Evolution Toward Sensing-Enabled RAN: Practical Bistatic Sensing via Uplink
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: We're looking at the paper: "Evolution Toward Sensing-Enabled RAN".
Dev: Here's the gist of the paper. This paper considers uplink channel state information (CSI)-based sensing as an opportunistic approach to enabling sensing in cellular networks.
Rosa: The approach is validated through simulation and practical experiments with a software-defined radio-based setup, utilizing open-source software frameworks and commercially available hardware components.
Dev: Results demonstrate feasibility of the uplink CSI-based sensing for detection, localization and tracking of moving targets over local areas.
Rosa: Uplink CSI-Based Sensing Fundamentals The paper explores uplink channel state information (CSI)-based sensing as a practical near-term route to enabling sensingenabled networks by reusing known pilot signals.
Paper summary: Rosa: Looking at the whole paper, the authors are essentially showing that uplink CSI estimates can enable bistatic sensing in evolved 5G FR1 networks by only making minimum required modifications to the infrastructure <ref:2610.20923#pg1,minimum required modifications to the infrastructure>. This isn't some futuristic concept; it's something they have validated through simulation and real measurements using an NVIDIA Aerial open-source gNB implementation.
Dev: The key is that they manage to use the 5G UE positioning function alongside the gNB antenna array for both absolute range and angle of arrival estimation, which is what makes this bistatic sensing possible <ref:2610.20923#pg1>. It moves the sensing capability from just detecting things to actually localizing them over local areas.
Taro: So, for someone just listening in, what's the big picture here? Does this mean we can start seeing movement detected by cell towers without installing entirely new radar equipment everywhere?
Rosa: It suggests that because of how the 5G system is already set up with these uplink signals, we can start enabling sensing applications simply by processing those channel estimates in a specific way <ref:2610.20923#pg1>. It opens up possibilities for things like security or environmental monitoring where knowing what's moving nearby is important.
Dev: The implication is that the effort needed to get this kind of sensing capability isn't as high as you might think, because they are showing how to do it with existing components and just a few processing tweaks. The feasibility was confirmed by detecting moving targets walking towards the radio unit within about zero point five to one meter distance in their experiments.
Taro: So it's a way to add a sensing dimension onto cellular communication without needing huge new hardware deployments right away, just by exploiting the channel information they already collect.
Rosa: Exactly, and that's what the paper about "Evolution Toward Sensing-Enabled RAN: Practical Bistatic Sensing via Uplink" is about—it shows how to make this happen in a practical way for modern 5G networks <ref:2610.20923#pg1,Evolution Toward Sensing-Enabled RAN: Practical Bistatic Sensing via Uplink>.
Conclusion: Rosa: So this paper is talking about how we can actually get sensing working in 5G networks by using the information they already collect from the uplink channel, which is what this "Evolution Toward Sensing-Enabled RAN" is all about <ref:2610.20923#pg1,Evolution Toward Sensing-Enabled RAN>.
Dev: Yeah, it's not about building some giant new radar system; it's leveraging what you already have on the airwaves to detect things. The authors are focusing on making bistatic sensing practical right now.
Taro: So, if I understand this right, they aren't proposing a total overhaul of the network architecture to add sensing capabilities; they’re looking at how to add it with just a few changes.
Rosa: Exactly, they show feasibility through simulation and actual tests using an open-source setup and even real hardware like an NVIDIA Aerial gNB. It proves you can detect moving targets over local areas.
Dev: The core mechanism they use is exploiting the channel state information, or CSI, from two specific 5G signals: the DMRS and the sounding reference signal <ref:2610.20923#pg1>. Those are just standard pilot signals used for communication.
Taro: And they use those pilots to figure out where a target is by combining estimates from both the ground station and maybe even multiple antennas on that station. That's how they get that localization down to a few meters, as the experiment showed.
Rosa: Right, the results show target positions clustered around where the actual subject was walking, within about five to ten centimeters of error in their practical test. The caveat is that they have to account for things like limited power and motion compensation because the user is moving around.
Dev: The paper mentions that this approach needs minimal modifications to existing 5G FR1 networks, which is a big deal for deployment timelines <ref:2610.20923#pg1>. It's not some massive infrastructure project waiting years to happen.
Taro: So, what does this mean for the world beyond just detecting a person walking near a tower? It opens the door for things like enhanced security monitoring in urban areas or even better tracking for autonomous systems operating locally.
