Sensing-Aided Beam Selection for UAV Networks

arXiv:2606.00277 · eess.SP · Submitted 2026-05-29 · Read on arXiv

Listen

Radio episode about this paper

Transcript

Introduction to the show: ident: Electrical Engineering. The day in electrical engineering: signals, systems and the hardware that carries them.

Rosa: I'm Rosa.

Dev: And I'm Dev, here with Taro.

Rosa: We're looking at the paper: "Sensing-Aided Beam Selection for UAV Networks".

Dev: Here's the gist of the paper. Fast and low-overhead beam management is a critical requirement for the practical deployment of nonterrestrial networks (NTNs) operating at millimeter-wave and higher frequencies.

Rosa: The proposed radar-assisted beam selection framework limits the set of candidate beams by utilizing spatial sensing information such as the angle-of-departure (AoD) and distance estimations.

Dev: Framework Overview The paper proposes a novel radar-assisted beam selection framework for the initial access stage of NTNs.

Rosa: This approach relies exclusively on radar for sensing assistance, and it utilizes distance information provided by the radar to impose an early stopping condition based on received power feedback to reduce beam search overhead.

Paper summary: Rosa: So we're wrapping up this look at "Sensing-Aided Beam Selection for UAV Networks". Basically, they've built a system where radar helps decide which radio beams to use, and that helps keep the network fast and low overhead for those high-frequency millimeter-wave systems.

Dev: Right. The authors are Bora Bozkurt, Cevdet Tosun, Mehmet Nuri Akıncı, Ali Gorcin, Ibrahim Hokelek and Mehmet Kemal Özdemir. They put a lot of work into this framework to make beam management less of a headache for the network control loop.

Rosa: It's interesting because they use the radar to get that angle of departure information and distance estimations to narrow down the search area before sweeping through all the possible beams.

Taro: So it’s not just testing every possible beam blindly anymore, it’s using prior knowledge from the radar to focus only where targets might be. That should cut down on searching time significantly in a dynamic environment.

Dev: Exactly. And they even looked at what happens if you do this search suboptimally, providing some math on the worst case overhead probability under those bad conditions. That gives us a safety case for how much time we're actually wasting.

Rosa: And it’s not just about speed; they extended it for physical layer security too. The radar can spot passive targets, which means you can suppress interference from eavesdroppers while still boosting the signal for the intended users.

Taro: That suppression result, where they null an unintended user down to negative one hundred thirty five decibels while getting a two decibel gain for others, that’s a tangible improvement in signal quality when things go wrong.

Dev: Yeah, it shows how this radar sensing actually helps the overall system performance in practice. It moves us closer to having beam management that's efficient and secure for real nonterrestrial networks. But we still have to figure out how robust this works when the world gets really messy with movement and interference on a larger scale.

Conclusion: Rosa: So we're wrapping up this look at "Sensing-Aided Beam Selection for UAV Networks." Basically, they built a system where radar helps decide which radio beams to use, and that keeps the network fast and low overhead for those high-frequency millimeter-wave systems.

Dev: Right. The authors are Bora Bozkurt, Cevdet Tosun, Mehmet Nuri Akıncı, Ali Gorcin, Ibrahim Hokelek, and Mehmet Kemal Özdemir. They put a lot of work into this framework to make beam management less of a headache for the network control loop.

Rosa: It’s interesting because they use the radar to get that angle of departure information and distance estimations to narrow down the search area before sweeping through all possible beams.

Taro: So it’s not just testing every possible beam blindly anymore, it’s using prior knowledge from the radar to focus only where targets might be. That should cut down on searching time significantly in a dynamic environment.

Dev: Exactly. And they even looked at what happens if you do this search suboptimally, providing some math on the worst case overhead probability under those bad conditions. That gives us a safety case for how much time we're actually wasting.

Rosa: And it’s not just about speed; they extended it for physical layer security too. The radar can spot passive targets, which means you can suppress interference from eavesdroppers while still boosting the signal for the intended users.

Taro: That suppression result, where they null an unintended user down to negative one hundred thirty five decibels while getting a two decibel gain for others, that’s a tangible improvement in signal quality when things go wrong.

