Papers

Temporal detection problems appear in many fields including time-series estimation, activity recognition and sound event detection (SED). In this work, we propose a new approach to temporal event modeling by explicitly modeling event onsets and offsets, and by introducing boundary-aware optimization and inference strategies that substantially enhance temporal event detection. The presented methodology incorporates new temporal modeling layers - Recurrent Event Detection (RED) and Event Proposal Network (EPN) - which, together with tailored loss functions, enable more effective and precise temporal event detection. We evaluate the proposed method in the SED domain using a subset of the temporally-strongly annotated portion of AudioSet. Experimental results show that our approach not only outperforms traditional frame-wise SED models with state-of-the-art post-processing, but also removes the need for post-processing hyperparameter tuning, and scales to achieve new state-of-the-art performance across all AudioSet Strong classes.

Emergency vehicles require rapid passage through congested traffic, yet existing strategies fail to adapt to dynamic conditions. We propose a novel hierarchical graph neural network (GNN)-based multi-agent reinforcement learning framework to coordinate connected vehicles for emergency corridor formation. Our approach uses a high-level planner for global strategy and low-level controllers for trajectory execution, utilizing graph attention networks to scale with variable agent counts. Trained via Multi-Agent Proximal Policy Optimization (MAPPO), the system reduces emergency vehicle travel time by 28.3% compared to baselines and 44.6% compared to uncoordinated traffic in simulations. The design achieves near-zero collision rates (0.3%) while maintaining 81% of background traffic efficiency. Ablation and generalization studies confirm the framework's robustness across diverse scenarios. These results demonstrate the effectiveness of combining GNNs with hierarchical learning for intelligent transportation systems.

Grant-free cell-free massive multiple-input multiple-output (GF-CF-MaMIMO) systems are anticipated to be a key enabling technology for next-generation Internet-of-Things (IoT) networks, as they support massive connectivity without explicit scheduling. However, the large amount of connected devices prevents the use of orthogonal pilot sequences, resulting in severe pilot contamination (PC) that degrades channel estimation and data detection performance. Furthermore, scalable GF-CF-MaMIMO networks inherently rely on distributed signal processing. In this work, we consider the uplink of a GF-CF-MaMIMO system and propose two novel distributed algorithms for joint activity detection, channel estimation, and data detection (JACD) based on expectation propagation (EP). The first algorithm, denoted as JACD-EP, uses Gaussian approximations for the channel variables, whereas the second, referred to as JACD-EP-BG, models them as Bernoulli-Gaussian (BG) random variables. To integrate the BG distribution into the EP framework, we derive its exponential family representation and develop the two algorithms as efficient message passing over a factor graph constructed from the a posteriori probability (APP) distribution. The proposed framework is inherently scalable with respect to both the number of access points (APs) and user equipments (UEs). Simulation results show the efficient mitigation of PC by the proposed distributed algorithms and their superior detection accuracy compared to (genie-aided) centralized linear detectors.

Pathology foundation models (PFMs) have become central to computational pathology, aiming to offer general encoders for feature extraction from whole-slide images (WSIs). Despite strong benchmark performance, PFM robustness to real-world technical domain shifts, such as variability from whole-slide scanner devices, remains poorly understood. We systematically evaluated the robustness of 14 PFMs to scanner-induced variability, including state-of-the-art models, earlier self-supervised models, and a baseline trained on natural images. Using a multiscanner dataset of 384 breast cancer WSIs scanned on five devices, we isolated scanner effects independently from biological and laboratory confounders. Robustness is assessed via complementary unsupervised embedding analyses and a set of clinicopathological supervised prediction tasks. Our results demonstrate that current PFMs are not invariant to scanner-induced domain shifts. Most models encode pronounced scanner-specific variability in their embedding spaces. While AUC often remains stable, this masks a critical failure mode: scanner variability systematically alters the embedding space and impacts calibration of downstream model predictions, resulting in scanner-dependent bias that can impact reliability in clinical use cases. We further show that robustness is not a simple function of training data scale, model size, or model recency. None of the models provided reliable robustness against scanner-induced variability. While the models trained on the most diverse data, here represented by vision-language models, appear to have an advantage with respect to robustness, they underperformed on downstream supervised tasks. We conclude that development and evaluation of PFMs requires moving beyond accuracy-centric benchmarks toward explicit evaluation and optimisation of embedding stability and calibration under realistic acquisition variability.

