Journal of Transportation Engineering, Part A: Systems

The development of an efficient traffic incident detection system is essential for road safety risk warning and active safety control. Despite the fact that a large amount of traffic data can be collected from various detectors installed on expressways, the utilization rate of multisource data is still low, and the spatiotemporal traffic flow data need to be further mined. We propose a multilevel fusion method for the detection of traffic incidents, consisting of both data-level fusion and feature-level fusion. Accordingly, a macro and micro data-level fusion framework was developed, which created virtual detectors by converting video data into virtual loop data, thereby densifying the layout of the original loop detectors without increasing traffic facilities. Both sectional traffic flow data from loop detectors and single vehicle behavior data from video detectors were considered. Based on the fused data, several spatiotemporal variables are constructed to extract the spatiotemporal variation characteristics of traffic flow before and after an incident. The feature-level fusion framework made use of neural networks with a bidirectional encoding strategy to extract multisource features from various detectors and jointly encoded them to generate a comprehensive representation. Specifically, the inner networks extracted features from the multisource data, whereas the outer networks combined features from multiple single-source data to create a comprehensive representation for traffic incident detection. The results demonstrate that the model based on multisource data fusion was superior to single-source models. In addition, the performance of the proposed model was superior to that of the five common methods for detecting traffic incidents. Furthermore, the ablation experiments validated the advantages of key model components.

Improving the efficiency and safety of highway traffic relies heavily on accurately modeling the complex dynamics of traffic flow. This study aims to develop a novel method for modeling highway traffic flows by integrating data from multiple sources, such as roadside cameras, gantry cameras, and communication devices. The method leverages heterogeneous sensor data characteristics, temporal information, and spatial structures to achieve deep fusion of latent features at the sensor level. To minimize human intervention, a multistage training approach is employed, combining large-scale self-supervised learning with supervised fine-tuning, leveraging abundant unlabeled unstructured data such as monitoring videos recorded by roadside cameras and snapshots captured by gantry cameras, alongside limited accurate structured traffic flow data aggregated by communication devices from gantries and toll stations. We demonstrate the effectiveness and stability of the proposed method on a case study, the G92 ring highway around the Hangzhou Bay in Zhejiang Province, China, achieving the mean absolute percentage error of traffic flow within 7.5% and 8.3% for fixed and variable highway sensor networks, respectively. Ablation studies further demonstrate the significant improvement in predictive accuracy achieved by the designed self-supervised pretraining task. To summarize, our approach provides a promising solution for efficient and safe management of highway traffic flow, with potential applicability to real-world scenarios.

Abstract

Accurate and proactive lane-changing (LC) intention recognition can assist drivers in making LC decisions to improve driving safety. However, the mechanism of drivers’ LC decisions in dynamically changing environments is still not fully understood, which makes it difficult for advanced driver-assistance systems (ADASs) to make accurate LC decisions under different working conditions. To accurately capture the dynamic features before the generation of LC intention, relying on the multivehicle interaction capability of the connected environment, a LC intention recognition framework using graph theory to model the interaction relationship among multiple vehicles, i.e., the Multivehicle Interaction Dynamic Time Graph (MIDTG) framework, is proposed. First, the interaction relationship between LC vehicles and their surrounding vehicles in the connected communication range is modeled by graph theory. Second, the graph convolutional network (GCN) is used to extract spatial features of multivehicle interactions, and a long short-term memory (LSTM) neural network is used to learn the association of multivehicle interaction graphs in a time series. Finally, LC intentions are output through the Softmax function. The highD data set is used to validate the proposed model. Results show that the model can accurately extract the dynamic features of multivehicle interactions within a time window of 2.5 to 3.5 s, and the accuracy of LC intention recognition reaches 98%, which is an average improvement of 3.5% compared with other baseline models. The study provides a new way to model multivehicle interactions in the connected environment, which can be helpful for ADASs’ LC decision-making.

Practical Applications

The significance of this study is to provide guidance for advanced driver-assistance systems (ADASs) to better assist drivers in making lane-changing decisions. In this study, a novel multivehicle interaction modeling method based on graph theory is proposed, combined with an advanced deep learning method to extract the spatiotemporal characteristics of multivehicle interaction, and finally applied to lane-changing intention recognition. The results show that the proposed method achieves good lane-changing intention recognition accuracy and can accurately extract key information from multivehicle interactions. This will provide a new way to predict driving behavior in a connected environment. This study will be meaningful for traffic operators and can spur a new wave of innovative applications in ADASs, such as lane-changing alerts.

The advent of automated vehicles (AVs) and the sharing of these vehicles in traffic flow has raised researchers’ interest in low-cost solutions to traffic congestion, especially at intersections as points of traffic flow interlace. Also, as AVs have not been used commercially yet, it is impossible to analyze the effects of their sharing on traffic flow. However, different scenarios of AVs alongside human-driven vehicles (HDVs) can be modeled and simulated. By defining three levels of automated vehicles along with HDVs, this paper investigates the impact of AVs on intersection capacity by applying a series of hypotheses to the fundamental traffic flow formula. Different scenarios with an AV penetration coefficient of 10% to 70% showed an increase in intersection capacity from 10% to more than 50%. Finally, an innovative interactive framework between intelligent intersections and vehicles is presented to control intersection capacity and performance at of service level B with various percentages of AVs.