A Triple-Layer HFM-LFM-CAZAC Preamble Framework for Underwater Acoustic Integrated Sensing and Communication

Area of Science:

• Underwater Acoustic (UWA) Communications
• Integrated Sensing and Communication (ISAC)
• Signal Processing
• Waveform Design

Background:

• Existing UWA systems often require separate hardware for sensing and communication, increasing complexity and cost.
• There is a need for efficient waveforms that can simultaneously support sensing (e.g., ranging) and communication functions in UWA environments.
• Challenges include Doppler effects, multipath propagation, and interference in UWA channels.

Purpose of the Study:

• To propose and evaluate a novel three-functional-layer decomposition framework for UWA ISAC preambles, named P5.
• To assess P5's performance in terms of sensing accuracy, communication capabilities, and security against potential adversaries.
• To compare P5 against other waveforms in various underwater acoustic environments.

Main Methods:

• Developed a P5 preamble using a three-layer decomposition: hyperbolic frequency modulation (HFM) for Doppler invariance, linear frequency modulation (LFM) for ranging, and a constant-amplitude zero-autocorrelation (CAZAC) envelope for cell identification and despreading.
• Utilized closed-form screening conditions and direct cross-ambiguity measurements to ensure near-orthogonality between the chirp layers.
• Conducted matched-filter Monte Carlo simulations to evaluate P5's ranging accuracy (σR) and performance against classical structure-aware attackers, assessing low-probability-of-intercept (LPI) capabilities.
• Employed a ResNet-18 deep learning model to determine detection probabilities (Pd) against P5 under varying signal-to-noise ratios (SNR).
• Analyzed performance trade-offs, including composite peak-sidelobe level (CPSL) and cell identification accuracy under interference.
• Performed replay attacks using the WATERMARK dataset to evaluate P5's resilience against replay-based adversaries.

Main Results:

• P5 meets the continuous-sensing target of range root mean square error (σR) ≤ 1 m at 10 dB SNR.
• P5 demonstrated the smallest normalized matched-filter peak loss across twelve modeled UWA environments compared to three other waveforms.
• P5 maintained a low-probability-of-intercept (Pd ≤ 0.1) against four classical attackers at 0 dB attacker-input SNR.
• A ResNet-18 model achieved Pd=0.5 against P5 at +21.24 dB total-energy SNR, indicating a lower bound on adversary capability.
• Performance trade-offs were observed, with P5 exhibiting inferior composite peak-sidelobe levels compared to single-waveform baselines and reduced cell identification accuracy under near-far interference.

Conclusions:

• The proposed P5 waveform offers a promising solution for UWA ISAC, effectively integrating wideband Doppler invariance, sub-meter ranging, and robust security features.
• P5 shows superior performance in sensing accuracy and LPI characteristics compared to other tested waveforms in simulated UWA environments.
• While integration introduces performance trade-offs, P5's benefits in combined sensing and communication capabilities warrant further investigation, with future work focusing on real-world sea trials.