PaperlayerView exact revision on arXiv

Source-linked AI summary

RadarCom: Unlocking Simultaneous Communication and Sensing on Commodity Automotive mmWave Radars

Yi Wei, Yue Zhang, Peijun Hou, Qiuyu Liao, Nan Cen

arXiv:2610.36622v1cs.NI

TL;DR

V2V can improve road safety and traffic efficiency, but existing systems do not meet all requirements for direct communication on unmodified COTS automotive radars while preserving sensing. RadarCom repurposes inter-radar interference for communication and reports an average effective data rate of 5.55 kbps alongside sensing F1 scores of up to 80.5% during concurrent operation.

  • Problem

    Existing systems do not satisfy all requirements for direct communication on unmodified COTS automotive radars without added hardware, customized waveforms, or centralized coordination while preserving sensing.

    Source: §1 · p. 2

  • Method

    RadarCom repurposes mutual inter-radar interference as a communication carrier and uses radar-aware discovery, synchronization, coding, and interference suppression on unmodified COTS radars.

    Source: §1 · p. 3

  • Results

    RadarCom achieves an average effective data rate of 5.55 kbps and a sensing F1 score of up to 80.5% during simultaneous communication and sensing.

    Source: §1 · p. 3, §8 · p. 13

  • Takeaways & Limitations

    The TI AWR2243 prototype demonstrates direct communication between independent radars without hardware modification, external synchronization support, or additional communication radios.

    Source: §8 · p. 13

    RadarCom relies on direct LoS propagation, and its effective data rate drops significantly when a blocking vehicle obstructs that path.

    Source: §6 · p. 12

Abstract

from arXiv · show

Vehicle-to-vehicle (V2V) communication can improve road safety and traffic efficiency, yet its deployment remains limited by the need for dedicated communication hardware and supporting infrastructure. Meanwhile, production vehicles are already widely equipped with millimeter-wave radars for sensing. To unlock their communication potential, this paper presents RadarCom, a software-defined system that enables direct radar-to-radar communication on unmodified commercial off-the-shelf (COTS) automotive mmWave radars while preserving their native sensing functionality. The key insight is that mutual inter-radar interference, traditionally viewed as harmful, can be repurposed as a communication carrier. Based on this insight, RadarCom introduces slope-mismatch-based discovery, adaptive chirp synchronization, differential coding on range-FFT, and sensing-aware interference suppression to enable simultaneous communication and sensing without hardware modification, waveform redesign, or infrastructure support. We implement RadarCom on TI AWR2243 automotive radars and evaluate it under static, dynamic, and challenging real-world conditions. Results show an average effective data rate of 5.55 kbps with symbol error rates below 1.5%, while maintaining sensing F1 scores of up to 80.5% during concurrent communication. These results demonstrate the feasibility of transforming commodity automotive radars into dual-purpose communication and sensing devices.

1 Introduction

RadarCom repurposes naturally received inter-radar signals as a communication carrier, targeting direct communication and simultaneous sensing on unmodified COTS automotive radars. It addresses discovery, synchronization, data exchange, and sensing interference through software-level mechanisms.

1 Introduction

  • Existing approaches face limitations involving customized waveforms, additional hardware, degraded native sensing, or centralized synchronization and infrastructure.The paper states that prior systems do not simultaneously satisfy these requirements for decentralized vehicular deployment.Source: §1 · pp. 1–2
  • RadarCom uses naturally received inter-radar signals as a carrier, encoding information through software reconfiguration of standard FMCW chirp parameters on unmodified COTS radars.The system aims to preserve native sensing without dedicated communication hardware, customized waveforms, or centralized infrastructure.Source: §1 · pp. 2–3
  • Independent radars must discover peers despite unsynchronized chirps potentially falling outside the receiver’s narrow IF observation window.RadarCom’s slope-mismatch-based discovery uses beacon chirps with distinctive frequency-sweeping signatures to establish an initial channel.Source: §1 · p. 2
  • Adaptive synchronization and differential coding on range-FFT outputs address timing and frequency offsets, hardware imperfections, and mobility-related propagation-delay changes.The synchronization mechanism continuously estimates and compensates for offsets; coding uses relative spectral shifts between adjacent chirps.Source: §1 · p. 2
  • Communication-bearing signals can dominate sensing measurements, raising the noise floor, creating false detections, and masking weak targets.RadarCom reconstructs corrupted range-FFT samples across chirps and removes residual false detections through sub-frame consistency checks.Source: §1 · p. 2
  • Up to 5.55 kbps effective data rate and up to 80.5% sensing F1 score were achieved during simultaneous communication and sensing on TI AWR2243 radars.The prototype was evaluated in static, dynamic, and robustness scenarios.Source: §1 · p. 3

2 Preliminary Analysis and Observations

Preliminary experiments establish that one COTS radar can capture another’s signal under controlled timing, but capture is unreliable without synchronization. They also identify timing drift and sensing interference as practical obstacles.