A single stage active AC/DC interface able to emulate the optimal load impedance of a Resonant Piezoelectric Vibration Energy Harvester (RPVEH) is proposed. As theoretically shown, unlike an electronic interface that emulates an optimal load generator, an interface that emulates an optimal load impedance does not require adaptation to the acceleration of input vibrations. This allows the use of a very simple control, avoiding the implementation of Maximum Power Point Tracking (MPPT) algorithms that require lossy microcontrollers. Thus, the proposed interface is equipped with a simple analog controller allowing the RPVEH to work in its Maximum Power Point (MPP) in both steady-state and variable conditions of vibrations, without recurring to multivariable perturbative approaches, as it happens for the most of single stage AC/DC interfaces proposed in the literature. The absence of perturbative techniques allows a significant improvement of both stationary and dynamic performances. Experimental tests of a prototype of the proposed interface confirm the theoretical findings and the predicted behavior.

Efficient navigation in swarms often relies on the emergence of decentralized approaches that minimize traversal time or energy. Stigmergy, where agents modify a shared environment that then modifies their behavior, is a classic mechanism that can encode this strategy. We develop a theoretical framework for stigmergic transport by casting it as a stochastic optimal control problem: agents (collectively) lay and (individually) follow trails while minimizing expected traversal time. Simulations and analysis reveal two emergent behaviors: path straightening in homogeneous environments and path refraction at material interfaces, both consistent with experimental observations of insect trails. While reminiscent of Fermat's principle, our results show how local, noisy agent+field interactions can give rise to geodesic trajectories in heterogeneous environments, without centralized coordination or global knowledge, relying instead on an embodied slow fast dynamical mechanism.

We consider energy-efficient multi-user hybrid downlink beamforming (BF) and power allocation under imperfect channel state information (CSI) and probabilistic outage constraints. In this domain, classical optimization methods resort to computationally costly conic optimization problems. Meanwhile, generic deep network (DN) architectures lack interpretability and require large training data sets to generalize well. In this paper, we therefore propose a lightweight model-aided deep learning architecture based on a greedy selection algorithm for analog beam codewords. The architecture relies on an instance-adaptive augmentation of the signal model to estimate the impact of the CSI error. To learn the DN parameters, we derive a novel and efficient implicit representation of the nested constrained BF problem and prove sufficient conditions for the existence of the corresponding gradient. In the loss function, we utilize an annealing-based approximation of the outage compared to conventional quantile-based loss terms. This approximation adaptively anneals towards the exact probabilistic constraint depending on the current level of quality of service (QoS) violation. Simulations validate that the proposed DN can achieve the nominal outage level under CSI error due to channel estimation and channel compression, while allocating less power than benchmarks. Thereby, a single trained model generalizes to different numbers of users, QoS requirements and levels of CSI quality. We further show that the adaptive annealing-based loss function can accelerate the training and yield a better power-outage trade-off.

