Struggling with microseismic signal classification in deep underground engineering? Researchers from Sichuan University developed the LRE-UDAF framework, a lightweight, noise-robust solution for cross-domain identification. It excels in resource-poor, noisy settings, boosts classification accuracy significantly across real projects, and paves the way for better disaster early warning in underground construction.

Researchers have evaluated how Vision Transformers and convolutional neural networks can support faster and more accurate defect detection in railway track fasteners, a key task for smart railway maintenance. The study focuses on enhancing Non-Destructive Evaluation, or NDE, by using pre-trained deep learning models and transfer learning to identify irregularities without damaging transport infrastructure.

Researchers have proposed a SENet-CNN-Transformer model for predicting electric vehicle charging duration, aiming to improve estimates of how long it will take a vehicle to charge from its current state of charge to a target state of charge. The method combines data enhancement, channel attention, convolutional neural networks, Transformer modeling, and transfer learning to address real-world data scarcity and nonlinear battery behavior.

Researchers have developed a feature-enhanced ensemble learning method for rapidly estimating the capacity of lithium-ion batteries using only a short partial discharge segment from the initial stage of testing. The approach is designed to support battery capacity grading in industrial settings, where conventional full-discharge procedures can take several hours per cell and create a bottleneck for manufacturing, pack assembly, and second-life screening.

Researchers have examined the challenge of detecting and classifying dynamic road obstacles for autonomous driving systems and presented a deep learning-driven convolutional neural network approach for the task. The review article focuses on mobile obstacles such as pedestrians, vehicles, bicycles, buses, trucks, cars, motorbikes, and animals, which can pose serious safety risks in complex driving environments.

Researchers have proposed a personalized longitudinal motion planning policy for intelligent vehicles that combines reinforcement learning with imitation learning. The approach is designed to reduce the gap between human driving behavior and automated vehicle decision-making by allowing a vehicle to adapt its longitudinal driving style to a target driver while still meeting performance requirements.

Researchers have developed an economical vehicle-side strategy for electric bus charging stations participating in vehicle-to-grid services, using reinforcement learning to optimize when and how long buses provide grid support. The study focuses on a real electric bus charging station in China?s Pearl River Delta region and evaluates how health-aware vehicle-to-grid operation can reduce lifecycle cost while extending battery life.