Preliminary Analysis and Observations

  • With proper timing alignment, a radar captured another radar’s transmission, whose slope-mismatched chirps formed a distinctive frequency-sweeping pattern unlike relatively stable environmental echoes.The controlled experiments used synchronized radars and confirmed physical feasibility of direct inter-radar capture.Source: §2 · p. 3
  • 0.75% of 800 frames contained a captured inter-radar signal when two radars started independently without hardware synchronization.A frame counted as captured when the dechirped FFT showed a prominent peak corresponding to the received chirp.Source: §2 · p. 3
  • Capturing another radar’s signal does not by itself enable reliable communication; experiments identify hardware-induced timing drift and communication-induced sensing interference as further challenges.The sensing interference can introduce false detections and mask real targets.Source: §2 · pp. 3–4
  • Matched chirp slopes and fixed relative timing should yield a stable single-tone FFT peak, but the observed peak drifted approximately linearly across chirps.The drift indicates continuously changing relative timing, so practical communication requires tracking and compensation rather than one-time synchronization.Source: Figure 7 · p. 4
  • As Radar 2 transmit gain increased from −30 dB to 0 dB, false positives rose from 19.7% to 92.4%, while false negatives rose from 17.1% to 35.6%.The synchronized, matched-slope interference case also raised the noise floor in Radar 1’s Range-Doppler map.Source: §2 · p. 4

3 System Overview

RadarCom operates within the native FMCW workflow on unmodified COTS radars, using software-level chirp reconfiguration for simultaneous communication and sensing. Its four modules cover radar discovery, chirp synchronization, data transmission, and sensing-aware interference suppression.

3 System Overview

  • RadarCom uses four integrated modules: target radar discovery, adaptive chirp synchronization, differential coding on range-FFT outputs, and sensing-aware interference suppression.It requires software-level reconfiguration of standard chirp parameters, without hardware modification, customized waveform design, or centralized coordination or infrastructure.Source: §3 · p. 4
  • Target radar discovery detects beacon chirps by their frequency-sweeping signature, then uses device-specific SFO-induced spectral drift to distinguish radars sharing beacon parameters.The module names the initiating radar the local radar and its intended peer the target radar.Source: §3 · p. 4
  • Adaptive chirp synchronization alternates two beacon slopes so the local radar can jointly estimate initial timing offset, per-chirp drift, and carrier-frequency offset.The estimates support chirp-level alignment without external timing references despite hardware-induced drift and relative motion.Source: §3 · p. 4
  • Differential coding varies chirp start frequencies and decodes information from relative range-FFT peak shifts between adjacent chirps, reusing the native sensing pipeline.The relative shifts suppress common timing and distance-dependent effects.Source: §3 · p. 4
  • Sensing-aware suppression detects abrupt power jumps across chirps to remove dominant communication interference, then checks within-frame consistency to remove residual false detections.The design exploits communication-induced interference’s temporal inconsistency with real target reflections.Source: §3 · p. 4

4 Detailed Design

The detailed design describes how RadarCom turns radar interference into a communication signal through beacon-based discovery and synchronization, differential coding on native range-FFT outputs, and sensing-aware interference suppression. It also details range-FFT sample reconstruction and sub-frame detection-consistency filtering to suppress communication interference in sensing.

4.1 Target Radar Discovery

  • RadarCom detects target beacons by deliberately mismatching chirp slopes, concentrating beacon energy while dispersing reflections and non-target radar signals across frequency bins.The local radar uses a dechirping reference matched to the residual chirp to form a concentrated spectral peak.Source: §4.1 · p. 5
  • Device-specific SFO produces distinct linear beacon-peak drift trajectories, which RadarCom uses to identify the intended radar even when nearby radars share beacon parameters.The drift slope depends on relative SFO, given a predefined beacon slope.Source: §4.1 · p. 5, Figure 9 · p. 5

