The increasing complexity of modern urban traffic networks demands intelligent control strategies that can anticipate and adapt to dynamic traffic conditions. Model Predictive Control (MPC) is a framework that optimizes vehicle control by predicting future states and respecting real-time constraints, such as traffic signals at intersections. However, the computational complexity of MPC increases significantly with the number of decision variables and constraints, which is directly proportional to the length of the prediction horizon, creating a critical trade-off between control performance and computational efficiency. To address this challenge, this paper proposes an adaptive-horizon optimal driving (AHOD) bi-level optimization framework that incorporates a novel time-step discretization for real-time trajectory optimization and integrates it into a full traffic signal cycle. Unlike conventional MPC, which employs uniform time discretization leading to exponential growth in decision variables with horizon length, the proposed AHOD framework assigns finer time steps near signal phase transitions and coarser steps in the distant horizon, maintaining a fixed number of optimization nodes regardless of cycle length. The proposed framework comprises two controllers: the upper and lower controllers. The Upper controller employs finer resolution at critical times of signal change and coarser resolution in distant horizons, thereby reducing computational cost while maintaining prediction accuracy. The lower controller applies a practical MPC scheme to generate realtime control actions that are consistent with the long-term constraints of the upper controller. Simulation results demonstrate that the proposed framework achieves up to 17.6% fuel savings compared to traditional human driving and reduces computation time by approximately 61% compared to long-horizon MPC, while maintaining comparable control performance. The proposed framework enables real-time, cycle-aware predictive control for connected and automated vehicles (CAVs), and establishes a practical basis for embedding long-horizon prediction within an MPC-based trajectory-planning framework.

This paper presents a genetic algorithm (GA) tuned Mamdani type fuzzy logic control (FLC) framework for trajectory tracking of a quadrotor unmanned aerial vehicle (UAV) using a nonlinear rigid body model. The proposed architecture adopts a cascaded structure in which an outer loop position controller generates attitude and thrust references $(\phi_{\mathrm{ref}},\theta_{\mathrm{ref}},T_{\mathrm{ref}})$, while an inner loop attitude controller generates body torques $(\tau_\phi,\tau_\theta,\tau_\psi)$. Both loops employ a shared Mamdani fuzzy inference system with normalized inputs (tracking error and error-rate) and a normalized control output. The GA automatically tunes scaling gains $(K_e,K_d,K_u)$ across all axes to minimize a robust objective that averages tracking error, control effort, and constraint violations over multiple scenarios with mass uncertainty and wind disturbances. Simulation results on a three dimensional figure eight trajectory indicate that GA tuning can reduce position and attitude errors while respecting actuator saturation and tilt safety limits, demonstrating a practical route to performance enhancement without requiring a high fidelity aerodynamic model. The methodology leverages the interpretability of fuzzy rules and the global search capabilities of evolutionary optimization within a UAV modeling framework consistent with established quadrotor dynamics literature.

Rolling bearings are critical components of marine shafting power transmission systems, and accurate prediction of their vibration signal trends is essential for predictive maintenance. To address the limited adaptability of conventional time-series forecasting models under varying operating conditions and their insufficient ability to capture strong noise and abrupt changes, this study proposes a vibration signal prediction method that integrates particle swarm optimization (PSO) with an improved Informer model. PSO is used to adaptively optimize key Informer hyperparameters for different operating conditions, while a rolling time-window mechanism is introduced to enhance the capture of abrupt signal variations. In addition, a mixture of sparse attention (MoSA) encoder with a collaborative dense-head/sparse-head structure is designed to balance global temporal dependency modeling and local fault feature extraction. Experimental results on the Case Western Reserve University (CWRU) bearing fault dataset show that the proposed model outperforms Long Short-Term Memory (LSTM), Transformer, Informer, iTransformer, and Flowformer in terms of Mean Squared Error (MSE), Mean Absolute Error (MAE), and Root Mean Squared Erro (RMSE). The model achieves an MSE of 0.2015, which is 25.5% lower than that of the second-best iTransformer model. It also demonstrates robust performance under four different bearing operating states, confirming its adaptability to complex operating conditions. The proposed method provides a promising technical route for the predictive maintenance of rolling bearings in marine shafting systems.

