Papers

Reliable low-latency communication is a key requirement for mission-critical and mobile autonomous systems, including teleoperation, autonomous navigation, and real-time uplink-dominant telemetry applications. While commercial 5G networks often provide adequate downlink performance, uplink performance in rural deployments may be constrained by radio-resource limitations and uplink power-control mechanisms. This paper presents a comprehensive experimental evaluation of multi-connectivity strategies over commercial 5G Non-Standalone networks, based on measurement campaigns conducted in urban, suburban, and rural environments. The study analyzes per-packet uplink and downlink latency, packet loss, and radio-layer KPIs across two mobile network operators. The measurements indicate that latency and reliability cannot be inferred solely from coverage indicators such as RSRP. In coverage-constrained scenarios, performance appears to be strongly influenced by uplink power-limited operation and partially correlated impairments across operators. Several multi-connectivity strategies are evaluated, including link aggregation, switching-based policies, and conditional packet duplication. A Primary-Anchored Adaptive Failover (PAAF) framework is introduced to selectively activate redundancy based on radio, latency and service cost considerations. The results suggest that Partial Duplication (PD) approaches can approach the reliability of multi-connectivity while substantially reducing duplication overhead in the evaluated rural scenario.

In pathology, the spatial distribution and proportions of tissue types are key indicators of disease progression, and are more readily available than fine-grained annotations. However, these assessments are rarely mapped to pixel-wise segmentation. The task is fundamentally underdetermined, as many spatially distinct segmentations can satisfy the same global proportions in the absence of pixel-wise constraints. To address this, we introduce Variational Segmentation from Label Proportions (VSLP), a two-stage framework that infers dense segmentations from global label proportions, without any pixel-level annotations. This framework first leverages a pre-trained transformer model with test-time augmentation to produce a pixel-wise confidence estimate. In the second stage, these estimates are fused by solving a variational optimization problem that incorporates a Wasserstein data fidelity term alongside a learned regularizer. Unlike end-to-end networks, our variational method can visualize the fidelity-regularization energy, resulting in more interpretable segmentation. We validate our approach on two public datasets, achieving superior performance over existing weakly supervised and unsupervised methods. For one of these datasets, proportions have been estimated by an experienced pathologist to provide a realistic benchmark to the community. Furthermore, the method scales to an in-house dataset with noisy pathologist labels, severely outperforming state-of-the-art methods, thereby demonstrating practical applicability. The code and data will be made publicly available upon acceptance at https://github.com/xiaoliangpi/VSLP.

Microwave linear analog computer (MiLAC)-aided transmit beamforming, which processes transmitted symbols entirely in the analog domain, has recently emerged as a promising alternative to fully digital or hybrid beamforming architectures for single-user multi-antenna systems. However, recent studies have shown that deploying a single lossless and reciprocal MiLAC at the transmitter cannot achieve the same capacity as fully digital beamforming in multi-user scenarios. To address this limitation, we propose a novel two-layer MiLAC-aided beamforming architecture at the transmitter for a downlink multi-user multiple-input single-output (MISO) network. Leveraging microwave network theory, we first prove that lossless and reciprocal two-layer MiLAC-aided beamforming can achieve the same performance as digital beamforming, and we derive a closed-form mapping from digital beamforming to two-layer MiLAC analog beamforming. Furthermore, we formulate a sum-rate maximization problem and develop an efficient optimization framework to jointly optimize the power allocation and the scattering matrices for the proposed two-layer MiLAC architecture. Numerical results validate our theoretical findings and demonstrate that two-layer MiLAC achieves the same sum-rate performance as fully digital beamforming.

