Papers

Accurate vehicle localization from monocular roadside surveillance cameras is important for intelligent transportation systems, traffic monitoring, and traffic conflict analysis. Standard approaches often estimate vehicle position from the center of the detector bounding box, which can produce large errors due to perspective distortion and parallax, especially for elevated cameras and large vehicles. This paper proposes a two-stage geometry-aware localization pipeline that estimates the projection of the vehicle footprint onto the road plane. First, vehicles are detected using a YOLO26-based detector. Second, a dedicated ResNet34 regression network predicts four corner points corresponding to the projected vehicle base. The final position is computed as the geometric center of the predicted quadrilateral. The method was trained on synthetic data generated in CARLA and fine-tuned on real-world roadside imagery from DAIR-V2X. Experiments on synthetic and real data showed clear improvements over naive bounding-box-center localization. On DAIR-V2X, the mean image-space localization error decreased from 31.77 px to 15.30 px, a 51.8% improvement, while the median error decreased to 4.29 px. Median ground-plane error for medium-range vehicles decreased from 5.52 m to 0.90 m, and for far-range vehicles from 8.67 m to 1.84 m. The results also show that contextual information surrounding the detector bounding box is important for geometric localization. The largest gains were observed for distant vehicles and geometrically challenging cases affected by strong perspective distortion and parallax.

Applying foundation models to the radio frequency (RF) domain presents unique challenges due to the intrinsic physical complexity of raw I/Q signals and the extreme heterogeneity of spectral data. In this paper, we present Radio-FM, a scalable family of foundation models designed for universal radio signal representation learning. Unlike standard architectures, Radio-FM employs dual-channel processing specifically optimized for I/Q independence while capturing cross-channel interactions through a lightweight attention mechanism. To scale pretraining across heterogeneous multi-source corpora with highly variable sequence lengths, we propose a token-budgeted dynamic batching strategy coupled with channel-independent masked reconstruction. We pretrain Radio-FM on a diverse collection of 15 datasets spanning modulation, radar, and communication domains, and rigorously evaluate it on 15 downstream benchmarks. Experimental results show Radio-FM achieves state-of-the-art performance on 13 of 15 benchmarks, consistently improving across modulation, radar, emitter identification, wireless technology recognition, and wireless interference identification. Notably, it exhibits superior few-shot transferability, significantly outperforming existing baselines in data-scarce regimes, validating its potential as a general-purpose backbone for radio signal understanding.

Neural Audio Codecs are widely adopted in speech generation and editing. However, existing neural audio codecs are not idempotent: across the paper's twelve baseline systems, every configuration tested rewrites, on average, at least 15% of its tokens in a single decode-re-encode pass. This poses a problem for utilizing Neural Audio Codecs as token interfaces in pipelines where re-encoding decoded outputs can occur. We present LILAC, a fully convolutional 24 kHz speech codec at 9.375 Hz and 0.75 kbit/s that is codec idempotent by construction; re-encoding the decoded audio of any valid token stream returns the identical stream. LILAC achieves idempotency while maintaining competitive quality, reaching UTMOS 4.14 and 4.24 on LibriSpeech and LibriTTS-R test sets, comparable to SOTA sub-1 kbit/s Neural Audio Codecs.

Dual-polarized (DP) microstrip antennas are utilized in wireless systems for efficient data transmission. However, their bandwidth is typically very limited. In this contribution, we propose to design DP microstrip antennas with enhanced bandwidth using a density-based topology optimization approach. We formulate an optimization problem that simultaneously accounts for feeding port matching, the ports' isolation, and far-field dual-polarized performance. To enhance the bandwidth, we employ an FR4 stack-up, in which the copper on two layers is optimized simultaneously. We present two antenna designs operating around 5.7 GHz, which show a compromise in performance between a high isolation (more than 40 dB) and enhanced impedance bandwidth (around 10%). The optimized designs are experimentally validated, showing an excellent agreement between the simulated and measured performance.