Rosa: It means we can start thinking about networks that are not just for data transfer but also for environmental awareness and situational awareness using existing radio infrastructure. We need to look at how much more useful this becomes when we expand the sensing targets beyond just a walking person.
Fraunhofer Heinrich Hertz Institute · Technische Universitat Berlin
eess.SP
Submitted: 2026-10-14
Updated: 2026-11-16
Comments: Presented at IEEE ISAC 2026
Code: https://github.com/NVIDIA/aerial-cuda-accelerated-ran
License: http://creativecommons.org/licenses/by/4.0/
Importance score: 84/100
The gist: This paper considers uplink channel state information (CSI)-based sensing as an opportunistic approach to enabling sensing in cellular networks.
Key concepts
- Uplink CSI-Based Sensing
- This method uses information about the channel quality (CSI) from signals sent from a user device (UE) back to the base station (gNB). It leverages known pilot signals, specifically DMRS and SRS, to infer properties of the environment and detect targets in real-time.
- DMRS and SRS
- These are two types of 5G uplink pilot signals used for sensing. DMRS is embedded in every data transmission to help with channel equalization. SRS is sent independently to assist scheduling decisions, both providing valuable channel insights for target detection.
- Target Localization Methodology
- To find a target's position, the system needs two main pieces of information: an absolute LoS delay and the Angle-of-Arrival (AoA). The paper proposes using 5G UE positioning to estimate the physical LoS path length and gNB antenna arrays to measure AoA, which together pinpoint the target's location.
Terminology
Summary
This paper considers uplink channel state information (CSI)-based sensing as an opportunistic approach to enabling sensing in cellular networks. p. 0 fraunhofer. de
Abstract—This paper considers uplink channel state information (CSI)-based sensing as an opportunistic approach to
enabling sensing in cellular networks, with minimum required
modifications to the infrastructure. A practical sensing signal
processing procedure is presented, combining 5G positioning
function with multi-antenna array processing for passive target
localization. The approach is validated through simulation and
practical experiments with a software-defined radio-based setup,
utilizing open-source software frameworks and commercially
available hardware components. The approach is validated through simulation and
practical experiments with a software-defined radio-based setup,
utilizing open-source software frameworks and commercially
available hardware components. Results demonstrate feasibility
of the uplink CSI-based sensing for detection, localization and
tracking of moving targets over local areas.Where the paper says it
Where the paper says it
Where the paper says it
Uplink CSI-Based Sensing Fundamentals
The paper explores uplink channel state information (CSI)-based sensing as a practical near-term route to enabling sensingenabled networks by reusing known pilot signals. Two types of 5G uplink pilot signals are suitable for sensing, namely the DMRS and sounding reference signal (SRS). p. 0 Two types of 5G uplink pilot signals are suitable for sensing,
i. , DMRS and sounding reference signal (SRS) [14]. , DMRS and sounding reference signal (SRS) [14]. The
DMRS is embedded within every uplink data transmission
and the obtained CSI is used for channel equalization prior to
symbol detection, while the SRS is transmitted independently
of user data and is used to support scheduling decisions. , DMRS and sounding reference signal (SRS) [14]. The
DMRS is embedded within every uplink data transmission
and the obtained CSI is used for channel equalization prior to
symbol detection, while the SRS is transmitted independently
of user data and is used to support scheduling decisions.Where the paper says it
Where the paper says it
Where the paper says it
Channel Modeling and Sensing Context
The channel transfer function (CTF) observed at the gNB for a single-antenna transmitter and receiver can be expressed mathematically. This channel includes contributions from multiple propagation paths, including the lineof sight (LoS) and multipath components associated with objects in the environment. p. 1 , fn = f0 + n∆f;
