"Visualization of congestion in trains" Tokyu and Hankyu, Tokyo Institute of Technology perform demonstration experiment

Tokyu Corporation and Hankyu Corporation have commenced a demonstration experiment on the visualization of congestion in trains in collaboration with the Tokyo Institute of Technology. The accuracy of the "congestion degree analysis technology in trains," developed by Tsujimoto Laboratory of the School of Environment and Society, Tokyo Institute of Technology, was verified. The research group aims to realize a service that visualizes train congestion information and provides it prior to boarding.

In the demonstration experiment, the Bluetooth signal of a passenger's smartphone is acquired by the congestion analysis device installed at the station, and the congestion situation is analyzed using AI on the cloud. To improve the accuracy of AI analysis, it was tuned by combining the congestion situation and measuring it with a high-speed camera from the platform of the station. All of these data are acquired without including information that can identify the terminal or individual.

Specifically, for the period of January 17^th to February 28^th, Tokyu Corporation installed a congestion analyzer and a high-speed camera on the inbound platform of the Den-en-toshi Line, Komazawa-Daigaku Station (Setagaya-ku, Tokyo). Once this technology is established, the routewill be expanded to the real-time congestion situation currently provided through the "Tokyu Line app" on certain Den-en-toshi Line vehicles; simultaneously, accumulating these data can afford a better understanding of the distribution of tendency values.

For the period of January 12^th to March 31^st, Hankyu Corporation installed a congestion analyzer, high-speed camera (Nakatsu Station only), and congestion level display signage (two units at Juso Station) on the outbound platforms of Nakatsu Station and Juso Station on the Kobe Main Line. From February to early March, a congestion level display signage was installed at Juso Station. Information regarding the congestion status will be provided based on the analysis data collected by the equipment installed at the station up to that point. In addition, the congestion status is compared before and after the installation of the congestion level display signage, and the extent to which the provision of information is associated with the behavior of passengers is analyzed. It is expected that commercialization will be realized via a venture from the Tokyo Institute of Technology.