Respiration provides a continuously available window into physiological state and behavior. However, monitoring it outside controlled settings remains challenging because a wearable system must capture small body deformations while remaining comfortable, low power, and robust to changes in posture and motion. We present a compact non-invasive respiratory sensing system based on a force-sensitive resistor (FSR) embedded in an abdominal belt and integrated with a custom Bluetooth Low Energy acquisition board. The system combines a simple piezoresistive readout with a mechanical holder designed to transfer abdominal expansion to the sensor without analog amplification. We evaluate the complete sensing pipeline across multiple breathing patterns and body positions. In stationary settings, the recorded signals exhibit consistent amplitude changes and recurring peak-to-peak timing across breathing maneuvers; under light movement, these variations remain visible despite motion-induced baseline shifts. We further design a five-phase stress-induction protocol and collect respiratory recordings from 12 participants. Using interpretable time-domain features and standard classifiers, we examine whether the acquired signals distinguish relaxation from stress-induction phases. In this preliminary experiment, the best-performing model achieves 88.0% test accuracy, indicating that the extracted respiratory features distinguish stress-induced phases from relaxation phases in this dataset. Overall, our results show that the proposed platform enables real-time respiratory monitoring across diverse daily-life scenarios and captures respiratory changes that distinguish stress-induction from relaxation phases, supporting its potential for affective-computing applications.

Nonvisual classification of ground condition based on a multimodal sensing approach was investigated for an amoeba-inspired autonomous walking robot. To classify ground condition without image sensing and processing, we implemented artificial proprioception by integrating a three-axis accelerometer, eight foot pressure sensors, and reservoir computing (RC). Even when large fluctuations in the sensor outputs are caused by dynamic motions of a four-legged robot in walking, our system can classify the ground condition, flat or rough, with high accuracy. We demonstrate on-site switching of walking gait depending on ground condition in the robot. We also discuss the contribution of each sensor to ground condition classification.

We present a case study in application-driven specialization of a five-stage soft processor, evaluated on the inner control loop of a reaction-wheel self-balancing bicycle. Starting from a custom 32-bit RISC core in the MIPS tradition, we specialize the design in two ways. First, two frequently accessed peripheral inputs are mapped directly into architectural register state, written every cycle by hardware and owned exclusively through the register file's write-port structure rather than by arbitration. Second, four periodic PWM channels are offloaded to hardware and driven continuously from four exported registers, removing periodic actuation from software entirely. Because peripheral values are addressable as ordinary register operands, all ten sensor reads in the control loop cost no dedicated instruction and no dedicated cycle, folding into arithmetic that executes anyway; the memory-mapped equivalent requires an explicit load per snapshot and costs five extra instructions and cycles. The actuation path likewise removes waveform maintenance from software entirely. We report two configurations, because the extensions and the single-cycle array multiplier they were deployed alongside are not present together in a single archived build: an archived configuration, whose worst-case loop is 91 cycles, and the integrated configuration matching the deployed system, at 43 cycles. Against a 20 ms actuation frame these are margins of roughly 7,300x and 15,000x. The deadline is met by so wide a margin in either case that the specialization was not necessary for real-time compliance; its value lies in instruction count and software simplicity, not in determinism, which an on-chip single-cycle I/O region already provides. The zero-instruction sensor read is independent of that choice: the multiplier cannot affect whether a peripheral read needs an instruction of its own.

Real-time extended reality (XR) services impose stringent throughput, latency, and reliability requirements. An XR media unit is commonly segmented into a protocol data unit (PDU) set that is useful only when all required PDUs are delivered within its delay budget. Existing PDU set-aware medium access control (MAC) schedulers generally prioritize transmissions according to urgency or progress, without explicitly accounting for whether a PDU set can still be completed on time. In this paper, we propose a Lyapunov-based completion-aware MAC scheduler that jointly captures the set-level completion dependency, deadline urgency, and completion feasibility of each head-of-line PDU set. Long-term quality-of-service (QoS) requirements are represented through virtual debt queues, and resources are allocated according to their marginal contribution to the timely-completion probability of each head-of-line PDU set. These components are integrated into a drift-plus-penalty scheduling metric that balances long-term reliability requirements against current completion opportunities. System-level simulations demonstrate that the proposed scheduler improves PDU set delivery reliability and lower-tail performance under resource contention, while providing robust performance across heterogeneous traffic demands and network conditions.