4.2 Adaptive Chirp Synchronization

  • Dual-slope beacons separate timing offset, SFO drift, and CFO estimates; RadarCom uses them to adjust chirp idle time and start frequency, periodically re-estimating for stable synchronization.Alternating slopes slightly below and above the local slope produce interleaved peaks; least squares estimates the synchronization parameters.Source: §4.2 · p. 6, Figure 10 · p. 6

4.3 Differential Coding on Range-FFT Outputs

  • Reference-data chirp pairs encode symbols as start-frequency shifts, and adjacent-spectrum subtraction cancels distance-dependent timing and common reflections during decoding.The receiver maps the measured peak-frequency gap to the nearest symbol shift.Source: §4.3 · p. 7
  • RadarCom limits effective modulation bandwidth to half the receiver IF bandwidth; reference-data pairing and periodic re-synchronization constrain throughput, while start-frequency-induced sensing distortion is negligible.The known per-chirp frequency shift can be used to compensate sensing effects when higher accuracy is desired.Source: §4.3 · p. 8

4.4 Sensing-Aware Interference Suppression

  • 11.43 false detections per frame accompanied 81.0% recall, 20.1% precision, and 32.2% F1 when communication-induced interference was present.These static-environment measurements motivated interference suppression; the table is identified as an ablation study of that suppression.Source: §4.4 · p. 8, Table 2 · p. 8
  • RadarCom exploits abrupt adjacent-chirp ghost-peak shifts versus smoother real-target reflections in a two-stage suppression framework.Power-jump detection and sample reconstruction remove dominant interference in range-FFT profiles; residual false detections are then removed in the range-Doppler domain.Source: §4.4 · pp. 8–9
  • RadarCom reconstructs marked interference segments by linear interpolation between nearest unmarked samples, preserving smooth slow-time evolution of real target reflections.Marked samples include upward power jumps and adjacent samples at the same range bin with similar elevated power.Source: §4.4 · p. 9
  • RadarCom retains a CFAR detection only if it appears at a similar range-Doppler location in most sub-frames, filtering inconsistent residual interference.Each 128-chirp frame is divided into equal-length sub-frames, with RD-FFT and OS-CFAR applied independently to each.Source: §4.4 · p. 9
  • With 8 sub-frames, the two suppression stages achieved the best sensing performance, with an F1 score of 80.5%.The 8-sub-frame setting balanced residual false-detection removal and target preservation.Source: Figure 18 · p. 9

5 Implementation and Evaluation

RadarCom runs on two TI AWR2243 radars with independent hosts, enabling direct communication during sensing without shared timing or centralized synchronization. Its evaluation covers static and moving vehicles, cold starts, radar interference, deployment environments, and API-latency constraints.

5.1 Implementation Details

  • On TI AWR2243BOOST radars with DCA1000EVM capture cards, independent hosts run RadarCom while the two vehicles communicate directly and sense without shared timing or centralized synchronization.The prototype uses the vendor dynamic chirp API; its 94 ms average update latency leads to long frame and chirp periods.Source: §5.1 · p. 9
  • The API-constrained configuration achieved 617.37 log2(M) bps, or 5.56 kbps at M=512, with approximately 0.55 m native range resolution and a 4.7 km/h maximum unambiguous Doppler velocity.The paper attributes these limitations primarily to API reconfiguration latency rather than RadarCom’s design.Source: Figure 19 · p. 10, §5.3 · p. 12

5.2 Basic Evaluation Results

  • Across 10–400 m static LoS distances, effective data rate stayed stable, rising from approximately 3.7 kbps at M=64 to 5.5 kbps at M=512.The 400 m distance was limited by the test site; in four increasingly complex sensing scenes, F1 stayed above 74%, range RMSE below 0.55 m, and target miss rate below 0.5%.Source: §5.2 · p. 10, Figure 22 · p. 10, Figure 23 · p. 11
  • At relative speeds of 5–25 km/h, effective data rate remained nearly constant, while M=512 throughput rose from approximately 3.7 kbps to 5.5 kbps and SER increased from 1.27% to 1.51%.Differential coding cancels the distance-dependent shift between adjacent reference and data chirps; velocity-estimation MAE remained below 0.25 m/s across tested speeds.Source: §5.2 · p. 10, Figure 25 · p. 11

