in practice than the tool originally used to prepare commercial drone missions. "It's great when research isn't shelved, but put into practice. It is precisely such projects that show where the limits are and what has a real impact," adds Dr. Pěnička.
Three entities with clearly defined roles collaborated on the project. The MRS group from FEE ČVUT was responsible for autonomous algorithms, system integration, and field verification, including BVLOS operation. E.ON Group Innovation commissioned the contract research and developed an operator application for controlling multiple drones and integrating them into operations. EG.D, as the distribution network operator in part of the Czech Republic, provided experienced drone operators, technical requirements, test sections, and operational validation.
