Waveform Randomization for Secure ISAC
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: "Waveform Randomization for Secure ISAC".
Paper summary: Rosa: To wrap up this paper, "Waveform Randomization for Secure ISAC" by Zexin Fang and colleagues, they demonstrate a way to secure integrated sensing and communication by randomly scrambling the sensing waveform itself using phase keys and tangent artificial noise.
Dev: The main point is that coordinated legitimate nodes can recover the key to make sense of the randomized signal, while an eavesdropper only gets a scrambled waveform that they can't reconstruct coherently.
Taro: It shows that we can use this randomization to protect communication even against a stronger adversary who tries to figure out the phase keys over long periods, because residual errors still hurt their ability to decode correctly.
Rosa: It suggests that for next-generation wireless networks, reusing radio resources for both data and sensing through waveform randomization is a promising approach.
Dev: But they also highlighted that accurate key recovery is computationally demanding because the phase keys are refreshed independently every time.
Conclusion: Rosa: So, to wrap up this paper, "Waveform Randomization for Secure ISAC," these authors are showing how you can scramble the sensing signal itself using phase keys and some artificial noise to keep eavesdroppers out while letting the right receivers figure it out.
Dev: Exactly. The authors are mapping a standard phase key into a bunch of harmonic symbols, and they use something called a tangent vector to add that rotational change without changing the main symbol itself.
Taro: What I find interesting is how they handle the signal transmission, where the real scalar is drawn randomly based on that keyed symbol and the tangent direction. It sets up this whole system for detection.
Rosa: And when we look at what Eve does, she tries non-coherent detection and cross-burst coherent reference reconstruction to steal those phase keys or recover the signal.
Dev: The security metrics they track are the correct bin probability, which is how often you guess the right symbol, and false lock probability, which is when you get stuck.
Taro: The numerical results show that their unprotected sequence has a high correct bin probability, like one point zero versus maybe just zero point five for the random baseline.
Rosa: But adding the keying alone drops that down to three eighty-six percent, and adding the artificial noise further reduces it, bringing it to zero when they use about six tenths of that noise.
Dev: That tells us communication can be protected this way without needing some extra complicated security layer on top.
Taro: But there's a catch because the authors mention that residual key mismatch and imperfect removal of that artificial noise still limit how well coherent sensing works when Eve tries to reconstruct the keys.
Rosa: Right, so while it’s a solid way to hide things, getting the key back perfectly is still tough computationally.
Dev: Yeah, and they flag that trying an exhaustive search for just one burst within a short observation time is really demanding because those phase keys are refreshed randomly and independently every time.
Taro: It sounds like the real challenge isn't just hiding the signal, but making sure we can still use it reliably when we actually need to sense something accurately.
Rosa: That leads us into how this design does things in a bypass scenario where Eve listens on one path while trying to find another.
Dev: We’ll see how that plays out next, and then we'll look at what the structural characteristics of those waveforms actually mean for ranging reliability, like the peak sidelobe ratio.
RPTU University Kaiserslautern-Landau · German Research Center for Artificial Intelligence (DFKI), Germany
eess.SP
Submitted: 2026-10-06
Updated: 2026-11-18
Importance score: 78/100
The gist: This paper studies waveform-level security for Integrated Sensing and Communication (ISAC) by randomizing sensing waveforms using phase keys and tangent Artificial Noise (AN) to hide the sensing…
Key concepts
- Tangent Vector
- This vector is derived by differentiating the symbol vector with respect to a phase index. It represents a purely rotational (phase) perturbation of the original signal, meaning it changes the phase without altering the fundamental component along its direction.
- Phase Key Vector
- A shared key applied across all harmonic components of a signal. This key is used to create a keyed Fourier-curve symbol vector, which helps in both hiding the sensing reference and allowing legitimate receivers to recover it.
- Tangent Artificial Noise (AN)
- This noise is strategically added to the signal. It serves two purposes: it helps hide the sensing reference from eavesdroppers, and it aids in the recovery process for legitimate receivers by providing a measurable perturbation.