Artificial intelligence is increasingly described as a candidate next generation general purpose technology (GPT). However, existing interpretations predominantly emphasize performance scaling rather than structural transformation. This paper introduces a formal framework for evaluating AI as a systemic infrastructural transition rather than merely a computational breakthrough. We propose the concept of AI Native Autonomous Infrastructure (ANAI), defined as a regime in which decision autonomy becomes embedded within critical infrastructures. The framework operationalizes this transition through three quantitative constructs: the Autonomy Index (AIx), the Infrastructure Coupling Coefficient (ICC), and the Technological Transition Potential (TTP). We formalize the joint scaling dynamics of autonomy and infrastructural embedding, derive threshold conditions for paradigm transition, and introduce a phase-space representation of systemic transformation. A temporal transition model further illustrates how nonlinear coevolution between autonomy and infrastructure integration produces super linear growth in transition potential. Unlike prior GPT cycles, the ANAI regime exhibits a recursive energy computation feedback loop in which AI systems both increase computational demand and optimize the infrastructures that sustain them. This feedback mechanism accelerates infrastructural embedding and differentiates AI driven transformation from previous technological revolutions. By shifting analytical focus from model performance to infrastructural autonomy and coupling intensity, this study offers a conceptual and mathematical foundation for assessing whether artificial intelligence constitutes the next general purpose technology.

We establish a framework for realizing back-action-evading (BAE) measurements and quantum non-demolition (QND) variables in linear quantum systems. The key condition, a purely imaginary Hamiltonian with a real or imaginary coupling operator, enables BAE measurements of conjugate observables. Symmetric coupling further yields QND variables. For non-compliant systems, coherent feedback is designed to engineer BAE measurements. Crucially, the QND interaction condition simultaneously ensures BAE measurements and promotes the coupling operator to a QND observable.

The escalating climate crisis and ecosystem degradation demand intelligent, low-cost sensors capable of robust, long-term monitoring in real-world environments. Absolute dissolved oxygen (DO) concentration is a key parameter for predicting climate tipping points. Inexpensive optoelectronic sensors based on microstructured polymer films doped with phosphorescent dyes could be readily deployable; however, signal drift and marine biofouling remain major challenges. Here, we introduce a sensing paradigm that combines camera-based DO sensors with a visual transformer (ViT)-based physics-informed neural network (PINN) for high-fidelity sensing under biofouling conditions. Training and testing data were obtained from an algae-laden water tank over 14 days to capture accelerated biofouling. The ViT-PINN, which embeds the Stern-Volmer (SV) equation into the loss function, reduces mean average error (MAE) by 92% and 89% compared to classical statistical and ML approaches, achieving ~2 umol/L absolute error. A deep ensemble further quantifies predictive uncertainty, enabling self-diagnostic sensing.

The rapid growth of the low-altitude economy drives increasingly autonomous unmanned aerial vehicle (UAV) operations, giving rise to low-altitude embodied intelligence (LAEI), in which sensing, communication, computation, and control (SC$^3$) are tightly integrated to enable closed-loop interaction, ensuring timely, effective, and safe responses in complex or unknown environments. This article systematically explores the LAEI networks, from its fundamental architecture to the diverse scenarios that it can support. We examine key enabling techniques that sustain timely information exchange and effective decision feedback within the $\text{SC}^3$ closed loop. A representative low-altitude UAV mission in an unknown urban area is presented as a case study, where the UAV provides communication services and performs environmental sensing to inform closed-loop control, illustrating how coordinated $\text{SC}^3$ capabilities enable efficient and responsive operation. By identifying major challenges and outlining future research directions, this work serves as a cornerstone for developing next-generation low-altitude intelligent systems.

With the shift towards decentralized energy generation, the increasing complexity of power systems renders physics-based modeling challenging. At the same time the growing amount of available measurement data opens the door for obtaining data-driven models. A modern method to do so are neural ordinary differential equations (NODEs), offering a framework for continuous-time system identification. Recent extensions, so called graph NODEs impose a structural inductive bias that has the potential to improve generalization of the learned representations. In this work, we employ and extend them to develop message-passing graph NODEs (MPG-NODEs) for identification of coupled systems with heterogeneous node dynamics and heterogeneous edge couplings. This includes state-of-the-art machine learning architectures to infer latent representations of unmeasured states, local node and edge embeddings to account for heterogeneity as well as an autoregressive scheme to allow for piecewise constant control inputs. We apply MPG-NODEs to identify voltage and frequency dynamics of power systems and compare them to a monolithic NODE under identical measurement assumptions. Our case study indicates that while achieving similar performance, our proposed MPG-NODEs are more flexible with transfer learning options that allow to add or remove edges and nodes with little retraining.

