Technical depth
International engagement
Availability
When I tell colleagues that the antenna system of the future could cost 15 times less than current ones while working even better, the first reaction is usually skeptical. It sounds too good to be true. But that is exactly the result delivered by a new development – the cylindrical directly-connected antenna array. And there is no magic here, just sound engineering.
The problem that needs solving
Imagine you are building an apartment complex and want every unit to have a perfect view. The logical solution is to make the building round with panoramic glazing. Modern wireless communication works in roughly the same way: the more antennas you have, the better the coverage, the higher the speed, and the more precise the user positioning.
MIMO technology – Multiple Input, Multiple Output – is exactly about this. 4G used dozens of antennas; 5G uses hundreds. For 6G, we are already talking about thousands of antennas on a single base station. This is called XL-MIMO – extremely large antenna systems. It sounds impressive, but there is a catch: every antenna requires electronics.
In a standard system, every antenna or group of antennas is connected via an RF chain, which includes amplifiers, filters, and – crucially – phase shifters. These devices control the signal phase to form a directed beam. A single phase shifter for high frequencies can cost several euros. When you need tens of thousands of them, the bill runs into the millions.
Plus power consumption. Plus cooling. Plus the complexity of manufacturing and calibration. As a result, XL-MIMO remains an expensive dream, even though technically everything works perfectly.
First attempt to simplify: The Ray Antenna
A few years ago, an idea emerged to radically simplify the schematic. Instead of complex electronics – simple physics. The concept is called RAA – Ray Antenna Array. The essence is simple: we take several linear chains of antennas, point each in its own direction, and connect the elements directly via cables of varying lengths.
Sound unusual? Let me explain with an example. When a radio wave arrives at an angle to a chain of antennas, it reaches each antenna with a slight delay. If you select the lengths of the connecting cables correctly, these delays are compensated, and all signals add up in phase – creating a directed beam without any electronics at all.
It’s as if you lined up several microphones and connected them with wires of different lengths so that sound coming strictly from the front arrives at all microphones simultaneously after passing through the wires. Sound from the side would arrive with different delays and get suppressed.
RAA works, costs pennies, and gives good angular resolution. But there is a serious problem: all these chains lie in the same plane. They physically obstruct each other. An antenna pointing east blocks signals for an antenna pointing west. In the end, half the system is practically useless.
The solution: Going into the third dimension
Engineers from the research group proposed a logical solution: if a flat design creates blockage, we need to make it volumetric. Thus appeared the cylindrical DCAA – Directly-Connected Antenna Array.
Imagine a cylinder several meters high. Antennas are placed around the circumference of each level – say, 128 units, evenly spaced in a circle. But this isn’t just a circular array. Each circle is split in half into two semi-circles. One semi-circle «looks» one way, the other strictly the opposite.
The antennas in each semi-circle are interconnected by cables with calculated delays – just like in RAA, only now instead of a straight line, we have an arc. The geometry is circular, but the principle is the same: the correct delays create a directed beam in the direction the semi-circle is «facing».
There can be several dozen such levels in the cylinder. Each level is rotated relative to the previous one by a small angle. Why? So that the main beam of one subarray hits precisely the null of the adjacent one's radiation pattern. This way, we avoid mutual interference.
The result is a 3D structure where hundreds of subarrays evenly cover the entire space around the base station with no dead zones and no blockages.
How it works in practice
Let's break down the details. Take a concrete example: a circle of 128 antennas. We split it in half – two subarrays of 64 elements each. One is oriented, say, north, the other south.
The distance between antennas is half a wavelength. For millimeter waves (e.g., 28 GHz frequency), this is about 5 millimeters. Small antennas, dense packing. The radius of the circle ends up being around 20 centimeters.
Now, about the delays. When a signal arrives perpendicularly to the center of our arc, the outer antennas «see» the wave slightly earlier or later than the central one. The phase difference depends on the antenna's position on the arc. We calculate the cable length required from each antenna to the summation point to compensate for this difference.
The formula isn't complicated, but it requires precision. For every antenna with index m in a subarray of M elements, a delay proportional to the sine of its angular position is needed. Physically, this is realized by segments of microstrip lines on a printed circuit board or coaxial cables. Cheap and reliable.
