Systems and Control · Cross-listed from cs.AI
Cutting AI Datacenter Energy with Reinforcement Learning: Measured Power Control of LLM Training from One GPU to the Fleet
Eliseo Curcio
Reinforcement-learning post-training dominates modern
language-model development, yet its power behavior on GPU hardware has not been characterized, and datacenters manage GPU power with workload-blind mechanisms, static caps and reactive throttling, that slow hardware indiscriminately. We instrument GRPO training with half-second power telemetry at 7B, 14B, and 72B scales on one to four A100s (380,000+ samples), and train a PPO meta-controller that adapts the workload's own generation parameters to measured power. Against the full 500-step 7B trace, the controller cuts power-limit violations by 89.8% while increasing
token output by 18.1% and energy efficiency by 26.2% (
tokens per MWh). Deployed live at 72B, the same controller family yields replicated null results, diagnosed as the group-size actuator losing authority under model sharding. An actuator-authority sweep shows the same parameters applied as generation concurrency retain 17-22% power authority, isolating an occupancy-versus-volume principle; a controller rebuilt on that actuator controls a live 72B rollout-generation workload across three replications: 35.7% more output than a static safe baseline at 2.27 +/- 1.08% budget violations, 87.2% fewer violations than uncontrolled operation, and the best mean throughput and energy per
token among constrained controllers, with an adaptive threshold rule matching it in one of three operating conditions. Under realistic measurement windows the original 72B transients fall from 23.6% at half-second resolution to 1.6% at 30 s and zero at 5 min; a composed 16-GPU fleet shows zero violations at 30 s and longer, with peak demand at 50-56% of nameplate. For this fleet mix, roughly twofold oversubscription of nameplate appears feasible, subject to operator validation. We quantify the economic and carbon consequences and specify a low-cost operator pilot.
Read the full abstract ▾
Signal Processing
Automated ECG Interval Measurement and Wave Delineation Using Fast Fourier Convolution ResNet
Farhan Adam Mukadam, Harshit Mishra, Nachiket Makwana et al. (+3 more)
Accurate measurement of ECG intervals, including PR, QRS duration, and QT/QTc, is central to cardiac diagnosis, yet the published ECG delineation literature evaluates performance almost exclusively as fiducial-point timing errors on small curated databases, rather than as clinical interval accuracy on large unselected cohorts. We bridge this gap by evaluating a complete end-to-end pipeline on 10,646 clinical 12-lead ECGs and reporting the first large-scale interval measurement accuracy study with full statistical characterisation, including bias, 95% limits of agreement (Bland-Altman), bootstrap confidence intervals, and rhythm-stratified error analysis. The underlying delineation is performed by a Fast
Fourier Convolution ResNet (FFCResNet), adapting local temporal convolutions with global spectral processing via FFT and augmented with register
tokens for contextual feature learning. Three per-wave models (P, QRS, and T) are trained on six public databases with ECG-specific augmentation. On 10,646 ECGs, the system achieves a QT MAE of 17.5 ms [95% CI: 16.9-18.2], with a Bland-Altman bias of +8.5 ms (LoA: -68.5 to +85.5 ms); a QRS duration MAE of 14.8 ms [95% CI: 14.6-15.0], with a bias of +12.6 ms (LoA: -12.3 to +37.6 ms); and a ventricular rate MAE of 0.8 beats/min. All biases are statistically significant by the Wilcoxon signed-rank test (p < 0.001) but remain within or near published inter-observer variability bounds for sinus rhythms. Rhythm-stratified analysis reveals substantially higher QT errors for supraventricular tachycardias (SVT MAE: 75.0 ms; AVRT MAE: 85.3 ms) than for sinus bradycardia (SB MAE: 9.3 ms) and sinus rhythm (SR MAE: 8.9 ms), providing an honest characterisation of the deployment scope. Wave
segmentation achieves internal Dice scores of 95.5%, 98.2%, and 96.1% for P, QRS, and T waves, respectively, and cross-database Dice scores of 78.1%, 85.5%, and 74.2%.
Read the full abstract ▾
All authors Farhan Adam Mukadam, Harshit Mishra, Nachiket Makwana, Pradyot Tiwari, Subramani Kandasamy, KVS Hari
Signal Processing
Domain-Generalized Adaptive Semantic Communication for Collaborative Perception
Fan Gao, Youzheng Wang and Ning Ge
We propose RSTA, a domain-generalized semantic communication framework enabling source-free V2X collaborative perception under both observation-domain shift and unseen wireless channel conditions. In V2X, received semantic
tokens suffer coupled degradation from pre-transmission domain drift and in-transit channel corruption; existing methods address only one source, leaving adaptation misled by
tokens that are simultaneously off-domain and physically degraded. RSTA trains a pre-deployment semantic encoder for transmission stability via cross-domain prototype
alignment and cross-channel gradient consistency, and updates a lightweight in-deployment decoder adapter through reliability-gated
entropy minimization that restricts gradients to
tokens ranked high in both semantic relevance and channel fidelity. A theoretical task robustness decomposition links each loss term to a distinct degradation source, grounding each algorithmic component in a measurable error mode. Trained on AWGN and tested on unseen Rayleigh fading, RSTA achieves +7.2 AP@0.7 over pre-deployment domain generalization on cross-weather tasks and +5.5 on cross-
dataset tasks across four V2X
benchmarks, updating only 0.21% of parameters in-deployment with zero inter-
agent synchronization overhead.
