Papers

A control strategy for the electrical current in GMAW processes is proposed. The control is in closed-loop, designed by formal methods, based on a mathematical model of the electrical behavior of the GMAW process, and implemented in C+ language in a microcontroller. The model consists of a switched equivalent electrical circuit whose parameters are obtained in a data-driven manner. The strategy is tested in numerous experiments with both manual and robot welding, showing improvements in the overall welding process.

CAPTCHAs are widely used by websites to block bots and spam by presenting challenges that are easy for humans but difficult for automated programs to solve. To improve accessibility, audio CAPTCHAs are designed to complement visual ones. However, the robustness of audio CAPTCHAs against advanced Large Audio Language Models (LALMs) and Automatic Speech Recognition (ASR) models remains unclear. In this paper, we introduce AI-CAPTCHA, a unified framework that offers (i) an evaluation framework, ACEval, which includes advanced LALM- and ASR-based solvers, and (ii) a novel audio CAPTCHA approach, IllusionAudio, leveraging audio illusions. Through extensive evaluations of seven widely deployed audio CAPTCHAs, we show that most existing methods can be solved with high success rates by advanced LALMs and ASR models, exposing critical security weaknesses. To address these vulnerabilities, we design a new audio CAPTCHA approach, IllusionAudio, which exploits perceptual illusion cues rooted in human auditory mechanisms. Extensive experiments demonstrate that our method defeats all tested LALM- and ASR-based attacks while achieving a 100% human pass rate, significantly outperforming existing audio CAPTCHA methods.

Internet of underwater things (IoUT) is increasingly gathering attention with the aim of monitoring sea life and deep ocean environment, underwater surveillance as well as maintenance of underwater installments. However, conventional IoUT devices, reliant on battery power, face limitations in lifespan and pose environmental hazards upon disposal. This paper introduces a sustainable approach for simultaneous information uplink from the IoUT devices and acoustic energy transfer (AET) to the devices via an autonomous underwater vehicle (AUV), potentially enabling them to operate indefinitely. To tackle the time-sensitivity, we adopt age of information (AoI), and Jain's fairness index. We develop two deep-reinforcement learning (DRL) algorithms, offering a high-complexity, high-performance frequency division duplex (FDD) solution and a low-complexity, medium-performance time division duplex (TDD) approach. The results elucidate that the proposed FDD and TDD solutions significantly reduce the average AoI and boost the harvested energy as well as data collection fairness compared to baseline approaches.

Roll-to-roll slot die coating is a widely used industrial manufacturing technique applied in a diverse range of applications such as the production of lithium-ion batteries, solar cells and optical films. The efficiency of roll-to-roll slot die coating depends on the precise control of various input parameters such as pump rate, substrate velocity and coating gap. However, these inputs are sensitive to disturbances in process conditions, leading to inconsistencies in the various characteristics of the produced film. To address this challenge, a \gls{DO} is utilized for detecting disturbances, which may occur in the same or different channels as the control signal within the system. A generalized compensator is then implemented to mitigate the impact of these disturbances on the output, thereby enhancing uncertainty suppression. Additionally, integrating the disturbance rejection system with an output tracking controller enables the coating system to maintain the desired thickness under varying input conditions and disturbances. The effectiveness of this approach is then validated using a test rig equipped with a camera system, which facilitates the development of a data-driven model of the dynamic process, represented by state-space equations. The simulation results were demonstrated to showcase the effectiveness of the DOBOTC system, which provides a resilient solution for the output tracking issue in a data-driven model with generalized disturbances.

This paper introduces a scheme for drone surveillance coordinated with the fifth generation (5G) synchronization signal block (SSB) cell-search procedure to simultaneously detect low-altitude drones within a volumetric surveillance grid. Herein, we consider a multistatic configuration where multiple access points (APs) collaboratively illuminate the volume while independently transmitting SSB broadcast signals. Both tasks are performed through a beam sweeping. In the proposed scheme, coordinated APs send sensing beams toward a grid of voxels within the volumetric surveillance region simultaneously with the 5G SSB burst. To prevent interference between communication and sensing signals, we propose a precoder design that guarantees orthogonality of the sensing beam and the SSB in order to maximize the sensing signal-to-interference-plus-noise ratio (SINR) while ensuring a specified SINR for users, as well as minimizing the impact of the direct link. The results demonstrate that the proposed precoder outperforms the non-coordinated precoder and is minimally affected by variations in drone altitude.