Low-altitude wireless networks (LAWNs) are expected to play a central role in future 6G infrastructures, yet uplink transmissions of uncrewed aerial vehicles (UAVs) remain vulnerable to eavesdropping due to their limited transmit power, constrained antenna resources, and highly exposed air-ground propagation conditions. To address this fundamental bottleneck, we propose a flexible-duplex cell-free (CF) architecture in which each distributed access point (AP) can dynamically operate either as a receive AP for UAV uplink collection or as a transmit AP that generates cooperative artificial noise (AN) for secrecy enhancement. Such AP-level duplex flexibility introduces an additional spatial degree of freedom that enables distributed and adaptive protection against wiretapping in LAWNs. Building upon this architecture, we formulate a max-min secrecy-rate problem that jointly optimizes AP mode selection, receive combining, and AN covariance design. This tightly coupled and nonconvex optimization is tackled by first deriving the optimal receive combiners in closed form, followed by developing a penalty dual decomposition (PDD) algorithm with guaranteed convergence to a stationary solution. To further reduce computational burden, we propose a low-complexity sequential scheme that determines AP modes via a heuristic metric and then updates the AN covariance matrices through closed-form iterations embedded in the PDD framework. Simulation results show that the proposed flexible-duplex architecture yields substantial secrecy-rate gains over CF systems with fixed AP roles. The joint optimization method attains the highest secrecy performance, while the low-complexity approach achieves over 90% of the optimal performance with an order-of-magnitude lower computational complexity, offering a practical solution for secure uplink communications in LAWNs.

Neural networks commonly employ the McCulloch-Pitts neuron model, which is a linear model followed by a point-wise non-linear activation. Various researchers have already advanced inherently non-linear neuron models, such as quadratic neurons, generalized operational neurons, generative neurons, and super neurons, which offer stronger non-linearity compared to point-wise activation functions. In this paper, we introduce a novel and better non-linear neuron model called Padé neurons (Paons), inspired by Padé approximants. Paons offer several advantages, such as diversity of non-linearity, since each Paon learns a different non-linear function of its inputs, and layer efficiency, since Paons provide stronger non-linearity in much fewer layers compared to piecewise linear approximation. Furthermore, Paons include all previously proposed neuron models as special cases, thus any neuron model in any network can be replaced by Paons. We note that there has been a proposal to employ the Padé approximation as a generalized point-wise activation function, which is fundamentally different from our model. To validate the efficacy of Paons, in our experiments, we replace classic neurons in some well-known neural image super-resolution, compression, and classification models based on the ResNet architecture with Paons. Our comprehensive experimental results and analyses demonstrate that neural models built by Paons provide better or equal performance than their classic counterparts with a smaller number of layers. The PyTorch implementation code for Paon is open-sourced at https://github.com/onur-keles/Paon.

Allowing less capable devices to offload computational tasks to more powerful devices or servers enables the development of new applications that may not run correctly on the device itself. Deciding where and why to run each of those applications is a complex task. Therefore, different approaches have been adopted to make offloading decisions. In this work, we propose a decentralized Deep Reinforcement Learning (DRL) agent to address the selection of computing locations. Unlike most existing work, we analyze it in a real testbed composed of various edge devices running the agent to determine where to execute each task. These devices are connected to a Multi-Access Edge Computing (MEC) server and a Cloud server through 5G communications. We evaluate not only the agent's performance in meeting task requirements but also the implications of running this type of agent locally, assessing the trade-offs of training locally versus remotely in terms of latency and energy consumption.

In large-scale Bayesian inverse problems, it is often necessary to apply approximate forward models to reduce the cost of forward model evaluations, while controlling approximation quality. In the context of Bayesian inverse problems with linear forward models, Gaussian priors, and Gaussian noise, we use perturbation theory for inverses to bound the error in the approximate posterior mean and posterior covariance resulting from a linear approximate forward model. We then focus on the smoothing problem of inferring the initial condition of linear time-invariant dynamical systems, using finitely many partial state observations. For such problems, and for a specific model order reduction method based on balanced truncation, we show that the impulse response of a certain prior-driven system is closely related to the prior-preconditioned Hessian of the inverse problem. This reveals a novel connection between systems theory and inverse problems. We exploit this connection to prove the first a priori error bounds for system-theoretic model order reduction methods applied to smoothing problems. The bounds control the approximation error of the posterior mean and covariance in terms of the truncated Hankel singular values of the underlying system.