Read the full abstract ▾
Image and Video Processing
Shape-Based Inductive Bias for Glioma Grading from Tumor Contours
Puneet Velidi, Michelle F. Miranda, Farouk Nathoo et al. (+2 more)
Glioma grading from tumor contours is often treated as a pixel problem even when the signal of interest is shape. We align closed contours with a functional shape-
alignment framework, separate global deformation from residual
Fourier shape, and organize these quantities as frequency-ordered
tokens. In five-fold patient-disjoint cross-validation on BraTS~2020 tumor contours, with model selection performed using grouped inner validation, a compact multilayer perceptron (MLP) achieves the highest mean balanced accuracy at 71.5%, compared with 65.9% for ResNet-18 and 63.3% for ViT-Tiny. It also gives the highest mean low-grade glioma F1 at 54.9%. Its pooled out-of-fold balanced accuracy is 72.4% (patient-bootstrap 95% CI: 66.4–77.8%). The selected MLPs use 2.9k–117.3k parameters across folds, at least 46 times fewer than the pixel baselines. In a controlled noise-free simulation, shape-based models reach 56.3–71.5% balanced accuracy while the pixel models remain at 50.0–52.5%. This work demonstrates how incorporating a shape-based inductive bias at the representation level can improve interpretability and scalability while enabling substantial dimensionality reduction.
Read the full abstract ▾
All authors Puneet Velidi, Michelle F. Miranda, Farouk Nathoo, Ashery Mbilinyi, Cédric Beaulac
Systems and Control · Cross-listed from cs.SE
Specification-Driven DevOps for Multi-Service Environments
Oleg Grynets, Kyrylo Fursov, Vasyl Lyashkevych et al. (+1 more)
Large
Language Models (LLMs) are increasingly used to generate executable software environments from repository artifacts. However, functional executability does not necessarily imply conformity with architectural, security, workflow, and production intent. This study investigates whether a frontier LLM can generate Dockerfiles and Docker Compose configurations for multi-service applications using repository contents without access to developer-authored deployment artifacts. Three heterogeneous repositories combining Python, Node.js, .NET, React, Rust, Java, Redis, PostgreSQL, and MySQL-compatible infrastructure were evaluated using deterministic end-to-end HTTP oracles and manual structural comparison. All three generated environments became functionally operational, although one required a Rust base-image update from version 1.85 to 1.88. The model correctly reconstructed service
topology, application ports, infrastructure dependencies, service hostnames, a background worker, hidden proxy configurations, and a file-based Docker secrets mechanism. However, it consistently omitted network
segmentation, multi-stage builds, dependency-layer caching, live-reload volumes, production frontend serving, restrictive backend-port policies, and cross-platform build logic. Based on these observations, the study formalizes the distinction between functional correctness and deployment-intent fidelity and analytically derives a minimal explicit deployment specification for information that cannot be reliably inferred from repository artifacts.
Read the full abstract ▾
All authors Oleg Grynets, Kyrylo Fursov, Vasyl Lyashkevych, Volodymyr Veres
Image and Video Processing
Analysis of the Shortcut Learning and Clever Hans Effect in CNN based ECG Image Classification
Abhay Kumar Pathak, Mrityunjay Chaubey, Manjari Gupta et al. (+1 more)
Deep learning models for ECG image classification may achieve high accuracy by exploiting non-physiological visual cues instead of ECG waveform morphology. Given the black-box nature of deep learning models, their promise of high predictive performance often remains insufficiently translated into clinical or real-world trust, interpretability, and actionable decision-making. In this study, we examine shortcut learning and Clever Hans effect in a publicly available ECG image
dataset using convolutional neural networks. In process we have created six image-derived feature sets (FSs), FS1: raw full ECG images, FS2: cropped waveform-only images, FS3: waveform-masked metadata images, FS4: red-arrow artifact images for the myocardial infarction class, FS5: contrast-enhanced images for the abnormal heartbeat class and FS6: Gaussian-blurred images for the normal class. These controlled representations were used to test whether classification performance persists when waveform information is removed or when artificial class-specific artifacts are introduced. Shortcut retention score, prediction consistency and confidence divergence across Feature-Set Representations have been calculated to assess the transparency about the learning pattern. Along with factual results, average Integrated Gradients and occlusion sensitivity test results are presented to inspect whether model attribution focused on ECG-relevant waveform regions or on non-clinical artifacts. Performance changes across feature sets and attribution patterns were used to identify potential Clever Hans behavior. This study evaluates whether ECG image classifiers learn clinically meaningful morphology or shortcut cues introduced by report layout, metadata, contrast, blur, or artificial markers.
Read the full abstract ▾
All authors Abhay Kumar Pathak, Mrityunjay Chaubey, Manjari Gupta, Deepti Mishra
Systems and Control
Integrated Eco-Driving and Powertrain Optimization for Hybrid Vehicles in Complex Urban Traffic
Milad Hasanzadeh, Amin Kargarian and Mehdi Farasat
Urban eco-driving requires the simultaneous planning of speed, acceleration, lane decisions, surrounding-vehicle safety, intersection rules, and vehicle energy use. Existing studies commonly optimize only selected aspects of urban driving, such as longitudinal motion, lane changing, intersection crossing, or powertrain energy management, while treating the remaining decisions separately. Unlike these studies, this paper develops an integrated finite-horizon eco-driving planning framework for urban hybrid vehicles that jointly optimizes traffic behavior and powertrain operation within a unified mixed-integer formulation. The novelty of the proposed method lies in simultaneously modeling longitudinal motion, lane occupancy and changes, car-following and lane-change safety, signalized and unsignalized intersection behavior, and hybrid powertrain operation. The formulation captures lane availability, lane-dependent speed limits, mandatory and emergency lane changes, safe-to-clear signal conditions, downstream clearance, stop-and-yield rules, engine and motor power, battery state of charge, regenerative braking, and fuel consumption. Simulation studies compare the proposed method with rule-following, signal-aware, overtaking-enabled, and kinematic-only optimization baselines. The results demonstrate improved energy performance while maintaining feasible, safe, comfortable, and traffic-rule-compliant urban driving plans.