Dev: Yeah, it shows how this radar sensing actually helps the overall system performance in practice. It moves us closer to having beam management that's efficient and secure for real nonterrestrial networks. But we still have to figure out how robust this works when the world gets really messy with movement and interference on a larger scale.

Rosa: The paper they're talking about is titled "Sensing-Aided Beam Selection for UAV Networks." It’s about using radar data to smart filter the search space for communication beams.

Dev: Right. And they did this by tying the distance information from the radar directly into an early stopping rule based on how strong the power feedback gets. It’s all about cutting down on wasted beam searching time during initial access.

Taro: What this means for us is that for those high-frequency networks, we can do more beams in less time, which is crucial when a UAV has to quickly find a link before it moves out of range or gets blocked by something else.

Rosa: It's also about security. They show you can use the radar to spot potential eavesdroppers passively and then use that information to actively suppress their signal while helping your actual users get a better connection.

Dev: So, this whole framework is designed to be low overhead, fast, and secure for those non-terrestrial networks we're deploying high up in the air.

Communications and Signal Processing Research (HÍSAR) Lab. · Department of Electronics and Communication Engineering, Istanbul Technical University · Department of Electrical and Electronics Engineering, Medipol University

eess.SP

Submitted: 2026-05-29

Updated: 2026-11-13

Importance score: 80/100

The gist: Fast and low-overhead beam management is a critical requirement for the practical deployment of nonterrestrial networks (NTNs) operating at millimeter-wave and higher frequencies.

Key concepts

Radar-Assisted Beam Selection
This framework uses a radar system to help the network choose the best communication beams. The radar first finds rough directions of potential users, and then the network only searches for narrow beams in those specific areas, making the search much faster and more efficient.
Early Stopping Condition
The system stops searching for communication beams early if it receives feedback about the power level. This feedback is used alongside distance information from the radar to decide when to stop searching, which saves time and computational resources during beam selection.
Physical Layer Security (PLS)
This extension uses the radar's ability to detect passive targets, such as eavesdroppers. By nulling the signal of an unintended user, the system can suppress their power while simultaneously boosting the signal quality for legitimate users, improving security and performance.
Worst Case Overhead Approximation (WCOA)
This is a mathematical formula that estimates the maximum possible probability of having too much beam search overhead under a non-ideal searching strategy. It helps researchers understand the worst-case efficiency of the proposed beam selection method.

Terminology

Summary

Fast and low-overhead beam management is a critical requirement for the practical deployment of nonterrestrial networks (NTNs) operating at millimeter-wave and higher frequencies. p. 0 The proposed radar-assisted beam selection framework limits the set of candidate beams by utilizing spatial sensing information such as the angle-of-departure (AoD) and distance estimations. p. 0

Framework Overview

The paper proposes a novel radar-assisted beam selection framework for the initial access stage of NTNs. This approach relies exclusively on radar for sensing assistance, and it utilizes distance information provided by the radar to impose an early stopping condition based on received power feedback to reduce beam search overhead. p. 1 p. 1 The framework involves the radar first identifying coarse target directions, followed by narrow-beam sweeping within those detected regions. p. 1 p. 1 p. 1

Worst Case Overhead Analysis

An analysis on the worst case beam selection overhead probability under a suboptimal beam search strategy is provided. p. 1 Under idealized conditions, an approximation for the worst case beam selection overhead is proposed and its statistics are derived under Gaussian error. p. 0 The suggested Worst Case Overhead Approximation (WCOA) is given by Ls ≈ L˜s = O 2NxNyπδ 2/4. p. 3

Physical Layer Security Application

Leveraging the fact that the radar can detect passive targets that are potentially eavesdroppers, the proposed scheme is extended for physical layer security (PLS) application. p. 1 The performance of this security application is evaluated through comprehensive simulations, showing that an unintended user’s power is suppressed below −135 dBm while an additional beamforming gain of roughly 2 dB is attained for the legitimate users. p. 0

Simulation Results and Statistics

The simulation environment involves a radar-assisted gNB at 500 m altitude and three UEs, demonstrating that upon nulling the unintended user, the first and third UEs both experience an average RSRP increase of 2.218 and 2.174 dBs, respectively. p. 5 p. 5 The PDF of L˜s in the case of equal variances is given by fL˜s(z) = 4O 2NxNyς squared exp − 2zO 2NxNyς squared, z ≥ 0, 0, z < 0. p. 4 p. 4