ASVspoof 5 is the fifth edition in a series of challenges which promote the study of speech spoofing and deepfake detection solutions. A significant change from previous challenge editions is a new crowdsourced database collected from a substantially greater number of speakers under diverse recording conditions, and a mix of cutting-edge and legacy generative speech technology. With the new database described elsewhere, we provide in this paper an overview of the ASVspoof 5 challenge results for the submissions of 53 participating teams. While many solutions perform well, performance degrades under adversarial attacks and the application of neural encoding/compression schemes. Together with a review of post-challenge results, we also report a study of calibration in addition to other principal challenges and outline a road-map for the future of ASVspoof.

Image deblurring is a challenging problem in imaging due to its highly ill-posed nature. Deep learning models have shown great success in tackling this problem but the quest for the best image quality has brought their computational complexity up, making them impractical on anything but powerful servers. Meanwhile, recent works have shown that mobile Lidars can provide complementary information in the form of depth maps that enhance deblurring quality. In this paper, we introduce a novel low-complexity neural network for depth-guided image deblurring. We show that the use of the wavelet transform to separate structural details and reduce spatial redundancy as well as efficient feature conditioning on the depth information are essential ingredients in developing a low-complexity model. Experimental results show competitive image quality against recent state-of-the-art models while reducing complexity by up to two orders of magnitude.

Many nonlinear optimal control and optimization problems involve constraints that combine continuous dynamics with discrete logic conditions. Standard approaches typically rely on mixed-integer programming, which introduces scalability challenges and requires specialized solvers. This paper presents an exact reformulation of broad classes of logical constraints as binary-variable-free expressions whose differentiability properties coincide with those of the underlying predicates, enabling their direct integration into nonlinear programming models. Our approach rewrites arbitrary logical propositions into conjunctive normal form, converts them into equivalent max–min constraints, and applies a smoothing procedure that preserves the exact feasible set. The method is evaluated on two benchmark problems, a quadrotor trajectory optimization with obstacle avoidance and a hybrid two-tank system with temporal logic constraints, and is shown to obtain optimal solutions more consistently and efficiently than existing binary variable elimination techniques.

In this paper, we introduce a novel pipeline for predicting chemotherapy response in pediatric brain tumors that are not amenable to complete surgical resection, using pre-treatment magnetic resonance imaging combined with clinical information. Our method integrates a state-of-the-art pediatric brain tumor segmentation framework with radiomic feature extraction and clinical data through an ensemble of a Swin UNETR encoder and XGBoost classifier. The segmentation model delineates four tumor subregions enhancing tumor, non-enhancing tumor, cystic component and edema which are used to extract imaging biomarkers and generate predictive features. The Swin UNETR network classifies the response to treatment directly from these segmented MRI scans, while XGBoost predicts response using radiomics and clinical variables including legal sex, ethnicity, race, age at event (in days), molecular subtype, tumor locations, initial surgery status, metastatic status, metastasis location, chemotherapy type, protocol name and chemotherapy agents. The ensemble output provides a non-invasive estimate of chemotherapy response in this historically challenging population characterized by lower progression-free survival. Among compared approaches, our Swin-Ensemble achieved the best performance (precision for non effective cases=0.68, recall for non effective cases=0.85, precision for chemotherapy effective cases=0.64 and overall accuracy=0.69), outperforming Mamba-FeatureFuse, Swin UNETR encoder, and Swin-FeatureFuse models. Our findings suggest that this ensemble framework represents a promising step toward personalized therapy response prediction for pediatric low-grade glioma patients in need of chemotherapy treatment who are not suitable for complete surgical resection, a population with significantly lower progression free survival and for whom chemotherapy remains the primary treatment option.

Sampling-based model predictive control (MPC) is effective for nonlinear systems but often produces non-smooth control inputs due to random sampling. To address this issue, we extend the model predictive path integral (MPPI) framework with deterministic sampling and improvements from cross-entropy method (CEM)–MPC, such as iterative optimization, proposing deterministic sampling MPPI (dsMPPI). This combination leverages the exponential weighting of MPPI alongside the efficiency of deterministic samples. Experiments demonstrate that dsMPPI achieves smoother trajectories compared to state-of-the-art methods.