Read the full abstract ▾
Image and Video Processing · Cross-listed from cs.CV
OPERA: Offline Policy-guided Expert Routing and Adaptation for Universal Biomedical Image Analysis
Zihan Li, Feiyang Liu, Dandan Shan et al. (+2 more)
Biomedical image analysis spans diverse modalities and tasks, yet real-world deployment is hindered by severe distribution shifts across scanners, protocols, and patient populations. High-performing models consequently require repeated domain-specific fine-tuning, which is a costly cycle that becomes impractical when labels are scarce or
privacy constraints limit data sharing. We propose OPERA (Offline Policy-guided Expert Routing and Adaptation), a multi-
agent ensemble framework that addresses this deployment bottleneck by treating expert weight assignment as an offline policy learning problem: a routing policy is learned from a small validation set without gradient updates to any expert
agent, then deployed with test-time adaptation to handle distribution shift. OPERA coordinates heterogeneous specialist
agents through complementary mechanisms. The expert profiling module learns selection policies offline, enabling informed allocation of expertise. Each
agent undergoes confidence calibration through temperature adjustment, ensuring more reliable probabilistic outputs. OPERA also incorporates distribution aware adaptation, where class weights are dynamically adjusted at the batch level using statistics derived from unlabeled test data. Instance level routing assigns each sample to the most suitable expert by leveraging inter model agreement and predictive
entropy. We evaluate OPERA on 9
datasets covering fundus photography, chest X-ray, CT, MRI, and
multimodal diagnostic
benchmarks, comparing against 30+ baselines across classification,
segmentation, and
multimodal settings. OPERA consistently improves performance and calibration quality, demonstrating that offline policy-guided expert
agents coordination is a practical path to deployable biomedical AI without retraining. Code is on \href{https://github.com/HUANGLIZI/OPERA}{GitHub}.
Read the full abstract ▾
All authors Zihan Li, Feiyang Liu, Dandan Shan, Ruibo Wang, Qingqi Hong
Audio and Speech Processing
Text-Prompted CLAP: Learning Text-Conditioned Audio Representations via Contrastive Learning
Mohan Li, Rama Doddipatla and Philip C. Woodland
Contrastive
Language-Audio Pretraining (CLAP) aligns text and audio in a shared
embedding space, but encoding each modality independently limits its ability to model cross-modal semantics in complex audio understanding and
retrieval tasks. To address this limitation, this paper proposes Text-Prompted CLAP (TP-CLAP), a parameter-efficient extension of CLAP that introduces a cross-
attention-based fusion module to incorporate textual
prompts into audio features. TP-CLAP is trained using an audio multiple-choice question answering (AMCQA) framework, where it learns to align text-conditioned audio representations with text
embeddings of correct answer choices via contrastive learning. Experiments demonstrate that TP-CLAP matches or exceeds several substantially larger audio-LLMs on audio question answering despite its compact size. After fine-tuning for attribute-focused audio-to-audio
retrieval, text-conditioned audio representations consistently outperform their unconditioned counterparts on music
retrieval benchmarks. In addition, TP-CLAP improves upon the base CLAP model on conventional audio-text
retrieval and zero-shot classification.
Read the full abstract ▾
Systems and Control · Cross-listed from cs.RO
Input Shaping for Point-to-Point Motion with a Continuum Robot Arm
Rodolfo Hdz. Ibarra, Karan Baker, Parsa Molaei et al. (+2 more)
A cable-driven continuum
robot arm is an underactuated mechanism and may suffer residual vibration at the end of a rest-to-rest maneuver. In this work, a time-delay filter is applied as an input shaper to the system to eliminate the excitation of vibratory modes. A non-robust and a robust time-delay filter are designed based on a linear system model and demonstrate improved response compared to a velocity-driven pulse input. Experimental results using the continuum
robot validate the application of the input shaper, with reduced overshoot and settling time exemplifying the reduction in oscillation at the end of the maneuver. It is also shown that utilizing the robust shaper further improves the response of the arm in comparison to applying the non-robust shaper. These results are significant towards the precise and robust implementation of continuum
robots in applications involving arbitrary end-effector trajectories.
Read the full abstract ▾
All authors Rodolfo Hdz. Ibarra, Karan Baker, Parsa Molaei, Adrian Stein, Hunter B. Gilbert
Signal Processing
Characterizing and Mitigating the Effects of Device Temperature on RF Fingerprinting Accuracy
Haytham Albousayri and Bechir Hamdaoui
Radio Frequency Fingerprinting (RFFP) has emerged as a promising approach for device authentication by exploiting hardware-specific impairments embedded in transmitted signals. Yet existing methods largely overlook a major drawback: RFFP sensitivity to temperature–a critical factor influenced by both internal and environmental conditions–which can significantly alter device signatures and degrade classification performance. In this paper, we propose a novel temperature-aware RFFP framework that explicitly incorporates device temperature information into the learning process to improve robustness and generalization. We evaluate the proposed method on a real-world Bluetooth Low Energy (BLE)
dataset collected across multiple devices and environmental conditions. Experimental results demonstrate that temperature-aware modeling consistently outperforms other temperature mitigation baselines, achieving significant improvements in classification accuracy, particularly under unseen temperature and environmental conditions.