We propose Speech Enhancement based on Drifting Models (DriftSE), a novel generative framework that formulates denoising as an equilibrium problem. Rather than relying on iterative sampling, DriftSE natively achieves one-step inference by evolving the pushforward distribution of a mapping function to directly match the clean speech distribution. This evolution is driven by a Drifting Field, a learned correction vector that guides samples toward the high-density regions of the clean distribution, which naturally facilitates training on unpaired data by matching distributions rather than paired samples. We investigate the framework under two formulations: a direct mapping from the noisy observation, and a stochastic conditional generative model from a Gaussian prior. Experiments on the VoiceBank-DEMAND benchmark demonstrate that DriftSE achieves high-fidelity enhancement in a single step, outperforming multi-step diffusion baselines and establishing a new paradigm for speech enhancement.

Separating multiple graph signals from a single observed mixture is an inherently ill-posed problem that traditionally relies on restrictive and handcrafted priors. This letter addresses this challenge by proposing an unsupervised learnable spectral filtering framework. Our approach reconstructs latent components by passing a fixed random input through learnable spectral filters, operating within the low-frequency eigenspace of each source-specific graph Laplacian. The architecture implicitly biases the recovered signals toward smooth patterns by confining reconstruction to these low-frequency subspaces. This acts as a structural prior, establishing a principled bridge between classical graph spectral analysis and modern neural decomposition. Numerical experiments confirm that this framework successfully isolates individual sources using solely the observed mixture and the underlying graph topology.

Pretrained diffusion models have revolutionized real-world image super-resolution (Real-ISR), but their iterative sampling is computationally prohibitive, driving efforts to distill it into a single step. General one-step methods fine-tune the generative prior into a deterministic mapping, restoring efficiency but discarding its stochastic nature. Conversely, recent attempts re-engage generation by shifting the timestep or injecting random noise, adjusting either the position or the state while the other stays fixed. Because only one side is controlled, the two align at isolated preset timesteps but drift apart once steered, leaving generation unstable. To address this, we present one-step diffusion via Inversion and Degradation-aware Sampling for Real-ISR (IDaS-SR), a one-step framework that bridges deterministic restoration and stochastic generation. At its core, Manifold Anchoring grounds the low-quality latent on the pretrained trajectory through two operations jointly estimated by the Manifold Inversion Noise Estimator (MINE): positioning declares where the latent lies and how it deviates from the clean state, while inversion aligns the latent to the declared position. Upon the anchor, CHARIOT reintroduces controlled stochasticity by jointly rescheduling the trajectory and interpolating the noise, enabling a single scalar to smoothly navigate the fidelity-realism trade-off. Extensive experiments demonstrate that IDaS-SR effectively unleashes the generative prior, achieving state-of-the-art performance under explicit control in a single inference step.

This paper addresses the planning and control problem for nonlinear systems under Signal Temporal Logic (STL) specifications. We first decompose an STL task into finite local tasks. A sampling-based method generates sequences of local waypoints to satisfy all local tasks, from which the corresponding satisfaction pair sets are derived. Following a local-to-global strategy, all sequences of local waypoints are synthesized into a global one, based on which a safe corridor is then constructed. Leveraging the safe corridor and the satisfaction pair sets, an optimization problem is formulated and solved to derive a position trajectory that satisfies the STL task. Finally, numerical examples and comparative results are presented to demonstrate the efficacy of the proposed approach.

