Paper: "UIT-VehPredTraj: The Pioneer Dataset for Vehicle Trajectory Prediction Problem in Vietnamese Traffics"
Paper link: UIT-VehPredTraj
Authors:
- Chau Tan – 20520926 – KHDL2020: Main author.
- Ngo Duc Vu – 20520950 – KHDL2020: Co-author.
- Nguyen Tran Anh Duc – 20521198 – KHDL2020: Co-author.
Supervisor: M.S. Do Trong Hop
Summary of the paper:
Predicting the trajectory of vehicles plays a crucial role in the field of intelligent transportation systems, autonomous driving, and advanced driver assistance systems. This article documents the development process of the UIT-VihPredTraj dataset, where advanced methods such as YOLOv7 for accurate object detection and DeepSORT for precise object tracking were utilized. Using these techniques, we effectively captured and recorded the precise coordinates of vehicles in the landscape, laying the foundation for a comprehensive vehicle trajectory dataset, specifically tailored for Vietnamese traffic.
We also conducted a series of experiments using various deep learning models, including LSTM, GRU, CNN-LSTM, CNN-GRU, ResNet, and Transformer, to predict the future trajectories of vehicles. After analysis, the results showed that CNN-LSTM outperformed other methods, demonstrating superiority in accurately predicting the vehicle trajectories based on the collected dataset. Through this effort, we not only aim to advance in the field of vehicle trajectory prediction but also provide a valuable resource for the development of innovative automotive applications in Vietnam.
"We would like to express our gratitude to Mr. Do Trong Hop for accompanying and guiding our group throughout the research and publication process of this scientific paper at the national conference."
VNICT 2023 is the 26th National Conference on "Selected Issues in Information Technology and Communications" organized by the Institute of Information Technology - Vietnam Academy of Science and Technology and the Technical University - Logistics Department of the Ministry of Public Security. The conference took place in Bac Ninh on October 5-6, 2023, with the collaborative participation of the Club of IT Schools and Institutes in Vietnam (FISU). The conference is an annual forum for researchers, implementers, educators, and managers in the field of information technology and communications to exchange academic knowledge, share application experiences, and seek collaboration. The conference particularly encourages young scientists, researchers, and postgraduate students to present and exchange their research and study results. It also promotes the exhibition of scientific and technological products and the application of research results into practice.
For more details, visit: https://www.facebook.com/UIT.Fanpage/posts/pfbid0uLeznBTnHNP...
Hai Bang - Media Collaborator, University of Information Technology
Nhat Hien - Translation Collaborator, University of Information Technology