Read the full abstract ▾
Signal Processing · Cross-listed from cs.AI
CogEEGAgent: Toward Autonomous Cognitive EEG Analysis with Grounded Execution and Selection-Aware Verification
Dengzhe Hou, Lingyu Jiang, Fangzhou Lin et al. (+1 more)
Electroencephalography (
EEG) analysis in cognitive studies requires specialized expertise and involves many defensible choices over contrasts, channels, time windows, and statistical tests. LLM
agents can translate varied natural-
language questions into analysis choices, offering a flexible interface for automation. Yet fluent reports alone cannot establish that an
agent selected the requested analysis or evaluated a confirmatory claim independently of adaptive search. We present CogEEGAgent, a cognitive-
EEG analysis
agent grounded in MNE-Python. Its
EEG-specific scientific harness separates semantic from scientific authority. The LLM interprets intent and proposes registered analyses, while deterministic components validate typed contracts, control confirmation access, and authorize evidence-bound release. On a prespecified routing
benchmark, CogEEGAgent maps
language to registered analyses more accurately than a matched deterministic router, while matched preflight makes both systems abstain whenever required. In an externally model-authored, outcome-blind campaign, the complete system releases supported analyses with participant-disjoint confirmation and blocks prespecified capability hazards and lifecycle-reuse requests. Policy stress testing shows that held-out confirmation curbs false positives from uncorrected adaptive search. Together, these studies establish bounded autonomy and an auditable automation framework for cognitive-
EEG workflows. More broadly, they show how scientific
agents can combine flexible
language understanding with fail-closed control over inference and release.
Read the full abstract ▾
All authors Dengzhe Hou, Lingyu Jiang, Fangzhou Lin, Kazunori D Yamada
Systems and Control
Ultra-Low-Energy Open-Circuit Fault Diagnosis for Three-Phase Inverters
Xiaoyi Lei, Fanfu Wu and Yunting Liu
Embedded fault diagnosis in three-phase inverters must satisfy the sub-watt power budget of converter control hardware, but conventional convolutional neural network (CNN)-based methods require dense multiply-accumulate operations and impose substantial inference energy. This work proposes an event-driven neuromorphic framework for energy-efficient open-
circuit (OC) fault diagnosis. A CNN trained on current-vector trajectory
matrices is converted into a spiking neural network (SNN) and evaluated using the NengoLoihi framework with Loihi-based neuromorphic energy estimation. By exploiting the sparse structure of trajectory
matrices, the SNN activates computation only in informative regions instead of processing the full feature map densely. Experiments on a three-phase inverter platform show that the proposed method achieves 11 microjoules per diagnosis, corresponding to a 382 times inference-energy reduction compared with a GPU-based CNN, while maintaining 100% diagnostic accuracy. Robustness is further validated under unbalanced loading, current amplitude step changes, and injected measurement noise.
Read the full abstract ▾
Signal Processing
Joint Access and Fronthaul Resource Allocation for Cell-Free Massive MIMO with Wireless Fronthaul
Ozan Alp Topal, Özlem Tuğfe Demir, Emil Björnson et al. (+1 more)
Wireless fronthaul is a key enabler of flexible and scalable
cell-free massive
MIMO systems, but its limited capacity poses significant challenges for maintaining high and uniform user performance. In this work, we analyze the performance of a
cell-free massive
MIMO network with wireless fronthaul under realistic low physical layer functional splits. We propose a joint access and fronthaul resource allocation
algorithm that maximizes the minimum user equipment (UE) spectral efficiency while satisfying fronthaul load constraints. Our analysis reveals that power allocation over the wireless fronthaul follows a modified water-filling structure, where the water level is jointly determined by the access and fronthaul channel gains. Furthermore, we show that severe fronthaul limitations not only reduce UE rates but also introduce spatial performance disparities depending on the cloud location. Finally, we demonstrate that split option 8 is impractical under wireless fronthaul constraints, underscoring the importance of dynamic fronthaul bit allocation to reduce fronthaul load and enable efficient system operation.
Read the full abstract ▾
All authors Ozan Alp Topal, Özlem Tuğfe Demir, Emil Björnson, Cicek Cavdar
Signal Processing
Conceptual and Signal Processing Principles of Spread-Spectrum Color Doppler Ultrasound
Kian Esmailian and James C. Lacefield
The spread-spectrum color Doppler method provides an alternative to single-plane-wave color Doppler for ultrafast blood-flow imaging with a high Nyquist velocity, high frame rate, and long ensembles. Spread-spectrum images are acquired by transmitting plane waves at a sequence of azimuth angles that is repeated multiple times in a different random order during each repetition. Randomization of the plane-wave angle modulates in-focus and off-focus echoes in distinct manners that enable stationary clutter to be suppressed without compounding by application of a notching comb filter. In this paper, a simplified mathematical model for the slow-time signal acquired with a sequence of plane waves is used to explain the mechanics of the spread-spectrum method, investigate the effect of different discrete-time comb filter realizations on the accuracy of the resulting Doppler frequency estimates, compare the effectiveness of spread-spectrum to single-plane-wave imaging for spatial localization of a Doppler signal source, and characterize the method's sensitivity to background tissue motion. The results establish a foundation to further improve the utility of the spread-spectrum method for practical applications involving non-stationary clutter and physiologically relevant flow patterns.
