Search Results (61)

Networked Cyber–Physical Systems (CPSs), like autonomous quadrotor swarms, tightly couple continuous physical kinematics, wireless information exchange, and discrete real-time task scheduling. While conventional consensus security architectures focus exclusively on data-layer falsification, they fundamentally decouple adversarial behavior from onboard computational resource state profiles. This paper addresses a core CPS vulnerability termed Computational Weaponization, the deliberate injection of complex computational workloads (adversarial LLM token parsing or cryptographic verification) to intentionally manipulate hardware execution delays. Through this exploit, strategic cyber–physical perturbations force resource-constrained embedded microcontrollers to saturate their task execution queues, inducing real-time scheduling starvation and physical tracking divergence. To mitigate this without optimization bottlenecks, we present a state-dependent, Self-Triggered Control (STC) and Prospect Theoretic Alignment (PTA) co-design framework. The proposed protocol models the hardware microprocessor’s execution delay as an endogenous dynamic state coupled directly to continuous tracking spaces. By mapping discrete topology reconfigurations and variable task delays to a switched impulsive time-delay system, we leverage an Input-to-State Stability (ISS) to derive sufficient linear matrix inequality conditions. We prove that the coupled cyber–physical–computational loop achieves asymptotic consensus and bounded trajectory containment under adversarial actions. Full article

Understanding pore-scale interactions among CO₂, the aqueous phase, and heavy oil is crucial for optimizing surfactant-assisted CO₂-enhanced oil recovery. In this study, a three-phase volume-of-fluid (VOF) framework is established in OpenFOAM to simulate immiscible displacement in two-dimensional porous media composed of circular solid grains. Two pore geometries are considered: a relatively uniform geometry and a preferential-channel geometry designed to promote early breakthrough. Sixteen injection schemes are compared, including pure CO₂ flooding, ordinary-water flooding, surfactant–water flooding, aqueous-phase preflush followed by CO₂ injection, and cyclic aqueous-phase/CO₂ injection with different aqueous-slug durations and switching frequencies. The effects of pore geometry, injection strategy, capillary number, viscosity ratio, interfacial tension, and wettability are evaluated using pore-volume-normalized oil recovery, breakthrough PV (the ratio of injected volume to pore volume), cumulative injected CO₂ PV at breakthrough, and phase-distribution indicators. The results show that pore geometry strongly affects macroscopic sweep and breakthrough behavior. In the preferential-channel geometry, pure CO₂ flooding and continuous ordinary-water flooding suffer from early breakthrough and poor sweep, whereas continuous surfactant–water flooding maintains high recovery because reduced oil–water interfacial tension and a more water-wet wall condition promote oil-film detachment and residual-oil mobilization. At 2.5 injected PV, the high-frequency short-slug and 1 s surfactant–water cyclic schemes give the highest or near-highest recovery in the relatively uniform geometry, while continuous surfactant–water flooding remains the highest-recovery scheme in the preferential-channel geometry. Considering the higher chemical demand of continuous surfactant injection and the carbon-utilization objective of CO₂-EOR, the combined surfactant–water/CO₂ schemes are evaluated to clarify the coupling between surfactant-induced oil mobilization and CO₂ displacement. Surfactant–water preflush followed by CO₂ injection becomes more effective as the preflush duration increases, and its recovery advantage over ordinary-water preflush is especially large in the preferential-channel geometry. However, the increase in CO₂ breakthrough PV in this geometry is limited compared with the recovery increment, indicating that the main benefit of surfactant–water is not only delayed gas breakthrough but also enhanced microscopic oil mobilization in poorly swept regions. Increasing the switching frequency slightly improves the cyclic response in the relatively uniform geometry under the tested schedules, whereas all cyclic schemes remain strongly constrained by reconnection with the dominant gas pathway in the preferential-channel geometry. Within the present idealized immiscible VOF model, these comparisons provide a controlled pore-scale comparison for distinguishing surfactant-induced residual-oil mobilization from CO₂ gas-channeling effects in heavy-oil porous media. Full article

