Accurate estimation of state of charge (SOC) is crucial for operation performance promotion of lithium-ion batteries. However, the variations of temperature and loading current directly impact the estimation accuracy of SOC. To fully account for these influences, this study proposes a hybrid compensation model and exploits an advanced algorithm for high-performance SOC estimation. First, a fractional-order model (FOM) is constructed to delineate the electrochemical behaviors of batteries with higher accuracy, compared with traditional integral-order model (IOM). Then, the relationship among discharge rate, temperature, and available capacity is explored, and a capacity compensation model is established via the random forest (RF) algorithm. Based on the trustworthy parameter identification and capacity recognition, the SOC is estimated by the adaptive H-infinity filter (AHIF) to fully cope with the model and operation condition variations raised by different temperatures and loading currents. By this manner, the presented method enhances the robustness to parameter uncertainty and modeling errors and promotes the estimation accuracy of SOC in wide temperature range. The experimental results highlight that compared with the traditional IOM and adaptive extended Kalman filter (AEKF), the proposed method can highly boost the temperature adaptability, convergence speed, and estimation accuracy of SOC.

Traffic accident severity prediction is essential for dynamic traffic safety management. To explore the factors influencing the severity of traffic accidents on mountain freeways and to predict the severity of traffic accidents, four models based on machine learning algorithms are constructed using support vector machine (SVM), decision tree classifier (DTC), Ada_SVM and Ada_DTC. In addition, random forest (RF) is used to calculate the importance degree of variables and the accident severity influences with high importance levels form the RF dataset. The results show that rainfall intensity, collision type, number of vehicles involved in the accident and road section type are important variables influencing accident severity. The RF feature selection method improves the classification performance of four machine learning algorithms, resulting in a 9.3%, 5.5%, 7.2% and 3.6% improvement in prediction accuracy for SVM, DTC, Ada_SVM and Ada_DTC, respectively. The combination of the Ada_SVM integrated algorithm and RF feature selection method has the best prediction performance, and it achieves 78.9% and 88.4% prediction precision and accuracy, respectively.

Railroad tunnel linings may have voids filled with water or air, or areas that lack compaction due to construction processes. These structural defects are hard to detect and require regular inspections to ensure safety. Existing CNN-based inspection systems that employ ground penetrating radar (GPR) imagery sometimes cannot fully assess the defects and ensure the structural integrity of the linings. This study develops an internal defect evaluation system based on an innovative computer vision architecture named YolaTC (You Only Look at Tunnel Coefficients) for monitoring railroad tunnel linings. First, YolaTC is equipped with a novel backbone that has multiple receptive fields available to strengthen multi-granularity feature representation and enrich semantic information in the feature space. Then, a new strip pooling module is developed and incorporated to model long-range context dependencies, improving the segmentation of long but narrow defect regions. Second, the system integrates a novel skeleton extraction-based Deterioration Level Evaluation (DLE) method that can calculate the actual defect length to quantify the level of void degradation. Finally, extensive comparative experiments conducted on a customized dataset established with GPR imagery demonstrate that the proposed system can effectively convert GPR images into valuable information with high accuracy, rapid processing speed, and good adaptation capability in complex tunnel structures.

With their advantages of high experimental safety, convenient setting of scenes, and easy extraction of control parameters, driving simulators play an increasingly important role in scientific research, such as in road traffic environment safety evaluation and driving behavior characteristics research. Meanwhile, the demand for the validation of driving simulators is increasing as its applications are promoted. In order to validate a driving simulator in a complex environment, curve road conditions with different radii are considered as experimental evaluation scenarios. To attain this, this paper analyzes the reliability and accuracy of the experimental vehicle speed of a driving simulator. Then, qualitative and quantitative analysis of the lateral deviation of the vehicle trajectory is carried out, applying the cosine similarity method. Furthermore, a data-driven method was adopted which takes the longitudinal displacement, lateral displacement, vehicle speed and steering wheel angle of the vehicle as inputs and the lateral offset as the output. Thus, a curve trajectory planning model, a more comprehensive and human-like operation, is established. Based on directional long short-term memory (Bi–LSTM) and a recurrent neural network (RNN), a multiple Bi–LSTM (Mul–Bi–LSTM) is proposed. The prediction performance of LSTM, MLP model and Mul–Bi–LSTM are compared in detail on the validation set and testing set. The results show that the Mul–Bi–LSTM model can generate a trajectory which is very similar to the driver’s curve driving and have a preferable generalization performance. Therefore, this method can solve problems which cannot be realized in real complex scenes in the simulator validation. Selecting the trajectory as the validation parameter can more comprehensively and intuitively reflect the simulator’s curve driving state. Using a speed model and trajectory model instead of a real car experiment can improve the efficiency of simulator validation and lay a foundation for the standardization of simulator validation.

Traffic congestion remains a significant challenge in urban management, with traditional fixed-cycle traffic signal systems struggling to adapt to dynamic traffic conditions. This paper proposes an adaptive traffic signal control method based on a Graph Neural Network (GNN) and a dynamic entropy-constrained Soft Actor–Critic (DESAC) algorithm. The approach first extracts both global and local features of the traffic network using GNN and then utilizes the DESAC algorithm to optimize traffic signal control at both single and multi-intersection levels. Finally, a simulation environment is established on the CityFlow platform to evaluate the proposed method’s performance through experiments involving single and twelve intersection scenarios. Simulation results on the CityFlow platform demonstrate that G-DESAC significantly improves traffic flow, reduces delays and queue lengths, and enhances intersection capacity compared to other algorithms. In single intersection scenarios, G-DESAC achieves a higher reward, reduced total delay time, minimized queue lengths, and improved throughput. In multi-intersection scenarios, G-DESAC maintains high rewards with stable and efficient optimization, outperforming DQN, SAC, Max-Pressure, and DDPG. This research highlights the potential of deep reinforcement learning (DRL) in urban traffic management and positions G-DESAC as a robust solution for practical traffic signal control applications, offering substantial improvements in traffic efficiency and congestion mitigation.