f0 carrier frequency;
∆f subcarrier spacing;
βl complex path gains;
τl propagation delays;
f
l
D exhibited Doppler shifts. The channel includes contributions from multiple propagation paths between the Tx and the Rx, including the lineof-sight (LoS) (if unobstructed) and multipath components
associated with objects in the environment, i. The sensing context follows from the uplink propagation
geometry, illustrated in Fig. A simple scenario is considered
for simplicity, including the gNB, UE and a single scatterer
assumed to be the target of interest. As shown in the figure,
for a given echo signal delay, the corresponding target lies on
an ellipse with the UE and gNB at its foci, and the semi-major
axis being twice the propagation path length [16]. The Doppler
shift corresponds to the rate at which the total propagation path
length changes over time and is proportional to the relative
speed in the direction normal to the propagation ellipse [17].Where the paper says it
Where the paper says it
Where the paper says it
Target Localization Methodology
Total bistatic range recovery requires knowing whether an absolute LoS delay is known, which can be achieved via timing advance (TA) or 5G UE positioning. The 5G uplink-based UE positioning function is presented as a more accurate alternative for estimating the physical LoS path length. p. 2 the narrow bandwidth of the utilized Physical Random Access
Channel (PRACH) signal, resulting in a distance error on
the order of tens of meters, hence not suitable for target
localization. The 5G uplink-based UE positioning function is a
more accurate alternative for estimating the physical LoS path
length [19]. 1), can be arbitrarily oriented in space and the
scatterer position on this ellipse remains unknown. Target
localization additionally requires angle of arrival (AoA) estimation at the gNB, by exploiting multiple Rx antennas. 1Here we adopt the typical assumption that the target speed is low enough
such that the fast-time Doppler shift is negligible. Target Position Mapping
With the absolute target echo delay and AoA estimated,
and the absolute LoS delay and AoA used to establish the
propagation ellipse geometry, the target position is obtained
at the intersection of the AoA ray from the gNB and the
propagation ellipse.Where the paper says it
Where the paper says it
Where the paper says it
Sensing Procedure and Validation
The gNB-side sensing signal processing procedure involves two branches: one for LoS parameters to support synchronization and hardware impairments compensation, and another for target detection and localization. The rangeDoppler-angle map is processed through successive Fourier transforms to obtain the delay-Doppler-angle map. p. 2 The first one estimates LoS parameters to
support synchronization, hardware impairments compensation,
LoS suppression and target localization. The second one performs LoS and clutter suppression, and Fourier transform steps
to obtain the range-Doppler-angle map for target detection
and localization, utilizing inputs from the first branch and UE
positioning mechanisms. Alternatively,
a more computationally-efficient second-order polynomial interpolation can be applied instead [20, Sec. Finally, by employing the absolute LoS delay estimate
obtained via the 5G UE positioning function, the relative
target echo delays are translated into the target distances
from the gNB, according to (6), thereby completing the target
localization procedure.Where the paper says it
Where the paper says it
Practical Implementation Challenges
Several practical challenges are associated with this approach, including limited Tx power and positioning error. The method assumes the UE is nomadic, requiring compensation for its motion by estimating the LoS path Doppler shift. p. 3 3) UE mobility: The presented method assumes that the
UE is nomadic and remains static over the CPI, such that the
LoS component and clutter can be treated as time-invariant. In
general, the impact of UE motion must be compensated for,
i. 4) CSI availability and signaling overhead: The DMRSbased estimates are available whenever data is being transmitted on the uplink, and thus can be sparse or irregular
under low traffic load conditions. In contrast, SRS is scheduled
