Jun 12, 2023
The student paper competition for the symposium "The Future of Traffic Management" in Vienna has closed. After reviewing the numerous submissions received from talented young researchers spanning across the globe, the six-member committee is happy to announce the names of the three winners.
Xuerun Yan and his co-authors, affiliated with Tongji University in Shanghai, emerged as the winners of the student competition for their outstanding paper on a cutting-edge simulation platform designed for evaluating truck platooning. The proposed platform has been successfully commercialized and put into use by the Shanghai Automotive Industry Corporation (SAIC) at the Yangshan Port, the second largest port globally. To validate the platform's credibility, a comparison was conducted against an actual field test, and additional sample tests were performed to assess truck platoon performance and the impact on mixed traffic. The thorough analysis of experimental results highlighted the need for upgrading existing platoon lane-change technologies to be compatible with high traffic demand scenarios. Furthermore, it emphasized the necessity of conducting localized and up-to-date assessments before permitting truck platooning.
Securing second place in the student paper competition, Chintaman Bari, a researcher at the Sardar Vallabhbhai National Institute of Technology in India, presented a paper focusing on optimizing the distribution between manual and electronic toll collection (ETC) lanes to maximize throughput when both vehicles, with and without ETC, are on the road. By considering factors such as approach volume, ETC penetration rate, and delay experienced in toll lanes, Bari devised different scenarios and developed a delay equation for estimating delays in the field. Additionally, he established delay-based Level of Service (LOS) thresholds specific to ETC lanes. This study will be useful to field practitioners to determine the number of lanes to be operated for efficient traffic management and to maintain prescribed LOS.
Junlan Chen from Southeast University and Monash University reached third place of the student paper competition. She and her co-authors engaged with imbalanced crash data that can lead to poor performance for data-driven methods. Commonly used data resampling methods are often inadequate as they only handle continuous variables and cannot account for the correlation between risk factors. Therefore, Junlan and her team proposed a novel approach that utilizes a deep generative model based on Conditional Tabular Generative Adversarial Networks (CTGAN) to generate synthetic crash data. The results of their study demonstrate that CTGAN-RU is comparable to other resampling methods, and the generated data is consistent in terms of prediction accuracy and statistical inference.
These exceptional papers will be honored and awarded during the Plenary Session 3 on Wednesday, June 28, at the conference. The student paper competition is generously sponsored by TM2.0.
Congratulations to all the winners!