Federated learning (FL) offers a promising distributed learning paradigm for internet of vehicles (IoV) applications. However, it faces challenges from communication overhead and dynamic environments. Model compression techniques reduce computing and communication burden yet create trade-offs between compression ratios and vehicle participation strategies. In this paper, we propose a lightweight FL algorithm named federated learning with dynamic low-rank adaptation (Fed-DLoRA), which is combined with low-rank adaptation (LoRA) to effectively reduce parameters and communication costs while enhancing training efficiency. The convergence analysis of Fed-DLoRA is conducted through stochastic gradient descent optimization coupled with singular value decomposition. This analysis establishes the theoretical relationships among LoRA rank, vehicular scheduling strategies and the model's convergence characteristics. Building on these insights, we formulate a joint optimization problem aimed at maximizing system performance. To address this problem, we propose an adaptive rank, bandwidth and vehicle selection (ARBVS) algorithm that integrates enumeration with greedy optimization strategies. The algorithm provides efficient rank selection and resource scheduling strategies for each FL communication round, thereby achieving effective performance improvements for the FL system. Experimental results demonstrate that Fed-DLoRA achieves superior performance compared to conventional federated learning approaches, exhibiting enhanced accuracy, faster convergence, and improved communication efficiency.

Precise synchronization is essential in various technical disciplines, being especially challenging in mobile scenarios. Unfortunately, state-of-the-art global navigation satellite system (GNSS) disciplined oscillators (GNSSDOs) are designed and optimized for stationary operation. We present a novel solution that is optimized for mobile use from the ground up. The centerpiece is a precise oven-controlled crystal oscillator (OCXO) that is optimized for low sensitivity to dynamic accelerations. A state-of-the-art GNSS timing module is used to discipline it. We evaluate the system by comparing it with state-of-the-art test equipment in a real-world test drive through diverse environments. After compensating for the stationary offset, the state-of-the-art devices deviated by up to 2315 ns, while with our devices, the deviation never exceeded 22.6 ns. It is evident that the devices designed for laboratory use perform inadequately in mobile operation and that our novel solution enables a significant leap in accuracy.

This paper studies economic model predictive Control (EMPC) schemes, where the stage cost depends only on control inputs. Such problems arise in applications like water distribution networks and differ from standard EMPC since multiple steady states can correspond to the unique optimal steady input. We show that, under a strict dissipativity assumption related to the set of optimal steady states, the closed-loop trajectories converge asymptotically to this set, ensuring convergence of the economic cost to the optimal steady state cost. To enhance Lyapunov stability, we propose a modified stage cost that preserves the optimal input while guaranteeing asymptotic stability of a specific equilibrium with a slight performance loss. The approach is further extended to EMPC of a class of linear systems with periodic costs and disturbances by lifting it to a multi-step EMPC problem for periodic operations. A case study with a water distribution network demonstrates the effectiveness of the proposed methods in achieving both asymptotic convergence and stability.

While Multimodal Large Language Models (MLLMs) have enhanced grounding capabilities in general scenes, their robustness in crowded scenes remains underexplored. Crowded scenes entail visual challenges (i.e., occlusion and small objects), which impair object semantics and degrade grounding performance. In contrast, language expressions are immune to such degradation and preserve object semantics. In light of these observations, we propose a novel method that overcomes such constraints by leveraging Language-Guided Semantic Cues (LGSCs). Specifically, our approach introduces a Semantic Cue Extractor (SCE) to derive semantic cues of objects from the visual pipeline of an MLLM. We then guide these cues using corresponding text embeddings to produce LGSCs as linguistic semantic priors. Subsequently, they are reintegrated into the original visual pipeline to refine object semantics. Extensive experiments and analyses demonstrate that incorporating LGSCs into an MLLM effectively improves grounding accuracy in crowded scenes.

Radio Frequency Fingerprinting (RFF) is a key technology for identity authentication in wireless networks. However, due to the rapid dynamics of Autonomous Aerial Vehicles (AAVs) in low-altitude wireless networks, RFF models require parameter updates to maintain authentication performance, posing a major challenge to existing schemes. Conventional retraining approaches for handling departed or compromised AAVs are computationally prohibitive and risk retaining polluted features, which compromises both authentication security and user privacy. To address these limitations, we propose an Input-Perturbation-based RFF Unlearning (IPRU) scheme. By optimizing a universal Fingerprint Forget Vector (FFV) as a lightweight input perturbation, IPRU successfully erases the fingerprints of target AAVs without modifying the RFF model parameters, achieving an effective balance between efficient unlearning and preserved authentication performance. A combinatorial optimization strategy further enables multi-AAV forgetting on demand. The simulation results demonstrate that IPRU achieves 1.41% unlearning accuracy, 99.41% remaining accuracy, and 100% resistance to membership inference attack, while running 5.79X faster than retraining and 2.1X faster than the baseline scheme.