independently of data transmissions and offers more flexible
time-frequency configuration, at the expense of occupying
additional uplink resources. A combined opportunistic strategy
could be of interest, where DMRS-based CSI estimates are
exploited in high traffic conditions, while SRS transmissions
are scheduled during periods of low data activity. 5) Sensing with communication hardware: The hardware
impairments, including timing offset, carrier frequency offset,
and sampling frequency offset (SFO) between the Tx and thWhere the paper says it
Where the paper says it
Where the paper says it
Simulation and Experimental Outcomes
Simulation results validate the approach using a single-antenna UE, a four-element ULA, and a static target scenario. Experimental results with an NVIDIA Aerial SDR gNB demonstrate feasibility in detecting moving targets walking towards the RU. p. 5 The presented sensing procedure is demonstrated based on
simulation and measurements obtained with a real-time SDRbased setup, developed using the NVIDIA Aerial open-source
gNB implementation and commercially-available RU and UE. By considering an indoor scenario with a person walking, the
results demonstrate the feasibility of target localization via
uplink communication in 5G FR1 networks, with minimum
required modifications. The detected target positions are clustered
around the walking path, typically remaining within 0. 5 - 1 m
distance from the actual subject position. the narrow bandwidth of the utilized Physical Random Access
Channel (PRACH) signal, resulting in a distance error on
the order of tens of meters, hence not suitable for target
localization. The 5G uplink-based UE positioning function is a
more accurate alternative for estimating the physical LoS path
length [19]. the narrow bandwidth of the utilized Physical Random Access
Channel (PRACH) signal, resulting in a distance error on
the order of tens of meters, hence not suitable for target
localization. The 5G uplink-based UE positioning function is a
more accurate alternative for estimating the physical LoS path
length [19]. The presented sensing procedure is demonstrated based on
simulation and measurements obtained with a real-time SDRbased setup, developed using the NVIDIA Aerial open-source
gNB implementation and commercially-available RU and UE. By considering an indoor scenario with a person walking, the
results demonstrate the feasibility of target localization via
uplink communication in 5G FR1 networks, with minimum
required modifications.Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Conclusion
The paper concludes that the presented sensing approach enables bistatic sensing in evolved 5G FR1 networks by exploiting uplink CSI estimates. The feasibility is demonstrated through simulation and measurements using an NVIDIA Aerial open-source gNB implementation. p. 5 The presented sensing procedure is demonstrated based on
simulation and measurements obtained with a real-time SDRbased setup, developed using the NVIDIA Aerial open-source
gNB implementation and commercially-available RU and UE. By considering an indoor scenario with a person walking, the
results demonstrate the feasibility of target localization via
uplink communication in 5G FR1 networks, with minimum
required modifications. CONCLUSIONS
This paper considers an opportunistic approach to enable
bistatic sensing in evolved 5G FR1 networks, by exploiting
uplink CSI estimates obtained from pilot signals. The 5G UE
positioning function and gNB antenna array are explored for
LoS absolute range and AoA estimation, effectively allowing
for target localization.Where the paper says it
Where the paper says it
References
[1] N. Gonzalez-Prelcic, M. p. 6 REFERENCES
[1] N. Gonzalez-Prelcic, M. REFERENCES
[1] N. Gonzalez-Prelcic, M. Gonzalez-Prelcic, M. Furkan Keskin, O. Gonzalez-Prelcic, M. Furkan Keskin, O. Furkan Keskin, O. Kaltiokallio, M. Furkan Keskin, O. Kaltiokallio, M. Kaltiokallio, M. Valkama, ´
D. Kaltiokallio, M. Valkama, ´
D. Valkama, ´
D. Dardari, X. Valkama, ´
D. Dardari, X. Dardari, X. Bayraktar, and H.Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
[2] F. Liu, Y.-F. Liu, Y.-F., Cui, C., Masouros, J. p. 6 676–723,
Jul. Masouros, J. Masouros, J. Xiao Han, S. Xiao Han, S. Eldar, and S.Where the paper says it
