Subject: Cutting-Edge Advancements in Wireless Communication, Sensing, and Signal Processing

Hi Elman,

This newsletter dives into the latest breakthroughs in wireless communication, sensing, and signal processing, focusing on the exciting realms of intelligent surfaces, semantic communication, and deep learning.

General Overview

This collection of papers explores cutting-edge advancements in wireless communication, sensing, and signal processing, with a particular emphasis on intelligent surfaces, semantic communication, and deep learning applications. Several works focus on integrated sensing and communication (ISAC), proposing novel beamforming strategies and architectures. Jiang et al. (2025) introduce a movable antenna (MA) framework for ISAC, optimizing sensing signal-to-interference-plus-noise ratio (SINR) while maintaining communication quality of service. Similarly, Yang et al. (2025) leverage deep reinforcement learning (DRL) for intelligent antenna positioning in fluid antenna systems (FAS)-aided ISAC, demonstrating scalability and improved performance in multi-target sensing scenarios. Zhang et al. (2025) address the challenges of wideband near-field communications with reconfigurable refractive surfaces (RRS), proposing a delayed-RRS structure to mitigate beam-split effects caused by frequency selectivity and near-field propagation. The fundamental limits of ISAC are explored by Wang & Wang (2025), who derive optimal input and output distributions for modified minimum-mean-square-error (MMSE) and Shannon rate, providing insights into the inherent trade-offs between sensing and communication performance.

Semantic communication emerges as another prominent theme, with several papers investigating its security and efficiency. Zhou et al. (2025) propose ROME, a robust model ensembling framework to defend against semantic jamming attacks, adapting its robustness based on the detected attack power. Xie et al. (2025) investigate power-efficient optimization for coexisting semantic and bit-based users in non-orthogonal multiple access (NOMA) networks, proposing clustered frameworks and adaptive multiple access schemes. Meng et al. (2025) provide a comprehensive survey of secure semantic communication, highlighting security and privacy concerns across the SemCom lifecycle and summarizing mitigation techniques. Furthering the exploration of semantic communication, Zhang et al. (2025) introduce semantics-guided diffusion DeepJSCC (SGD-JSCC), leveraging semantic information and diffusion models for enhanced image transmission robustness in varying channel conditions.

Beyond ISAC and semantic communication, the papers also delve into diverse areas such as localization, channel estimation, and signal processing techniques. Inamdar (2025) presents exact linear solutions for the general 3D time difference of arrival (TDOA) source localization problem, offering computationally efficient and accurate localization capabilities. Zhao et al. (2025) introduce a confined orthogonal matching pursuit (OMP) algorithm for sparse signal recovery with reduced complexity. Oh & Choi (2025) propose a scalable beamforming design for multi-RIS-aided MU-MIMO systems with imperfect CSIT, utilizing a generalized power iteration approach. Xia (2025) re-examines delay Doppler channels and time-frequency coding, arguing against the effectiveness of current modulation schemes in compensating for non-trivial Doppler spread.

Several papers explore the application of deep learning in various domains. Dip et al. (2025) propose NeuroSleepNet, a multi-head self-attention based automatic sleep scoring scheme. Liu et al. (2025) introduce a model-driven deep neural network for enhanced AoA estimation using 5G gNB. Khalfin Fekson et al. (2025) analyze data-driven techniques for improving neural inertial regression networks. Liu et al. (2025) explore multipath component-aided signal processing for ISAC. Ge et al. (2025) propose linear precoding solutions for OTFS systems. Ragab et al. (2025) investigate automotive speed estimation using various sensor types. Aboueleneen et al. (2025) use multi-agent deep RL for distributed traffic control. Abanto-Leon & Maghsudi (2025) address hierarchical functionality prioritization in multicast ISAC. Osman & Nadeem (2025) present LoRaFlow for high-quality LoRa signal reconstruction.

Collectively, these papers highlight the ongoing research efforts to develop advanced signal processing and communication techniques for next-generation wireless systems. The focus on intelligent surfaces, semantic communication, and deep learning reflects the growing importance of these technologies in addressing the challenges of increasing data rates, spectral efficiency, and robust performance in complex wireless environments. The diverse range of applications explored, from localization and sensing to traffic control and healthcare, underscores the broad potential impact of these advancements.

Paper Highlights

Fundamental MMSE-Rate Performance Limits of ISAC Systems

Fundamental MMSE-Rate Performance Limits of Integrated Sensing and Communication Systems by Zijie Wang, Xudong Wang https://arxiv.org/abs/2501.01053

Caption: MMSE-Rate Trade-off in an ISAC System

This paper provides a rigorous investigation into the fundamental performance limits of Integrated Sensing and Communication (ISAC) systems, a critical technology for future 6G networks. Unlike previous studies that relied on simplifications, this work tackles the complex problem of Pareto stochastic optimization by considering a random ISAC signal designed for both sensing channel estimation and information transmission. The authors utilize modified Minimum Mean-Square Error (MMSE) and Shannon rate as the respective sensing and communication (SAC) performance metrics.

The study derives conditions for optimal channel input and output distributions on the MMSE-Rate limit using variational approaches, resulting in high-dimensional convolutional equations. A detailed analysis of the single-input-single-output (SISO) ISAC case reveals a key insight: the MMSE-Rate limit is achievable only at the sensing-optimal and communication-optimal operating points, effectively making it a supremum limit. This suggests a similar unattainability for general MIMO systems, although a rigorous proof remains an open problem due to the high dimensionality.