Quantum state tomography (QST) has attracted considerable attention due to its fundamental role in quantum information processing. In this paper, we develop a unified theoretical framework for analyzing the sample complexity of structured QST under noisy observations arising from state preparation noise, measurement noise, and finite-shot statistical noise. The proposed framework applies to a broad family of structured quantum-state classes, including general mixed states, sparse states, low-rank states, matrix product states (MPSs), matrix product operators (MPOs), projected entangled-pair states (PEPSs), and projected entangled-pair operators (PEPOs), while further introducing physically consistent structured models—including low-rank and sparse states, low-rank MPOs (LR-MPOs), and low-rank PEPOs (LR-PEPOs)—that simultaneously exploit low-dimensional structures and preserve the physical constraint. Within this framework, we derive unified non-asymptotic sample complexity guarantees for two constrained least-squares estimators under noisy observations: a noise-aware estimator that incorporates the calibrated noise model and a noise-unaware estimator based on the ideal Born measurement model. For the noise-aware estimator, we derive unified trace-norm recovery guarantees that explicitly characterize the dependence of the sample complexity on three fundamental quantities: the complexity of the underlying structured state class, the complexity of the measurement ensemble, and the state preparation and measurement noise levels. For the noise-unaware estimator, we establish a unified non-asymptotic recovery guarantee consisting of a statistical error term and an additional deterministic bias term arising from the mismatch between the assumed reconstruction model and the noisy observation process.

Accurate channel estimation is essential for coherent transmission in ultra-massive multiple-input multiple-output (UM-MIMO) systems, where near-field propagation and high-dimensional spatial channels impose substantial signal processing challenges. Movable antenna architectures increase the estimation complexity further due to geometry-dependent channel variations. Existing approaches struggle to balance accuracy and complexity, motivating the use of environment-dependent propagation structures for efficient estimation. To this end, this paper proposes a unified channel map-based channel estimation framework for UM-MIMO systems, which integrates movable planar array reconfiguration and near-field spherical-wave modeling to support geometry-aware line-of-sight (LoS) estimation and efficient non-LoS (NLoS) recovery. A channel map-based LoS estimator is proposed combining coarse user position information with a Fisher information-guided antenna placement strategy. Two efficient NLoS estimation methods are also presented, including a sketch-based reduced-subspace estimator for low-complexity processing and a channel map-based estimator that leverages scatterer location information for near-optimal performance. The framework further incorporates visibility-region modeling and a structural similarity-based pilot assignment strategy for multi-user scenarios. Simulation results show that the proposed channel map-based framework improves estimation accuracy, reduces computational overhead, and enhances scalability compared with state-of-the-art benchmarks without channel maps.

Electric guitar tone transfer (EGTT) and tone removal (EGTR) are two fundamental tasks in guitar tone modeling: EGTT replaces a recording's tone with that of a reference, while EGTR recovers the dry direct-input (DI) signal from a wet, processed recording. Despite their highly related nature, prior work has addressed them independently, and both works have yet to achieve satisfactory results. In this paper, we propose EG-VAE, a unified framework that jointly models EGTT and EGTR by disentangling frame-level content and global tone representations from wet recordings with a variational autoencoder. EGTT is achieved by recombining a source's content with a reference's tone, while EGTR is attained by a novel tone masking objective that enforces content-tone disentanglement during training and realizes the removal procedure at inference. To improve transfer to tones unseen in training, a second training stage shapes a smooth tone space through variational sampling and audio-effects augmentation. Experimental results from both objective and subjective evaluations demonstrate that EG-VAE outperforms task-specific baselines on transfer and removal. Demos are available at https://guitar-tone-demo.vercel.app/.