Where the paper says it
Where the paper says it
[3] E. Dahlman, S. p. 6 1–30, Oct. Dahlman, S. 1–30, Oct. Dahlman, S. Dahlman, S. Parkvall, and J. Dahlman, S. Parkvall, and J.Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
[4] L. Giroto, M. Henninger, A. p. 6 London, UK: Academic Press,
2021. Henninger, A. Henninger, A. Bauhofer, T. Bauhofer, T. ten Brink, T. Nuss, and S. Mandelli, “Bistatic
ISAC: Practical Challenges and Solutions,” Jan.Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
[5] K. Khosroshahi, P. p. 6 org/abs/2601. Khosroshahi, P. org/abs/2601. Khosroshahi, P. Khosroshahi, P. Sehier, and S. Khosroshahi, P. Sehier, and S.Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
[6] N. K. Nataraja, S. p. 6 4702–4707. Nataraja, S. 4702–4707. Nataraja, S.Where the paper says it
Where the paper says it
[7] V. Tapio, N. Tervo, M. E. Leinonen, and A. p. 6 6121–6137, Apr. Leinonen, and A. 6121–6137, Apr. Leinonen, and A.Where the paper says it
Where the paper says it
[8] D. Brunner, L. p. 6 (JC&S), Leuven, Belgium, Mar. Brunner, L. (JC&S), Leuven, Belgium, Mar. Brunner, L. Brunner, L. Giroto de Oliveira, C. Nuss, and S. Mandelli, “Bistatic
ISAC: Practical Challenges and Solutions,” Jan. Diewald, Y. Basim Alabd, L.Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
[9] R. C. S. J. Sanson and V. p. 6 3016–3029, May 2025. Sanson and V. 3016–3029, May 2025. Sanson and V. Sanson and V. Frascolla, “Human Presence Detection via
Wi-Fi Range-Filtered Doppler Spectrum on Commodity Laptop,” in
Proc. Sanson and V. Frascolla, “Human Presence Detection via
Wi-Fi Range-Filtered Doppler Spectrum on Commodity Laptop,” in
Proc.Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
[10] J. Pegoraro, J. p. 6 Workshops (PerCom
Workshops), Pisa, Italy, Mar. Pegoraro, J. Workshops (PerCom
Workshops), Pisa, Italy, Mar. Pegoraro, J. Pegoraro, J. Mezzavilla, M.Where the paper says it
Where the paper says it
Where the paper says it
[11] R. K. Kanth, W. Khawaja, and A. p. 6 9759–9775, Aug. Khawaja, and A. 9759–9775, Aug. Khawaja, and A.Where the paper says it
Where the paper says it
[14] 3GPP, “5G; NR; physical channels and modulation,” 3rd Generation Partnership Project (3GPP), Technical Specification (TS) 38. p. 6 org/abs/2512. [14] 3GPP, “5G; NR; physical channels and modulation,” 3rd Generation
Partnership Project (3GPP), Technical Specification (TS) 38. org/abs/2512. [14] 3GPP, “5G; NR; physical channels and modulation,” 3rd Generation
Partnership Project (3GPP), Technical Specification (TS) 38. [14] 3GPP, “5G; NR; physical channels and modulation,” 3rd Generation
Partnership Project (3GPP), Technical Specification (TS) 38. 2026, version 16. [14] 3GPP, “5G; NR; physical channels and modulation,” 3rd Generation
Partnership Project (3GPP), Technical Specification (TS) 38. 2026, version 16. [14] 3GPP, “5G; NR; physical channels and modulation,” 3rd Generation
Partnership Project (3GPP), Technical Specification (TS) 38. 2026, version 16. [14] 3GPP, “5G; NR; physical channels and modulation,” 3rd Generation
Partnership Project (3GPP), Technical Specification (TS) 38. 2026, version 16. 2026, version 16. 0, Release 16. 2026, version 16. 0, Release 16. 2026, version 16. 0, Release 16. 2026, version 16. 0, Release 16.Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
[15] C. Sturm and W. p. 6 0, Release 16. Sturm and W. 0, Release 16. Sturm and W.Where the paper says it
Where the paper says it
[16] J. D. Parsons, The Mobile Radio Propagation Channel, 2nd ed. p. 6 1236–1259, May 2011. Parsons, The Mobile Radio Propagation Channel, 2nd ed. 1236–1259, May 2011. Parsons, The Mobile Radio Propagation Channel, 2nd ed. Parsons, The Mobile Radio Propagation Channel, 2nd ed. London,
UK: John Wiley & Sons, 2000. Parsons, The Mobile Radio Propagation Channel, 2nd ed. London,
UK: John Wiley & Sons, 2000.Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
[17] N. J. Willis, Bistatic Radar. p. 6 London,
UK: John Wiley & Sons, 2000. Willis, Bistatic Radar. London,
UK: John Wiley & Sons, 2000. Willis, Bistatic Radar. Willis, Bistatic Radar. Raleigh, NC, USA: SciTech Publishing
Inc. Willis, Bistatic Radar. Raleigh, NC, USA: SciTech Publishing
Inc.Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