Embedded systems increasingly require on-device deep learning, yet their processors must simultaneously handle real-time sensing, networking administration, and data control. Existing Field-Programmable Gate Array (FPGA) accelerators typically target peak per-operator throughput without addressing concurrent execution demands of real-time embedded platforms. This paper presents a systolic array-based accelerator prototype implemented on the Zynq-7000 SoC integrated directly into PyTorch, enabling dense linear algebra to be delegated to the FPGA chip while Cortex-A9 continues executing the software stack uninterrupted. Unlike traditional accelerators optimized for peak per-operator speed, this design prioritizes asynchronous co-executionbetween the processing system (PS, the dual-core Cortex-A9) and the programmable logic (PL): The PL performs tiled matrix multiplication, while the PS executes preprocessing, orchestration, and I/O data concurrently, increasing effective end-to-end throughput regardless of the relative isolated performance of CPU and FPGA. The proposed module includes high-level-synthesis (HLS)-based matrix multiplication, activation functions, and Advanced eXtensible Interface (AXI)-Stream Direct Memory Access (DMA) interfaces, wrapped as custom PyTorch kernels under the PetaLinux operating system. The results obtained on the PYNQ-Z2 board show that, once the DMA transfer time is included in the measurement, the FPGA path does not surpass Cortex-A9 in isolated per-operator latencies across the evaluated range; the benefit lies instead in delegating the matrix compute to the fabric at low incremental power while the host CPU cores stay available for concurrent tasks. A concurrent workload sweep across matrix sizes from $8\times 8$ to $256\times 256$ confirms that the co-execution mode sustains 98–99% of available PS compute throughput compared with a constant ≈50% in single-core blocking mode; the difference is statistically significant for all evaluated sizes (see Mann–Whitney U: $U=25$, $p=3.97\times {10}^{-3}$, perfect discrimination, $n=5$). A fair dual-core CPU-only baseline attains comparable PS availability, so this figure reflects the dual-core scheduling that co-execution enables rather than a per-operator advantage of the fabric; the accelerator’s distinct role is to perform the matrix arithmetic off the general-purpose cores at low incremental power. The design occupies only 8% of available look-up tables (LUTs) and 5% of digital signal processing (DSP) blocks, maintains 1.69 W power with a junction temperature of 44.5 °C, and achieves 96.10% MNIST accuracy under fixed-point arithmetic. Full article

Traditional windcatchers are often discussed as passive cooling devices for hot–arid climates, yet their sustainability performance under contemporary comfort-controlled operation remains insufficiently understood. This study evaluates the energy and load-based carbon implications of integrating a windcatcher into a high-thermal-mass courtyard house in Şanlıurfa, Türkiye. A combined DesignBuilder v6.1/EnergyPlus v8.2 and CFD-based assessment was used: annual heating and cooling loads were calculated through EnergyPlus-based building energy simulation, while CFD analyses were used to interpret representative airflow behavior and localized thermal effects within the semi-open iwan. Scenarios varied operational schedule, geometry, material configuration, ventilation openings, and water pool integration. The results show strongly context-dependent performance rather than uniform energy or carbon benefit. Continuous operation weakened annual performance, whereas seasonal operation produced more balanced outcomes. The P.1 configuration produced the lowest total annual energy demand among the tested scenarios, decreasing total demand from 70,929.99 to 70,806.65 kWh/a, corresponding to a reduction of 123.34 kWh/a or 0.17% relative to the baseline. However, this limited reduction was accompanied by a 6.02% increase in cooling demand and a 2.52% decrease in heating demand. Consequently, the total load-based carbon indicator increased from 18.32 to 18.60 tCO₂/year, corresponding to an increase of 0.28 tCO₂/year or 1.53%. CFD results indicate that the semi-open iwan geometry and its orientation relative to prevailing winds constrained airflow effectiveness and limited the transfer of local cooling effects to conditioned zones. This study demonstrates that vernacular passive systems should be evaluated through integrated annual energy, airflow, and load-based carbon analyses before being adopted in sustainable renovation or climate-responsive design. Full article

