Search Results (949)

Integrating complementary attention mechanisms into standard Convolutional Neural Networks (CNNs) is a promising strategy for improving feature discrimination without substantial computational overhead. This paper presents a controlled empirical study of a hybrid attention module that combines Squeeze-and-Excitation Networks (SENet) and the Convolutional Block Attention Module (CBAM) through an adaptive element-wise summation with a learnable weighting parameter $\alpha$ and a residual connection. This work contributes a systematic and statistically rigorous evaluation of attention fusion across four CNN backbones (ResNet18, VGG16, AlexNet, and SqueezeNet) on the CIFAR-10 benchmark at $32\times 32$ resolution. All models were trained from scratch under a deliberately conservative protocol (50 epochs, no pretrained weights, standard augmentation) to isolate the incremental effect of attention mechanisms under controlled conditions. Under this protocol, the hybrid SENet+CBAM configuration achieves statistically significant accuracy improvements over the corresponding baselines ($p<0.001$, 5-fold cross-validation): ResNet18 improves from 77.93% to 90.71% (+12.78%), VGG16 from 55.78% to 70.17% (+14.39%), AlexNet from 62.67% to 71.82% (+9.15%), and SqueezeNet from 71.91% to 78.29% (+6.38%). These gains must be interpreted within the scope of this controlled setting. Absolute accuracy values are below fully optimized literature benchmarks. For VGG16 in particular, part of the improvement likely reflects correction of underfitting under the conservative protocol, not the full potential of the hybrid mechanism. Parameter overhead remains modest at 1.5–5.8%, and training convergence improves by 16.5% on average. The hybrid approach outperforms the best previously reported SENet+CBAM result for each architecture by an average of 2.32%. Grad-CAM visualizations and attention entropy analysis provide qualitative evidence of more concentrated spatial attention patterns under the hybrid configuration. These should be understood as proxy indicators rather than rigorous interpretability measures. Validation on higher-resolution benchmarks such as CIFAR-100, STL-10, and ImageNet subsets is a necessary next step before broader applicability can be claimed. Full article

Healthcare analytics is often limited by data silos and strict privacy requirements, which make it difficult to share patient-level records across organisations and to build robust predictive models. Federated learning (FL) provides an alternative by keeping data local and exchanging model updates instead of raw records. However, many existing FL solutions remain difficult to deploy in healthcare settings, as they provide limited support for auditability, governance-oriented evidence, and system-level transparency. This paper presents MediVault, an auditable and security-aware federated learning-based system for privacy-preserving healthcare collaboration. MediVault combines round-based federated training, prototype-level protected update exchange, audit-ready telemetry, and an interactive dashboard that exposes non-sensitive evidence of collaboration, model progress, and protocol execution. In addition, the system supports controlled reporting to improve stakeholder communication during pilot deployments. We evaluate MediVault on two public healthcare classification datasets, Breast Cancer Wisconsin (Diagnostic) and Heart Disease, under IID and label-skewed Non-IID settings. Experiments are conducted using logistic regression, linear SVM, and an additional lightweight MLP under matched settings. The observed results suggest that federated training remains competitive with centralised training under the evaluated settings. A prototype-level overhead analysis further shows that protected update exchange introduces measurable computational and communication costs, especially for larger update vectors. These findings indicate that MediVault can support initial system-level validation of auditable, privacy-preserving healthcare FL workflows, while further work is needed for larger-scale deployment, stronger adversarial evaluation, and real-world clinical validation. Full article

Fault tolerance in distributed real-time systems has, up till now, relied on static redundancy, replication, or predictive mechanisms, which introduce latency, resource overhead, and inadaptability under dynamic failure conditions. This paper presents Chrono Weave (CW) as a revolutionary new idea that describes how a system is working as a flow of a time-ordered field of states, so that even if the system is broken, it can recover without explicit redundancy or replication. CW does not replicate computation but rather encodes system evolution into temporally entangled microstates; therefore, recovery is made possible through deterministic temporal interpolation. The Temporal Consistency Field (TCF), a new concept, is presented to measure system integrity over time, enabling fault localization and instant reconstruction. The new system does not require standby replicas, and recovery is achieved just by way of using temporal coherence that is inherent. From a theoretical viewpoint, it is shown that CW can reduce recovery latency asymptotically towards zero as long as the drift is bounded. From the perspective of distributed control, simulation experiments have still managed to show great recovery speed and system reliability improvements over the traditional ones. This paper opens fault-tolerant computing to a new mode of operation where instead of being based on redundancies, time-structured, self-healing systems are used. Full article