Deep learning has achieved significant advancements in medical image segmentation. Currently, obtaining accurate segmentation outcomes is critically reliant on large-scale datasets with high-quality annotations. However, noisy annotations are frequently encountered owing to the complex morphological structures of organs in medical images and variations among different annotators, which can substantially limit the efficacy of segmentation models. Motivated by the fact that medical imaging annotator can correct labeling errors during segmentation based on prior knowledge, we propose an end-to-end Staged Voxel-Level Deep Reinforcement Learning (SVL-DRL) framework for robust medical image segmentation under noisy annotations. This framework employs a dynamic iterative update strategy to automatically mitigate the impact of erroneous labels without requiring manual intervention. The key advancements of SVL-DRL over existing works include: i) formulating noisy annotations as a voxel-dependent problem and addressing it through a novel staged reinforcement learning framework which guarantees robust model convergence; ii) incorporating a voxel-level asynchronous advantage actor-critic (vA3C) module that conceptualizes each voxel as an autonomous agent, which allows each agent to dynamically refine its own state representation during training, thereby directly mitigating the influence of erroneous labels; iii) designing a novel action space for the agents, along with a composite reward function that strategically combines the Dice value and a spatial continuity metric to significantly boost segmentation accuracy while maintain semantic integrity. Experiments on three public medical image datasets demonstrates State-of-The-Art (SoTA) performance under various experimental settings, with an average improvement of over 3% in both Dice and IoU scores.

Small-scale wind turbine research in resource-constrained academic settings frequently produces unreliable or unpublishable datasets due to ad-hoc instrumentation, inadequate time synchronization, storage failures, and weak data governance. This paper presents a systematic data acquisition (DAQ) methodology and ESP32-based reference implementation design for field characterization of small wind turbines (100~W–5~kW), emphasizing tropical/coastal deployment constraints typical of Low- and Middle-Income Countries (LMIC). We integrate (i)~a student-adapted V-model with requirements traceability, (ii)~hardware selection strategies for high-humidity and salt-spray environments, (iii)~an embedded firmware architecture featuring interrupt-driven rotor speed measurement, state-machine fault handling, and NTP-based time synchronization, (iv)~a local-first hybrid storage design combining SD-card persistence with optional MQTT cloud telemetry, and (v)~a data-management workflow adapting CRISP-DM and FAIR principles with explicit quality dimensions and publication templates. A detailed helical vertical-axis wind turbine (VAWT) design scenario for coastal Sri Lanka illustrates the complete methodology, targeting $>90\%$ data completeness over six-month campaigns. The methodology is accompanied by open-source firmware, hardware templates, and data-publication workflow artifacts released via GitHub and Zenodo.

Reconfigurable intelligent surfaces (RISs) enable programmable control of the wireless propagation environment and are key enablers for future networks. Beyond-diagonal RIS (BD-RIS) architectures enhance conventional RIS by interconnecting elements through tunable impedance components, offering greater flexibility with higher circuit complexity. However, excessive interconnections between BD-RIS elements require multi-layer printed circuit board (PCB) designs, increasing fabrication difficulty. In this letter, we use graph theory to characterize the BD-RIS architectures that can be realized on double-layer PCBs, denoted as planar-connected RISs. Among the possible planar-connected RISs, we identify the ones with the most degrees of freedom, expected to achieve the best performance under practical constraints.