Anomalous sound detection (ASD) typically involves self-supervised proxy tasks to learn feature representations from normal sound data, owing to the scarcity of anomalous samples. In ASD research, proxy tasks such as AutoEncoders operate under the explicit assumption that models trained on normal data will increase the reconstruction errors related to anomalies. A natural extension suggests that improved proxy task performance should improve ASD capability; however, this relationship has received little systematic attention. This study addresses this research gap by quantitatively analyzing the relationship between proxy task metrics and ASD performance across five configurations, namely, AutoEncoders, classification, source separation, contrastive learning, and pre-trained models. We evaluate the learned representations using linear probe (linear separability) and Mahalanobis distance (distributional compactness). Our experiments reveal that strong proxy performance does not necessarily improve anomalous sound detection performance. Specifically, classification tasks experience performance saturation owing to insufficient task difficulty, whereas contrastive learning fails to learn meaningful features owing to limited data diversity. Notably, source separation is the only task demonstrating a strong positive correlation, such that improved separation consistently improves anomaly detection. Based on these findings, we highlight the critical importance of task difficulty and objective alignment. Finally, we propose a three-stage alignment verification protocol to guide the design of highly effective proxy tasks for ASD systems.

Distracted driving is a major cause of traffic collisions, calling for robust and scalable detection methods. Vision-language models (VLMs) enable strong zero-shot image classification, but existing VLM-based distracted driver detectors often underperform in real-world conditions. We identify subject-specific appearance variations (e.g., clothing, age, and gender) as a key bottleneck: VLMs entangle these factors with behavior cues, leading to decisions driven by who the driver is rather than what the driver is doing. To address this, we propose a subject decoupling framework that extracts a driver appearance embedding and removes its influence from the image embedding prior to zero-shot classification, thereby emphasizing distraction-relevant evidence. We further orthogonalize text embeddings via metric projection onto Stiefel manifold to improve separability while staying close to the original semantics. Experiments demonstrate consistent gains over prior baselines, indicating the promise of our approach for practical road-safety applications. Code is available at https://github.com/mtakamichi/ZVL-DDD

Learning-based wireless sensing has made rapid progress, yet the field still lacks a unified and reproducible experimental foundation. Unlike computer vision, wireless sensing relies on hardware-dependent channel measurements whose representations, preprocessing pipelines, and evaluation protocols vary significantly across devices and datasets, hindering fair comparison and reproducibility. This paper proposes the Sensing Data Protocol (SDP), a protocol-level abstraction and unified benchmark for scalable wireless sensing. SDP acts as a standardization layer that decouples learning tasks from hardware heterogeneity. To this end, SDP enforces deterministic physical-layer sanitization, canonical tensor construction, and standardized training and evaluation procedures, decoupling learning performance from hardware-specific artifacts. Rather than introducing task-specific models, SDP establishes a principled protocol foundation for fair evaluation across diverse sensing tasks and platforms. Extensive experiments demonstrate that SDP achieves competitive accuracy while substantially improving stability, reducing inter-seed performance variance by orders of magnitude on complex activity recognition tasks. A real-world experiment using commercial off-the-shelf Wi-Fi hardware further illustrating the protocol's interoperability across heterogeneous hardware. By providing a unified protocol and benchmark, SDP enables reproducible and comparable wireless sensing research and supports the transition from ad hoc experimentation toward reliable engineering practice.

A key challenge with large battery systems is heterogeneous currents and temperatures in modules with parallel-connected cells. Although extreme currents and temperatures are detrimental to the performance and lifetime of battery cells, there is not a consensus on the scale of typical imbalances within grid storage modules. Here, we quantify these imbalances through simulations and experiments on an industrially representative grid storage battery module consisting of prismatic lithium iron phosphate cells, elucidating the evolution of current and temperature imbalances and their dependence on individual cell and module parameter variations. Using a sensitivity analysis, we find that varying contact resistances and cell resistances contribute strongly to temperature differences between cells, from which we define safety thresholds on cell-to-cell variability. Finally, we investigate how these thresholds change for different applications, to outline a set of robustness metrics that show how cycling at lower C-rates and narrower SOC ranges can mitigate failures.