Integratingartificial intelligence (AI), the Internet of Things (IoT), and Building Information Modeling (BIM) holds considerable promise for modernizing construction management, yet a unified real-time framework connecting these technologies for heavy civil earthmoving remains lacking. This paper presents BIM-iDT, a BIM-Integrated Digital Twin framework that couples multi-source IoT sensing with an IFC-based BIM model to enable intelligent fleet coordination and automated progress control. The research follows a design-science methodology comprising framework formulation, modular development, field deployment, and multi-project validation. The framework comprises a heterogeneous sensor fusion layer aligning GPS, IMU, fuel-consumption, and LiDAR data within the BIM coordinate system; a spatio-temporal graph attention network (ST-GAT) that recognizes equipment states and predicts short-horizon productivity by modeling fleet-level spatial dependencies; a temporal point cloud differencing module that quantifies cut/fill volumes against BIM design surfaces; and a constrained multi-objective evolutionary optimizer (CMOEO) that generates Pareto-optimal dispatch plans balancing fuel, cycle time, utilization, and schedule adherence. Validation on a highway project with instrumented machines shows that ST-GAT achieves a macro-averaged F1 of 0.943, volume MAPE stays below 3%, and CMOEO reduces fuel consumption by 12.6% and cycle time by 9.3% while maintaining schedule adherence above 96%, yielding an estimated 168-ton CO₂ emission reduction. End-to-end latency averages 600 ms, satisfying real-time requirements. Cross-project transfer experiments on a secondary dam construction site further confirm framework generalizability, establishing BIM-iDT as a scalable paradigm for AI-and-IoT-enabled smart construction in infrastructure engineering. Full article

Surface congestion and taxiing delays at high-density airports increasingly constrain aviation sustainability, as ground-phase fuel consumption and emissions constitute a significant share of total airport emissions. Existing studies typically decouple air traffic flow management from ground resource scheduling, hindering coordinated optimization of punctuality, environmental benefits, and resource utilization. This paper proposes a multi-objective optimization method for runway-taxiway systems oriented toward air–ground collaborative decision-making, integrating Calculated Take-Off Time (CTOT) compliance constraints. A tri-objective mixed-integer programming model is formulated to minimize CTOT deviation, total taxiing time, and runway workload imbalance. A hybrid intelligent algorithm, SSA-SCA-NSGA-II, is designed with a bidirectional elite feedback mechanism to address this NP-hard problem. Validation uses real operational data of 58 departure flights during a peak period at Beijing Daxing International Airport. The results demonstrate that the proposed method achieves effective trade-offs on the Pareto front: CTOT compliance rate increased from 77.6% to 89.7–96.6%; total taxiing time decreased from 692 min to 551–635 min; and dual-runway utilization imbalance declined from 5.2% to 1.7–3.8%. These improvements translate into quantifiable sustainability gains: fuel consumption is reduced by 1425–3525 kg and CO₂ emissions by 4503–11,139 kg per peak hour, alongside a 19-percentage point improvement in punctuality that lowers passenger delay costs and reduces controller coordination workload. By simultaneously advancing environmental sustainability (carbon footprint reduction), economic sustainability (fuel and operational cost savings), and social sustainability (service punctuality and labor efficiency), the framework provides a measurable, monitorable, and policy-relevant decision-support tool for green airport surface operations aligned with sustainable development goals (SDGs). Full article

The high-performance control of active suspension electro-hydraulic actuators (ASEHA) is limited by a timing mismatch: the primary internal physical state (load pressure) responds to disturbances almost instantaneously, whereas the tracking error used for feedback lags behind. To address this issue, a physics-aware co-design framework introduces three innovations: (i) a pressure-adaptive bandwidth ESO that directly schedules the observer bandwidth via load pressure, enabling faster disturbance estimation; (ii) a disturbance-adaptive constraint controller whose safety boundary is adjusted in real time using the observer‘s disturbance estimates, balancing tracking precision and safety; and (iii) a structured disturbance-separation architecture that reduces observer burden via model-based feedforward. By leveraging load pressure as a feedforward signal, this framework overcomes the latency inherent in error-feedback methods. Comparative simulations show that the proposed method outperforms conventional error-feedback methods by achieving a significant reduction in estimation error, as well as 2.3-times faster convergence, while ensuring both high tracking accuracy and strict constraint satisfaction. Full article