Assigning individuals with hearing impairment to auditory profiles can support a better understanding of the causes and consequences of hearing loss and facilitate profile-based hearing-aid fitting. However, the factors influencing auditory profile generation remain insufficiently understood, and existing profiling frameworks have rarely been compared systematically. This study therefore investigated the impact of two key factors - the clustering method and the number of profiles - on auditory profile generation. In addition, eight established auditory profiling frameworks were systematically reviewed and compared using intrinsic statistical measures and manifold learning techniques. Frameworks were evaluated with respect to internal consistency (i.e., grouping similar individuals) and cluster separation (i.e., clear differentiation between groups). To ensure comparability, all analyses were conducted on a common open-access dataset, the extended Oldenburg Hearing Health Record (OHHR), comprising 1,127 participants (mean age = 67.2 years, SD = 12.0). Results showed that both the clustering method and the chosen number of profiles substantially influenced the resulting auditory profiles. Among purely audiogram-based approaches, the Bisgaard auditory profiles demonstrated the strongest clustering performance, whereas audiometric phenotypes performed worst. Among frameworks incorporating supra-threshold information in addition to the audiogram, the Hearing4All auditory profiles were advantageous, combining a near-optimal number of profile classes (N = 13) with high clustering quality, as indicated by a low Davies-Bouldin index. In conclusion, manifold learning and intrinsic measures enable systematic comparison of auditory profiling frameworks and identify the Hearing4All auditory profile as a promising approach for future research.

Although machining chatter can be suppressed by the choice of stable cutting parameters through means of stability lobe diagram (SLD), surface roughness still remains due to the forced vibration, which limits surface quality, especially in the surface finish. Better cutting parameters can be achieved considering surface location error (SLE) together with SLD. This paper proposes an innovative modeling framework of the machining dynamic system that enables efficient computation of the chatter stability and SLE. The framework mainly embodies two techniques, namely semi-discretization method (SDM) and lifting method. The machining dynamics system is mathematically expressed as an angle-varying delay differential equation (DDE). The SDM approximates the angle-varying and delayed terms to ordinary terms using zero-phase interpolations and governs the discrete angle-varying dynamics system. Then, the system is merged over the tooth passing angle using the lifted approach to establish an explicit dynamic system in the compact state-space form. Based on the compact state-space model, the chatter stability and SLE prediction are easily and efficiently conducted. Simulation results show the improved efficiency of the proposed method over other well-known methods.

Self-Supervised Learning (SSL) has emerged as a key technique in machine learning, tackling challenges such as limited labeled data, high annotation costs, and variable wireless channel conditions. It is essential for developing Channel Foundation Models (CFMs), which extract latent features from channel state information (CSI) and adapt to different wireless settings. Yet, existing CFMs have notable drawbacks: heavy reliance on scenario-specific data hinders generalization, they focus on single/dual tasks, and lack zero-shot learning ability. In this paper, we propose CSI-MAE, a generalized CFM leveraging masked autoencoder for cross-scenario generalization. Trained on 3GPP channel model datasets, it integrates sensing and communication via CSI perception and generation, proven effective across diverse tasks. A lightweight decoder finetuning strategy cuts training costs while maintaining competitive performance. Under this approach, CSI-MAE matches or surpasses supervised models. With full-parameter finetuning, it achieves the state-of-the-art performance. Its exceptional zero-shot transferability also rivals supervised techniques in cross-scenario applications, driving wireless communication innovation.

While multimodal mobility systems have the potential to bring many benefits to travelers, drivers, the environment, and traffic congestion, such systems typically involve multiple non-cooperative decision-makers who may selfishly optimize their own objectives without considering the overall system benefits. This paper aims to investigate market-based interactions of travelers and ride-sourcing drivers in the context of multimodal mobility systems. We propose a unified mathematical modeling framework to capture the decentralized travelers and drivers' decision-making process and balance the network's demand and supply by equilibrium pricing. Such a model allows analyses of the impact of decentralized decision-making on multimodal mobility efficiencies. The proposed formulation can be further convexified to efficiently compute the equilibrium ride-sourcing prices. We conduct numerical experiments on different settings of transportation networks to gain policy insights. We find that travelers prefer ride-sourcing and multimodal transportation more than the driving option when they are more sensitive to prices. We also find that travelers may need to be subsidized to use multimodal transportation when there is fewer transit hubs in the network or, ride-sourcing drivers become too sensitive to the prices. However, we find that more transit hubs in the network increases the total empty VMT of ride-sourcing drivers by increasing the total relocation time. The proposed model can be used by policymakers and platform operators to design pricing and subsidy schemes that align individual decision-making with system-level efficiency and evaluate the trade-offs between accessibility and environmental impacts in multimodal transportation networks.

