This article is a part of Poland Unpacked. Weekly intelligence for decision-makers
The growing threat from drones is increasing the need to protect airports, refineries and other critical infrastructure. Asseco Poland has identified this demand and is developing Puszczyk, a system combining different drone detection technologies. Ericsson has joined the project as a partner.
Until recently, the need for protection against unmanned aerial vehicles in Poland was viewed primarily from a civilian perspective. The war across Poland's eastern border and Russia's provocative actions towards NATO countries have changed that. Today, civilian and military applications increasingly overlap, while drone detection systems are becoming more important. New technological solutions are entering the market.
Asseco Poland has been developing an integrated drone detection system called Puszczyk for approximately a year.
A multilayered drone detection system
The system is being developed in cooperation with several technology partners. Its creators concluded that combining solutions already available on the market would be faster and more efficient than developing every component from scratch.
Puszczyk's individual modules come from several companies, including Cellnex, CodeQuest and DGT Zvook. Ericsson has also contributed its technology to the project.
During Mobile Networks Defence Day at its research and development center in Krakow, Ericsson demonstrated the capabilities of Integrated Sensing and Communication (ISAC) technology for real-time drone detection. Play's mobile network was used to transmit the data.
"Combining Ericsson's technology with Asseco's acoustic detection system, which uses CodeQuest and DGT Zvook sensors, within the Puszczyk drone detection platform, where the data is visualized, is intended to enable the development of a multilayered system for accurately detecting and locating low-flying drones," the companies said in a statement.
Response time is critical
Puszczyk can be expanded with additional technological layers.
According to Grzegorz Bartler, vice president of Asseco Poland and the originator of the system, combining different types of sensors – acoustic, optical, radio-frequency and radar – already improves drone detection accuracy.
One of the biggest challenges, particularly in the current geopolitical environment, remains the time available to respond to a threat.
"Suppose we install sensors around an airport in Warsaw or Rzeszów. If a drone is traveling at 300 km/h, it covers two kilometers in 24 seconds. If it is detected only one kilometer from the airport, there are approximately 12 seconds left to react.
"Moreover, under Polish regulations currently in force, critical infrastructure operators are not authorized to destroy detected drones themselves. They must notify the relevant authorities and wait for their intervention," Bartler explains.
This makes it essential to detect drones much earlier and monitor their flight paths.
"To make full use of Puszczyk's potential, we need to build an interconnected network.
"Cellnex, a telecommunications infrastructure operator that previously acquired mobile towers from Play and Polkomtel, brings an extremely important element to the partnership.
"It gives us access to approximately 15,800 base stations across Poland," says Bartler.
From communications to object detection
Ericsson, a provider of technologies and services for telecommunications operators, joined the Puszczyk project to demonstrate how mobile networks can help detect drones and monitor threats near critical infrastructure.
The project fits into Ericsson's strategy of developing mobile networks that serve not only as communications infrastructure but also as systems capable of detecting objects.
This is one of the areas being explored in the development of 6G technology.
Working with partners allows Ericsson to examine how such solutions could complement other drone detection systems using existing telecommunications infrastructure.
Ericsson's solution will not be customized specifically for Puszczyk. Instead, it will be integrated as an object-detection layer based on mobile network technology.
Asseco's platform will combine data from Ericsson's system with acoustic detection data and provide visualization capabilities.
Any future deployments will be adapted to customers' local requirements and the relevant operational and regulatory conditions.
A combined layer for detecting and tracking objects
Bartler argues that Ericsson Sensing and Puszczyk complement each other.
Together, they could form a multilayered system capable not only of detecting and tracking objects but also of determining their size, speed and type.
Ericsson Sensing uses existing mobile network infrastructure. According to Bartler, once the technology becomes commercially available, it could be deployed rapidly across large geographical areas.
Good to know
Existing infrastructure in the service of security
Telecommunications networks are a key component of Integrated Sensing and Communication (ISAC) solutions because existing radio infrastructure can simultaneously provide connectivity and monitor the surrounding environment.
By using signals emitted by base stations, it is possible to detect objects, analyze movement and identify changes in the environment without building a separate network of sensors.
This opens up new applications in security, crisis management, critical infrastructure protection and intelligent transport systems.
As Michał Ziółkowski, board member and chief technology officer at Play, emphasizes, a modern telecommunications network is much more than a communications tool.