Running neural networks on battery-powered Internet of Things (IoT) sensor nodes is difficult because flash memory, SRAM, latency, and energy per inference are limited at the same time. Existing TinyML co-design methods usually improve model size or memory use, but runtime voltage–frequency control is often handled as a separate step. This separation limits energy saving because the power policy does not use the layer-wise compute profile of the final compressed model. We propose the Cross-Layer Resource Optimizer (CLRO), a three-stage resource optimization pipeline for TinyML inference on an ARM Cortex-M7 target. The first stage, Mixed-Precision Aware Pruning and Distillation (MPAD), assigns per-layer bit widths and pruning ratios using calibration-set sensitivity scores. The second stage, consisting of the Activation Lifetime-Aware Tensor Scheduler (ALTS), uses the compressed graph to find an execution order that reduces peak live static random-access memory (SRAM). The third stage, Reinforcement Learning-Based Dynamic Voltage and Frequency Scaling (DVFS-RL), trains a tabular Q-learning policy from the multiply–accumulate (MAC) utilization profile of the compressed and scheduled model. The learned voltage–frequency policy is stored as a small flash lookup table, so it adds no runtime decision cost during inference. We evaluate the CLRO on all four MLPerf Tiny tasks using an STM32H743ZI microcontroller with 512 kB SRAM and 2 MB flash. The CLRO reaches 91.7% image classification accuracy, 95.4% keyword-spotting accuracy, 89.6% visual wake words accuracy, and 0.913 anomaly detection AUC. The final deployment uses 198 kB flash and 174 kB peak SRAM, with 387 μJ energy per inference and 38 ms latency. Compared with the MCUNet baseline, the CLRO reduces energy by 58.1% and peak SRAM by 39% while keeping the same accuracy level. Full article

Limited by the computational performance limits of the CPU(Central Processing Unit), the traditional Spark architecture struggles to achieve high throughput and low latency under the dual pressure of a large data scale and real-time requirements in centralized control systems. This work uses a publicly available CNC(Computer Numerical Control) milling dataset as a functional validation proxy for time-series data processing, then extends validation to a large-scale synthetic power transmission grid dataset. Furthermore, Spark-GPU(Graphics Processing Unit) collaboration suffers from load balancing failure due to heterogeneous resource scheduling and communication overhead, thus failing to unleash its performance potential. This paper proposes a Spark-GPU fusion acceleration technology path. The path consists of three key components: first, it integrates the RAPIDS accelerator; second, it designs a GPU-aware partitioning and task co-scheduling strategy; and third, it optimizes the zero-copy data path. Together, these components realize an integrated collaboration of heterogeneous resources. Validation on real-world datasets yields the following results. In real-time aggregation scenarios, the proposed solution improves throughput by a factor of 3.7 over the pure CPU baseline and reduces end-to-end latency by 62%. Compared with the basic GPU solution, GPU utilization rises from 51.7% to 72.3%, representing a relative improvement of 39.8%. Furthermore, the solution meets industrial-grade high availability requirements. This research significantly improves the processing throughput and reduces end-to-end latency in typical centralized control scenarios, thus providing a feasible technical route for demanding concurrent centralized control scenarios such as electric power industry manufacturing with high real-time demands. Full article