Autoregressive speech synthesis often adopts a left-to-right order, yet generation order is a modelling choice. We investigate decoding order through masked diffusion framework, which progressively unmasks positions and allows arbitrary decoding orders during training and inference. By interpolating between identity and random permutations, we show that randomness in decoding order affects speech quality. We further compare fixed strategies, such as l2r and r2l with adaptive ones, such as Top-$K$, finding that fixed-order decoding, including the dominating left-to-right approach, is suboptimal, while adaptive decoding yields better performance. Finally, since masked diffusion requires discrete inputs, we quantise acoustic representations and find that even 1-bit quantisation can support reasonably high-quality speech.

Residential users in demand response programs must balance electricity costs and user dissatisfaction under real-time pricing. This study proposes a multiobjective model predictive control approach for home energy management systems with battery storage, aiming to minimize both objectives while mitigating uncertainties. Laguerre functions parameterize control signals, transforming the optimization problem into one with linear inequalities for efficient exploration. A constrained multiobjective evolutionary algorithm, incorporating convex sampler-based crossover and mutation, is developed to ensure feasible solutions. Simulations show that the proposed method outperforms existing approaches, limiting cost increases to 0.52% under uncertainties, compared to at least 2.3% with other methods.

Predicting pathloss by considering the physical environment is crucial for effective wireless network planning. Traditional methods, such as ray tracing and model-based approaches, often face challenges due to high computational complexity and discrepancies between models and real-world environments. In contrast, deep learning has emerged as a promising alternative, offering accurate path loss predictions with reduced computational complexity. In our research, we introduce a ResNet-based model designed to enhance path loss prediction. We employ innovative techniques to capture key features of the environment by generating transmission (Tx) and reception (Rx) depth maps, as well as a distance map from the geographic data. Recognizing the significant attenuation caused by signal reflection and diffraction, particularly at high frequencies, we have developed a weighting map that emphasizes the areas adjacent to the direct path between Tx and Rx for path loss prediction. {Extensive simulations demonstrate that our model outperforms PPNet, RPNet, and Vision Transformer (ViT) by 1.2-3.0 dB using dataset of ITU challenge 2024 and ICASSP 2023. In addition, the floating point operations (FLOPs) of the proposed model is 60% less than those of benchmarks.} Additionally, ablation studies confirm that the inclusion of the weighting map significantly enhances prediction performance.

This work presents Bio-RV, a compact and resource-efficient RISC-V processor intended for biomedical control applications, such as accelerator-based biomedical SoCs and implantable pacemaker systems. The proposed Bio-RV is a multi-cycle RV32I core that provides explicit execution control and external instruction loading with capabilities that enable controlled firmware deployment, ASIC bring-up, and post-silicon testing. In addition to coordinating accelerator configuration and data transmission in heterogeneous systems, Bio-RV is designed to function as a lightweight host controller, handling interfaces with pacing, sensing, electrogram (EGM), telemetry, and battery management modules. With 708 LUTs and 235 flip-flops on FPGA prototypes, Bio-RV, implemented in a 180 nm CMOS technology, operate at 50 MHz and feature a compact hardware footprint. According to post-layout results, the proposed architectural decisions align with minimal energy use. Ultimately, Bio-RV prioritises deterministic execution, minimal hardware complexity, and integration flexibility over peak computing speed to meet the demands of ultra-low-power, safety-critical biomedical systems.