DC–DC converters are ubiquitous in consumer, industrial, commercial, and medical applications. In such voltage-, power-, and area-constrained systems, guaranteeing accurate output voltage remains a key challenge. Investigation of the fundamental cause of steady-state output errors in DC–DC converters, however, is largely absent in the literature. This work identifies systemic voltage offset error as one of the key contributors to steady-state output inaccuracy in PWM and peak–valley-controlled switched-inductor voltage regulators. It uses an insightful reverse-feedback translation framework to quantify the systemic offset as a function of the duty cycle, input voltage, sawtooth amplitude, propagation delays, load conditions, error amplifiers, and comparator. Furthermore, with the derived offset expressions, the paper develops accurate and low-overhead design guidelines to remove systemic errors by aligning the regulator’s steady-state equilibrium with its operating conditions. With the proposed offset “centering” and “elimination” techniques, the systemic error (that accounts for up to 2.1% variation in the steady-state output) is reduced by over 70% when centered and to zero when eliminated at room temperature. Overall, this work provides an insightful and generalized quantification of systemic offsets and describes low-overhead strategies to restore steady-state accuracy in practical PWM, hysteretic and peak/valley-controlled voltage regulators. Full article

Blockchain-based crowdsourcing logistics is a promising decentralized paradigm for solving the “last-mile delivery” problem, in which smart contracts automatically execute the business logic. Since crowdsourcing logistics inherently involves frequent fund transfers, its smart contracts are particularly susceptible to reentrancy vulnerabilities. Existing works address reentrancy by inserting a lock mechanism at design-time, which lacks dynamic responsiveness and incurs additional gas overhead. To overcome this limitation, we propose RE4SC, the first runtime enforcement framework for vulnerable smart contracts. RE4SC contains two components: off-Blockchain granularity segmentation and on-Blockchain granular block reordering. At the off-Blockchain level, bytecode is segmented into granular blocks through control flow analysis. This yields a finer granularity than conventional basic blocks in a control flow graph. These granular blocks are then organized into a tree structure that captures their hierarchical nesting relationships. A data flow analysis further ensures data dependency consistency after reordering. At the on-Blockchain level, a runtime enforcer retrieves the pre-computed reordering specifications from off-Blockchain analysis. It applies a depth-first reordering algorithm to reposition key state variable assignments before transfer operations, eliminating reentrancy vulnerabilities without introducing additional bytecode. We implement a prototype tool and make it open-source. Experiments on self-constructed crowdsourcing logistics contracts and three public datasets demonstrate that RE4SC repairs vulnerable contracts with zero gas overhead, outperforming existing approaches. Full article

Side-scan sonar (SSS) imaging plays a critical role in underwater perception for autonomous underwater vehicles (AUVs). However, the spatial sparsity of targets and the limited computational resources remain challenging for real-time object detection. Existing methods typically adopt dense inference strategies, leading to substantial computational redundancy and limited deployment feasibility. In this work, we propose a lightweight and ultra-fast SSS object detection framework based on target presence awareness. The proposed framework follows a coarse-to-fine inference paradigm, in which a target presence analysis module is first employed to rapidly filter out target-absent image patches, and only target-positive patches are forwarded to an Object Forward Detection (OFD) module for fine-grained detection. The TPA module integrates spatial–frequency convolution to efficiently capture both local structural cues and global contextual information with minimal computational overhead. Furthermore, an AttnConv-enhanced detection module is introduced in the OFD stage to strengthen high-frequency target features and improve fine-grained detection performance. Extensive experiments on public SSS datasets demonstrate that the proposed method achieves an mAP of 74.63% on the AI4Shipwrecks dataset and 63.02% on the SSS-Mine dataset. Notably, the framework delivers an ultra-fast inference speed of 174.74 FPS on embedded hardware, representing a 5.2× speedup over conventional dense-processing detection methods. Full article

Accurate short-term traffic prediction is a key requirement for modern traffic management systems, yet many existing approaches remain focused on offline evaluation and do not address the challenges of continuous real-time deployment. In this work, we present a context-aware spatiotemporal graph convolutional network (STGCN) framework designed for low-latency, scalable traffic forecasting under operational conditions. The proposed approach integrates structural information from the road network, temporal regularities derived from historical data, and a residual correction mechanism trained on systematic prediction errors observed during real-time operation. The framework is designed to remain lightweight, enabling continuous minute-level inference without computational overhead that would hinder long-term deployment. The methodology is evaluated in two real-world case studies of different scale and complexity. In Thessaloniki, Greece, multiple forecasting models are evaluated across different temporal resolutions using one-minute speed data, with the proposed STGCN selected for real-time deployment. A residual correction module trained on historical prediction errors further improves real-time forecasting accuracy compared to the baseline STGCN deployment. Scalability is further demonstrated in the South Holland region of the Netherlands, where the same architecture is applied to a larger network and extended to multi-horizon forecasting. Results show that the proposed framework achieves competitive predictive performance while maintaining low computational cost, and that incorporating residual error learning provides a robust and practical solution for improving forecasting accuracy in real-world deployments. These findings highlight the importance of combining domain-specific modeling with operational considerations in traffic prediction systems. Full article