Speech deepfake detection (SDD) systems achieve strong performance on conventional benchmarks; however, existing datasets provide limited coverage of emotionally expressive and recent large audio-language model (LALM)-based attacks. Existing emotional spoofing datasets are also limited in scale and attack diversity, typically covering only voice conversion (VC) or text-to-speech (TTS) attacks. We introduce AffectDF, the most comprehensive benchmark for emotionally expressive speech deepfakes, spanning TTS, VC, emotional VC, and LALM-based spoofing attacks across both acted and spontaneous emotional speech. AffectDF contains approximately 260 hours of speech generated using 21 spoofing attacks across five emotional states. We benchmark state-of-the-art SDD systems under conventional and emotional spoofing conditions, including LALM-based detectors evaluated with both inference-only prompting and supervised fine-tuning. Our experiments reveal severe robustness degradation when models trained on conventional benchmarks are evaluated on AffectDF, with several systems approaching near-random performance. Surprisingly, even large-scale emotional training does not consistently improve cross-domain robustness, indicating that current SDD systems fail to learn generalized spoof representations under emotional and prosodic variability. Robustness further varies substantially across emotional states, attack families, and acted vs spontaneous emotional speech conditions. These findings expose fundamental limitations of current SDD systems and establish AffectDF as a benchmark for developing more robust spoof detection models.

Automatic music mixing, the task of automatically combining individual audio tracks into a cohesive mixture, is typically addressed by parallelized architectures that process all input tracks in a single pass. In this work, inspired by how human mix engineers process stems one at a time, we propose a paradigm shift and ask whether automatic music mixing can be reformulated as a sequential stem blending task, where each stem is blended into a growing submix. Specifically, we train a latent flow matching model conditioned on the submix context, enabling sequential processing of an arbitrary number of input tracks. To train the model, we introduce a degradation-based data synthesis strategy that simulates realistic stem blending scenarios from existing multitrack and source separation datasets. Experimental results on both stem blending and automatic music mixing benchmarks demonstrate the effectiveness of the proposed approach. We provide audio examples on the accompanying demo page\footnote{https://sequential-mixing-demo.vercel.app/}.

This paper presents a discrete-time passivity-based analysis of first-order momentum-based methods for a class of functions whose gradient has lower and upper sector bounds of $0$ and $L$, respectively. Through a loop transformation, it is shown that momentum-based methods can be represented as a passive controller in negative feedback with an output strictly passive (OSP) system. The weak passivity theorem is then used to derive explicit hyperparameter conditions under which the shifted gradient asymptotically vanishes. Under an additional assumption that requires the existence of a unique stationary point and excludes arbitrarily small gradients far from that point, convergence of the iterates to the global minimizer is established.

Electroencephalography (EEG) is a low-cost and non-invasive signal source for dementia screening, yet existing EEG-based studies remain difficult to compare because preprocessing, EEG segmentation, feature design, classifier choice, and validation protocols vary across studies and are often evaluated in isolation. This variability limits the derivation of robust pipeline recommendations. This paper presents a leakage-safe empirical benchmark for resting-state EEG-based dementia classification and uses it to identify a practical best-practice pipeline. Using the public OpenNeuro ds004504 dataset, the benchmark evaluates artifact correction, fixed-length EEG segmentation, training-only augmentation, multi-domain feature extraction, fold-internal feature selection, classical machine-learning (ML) classifiers, subject-level aggregation, and interpretation under leave-one-subject-out (LOSO) validation. The best-performing pipeline in this benchmark combines Artifact Subspace Reconstruction (ASR) followed by Independent Component Analysis (ICA), 10 s EEG epochs with training-only amplitude scaling and Gaussian-noise augmentation, spectral, complexity, and pairwise connectivity features, mutual-information top-100 selection, linear support vector machine classification, and mean-probability aggregation. It achieves 87.69 percent accuracy, 88.89 percent F1-score, and 91.20 percent AUC on Alzheimer disease (AD) versus cognitively normal controls (CN) classification while retaining 100 features from 1596 raw descriptors. The results support compact multi-domain EEG features as an interpretable and leakage-safe baseline for subject-level dementia classification.

