Source-linked AI summary

Affine Frequency Division Multiplexing With Pre-Chirp Spreading Index Modulation: A CSI-Driven Secure Transmission Framework

Yiwei Tao, Miaowen Wen, Yao Ge, Merouane Debbah, Erdal Panayirci

arXiv:2611.15045v1eess.SP

TL;DR

High-mobility AFDM security designs have not fully investigated structured information-bearing pre-chirp codebooks jointly encrypted with CSI and paired with a structure-aware detector. AFDM-PCS-IM uses an orthogonal, CSI-perturbed pre-chirp codebook as an index dimension, with noise-aware quantization and TLMP-WC detection. Simulations report BER advantages over AFDM-IM and conventional AFDM, alongside favorable reliability, security, and complexity balance.

  • Problem

    AFDM security methods have used pre-chirp parameters for permutation, hopping, or secure waveform variables, while their joint use as orthogonal index codebooks with CSI encryption and structure-aware detection remains underexplored.

    Source: B. Motivation and Contribution · p. 2
  • Method

    AFDM-PCS-IM uses an orthogonal pre-chirp codebook encrypted by legitimate-CSI-driven perturbations as an index dimension, with noise-aware CSI quantization and a TLMP-WC detector.

    Source: B. Motivation and Contribution · p. 2, §IV · p. 6, A. Equivalent Received Signal Model… · p. 7
  • Results

    At BER 10^-4 and P = 4, AFDM-PCS-IM provides approximately 0.8 dB gain over AFDM-IM and about 1.8 dB over conventional AFDM.

    Source: B. Effectiveness of the Proposed… · p. 14
  • Takeaways & Limitations

    The findings support pre-chirp parameters as a CSI-adaptive modulation dimension for combining waveform design and physical-layer security in time-varying wireless environments.

    Source: §VII · p. 15

Abstract

from arXiv · show

The flexible configurability of pre-chirp parameters in affine frequency division multiplexing (AFDM) provides new degrees of freedom for designing covert communication mechanisms. Leveraging this property, this paper proposes a covert AFDM pre-chirp spreading index modulation (IM) framework referred to as AFDM-PCS-IM, tailored for high-mobility scenarios. Specifically, an orthogonal pre-chirp parameter codebook is constructed by exploiting the tunability of the pre-chirp parameters. To enhance covertness, the legitimate channel state information (CSI) is utilized to generate controlled perturbations that encrypt the codebook, thereby tightly coupling the modulation structure with instantaneous channel characteristics. The encrypted pre-chirp codebook is then employed as an IM dimension for information transmission, where the index bits are conveyed through the selection of pre-chirp parameter sequences. Since eavesdroppers cannot obtain consistent CSI, they are unable to reconstruct the encrypted codebook and thus cannot reliably recover any useful information. Moreover, to address channel estimation errors inherent in practical communication environments, a noise-aware CSI quantizer is further introduced to improve robustness under imperfect CSI conditions. Finally, a low-complexity two-layer message passing detector with Walsh code constraints (TLMP-WC) is developed to enable efficient large-scale data detection while simultaneously enhancing the detection reliability. Simulation results demonstrate that the proposed AFDM-PCS-IM framework with the TLMP-WC detector achieves superior covertness, improved bit error rate (BER) performance, and reduced computational complexity compared with the state-of-the-art schemes.

I. INTRODUCTION

High-mobility wireless links make reliable covert communication difficult as channels vary rapidly. This section positions AFDM pre-chirp parameters as an information-bearing, CSI-encrypted index dimension, addressing a gap in AFDM security designs.

I. INTRODUCTION

  • AFDM is motivated by high-mobility settings where rapidly varying channels challenge reliable and covert transmission; its chirp structure is described as robust to severe delay and Doppler shifts.Prior AFDM work in the section also covers receivers, channel estimation, waveform efficiency, and index modulation, with communication metrics as the main focus.Source: §I · p. 1, A. Related Work · pp. 1–2