This paper considers multiple extended object tracking based on Poisson multi-Bernoulli mixture (PMBM) filtering, which gives the closed-form Bayesian solution for standard multiple extended object models with Poisson birth. To efficiently address the challenging extended object data association problem in PMBM filtering, we develop implementations of the extended object PMBM filter using blocked Gibbs sampling. By formulating the PMBM density on an augmented state space with auxiliary variables and leveraging the Poisson object measurement model, we first derive a joint posterior over potential objects, previous global hypotheses, and current measurement association variables, together with its corresponding factorization. This factorized representation leads to blocked Gibbs samplers that can efficiently generate high-weight global hypotheses and thereby provide an efficient implementation of the PMBM update step. We further introduce a collapsed Gibbs sampling variant, in which the Bernoulli object existence variables are marginalized out, yielding higher sampling efficiency, especially for the initiation of newly detected objects. The proposed methods, implemented under the gamma Gaussian inverse-Wishart model, are compared with existing extended object PMBM baselines with different data association strategies. Simulation results demonstrate that the proposed approaches achieve comparable tracking performance while requiring substantially less runtime.

The rapid growth of large, power electronics-rich data center loads (DCLs) is creating new operational challenges for bulk power systems. A key risk arises when a DCL's uninterruptible power supply (UPS) disconnects the facility during voltage/frequency disturbances and then reconnects it while the bulk grid is still dynamically settling to a new equilibrium point. Poorly timed reconnection can amplify electromechanical oscillations, deepen frequency deviations, and lead to repeated connect-disconnect flapping. In this paper, we develop an analytical framework to characterize the safe reconnection windows for large DCLs after a disturbance-induced disconnection that satisfy the prescribed limits. Using a model in the spirit of the classical single-machine infinite-bus system, we capture (i) swing dynamics during the disconnection interval and (ii) voltage-angle coupling at the load bus, which determines the electrical power step at reconnection under constant power load assumptions. Using the energy function method, we characterize the certified safe reconnection windows such that the post-reconnection trajectory is guaranteed to remain within operational limits (rotor speed/angle/voltage). Time-domain simulations validate the effectiveness of the proposed analytical approach. The results provide a simple physics-informed criterion that can be used to identify certified reconnection windows for large DCL facilities and inform UPS reconnection logic.

Line spectral estimation (LSE) involves estimating both spectral frequencies and their associated complex amplitudes. Existing Fisher-information-based benchmarks are local and therefore do not capture either the threshold behavior of frequency estimation or the propagation of frequency errors to subsequent amplitude reconstruction. This paper develops explicit performance benchmarks for LSE from two complementary perspectives: ordered frequency estimation and plug-in amplitude reconstruction. On the frequency side, we develop a computable Ziv-Zakai bound (ZZB)-type benchmark under an ordered prior by combining a generalized-likelihood-ratio-test (GLRT)-based surrogate for the unavailable pairwise kernel with an ordered-prior correction. The resulting benchmark recovers the ordered a priori bound at low signal-to-noise ratio (SNR) and the marginalized frequency-side Cramer-Rao bound (CRB) at high SNR. On the amplitude side, we derive a local transfer characterization for the sequential plug-in estimator and obtain a computable benchmark for the induced amplitude error. The resulting framework explicitly characterizes threshold behavior on the frequency side and error propagation on the amplitude side. Numerical results support the proposed benchmarks across different SNR regimes, snapshot numbers, and model orders.