Read the full abstract ▾
Systems and Control · Cross-listed from cs.RO
Amortising Trajectory Optimisation for Residual MPC via Implicit Contact Differentiation
Daniel Layeghi, Thomas Corbères, Calum Arnott et al. (+4 more)
Differentiable simulation can accelerate contact-rich trajectory optimisation by exposing local sensitivities of task outcomes to controls. Existing approaches either use finite differences, which are expensive and step-size sensitive; differentiate iterative contact solvers by unrolling automatic differentiation (AD), which stores a growing computation trace; or require intricate, solver-specific KKT sensitivity derivations. We introduce an AD-assisted implicit derivative for regularised smooth contacts and apply it to Mujoco MJX, based on the Implicit Function
Theorem (IFT). The method differentiates the stationarity residual at the tolerance-converged solution, avoiding both solver unrolling and hand-assembled KKT systems. IFT keeps compiled temporary
memory nearly constant with solver effort, changing by less than 4$\%$ from one to ten iterations versus 10.6$\times$ growth for unrolled AD. IFT
memory grows slower with active contacts and model dimension, using 20$\times$ less
memory at 256 contacts and 6$\times$ less at 16 contacts and 96 DoF. We further introduce optimiser
distillation for residual MPC, amortising batched full-horizon iLQR into a policy that guides short-horizon residual iLQR. Across Finger, Franka, and Unitree, this raises six-step success by 28-98 percentage points over standard iLQR.
Read the full abstract ▾
All authors Daniel Layeghi, Thomas Corbères, Calum Arnott, Aditya Kamireddypalli, Hashim Al-Obaidi, Steve Tonneau, Michael Mistry
Audio and Speech Processing
Towards Operational Conversational Intelligence: A Speech Intelligence Framework
C. Vishnoi, S. Khurana, A. Timmapur et al. (+2 more)
Body-worn camera (BWC) audio presents unique challenges including high ambient noise, variable recording conditions, and multiple overlapping speakers that make automated transcription and speaker labeling challenging. We propose a dual-path conversational intelligence framework that preprocesses raw BWC audio, separates the processing pipeline into a diarization branch and an ASR branch, and fuses their outputs. The diarization branch uses a denoising front-end (DeepFilterNet), voice activity detection (VAD), and NVIDIA's Multi-Scale Speaker Diarization Decoder (MSDD) with TitaNet
embeddings. The transcription branch uses loudness normalization and WhisperX (Large-v3) with forced
alignment and probability-guided
speech segmentation. Finally, word-level speaker attribution is performed by assigning each recognized word to the speaker segment with the greatest temporal overlap. We evaluate the proposed framework on a curated body-worn camera
dataset constructed from publicly available U.S. and U.K. police body-worn camera recordings. Experimental results demonstrate that task-specific acoustic conditioning and probability-guided
speech segmentation improve speaker diarization, transcription, and word-level speaker attribution under challenging body-worn camera recording conditions. The proposed modular architecture provides an extensible foundation for future speaker-aware conversational intelligence systems.
Read the full abstract ▾
All authors C. Vishnoi, S. Khurana, A. Timmapur, S. Rai, S. Mohanty
Signal Processing
Sensing-Aware Backscatter Communications: A Survey on Envelope Stability, Waveform Design, and Selection Diversity
Rahul Gulia, Priyanka Mann, Ashish Sheikh et al. (+2 more)
The
vision of ubiquitous battery-free connectivity has positioned backscatter communication as a foundational technology for next-generation Internet of Things (IoT) and industrial sensing applications. However, as backscatter systems evolve from simple identification tags toward high-fidelity integrated sensing and communication (
ISAC) devices, fundamental challenges emerge at the intersection of waveform design, hardware nonlinearity, and channel dynamics. This paper presents a comprehensive survey of sensing-aware backscatter communications, with a focus on three tightly coupled dimensions: envelope stability, spatial diversity, and temporal channel robustness. We introduce the concept of the “Illuminator's Dilemma” to describe the inherent conflict between high-peak-to-average power ratio (PAPR) multicarrier waveforms optimized for active communication and the stringent dynamic range requirements of passive tags. The survey provides a unified treatment of multi-objective waveform design strategies, a comprehensive taxonomy of near-far mitigation techniques, and an in-depth analysis of channel state information (CSI) aging in backscatter networks. Key contributions include tag-centric metrics such as the Backscatter Crest Factor (BCF), Envelope Stability Factor (ESF), and Sensing Fidelity Index (SFI), along with a comparative evaluation of PAPR reduction techniques, spatial diversity methods, and non-coherent detection schemes. This survey is intended for researchers, engineers, and graduate students working in wireless communications, IoT, RF sensing, and integrated sensing and communications.
Read the full abstract ▾
All authors Rahul Gulia, Priyanka Mann, Ashish Sheikh, Feyisayo Favour Popoola, Serisha Vadlamudi
Image and Video Processing · Cross-listed from cs.IT
Low-Latency Generative Semantic Communication via Channel-Realization Flow Matching
Fan Gao, Youzheng Wang, Zhijin Qin et al. (+1 more)
Generative semantic communication receivers deliver high perceptual quality but suffer from prohibitive decoding latency. This bottleneck arises because
diffusion receivers rely on
stochastic iterative decoding, while existing flow matching receivers employ independent endpoint coupling that ignores the physical source–channel link, yielding unnecessarily long and curved sampling trajectories. In this paper, we reformulate receiver-side recovery as a realization-coupled bridge flow matching problem under explicit bandwidth and power constraints. Specifically, we propose Realization-Coupled Bridge Flow Matching (RC-BFM), where the decoder initializes from a channel-induced semantic state rather than isotropic noise. Crucially, training pairs are linked via a realization-coupled entropic optimal transport (RC-OT) plan that preserves the physical channel realization of each transmission while maintaining robustness to
stochastic fading. Furthermore, we identify independent coupling as the fundamental source of a conditional train–test distribution shift in conditional flow matching-based receivers, and derive an end-to-end distortion bound whose discretization error decays as \(O(K^{-2})\). Experiments on CIFAR-10 and FFHQ-64\(\times\)64 over AWGN and Rayleigh fading channels demonstrate that RC-BFM achieves a superior fidelity–perception trade-off, reducing decoding latency by over 10\(\times\) compared to
diffusion-based receivers.