The result: a beam width of approximately 7.66/M radians. For M=64, that’s about 0.12 radians or roughly 7 degrees. Narrow, directional, and well-localized in space.
Layers and coverage
One circle gives us two subarrays – two beams in opposite directions. To cover the full 360 degrees with a resolution of 7 degrees, we need about 50 such pairs. That means 50 levels along the height of the cylinder.
The distance between levels is also half a wavelength. At 28 GHz, that’s the same 5 mm. So, the cylinder height is about 25 centimeters. A compact structure that can be placed on a mast or building roof.
But there is an important nuance: you can't just stack circles on top of each other with the same orientation. You need to rotate each subsequent level by a small angle so that the main lobes don't overlap but fall into the nulls of adjacent patterns.
The rotation angle is calculated from the geometry of the radiation pattern. For M=64, it’s about 2.3 degrees between levels. The exact formula accounts for the position of the first null of the pattern – that very «dip» where we aim the neighbor's beam.
As a result, 50 levels with a 2.3-degree step cover the range from 0 to 115 degrees. The second semi-circles, pointing in opposite directions, cover the remaining sector. Full coverage, uniform resolution, no dead zones.
Connecting to RF chains
Naturally, we cannot connect all 100 subarrays (50 levels × 2 directions) to separate receivers – that would be too expensive. Instead, a matrix of selectors – simple switches – is used.
At any given moment, the base station selects, say, 10 subarrays out of 100 and connects them to 10 RF chains. The choice depends on where the users are. If subscribers are concentrated in the north and east, we connect the subarrays pointing there. If users move later, we switch the matrix.
Switches are inexpensive components compared to phase shifters. One RF switch costs a few cents; a phase shifter costs several euros. A difference of hundreds of times.
From there, signals from the selected subarrays are processed by a digital processor. Here, standard methods are used: digital beamforming, MIMO processing, interference cancellation. These are well-studied technologies operating in modern 5G systems.
Comparison with the traditional approach
The classic architecture for XL-MIMO is called Hybrid Beamforming, or HBF. It is a combination of analog phase shifters and digital processing.
A typical setup: three sectors of 120 degrees, with a linear antenna array of hundreds of elements in each sector. Each group of antennas is connected via an analog block with phase shifters, which roughly steers the beam, and then the digital processor performs the fine-tuning.
Does it work? Yes. Is it expensive? Very. A system for 6G with thousands of antennas and tens of thousands of phase shifters could cost 1.5 million euros just for the components – excluding assembly, calibration, and power consumption.
Another problem with HBF is uneven resolution. A linear array distinguishes angles well across its axis, but poorly along it. Three sectors create hard boundaries. As a result, users at the sector edge get worse connection quality than those in the center.
The cylindrical DCAA is free of these flaws. Circular symmetry provides uniform resolution in all directions. The absence of phase shifters reduces costs by an order of magnitude. According to developers' estimates, a DCAA system for a similar configuration costs about 90,000 euros. That is a 94% saving.
Simulation and results
To verify the theory, the authors conducted a detailed simulation according to 3GPP standards – these are industrial radio propagation models used in designing cellular networks.
Scenario: an office building, high frequencies (28 GHz), no direct line of sight between the base station and users (NLOS). The signal reflects off walls, furniture, ceilings. Complex conditions, close to reality.
Two cases were considered. First – moderate load: 10 users, 64 antennas per level, 10 RF chains. Second – high density: 30 users, 128 antennas per level, 30 chains.
The results are impressive. In uplink mode (from user to base station), the cylindrical DCAA shows a total speed 30–40% higher than traditional ULA+HBF at the same signal-to-noise ratio. In downlink mode (from station to user), the gain is even more noticeable, especially with high user density.
Why is that? Uniform angular resolution means that every user, wherever they are, receives quality service. There are no «bad» zones at sector boundaries. The system effectively uses spatial separation of subscribers, minimizing mutual interference.
An important point is optimization algorithms. Subarray selection, digital precoding, power allocation among users – these are three interconnected tasks. They can be solved iteratively: first optimize subarray selection at fixed power, then adjust precoding, then reallocate power. A few cycles – and the solution converges.
Simulation showed that the algorithm converges in 5–6 iterations. This is fast – fractions of a millisecond on a modern processor. The system can adapt to changes in user positions almost in real-time.