The shift to the radiative near field region due to large antenna arrays necessitates beamforming that accounts for both angle and range, evolving mobility management into a joint angular range tracking challenge. Conventional schemes rely on rigid pilot payload structures with dedicated training slots, which interrupt data transmission and degrade spectral efficiency. To address this, we propose a pilot-free beam tracking framework leveraging Thompson sampling(TS). Within each sliding window, the user trajectory is modeled by local low-order polynomials in angle and range, and the motion parameters are estimated by maximum likelihood with uncertainty quantified via the Fisher information matrix. TS adaptively probes uncertain trajectory regions using beams that simultaneously serve as payload beams. Simulations demonstrate that the proposed framework maintains reliable connectivity while eliminating the overhead of dedicated pilot-based beam sweeping.

Electrooculogram (EOG) is a non-invasive bio-signal generated by the potential difference between the retina and cornea during eye movement, and is widely utilized in Human-Computer Interaction (HCI) systems. Expanding the range of detectable eye movements enhances system capability. However, increasing the number of classes typically degrades classification performance. While AI-based approaches can mitigate this limitation, their complexity increases significantly when operating on single-cycle EOG signals. Although single-cycle signals offer advantages such as low latency, reduced power consumption, and improved responsiveness, they are inherently limited by reduced informational content and higher susceptibility to noise. Ensuring low latency remains critical for real-time HCI applications, where system response must remain below human reaction thresholds. In this experimental study, using explainable AI, we address these challenges by developing 1-dimensional (1D) and cascaded ANN and CNN architectures capable of highly accurate classification across ten EOG classes (Stare, Blink, Up, Down, Right, Left, Up-left, Up-right, Down-left, and Down-right) using single-cycle signals, while simultaneously achieving latency substantially lower than human reaction time. The study achieved an accuracy of around 99% for all the models with a latency of 38.6 ms for the 1D ANN, and 82.85 ms for the cascaded CNN. These findings confirm that cascaded neural network architectures, can effectively balance high classification accuracy and low latency for single-cycle, multi-class EOG-based HCI systems under limited data availability.

The Laplacian operator transforms the image into its Laplacian field, which usually is sparse and satisfies a stable distribution. On the other hand, an image can be uniquely reconstructed from its Laplacian field via solving a Poisson equation with a proper boundary condition. Such uniqueness is mathematically guaranteed. Thanks to these properties, we propose to use the sparse Laplacian field to present the image. We first show that the Laplacian field is sparse and satisfies a stable distribution on hundreds images. Then, we show that the image can be accurately reconstruct from its Laplacian field. For the reconstruction task, we propose a shared-kernel wavelet neural network, which solves the Poisson equation and has three advantages. First, it has less than 0.0002M parameters, which is compact enough for most of devices. Second, it has linear computation complexity, leading to a real-time reconstruction. Third, it achieves higher accuracy than previous methods. Several numerical experiments are conducted to show the effectiveness and efficiency of the sparse Laplacian field and the proposed Poisson solver. The proposed method can be applied in a large range of applications such as image compression, low light enhancement, object tracking, etc.

A safety filter based on a neural control barrier function (CBF) deployed in an embedded control loop evaluates, at each control cycle, the trained network and its Lie derivatives along the system vector fields, under the memory and worst-case execution time (WCET) constraints that safety-oriented coding standards impose. Reverse-mode automatic differentiation, by which training frameworks obtain these derivatives, retains an activation cache whose size grows with the sum of the layer widths, and general-purpose differentiation runtimes allocate the computational graph from the heap at each call. This paper presents a compiler that evaluates a neural CBF and its exact Lie derivatives by forward-mode dual-number arithmetic. The compiler emits self-contained C++ code in which a single forward pass, without backpropagation, returns the barrier value and its exact Lie derivative along a given vector field; the drift and input Lie derivatives of the safety constraint are obtained from one such pass per vector field, and a second-order extension based on hyper-dual numbers returns the exact second-order Lie derivatives required by CBFs of relative degree two. The dual forward pass requires a workspace bounded by four times the widest layer, independent of network depth, and the emitted code contains no allocation call sites, so the absence of dynamic allocation is verifiable by inspection of the code. On an ESP32-S3 microcontroller, the compiled filter assembles the complete safety constraint in under one millisecond from statically allocated buffers of at most 768 bytes, and the maximum execution time over 1000 evaluations lies within 5% of the median in all three examples, whereas a heap-allocating reverse-mode baseline shows maxima 33% and 70% above its median in the two first-order examples. The compiler and the embedded experiments are released as open-source software.

