The automatic detection of foreign objects between platform screen doors (PSDs) and metro train doors significantly affects personnel and property safety and maintains the train’s normal operation. However, some existing works only determine the presence of foreign objects but cannot indicate their categories. Besides, although deep-learning-based object detection algorithms can indicate the presence and categories of foreign objects, most of them only harness the information in region proposals, ignoring global contextual information. Furthermore, their performance comes at the considerable cost of computational complexity, and leading cannot be well deployed in the metro environment. To address these issues and better implement foreign object detection (FOD), we present You Only Look Once-Transformer (YOLO-Former), a simple but efficient model. YOLO-Former is accomplished based on YOLOv5 through the following procedure. First, the vision transformer (ViT) is introduced for dynamic attention and global modeling, thereby solving the problem that the original YOLOv5 only utilizes information in region proposals and has insufficient ability to capture global information. Second, the convolutional block attention module (CBAM) and Stem module are used to improve feature expression ability further and reduce floating point operations (FLOPs). Finally, we design various variants with different widths and depths to meet every need. Experiments on the foreign object detection dataset (FODD) and PASCAL VOC dataset demonstrate that YOLO-Former-x consistently outperforms other state-of-the-arts with significant margins (0.5 to 11.3 mean average precision, mAP, on FODD and 0.6 to 13.6 on PASCAL VOC dataset). Last but not least, YOLO-Former-x maintains real-time processing speed (27.32 and 28.17 frame per second, FPS, on TITAN Xp).

Cooperative Adaptive Cruise Control (CACC) systems can significantly improve traffic safety and roadway capacity utilizing short following gaps of vehicles enabled by inter-vehicle communications. However, due to merging processes occurring at freeway merging areas, existing CACC operation approaches are generally not applicable and the operation will have to revert back to Adaptive Cruise Control (ACC) or human-driven mode, which in turn will result in a capacity drop. This paper proposes an optimal trajectory optimization strategy for Connected and Automated Vehicles (CAVs) to cooperatively carry out mainline platooning and on-ramp merging. Firstly, a control framework of the CACC is adopted for a longitudinal control of CAVs, which helps individual CAVs to join platoons and to maintain platoon operations. Secondly, to ensure smooth lane-changing executions while achieving stable platoons, an optimal controller that considers lane-changing motivation of merging vehicles and impact of merging on platoons, is proposed. Third, a Legendre pseudo-spectral algorithm is applied to transform the controller into a simpler nonlinear programming problem and to efficiently solve it. Simulation assessments of the proposed method are conducted at both individual vehicle level and traffic-flow level. At the individual vehicle level, the proposed method has the potential to improve the traffic safety without compromising fuel consumption and emissions compared with unoptimized feasible schemes. At a traffic-flow level, an online evaluation platform is implemented, and a typical freeway on-ramp area is studied. The simulation results have demonstrated that the proposed controller provides significant improvements in terms of efficiencies in traffic operations.

This paper comprehensively discusses the cooperative communication and computation of vehicular system. Based on the cooperative transmission, an stochastic model of vehicle-to-vehicle (V2V) communication reliability is established using probability theory. Furthermore, the computation reliability is defined as a new metric for computation offloading, and a vehicle computational performance evaluation model is also established. In order to effectively compute the required data, we combine V2V communication and vehicle computing to further characterize the coupling reliability of cooperative communications and computation systems. In addition, we propose a virtual queue model that combines queue length and vehicle privacy entropy to optimize partitioning. Finally, considering the amount of processing data and cut-off time of vehicle applications, we establish the optimal partition model of vehicle computing with the goal of maximizing the coupling reliability, and propose the coupling-oriented reliability calculation for vehicle collaboration using dynamic programming methods. Simulations show that the proposed scheme outperforms traditional approaches in terms of coupling reliability and completion rate. In addition, the allocation between local computing and data offloading is controlled by the server's privacy perception of collaboration events.

Metros are usually built and added on the basis of a completed bus network in Chinese cities. After the metro construction, it is faced with the problem of how to adjust and optimize the original bus lines based on the new metro system. This research mainly proposes a bus line optimization method based on bus and metro integration. In the consideration of the geographical space, the cooperation and competition relationship between bus and metro lines is qualitatively introduced according to the geographical location and service range of metro (800 m radius) and bus (500 m radius) stations. The competition and cooperation indexes are applied to define the co-opetition relationship between bus and metro lines. The bus line optimization model is constructed based on the co-opetition coefficient and Changsha Metro Line Number 2 is chosen as a case study to verify the optimization model. The results show that the positive competition, efficient cooperation, and travel efficiency between metro and bus has been significantly enhanced after optimization. Moreover, this paper provides a reasonable reference for public transport network planning and resource allocation.

Real-time prediction of vehicle trajectory at unsignalized intersections is important for real-time traffic conflict detection and early warning to improve traffic safety at unsignalized intersections. In this study, we propose a robust real-time prediction method for turning movements and vehicle trajectories using deep neural networks. Firstly, a vision-based vehicle trajectory extraction system is developed to collect vehicle trajectories and their left-turn, go straight, and right-turn labels to train turning recognition models and multilayer LSTM deep neural networks for the prediction task. Then, when performing vehicle trajectory prediction, we propose the vehicle heading angle change trend method to recognize the future move of the target vehicle to turn left, go straight, and turn right based on the trajectory data characteristics of the target vehicle before passing the stop line. Finally, we use the trained multilayer LSTM models of turning left, going straight, and turning right to predict the trajectory of the target vehicle through the intersection. Based on the TensorFlow-GPU platform, we use Yolov5-DeepSort to automatically extract vehicle trajectory data at unsignalized intersections. The experimental results show that the proposed method performs well and has a good performance in both speed and accuracy evaluation.