To address this complexity, a novel Blahut-Arimoto-type algorithm is proposed to numerically determine the optimal input distribution on the MMSE-Rate limit and compute SAC performance. The algorithm's convergence to the limit is proven, providing a practical tool for evaluating ISAC performance. Furthermore, closed-form SAC-optimal waveforms are derived for scenarios with multiple widely-separated sensing receivers. These waveforms are characterized by power allocation according to channel statistics/realization (water-filling tradeoff) and waveform selection (waveform uncertainty tradeoff), correcting some previously established loose bounds.

For scenarios where sensing and communication channels coincide (G = H), a compound signaling strategy is introduced. This strategy sequentially employs sensing-optimal waveforms for channel estimation and communication-optimal waveforms for data transmission over the estimated channel. Numerical results demonstrate significant rate improvements over non-coherent "capacity," highlighting the benefits of ISAC integration.

Uncovering the Iceberg in the Sea: Fundamentals of Pulse Shaping and Modulation Design for Random ISAC Signals

Uncovering the Iceberg in the Sea: Fundamentals of Pulse Shaping and Modulation Design for Random ISAC Signals by Fan Liu, Yifeng Xiong, Shihang Lu, Shuangyang Li, Weijie Yuan, Christos Masouros, Shi Jin, Giuseppe Caire https://arxiv.org/abs/2501.01721

Caption: Impact of Coherent Integration and "Iceberg Shaping" on Ambiguity Function

This paper delves into the critical aspect of exploiting the inherent randomness of communication signals for sensing in Integrated Sensing and Communication (ISAC) systems, a key technology for 6G. The authors focus on pulse shaping and modulation design to improve target ranging performance by reshaping the statistical properties of the auto-correlation function (ACF).

A key contribution is the derivation of a closed-form expression for the expectation of the squared ACF of random ISAC signals, considering arbitrary modulation bases and constellation mappings within the Nyquist pulse shaping framework. This expression reveals a structure metaphorically described as an "iceberg hidden in the sea," where the "iceberg" represents the squared mean of the ACF, determined by the pulse shaping filter, and the "sea level" characterizes its variance due to data randomness. Mathematically, E(|R_k|²) = |E(R_k)|² + var(R_k), where |E(R_k)|² is the "iceberg" and var(R_k) is the "sea level." The authors show that the "iceberg" corresponds to the squared ACF of the pulse, and coherent integration reduces the "sea level" by a factor of M (integration times).

The paper establishes that for QAM/PSK constellations with Nyquist pulse shaping, Orthogonal Frequency Division Multiplexing (OFDM) uniquely achieves the lowest ranging sidelobe level at every lag. This extends previous work that focused solely on discrete time-domain samples. The optimal modulation basis also depends on the constellation's excess kurtosis (μ₄-2): OFDM is optimal for sub-Gaussian constellations (μ₄ < 2) like QAM/PSK, while single-carrier is preferred for super-Gaussian constellations (μ₄ > 2).

Finally, a novel "iceberg shaping" technique for pulse shaping design is proposed. By recognizing that coherent integration primarily reduces the "sea level," the authors optimize the pulse's ACF sidelobe level within a specified delay region, effectively shaping the "iceberg" itself for significant sidelobe suppression.

Semantics-Guided Diffusion for Deep Joint Source-Channel Coding in Wireless Image Transmission

Semantics-Guided Diffusion for Deep Joint Source-Channel Coding in Wireless Image Transmission by Maojun Zhang, Haotian Wu, Guangxu Zhu, Richeng Jin, Xiaoming Chen, Deniz Gündüz https://arxiv.org/abs/2501.01138

Caption: The architecture of Semantics-Guided Diffusion DeepJSCC (SGD-JSCC) depicts the integration of a Diffusion Transformer (DiT) within a Joint Source-Channel Coding (JSCC) framework. The DiT leverages semantic information from a CLIP model and side information interpolator, guiding the image reconstruction process through multiple DiT blocks and a linear projection before decoding. The DiT block detail showcases multi-head self-attention, cross-attention with CLIP features, and scaling mechanisms influenced by channel conditions.

This paper introduces Semantics-Guided Diffusion DeepJSCC (SGD-JSCC), a novel approach to enhance wireless image transmission by leveraging diffusion models (DMs) and semantic side information within a Deep Joint Source-Channel Coding (DeepJSCC) framework. SGD-JSCC addresses the challenges of existing DM integrations with JSCC, such as the randomness of generative processes and adaptation to dynamic channel conditions.

SGD-JSCC's innovation lies in its two key features. First, it incorporates semantic side information like text descriptions or edge maps to guide the DM denoising process, transforming unconditional denoising into a conditional one and ensuring semantic alignment. Second, it introduces a channel-tailored DM that adapts to varying channel conditions. In slow fading channels, SGD-JSCC dynamically estimates the instantaneous Signal-to-Noise Ratio (SNR) directly from the channel output, eliminating the need for pilot transmissions using the formula m = S⁻¹( σ² / (σ² + |h|²/2) ), where m is the timestep, S is the sigmoid scheduling function, σ² is the noise variance, and h is the channel gain. In fast fading scenarios, a training-free, water-filling-inspired denoising strategy allows adaptation to fluctuating channel gains.

Conclusion

This newsletter highlights a convergence of innovative approaches to enhance wireless communication and sensing. The exploration of fundamental ISAC performance limits provides a theoretical foundation for optimizing joint sensing and communication systems, while the novel pulse shaping techniques offer practical solutions for improving ranging capabilities by leveraging signal randomness. Furthermore, the introduction of semantics-guided diffusion in DeepJSCC demonstrates the potential of incorporating semantic information and advanced deep learning models to achieve robust and high-fidelity image transmission in challenging wireless environments. These advancements collectively pave the way for more efficient, reliable, and intelligent next-generation wireless systems.