Full-duplex voice assistants must continuously listen while speaking, handling user interruptions under low-latency and resource-constrained streaming conditions. Existing end-to-end full-duplex models can compromise reasoning-related capabilities after speech-domain adaptation, whereas cascaded pipelines introduce extra inference overhead and handcrafted control logic. We propose TurnFSM, an LLM-based state prediction framework that internalizes turn control as explicit finite-state transitions, unifying streaming semantic VAD and utterance-level rejection. TurnFSM decomposes submission and rejection into a serial decision process, reducing multi-task interference while maintaining performance comparable to single-task models. We further introduce a first-order state transition mechanism that enforces the dependency on only the previous state during training, enabling compact inference with the standard causal mask and original LLM positional encoding while avoiding historical state-token accumulation and unnecessary step-by-step state generation. Experimental results show that TurnFSM consistently outperforms the binary-head baseline and remains competitive with task-specific models.

Passive Human Sensing (PHS) is an approach to collecting data on human presence, motion or activities that does not require the sensed human to carry devices or participate actively in the sensing process. In the literature, PHS is generally performed by exploiting the Channel State Information variations of dedicated WiFi, affected by human bodies obstructing the WiFi signal propagation path. However, the adoption of WiFi for PHS has some drawbacks, related to power consumption, large-scale deployment costs and interference with other networks in nearby areas. Bluetooth technology and, in particular, its low-energy version Bluetooth Low Energy (BLE), represents a valid candidate solution to the drawbacks of WiFi, thanks to its Adaptive Frequency Hopping (AFH) mechanism. This work proposes the application of a Deep Convolutional Neural Network (DNN) to improve the analysis and classification of the BLE signal deformations for PHS using commercial standard BLE devices. The proposed approach was applied to reliably detect the presence of human occupants in a large and articulated room with only a few transmitters and receivers and in conditions where the occupants do not directly occlude the Line of Sight between transmitters and receivers. This paper shows that the proposed approach significantly outperforms the most accurate technique found in the literature when applied to the same experimental data.

In this paper, we address the certification of datadriven feedforward control for periodic tracking of unknown nonlinear systems under partial state measurements. To this end, we adopt an invertible neural network (INN) as a surrogate for the unknown system. This choice allows us to bypass solving a nonconvex inversion problem, eliminating the associated inversion errors and reducing tracking error certification to a surrogate modeling problem. We then apply conformal prediction to provide finite-sample probabilistic guarantees on the surrogate modeling error which, through the derived tracking error bound, yield marginal certificates on feedforward tracking error. Finally, we demonstrate the approach on a DC-motor-driven mechanical load with nonlinear friction.

Understanding which surfaces in a scene are physically accessible to a given tool is fundamental for robotic interaction, yet 3D perception systems typically stop at geometric reconstruction or visibility estimation. Existing geometric accessibility methods require complete, noise-free meshes and fixed kinematic bases, assumptions that fail for mobile platforms mapping incrementally from live data; visibility estimation cannot account for tool geometry or approach-corridor clearance. We propose the Accessibility Field: a per-point labelling of surface accessibility for a given tool, produced in real time from streaming sparse LiDAR and updated at sensor rate as the platform moves. Running entirely on GPU, our method evaluates each surface point against precomputed geometry kernels representing the tool at a set of rotated approach orientations, checking tool collisions and approach-corridor clearance. A scan-centric Truncated Signed Distance Field integration scheme underpins our system, updating only voxels near each observed return rather than projecting every frustum voxel each frame – critical for nonrepetitive sensors like the Livox Mid-360, where some bins contain no returns. Our system is tool-agnostic, needs no prior scene model, and runs on workstation and Jetson Orin edge hardware. We evaluate quantitatively on synthetic objects and mature-scale Pinus radiata models, showing visibility alone is insufficient as an accessibility proxy: our method achieves F1=90.8 vs. 69.8 for a Hidden Point Removal baseline on mixed-accessibility geometry, and correctly identifies 56.8% of pine branch surfaces as inaccessible despite being visible from the sensor. To our knowledge, this is the first method to estimate per-point surface accessibility in real time from streaming sparse LiDAR without a prior scene model or fixed base frame – a capability visibility estimation cannot provide.