Protected horticulture moves fragile pots, plug trays, seedlings, harvested products, and carriers through narrow, humid, and crowded spaces. Transport robots must therefore integrate locomotion, perception, localization, handling, placement, scheduling, and human–robot interaction rather than operate as simple carts. This structured narrative review reorganizes evidence from seedling transplanting, nursery operations, harvest support, manipulation, perception, and autonomous navigation around the complete transport chain: target recognition, pickup, loading, loaded navigation, docking, unloading or placement, payload protection, and workflow feedback. The synthesis covers mobile platforms, payload support, perception and localization, motion control, gentle handling, digital support, and fleet coordination. Three barriers remain: short laboratory tests rarely provide season-long evidence; many prototypes are too specialized for variable workflows; and benchmarks seldom combine motion accuracy, handling reliability, payload quality, and resilience. Progress will require modular platforms, robust sensing, payload-safe control, standardized interfaces, and closer co-design between robotics and horticultural operations. Full article

Unlike reactive systems, mechanical ventilation controlled by CO₂ concentration operates at a target efficiency that dynamically increases whenever the target CO₂ level is exceeded. This approach eliminates the typical ‘dead-time’ and prevents air quality degradation by ensuring the system adjusts its performance immediately in response to concentration changes. In this work, the study focuses on the development and evaluation of data-driven predictive models for near-term indoor CO₂ forecasting that can be integrated into pre-occupancy ventilation strategies, rather than designing a complete control scheme. Experimental data were collected over four months in a 48 m² smart laboratory configured as an open-plan office, where a heterogeneous IoT sensing architecture logged synchronized time-series measurements of CO₂ and microclimate variables (temperature, relative humidity, PM_2.5, TVOCs), together with acoustic noise levels and appliance-level energy consumption used as indirect occupancy-related signals. Raw telemetry was transformed into a 22-feature state vector using a structured feature engineering method incorporating z-score standardization, cyclic time encodings, multi-horizon CO₂ lags, rolling statistics, momentum features, and non-linear interactions to represent temporal autocorrelation and daily periodicity. The study benchmarks multiple regression paradigms, including simple baselines and ensemble methods, and found that an automated multi-level stacked ensemble achieved the highest predictive fidelity for short-term forecasting, with an Mean Absolute Error (MAE) of 32.97 ppm across an observed CO₂ range of 403–2305 ppm, representing improvements of approximately 24% and 43% over Linear Regression and K-Nearest Neighbors (KNN), respectively. Temporal diagnostics showed strong phase alignment with observed CO₂ rises during occupancy transitions and statistically reliable prediction intervals. Five-fold walk-forward cross-validation confirmed the temporal stability of these results, with top models achieving consistent R² values of 0.93–0.95 across Folds 2–5. These results demonstrate that, within a single-room university laboratory setting, historical sensor data from low-cost IoT devices can support accurate short-term CO₂ forecasting, providing a predictive layer that could support future proactive ventilation scheduling aimed at reducing CO₂ lag at the start of occupancy while avoiding unnecessary ventilation runtime. Generalization to other building types and occupancy profiles requires further validation. Full article

With the widespread adoption of electric vehicles (EVs), the deep integration of transportation and power grids has emerged as a significant trend. EV charging stations, acting as dynamic loads, present challenges to real-time power balance and economic dispatch in microgrids, while EVs serving as mobile energy storage units offer new opportunities for system flexibility. To address these issues, this paper proposes a hardware-in-the-loop (HIL) co-design method for vehicle–grid-integrated microgrid energy management systems, covering the entire workflow from simulation to embedded deployment. This method resolves the core challenges of multi-objective optimization algorithm deployment on embedded platforms (i.e., high computational complexity, strict real-time constraints, and heterogeneous communication protocol integration) via deployability analysis, hybrid code generation, real-time task restructuring, and consistency validation. A prototype microgrid system integrating photovoltaic panels, wind turbines, diesel generators, an energy storage system, and EV charging loads was built on the RK3588 embedded platform. An improved multi-objective particle swarm optimization (MOPSO) algorithm is employed to optimize operational costs. Experimental results verify the effectiveness of the proposed co-design method. Compared with traditional rule-based control strategies, the MOPSO algorithm reduces the total daily operating cost of the VGIM system by approximately 50%. After integrating vehicle-to-grid (V2G) scheduling, the operating cost is further reduced. In addition, this method ensures the consistency of algorithm functionality and performance during the migration from HIL simulation to embedded deployment, and the RK3588-based embedded system can complete a single optimization iteration within 60 s, which fully satisfies the real-time requirements of industrial applications. This work provides a feasible technical pathway for the reliable deployment of vehicle–grid-integrated microgrids in practical industrial scenarios. Full article