5.3 Robustness Evaluation Results

  • After independent cold starts under clear LoS, synchronization succeeded in 99% of approximately 300 trials, with more than 94% completing within 14 seconds.The setup used no shared triggers, clocks, or centralized coordination.Source: §5.3 · p. 11, Figure 26 · p. 12
  • With multi-radar interference, synchronization success was 98% for local-like chirps, 90% for one target-like beacon, and 86% for two target-like beacons.Target-like beacon interference increased average latency to 22.4 s and 38.9 s, while post-synchronization verification helped identify the intended target.Source: §5.3 · p. 11, Figure 27 · p. 12
  • Across indoor parking, outdoor parking, and outdoor-road scenarios, SER was 1.29%, 1.21%, and 1.35%, while synchronization success was 99%, 99%, and 90%, respectively.The authors attribute the lower road synchronization rate to stronger environmental dynamics and less predictable vehicle motion.Source: §5.3 · p. 12, Figure 28 · p. 12
  • Reducing emulated API latency from 94 ms to 0 ms lowered cold-start synchronization latency from 10.9 s to 0.15 s and raised effective data rate from 5.55 kbps to 61.92 kbps.The prototype’s approximately 94 ms API update latency and 7 ms host computation produce a 101.592 ms frame period; sparse ADC observation also reduces opportunities to capture asynchronous beacons.Source: §5.3 · p. 12, Figure 29 · p. 12

6 Discussion •

RadarCom is evaluated across deployment conditions and possible V2V uses, with placement results showing sensitivity to angular misalignment but tolerance to height and lane offsets. The discussion also identifies direct-path blockage and the current prototype’s half-duplex design as boundaries.

Potential Use Cases

  • 5.55 kbps supports the 4.8-kbps BSM rate at 2 Hz, while 61.92 kbps without emulated API latency exceeds the 24-kbps requirement at 10 Hz.The 2-Hz rate is considered for dense freeway traffic to avoid channel congestion.Source: §6 · p. 12

Operation under Non-LoS

  • Direct LoS propagation is required currently; severe blockage can significantly reduce effective data rate, although partial blockage retains near-LoS performance.Reflected paths are proposed as a potential solution; differential coding can mitigate path-dependent delay variations because reference and data chirps share a path.Source: §6 · pp. 12–13

Operation under Diverse Placements

  • Within 30° of angular misalignment, effective data rate remains close to LoS; it falls significantly at 40° and nearly zero at 50°, while height offsets up to 2 m have negligible impact.The local radar moves outside the FoV at 50°; higher-order modulation is more sensitive to angular misalignment.Source: §6 · p. 13, Figure 31 · p. 13

Duplex and Multi-Radar Communication

  • The prototype uses half-duplex time division, while multi-radar communication can build on target identification, angle separation, TDMA, or FDMA.RadarCom identifies and synchronizes with a target amid two additional interfering radars; time-frequency coordination is proposed when angular separation is insufficient.Source: §6 · p. 13

7 Related Work

Prior COTS automotive radar communication approaches face barriers involving customized hardware or waveforms, centralized synchronization, or additional radios and infrastructure. RadarCom is contrasted as enabling integrated sensing and communication on unmodified radars without those additions.

Deployment Barriers

  • Hardware- or waveform-centric systems require customized architectures, cooperative systems assume centralized time-frequency synchronization, and hybrid approaches add radios or infrastructure.The cooperative-system citations are [2] [25] [47], and hybrid-approach citations are [12].Source: §7 · p. 13

RadarCom

  • Unlike a recent commercial-device prototype that uses dedicated modulation and separates sensing from communication, RadarCom integrates both on unmodified COTS radars without external synchronization or additional radios.RadarCom is described as software-defined and does not require hardware modification.Source: §7 · p. 13

8 Conclusion

RadarCom repurposes inter-radar interference to provide direct communication on unmodified COTS automotive mmWave radars while preserving native sensing. Its TI AWR2243 prototype averages 5.55 kbps while maintaining high sensing accuracy, supporting the feasibility of dual-purpose automotive radars.

Conclusion

  • RadarCom enables direct radar-to-radar communication on unmodified COTS automotive mmWave radars while preserving native sensing functionality.It uses mutual inter-radar interference as a communication carrier and includes mechanisms for discovery, synchronization, communication, and sensing-aware interference suppression.Source: §8 · p. 13
  • 5.55 kbps is the prototype’s average effective data rate while maintaining high sensing accuracy.The system was implemented on TI AWR2243 radars and evaluated under static, dynamic, and real-world conditions.Source: §8 · p. 13
  • The experiments demonstrate the feasibility of transforming widely deployed automotive radars into dual-purpose communication and sensing devices.Source: §8 · p. 13
Loading 2610.36622v1…