Autonomous driving systems often infer pedestrian yielding behavior from geometric and kinematic cues alone, limiting their ability to reason about visual scene context and age-dependent behavioral variability. This limitation can produce delayed interventions in safety-critical encounters and unnecessary braking in benign interactions. This work introduces Vision-Language Model-based Vehicle-Pedestrian Interaction (VLM-VPI), a multimodal reasoning framework for pedestrian intent understanding and yielding-aware control in autonomous driving. The system combines three components: a multimodal perception layer that captures visual and kinematic observations, a reasoning layer that uses Qwen3-VL 8B for visual scene understanding and GPT-OSS 20B for few-shot intent reasoning, and a tiered safety controller that applies age-specific braking margins for children, adults, and seniors. In 112 CARLA scenarios, VLM-VPI achieves 92.3% intent classification accuracy, outperforming a rule-based baseline (78.4%), supervised trajectory models (73.5-82.4%), and a zero-shot LLM configuration (88.4%). Validation on 24 real-world PIE scenarios yields 87.5% accuracy, indicating functional sim-to-real transferability. Across 200 simulation cases, VLM-VPI reduces the false-alarm rate from 7.4% to 2.8% and mean intersection traversal time from 13.5 s to 11.8 s. Conflict occurrences decrease from 124 to 33, while mean minimum time-to-collision improves from 1.92 s to 4.47 s. Demographic-adaptive control further reduces conflicts by 60% for children and 54.5% for seniors compared with uniform control. These results show that an explicit vision-language reasoning layer can improve both safety and efficiency by linking pedestrian intent, demographic context, and vehicle control decisions.

Developing robust and clinically reliable EEG biomarkers requires evaluation frameworks that explicitly address cross population generalization in multi site settings such as Parkinsons disease (PD) detection. Models trained under i.i.d. assumptions often capture population specific artifacts rather than disease relevant neural structure, leading to poor generalization across clinical cohorts. EEG further amplifies this challenge due to low signal to noise ratio and heterogeneous acquisition conditions. We propose a population aware evaluation framework to assess the robustness and clinical reliability of EEG biomarkers under distribution shift. Using an n gram expansion strategy, we enumerate all cross population train test configurations across five independent cohorts, resulting in 75 directional evaluations. A nested cross validation design with integrated channel selection ensures prospective biomarker identification without population leakage. Results show that cross population transfer is asymmetric and that both accuracy and biomarker stability improve with increasing training population diversity, achieving up to 94.1% accuracy on held out cohorts. A theoretical analysis based on mixture risk optimization and hypothesis space contraction explains these trends, showing that multi population training promotes population robust representations. This work establishes a principled framework for learning robust, generalizable, and clinically reliable EEG biomarkers for multi site biomedical applications.

This paper investigates the cross-frequency structure of background clutter induced by random dispersive media in single-snapshot FDA-MIMO-GPR. Representative media are modeled by the Cole–Cole formulation to relate dispersive constitutive behavior to the reference propagation environment and observation-domain statistics. A normalized incremental contrast function is introduced under a reference-medium framework, and a single-snapshot background-response expression with first-order propagation-kernel feedback is derived. Based on this expression, a cross-frequency coupling strength of the leading-order background covariance is defined. Numerical results show that, in weakly dispersive scenes, the proposed analysis remains consistent across constitutive mapping, the zeroth-order propagation skeleton, first-order distorted-Born truncation, propagation-kernel feedback, and single-channel response closure. The proposed metric distinguishes uncoupled and explicitly coupled constructions, remains stable under pure energy scaling, responds clearly to correlation length and relaxation-location parameters, and corresponds directly to the error of the frequency block-diagonal approximation. Additional experiments show that the resulting cross-frequency structure affects whitening and principal-subspace extraction. In scenes with pronounced relaxation, abrupt breakdown under strong perturbations and high-error plateaus indicate that the present theory is mainly applicable within the validity range of first-order feedback.

