1593 papers
This study demonstrates that strategically retiring electric vehicle batteries 1–2 years earlier can transform 40–50% of currently unreusable units into valuable second-life energy storage resources, significantly enhancing circularity and environmental sustainability through a proposed dynamic trade-in framework.
This study demonstrates that mobile point-to-point speed enforcement is operationally feasible in Queensland, revealing that vehicle re-identification rates are primarily constrained by the number of road exits rather than camera separation distance, while also highlighting significant discrepancies between spot and average speed measurements.
This paper introduces CPFxDRL, a novel airfoil optimization framework that couples Deep Reinforcement Learning with Constructive Piecewise Functions to create a self-intersection-free, low-dimensional search space that outperforms traditional CST methods in convergence speed and geometric fidelity while highlighting the critical need for high-fidelity CFD validation to prevent reward hacking.
This paper demonstrates that an optimized Artificial Neural Network-based MPPT algorithm, trained on PVGIS-SARAH3 data, significantly outperforms traditional Perturb and Observe and Incremental Conductance methods by achieving 98.7% tracking efficiency and rapid convergence for real-time PV power optimization.
This paper introduces the BanglaVowelDataset, comprising 120 recorded air-conducted (AC) and 120 synthetic bone-conducted (BC) Bangla vowels from ten speakers, to address the lack of such resources and facilitate research in noise-robust speech processing, recognition, and speaker identification.
This study utilizes a three-dimensional finite element model to demonstrate that an aluminum die-cast metal matrix composite brake drum for a three-wheeler vehicle can safely withstand extreme fatigue loading for 67,500 cycles.
This paper proposes a continuous mid-course Weapon Target Assignment framework that integrates dynamic reassignment with trajectory guidance to enhance the autonomy, reactivity, and robustness of air-defense systems against coordinated swarm attacks, utilizing the Hungarian method for efficient real-time optimization in many-on-many scenarios.
This study utilizes a hybrid global/local optimization algorithm coupled with high-fidelity CFD simulations to design a foldable box-wing drone configuration that achieves a 16% improvement in aerodynamic efficiency and a 22% increase in the combined lift–efficiency objective compared to the initial design.
This paper proposes a wide area protection method for AC networks with HVDC-connected inverter-based resources that utilizes PMU data to identify suspected buses and employs superimposed current polarity analysis at line terminals to accurately detect and locate faults while distinguishing them from power swings, as validated through PSCAD/EMTDC simulations on a modified IEEE 39-Bus system.
This study proposes an enhanced hyperspectral oil spill detection framework that integrates 3D-CNNs with anomaly-aware feature fusion using PCA and Isolation Forest, achieving superior accuracy and robustness in distinguishing oil spills from look-alike phenomena compared to baseline methods on the HOSD dataset.