Benefiting from the rapid advancements in large language models (LLMs), human-drone interaction has reached unprecedented opportunities. In this paper, we propose a method that integrates a fine-tuned CodeT5 model with the Unreal Engine-based AirSim drone simulator to efficiently execute multi-task operations using natural language commands. This approach enables users to interact with simulated drones through prompts or command descriptions, allowing them to easily access and control the drone's status, significantly lowering the operational threshold. In the AirSim simulator, we can flexibly construct visually realistic dynamic environments to simulate drone applications in complex scenarios. By combining a large dataset of (natural language, program code) command-execution pairs generated by ChatGPT with developer-written drone code as training data, we fine-tune the CodeT5 to achieve automated translation from natural language to executable code for drone tasks. Experimental results demonstrate that the proposed method exhibits superior task execution efficiency and command understanding capabilities in simulated environments. In the future, we plan to extend the model functionality in a modular manner, enhancing its adaptability to complex scenarios and driving the application of drone technologies in real-world environments.

This paper develops a data-driven time-limited h2 model reduction method for discrete-time linear time-invariant systems. Specifically, we formulate and solve a regularized time-limited h2 model reduction problem using only noisy impulse response data. Furthermore, we show that the objective function and its gradient can be represented using only noisy impulse response data. Numerical experiments using SLICOT benchmarks demonstrate that the proposed regularized method achieves lower relative time-limited h2 errors than the tested alternatives and is effective in situations where the unregularized method may deteriorate under noise.

Even though the ocean covers the majority of the planet's surface, it remains the least explored ecosystem. As light and radio waves do not propagate through water, underwater acoustics is the main choice for various ocean applications ranging from marine biology to pollution monitoring. Increasing levels of anthropogenic noise from ships contribute significantly to underwater sound pollution, posing risks to marine ecosystems. This makes monitoring crucial to understand and quantify the impact of the ship radiated noise. Passive Acoustic Monitoring (PAM) systems are widely deployed for this purpose, generating years of underwater recordings across diverse soundscapes. Manual analysis of such large-scale data is impractical, motivating the need for automated approaches based on machine learning. Recent advances in automatic Underwater Acoustic Target Recognition (UATR) have largely relied on supervised learning, which is constrained by the scarcity of labeled data. Transfer Learning (TL) offers a promising alternative to mitigate this limitation. In this work, we conduct the first empirical comparative study of transfer learning for UATR, evaluating multiple pretrained audio models originating from diverse audio domains. The pretrained model weights are frozen, and the resulting embeddings are analyzed through classification, clustering, and similarity-based evaluations. The analysis shows that the geometrical structure of the embedding space is largely dominated by recording-specific characteristics. However, a simple linear probe can effectively suppress this recording-specific information and isolate ship-type features from these embeddings. As a result, linear probing enables effective automatic UATR using pretrained audio models at low computational cost, significantly reducing the need for a large amounts of high-quality labeled ship recordings.

Movable antenna (MA) introduces a new degree of freedom for future wireless communication systems by enabling the adaptive adjustment of antenna positions. Its large-range movement renders wireless channels transmission into the near-field region, which brings new performance enhancement for integrated sensing and communication (ISAC). This paper proposes a novel multi-stage design framework for broadband near-field ISAC assisted by MA. The framework first divides the MA movement area into multiple subregions, and employs the Newtonized orthogonal matching pursuit algorithm (NOMP) to achieve high-precision angle estimation in each subregion. Subsequently, a method called near-field localization via subregion ray clustering (LSRC) is proposed for identifying the positions of scatterers. This method finds the coordinates of each scatterer by jointly processing the angle estimates across all subregions. Finally, according to the estimated locations of the scatterers, the near-field channel estimation (CE) is refined for improving communication performance. Simulation results demonstrate that the proposed scheme can significantly enhance MA sensing accuracy and CE, providing an efficient solution for MA-aided near-field ISAC.

In the 21st century, transitioning to renewable energy sources is imperative, with fossil fuel reserves depleting rapidly and recognizing critical environmental issues such as climate change, air pollution, water pollution, and habitat destruction. Embracing renewable energy is not only an environmental necessity but also a strategic move with multiple benefits. By shifting to renewable energy sources and supporting their production through the acquisition of renewable energy certificates, we foster innovation and drive economic growth in the renewable energy sector. This, in turn, reduces greenhouse gas emissions, aligning with global efforts to mitigate climate change. Additionally, renewable energy certificates ensure compliance with regulations that mandate the use of renewable energy, enhancing legal adherence while promoting transparency and trust in energy sourcing. To monitor the uptake of renewable energy, governments have implemented Renewable Energy Certificates (RECs) as a tracking mechanism for the production and consumption of renewable energy. However, there are two main challenges to the existing REC schema: 1) The RECs have not been globally adopted due to inconsistent design; 2) The consumer privacy has not been well incorporated in the design of blockchain. In this study, we investigate the trading of RECs between suppliers and consumers using the directed acyclic graph (DAG) blockchain system and introduce a trading schema to help protect consumer information. Our results demonstrate lower transaction time by 41% and energy consumption by 65% compared to proof-of-stake.