Conclusion

Overall, with this work, we have provided a viable and low-overhead solution for beam management in practical NTN systems operating at mmWave and higher frequency bands. Future work will consider dynamic environments, hybrid precoding extensions, and experimental validation on hardware platforms. p. 5

Bora Bozkurt, Cevdet Tosun, Mehmet Nuri Akıncı, Ali Gorcin, Ibrahim Hokelek, Mehmet Kemal Özdemir. p. 0 The paper is 2606. p. 000277 5 May 29, 29 May 5, and the single most informative sentence about the paper's subject and finding is Fast and low-overhead beam management is a critical requirement for the practical deployment of nonterrestrial networks (NTNs) operating at millimeter-wave and higher frequencies. The framework limits the set of candidate beams by utilizing spatial sensing information such as the angle-of-departure (AoD) and distance estimations. p. 0

The proposed radar-assisted beam selection framework for NTN initial access relies exclusively on radar for sensing assistance. It utilizes distance information provided by the radar to impose an early stopping condition based on received power feedback to reduce beam search overhead. p. 1 The framework involves the radar first identifying coarse target directions, followed by narrow-beam sweeping within those detected regions.

Pruning redundant reasoning tokens

The paper proposes a novel radar-assisted beam selection framework for the initial access stage of NTNs. This approach relies exclusively on radar for sensing assistance, and it utilizes distance information provided by the radar to impose an early stopping condition based on received power feedback to reduce beam search overhead. The framework involves the radar first identifying coarse target directions, followed by narrow-beam sweeping within those detected regions.

How the training signal is built

The system model considers an aerial 5G mmWave BS equipped with two colocated uniform rectangular arrays (URAs) serving aerial user equipments (UEs). The gNB uses its radar to roughly estimate the dominant AoD and the distance of the UE bearing this AoD. p. 2 Once the AoD of a UE, (ϕ, ˆ ˆθ), is estimated with the radar, the gNB performs a sweep using an appropriate subset of the codebook F to select the best beam maximizing received power. p. 2

Where the gains come from

The main contributions of this paper are summarized as follows:

  1. We propose a radar-assisted beam selection framework for NTN initial access, where the radar first identifies coarse target directions and the communication array then performs narrow-beam sweeping only within the detected regions. p. 1

  2. An analysis on the worst case beam selection overhead probability under a suboptimal beam search strategy is provided.

  3. Leveraging the fact that the radar can detect passive targets that are potentially eavesdroppers, the proposed scheme is extended for physical layer security (PLS) application.

The simulation environment consists of a radar-assisted gNB at 500 m altitude and three UEs, demonstrating that upon nulling the unintended user, the first and third UEs both experience an average RSRP increase of 2.218 and 2.174 dBs, respectively. The PDF of L˜s in the case of equal variances is given by fL˜s(z) = 4O 2NxNyς squared exp − 2zO 2NxNyς squared, z ≥ 0, 0, z < 0.

Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.

Improvements for AI systems

  1. Radar-assisted beam selection framework for initial access: The improved system can skip initial beam sweeping by using wider beams from radar and performing refinement only within detected regions, reducing beam search overhead while preserving narrow communication beams and low RFchain complexity.

  2. Distance-aware early stopping criterion: The system can reduce beam selection overhead by utilizing distance information provided by the radar to impose an early stopping condition based on received power feedback, which is achieved by transmitting a pre-determined threshold value Pth to the UE.

  3. Physical layer security enhancement: The system can enhance physical layer security (PLS) by leveraging radar's capability to detect passive targets and using regularized zero forcing to null these unintended users via the beam formula fRZF/kfRZFk, suppressing their signal below-135 dBm.

  4. Worst case overhead approximation: The theoretical analysis provides a performance benchmark for similar sensing-assisted schemes by deriving the worst case overhead approximation Ls approximately as Ltilde s = O2NxNypi delta squared / 4 under Gaussian error assumptions.

  5. Statistical characterization of overhead: The system can utilize the derived probability density function fLtilde s(z) to characterize the statistics of the approximation, providing insight into the expected beam selection overhead for systems with equal variances in estimation errors.

Sources

Related papers