"It is strategic and critical infrastructure that underpins economic development, digitalization and, crucially, the security of the country and its citizens.
"Thanks to its scale, availability and resilience, Play's network can support the development of advanced airspace monitoring and protection systems using resources that already exist," says Ziółkowski.
Puszczyk was built by combining existing solutions
The idea for Puszczyk emerged on September 10, 2025.
"Russian drones entered Polish airspace. But it wasn't clear exactly what objects we were dealing with or where they were located.
"At Asseco, I'm responsible for data center infrastructure, among other things. The company develops systems used by major institutions and businesses.
"We certainly wouldn't want to lose them because of an explosion or a falling drone.
"I realized that we needed a solution capable of determining the flight paths of such unmanned aircraft," Bartler recalls.
At the time, Asseco was already working on a platform for managing Internet of Things (IoT) sensors for a mobile network operator.
The platform can handle very large numbers of devices, including those used to collect data from electricity meters.
"It can also be used to manage street lighting.
"For example, Warsaw alone has around 250,000 streetlights, approximately 120,000 of which can be controlled and fitted with sensors.
"So we already had the platform, and we were working with telecommunications partners and suppliers of acoustic sensors.
"That led to an idea: why not combine these tools and capabilities to create a drone detection system?" Bartler explains.
Puszczyk tests new technologies
The Puszczyk system is gradually being expanded with additional technologies and object-detection capabilities.
"We now want to test acoustic sensors mounted on streetlights.
"Until now, devices of this kind have been used, for example, to analyze urban traffic and count passing vehicles.
"We're adapting them so they can recognize approaching drones based on their acoustic signatures. Solutions from companies such as CodeQuest could work well here," says Bartler.
Puszczyk remains an open system.
Bringing in additional technology partners and making new testing areas available in Poland are intended to support its continued development.
Good to know
How Lithuania plans to build an early-warning drone detection system
Lithuania intends to deploy the Ukrainian acoustic drone detection system SkyFortress, which can identify and track low-flying objects based on the sounds they produce.
According to Lithuania's Ministry of National Defence, the technology, developed and used in Ukraine, employs artificial intelligence to detect drones and cruise missiles.
The system classifies targets in real time based on their acoustic or electromagnetic signatures.
Defence Minister Robertas Kaunas emphasized that Lithuania wants to use technologies already proven in Ukraine rather than develop its own solutions from scratch.
The agreement signed in the summer covers not only the supply of the system but also the processing of sensor data, access to libraries of unclassified acoustic and radio-frequency signatures, and personnel training.
The planned tests will assess SkyFortress's effectiveness in detecting and identifying low-flying drones under Lithuanian conditions.
The system is intended to become part of Lithuania's developing multilayered air defense architecture, providing early warning of threats.
Lithuania's Ministry of National Defence did not respond to questions from XYZ concerning the criteria used to select SkyFortress, whether other acoustic solutions had been considered, the procedure for choosing the system for testing, the duration of the trials or the planned coverage of the sensor network.
The questions also addressed whether the network could eventually cover the entire country and critical infrastructure or remain limited to border areas.
A nationwide system requires political and regulatory decisions
More effective protection of critical infrastructure requires monitoring not only the immediate surroundings of individual facilities but also much larger areas.
Cooperation between operators of different drone detection systems is therefore essential.
Bartler cites the example of Modlin Airport, which lies along a potential drone flight path towards Płock.
A system monitoring the area around the airport could therefore detect an object heading towards the refinery at an earlier stage.
"It makes sense to integrate these solutions rather than create independent systems that don't exchange information," Bartler emphasizes.
Ultimately, a single nationwide system managed by public authorities would be preferable.
"We've started discussions. But this involves not only political decisions. A number of formal and legal issues also need to be resolved," says Bartler.
Detecting a drone is only half the battle
Expert's perspective
Counter-drone systems need more than sensors. Data and maintenance costs matter too
Acoustic sensors are relatively inexpensive and can detect drones flying low and relatively slowly, which may be invisible to radar.
Under favorable conditions, they can detect a small electric drone, such as an FPV drone, from several dozen to several hundred meters away. A combustion-engine drone such as a Shahed can be detected from several kilometers away.
Wind and street noise reduce these ranges.