Selecting a few available actuators to ensure the controllability of a linear system is a fundamental problem in control theory. Previous works either focus on optimal performance, simplifying the controllability issue, or make the system controllable under structural assumptions, such as in graphs or when the input matrix is a design parameter. We generalize these approaches to offer a precise characterization of the general minimal actuator selection problem where a set of actuators is given, described by a fixed input matrix, and goal is to choose the fewest actuators that make the system controllable. We show that this problem can be equivalently cast as an integer linear program and, if actuation channels are sufficiently independent, as a set multicover problem under multiplicity constraints. The latter equivalence is always true if the state matrix has all distinct eigenvalues, in which case it simplifies to the set cover problem. Such characterizations hold even when a robust selection that tolerates a given number of faulty actuators is desired. Our established connection legitimates a designer to use algorithms from the rich literature on the set multicover problem to select the smallest subset of actuators, including exact solutions that do not require brute-force search.

This paper presents finite-time and fixed-time stabilization results for inhomogeneous abstract evolution problems, extending existing theories. We prove well-posedness for strong and weak solutions, and estimate upper bounds for settling times for both homogeneous and inhomogeneous systems. We generalize finite-dimensional results to infinite-dimensional systems and demonstrate partial state stabilization with actuation on a subset of the domain. The interest of these results are illustrated through an application of a heat equation with memory term.

Current cellular systems achieve high spectral efficiency through Massive MIMO, which leverages an abundance of antennas to create favorable propagation conditions for multiuser spatial multiplexing. Looking towards future networks, the extrapolation of this paradigm leads to systems with many hundreds of antennas per base station, raising concerns regarding hardware complexity, cost, and power consumption. This article suggests more intelligent array designs that reduce the need for excessive antenna numbers. We revisit classical uniform array design principles and explain how their uniform spatial sampling leads to unnecessary redundancies in practical deployment scenarios. By exploiting non-uniform sparse arrays with site-specific antenna placements – based on either pre-optimized irregular arrays or real-time movable antennas – we demonstrate how superior multiuser MIMO performance can be achieved with far fewer antennas. These principles are inspired by previous works on wireless localization. We explain and demonstrate numerically how these concepts can be adapted for communications to improve the average sum rate and similar metrics. The results suggest a paradigm shift for future antenna array design, where antenna intelligence replaces sheer antenna count. This opens new opportunities for efficient, adaptable, and sustainable Gigantic MIMO systems.

Battery-free Internet of Things (BF-IoT) enabled by backscatter communication is a rapidly evolving technology offering advantages of low cost, ultra-low power consumption, and robustness. However, the practical deployment of BF-IoT is significantly constrained by the limited communication range of common backscatter tags, which typically operate with a range of merely a few meters due to inherent round-trip path loss. Meta-backscatter systems that utilize metamaterial tags present a promising solution, retaining the inherent advantages of BF-IoT while breaking the critical communication range barrier. By leveraging densely paved sub-wavelength units to concentrate the reflected signal power, metamaterial tags enable a significant communication range extension over existing BF-IoT tags that employ omni-directional antennas. In this paper, we synthesize the principles and paradigms of metamaterial sensing to establish a unified design framework and a forward-looking research roadmap. Specifically, we first provide an overview of backscatter communication, encompassing its development history, working principles, and tag classification. We then introduce the design methodology for both metamaterial tags and their compatible transceivers. Moreover, we present the implementation of a meta-backscatter system prototype and report the experimental results based on it. Finally, we conclude by highlighting key challenges and outlining potential avenues for future research.

