Multimodal Machine Learning Method for Predicting Road Congestion Using Bus Speed

View at publisher Summarise with VIRAT
Authors: Mikiko Sode Tanaka
Cited by: 0
References: 5

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

Currently, research and development of systems that simulate urban traffic on a digital twin platform, based on real-world traffic data, are actively underway to enable real-time monitoring of urban road traffic and to plan traffic congestion mitigation measures and optimal routes. To accurately simulate real-world traffic flow and predict future traffic patterns using a digital twin, high-precision traffic data, both temporally and spatially, is required. Therefore, research and development utilizing 5G and next-generation communication technologies (Beyond 5G) that enable high-volume data transmission are being pursued. However, this approach requires continuous collection of massive amounts of data, making it cost-prohibitive for implementation in smaller, rural cities. Therefore, this research proposes a multi-objective machine learning model that utilizes various data sources, such as bus speed and traffic congestion information displayed on digital maps, to construct a system for predicting traffic congestion levels. This method can create a digital twin with a small amount of data, making it a technology applicable to various regions, including smaller cities where collecting large amounts of data is difficult.

Fields of Study & Subjects

Traffic Prediction and Management Techniques Traffic control and management Transportation Planning and Optimization

Publication details

Journal / source
International Conference on Consumer Electronics
Pages
1–4
Publication date
3 February 2026
Language
EN