This work addresses the output consensus problem of constrained heterogeneous multi-agent systems under a switching network with potential communication delays, where outputs are periodic and characterized by an exosystem. Since periodic references have more complex dynamics, it is more challenging to track periodic references and achieve consensus on them. In this paper, a model predictive control method incorporating an artificial reference and a modified cost function is proposed to track periodic references, which maintains recursive feasibility even when references switch. Moreover, consensus protocols are proposed to achieve consensus on periodic references in different scenarios, in which global information such as the set of globally admissible references and the global time index are not involved. Theoretical analysis proves that constrained output consensus is asymptotically achieved with the proposed algorithm as the references of each agent converge and agents track their references while maintaining constraint satisfaction. Finally, numerical examples are provided to verify the effectiveness of the proposed algorithm.

In this letter, we study an efficient multi-beam training method for multiuser millimeter-wave communication systems. Unlike the conventional single-beam training method that relies on exhaustive search, multi-beam training design faces a key challenge in balancing the trade-off between beam training overhead and success beam-identification rate, exacerbated by severe inter-beam interference. To tackle this challenge, we propose a new two-stage multi-beam training method with two distinct multi-beam patterns to enable fast and accurate user angle identification. Specifically, in the first stage, the antenna array is divided into sparse subarrays to generate multiple beams (with high array gains), for identifying candidate user angles. In the second stage, the array is redivided into dense subarrays to generate flexibly steered wide beams, for which a cross-validation method is employed to effectively resolve the remaining angular ambiguity in the first stage. Last, numerical results demonstrate that the proposed method significantly improves the success beam-identification rate compared to existing multi-beam training methods, while retaining or even reducing the required beam training overhead.

In the context of joint communication and sensing JC&S, the challenge of obtaining accurate parameter estimates is of interest. Parameter estimates, such as the AoA can be utilized for solving the initial access problem, interference mitigation, localization of users or monitoring of the environment and synchronization of MIMO systems. Recently, TTD systems have gained attention for fast beam training during initial access and mitigation of beam squinting. This work derives the CRB for angle estimates in typical TTD systems. Properties of the CRB and the Fisher information are investigated and numerically evaluated. Finally, methods for angle estimation such as ML and established estimators are utilized to solve the angle estimation problem using a uniform linear array.

Active Alignment (AA) is a key technology for the large-scale automated assembly of high-precision optical systems. Compared with labor-intensive per-model on-device calibration, a digital-twin pipeline built on optical simulation offers a substantial advantage in generating large-scale labeled data. However, complex imaging conditions induce a domain gap between simulation and real-world images, limiting the generalization of simulation-trained models. To address this, we propose augmenting a simulation baseline with minimal unlabeled real-world images captured at random misalignment positions, mitigating the gap from a domain adaptation perspective. We introduce Domain Adaptive Active Alignment (DA3), which utilizes an autoregressive domain transformation generator and an adversarial-based feature alignment strategy to distill real-world domain information via self-supervised learning. This enables the extraction of domain-invariant image degradation features to facilitate robust misalignment prediction. Experiments on two lens types reveal that DA3 improves accuracy by 46% over a purely simulation pipeline. Notably, it approaches the performance achieved with precisely labeled real-world data collected on 3 lens samples, while reducing on-device data collection time by 98.7%. The results demonstrate that domain adaptation effectively endows simulation-trained models with robust real-world performance, validating the digital-twin pipeline as a practical solution to significantly enhance the efficiency of large-scale optical assembly.