Prenatal ultrasound imaging is key for assessing fetal health, but AI progress is limited by scarce, privacy-restricted, and hard-to-annotate datasets. We propose a high-resolution fetal ultrasound synthesis framework based on the EDM2 diffusion architecture, trained on multiple public datasets to generate 512x512 images across six anatomical classes. Our method achieved improved image quality with lower FID scores and enhanced downstream fetal plane classification, reaching 93.36% ensemble accuracy after fine-tuning, surpassing real-data-only training. Clinical evaluation by an experienced fetal ultrasound specialist (10+ years) on 100 images yielded a mean realism score of 2.67/5, with real images rated higher than synthetic. Artefacts included smoothing, speckle irregularities, and anatomical inconsistencies. Code, data, models and other resources to reproduce this work are available at https://github.com/xfetus/fetal-ultrasound-edm2.

Automatic music mixing aims to combine multitrack recordings into a balanced and coherent musical piece. Because the content of different songs and the subjective preferences of mixing engineers jointly shape the final outcome, a practical system should deliver well-balanced mixes while allowing for controllable stylistic variation. However, most existing methods treat automatic mixing and mixing style control as separate tasks, making it difficult for a single system to produce high-quality mixes while remaining editable and style-aware. To address this limitation, this paper presents Diff2Mix, a generative automatic mixing system based on diffusion models and a differentiable mixing console. This system offers two levels of optional user control: a reference audio enables overall production style control, and the differentiable mixing console provides explicit audio effects parameters for interpretability and fine-grained optimization. We demonstrate our system's competitive performance through both objective and subjective evaluations in terms of mixing quality and control ability. We provide code and audio samples at our project page https://zys711.github.io/Diff2Mix .

This paper deals with security allocation challenges for networked control systems represented by positive-weighted digraphs under stealthy false data injection attacks. These systems consist of interconnected subsystems, referred to as nodes in the underlying digraph, where an adversary aims to maximize network performance loss by stealthily attacking specific nodes. Meanwhile, a defender monitors several nodes to impose stealthiness constraints on the adversary's actions, thereby minimizing the network performance loss. We analyze the worst-case network performance loss of these stealthy attacks and make the following contributions: we (i) show that the worst-case network performance loss is upper-bounded by a tractable semi-definite programming (SDP) problem; (ii) establish the relationship between the SDP problem and the Katz centrality measure of the underlying digraph under a sufficient condition, resulting in a network-size-independent optimization problem; and (iii) provide a heuristic search based on the Katz centrality measure of the underlying digraph for selecting sub-optimal monitor nodes against all admissible attack scenarios without solving optimization problems. These results offer practical insights for safeguarding large-scale networked control systems against stealthy false data injection attacks. The obtained results are validated via extensive simulations on Erdos-Renyi random graphs with different network sizes.

Optimizing for both low Age of Information (AoI) and high throughput is critical for remote sensing applications that rely on multichannel wireless networks. However, jointly optimizing these two metrics is an analytically challenging problem, particularly in systems with heterogeneous and unreliable channels. To address this challenge, we propose TSDM, a Two-Stage Deficit Matching scheduling framework. TSDM is based on a second-order approach that characterizes the performance of each data flow by its mean and temporal variance. In the first stage, TSDM translates the high-level utility maximization objective into a concrete set of target mean and temporal variance statistics for transmissions over each node-channel pair. In the second stage, a low-complexity Weighted Matching Deficit (WMD) rule performs real-time channel assignment. We theoretically prove that TSDM achieves the desired mean and temporal variance for each flow. Furthermore, we conduct extensive simulations on two open joint throughput-AoI optimization problems. In both cases, TSDM significantly outperforms existing scheduling policies.

Text-controlled symbolic music generation has recently gained research attention due to its versatile, flexible and straightforward approach to music composition. However, previous approaches tend to generate symbolic music with compromising quality, diversity, controllability and limited duration. In this paper, we present Diff-Symbo, an innovative method that uses latent diffusion model (LDM) to generate high-quality, diverse and long-duration symbolic music. To address the lack of text-symbolic music dataset, we develop a comprehensive dataset with 19,345 text templates by employing large language model. Furthermore, we design a music information encoder to reduce the training overhead while extracting more effective control representations. Given textual descriptions, our proposed method leverages LDM to improve the quality and diversity of music generation. Our method also improves the duration and the compositional consistency of music generation through an autoregressive approach. Experimental results show significant improvements of Diff-Symbo in text controllability, duration, and the quality of generated music compared to the baseline models such as GPT-4, MuseCoco and Multitrack Music Transformer (MMT). As one of the pioneer models in this field, Diff-Symbo paves the way towards controllable and high-quality symbolic music composition based on LDM, offering valuable contributions to both music amateurs and practitioners.

