«I have always been occupied by a question: how do you explain to a client that a «perfect» solution on paper might turn out to be unworkable in practice? It is the same story here – the beautiful ISAC theory runs into a banal modulation conflict. The hybrid scheme looks logical and has been tested in simulations, but I would like to see its behavior in a real urban environment with dozens of interference sources. Only then can we say: «Yes, this works not only in the laboratory».» – Dr. Anna Muller
Imagine your mobile phone not only receiving calls but simultaneously working as a radar – tracking objects around you, measuring distances, determining speeds. Sounds like science fiction? In reality, this is a completely feasible task for sixth-generation networks. The technology is called Integrated Sensing and Communication, or ISAC. Its essence is simple: the same equipment performs two functions – communications and radar.
The idea is elegant, but there is a problem. High-speed communications require complex signals with a large amount of information per symbol – these are the so-called high modulation orders, such as 64-QAM or 256-QAM. Radar, however, prefers simple signals like QPSK – they have higher energy per symbol, which means a better signal-to-noise ratio and more accurate target detection. What we have here is a classic conflict of interests.
Why Combine Communications and Radar?
Why combine communications and radar at all?
Before discussing the solution, it is worth understanding why we need this. ISAC is not an engineer's whim but a logical answer to several practical challenges.
First, the radio frequency spectrum is limited. Every megahertz counts. If the same frequency band can simultaneously transmit data and perform radar location, we save resources.
Second, the equipment. Why install a separate base station for communications and a separate radar if you can combine them in one device? Less hardware means lower costs for installation, maintenance, and power consumption.
Third, performance quality. When the system knows where objects are and how fast they are moving, it can manage the connection more precisely: steering the antenna beam, compensating for interference, and adapting transmission power.
Most early works on ISAC focused on monostatic systems – where the transmitter and receiver are in the same place, like in a classic radar. But there is another option: bistatic sensing. Here, the transmitter and receiver are spatially separated. This provides additional advantages: better coverage of the observation zone, resistance to interference, and difficulty in intercepting the signal. It is for the bistatic scheme that it is particularly important to use not only special pilot signals but also the communication data itself – it contains more information that helps detect and evaluate targets more accurately.
Communications Radar Conflict Speed vs Accuracy
The Conflict: Speed vs Accuracy
Now, to the heart of the problem. Modern communication systems use Orthogonal Frequency-Division Multiplexing – OFDM. Data is transmitted on multiple narrow subcarriers, each modulated separately. The more complex the modulation, the more bits can be packed into a single symbol.
For example, QPSK encodes 2 bits per symbol, 16-QAM encodes 4 bits, 64-QAM encodes 6 bits, and 256-QAM encodes 8 bits. For communications, this is excellent: more bits mean higher data transmission speed. But each bit becomes less energetically protected. With a fixed transmitter power, the energy is spread across all bits, and each of them becomes more vulnerable to noise.
For radar, this is bad. Radar detection works like this: the transmitter emits a signal, it reflects off a target, and returns to the receiver. The receiver must catch this weak reflected signal against a background of noise and interference. The higher the energy of each symbol, the easier this is to do. Which means radar requires simple modulations with a low number of bits per symbol.
The traditional approach is to use only pilot symbols for sensing. These are special symbols inserted into the data stream to estimate the communication channel state. They are modulated with simple schemes like QPSK, and there are few of them. This is enough for a basic channel estimate but insufficient for a high-quality radar image.
We could go another way: use all symbols for sensing, including data. But if this data is modulated with 64-QAM or higher, the quality of the radar picture drops. Symbol energy is low, the signal-to-noise ratio is poor, and target detection becomes unreliable.
The third option is to transmit everything in QPSK. The radar will work perfectly, but communication throughput will drop several times over. This is not a solution either.
Hybrid Scheme Using Pseudo-Pilots for ISAC
The Hybrid Scheme: Pseudo-Pilots in the Data Stream
Our solution is a hybrid resource allocation scheme. The main idea: we transmit part of the symbols in simple modulation specifically for the radar, and the rest in complex modulation for communications. We call the symbols with simple modulation pseudo-pilots.
It works like this. Let's say a base station transmits data in 64-QAM for maximum speed. But every eighth symbol (or every sixteenth – the ratio is adjustable) is modulated not with 64-QAM but with QPSK. These QPSK symbols are placed uniformly across the time-frequency grid: across subcarriers and across OFDM symbols in time.
On the receiving side, these pseudo-pilots are used to build the radar map. They have high energy, a low error rate, and a good signal-to-noise ratio. The remaining symbols are decoded as regular communication data.
Yes, we sacrifice data transmission speed slightly. If one-eighth of the symbols are transmitted in QPSK instead of 64-QAM, total spectral efficiency drops by about 5-10 percent. But in return, we get a significantly better radar image.
The formula is simple. Let the share of pseudo-pilots be α, the high modulation order be M_H, and the low one be M_L. Then the average efficiency is:
(1 − α) × log₂(M_H) + α × log₂(M_L)
For example, if α = 1/8, M_H = 64, M_L = 4, then the efficiency will be approximately 5.5 bits per symbol instead of 6 bits with pure 64-QAM. The loss is small, while the gain in radar is substantial.
Practical Application of ISAC Hybrid Scheme
How This Applies in Practice
The scheme is flexible. The share of pseudo-pilots can be changed dynamically depending on the tasks. If connection speed is more important at the moment – we lower α. If radar accuracy is more important – we raise it. The system adapts to current conditions.
Pseudo-pilots are positioned to ensure good coverage along the delay and Doppler shift axes. Delay relates to the distance to the target: the further the object, the later the reflected signal arrives. Doppler shift relates to speed: a moving target causes a change in the frequency of the received signal.