Connected and autonomous vehicles (CAVs) increasingly rely on vehicle-to-everything (V2X) communication and distributed sensing infrastructures to support cooperative driving and intelligent transportation services. While these capabilities improve traffic efficiency and safety, they also expand the attack surface of vehicular networks and expose in-vehicle communication systems such as the Controller Area Network (CAN) bus to a wide range of cyber threats. Machine learning-based anomaly detection has emerged as a promising approach for identifying malicious CAN traffic patterns; however, conventional centralized learning requires large-scale data aggregation from vehicles, which raises privacy and scalability concerns. Federated learning (FL) enables collaborative model training across distributed vehicles without requiring the exchange of raw in-vehicle data, making it attractive for privacy-preserving vehicular security applications. Nevertheless, FL systems remain vulnerable to adversarial participants that manipulate local training data or model updates to poison the global model during aggregation. In this work, we present a systematic robustness evaluation of federated anomaly detection in connected vehicular networks under adversarial conditions. The study compares six aggregation strategies, including Federated Averaging (FedAvg), coordinate-wise Median, Trimmed Mean, Krum, Multi-Krum, and Geometric Median (GeoMed), within a non-IID federated CAN bus anomaly detection setting. The evaluation covers label-flipping attacks, gradient-scaling attacks, and a feature-triggered backdoor attack. In addition, the analysis examines malicious client participation, attack-strength variation, learning-rate sensitivity, Trimmed Mean beta sensitivity, multi-seed reliability, and server-side aggregation time. The results show that FedAvg is vulnerable under strong adversarial manipulation, while Trimmed Mean is sensitive to the selected trimming fraction. Median and GeoMed provide strong robustness against gradient-scaling attacks, whereas Multi-Krum achieves the strongest resistance to label-flipping and backdoor attacks. These findings demonstrate that no single aggregation strategy is optimal across all threat models. Instead, robust aggregation for federated CAV anomaly detection should be selected according to the expected attack type, reliability requirement, and computational overhead. Full article

The rapid proliferation of Internet of Things (IoT) devices in cloud-based e-learning platforms has posed significant security risks, particularly in protecting learner information, authentication of devices, and safe communication in the highly heterogeneous learning settings. Current cryptographic solutions are largely based on classical public-key infrastructure (PKI) protocols such as RSA and ECC, which will become vulnerable with the advent of large-scale quantum computers capable of executing Shor’s algorithm. In addition, traditional perimeter-based security models are inadequate for handling the dynamics, scattered, and resource-limited characteristics of IoT-enabled educational systems. As a solution to these problems, this paper introduces ZeroTrustEdu, a scalable zero-trust cryptographic solution that combines lightweight post-quantum key management with adaptive trust scoring of cloud-connected IoT e-learning infrastructure. The proposed framework makes three fundamental contributions namely: (1) a hierarchical zero-trust security model with no implicit trust, operating across device, edge, and cloud layers; (2) a lightweight key distribution protocol based on the Module-Lattice Key Encapsulation Mechanism (ML-KEM) compliant with NIST FIPS 203 standards and (3) an adaptive behavioral trust scoring engine that dynamically adjusts device and user trust levels based on real-time interaction analytics. The architecture is evaluated using extensive NS-3 network simulations with up to 100,000 concurrent IoT nodes with formal security analysis under Chosen Plaintext Attack (CPA) and Chosen Ciphertext Attack (CCA) threat models. Comparative evaluation against RSA-2048, ECC-P256, and AES-256 baselines demonstrates that, ZeroTrustEdu delivers a 62% ± 3% (95% CI, 10 independent runs) reduction in ML-KEM encapsulation latency (12.8 ms for key encapsulation/decapsulation, contributing to a complete device authentication latency of 47.3 ms including ML-DSA signature operations), 45% reduced communication overheads, and 38% reduction in energy consumption on ARM Cortex-M4 constrained devices compared to RSA-2048 and achieves provable post-quantum security reducible to the hardness of the Module Learning With Errors (MLWE) problem. These findings demonstrate that the proposed architecture provides a viable, scalable, and quantum-resilient security solution for next-generation IoT-enabled e-learning environments. The cryptographic security of ZeroTrustEdu is guaranteed at the primitive level through NIST-standardized ML-KEM (FIPS 203) and ML-DSA (FIPS 204), with IND-CCA2 and EUF-CMA security formally proven in the respective standards; full protocol-level formal verification using automated theorem provers (ProVerif, Tamarin) is identified as valuable future work to rule out protocol-composition vulnerabilities beyond primitive-level guarantees. Full article