This work presents a compressive sensing-based approach for estimating the micro-motion parameters of targets with rotating components, such as small unmanned aerial vehicles (UAVs) with propellers, using fewer measurements than conventional methods. A multiple-input-multiple-output (MIMO) frequency-modulated continuous-wave (FMCW) radar employing a randomly spaced sparse linear antenna array is utilized. Random sequences of linear frequency-modulated (LFM) chirps are transmitted to enable random sampling in the slow-time domain. At first, the range, velocity, and angle of arrival (AoA) of the targets are estimated to identify the bulk motion. A three-dimensional point target response (3D-PTR) is then constructed using the estimated parameters and subtracted from the total radar return to extract the micro-Doppler signatures associated with target rotation. These residual signals are processed within a compressive sensing (CS) framework using the one-dimensional orthogonal matching pursuit (1D-OMP) algorithm to jointly estimate the propellers' rotation frequencies and blade lengths with the help of a parametric dictionary. The proposed approach is also extended to a multi-target scenario. Simulation results demonstrate that the proposed approach accurately estimates micro-motion parameters with limited measurements in the slow-time and the spatial dimensions, validating its potential for UAV detection applications. These estimated parameters are then used to distinguish among various UAV motion types using a classification framework based on a set of decision rules.

Modern Greek is absent from NVIDIA's Nemotron retrieval models and from major multilingual retrieval benchmarks, despite being important for retrieval-augmented generation (RAG) in legal, energy, financial, and medical applications. We present an end-to-end adaptation of the Nemotron retrieval stack for Modern Greek, including corpus mining, synthetic supervision, retrieval model training, reranker adaptation, reader fine-tuning, and a new benchmark called HERA. Our study shows that a parameter-free BM25 baseline outperforms several off-the-shelf multilingual dense retrieval models on specialist Greek corpora. After fine-tuning on 65,773 Greek retrieval pairs, a Nemotron 1B embedder improves nDCG@10 from 0.362 to 0.835 and substantially outperforms its unadapted counterpart. The learned language competence transfers to general-domain Greek, although the advantage over BM25 remains domain-dependent. We further adapt a cross-encoder reranker and demonstrate consistent improvements across specialist domains. Finally, we LoRA-tune a Nemotron 30B-A3B mixture-of-experts reader for grounded generation, increasing judged answer correctness from 29.4% to 66.9% while significantly improving faithfulness and citation quality. We also introduce HERA, the first large-scale Greek benchmark for retrieval-augmented generation, and release our adapted models and benchmark to support future research on Greek-language RAG systems.

In Orthogonal Frequency Division Multiplexing (OFDM)-based massive Multiple-Input Multiple-Output (MIMO) near-field (NF) sensing, target motion induces an antenna-dependent bistatic Doppler variation across the array aperture. Ignoring this spatial Doppler variation leads to a model mismatch that degrades NF localization. In this paper, we propose a low-complexity recursive framework for joint radial/transverse velocity estimation and Doppler-aware localization. Initialized by a constant-Doppler coarse localization, the method alternates between closed-form Least Squares Estimator (LSE)-based velocity estimation and antenna-dependent Doppler-aware localization refinement. Simulation and measurement results demonstrate the effectiveness of the proposed framework against two benchmark methods. Compared with a low-complexity constant-Doppler baseline method, the proposed algorithm improves range, angle, and radial velocity estimation results, while also enabling transverse velocity estimation. In the measurement results, the overall localization error decreases from 0.268 m to 0.064 m. The radial and transverse velocity estimation errors are 0.032 m/s and 0.069 m/s, respectively. Compared with a high-complexity exhaustive four-dimensional (4D) Maximum Likelihood Estimator (MLE), the proposed method achieves comparable velocity estimation results while yielding a more accurate localization result when the 4D MLE has a practical finite search grid.