Read the full abstract ▾
All authors Fan Gao, Youzheng Wang, Zhijin Qin, Feifei Gao
Systems and Control · Cross-listed from cs.IT
Self-Supervised Pretraining for Transmission-Relevant MIMO CSI Representation
Sehyun Ryu, Yumin Kim, Minjae Lee et al. (+2 more)
Self-supervised representation learning from unlabeled channel state information (CSI) can reduce labeling and adaptation overhead in learning-based multiple-input multiple-output (
MIMO) systems. Existing CSI pretraining methods typically use reconstruction objectives or contrastive pairs from generic augmentations, which do not explicitly reflect transmission-relevant channel similarity. This paper proposes beam-response contrastive learning (BRCL), a self-supervised CSI pretraining framework based on the transmit-side Gram
matrix. For a channel
matrix $\mathbf{H}$, $\mathbf{R}=\mathbf{H}^{\mathrm{H}}\mathbf{H}$ determines the received power of any unit-norm transmit beam $\mathbf{w}$ as $|\mathbf{H}\mathbf{w}|_2^2=\mathbf{w}^{\mathrm{H}}\mathbf{R}\mathbf{w}$. BRCL maps each CSI sample to a beam-response profile and uses the induced soft similarity as a label-free relational target for contrastive pretraining. Combined with reconstruction learning, BRCL enforces both sample-level CSI recovery and beam-response-level consistency, yielding transferable CSI representations without task-specific labels or manual positive pairs. Experiments on diverse
MIMO channel
datasets show that BRCL improves label efficiency and outperforms autoencoder- and channel-charting-based pretraining across beam selection, user selection, and future beam selection tasks.
Read the full abstract ▾
All authors Sehyun Ryu, Yumin Kim, Minjae Lee, Hyun Jong Yang, John M. Cioffi
Signal Processing · Cross-listed from quant-ph
Quantum-Limited Symbol-Blind Channel Estimation for Coherent State Discrimination
Hongxu Chen, Renzhi Yuan, Haifeng Yao et al. (+1 more)
Residual dispersion breaks temporal-mode matching in
photon-starved coherent links. For equiprobable $M$-ary PSK coherent states in a known spectral mode, with unknown symbols and carrier phase, we establish the
quantum limit for blind joint estimation of group delay and second-order dispersion: after eliminating the common phase, it is $4N_s\mathbf{C}$, set by the covariance of the centered generators alone. A multi-output
quantum pulse gate with
photon-number-resolving detection locally attains it and supports reception below the standard
quantum limit under turbulent fading.
Read the full abstract ▾
All authors Hongxu Chen, Renzhi Yuan, Haifeng Yao, Mugen Peng
Systems and Control
Current References for Minimum Copper Loss and Torque Ripple in the Full Torque-Speed Range for Symmetrical Six-Phase PMSMs With Nonsinusoidal Back-EMF Under an Open-Phase Fault
Alejandro G. Yepes, Wessam E. Abdel-Azim, Oscar López et al. (+3 more)
Current-reference generation based on lookup tables (LUTs) is here proposed for
star-connected symmetrical six-phase nonsalient PMSMs with nonsinusoidal back-EMF under an open-phase fault.
Fourier coefficients for all healthy phases are computed offline, enabling unbalanced nonsinusoidal currents. A lexicographic optimization minimizes torque ripple and then copper loss, subject to torque, zero-current-sum, peak-current, peak-voltage, and torque-ripple constraints. Cogging torque can be included for compensation. Simulations show a higher feasible speed limit: about 30% at low torque and 23% along an example load curve. Finite-element analysis shows a 77% reduction in peak-to-peak torque ripple with cogging-torque compensation.
Read the full abstract ▾
All authors Alejandro G. Yepes, Wessam E. Abdel-Azim, Oscar López, Petros Karamanakos, Ayman S. Abdel-Khalik, Jesús Doval-Gandoy
Systems and Control · Cross-listed from cs.RO
Development of a Handheld Actuation Mechanism for a Tendon-driven Robotically Steered Guidewire
Saima Chavenet, Timothy A. Brumfiel, Revanth Konda et al. (+1 more)
An endovascular intervention begins with a skilled clinician manually navigating a long, slender wire, called a guidewire, to the target location within the vasculature. Due to factors, such as vessel tortuosity and lack of steerability at the guidewire tip, manual
navigation of a guidewire could be challenging, potentially resulting in vessel damage, perforation, dissection, and occlusion, as well as postsurgical complications, such as thrombosis. This work details the development of a handheld actuation mechanism for tendon-driven robotic guidewires by utilizing the aforementioned spooling mechanism. Incorporating the compact spooling mechanism into a handheld device enables clinicians to leverage both motorized guidewire steering and manual rotation and translation, if desired. The proposed device is designed such that, while holding the device, the clinician can execute feeding, rotation, and bending motions for up to 1.5 m of the robotically steerable guidewire using a joystick and momentary switch on the handle. Furthermore, the proposed device is demonstrated through
navigation in an anatomically accurate phantom aorta model.