Terminology
Summary
This paper studies waveform-level security for Integrated Sensing and Communication (ISAC) by randomizing sensing waveforms using phase keys and tangent Artificial Noise (AN) to hide the sensing reference from eavesdroppers while allowing legitimate receivers to recover it Waveform Randomization for Secure ISAC is a promising paradigm for next-generation wireless networks, reusing radio resources for both data transmission and environmental sensing within a unified system [1]. p. 0
Waveform Design and Keying Mechanism
The core of the design involves mapping a conventional M-PSK phase to a k-dimensional harmonic symbol vector, where the components are jointly determined by a single index i. The tangent vector tˆi is derived from differentiating this symbol vector with respect to θi and normalizing it to unit norm. p. 1 This tangent direction tˆi describes a purely rotational (phase) perturbation of xi, with no component along xi itself. p. 1 The transmitter applies a shared phasekey vector ϕ across the k harmonic components to create the keyed Fourier-curve symbol vector xi(ϕ). p. 1
Signal Transmission and Channel Modeling
The transmitted real scalar ξr is drawn independently for each symbol vector, defined by x tx i,r = p 1 − β xi(ϕ) + p β ξrtˆi(ϕ). The received symbol vector Yr is modeled as Yr = p 1 − β xi(ϕ) + p β ξrtˆi(ϕ) + Nr, where Nr ∼ N (0, σ2 c I2k). p. 1 The total symbol-vector SNR is defined as ρ = E[kx tx i,rk 2] = 1/2kσ 2c. p. 1
Detection and Security Metrics
The threat model considers Eve performing non-coherent detection, exploiting long-term observations to attempt cross-burst coherent reference reconstruction, and searching the phase-key space to reconstruct phase keys. The detection statistic z[l] is computed by correlating the direct and reflected signals, z[l] = X n r[n] d∗[n − l], l = 0,. p. 2.., L − 1. The performance metrics include the correct-bin probability Pcorr and the false-lock probability Pfalse. p. 2
Validation of Security Performance
Numerical results validate the security design, showing that the unprotected optimized sequence achieves PC = 1.0 versus 0. p. 305 for the unprotected random baseline. p. 3 The keying alone drops PC to 0. p. 3386, and adding AN further reduces it, falling below the unprotected random baseline by β ≈ 0. p. 33 and reaching PC = 0 from β ≈ 0. p. 36. The results indicate that communication can be protected by the proposed design, with Eve’s BER limited to 0. p. 35 without phase key. p. 0
Conclusion on Key Recovery Complexity
The paper concludes that residual key mismatch and imperfect AN removal still limit coherent sensing when Eve attempts phase-key reconstruction, while accurate key recovery incurs high computational complexity. The cost of exhaustive search for a single burst within a short observation time is computationally demanding because the phase keys are randomly and independently refreshed. p. 5
Bypass Scenario Analysis
In the bypass case, where Eve listens on the direct path while attempting to detect the reflected path, both Pcorr and PD are similar for the random and the optimized sequences. This similarity follows because Algorithm 1 optimizes the mainlobe/sidelobe trade-off for the clean, unkeyed, AN-free case through a sensitive phase alignment across chips. p. 4 The legitimate receiver reaches 0. p. 592 with power accumulation of 10 echoes under favorable geometry. p. 5
Sidelobe Characterization
The peak sidelobe ratio (PSR) and the mainlobe alignment χ[0] squared characterize ranging reliability more completely than PD alone. A lower-sidelobe waveform can exhibit lower PD than a higher-sidelobe one while achieving equal or higher Pcorr. p. 3 The structural effect of the QPSK alphabet causes dips at k = 4 and k = 8 due to harmonic sums vanishing when k is a multiple of the period. p. 3
Future Research Directions
Future work will investigate joint optimization of the phase key, tangent-AN power, and harmonic dimension k under explicit sensing accuracy and communication reliability constraints. This research will also validate the design in multi-target and multiEve scenarios closer to practical B5G/6G deployments. p. 5 The paper considers a stronger Eve that attempts phase-key reconstruction from longterm observations, showing that residual key mismatch structurally degrades correlation and AN cancellation. p. 0 The high computational complexity associated with accurate key recovery is also noted. p. 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
-
Waveform Randomization for Secure ISAC: The system can hide sensing references from an eavesdropper by randomizing a Fourier-curve constellation using phase keys and tangent Artificial Noise (AN) applied to a Fourier-curve constellation, ensuring Eve observes
only a randomized waveform.
-
Enhanced Eavesdropper Resilience: Even if Eve attempts phase-key reconstruction, the design ensures
residual key mismatch structurally degrades correlation and AN cancellation,
preventing accurate key recovery from leading to successful sensing reference recovery. -
Robust Detection in Bypass Scenarios: The system maintains detection performance even when a UAV-mounted Eve exploits favorable geometry through a bypass scenario where she listens to both the direct and echo sides, as the legitimate receiver
reliably locks onto the reference through echo accumulation.
-
Covariance-Aware Demodulation for Communication Security: The communication channel can be protected by using a covariance-aware demapper that accounts for candidate-dependent noise covariance, which
reduces the penalty when the residual lies along the AN direction expected under candidate i.
-
Improved Ranging Reliability: The system can achieve better ranging reliability by optimizing the mainlobe/sidelobe trade-off, as
a lower-sidelobe waveform can exhibit lower PD than a higher-sidelobe one while achieving equal or higher Pcorr.
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