Latent Thermal Energy Storages (LTES) can store thermal energy in a narrow temperature range. Therefore, they are favorable for integration into Rankine-based Carnot Batteries. For the design of such systems, simulations based on accurate models are desirable. However, physical phenomena such as natural convection in LTES units cannot be modeled directly in transient system models. Simplified models are required. Therefore, the objective of this work is to derive simplified LTES unit models for use in system models. In transient simulations the state of charge of the LTES influences its temperature profile. The temperature profile depends on the geometry of the LTES unit. Therefore, the geometry must be considered to model the transient behavior of an LTES unit. The LTES unit under investigation has a shell and tube heat exchanger structure. The phase change material (PCM) is located between the hexagonal fins and in the space between the finned tubes. Aluminum fins are used. They have a high thermal conductivity and thus compensate for the low thermal conductivity of the sodium nitrate used as PCM. The interaction between fins and PCM is complex. Therefore, a numerical approach can be used to gain insight into the behavior of the LTES unit. To transfer the results of a complex model to a simplified model where fins and PCM are not considered individually, the effective thermal conductivity of a single finned tube can be used to approximate the performance of the LTES unit. In this study, a model of a section with a single finned tube is developed using the COMSOL software. The effective thermal conductivity of the system is determined by varying the effective thermal conductivity in a simplified model and comparing the results with reference cases based on a complex modeling approach. The results can serve as model input for simplified system models of Carnot Batteries, among others.

Multi-speaker automatic speech recognition (MASR) aims to predict ”who spoke when and what” from multi-speaker speech, a key technology for multi-party dialogue understanding. However, most existing approaches decouple temporal modeling and speaker modeling when addressing ”when” and ”who”: some inject speaker cues before encoding (e.g., speaker masking), which can cause irreversible information loss; others fuse identity by mixing speaker posteriors after encoding, which may entangle acoustic content with speaker identity. This separation is brittle under rapid turn-taking and overlapping speech, often leading to degraded performance. To address these limitations, we propose TellWhisper, a unified framework that jointly models speaker identity and temporal within the speech encoder. Specifically, we design TS-RoPE, a time-speaker rotary positional encoding: time coordinates are derived from frame indices, while speaker coordinates are derived from speaker activity and pause cues. By applying region-specific rotation angles, the model explicitly captures per-speaker continuity, speaker-turn transitions, and state dynamics, enabling the attention mechanism to simultaneously attend to ”when” and ”who”. Moreover, to estimate frame-level speaker activity, we develop Hyper-SD, which casts speaker classification in hyperbolic space to enhance inter-class separation and refine speaker-activity estimates. Extensive experiments demonstrate the effectiveness of the proposed approach.

Decarbonization of isolated or off-grid energy systems through phase-in of large shares of intermittent solar or wind generation requires co-installation of energy storage or continued use of existing fossil dispatchable power sources to balance supply and demand. The effective CO2 emission reduction depends on the relative capacity of the energy storage and renewable sources, the stochasticity of the renewable generation, and the control of the isolated energy system. While the operation of the energy storage and dispatchable sources impacts the optimal sizing of the system, it is challenging to account for the effect of finite-horizon optimal control at the stage of system sizing. In this work, we present a flexible and computationally efficient sizing framework for energy storage and renewable capacity in isolated energy systems, accounting for uncertainty in the renewable generation and the optimal control. We implement an imitation learning approach to stochastic neural model predictive control (MPC) which allows us to relate the battery storage and wind peak capacities to the emissions reduction and investment costs while accounting for finite horizon, optimal control without solving an infeasible number of optimization problems. We evaluate the proposed sizing framework on a case study of an offshore energy system with a gas turbine, a wind farm and a battery energy storage system (BESS). In this case, we find a nonlinear, nontrivial relationship between the investment costs and the reduction in gas usage for different wind and BESS capacities.