1 May 2024 in IEEE Transactions on Intelligent Vehicles
As an important component of intelligent transportation systems, collaborative perception has made great progress in recent years. In collaborative perception, multiple devices rely on communication to interact and transfer perception information and use complementary information to overcome occlusion and long-distance perception problems. Although there have been some overviews and analyses of the development status and architecture design in this field, there has not been a comprehensive review of the specific task division and goals in perception and communication. In this paper, to facilitate the quick location of specific directions or tasks in vehicle-road collaboration, we first analyze the timing and role of each task in the workflow of cooperative perception. Then, based on perception goals, we divide perception tasks into detection, segmentation, and tracking and discuss communication tasks in terms of open issues such as delay robustness, communication bandwidth, positioning errors, countermeasures against attacks, and privacy preservation computation and comprehensively review the research in each task. We summarize research in downstream tasks of collaborative perception, which are important for safe decision-making and efficient operation of intelligent transportation. We summarize the domain-relevant datasets and discuss future research trends and challenges.
21 August 2023 in IEEE Transactions on Computational Social Systems
Driving behavior classification is an important direction in the field of social transportation systems and advanced driving assistance system (ADAS), which has attracted more and more attention in recent years. An accurate driving behavior classification algorithm plays a great role in traffic safety, energy saving, and other fields. In this article, we propose a novel vehicle speed and image fusion for driving behavior classification network (SIFDriveNet), which classifies driver behaviors into normal driving, aggressive driving, and drowsy driving. Our method has the following key advantages. First, in the research of driving behavior classification, we are the first to introduce a 2-D image with rich roadside information and convert speeds into a 2-D spectrogram expressing the time–frequency characteristics of speeds through short-time Fourier transform (STFT) while unifying the data space of image information and speed information. Second, we propose a tensor fusion method based on weight decomposition to fully fuse the vectors of the two modalities. This method maps the tensor outer product results to the low-dimensional space through weight decomposition and has a low computational cost while maintaining the fusion effect of the tensor outer product. In addition, we evaluated our model on the public UAH-DriveSet and compared it with the most advanced model. Experimental results show that our model has a better performance, and F1-score is 97.9% on all roads. Especially on the secondary road, our F1-score is 99.4%. Also, our model has strong generalization, and we have reached 99.3% F1 in distracted driving multimodal dataset. In addition, the inference speed reaches 411 FPS, enabling real-time needs. The code is available on https://github.com/alu222/SIFDriveNet .
16 March 2022 in Measurement
5 July 2023 in IEEE Transactions on Intelligent Vehicles
Vehicle re-identification is the task of identifying the same vehicle in different environments and from different angles and cameras. It is more challenging than re-identification of humans: 1)small differences between vehicles of the same model make it difficult to capture their subtle characteristics; 2)vehicles of different types and colors may have similar characteristics from different viewpoints or external conditions. To address these challenges, we propose a TVG-ReID network, using a Transformer network to enhance features extracted from a CNN backbone network. A vehicle knowledge graph transfer method(Vehicle-Graph) is proposed, which treats each vehicle as a node in a graph, where simple information is transmitted through edges to constrain the distance of the nodes in a metric learning manner. Experiments on two vehicle re-identification datasets demonstrate the good performance of our proposed model.
28 July 2025 in IEEE Transactions on Geoscience and Remote Sensing
In recent years, methods leveraging global context modeling for continuous road extraction from remote sensing images have garnered significant attention. These methods can effectively address the issues of road discontinuity and missed detection caused by complex background interference, but they generally suffer from high computational complexity. Therefore, we propose CGCNet, a simple yet effective road extraction network based on global context modeling. CGCNet adopts an encoder–decoder architecture with a compact global context-aware block (CGCB) embedded in the center part to capture long-range dependencies among road segments. This block effectively enhances the model’s global modeling capability and significantly reduces computational complexity using a compact representation of the embedded Gaussian nonlocal block (NLB). Furthermore, we introduce SW-XJU, the first remote sensing road dataset constructed explicitly for the unique landscapes of Western China. Extensive experiments show that CGCNet achieves a superior balance between efficiency and accuracy compared with the state-of-the-art methods, and the extracted roads exhibit stronger connectivity. The source code and dataset are available at https://github.com/LPeng625/CGCNet.