A5068374815- Traffic control and management
- Transportation Planning and Optimization
- Traffic Prediction and Management Techniques
- Autonomous Vehicle Technology and Safety
- Advanced Neural Network Applications
- Traffic and Road Safety
- Transportation and Mobility Innovations
- Vehicle emissions and performance
- Video Surveillance and Tracking Methods
- Vehicular Ad Hoc Networks (VANETs)
- Human-Automation Interaction and Safety
- Robotics and Sensor-Based Localization
- Adversarial Robustness in Machine Learning
- Urban Transport and Accessibility
- Visual Attention and Saliency Detection
- Domain Adaptation and Few-Shot Learning
- Infrastructure Maintenance and Monitoring
- Vehicle Dynamics and Control Systems
- Anomaly Detection Techniques and Applications
- Human Mobility and Location-Based Analysis
- Explainable Artificial Intelligence (XAI)
- Privacy-Preserving Technologies in Data
- Indoor and Outdoor Localization Technologies
- Real-time simulation and control systems
- Risk and Safety Analysis
University of California, Los Angeles
2020-2025
Inner Mongolia University
2024-2025
Tsinghua University
2022-2024
Shenyang University of Technology
2024
Harbin Medical University
2024
Second Affiliated Hospital of Harbin Medical University
2024
Samueli Institute
2022-2024
Civil Aviation University of China
2024
Changchun University of Science and Technology
2024
Delft University of Technology
2020-2023
Employing Vehicle-to-Vehicle communication to enhance perception performance in self-driving technology has attracted considerable attention recently; however, the absence of a suitable open dataset for benchmarking algorithms made it difficult develop and assess cooperative technologies. To this end, we present first large-scale simulated perception. It contains over 70 interesting scenes, 11,464 frames, 232,913 annotated 3D vehicle bounding boxes, collected from 8 towns CARLA digital town...
Although Cooperative Driving Automation (CDA) has attracted considerable attention in recent years, there remain numerous open challenges this field. The gap between existing simulation platforms that mainly concentrate on single-vehicle intelligence and CDA development is one of the critical gaps, as it inhibits researchers from validating comparing different algorithms conveniently. To end, we propose OpenCDA, a generalized framework tool for developing testing systems. Specifically,...
This letter reports on a TIV DHW (decentralized and hybrid workshop) that explores the prospective influence of ChatGPT research development in intelligent vehicles. To assess update capabilities ChatGPT, we conducted tests involving both basic technically relevant questions. Our preliminary testing revealed ChatGPT's information can be updated corrected at one time, but it may take some time for changes to reflected responses, so not always possess latest knowledge regarding specific...
3D-LiDAR-based cooperative perception has been generating significant interest for its ability to tackle challenges such as occlusion, sparse point clouds, and out-of-range issues that can be problematic single-vehicle perception. Despite effectiveness in overcoming various challenges, per-ception's performance still affected by the aforementioned when Connected Automated Vehicles (CAVs) operate at edges of their sensing range. Our proposed approach called HYDRO-3D aims improve object...
In 2014, IEEE Intelligent Transportation Systems Society established a Technical Committee on 5.0 with the mission of promoting and transforming deployment advanced innovative technologies, especially Artificial Intelligence in transportation. This paper briefly summarizes our main research findings over last decade. Foundation Models, Scenarios Engineering, Operating have been identified as directions for development next-generation intelligent transportation systems.
Modern perception systems of autonomous vehicles are known to be sensitive occlusions and lack the capability long perceiving range. It has been one key bottlenecks that prevents Level 5 autonomy. Recent research demonstrated Vehicle-to-Vehicle (V2V) cooperative system great potential revolutionize driving industry. However, a real-world dataset hinders progress this field. To facilitate development perception, we present V2V4Real, first large-scale multi-modal for V2V perception. The data...
Most object detection methods for autonomous driving usually assume a consistent feature distribution between training and testing data, which is not always the case when weathers differ significantly. The model trained under clear weather might be effective enough on foggy because of domain gap. This paper proposes novel adaptive framework weather. Our method leverages both image-level object-level adaptation to diminish discrepancy in image style appearance. To further enhance model's...
Deep learning has been widely used in intelligent vehicle driving perception systems, such as 3D object detection. One promising technique is Cooperative Perception, which leverages Vehicle-to-Vehicle (V2V) communication to share deep learning-based features among vehicles. However, most cooperative algorithms assume ideal and do not consider the impact of Lossy Communication (LC), very common real world, on feature sharing. In this paper, we explore effects LC Perception propose a novel...
Advances in Single-vehicle intelligence of automated driving has encountered great challenges because limited capabilities perception and interaction with complex trafic environments. Cooperative Driving Automation (CDA) been considered a pivotal solution to next-generation driv-ing smart transportation. Though CDA attracted much attention from both academia industry, exploration its potential is still infancy. In companies tend build their in-house data collection pipeline research tools...
This perspective paper delves into the concept of foundation intelligence that shapes future smart infrastructure services as transportation sector transitions era Transportation 5.0. First, discussion focuses on a suite emerging technologies essential for intelligence. These encompass digital twinning, parallel intelligence, large vision-language models, traffic simulation and systems modeling, vehicle-to-everything (V2X) connectivity, decentralized/distributed systems. Next, introduces...
In this paper, we present an overview and background on speed harmonization (SH). This paper reviews a number of representative studies that designed traffic control algorithms based variable limits, ramp metering, connected vehicle, or automated vehicle for SH. We summarize fundamental mechanisms, algorithms, evaluation results these studies. investigate the opportunities brought by portion vehicles communicating with each other using new technologies. also due to some having control....
Bird's eye view (BEV) semantic segmentation plays a crucial role in spatial sensing for autonomous driving. Although recent literature has made significant progress on BEV map understanding, they are all based single-agent camera-based systems. These solutions sometimes have difficulty handling occlusions or detecting distant objects complex traffic scenes. Vehicle-to-Vehicle (V2V) communication technologies enabled vehicles to share information, dramatically improving the perception...
Existing multi-agent perception algorithms usually select to share deep neural features extracted from raw sensing data between agents, achieving a trade-off accuracy and communication bandwidth limit. However, these methods assume all agents have identical networks, which might not be practical in the real world. The transmitted can large domain gap when models differ, leading dramatic performance drop perception. In this paper, we propose first lightweight framework bridge such gaps for...
Existing multi-agent perception systems assume that every agent utilizes the same model with identical parameters and architecture. The performance can be degraded different models due to mismatch in their confidence scores. In this work, we propose a model-agnostic framework reduce negative effect caused by discrepancies without sharing information. Specifically, calibrator eliminate prediction score bias. Each performs such calibration independently on standard public database protect...
Recent advancements in Vehicle-to-Everything communication technology have enabled autonomous vehicles to share sensory information obtain better perception performance. With the rapid growth of and intelligent infrastructure, V2X systems will soon be deployed at scale, which raises a safety-critical question: how can we evaluate improve its performance under challenging traffic scenarios before real-world deployment? Collecting diverse large-scale test scenes seems most straightforward...