[18] 3GPP, “5G; NR; Medium Access Control (MAC) protocol specification,” 3rd Generation Partnership Project (3GPP), Technical Specification (TS) 38. p. 6 Raleigh, NC, USA: SciTech Publishing
Inc. [18] 3GPP, “5G; NR; Medium Access Control (MAC) protocol specification,” 3rd Generation Partnership Project (3GPP), Technical Specification (TS) 38. Raleigh, NC, USA: SciTech Publishing
Inc. [18] 3GPP, “5G; NR; Medium Access Control (MAC) protocol specification,” 3rd Generation Partnership Project (3GPP), Technical Specification (TS) 38.Where the paper says it
Where the paper says it
[19] H.-S. Cha, G., Lee, A., GhoshM., BakerS., KelleyJ., and Hofmann. p. 6 Kelley, and J. Hofmann, “5G
NR Positioning Enhancements in 3GPP Release-18,” IEEE Commun. Kelley, and J. Hofmann, “5G
NR Positioning Enhancements in 3GPP Release-18,” IEEE Commun.Where the paper says it
Where the paper says it
[20] M. A. Richards, Fundamentals of Radar Signal Processing, 3rd ed. p. 6 22–27, Mar. Richards, Fundamentals of Radar Signal Processing, 3rd ed. 22–27, Mar. Richards, Fundamentals of Radar Signal Processing, 3rd ed. Richards, Fundamentals of Radar Signal Processing, 3rd ed. New
York, NY, USA: McGraw-Hill, 2022. Richards, Fundamentals of Radar Signal Processing, 3rd ed. New
York, NY, USA: McGraw-Hill, 2022.Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
[21] J. Capon, “High-Resolution Frequency-Wavenumber Spectrum Analysis,” Proc. p. 6 New
York, NY, USA: McGraw-Hill, 2022. Capon, “High-Resolution Frequency-Wavenumber Spectrum Analysis,” Proc. New
York, NY, USA: McGraw-Hill, 2022. Capon, “High-Resolution Frequency-Wavenumber Spectrum Analysis,” Proc. Capon, “High-Resolution Frequency-Wavenumber Spectrum Analysis,” Proc. 1408–1418, Aug. Capon, “High-Resolution Frequency-Wavenumber Spectrum Analysis,” Proc. 1408–1418, Aug.Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
[22] R. Schmidt, “Multiple Emitter Location and Signal Parameter Estimation,” IEEE Trans. p. 6 1408–1418, Aug. Schmidt, “Multiple Emitter Location and Signal Parameter Estimation,” IEEE Trans. 1408–1418, Aug. Schmidt, “Multiple Emitter Location and Signal Parameter Estimation,” IEEE Trans. Schmidt, “Multiple Emitter Location and Signal Parameter Estimation,” IEEE Trans. Antennas Propag. Schmidt, “Multiple Emitter Location and Signal Parameter Estimation,” IEEE Trans. Antennas Propag. Antennas Propag. 276–280, Mar. Antennas Propag. 276–280, Mar.Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
[23] B. Friedlander, “On Transmit Beamforming for MIMO Radar,” IEEE Trans. p. 6 276–280, Mar. Friedlander, “On Transmit Beamforming for MIMO Radar,” IEEE
Trans. 276–280, Mar. Friedlander, “On Transmit Beamforming for MIMO Radar,” IEEE
Trans. Friedlander, “On Transmit Beamforming for MIMO Radar,” IEEE
Trans. 3376–3388, Oct. Friedlander, “On Transmit Beamforming for MIMO Radar,” IEEE
Trans. 3376–3388, Oct.Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
[24] L Giroto de Oliveira, Y. p. 6 Brunner, L. Giroto de Oliveira, C. Brunner, L. Giroto de Oliveira, C. Henninger,
X. Jun Cui, T.Where the paper says it
Where the paper says it
Where the paper says it
[25] NVIDIA Corporation, “NVIDIA Aerial CUDA-Accelerated RAN,” GitHub repository, 2025, version 26. p. 6 4143–4161, Dec. [25] NVIDIA Corporation, “NVIDIA Aerial CUDA-Accelerated RAN,”
GitHub repository, 2025, version 26. 4143–4161, Dec. [25] NVIDIA Corporation, “NVIDIA Aerial CUDA-Accelerated RAN,”
GitHub repository, 2025, version 26.Where the paper says it
Where the paper says it
[26] S. D’Oro, M. p. 6 com/NVIDIA/aerial-cuda-accelerated-ran. D’Oro, M. com/NVIDIA/aerial-cuda-accelerated-ran. D’Oro, M.Where the paper says it
Where the paper says it
Simulation Results
The simulation scenario involves a single-antenna UE communicating with a gNB equipped with a four-element ULA. The target is simulated as static at a distance of 15 m and an azimuth angle of −30◦ relative to the array broadside. p. 4 The UE is placed at a distance of 10 m from
the gNB, in the direction of the array broadside (ϕ0 = 0◦
),
and is assumed to be static over the sensing interval. A single
passive target is considered at a distance of 15 m and an
azimuth angle of −30◦
relative to the array broadside. The signal propagation is assumed to be dominated by the