B. Motivation and Contribution

  • Existing AFDM security methods use pre-chirp parameters for permutation, hopping, or secure waveform variables, but not as structured information-bearing codebooks jointly designed with encryption and detection.The paper identifies this joint design as insufficiently investigated.Source: B. Motivation and Contribution · p. 2
  • AFDM-PCS-IM uses an orthogonal pre-chirp codebook and legitimate-CSI-driven perturbations to encrypt it, making pre-chirp sequence selection an additional index modulation dimension.The approach is presented for reliable transmission and enhanced covertness in high-mobility scenarios.Source: B. Motivation and Contribution · p. 2
  • The framework jointly maps information to constellation symbols and selected pre-chirp patterns, exploiting degrees of freedom in the pre-chirp parameter domain.The codebook construction rule generates mutually orthogonal pre-chirp parameter sets.Source: B. Motivation and Contribution · p. 2
  • CSI-driven perturbations encrypt the pre-chirp codebook while channel reciprocity lets legitimate users distinguish intended signals; eavesdroppers without legitimate CSI cannot reliably obtain useful information.The proposed noise-aware quantizer addresses channel-estimation errors and supports robust encrypted-codebook reconstruction at the legitimate receiver.Source: B. Motivation and Contribution · p. 2
  • The work also develops a low-complexity two-layer message-passing detector with Walsh code constraints for large-scale detection, reporting reliability gains with reduced computational complexity.The introduction states that simulations verify superior covertness, reliability, and robustness against state-of-the-art methods.Source: B. Motivation and Contribution · pp. 2–3

II. SYSTEM MODEL

The system model uses reciprocal CSI estimates at Alice and Bob, while Eve cannot reliably acquire the legitimate CSI. It then defines AFDM-PCS-IM transmission, the doubly selective channel, and receiver detection using the pre-chirp codebook.

II. SYSTEM MODEL

  • Alice and Bob estimate the reciprocal legitimate channel from bidirectional pilots, whereas Eve, at a different position, cannot reliably acquire that CSI in high-mobility scenarios.Neither instantaneous CSI nor the CSI-driven perturbation sequence needs explicit exchange between Alice and Bob.Source: §II · p. 3

A. Fundamentals of AFDM

  • AFDM is a DAFT-based waveform with adjustable post-chirp and pre-chirp parameters; the post-chirp choice supports full delay-Doppler diversity and reduces fractional-Doppler performance loss.The paper varies pre-chirp parameters across chirp subcarriers to exploit degrees of freedom left unused when conventional AFDM uses a shared parameter.Source: A. Fundamentals of AFDM · p. 3

B. Transmitter of the Proposed AFDM-PCS-IM

  • Each transmitter group selects one of K pre-chirp sequences for its index bits and spreads a constellation symbol across its M chirp subcarriers.The transmitter groups the N chirp subcarriers into G groups, and divides each group's bits between index and modulated bits.Source: B. Transmitter of the Proposed… · p. 3

C. Channel Model

  • The channel model represents each generalized LTV path by gain, normalized delay, and normalized Doppler, with Doppler separated into integer and fractional components.The model also describes a structured sparse DAFT-domain channel representation for path separation and low-complexity detection.Source: C. Channel Model · p. 4, §2 · p. 4

D. Receiver of the Proposed AFDM-PCS-IM

  • At reception, Bob and Eve equalize and perform maximum-likelihood detection over symbols and pre-chirp sequences; Eve's effective codebook is approximately disjoint from the legitimate codebook.The text attributes Eve's failure to recover transmitted bits to this codebook mismatch, while Bob's reliability depends on distinguishability of the pre-chirp sequences over the legitimate channel.Source: D. Receiver of the Proposed AFDM-PCS-IM · pp. 4–5

III. PRE-CHIRP PARAMETER CODEBOOK DESIGN AND ENCRYPTION PRINCIPLES

The section designs a pre-chirp codebook to support both reliable detection by Bob and secrecy against Eve. Walsh-Hadamard sequences provide orthogonality, while common CSI-driven perturbations encrypt the codebook without disrupting that structure.

A. Reliability Requirement and Orthogonal Codebook Construction

  • Walsh-Hadamard sequences construct an orthogonal pre-chirp codebook, meeting the reliability goal of distinguishing sequences at Bob.The design criterion seeks mutually uncorrelated effective matrices to minimize pairwise error probability.Source: A. Reliability Requirement and… · p. 5
  • The Walsh-to-pre-chirp mapping preserves mutual orthogonality, with the zeroth chirp subcarrier set to zero to avoid a singularity.The Walsh-Hadamard structure also supports alternative orthogonal or low-correlation codebooks, although residual correlation may reduce index-pattern distinguishability.Source: A. Reliability Requirement and… · p. 5, B. Secrecy Requirement and CSI-Driven… · p. 5

B. Secrecy Requirement and CSI-Driven Perturbation Design

  • A CSI-driven perturbation adds uncertainty for Eve, while a common public baseline and perturbation preserve codebook orthogonality.The perturbation uses a sign sequence derived from legitimate CSI; the public irrational baseline is selected offline.Source: B. Secrecy Requirement and CSI-Driven… · pp. 5–6

IV. ROBUSTNESS MECHANISM BASED ON CSI QUANTIZATION

A noise-aware quantizer addresses mismatches between Alice’s and Bob’s imperfect CSI estimates. Its noise-scaled dead zone and decision intervals support identical perturbations when the estimates agree within a quantization interval.

