WCSE 2020 Summer
ISBN: 978-981-14-4787-7 DOI: 10.18178/wcse.2020.06.009

Traffic Congestion Analysis Based on Deep Neural Networks

Di Zang, Xiaoke Qu, Yang Fang

Abstract— Traffic congestion analysis is a research hotspot in intelligent transportation system (ITS). However, existing research methods focus on real-time or short-term traffic congestion determination and prediction. As the time interval increases, noise and sudden abnormalities contained in traffic data increase. How to predict congestion state throughout the next day is still a challenging task that has not been well solved. In this paper, we propose a multi-branch congestion model (MC) based on the multi-branch speed prediction model (MSP) to realize long-term traffic congestion prediction. Experiments based on Shanghai elevated highways show that our method is robust and has better performance.

Index Terms— traffic congestion analysis, speed/congestion prediction, deep neural networks

Di Zang, Xiaoke Qu, Yang Fang
Department of Computer Science and Technology, Tongji University , CHINA

[Download]


Cite: Di Zang, Xiaoke Qu, Yang Fang, " Traffic Congestion Analysis Based on Deep Neural Networks " Proceedings of 2020 the 10th International Workshop on Computer Science and Engineering (WCSE 2020), pp. 52-56, Shanghai, China, 19-21 June, 2020.