To cover a large area, a country needs to deploy thousands of sensors, provide power and connectivity, lease space on rooftops and towers, and maintain the equipment throughout the year.
I estimate that over ten years, the cost of maintaining such a network would exceed the initial expenditure on hardware.
The second problem is sound recognition.
An AI model can distinguish a drone from a lawnmower, but Russia modifies its unmanned aircraft every few months.
It has already introduced jet propulsion and decoys that are increasingly difficult to recognize and classify by sound.
Unless companies and teams responsible for AI models regularly improve and retrain them, the systems will stop recognizing new types of drones or begin generating false alarms.
The altitude of a target can be estimated from sound, but usually with considerable error.
To locate a drone using multiple sensors, a large network needs a common time reference.
Sensors can synchronize through GPS, but Russia jams GPS signals along NATO's eastern flank.
For that reason, I would use acoustic detection primarily to cue other sensors, such as radars and cameras, towards potential targets.
The team on duty must receive an alert within seconds and know whom to dispatch.
Ukrainian forces shoot down drones detected by their acoustic network because alerts are transmitted to mobile fire teams.
Ukraine's Sky Fortress system comprises more than 14,000 acoustic sensors.
Choosing a system
Before selecting sensors, I would establish who will receive the alerts.
In peacetime, that could be the police; during wartime, the military.
The state therefore needs to design both response procedures from the outset.
The system should also distinguish registered, legal drone flights.
This requires integration with systems such as Poland's DroneTower, which is used to register unmanned aircraft operations.
The state should have guaranteed rights to the data and a contractual right to continue retraining AI models, even if the supplier disappears from the market.
A team capable of retraining a model using data on a new type of drone within a week should operate in Poland.
The system's interfaces should be open and compliant with SAPIENT, a standard adopted by NATO for counter-drone systems.
I would assess effectiveness through independent tests conducted under difficult conditions, such as winter weather or noisy urban environments.
Manufacturers should demonstrate how many drones the complete system detected and how many false alarms it generated.
Sensors should continue operating after a mains power failure and have a backup connection in case the primary communications channel fails.
I would also compare bids based on the ten-year cost of maintaining the network per square kilometer.
Multiple systems within one country
Dependence on a single supplier leaves a country vulnerable to that company's pricing and pace of technological development.
But the first step should be to establish a common data format and ensure control over the system that combines information from different sources into a single operational picture.
Acoustic sensors emit no signals, so equipment from different manufacturers can operate alongside one another without causing interference.
However, if two systems detect the same drone and their data are not combined, an operator may see two separate objects with different classifications on the map.
The military is building the SAN system.
Since this year, critical infrastructure operators have also gained certain capabilities to counter drones under an amendment to Poland's Crisis Management Act.
Operators will choose their own suppliers, as they have every right to do: they know the facilities they are protecting better than anyone.
However, the state should establish minimum requirements before they sign contracts.
Operators' sensors should transmit data in a uniform national format and meet cybersecurity requirements.
Alerts should be delivered in real time to the situational awareness systems used by the police and military.
If the state subsidizes the purchase of detection systems, it can include these requirements in funding agreements.
That way, when a power plant operator installs its own sensors, it would also strengthen the nationwide drone detection network.
Without a shared data format, security personnel at a refinery or port might see a drone on their screen, while the military just a few kilometers away could learn about it too late.
That is why the state should announce a data exchange standard before local authorities and infrastructure operators sign more contracts.
Key Takeaways
- Combining technologies could improve drone detection capabilities. Asseco's Puszczyk system integrates solutions from several partners, including acoustic sensors and Ericsson's mobile network-based sensing technology. Combining data from different sources could eventually make it easier to detect and track objects across larger areas.
- Effective protection requires not only early detection but also a rapid response. When a drone is moving at high speed, there may be only a dozen or so seconds to take action. A detection system must therefore not only identify the threat but also immediately transmit alerts to the relevant authorities. Regular updates to AI models, minimizing false alarms and ensuring reliable infrastructure are equally important.Developing counter-drone systems requires common standards and nationwide coordination. Independent solutions cannot provide a complete picture of the situation unless they exchange data. The state should establish rules for integrating systems, cybersecurity requirements and criteria for assessing performance. Long-term network maintenance costs must also be taken into account. This would allow systems developed by the military, critical infrastructure operators and technology providers to form a coherent national drone detection network.