IV. ROBUSTNESS MECHANISM BASED ON CSI QUANTIZATION

  • The quantizer scales its dead-zone width d and step size Δ with CSI estimation-error variance to address codebook mismatch.The passage specifies d = kσ_h and Δ = γσ_h, with adjustable k and γ.Source: §IV · p. 6
  • A zero-output dead zone suppresses unreliable low-magnitude CSI components, and channel estimation is repeated if all CSI metrics fall within it.Individual zero outputs are allowed; the repeat condition is max_{1≤n≤N−1}|μ̂[n]| < d.Source: §IV · p. 7
  • Alice and Bob generate the same encrypted codebook when their quantized CSI values produce identical perturbations.This holds when their local estimates fall within the same decision interval and have the same sign.Source: §IV · p. 7

V. PROPOSED TWO-LAYER MESSAGE PASSING DETECTOR WITH WALSH CODE CONSTRAINTS

The section introduces a low-complexity TLMP-WC detector for large-dimensional data detection. It outlines an equivalent received-signal model, the detector principle, and a computational-complexity analysis.

V. PROPOSED TWO-LAYER MESSAGE PASSING DETECTOR WITH WALSH CODE CONSTRAINTS

  • The detector section derives an equivalent received-signal model to explain the proposed TLMP-WC detector.The section then presents its detection principle for large-dimensional AFDM-PCS-IM data and analyzes computational complexity.Source: §V · p. 7

A. Equivalent Received Signal Model for Walsh-Constrained Detection

The received-signal model is reformulated to expose the Walsh-Hadamard structure of the transmitted signal. This yields a block-sparse representation that supports Walsh-constrained detection.

A. Equivalent Received Signal Model for Walsh-Constrained Detection

  • The model represents the transmitted signal as block-sparse, with one activated index in each group.The transformed signal z is obtained through a block-diagonal Walsh-Hadamard transform, and its elements retain the symbol alphabet.Source: A. Equivalent Received Signal Model… · p. 7
  • The optimal MAP detector’s complexity grows exponentially with signal dimension N, motivating a low-complexity suboptimal MAP approach.Because z = W_eff x′ is a deterministic one-to-one transformation, the joint posterior is equivalent to Pr(z | y_B).Source: A. Equivalent Received Signal Model… · p. 7
  • Walsh-constrained block-sparse priors restrict each group to a single valid activation pattern, reducing the effective detection search space.The formulation exploits the deterministic Walsh-Hadamard relationship between x′ and z.Source: A. Equivalent Received Signal Model… · pp. 7–8

B. Proposed TLMP-WC Detector

TLMP-WC combines Gaussian message-passing linear detection with Walsh-constrained grouped local MAP detection, exchanging soft information iteratively. Its update and stopping rules use a convergence indicator to regulate the iteration and produce final symbol decisions.

B. Proposed TLMP-WC Detector

1) Gaussian Message Passing-Based Linear Detection (GMP-LD):

  • The GMP-LD layer estimates symbol posteriors from received observations, approximating other-node interference as Gaussian before sending means and variances to the second layer.Its variable-node priors come from the previous iteration; the first iteration uses uniform probabilities.Source: 1) Gaussian Message Passing-Based… · p. 8

2) Walsh-constrained Grouped Local MAP Detection (WC-GLMD):

  • WC-GLMD exploits the Walsh block-sparse constraint: each group has one nonzero constellation symbol, so it evaluates valid activation patterns and returns hard and soft estimates.There are M|S| valid patterns per group.Source: 2) Walsh-constrained Grouped Local MAP… · pp. 8–9

3) Convergence indicator:

  • The convergence indicator ψ measures the proportion of symbol nodes whose posterior probabilities concentrate on one constellation point; larger values indicate a more stable detection state.Concentration is assessed against a predefined threshold ε.Source: 3) Convergence indicator · p. 9

4) Update criteria:

5) Stopping criteria and final decision:

  • Iteration stops when ψ reaches one or the maximum iteration count is reached, after which MAP decisions recover symbols and the Walsh mapping recovers the sparse vector.The final symbol decision selects the constellation point with the greatest posterior probability.Source: 5) Stopping criteria and final decision · p. 10

C. Computational Complexity Analysis

The analysis counts receiver-side FLOPs for TLMP-WC, GMP, and MMSE-ML under a stated operation-counting convention. With fixed graph degrees, group size, constellation size, and iteration counts, TLMP-WC and GMP scale linearly with N, while MMSE-ML is dominated by a cubic operation.