Speckle noise severely limits the quality of images acquired from coherent imaging systems such as Synthetic Aperture Radar (SAR) and medical ultrasound. Traditional second-order PDE-based despeckling approaches, although popular, often introduce staircase artifacts and blur fine details. To overcome these limitations, we present a nonlinear, fourth-order coupled hyperbolic-parabolic PDE model that effectively reduces noise while preserving the structure. The framework consists of two evolution equations: one governing fourth-order diffusion for effective speckle reduction and smooth intensity transitions, and another refining an edge indicator to protect textures and structural features. The diffusion coefficient is adaptively constructed using both the image intensity variable u and a grayscale-based indicator function, ensuring structure-aware denoising while avoiding blocky artifacts and preserving fine details. We also prove the existence of a weak solution to the proposed model by applying Schauder fixed-point theorem. A finite-difference scheme with Gauss Seidel iteration is employed for efficient implementation. We compare the proposed model with the existing coupled second-order PDE model (HPCPDE) and the fourth-order telegraph diffusion model (TDFM). The results show that our model consistently outperforms these approaches. Experiments on standard grayscale images, real SAR and ultrasound data, as well as speckle-corrupted color images, demonstrate that the proposed method achieves superior performance over conventional PDE-based techniques in terms of PSNR, MSSIM, and Speckle Index.

Rehabilitation aims to help patients with limited mobility regain their physical abilities through targeted movements, exercises, stimulation, and other therapeutic methods. Recent advances in technology have introduced sensor-based systems into rehabilitation and clinical practices, enabling real-time monitoring and providing accurate feedback on movement accuracy. Among these sensors, LiDAR has demonstrated strong potential, offering key advantages over conventional techniques such as camera-based systems, which raise privacy concerns, and wearable sensors, which can be uncomfortable and prone to errors. In this work, we review the applications of LiDAR in rehabilitation, post-injury care, and hospital environments, focusing on studies published between 2019 and 2025. Studies across several areas have been explored: 3D body scanning and gait analysis with standalone LiDAR, LiDAR mounted on robotic systems for rehabilitation, real-time monitoring and environment scanning for safe navigation, and activity and position recognition. We also analyze processing techniques, particularly learning-based approaches, and support the discussion with statistical analysis, highlighting trends, gaps, and future research opportunities. To the best of our knowledge, this is the first comprehensive survey dedicated to LiDAR for rehabilitation applications, providing an overview of current methods, AI-based processing techniques, and open challenges.

Short term electricity price forecast is essential in competitive power markets, yet electricity price series exhibit high volatility, irregularity, and non-stationarity. This phenomenon is pronounced in the South Australian region of the National Electricity Market, where high renewable penetration drives price volatility and frequent negative price intervals, while structural changes such as the transition to five-minute settlement further complicate forecast. To address these challenges, this study develops a unified benchmark framework. Under identical data preprocessing, feature engineering with lag features, rolling statistics, cyclic temporal encodings, and so on, and an 85% to 15% chronological train test split, six algorithms are systematically compared, including AWMLSTM, CatBoost, GBRT, LSTM, LightGBM, and SVR. The results show that for price prediction, tree-based models, especially GBRT with an R squared value of 0.88, generally outperform LSTM and SVR. However, all models achieve a mean absolute percentage error above 90%, and more than 65% of GBRT predictions have relative errors above 10%, which highlights the inherent difficulty of price forecast. For demand prediction, all models perform substantially better than in price prediction. AWMLSTM and GBRT achieve an R2 value of 0.96 with mean absolute percentage error below 32%, and GBRT has 74.37% of samples within 5% error, while LSTM and SVR perform less accurately in both tasks. Future improvements should focus on hybrid models such as tree plus transformers, data augmentation for extreme events, and error correction to better capture price spikes.