To build a radar map, the receiver performs coherent processing: accumulates several OFDM symbols, applies a Fourier transform in time and frequency, and obtains a 2D function – a Delay-Doppler map. Peaks on this map correspond to detected targets.
The key moment: pseudo-pilots with QPSK modulation provide a high signal-to-noise ratio at every point on this map. This increases the probability of target detection and the accuracy of estimating its parameters – distance and speed.
The Decoding Error Problem in ISAC
The Decoding Error Problem
Another important detail: decoding errors. If we use decoded data symbols with high order modulation for the radar, any decoding error turns the symbol into noise. This ruins the radar picture.
Pseudo-pilots solve this problem. QPSK is much more resistant to noise than 64-QAM or 256-QAM. The probability of decoding error for QPSK at the same noise level is orders of magnitude lower. This means pseudo-pilots can be used for radar with practically no risk of distortion.
If we tried to use 64-QAM symbols for sensing without decoding, their low energy would yield a poor signal-to-noise ratio. If we decode them and then use them – we run the risk of errors that distort the radar image. The hybrid scheme bypasses both of these problems.
ISAC Simulation Results Facts and Figures
Simulation Results: Facts and Figures
We tested the scheme on models. We compared three basic variants and our hybrid scheme.
The first variant involves only pilot signals. Sensing is performed exclusively by standard pilots, which are few and far between. The result: low probability of target detection, large error in parameter estimation.
The second variant entails only data with high modulation order (64-QAM). All symbols are used for communications and radar simultaneously. The connection works well, but the radar suffers due to low symbol energy.
The third variant involves only data with low modulation order (QPSK). The radar works perfectly, but communication throughput drops threefold.
The hybrid scheme showed the following. With a pseudo-pilot share of α = 1/8, the probability of target detection at a signal-to-noise ratio of 0 dB was 0.9. For comparison: the variant with pilots gave 0.6; the variant with 64-QAM gave 0.7. That is, we detect targets more reliably.
The root mean square error of delay estimation at a signal-to-noise ratio of 10 dB was about 10⁻⁶ seconds squared for the hybrid scheme versus 10⁻⁵ for the pilot variant. Accuracy improved by an order of magnitude.
Spectral efficiency of the connection decreased by about six percent compared to pure 64-QAM. This is an acceptable price for such a gain in radar performance.
The impact of decoding errors also turned out to be minimal. If we used decoded 64-QAM symbols for sensing, decoding errors would significantly distort the radar picture, especially with poor channel quality. With pseudo-pilots, this problem practically vanishes: QPSK is stable, there are almost no errors, and the radar works steadily.
ISAC Practical Application and Perspectives
Practical Application and Perspectives
Where can this be used? Anywhere simultaneous communication and radar location are needed.
Automotive networks. Cars exchange data with each other and simultaneously track surrounding objects – pedestrians, other vehicles, obstacles. The bistatic scheme is particularly good here: one car transmits a signal, another receives the reflection. This provides a more complete picture of the situation than a monostatic radar.
Smart cities. Mobile base stations can not only serve subscribers but also track traffic movement, monitor pedestrian flows, and detect anomalies like crowd gatherings or sudden changes in traffic.
Industrial networks. In factories, warehouses, and logistics centers, the same infrastructure can manage robots, transmit telemetry data, and simultaneously track object movement and monitor safety.
Bistatic sensing gives additional flexibility. Transmitters and receivers can be placed at different points, covering hard-to-reach zones, bypassing obstacles, and increasing resistance to interference. This is especially important in an urban environment where direct line of sight is often absent.
Future Research Directions for ISAC
What's Next
The hybrid scheme is not the final solution but a foundation for further development. There are several directions worth researching.
First is dynamic optimization of the pseudo-pilot share. Currently, we set α manually. But we can make a system that automatically adapts α depending on channel quality, communication requirements, and current sensing tasks. For example, if the channel is good and the connection is stable, we can increase the share of pseudo-pilots to improve the radar. If the channel is poor and every bit counts – we lower α.
Second is expansion to multi-user systems. In real networks, dozens and hundreds of users operate simultaneously. How do we distribute pseudo-pilots among them? Can we coordinate transmissions so that different users' pseudo-pilots complement each other and improve the overall radar picture?
Third involves MIMO systems with multiple antennas. Modern base stations use dozens or even hundreds of antennas to form directional beams. The hybrid scheme must account for this: placing pseudo-pilots not only across time and frequency but also across spatial directions.
Fourth is coding and decoding. We looked at basic modulation, but in real systems, data is encoded with error-correcting codes like Turbo codes or LDPC. How does the interaction of coding and pseudo-pilots affect overall performance? Can information from the decoder be used to improve sensing?
Finally, it is worth testing the scheme in real conditions. Modeling shows good results, but practice always throws up surprises: multipath propagation, non-stationary channels, interference from neighboring networks. Field tests will show how resistant the scheme is to these factors.
ISAC Summary
Summary
Integrated data transmission and sensing is one of the key technologies for future networks. It allows for the efficient use of spectrum, equipment, and energy. But to implement it, one must resolve the conflict between the requirements of communications and radar.
The hybrid resource allocation scheme offers a practical solution. A small share of pseudo-pilots with simple modulation ensures high-quality sensing while having almost no effect on connection speed. The scheme is flexible, adaptive, and resistant to decoding errors.
Modeling confirmed: the probability of target detection grows, parameter estimation accuracy increases, and throughput losses are minimal. This is a working tool, ready for implementation in real systems.
Reliable energy, much like air, should be. The same applies to both communications and radar. The hybrid scheme takes a step toward making both these functions work noticeably, efficiently, and without compromise.