In this paper, we propose a variable-length wideband channel state information (CSI) feedback scheme for Frequency Division Duplex (FDD) massive multiple-input multipleoutput (MIMO) systems in U6G band (6425MHz-7125MHz). Existing compressive sensing (CS)-based and deep learning (DL)- based schemes preprocess the channel by truncating it in the angular-delay domain. However, the energy leakage effect caused by the Discrete Fourier Transform (DFT) basis will be more serious and leads to a bottleneck in recovery accuracy when applied to wideband channels such as those in U6G. To solve this problem, we introduce the Loewner Interpolation (LI) framework which generates a set of dynamic bases based on the current CSI matrix, enabling highly efficient compression in the frequency domain. Then, the LI basis is further compressed in the spatial domain through a neural network. To achieve a flexible trade-off between feedback overhead and recovery accuracy, we design a rateless auto-encoder trained with tail dropout and a multi-objective learning schedule, supporting variable-length feedback with a singular model. Meanwhile, the codewords are ranked by importance, ensuring that the base station (BS) can still maintain acceptable reconstruction performance under limited feedback with tail erasures. Furthermore, an adaptive quantization strategy is developed for the feedback framework to enhance robustness. Simulation results demonstrate that the proposed scheme could achieve higher CSI feedback accuracy with less or equal feedback overhead, and improve spectral efficiency compared with baseline schemes.

Large AI Model (LAM) have been proposed to applications of Non-Terrestrial Networks (NTN), that offer better performance with its great generalization and reduced task specific trainings. In this paper, we propose a Deep Reinforcement Learning (DRL) agent that is guided by a Large Language Model (LLM). The LLM operates as a high level coordinator that generates textual guidance that shape the reward of the DRL agent during training. The results show that the LAM-DRL outperforms the traditional DRL by 40% in nominal weather scenarios and 64% in extreme weather scenarios compared to heuristics in terms of throughput, fairness, and outage probability.

Diabetic retinopathy (DR), affecting millions globally with projections indicating a significant rise, poses a severe blindness risk and strains healthcare systems. Diagnostic complexity arises from visual symptom overlap with conditions like age-related macular degeneration and hypertensive retinopathy, exacerbated by high misdiagnosis rates in underserved regions. This study introduces TIMM-ProRS, a novel deep learning framework integrating Vision Transformer (ViT), Convolutional Neural Network (CNN), and Graph Neural Network (GNN) with multi-modal fusion. TIMM-ProRS uniquely leverages both retinal images and temporal biomarkers (HbA1c, retinal thickness) to capture multi-modal and temporal dynamics. Evaluated comprehensively across diverse datasets including APTOS 2019 (trained), Messidor-2, RFMiD, EyePACS, and Messidor-1 (validated), the model achieves 97.8% accuracy and an F1-score of 0.96, demonstrating state-of-the-art performance and outperforming existing methods like RSG-Net and DeepDR. This approach enables early, precise, and interpretable diagnosis, supporting scalable telemedical management and enhancing global eye health sustainability.

In lifelong multi-agent path finding (MAPF) with many robots, unreliable wireless links and stochastic executions are the norm. Existing approaches typically either rely on centralized planning under idealized communication, or run fully distributed local controllers with fixed communication patterns; they rarely couple communication scheduling with policy learning, and thus struggle when bandwidth is scarce or packets are frequently dropped. We address this joint control–communication problem and propose a hybrid centralized–distributed scheme: a centralized cloud policy sends small residual corrections only when selected, while a lightweight on-board Gated recurrent unit (GRU) policy provides a safe default fallback when wireless connection is not available.

