Rail surface defects typically serve as early indicators of railway malfunctions, which may compromise the quality and corrosion resistance of rails, thereby endangering the safe operation of trains. The timely detection of defects is essential to ensure the safe operation of railways. To improve the classification accuracy of rail surface defect detection, this paper proposes a rail surface defects detection algorithm based on MobileNet-YOLOv7. By integrating lightweight deep learning algorithms into the engineering application of rail surface defect detection, a MobileNetV3 lightweight network is used as the backbone network for YOLOv7 to enhance both speed and accuracy in complex defect extraction. Subsequently, the efficient intersection over union (EIOU) loss function is utilized as the positional loss function to bolster system resilience. Finally, the k-means++ clustering algorithm is applied to obtain new anchor boxes. The experimental results demonstrate the effectiveness of the proposed method, achieving superior detection accuracy compared with traditional algorithms.

Increasingly stringent regulations regarding vehicle emissions have contributed to the diversification of vehicle technologies, resulting in the increasing complexity of typical vehicles that make up a fleet. In order to investigate the real gas emissions of different typical vehicles, tests were conducted using a portable emission measurement system (PEMS) in Beijing and emission studies were conducted on eight light-duty passenger vehicles (LDPVs, including light-duty gasoline passenger vehicles and hybrid electric vehicles), eight heavy-duty passenger vehicles (HDPVs), and four light-duty trucks (LDTs). The results show that the emissions of relevant pollutants from LDPV meet the emission standard limits. The emission factors of CO₂, CO, NO_X, and HC of China IV and China V hybrid electric vehicles (HEVs) are much smaller than the emission standard limits and the emission factors of other vehicles, which have better emission reduction effects. Among LDPV, heavy-duty passenger vehicles (HDPVs), and LDT, the emissions of HDPV and LDT are extremely high. Emission characteristics vary on different types of roads, with the highest emission factors generally occurring on secondary roads. The micro-trip method was used to explore the influence of speed on emission factors. HEV are less sensitive to speed changes and can still maintain a low emission level at low speeds. The average speed and emission factors of HDPV in micro-trip has a strong correlation.

Diesel pallet trucks, a type of heavy-duty diesel trucks (HDDTs), have historically been a vital component in logistics and transport due to their high payload capacity. However, they also present significant challenges, particularly in terms of emissions which contribute substantially to urban air pollution. Traditional HDDTs emission measurement methods, such as engine bench tests and those used in laboratory settings, often fail to capture real-world emission behaviors accurately. This study specifically examines the real-world emission characteristics of diesel pallet trucks exceeding 30 t under varying loads (unloaded, half loaded, and fully loaded) and different road conditions (urban, suburban, and high-speed). Considering that data quality is the key to the accuracy of the scheme, this research utilized a portable emission measurement system (PEMS) to capture real-time emissions data of carbon dioxide (CO₂), carbon monoxide (CO), nitrogen oxides (NO_X), and total hydrocarbons (THC). Key findings demonstrate a direct correlation between vehicle load and emission factors, with the emission factors for CO₂, CO, and NO_X increasing by 39.5%, 105.4%, and 22.7%, respectively, from unloaded to fully loaded states under comprehensive operating conditions. Regression analyses further provide an emission factor prediction model for HDDPTs, underscoring the continuous relationship between speed, load, and emission rates. These findings provide a scientific basis for pollution control strategies for diesel trucks.

Communication-based train control systems (CBTCs) have been widely used as crucial systems in urban rail transit networks. CBTCs typically utizes different levels of symmetry structure according to different geographic deployments. While, in practice, CBTCs crashes have destroyed the transportation systems of the whole city level for many times. Based on the extended object-oriented Petri net (EOOPN), this paper proposes a vulnerability model and an evaluation procedure, which are capable of considering the vulnerability factors in both inner system level and equipment level. On the system level, it establishes a complex dynamic communication structure model among the distributed subsystems, while on the equipment level, it details the equipment changing state during train operation. The searching algorithm of EOOPN depicts possible failed paths of CBTCs via the token transition among train¬–ground communication EOOPN subnets. The vulnerability calculation is applied to the metro company’s in situ CBTCs to illustrate the effectiveness of the approach.

With the rapid expansion of high-speed railways, maintaining track structural health is vital for modern railway systems. Although deep learning has improved defect detection, models still face problems such as varying defect scales, severe background noise (e.g., lubricant residues and ferruginous oxidation), and irregular defect boundaries. To solve these problems, we introduce a new network named Rail-Adaptive-RCNN (RA-RCNN). It uses a Large Selective Kernel (LSK) backbone to dynamically adjust the Effective Receptive Field (ERF) for capturing periodic corrugation. We also added an Efficient Multi-Scale Attention (EMA) module that purifies features by suppressing noise without lowering dimensions. Finally, combining Scylla-IoU (SIoU) Loss with K-means clustering optimizes the regression of odd-shaped defects. Our experiments indicate that RA-RCNN reaches a mean Average Precision (mAP_0.5) of 86.2%, outperforming the baseline Faster R-CNN by 8.8%. Corrugation detection specifically reached 91.4%. With a processing speed of 26 FPS, this method effectively meets the practical needs of real-time automated railway maintenance.