This newsletter explores a collection of papers focused on advanced signal processing and machine learning techniques across diverse applications. A recurring theme is the exploitation of structural information within data, ranging from the geometric structure of covariance matrices for radar clutter detection to the spatiotemporal dynamics of EEG signals for brain network analysis and the sparsity patterns in communication channels for efficient estimation.
For example, Hua et al. (2024) Hua et al. (2024) introduce LDA-MIG detectors. These detectors leverage matrix information geometry and linear discriminant analysis on the manifold of Hermitian positive-definite matrices to enhance maritime target detection in nonhomogeneous sea clutter. Li et al. (2024) Li et al. (2024) propose a grid evolution method for doubly fractional channel estimation in OTFS systems, dynamically adapting the delay-Doppler grid for improved accuracy and computational efficiency. Shao et al. (2024) Shao et al. (2024) exploit directional sparsity in their distributed channel estimation method for 6D movable antennas to reduce processing complexity.
Emerging technologies also take center stage. Jin et al. (2024) Jin et al. (2024) investigate cross-modality signal injection attacks, specifically the PhantomLiDAR attack, targeting LiDAR systems used in autonomous driving, highlighting the vulnerability of these sensors to electromagnetic interference and the urgent need for robust defenses. In 6G communications, Azizi et al. (2024) Azizi et al. (2024) explore HAPS-RIS integration in UAV-based networks, proposing a hierarchical optimization framework for maximizing coverage while minimizing the number of required UAVs. Ma et al. (2024) contribute two papers on massive MIMO systems, one on precoding designs with low-resolution ADCs (Ma et al., 2024) and another on max-min fairness beamforming in JCAS systems (Ma et al., 2024), both employing model-based machine learning for efficient optimization.
Machine learning applications extend beyond communication systems. Neumann et al. (2024) Neumann et al. (2024) introduce confounder-adjusted covariance estimation for structural health monitoring, mitigating the impact of environmental factors on damage detection. Ranjan and Kumar (2024) Ranjan & Kumar (2024) utilize crossplot transition entropy to analyze brain network organization in Alzheimer's disease and frontotemporal dementia based on EEG data. Li et al. (2024) Li et al. (2024) propose a deep learning model for translating mental imaginations into characters using EEG signals, offering a potential communication pathway for individuals with motor impairments.
Several papers address specific signal processing challenges. Eckrich et al. (2024) Eckrich et al. (2024) investigate fronthaul-constrained distributed radar sensing. Gu et al. (2024) Gu et al. (2024) introduce oversampled low ambiguity zone sequences for doubly selective channel estimation. Malekian and Maleki (2024) Malekian & Maleki (2024) explore speckle noise mitigation. Stroschein (2024) Stroschein (2024) uses prolate spheroidal wave functions for high-precision frequency estimation. Guo et al. (2024) Guo et al. (2024) present sub-Nyquist spectral estimation.
Finally, several papers introduce new datasets and frameworks, including the ECG-Image-Database by Reyna et al. (2024) Reyna et al. (2024), random codebooks for audio neural autoencoders by Giniès et al. (2024) Giniès et al. (2024), a deep learning model for THz channel estimation by Tarafder et al. (2024) Tarafder et al. (2024), and SEE, a semantically aligned EEG-to-text translation method by Tao et al. (2024) Tao et al. (2024). Other contributions include work on multi-agent Q-learning for wireless networks Bozkus & Mitra (2024), predictive covert communication Krishnan et al. (2024), a hardware-in-the-loop framework for wireless AI Redondo et al. (2024), calibration challenges in RIS-aided ISAC systems Ghazalian et al. (2024), secure enhancement for RIS-aided UAV with ISAC Xiu et al. (2024), near-field MIMO channel modeling Delbari et al. (2024), spatial modulation in LEO satellite communications Zhang et al. (2024), soil moisture sensing Zhang et al. (2024), and information freshness in mobile gossip networks Srivastava & Ulukus (2024).
Secure Enhancement for RIS-Aided UAV with ISAC: Robust Design and Resource Allocation by Yue Xiu, Wanting Lyu, Phee Lep Yeoh, Yonghui Li, Yi Ai, Ning Wei https://arxiv.org/abs/2409.16917
This paper delves into the security challenges and potential solutions within the increasingly complex landscape of RIS (Reconfigurable Intelligent Surface)-aided UAV (Unmanned Aerial Vehicle) communication systems integrated with ISAC (Integrated Sensing and Communications). The authors construct a scenario involving a multi-antenna UAV tasked with both sensing an untrusted target and communicating with a ground IoT device, all while contending with a potential eavesdropper.
The research acknowledges the practical limitations of obtaining perfect Channel State Information (CSI) for the eavesdropper and untrusted target, opting for a deterministic model to encapsulate CSI uncertainties. This approach adds a layer of realism to the problem formulation. The primary goal is to maximize the average worst-case secrecy rate. This is achieved through the joint optimization of several key parameters: the UAV's trajectory, the passive beamforming of the RIS, and both the transmit and receive beamforming. This joint optimization is a complex undertaking due to the inherent coupling between these variables and the added complexity of CSI uncertainties.
To tackle this intricate optimization problem, the authors employ a Block Coordinate Descent (BCD) method. This approach breaks down the problem into smaller, more manageable sub-problems. Each sub-problem is then addressed using a combination of Successive Convex Approximation (SCA) and Semidefinite Relaxation (SDR) techniques. Specifically, the original problem is decomposed into four sub-problems focusing on transmit beamforming, RIS beamforming, UAV trajectory, and receive beamforming, respectively. The transmit beamforming optimization leverages the triangle inequality to derive the optimal solution. The second sub-problem, dealing with RIS beamforming, utilizes SCA and SDR to handle the CSI uncertainty. SCA is also employed for optimizing the UAV trajectory, while the Majorize-Minimization (MM) algorithm is applied to the receive beamforming optimization.
The paper presents simulation results that validate the effectiveness of the proposed algorithm. These results demonstrate notable improvements in the secrecy rate as the UAV's transmit power increases. Furthermore, the simulations confirm the accuracy of the employed approximations even with a relatively limited number of RIS reflectors (in the tens). Crucially, the study reveals a fundamental trade-off between maximizing secrecy rates and maintaining satisfactory sensing performance. As the demand for higher sensing rates increases, the achievable secrecy rate tends to decrease, highlighting the need for strategic resource allocation to balance these competing objectives. The specific results mentioned, such as achieving an average secrecy rate of around 17 bps/Hz with Nt = 64, NR = 16, and γ = 0 dB, and the observed impact of increasing transmit power and the number of RIS reflectors, provide concrete evidence of the algorithm's performance.
This newsletter highlights the growing trend of leveraging structural information and advanced signal processing techniques to address challenges in diverse fields. From maritime target detection and LiDAR security to channel estimation and brain network analysis, researchers are finding innovative ways to extract meaningful insights from complex data. The highlighted paper on secure enhancement for RIS-aided UAV with ISAC demonstrates a sophisticated approach to optimizing resource allocation in a challenging communication scenario, emphasizing the importance of considering both security and sensing performance in system design. The collective efforts presented in this newsletter significantly advance the state-of-the-art in signal processing, communications, and related fields, paving the way for future breakthroughs.