The economics of the issue
Let’s calculate in more detail. Let’s take a configuration for 30 users: 52 levels of 128 antennas, totaling 6656 antenna elements, 30 RF chains.
Traditional ULA+HBF System:
- 6656 antennas × 5 € = 33,280 €
- 83,200 phase shifters (many connections in hybrid architecture) × 16 € = 1,331,200 €
- 30 RF chains × 800 € = 24,000 €
- Total: about 1.4 million €
Cylindrical DCAA:
- 6656 antennas × 5 € = 33,280 €
- 104 RF switches × 25 € = 2,600 €
- 30 RF chains × 800 € = 24,000 €
- Microstrip delay lines (integrated into PCBs) ≈ 30,000 € (estimate)
- Total: about 90,000 €
The difference is colossal. Even if the price of antennas rises to 50 euros apiece (which is unlikely with mass production), DCAA remains several times cheaper due to the absence of thousands of phase shifters.
Plus power consumption. Phase shifters require constant power for control. Switches consume energy only at the moment of switching – fractions of a watt versus tens of watts for every phase shifter. For a base station operating 24/7, electricity savings over a year amount to tens of thousands of euros.
Plus reliability. Fewer active components mean fewer failures. Delay lines are passive elements; they don't break. Phase shifters are complex electronics subject to wear and requiring periodic calibration.
Where this applies
Cylindrical DCAA shows the best results in the millimeter (mmWave) and terahertz (THz) ranges. Why? Because at high frequencies, the wavelength is small, antennas are compact, and large arrays can be realized within reasonable dimensions.
For 28 GHz, a cylinder 25 cm high and 40 cm in diameter is quite acceptable for installation on a mast or building. For frequencies around 100 GHz, the dimensions will become even smaller – convenient for compact base stations.
Applications are diverse:
- Urban 6G networks: High user density, need for uniform coverage, cost criticality in mass deployment.
- Industrial Internet of Things (IIoT): Many devices, strict requirements for reliability and latency, limited budgets.
- Joint communication and radar systems: The high angular resolution of DCAA allows not only data transmission but also precise object positioning – relevant for autonomous transport, drones, logistics.
The last point is particularly interesting. In 6G networks, the concept of ISAC – Integrated Sensing and Communication – is being considered. The same infrastructure is used both for data transfer and for radar sensing of the surrounding space. High resolution is critical here.
What next
The development is at the stage of theoretical justification and simulation. The next step is creating a prototype and field testing. Practical nuances will appear here: manufacturing precision of delay lines, the impact of weather conditions, integration with existing network protocols.
There are avenues for improvement. Subarray selection algorithms can be made more intelligent using machine learning to predict user movement. Power optimization can be linked not only to throughput but also to energy efficiency.
An interesting challenge is scaling to very large systems. What if we need not 100 subarrays, but 1000? How to effectively manage the selector matrix? Perhaps a hierarchical architecture with multiple switching levels makes sense.
It is also important to consider limitations. DCAA is effective for fixed beam directions defined by geometry. For fast scanning or tracking rapidly moving objects, a hybrid approach might be required, combining DCAA with a small number of phase shifters for adaptive fine-tuning.
Why this matters
Deploying next-generation networks is not just a technical challenge, but an economic one. Telecom operators invest billions in infrastructure, and equipment cost directly affects the speed and scale of new technology adoption.
If 6G equipment costs several times less than expected, it changes the entire economics of the industry. More base stations for the same money means better coverage, higher throughput, and more affordable services for end users.
The cylindrical DCAA is an example of how smart use of physics and geometry allows us to bypass expensive electronics. Instead of steering the phase of every antenna electronically, we use fixed delays built into the structure. Instead of a flat array – a volumetric one. Instead of universality – specialization, but with optimal parameter selection.
This solution is not universal. For some tasks, hybrid beamforming with phase shifters will remain preferable – where maximum flexibility or specific operating modes are needed. But for the mass deployment of 6G base stations in cities and buildings, DCAA looks like a very promising option.
Energy should be reliable like air. And connectivity – accessible, fast, and ubiquitous. Antenna technologies like the cylindrical DCAA bring this ideal closer to reality – not through revolution, but through methodical engineering improvement. When simulation results move to the stage of real systems, we will see how much theory aligns with practice. But if even half of the promised benefits are realized, it will be a serious step forward.