The exploitation of mineral resources in plateau regions is confronted with critical challenges including low blasting efficiency, excessive energy consumption,and compromised operational safety when dealing with low-temperature water-bearing frozen rock masses.This study systematically investigates the dynamic-static mechanical properties,deformation-fracture behaviors,and energy consumption characteristics of plateau frozen sandstone under the coupled effects of temperature and moisture content (5%-15%).The research methodology integrates field sampling, low-pressure low-temperature simulation tests, graded impact loading tests, and numerical inversion analysis. Results demonstrate that freezing significantly enhances the dynamic strength and brittleness of saturated sandstone.The pore structure undergoes substantial evolution with decreasing temperature, with the porosity increasing by 63.15%.Based on PFC3D microscopic simulations, the mechanism of frost heave damage and the regulatory effect of water-ice phase transition on rock mechanical behaviors are elucidated.A quantitative analysis method for energy dissipation is proposed, revealing that the energy absorption increment of frozen rocks is higher than that of room-temperature samples.The findings provide a theoretical basis and technical support for optimizing blasting parameters, realizing directional energy release,and promoting green construction of frozen rock masses in high-altitude areas.

Bilinear Matrix Inequalities (BMIs) are fundamental to control system design but are notoriously difficult to solve due to their nonconvexity. This study addresses BMI-based control optimization problems by adapting and integrating advanced evolutionary strategies. Specifically, a memetic Covariance Matrix Adaptation Evolution Strategy (memetic CMA-ES) is proposed, which incorporates a local refinement phase via a (1+1)-CMA-ES within the global search process. While these algorithmic components are established in evolutionary computing, their tailored integration and specific tuning for control design tasks represent a novel application in this context. Experimental evaluations on $H_{\infty}$ controller synthesis and spectral abscissa optimization demonstrate that the proposed method achieves superior performance compared to existing BMI solvers in terms of both solution quality and robustness. This work bridges the gap between evolutionary computation and control theory, providing a practical and effective approach to tackling challenging BMI-constrained problems.

Diffusion and flow policies are gaining prominence in online reinforcement learning (RL) due to their expressive power, yet training them efficiently remains a critical challenge. A fundamental difficulty that distinguishes online RL from standard generative modeling is the lack of direct samples from the target Boltzmann distribution defined by the Q-function. To address this, two seemingly distinct families of methods have been proposed for diffusion policies: a noise-expectation family, which uses a weighted average of noise as the training target, and a gradient-expectation family, which employs a weighted average of Q-function gradients. However, it remains unclear how these objectives are formally related, or whether they can be synthesized into a more general formulation. In this paper, we propose a unified framework, reverse flow matching (RFM), which rigorously addresses the problem of training diffusion and flow models without direct target samples. By adopting a reverse inferential perspective, we formulate the training target as a posterior mean estimation problem given an intermediate noisy sample. Crucially, we introduce Langevin Stein operators to construct zero-mean control variates, deriving a general class of estimators that share the same expectation. We show that existing noise-expectation and gradient-expectation methods are simply two specific instances within this broader class. This unified view yields two key advancements: it extends the capability of targeting Boltzmann distributions from diffusion to flow policies, and it enables the principled combination of Q-value and Q-gradient information to form an effective estimator, thereby improving training efficiency and stability. We instantiate RFM to train a flow policy in online RL and demonstrate improved performance on continuous-control benchmarks compared to diffusion policy baselines.

Ground Vibration Testing (GVT) supports aircraft certification but often requires lengthy and costly campaigns. Propeller-driven Vibration Testing (PVT) is assessed here as an output-only alternative, in line with Operational Modal Analysis approaches such as Taxi Vibration Testing and Flight Vibration Testing. A cantilever Aluminium 7075-T6 wing spar is instrumented with seven accelerometers and excited by an outboard electric motor and propeller. Seven runs are carried out: a motor-off baseline, five constant-throttle cases, and a manual up-down throttle sweep. The acquired spectra indicate that the dominant resonances remain observable under propeller excitation, while low-throttle conditions introduce narrowband harmonics that may mask structural peaks; the sweep reduces persistent overlap. Modal parameters are identified for the baseline and sweep cases using the Natural Excitation Technique with the Loewner Framework (NExT-LF). The first two modes remain closely matched (Modal Assurance Criterion (MAC) > 0.99), whereas the third mode shows reduced repeatability (MAC = 0.827) and a larger frequency shift, consistent with propeller-induced bending–torsion coupling and non-ideal sweep control. Overall, PVT provides a viable complement to GVT for extracting low-frequency modal information and motivates pursuing future work on automated throttle scheduling and coupling-aware test planning.