C. Computational Complexity Analysis

  • FLOP counts follow the stated convention: complex multiplication costs six FLOPs, while complex addition and complex-by-real multiplication each cost two.Real multiplication, addition, subtraction, and division each count as one FLOP.Source: C. Computational Complexity Analysis · p. 10
  • TLMP-WC’s per-iteration complexity is the sum of GMP-LD and WC-GLMD costs, which are calculated from sparse-channel edges, groups, constellation size, and related dimensions.The analysis gives separate expressions for the two layers and multiplies their combined cost by the number of TLMP-WC iterations.Source: C. Computational Complexity Analysis · p. 10
  • For fixed graph degrees, group size, constellation size, and iteration counts, TLMP-WC and GMP scale linearly with N, whereas MMSE-ML is dominated by the cubic MMSE operation.The comparison counts dominant receiver-side FLOPs and excludes exponentials, comparisons, indexing, final decisions, and one-time initialization.Source: C. Computational Complexity Analysis · p. 10

VI. SIMULATION RESULTS AND DISCUSSIONS

Simulations assess AFDM-PCS-IM’s reliability, covertness, quantizer, and detector in high-mobility channels. The scheme and TLMP-WC show favorable BER and complexity comparisons, while CSI quantization improves legitimate-link robustness under estimation error.

A. Effectiveness of The Proposed Codebook and Quantizer Design

  • With the encrypted pre-chirp codebook, Bob’s BER stays close to the fixed-parameter configuration across CSI errors, while Eve’s BER retains a high error floor.The paper attributes Eve’s near-random-guessing demodulation to her inability to acquire the codebook shared by Alice and Bob.Source: Figure 3 · p. 11, A. Effectiveness of The Proposed… · p. 11
  • At σ2 h = 10^-4, CSI quantization raises Bob’s pre-chirp parameter match rate from approximately 0.6 to 0.94, a gain of about 57%.The quantizer slows the match-rate decline as CSI estimation error increases.Source: Figure 4 · p. 11, A. Effectiveness of The Proposed… · p. 11
  • With imperfect CSI, the quantizer substantially improves Bob’s BER and suppresses high-SNR error floors, while introducing no noticeable BER change at Eve.With perfect CSI, the BER curves with and without quantization almost coincide.Source: Figure 5 · p. 11, A. Effectiveness of The Proposed… · pp. 11–12
  • The scaling factor α trades reliability against security: around α = 0.1, Bob’s BER remains low while Eve’s BER is significantly elevated.The paper reports limited security enhancement for α < 0.05 and increased Bob BER when α > 0.2.Source: Figure 6 · p. 12, A. Effectiveness of The Proposed… · p. 12
  • Across the tested CSI error range, AFDM-PCS-IM has lower BER than AFDM and AFDM-IM; near σ2 h = 10^-3, it reduces BER by about one order of magnitude versus AFDM.When σ2 h exceeds approximately 4 × 10^-3, AFDM-IM’s BER rises more rapidly and eventually becomes worse than conventional AFDM.Source: A. Effectiveness of The Proposed… · p. 12

B. Effectiveness of the Proposed TLMP-WC Detector

VII. CONCLUSION

The proposed AFDM-PCS-IM framework uses a CSI-encrypted pre-chirp codebook for covert high-mobility communication, with noise-aware quantization and TLMP-WC detection. Simulations report improvements in covertness, BER, and robustness, while the design balances reliability, security, and detection complexity.

VII. CONCLUSION

  • Simulations under high-mobility and imperfect-CSI conditions show superior covertness, improved BER, and enhanced robustness over conventional AFDM and AFDM-IM.The framework’s overall design tradeoff is divided between reliability-security balance and BER-complexity balance.Source: §VII · p. 14
  • CSI-driven perturbations encrypt an orthogonal pre-chirp codebook, which becomes an additional IM dimension for implicit data encryption without explicit secret-key exchange.The modulation structure is coupled with instantaneous legitimate-channel characteristics.Source: §VII · p. 14
  • Noise-aware CSI quantization mitigates channel-estimation errors to support reliable encrypted-codebook reconstruction under imperfect CSI.Quantization and perturbation parameters balance reliable reconstruction at Bob against preventing Eve from reconstructing the legitimate codebook.Source: §VII · p. 14
  • TLMP-WC supports large-scale data detection, achieving improved BER with fewer FLOPs than the considered benchmark detectors.The receiver design determines the BER-complexity tradeoff.Source: §VII · p. 14
  • Joint configuration provides a favorable balance among reliability, security, and detection complexity.The paper identifies multi-user scenarios, including user-specific codebook design and multi-user detection, as future work.Source: §VII · p. 15
Loading 2611.15045v1…