Long context reasoning in large language models (LLMs) is usually constrained by the fact that a single inference trajectory has to simultaneously explore the context, store intermediate state, verify evidence, and produce the final answer. This becomes particularly difficult in tasks that require extraction, counting, ordering, or multi-hop reasoning, where an early mistake can propagate until the final response. In this work, we propose Chained Recursive Language Models (Chained RLM), an inference-time architecture, in which the same underlying model is called repeatedly as a sequence of fresh reasoning roots. Each root receives the original problem and context, but does not inherit the full conversational history. Instead, it receives a compact plain-text summary, a plain-text blackboard, and some durable task-specific artifacts written by predecessor roots. The motivation is to manage the context by chopping into partial tasks rather than one large inference response; in each staged computation, intermediate artifacts can be inspected, corrected, and extended by a later fresh inference by the same model. We describe the system model, handoff mechanism, artifact workspace, and evaluation protocol for this system. We study when fresh-context artifact continuation gives a measurable gain in accuracy over direct LLM answering even with recursive tool-calling.

Significance. Accurate intraoperative depth perception is important for autonomous and semi-autonomous robotic laparoscopic surgery. Conventional fringe projection profilometry can achieve millimeter-scale accuracy but often requires multi-shot acquisition, digital-micromirror-device projection, and projector-camera synchronization, complicating integration into compact laparoscopic systems. Aim. To develop a synchronization-free, single-shot depth-sensing platform using a passive LED-illuminated binary mask and a VQ-VAE prior with a custom U-Net depth head. Approach. A compact projection module was coupled to one channel of a dual-channel laparoscope, while the second channel imaged the fringe-illuminated target. A Zivid 3D camera acquired reference depth for 722 paired phantom images. Zivid depth maps were reprojected into the SSLE image frame for supervised training and evaluation. The VQ-VAE encoded each input into a discrete latent representation, and a latent-space U-Net predicted depth without a separate mask-prediction branch. Results. Using a fixed train/validation/test split, the proposed model achieved an MAE of 3.70 mm, AbsRel of 0.0326, delta=1.1 accuracy of 0.962, and delta=1.1^2 accuracy of 0.970. It achieved lower MAE than the dual U-Net MaskNet + DepthNet baseline and outperformed off-the-shelf monocular depth models in MAE, AbsRel, and threshold accuracy. The pipeline operated at 26.0 Hz over 301 consecutive frames on an NVIDIA A100 GPU. Conclusions. The LED-illuminated binary-pattern platform with latent-space depth reconstruction enables synchronization-free, video-rate endoscopic depth estimation. Results demonstrate Zivid-referenced phantom reconstruction without an explicit segmentation stage, while emphasizing the importance of dataset size and SSLE-Zivid calibration accuracy.

Recent advances in machine learning have enabled training of wireless foundation models, which aim to support tasks such as channel estimation, beam prediction, and localization based on wireless signals. Existing wireless foundation models typically pretrain on channel tensors using masked reconstruction over subcarriers, antennas, or time but ignore the physical characteristics of wireless propagation. In this work, we propose to instead use multipath propagation as the fundamental pretraining object. We present MultiPathFormer, an autoregressive foundation model that represents each transmitter-receiver link as an ordered sequence of continuous-valued path tokens and pretrains with next-path prediction. We introduce an Environmental RAG (retrieval-augmented generation) mechanism and a first-path codebook on top of the transformer backbone, leveraging environment knowledge to improve path statistics estimation like delay and power by up to 59%. MultiPathFormer pretrained on 27 environments transfers to unseen users and, after scenario-specific fine-tuning, outperforms training the corresponding models from scratch in new environments. Across downstream tasks, it outperforms SOTA channel-based foundation models, achieving 5.57 m mean localization error, 0.914 top-3 beam accuracy, 0.994 line-of-sight classification accuracy, and 0.561 channel estimation NMSE. These results show that path-level pretraining can learn reusable representations of wireless propagation.