Intel Trust Domain Extensions (TDX) is a Confidential Virtual Machine (CVM) technology that provides hardware-enforced isolation through Trusted Execution Environments (TEEs). While TDX effectively mitigates interrupt-based stepping attacks, it remains vulnerable to fine-grained timing-based controlled-channel attacks such as T-Time, which exploit precise dwell-time measurements between consecutive page faults to infer secret-dependent control flows even within a single memory page. Existing page-level confinement defenses are insufficient against such timing attacks. In this paper, we propose RandDelay, a lightweight defense mechanism that raises the measurement budget required for a successful T-Time attack by injecting a cryptographically random latency into the SEAMCALL handler of the TDX module. We argue that SEAMCALL is the most practical and effective injection point among mandatory boundary handlers: it lies strictly between the attacker’s timestamp $T_s$ and the victim’s secret-dependent code execution, ensuring that every dwell-time measurement is corrupted by an independent random variable. We further integrate an anomaly-based page-fault rate limiter (RandDelay+) to prevent statistical averaging attacks. Security analysis shows that RandDelay raises the minimum number of measurements required for a successful attack beyond the budget enforced by RandDelay+, rendering the attack impractical under the analytical model and assumed parameter settings. We discuss implementation considerations within the TDX module firmware, expected performance overhead, and generalization to other TEE platforms. This paper contributes a design rationale, a quantitative tuning rule, and an analytical security–overhead model that together provide a deployable baseline for empirical follow-up. The proposed defense has not been experimentally validated on a deployed TDX system; a prototype, simulation, or trace-driven study is identified as essential future work. Full article

Industrial control systems (ICSs) increasingly rely on legacy communication protocols such as ModbusTCP, which lack built-in security mechanisms and remain widely exposed to network-based attacks. This paper investigates the security limitations of authentication mechanisms in ModbusTCP-enabled programmable logic controllers (PLCs) and demonstrates how plaintext credential transmission and limited connection handling capabilities can be exploited to perform brute-force and denial-of-service (DoS) attacks. An experimental testbed based on two industrial Delta PLC families (DVP-13SE and DVP-311SV3) was developed to systematically evaluate these vulnerabilities under realistic conditions. The results show that authentication credentials can be easily captured through network sniffing, while the PLC communication stack supports a maximum of 16 concurrent connections and can process up to approximately 8600 Modbus operations per second, making it susceptible to resource exhaustion and performance degradation under distributed attack scenarios. To address these limitations, this paper proposes a lightweight deception-based protection mechanism, termed the PLC misleading algorithm (PMA), which is implemented directly within the PLC ladder logic. Unlike traditional network-level defenses, PMA operates at the device level and dynamically misleads attackers by generating controlled randomized responses while preserving consistent behavior for legitimate clients. Experimental results demonstrate that PMA significantly mitigates brute-force effectiveness by preventing reliable password extraction while introducing minimal overhead (2.2% memory usage) and maintaining acceptable communication latency. Additionally, the proposed approach significantly reduces observable attack traffic, with only 0.246 Modbus operations per second observed during the attack phase, thereby limiting the effectiveness of automated exploitation tools. These findings highlight the potential of in-device deception mechanisms as a practical and deployable security layer for legacy industrial systems, and provide new insights into the resilience of PLC-based infrastructures against network-level attacks. This work bridges the gap between lightweight PLC-level protections and the growing need for robust cybersecurity mechanisms in industrial IoT environments. Full article

Integrating on-device learning into autonomous systems requires neural network frameworks that achieve both high energy efficiency and low latency. While spiking neural networks (SNNs) provide a promising event-driven paradigm, implementing hardware-efficient learning remains a challenge due to the computational overhead of error signaling and global gradients. This paper introduces a hardware-oriented hierarchical spiking predictive coding (SPC) framework designed for end-to-end event-driven systems. The proposed architecture implements an implicit prediction error encoding mechanism through local lateral and supervisory feedback connections, eliminating the need for dedicated error-storage memory or complex inter-layer error communication. The entire framework is structured and parameterized for physical implementation, utilizing digital-aligned simulations and arithmetic operations. We evaluate the system on neuromorphic datasets using a fixed 1 ms temporal resolution to mirror real-time hardware constraints. Experimental results demonstrate that the SPC framework can effectively identify stimuli from transient event streams, achieving stable on-device learning. Our work provides a practical path toward deploying low-power, scalable hierarchical spiking networks in resource-constrained environments. Full article