Read the full abstract ▾
All authors Saima Chavenet, Timothy A. Brumfiel, Revanth Konda, Jaydev P. Desai
Systems and Control · Cross-listed from cs.AI
Task-Conditional Faithfulness Auditing of Multimodal LLMs for Grid Diagnosis
Tianqiao Zhao, Meng Yue and Jianhui Wang
Multimodal large
language models (LLMs) can combine
topology, measurements, and incident text for grid diagnosis, yet answer accuracy does not establish that task-appropriate evidence was used. This letter proposes a general framework in order to conduct task-conditional faithfulness audit. It compares self-reported reliance, intervention-derived behavioral reliance, and preregistered engineering importance. The framework first registers task-specific evidence requirements and compares them with self-reported reliance and behavioral changes under controlled modality ablations. To resolve detected discrepancies, we design an evidence-gated correction and re-audit mechanism that regenerates failed responses under evidence constraints and independently re-ablates them to verify improved grounding without performance loss. Case studies evaluate three differently scaled LLMs on IEEE 39- and 118-bus scenarios. These results validate the framework ability to detect, diagnose, and correct task-conditional faithfulness failures.
Read the full abstract ▾
Image and Video Processing
Efficient Ultrasound Image Segmentation with Token-Conditioned Neural Cellular Automata
Fangyijie Wang, Tanya Akumu, Zi Ye et al. (+3 more)
Point-of-Care Ultrasound (POCUS) plays an important role in bedside diagnosis and clinical decision-making, particularly in resource-constrained settings. Recent deep learning methods have substantially improved ultrasound image
segmentation, enabling accurate diagnosis and biometric estimation. However, their computational cost limits deployment on portable and low-resource devices. To address this challenge, we propose LiteAdaNCA-Net (LANCANet), a lightweight ultrasound
segmentation framework that incorporates
token-conditioned Neural Cellular Automata (NCA) adapters for iterative feature refinement. Specifically, structure-aware
tokens guide local NCA refinement via
Token FiLM, enabling boundary-aware feature refinement with minimal computational cost. We evaluate LANCANet on HC18, CCA, and PSFHS, and assess robustness on two independent African fetal head
datasets collected from multiple clinical centers. Experimental results demonstrate that LANCANet achieves competitive or superior performance to recent lightweight CNN- and
transformer-based methods. On HC18 and CCA, LANCANet achieves the highest Dice Similarity Coefficient (DSC) of 96.62% and 92.86%, respectively. On PSFHS, it achieves the best performance on the challenging pubic symphysis structure while maintaining competitive fetal head
segmentation accuracy. Furthermore, despite being trained from scratch, LANCANet maintains competitive performance on the external KEN-FH and AFR-FH
datasets under substantial domain shifts. These results show that
token-conditioned NCA refinement improves
segmentation accuracy and boundary localization while maintaining computational efficiency for resource-constrained clinical deployment. Our code is on \href{https://anonymous.4open.science/r/LANCAN-21A0/README.md}{GitHub}.
Read the full abstract ▾
All authors Fangyijie Wang, Tanya Akumu, Zi Ye, Guénolé Silvestre, Karim Lekadir, Kathleen M. Curran
Image and Video Processing
Rapid quantitative chemical composition mapping using model-based MRI reconstruction with field inhomogeneity correction
Artyom Tsanda, Stefan Benders, Muhammad Adrian et al. (+2 more)
Magnetic resonance spectroscopic imaging methods are particularly attractive for chemical engineering applications, including the monitoring of chemical reactions, where a rapid assessment of spatial variations in chemical composition is required. Conventional approaches, such as chemical shift imaging, introduce an additional spectral-encoding dimension, which substantially increases acquisition time. Consequently, fast spatially resolved spectroscopy remains an active research topic. This work uses a model-based reconstruction framework that embeds a priori spectral knowledge of the involved chemical components into the forward model to accelerate composition mapping. It allows for the reconstruction of molar ratio maps for individual chemical components without acquiring high-resolution spectra. Extending from previous studies, the proposed model accounts for inhomogeneities of the main field, which become more pronounced in systems with larger bores relevant for process engineering. Phantom experiments employing a 2D multi-gradient echo sequence demonstrate the ability to determine molar ratios for chemical components with single peaks as well as multiple peaks in their spectra. The bias and precision of the method remain around 0.01 mol/mol and 0.09 mol/mol, respectively, for a 20 s scan, indicating suitability for dynamic processes. Finally, acquisition time can be reduced further by applying sparse k-space sampling, potentially shortening the scan to 5 s with only minor degradation in quantitative performance.
Read the full abstract ▾
All authors Artyom Tsanda, Stefan Benders, Muhammad Adrian, Alexander Penn, Tobias Knopp
Audio and Speech Processing · Cross-listed from cs.HC
Let Me Look at You: Advanced Facial Expression Modeling for Conversational Speech Synthesis
Yifan Hu, Shuwei He, Rui Liu et al. (+1 more)
Conversational
Speech Synthesis is a fundamental component of human-computer interaction, aiming to generate contextually appropriate, expressive, and empathetic
speech. However, facial expressions encode subtle and rich affective cues that are crucial for empathetic
speech interaction, whereas existing approaches often overlook this important modality. In addition, the lack of large-scale natural conversational
datasets with both
speech and visual modalities also limits the development of visual affect understanding in conversational settings.To address these limitations, we propose FacialTalker, a facial-expression-aware CSS framework built upon a large
language model backbone. To efficiently encode facial expressions, we propose AUTokenizer, a single-codebook visual tokenizer that discretizes each frame-level facial expression into a compact
token, trained with supervision from combinations of facial Action Units. We further introduce a dual direct preference optimization (DualDPO) strategy, which extends the DPO by jointly imposing preference constraints on both visual and
speech token sequences, to enhance the model's understanding of facial expressions and
speech semantics in
multimodal conversational contexts. Moreover, we construct VSDD-1K, a large-scale
multimodal dialogue
dataset collected through a fully automated pipeline from real-world Internet conversations, comprising over 1,033 hours of synchronized speaker
videos and
speech, with more than 85% of frames containing valid faces. Extensive objective and subjective experiments demonstrate that FacialTalker consistently outperforms strong baselines in facial-expression perception and
speech synthesis quality, generating
speech that is more natural, expressive, and better aligned with the conversational context. The results also validate the effectiveness of our training strategy and
dataset construction pipeline.