LoS path and the single-scatter target echo, with the CTF
at each antenna element simulated according to (5), with
the (relative) delays and AoAs determined by the scenario
geometry. The LoS path SNR at the Rx antenna output (prior
to signal processing gain) is set to 20 dB, with the target echo
being 12 dB weaker. The LoS path
is observed as the strongest component at 10 m, while the
target echo appears as a weaker component with path length
of around 23 m. The LoS suppression step effectively removes
the strong LoS component, thereby facilitating the subsequent
processing and target detection.Where the paper says it
Where the paper says it
Where the paper says it
Experimental Validation
Experimental results with an NVIDIA Aerial SDR gNB confirm feasibility using a Quectel RM500Q-GL 5G modem as the UE. The subject walking towards the RU was detected with positions clustered around the walking path, typically within 0. p. 5 6b) for the subject
walking towards the RU, with the dashed line indicating
the walking path. The detected target positions are clustered
around the walking path, typically remaining within 0. The presented sensing procedure is demonstrated based on
simulation and measurements obtained with a real-time SDRbased setup, developed using the NVIDIA Aerial open-source
gNB implementation and commercially-available RU and UE. By considering an indoor scenario with a person walking, the
results demonstrate the feasibility of target localization via
uplink communication in 5G FR1 networks, with minimum
required modifications.Where the paper says it
Where the paper says it
The paper demonstrates the feasibility of target localization via uplink communication in 5G FR1 networks by exploiting CSI estimates. The approach leverages the 5G UE positioning function and gNB antenna array for absolute range and AoA estimation. p. 5 CONCLUSIONS
This paper considers an opportunistic approach to enable
bistatic sensing in evolved 5G FR1 networks, by exploiting
uplink CSI estimates obtained from pilot signals. The 5G UE
positioning function and gNB antenna array are explored for
LoS absolute range and AoA estimation, effectively allowing
for target localization. The 5G UE
positioning function and gNB antenna array are explored for
LoS absolute range and AoA estimation, effectively allowing
for target localization. The presented sensing procedure is demonstrated based on
simulation and measurements obtained with a real-time SDRbased setup, developed using the NVIDIA Aerial open-source
gNB implementation and commercially-available RU and UE. The presented sensing procedure is demonstrated based on
simulation and measurements obtained with a real-time SDRbased setup, developed using the NVIDIA Aerial open-source
gNB implementation and commercially-available RU and UE. By considering an indoor scenario with a person walking, the
results demonstrate the feasibility of target localization via
uplink communication in 5G FR1 networks, with minimum
required modifications.Where the paper says it
Where the paper says it
Where the paper says it
[1] N. Gonzalez-Prelcic, M. Furkan Keskin, O. Kaltiokallio, M. Valkama, ´ D. Dardari, X., ShenY., ShenY., BayraktarM., and WymeerschH.
[2] F. Liu, Y.-F., LiuY., CuiC., MasourosJ., XuT., Xiao HanT, BuzziS, EldarYC, and JinS. p. 6 Masouros, J. Xiao Han, S.Where the paper says it
[3] E. Dahlman, S. Parkvall, and J. Skold.
[4] L. Giroto, M., HenningerM., FelixA., BauhoferM., JeongT., U., ErdemS, ten BrinkS, ZwickmT,, NussB,, and Mandelli. p. 6 Bauhofer, T. ten Brink, T.Where the paper says it
[5] K. Khosroshahi, P. Sehier, and S. Mekki.
[6] N. K. Nataraja, S., SharmaS., AliK., BaiF., WangR., and A. F.
[7] V. Tapio, N. Tervo, M. E. Leinonen, and A. Parssinen.
[8] D. Brunner, L., Giroto de OliveiraC. p. 6 Brunner, L. Giroto de Oliveira, C. Diewald, Y. Basim Alabd, L.Where the paper says it
Where the paper says it
[9] R. C. S. J. Sanson and V., Frascolla.
[10] J. Pegoraro, J., O., LacruzT., AzzinoM., MezzavillaM., RossiJ, WidmerJ,, and RanganS.
[11] R. K. Kanth, W. Khawaja, and A. C. Gurbuz.
[14] 3GPP, “5G; NR; physical channels and modulation,” 3rd Generation Partnership Project (3GPP), Technical Specification (TS) 38.211, Jun. 2026, version 16.10.0, Release 16.
[15] C. Sturm and W. Wiesbeck.