Reliable monocular depth estimation (MDE) is essential for agricultural robots and unmanned platforms, where low-cost visual perception is required for safe navigation and scene understanding in complex field environments. However, general-purpose depth foundation models remain limited by substantial domain gaps in agriculture, while full fine-tuning of large backbones is computationally expensive and less suitable for deployment on resource-constrained platforms. In this paper, an efficient agricultural MDE framework, termed AgriLoRA-DA, is proposed based on Depth-Anything-V2. Specifically, the pretrained DINOv2 encoder is kept frozen and adapted using LoRA in selected attention projections, while the original Dense Prediction Transformer (DPT) decoder is replaced with a lightweight Lite-FPNHead to reduce decoding overhead and improve deployment efficiency. Experiments conducted on the WE3DS dataset indicate that, although Depth-Anything-V3 provides the strongest zero-shot generalization among the evaluated baselines, target-domain adaptation is still necessary for WE3DS agricultural scenes. After adaptation, AgriLoRA-DA achieves the best overall performance with AbsRel = 0.0133, SqRel = 3.518, RMSE = 132.264, log10 = 0.0057, and delta1 = 0.9990, while requiring only 0.19 M (0.87%) trainable parameters. These results suggest that parameter-efficient adaptation and lightweight decoding provide a practical direction for deployable depth estimation in crop-row scenes similar to WE3DS, while broader cross-dataset validation remains an important direction for future work. Full article

This paper is the empirical validation companion to our prior methodology paper introducing the Automated Zero Trust Risk Management with DevSecOps Integration (AZTRM-D) methodology, conducted through multi-vector adversarial testing on physical NVIDIA Jetson Orin Nano hardware. AZTRM-D unifies DevSecOps automation, the NIST Risk Management Framework, and Zero Trust architecture with AI orchestration via Cybectr Sentinel, featuring six AI subsystems with formal specifications. Testing spanned three progressive hardening stages across seven attack categories under a blind three-tester protocol with inter-rater agreement analysis. Factory-default devices were fully compromised in under five minutes. After full hardening, zero successful breaches were recorded across any tested vector. The CI/CD pipeline achieved a vulnerability detection rate of 96.8% (Wilson 95% CI: [0.891, 0.991]). Sentinel delivered 94.1% precision, 91.8% recall, and 4.2 min average detection time within 12–18% CPU overhead on edge hardware. A 14-capability comparative analysis against five established frameworks found seven capabilities unique to AZTRM-D. The 93.7% adversarial detection rate is reported against DiCE-generated counterfactual inputs and is bounded by the black-box threat model used in evaluation; gradient-based white-box attack evaluation is documented as a scoped Stage 4 future-work item. All three testers are affiliated with Cybectr LLC, the developer of AZTRM-D and Cybectr Sentinel; this conflict of interest is the most significant limitation of the present work, and independent third-party laboratory validation is the highest-priority Stage 4 deliverable. Full article

Low-light enhancement technologies are of great significance for visual driver assistance applications and autonomous driving systems. Infrared vision can improve nighttime visibility but also faces challenges of low resolution and lack of color information. This paper presents a unified framework for RGB-guided infrared super-resolution and infrared-visible fusion that achieves high-resolution output under limited computational resources. Our approach employs a U-Net architecture with novel triple-grouped window attention (TGWA) encoding that captures global dependencies through grouped attention while reducing computational overhead, and adaptive multi-dilated convolutional (AMDC) decoding that adaptively selects optimal dilation rates using mixture-of-experts-inspired routing. Experiments on multiple datasets achieve competitive super-resolution and fusion results with minimal computational complexity, while real-world downstream object detection validation confirms robust performance in challenging nighttime scenarios. Quantitatively, the proposed method achieves 28.744 dB/0.872 SSIM on PBVS24 and 31.424 dB/0.882 SSIM on HDRT-Night for 8× infrared super-resolution, reaches competitive fusion quality on both MSRS and HDRT-Night, and attains 69.4% mAP@0.5 in downstream object detection on FLIR_aligned, while requiring only 1.12 M parameters and 85.44 G FLOPs. This work provides new possibilities for seeing clearly in the dark. Full article