Satellite constellations are transforming space systems from isolated spacecraft into networked, software-defined platforms capable of on-orbit perception, decision making, and adaptation. Yet many of the existing AI studies remain centered on single-satellite inference, while constellation-scale autonomy introduces fundamentally new algorithmic requirements: learning and coordination under dynamic inter-satellite connectivity, strict SWaP-C limits, radiation-induced faults, non-IID data, concept drift, and safety-critical operational constraints. This survey consolidates the emerging field of on-orbit space AI through three complementary paradigms: (i) federated learning for cross-satellite training, personalization, and secure aggregation; (ii) multi-agent algorithms for cooperative planning, resource allocation, scheduling, formation control, and collision avoidance; and (iii) collaborative sensing and distributed inference for multi-satellite fusion, tracking, split/early-exit inference, and cross-layer co-design with constellation networking. We provide a system-level view and a taxonomy that unifies collaboration architectures, temporal mechanisms, and trust models. Full article

This paper reviews recent work on dynamic virtual power plants (DVPPs) using an Energy–Information–Market framework. It addresses the important problem of how DVPPs can support low-inertia power system operation and feeder-level stability under high renewable penetration. First, system-level studies on low-inertia operation and frequency control are used to frame quantitative requirements on rate of change of frequency, nadir, and quasi-steady-state limits. Second, energy-layer models are surveyed, including participation-factor-based DVPP controllers, grid-forming architectures, model-free frequency regulation, and robust frequency-constrained scheduling for allocating virtual inertia and fast frequency response (FFR) across distributed energy resource fleets. Third, information-layer and market-layer models are reviewed, covering stochastic and robust bidding, distribution locational marginal price-based clearing, peer-to-peer and community markets, privacy-preserving coordination, and emerging governance and cybersecurity schemes for DVPP participation. Across these strands, much of the literature remains centred on steady-state active and reactive power dispatch, with dynamic security enforced as constraints rather than formulated as verifiable and tradable services. This review identifies gaps in dynamic metrics and benchmarks, forecasting of available inertia and FFR capacity, market-physics co-design, multi-aggregator interaction, and experimentally validated DVPP implementations. These findings suggest that DVPPs can “sell stability” at the feeder level only through co-designed control, information, and market mechanisms and outline a research roadmap for this purpose. Full article

Low Earth orbit (LEO) satellite networks require real-time routing to cope with dynamic topology variations caused by continuous orbital motion. As an alternative to conventional routing approaches, deep reinforcement learning (DRL) has recently gained attention as an effective means for optimizing routing paths. To solve routing problems modeled as a grid-based Markov decision process (grid-based MDP), DRL methods such as CNN-based Dueling DQN have been proposed. However, these approaches are difficult to implement in practice. In particular, the substantial floating-point computation and memory traffic of CNN inference make real-time onboard inference challenging under the stringent power and resource constraints of satellite platforms. To address these constraints, this paper proposes an INT8 quantization and hardware–software co-design framework using heterogeneous SoC FPGA acceleration. We offload compute-intensive CNN inference to the programmable logic (PL), while the processing system (PS) orchestrates overall control and data movement, forming a collaborative PS–PL architecture. Furthermore, we integrate the NITI-style two-pass scaling with PS–PL exponent propagation to preserve end-to-end integer consistency without floating-point conversion. To demonstrate its practical onboard feasibility, we employ standard accelerator implementation choices—such as output-stationary scheduling and on-chip prefetching—and conduct an ablation study over independently tunable axes (PE array size and PS-side buffer reuse) to quantify their incremental contributions. Experimental results show that the proposed PS–PL cooperative scheme dramatically reduces computation time compared to a PS-only reference implementation on the same platform. Full article