As early as the First World War, several countries experimented with unmanned, radio-controlled, or pre-programmed aircraft. During the 1950s and 1960s, this technology evolved into a popular hobby, with remote-controlled model aircraft becoming a common sight at airfields and flying clubs. Strictly speaking, the first modern drones were initially little more than advanced model aircraft equipped with cameras and additional electronics. Since then, however, a new class of machines has emerged: systems capable of perceiving their environment, processing data, and operating autonomously within defined limits. Technically speaking, according to international ISO standards, a modern drone is fundamentally a robot, explains Johanna Marie Haan, Head of the Robotics Technology Field at Siemens Energy. In practice, however, the distinction lies in the scope of functionality. “In this sense, a drone becomes a flying robot when it not only flies but also acts independently,” she says.
The drone market has been booming for years. Initially, drones were primarily regarded as flying camera platforms for aerial photography, surveying, or recreational use. Today, they are evolving into working machines that collect data, analyze it, and integrate it into operational processes. Flying robots combine sensors, software, navigation, and communication technologies into a system capable of far more than simply transporting itself through the air. They do not merely observe their surroundings—they derive actionable insights from what they perceive.
This development is particularly visible in the energy sector. According to Haan, robot-based solutions have become so well established that some processes could hardly function without them. Flying robots inspect power transmission lines, detect hazardous vegetation growth, and identify critical hotspots using thermal imaging technology. They can also detect geometric changes in transmission lines at an early stage. Such operations offer a significant advantage over traditional inspections because defects can be identified before they become serious problems. As a result, maintenance becomes more predictable and outages can be prevented more effectively.
A good example is the wind energy sector. Conducting detailed inspections of turbines at great heights is labor-intensive for human personnel and involves considerable risks. Flying robots can perform this work more quickly and without the same physical demands. At the same time, they generate high-resolution data that can be evaluated automatically. Haan reports that, in this field, the time required between analysis and diagnosis has been reduced by more than 90 percent. For operators, this means shorter downtime and more precise decisions regarding maintenance and repairs.
The technological progress driving this development comes from several directions at once. Sensors have become smaller, lighter, and more precise. Processors now provide significantly greater computing power while consuming less energy. In addition, advances in machine learning, computer vision, and navigation technologies have accelerated capabilities even further. As a result, today’s systems can perform tasks that only a few years ago would still have been considered experimental.
How reliable these capabilities are outside laboratory environments depends heavily on the operating conditions. This is emphasized by Holger Voos, a robotics researcher at the University of Luxembourg. While cameras, radar, and LiDAR systems can now operate very reliably under favorable conditions, the situation becomes significantly more challenging in darkness, rain, fog, or unstructured environments such as forests and indoor spaces. “Environmental perception and interpretation are still highly unreliable in such cases.”
Voos also identifies another fundamental challenge: seeing alone is not enough. A flying robot must be able to interpret sensor data and derive the correct decisions from it. Recognizing individual objects is now manageable in many cases. Complex situations involving numerous moving and stationary elements, however, remain highly demanding. For this reason, Voos believes that the greatest current limitations lie less in data acquisition itself and more in interpretation and autonomous decision-making.
“The field of environmental interpretation has developed rapidly in recent years thanks to new methods of artificial intelligence,” notes the engineering scientist. “However, the required computing power – and therefore the size and energy consumption of the onboard computers that must be integrated into a flying robot – still represents a significant constraint.”
However, the intelligence of modern flying robotics is not limited to the aircraft itself. Functions that require decisions within fractions of a second – such as stabilization, collision avoidance, and short-term course corrections – are still performed onboard. More extensive analyses, by contrast, are often carried out in ground stations or cloud platforms. There, data from multiple missions can be consolidated, compared with historical information, and integrated into broader operational workflows.
Jeroen Hanekamp, co-founder and CEO of the Dutch technology company Aerosophia, therefore describes the drone somewhat tongue-in-cheek as a “sensor with wings.”
For Hanekamp, the real economic value lies in this connectivity. “A drone that can fly autonomously but delivers its data into an isolated system is ultimately just a highly advanced tool,” he says, adding: “The most intelligent part of the system is not in the air—it is the software that interprets everything the drone observes.” Only when observations automatically feed into asset management systems, trigger maintenance processes, or support operational decision-making does genuine added value emerge. As a result, the market is also shifting. Demand is no longer focused solely on aircraft with excellent flight characteristics, but increasingly on integrated solutions that combine hardware, software, and process integration.
The fields of application for such systems now extend far beyond the energy sector. In construction, flying robots survey building sites and create three-dimensional models. In agriculture, they can detect drought stress, pest infestations, and nutrient deficiencies. In disaster response, they provide up-to-date situational awareness from areas that are difficult to access or hazardous for humans. Their value is growing wherever large areas must be inspected regularly or where human workers would otherwise be exposed to significant risks.
However, it is precisely in those areas with the greatest economic potential that many applications encounter regulatory limitations. “The largest gap between what is technically feasible and what is currently permitted lies in so-called BVLOS operations – ‘Beyond Visual Line of Sight’ flights, where the drone operates outside the pilot’s direct field of view,” explains Johanna Marie Haan of Siemens Energy. This challenge is particularly evident in inspections of extensive infrastructure such as railway corridors or power transmission lines. Long distances, varying geographic conditions, and differing administrative responsibilities create substantial bureaucratic complexity, especially in a federally organized country such as Germany. New documentation requirements or regulatory approvals may be necessary for each individual region.
The technology has already taken off – often faster than the regulatory framework can keep pace. Whether flying robots will be able to realize their full industrial potential therefore depends not only on the next generation of sensors or more advanced software. The decisive factor will be how quickly approval procedures, regulatory responsibilities, standards, and rules can adapt to technological reality. Only then will the full capabilities of flying robotics be able to move from technical possibility to widespread practical application.