[16] J. D. Parsons, The Mobile Radio Propagation Channel, 2nd ed., London, UK: John Wiley & Sons, 2000.
[17] N. J. Willis, Bistatic Radar., Raleigh, NC, USA: SciTech Publishing Inc., 2005.
[18] 3GPP, “5G; NR; Medium Access Control (MAC) protocol specification,” 3rd Generation Partnership Project (3GPP), Technical Specification (TS) 38.321, Jan. 2026, version 16.22.0, Release 16.
[19] H.-S. Cha, G., LeeA., GhoshM., BakerS., KelleyJ., and Hofmann.
[20] M. A. Richards, Fundamentals of Radar Signal Processing, 3rd ed., New York, NY, USA: McGraw-Hill, 2022.
[21] J. Capon, “High-Resolution Frequency-Wavenumber Spectrum Analysis,” Proc. IEEE, vol. 57, no. 8, pp. 1408–1418, Aug. 1969.
[22] R. Schmidt, “Multiple Emitter Location and Signal Parameter Estimation,” IEEE Trans. Antennas Propag., vol. 34, no. 3, pp. 276–280, Mar. 1986.
[23] B. Friedlander, “On Transmit Beamforming for MIMO Radar,” IEEE Trans. Aerosp. Electron. Syst., vol. 48, no. 4, pp. 3376–3388, Oct. 2012.
[24] L Giroto de Oliveira, Y., LiM, S., MandelliD., BrunnerM., HenningerX., WanT., Jun CuiT,, ZwickmT,, and NussmB. p. 6 Henninger,
X. Jun Cui, T.Where the paper says it
[25] NVIDIA Corporation, “NVIDIA Aerial CUDA-Accelerated RAN,” GitHub repository, 2025, version 26.1.
[26] S. D’Oro, M., PoleseM., L. BonatiH., ChengT., and MelodiaT.
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
-
Improve target detection capabilities by implementing a range-Doppler-angle map processing pipeline that utilizes constant false alarm rate CFAR methods for target detection and subsequent tracking across multiple coherent processing intervals (CPIs).
-
Enable absolute bistatic range recovery by integrating the 5G UE positioning function into the sensing procedure to accurately estimate the absolute LoS delay, which is then used in equation (6) to translate relative echo delays into target distances from the gNB.
-
Enhance target localization accuracy by employing multiple Rx antennas and applying an IDFT across the antenna dimension to estimate Angle of Arrival (AoA), followed by using the AoA ray and propagation ellipse geometry to determine the target position at the intersection point described in equation (6).
-
Improve clutter suppression effectiveness for moving targets by subtracting the CTF time-average from the estimated CSI, as described in equation (9), which effectively removes the zero-Doppler component to facilitate subsequent processing.
-
Increase system robustness against hardware impairments by incorporating a LoS-based over-the-air synchronization mechanism to compensate for timing offset, carrier frequency offset, and sampling frequency offsets as mentioned in Section VI.
Abstract
This paper considers uplink channel state information (CSI)-based sensing as an opportunistic approach to enabling sensing in cellular networks, with minimum required modifications to the infrastructure. A practical sensing signal processing procedure is presented, combining 5G positioning function with multi-antenna array processing for passive target localization. The approach is validated through simulation and practical experiments with a software-defined radio-based setup, utilizing open-source software frameworks and commercially available hardware components. Results demonstrate feasibility of the uplink CSI-based sensing for detection, localization and tracking of moving targets over local areas.
Sources
- Bistatic ISAC: Practical Challenges and Solutions
- CSI-Based User Positioning, Channel Charting, and Device Classification with an NVIDIA 5G Testbed
- Programmable GPU-Accelerated dApps for Real-Time Edge Inference and ISAC on NVIDIA Aerial Testbed
Related papers
- Open-H-RF: A Large-Scale Ultrasound Raw Channel Dataset Initiative
- Scalable Attention for 5G NR Channel Estimation
- Low-Altitude Wireless Networks: A Comprehensive Survey
- Sensing-Assisted OFDM with Index Modulation: A Waveform for Optical Wireless Integrated Communication and Localisation
- Iteration-Free Candidate Pruning and an Index-Domain VLSI Architecture for Sort-Free MIMO Box Decoding up to 1024-QAM
- A Foundation Model for Large-Scale Wireless Network Planning, Operation and Optimization