Efficient coordination of heterogeneous mobile resources is essential for delivering large-scale urban services, such as sanitation, infrastructure inspection, or last-mile delivery. This study addresses the problem of scheduling aerial and ground service vehicles—unmanned aerial vehicles (UAVs) and mobile ground crews—to cover spatially distributed demand points under operational constraints. We formulate the task as a multi‑objective optimization problem that simultaneously maximizes service coverage, minimizes total completion time, and optimizes resource utilization while respecting safety, capacity, and time‑window restrictions. A hierarchical solution framework is proposed: global task allocation first assigns demand zones to vehicle types according to their capabilities, and local path planning then generates efficient routes for each agent. A dynamic re‑optimization mechanism adjusts schedules in real time when disturbances occur, such as resource depletion or environmental changes. The method is evaluated on scenarios of increasing scale (51, 113, and 212 demand points) that emulate urban public spaces. Results from ten repeated experiments show that the cooperative strategy achieves coverage rates (CRs) above 97% across all scales, reduces total operation time (TOT) by up to 33% compared with single‑mode operations, and improves resource efficiency by 21.10% and 47.40% Statistical analysis confirms the robustness of the improvements. The framework offers a scalable, resource‑aware solution for coordinating heterogeneous service fleets, with direct applicability to intelligent transportation systems, particularly in demand‑responsive urban services and multimodal fleet management.

Urban public transport systems are required to respond to pronounced temporal variations in passenger demand driven by calendar effects, weather conditions, and evolving mobility patterns. Reliable short-term demand forecasts have therefore become an important role in supporting operational planning and service management in large-scale transit systems. This study examines the daily ridership dynamics of the Transjakarta bus rapid transit system and evaluates the forecasting performance of three modeling approaches: seasonal autoregressive integrated moving average with exogenous variables (SARIMAX), multilayer perceptron (MLP), and a dynamic moving-window model. The analysis is based on 851 daily observations from January 1, 2023 to April 30, 2025, with rainfall, working days, and national holidays included as exogenous variables. Each model is estimated using a training dataset and evaluated on a hold-out test set over a 30-day forecasting horizon. Forecast accuracy is assessed using the mean absolute percentage error (MAPE). The results indicate that the MLP model achieves the highest forecasting accuracy, with a MAPE of 8.547%, while SARIMAX and the dynamic model yield higher error levels of 33.345% and 37.754%, respectively. The findings suggest that non-linear modeling approaches are better suited to capturing the complex and irregular demand patterns observed in daily urban bus ridership data. The study provides empirical evidence that can support short-term planning and demand-aware operational decision-making in urban public transportation systems.

The rapid integration of technology, with increasing speeds, has transformed vehicles into cyber-physical systems by connecting them to each other Vehicle-to-Everything (V2X), significantly expanding the attack surface and leaving them vulnerable to network-based threats. Current cyber intrusion detection systems (CIDS) exhibit performance degradation due to significant class imbalance, limited resilience against adversarial attacks, and insufficient interpretability for security-critical environments. To overcome the identified issues in this study, we propose Hierarchical Classifier-Agnostic Boosted Stacking for Network Intrusion Detection (HCABS-NID), a hierarchical classifier-agnostic boosted stacking architecture for network intrusion detection in connected device ecosystems. The proposed framework adds the Synthetic Minority Over-sampling Technique for Nominal and Continuous features (SMOTENC)-based adaptive class balancing to increase minority attack detection and TreeSHAP to make it multi-level interpretable. As a hierarchical stacking strategy, a two-layer structure includes heterogeneous learners together with meta-learning, calibrated with LightGBM, XGBoost, CatBoost, and TabNet to take advantage of the complementary decision boundaries. Extensive experiments performed on the benchmark dataset from University of New South Wales Network-Based 15 (UNSW-NB15) should enhance generalization performance. HCABS-NID achieved 98.20% accuracy, 97.10% macro F1 score, and 0.989 macro Receiver Operating Characteristic Area Under the Curve (ROC-AUC), in contrast to the latest community-based methods found in the literature. The proposed model achieves 3.40 ms average inference latency, satisfying the real-time processing requirement of the V2X safety systems. Indeed, other analysis architectures show the same 96.8% accuracy at 5% corruption, which underscores their practicality. The results validate that hierarchical ensemble learning, with adaptive imbalance management artificial intelligence (AI) mechanisms, provides a sound, interpretable, and ready-to-use intelligent transportation security package.