Read the full abstract ▾
All authors Yifan Hu, Shuwei He, Rui Liu, Haizhou Li
Systems and Control
Does Central Planning Fail Locally? Evaluating EV Charging Flexibility Optimization Across Grid Levels
Ambra Van Liedekerke, Lorenzo Zapparoli, María Parajeles Herrera et al. (+3 more)
Electric vehicles (EVs) are a key enabler of global decarbonization, and their charging flexibility is crucial for integrating high shares of renewable energy into future power systems. Research on EV charging flexibility in future power systems has largely focused on either the transmission or the distribution level, with limited integration between the two. However, these two scales cannot be separated: EV charging optimized at the national transmission level for system-wide planning is ultimately activated in local distribution grids, where it may cause voltage and loading issues. This study bridges that gap by evaluating EV charging flexibility across both grid levels, using two different control approaches. First, a top-down approach disaggregates the optimal charging demand obtained from a centralized planner at the transmission grid down to the distribution grid, explicitly accounting for grid constraints and individual driving requirements. Second, a bottom-up approach optimizes EV charging locally under distribution-grid constraints using electricity price signals from the centralized planning model. Controlled charging from the top-down and bottom-up approaches is then compared with uncontrolled charging. Results show that the top-down approach produces EV charging schedules that are compatible with individual driving behavior and reduce grid violations compared to uncontrolled charging. The bottom-up approach identifies an optimal profile similar to the centrally optimized profile, showing that electricity price signals are a suitable coordination mechanism. These findings suggest that centrally optimized EV charging remains effective when implemented at the distribution level, in systems with high DER penetration.
Read the full abstract ▾
All authors Ambra Van Liedekerke, Lorenzo Zapparoli, María Parajeles Herrera, Blazhe Gjorgiev, Gabriela Hug, Giovanni Sansavini
Systems and Control
Optimization of Current Lookup Tables for Minimum Stator Copper Loss and Torque Ripple in the Full Torque-Speed Range for a Six-Phase PMSM With Nonsinusoidal Back-EMF
Alejandro G. Yepes, Shirin Rahmanpour, Oscar López et al. (+4 more)
A recently proposed method was able to generate current references with minimum stator copper loss (SCL) and torque ripple over the full torque-speed range of multiphase nonsalient permanent-magnet synchronous machines (PMSMs) with nonsinusoidal back-electromotive force (back-EMF). However, it relied on large lookup tables (LUTs) generated offline, requiring several hours for generation and large
memory for storage. Since on-chip
memory in industrial digital signal processors (DSPs) is limited, this hinders practical implementation. This paper addresses this problem by analyzing how the LUTs of that method can be simplified and optimized, focusing on the example of a symmetrical six-phase PMSM drive. Breakpoint selection (BS), LUT construction (LC), and interpolation approaches are evaluated to optimize the tradeoff between LUT simplification and performance. It is shown that the required
memory and LUT generation time can be greatly reduced while keeping nearly the same feasible torque-speed area and performance. Simulation results confirm negligible performance degradation.
Read the full abstract ▾
All authors Alejandro G. Yepes, Shirin Rahmanpour, Oscar López, Wessam E. Abdel-Azim, Petros Karamanakos, Ayman S. Abdel-Khalik, Jesús Doval-Gandoy
Systems and Control
Can sunlight replace LEDs in vertical farms? A critical assessment of optical-fiber daylighting strategies
Francesco Ceccanti, Aldo Bischi, Marco Antonelli et al. (+1 more)
Vertical farming is a promising approach for increasing food production in controlled environments while reducing land use, water consumption, and dependence on external climatic conditions. However, its large electricity demand, mainly driven by artificial lighting, still limits its energetic and economic sustainability. This study evaluates the techno-economic potential of an optical-fiber daylighting system designed to transport concentrated sunlight directly to the crop growing zones of a vertical farm. A validated transient vertical farm model was coupled with Tonatiuh ray-tracing simulations to assess a single-axis
solar-tracked Fresnel concentration system. Five optical fiber typologies and three daylighting strategies were compared in terms of crop production, specific electrical energy consumption, and light cost, and benchmarked against light-pipe and rooftop photovoltaic solutions. The optical-fiber system achieved an annual useful-PAR delivery efficiency of 19-27%, depending on fiber typology. By distributing daylight across all rack shelves and limiting heat gains through a UV-IR filter, it reduced electricity consumption by 35-89%, compared with 12% for the light-pipe configuration. The daylight-only strategy achieved the lowest SEEC, equal to 1.81 kWh kg^-1, but caused a 52% crop productivity reduction. Hybrid LED-optical-fiber strategies avoided this penalty and achieved electricity savings of about 50%, with a minimum SEEC of 3.16 kWh kg^-1. However, current optical fiber costs make the system economically unattractive, requiring an estimated capital cost reduction of about 90% to become competitive. Rooftop photovoltaics covered 22% of annual electricity demand and achieved an 11-12-year payback, making PV the most economically viable rooftop
solar option under current
market conditions.
Read the full abstract ▾
All authors Francesco Ceccanti, Aldo Bischi, Marco Antonelli, Andrea Baccioli