This study examines the energy efficiency of battery electric vehicles (BEVs) from a system-level and lifecycle perspective. The analysis highlights that the high mass of automotive battery packs, often an order of magnitude greater than conventional fuel tanks for equivalent energy storage, contributes to increased vehicle weight and may necessitate higher installed power to maintain performance. Consequently, larger battery capacities and associated increases in vehicle dimensions are commonly required, which in turn influences subsystems such as tires, resulting in higher rolling resistance. Operational advantages of BEVs are primarily observed in low-speed acceleration, while overall efficiency can be limited by frictional losses and auxiliary energy demands. Battery production is particularly energy-intensive, accounting for a substantial portion of the embodied energy in BEVs. Charge and discharge efficiencies vary with current rates and usage conditions. Slow charging, which takes approximately 10–12 hours, can reach an efficiency of around 95%, while fast charging from 25 to 75% of capacity over one hour typically achieves an efficiency of 85–90%. Discharge efficiency decreases from near 95% at low rates to roughly 70% at high C-rates (10–15 C). Despite the high efficiency of modern electric motors, including permanent magnet machines, system-level efficiency is further impacted by battery losses, power electronics, and auxiliary components. The effective energy delivery from the grid depends on the generation and distribution infrastructure; in contexts dominated by thermal power and limited renewable penetration, overall electricity efficiency may be around 40%, comparable to modern internal combustion engine vehicles, which operate at 40–50% thermal efficiency. Finally, current battery recycling technologies recover only 40–50% of materials and require additional energy input, highlighting limitations in end-of-life management. These findings suggest that BEVs may not always offer a clear energy efficiency advantage over conventional vehicles when evaluated on a comprehensive, life-cycle basis.

This study proposed a novel pin-level dynamic compensation strategy to combat the critical challenge of springback in the three-dimensional numerically controlled bending of ship hull plates. A collaborative prediction model combining convolutional and bidirectional recurrent networks (CNN-BiLSTM) was optimized using an improved metaheuristic algorithm, the Modified Sparrow Search Algorithm (MCSSA), to achieve millimeter-level precision in springback compensation. Based on the 225-pin independent control architecture, the system enabled real-time compensation with millisecond-level response ($\leq$ 50 ms) on standard industrial computing hardware, to overcome the limitations of conventionally fixed compensation methods. The optimized algorithm enhanced global search capability, population diversity, and convergence efficiency, hence yielding a prediction accuracy of RMSE = 4.41 $\times$ $10^{-5}$ mm. The integrated spatiotemporal learning framework effectively combined feature extraction, sequential modeling, and critical region emphasis, to achieve a test-set $R^2$ of 0.969. Industrial validation of the SKWB-1600 system demonstrated significant improvements in traditional stepwise approximation methods: (i) Post-compensation forming errors were reduced to 0.13–0.26 mm with a 47–62% improvement; and (ii) Curvature errors in high-stress zones were maintained within $\pm$ 0.02 mm, thus forming iterations decreased by 42% and energy consumption reduced by 35%. This certified pin-level dynamic compensation solution provides a new methodology for forming precision of complex curved ship hull plates under industrial conditions and establishes a technical paradigm for manufacturing related components requiring high precision and efficiency.