• PDF version [PDF: 766KB]

East Nippon Expressway Co., Ltd. (Headquarters: Chiyoda-ku, Tokyo, President and Chief Executive Officer: Hiroshi Hirose, NEXCO EAST) and NTT DoCoMo, Inc. (Headquarters: President and Chief Executive Officer, Tokyo): Kazuhiro Yoshizawa (hereinafter referred to as DOCOMO) is a demographic created using the mechanism of the mobile phone network. ^*1 (Hereinafter referred to as demographics) and NEXCO EAST 's past traffic congestion records and regulatory information, etc. are used to conduct a traffic congestion prediction demonstration experiment using "AI traffic congestion prediction" developed by DoCoMo using artificial intelligence (AI). First in a Japanese company Expressway ^*2, CA Tokyo Wan Aqua-Line Expressway (Aqualine) will start from December 2, 2017 (Saturday).

This demonstration is based on turnout in Boso Peninsula zone of noon of the day in field trials period, Aqua line from 14:00 to 24:00 In-bound line ^*3 It predicts traffic congestion (in the direction of Kawasaki) and distributes the information on the NEXCO EAST customer website "DraPla". By avoiding times when the aqua line is congested, we aim to improve customer satisfaction and revitalize the surrounding area.

"AI Congestion Prediction" is a technology developed by NTT DOCOMO that predicts congestion through artificial intelligence (congestion prediction model) that learns and patterns the relationship between demographics and congestion. NTT Group's AI "corevo®" The technology to compose. "AI Congestion Prediction" takes into account the demographic data of the day, so it is possible to accurately predict sudden congestion caused by weather or events.

Prior to this demonstration test, the accuracy of prediction by ``AI traffic jam prediction'' was estimated based on the traffic congestion on the aqualine In-bound line from January 2015 to April 2017 and the demographic statistics held by DOCOMO for the period. Review ^※Four Carried out, overlooked rate of traffic congestion prediction of 10km or more ^※Five Is a conventional congestion forecast calendar ^*6 Compared to 6% for “AI traffic jam prediction”, it was 1%.

In this demonstration test, from December 2, 2017 (Saturday) to March 31, 2018 (Saturday) (scheduled), the traffic congestion prediction of the aqua line In-bound predicted by "AI traffic congestion prediction" Results and "Yorutoku" coupon information that can be used during traffic jam forecast time ^*7 Will be distributed at "DraPla" to disperse traffic. Also, from February 2018, we plan to add a function to send notifications of traffic congestion prediction results using the "DraPla" application. Based on this information, we will verify changes in customer behavior and the effect of reducing traffic congestion.

In this year NEXCO EAST is developing "safe, secure, comfortable and convenient expressway service" in the new medium-term management plan. In addition, docomo is working on solving social issues with its partners in a new medium-term strategy 2020 "beyond declaration". Both companies will further strengthen the solution of traffic problems through prediction of congestion based on "AI congestion prediction" utilizing artificial intelligence.

1. In this experiment, "AI traffic jam prediction" technology is used to predict traffic jams based on real-time demographics (research and development) that accelerate mobile spatial statistics. The demographics used in this experiment show the number of people in each group by area and attribute, and do not include any information that can identify an individual customer. Therefore, this demographic does not reveal your behavior to others. The demographics used in this experiment comply with the Mobile Spatial Statistics Guidelines.
   Mobile spatial statistics guidelines
2. November 30, 2017 (Thursday) NEXCO EAST survey
3. CA Tokyo Wan Aqua-Line Expressway is a Expressway from Kawasaki City, Kanagawa Prefecture to Kisarazu City, Chiba Prefecture. In-bound line is the direction of Kawasaki in Kanagawa prefecture.
   CA Indicates the route number (numbering) of Expressway etc.
4. Based on the actual traffic congestion data from January 2015 to April 2017 and the population information of the entire Boso Peninsula during the period, a study was conducted to evaluate the prediction accuracy of AI traffic congestion prediction from May 2017 to September 2017. It was carried out during the period.
5. The overlooked rate of traffic congestion prediction is the percentage of days when traffic congestion actually occurred, although it was predicted that there will be no traffic congestion during the period.
6. The congestion forecast calendar is a congestion forecast created and published by NEXCO EAST on past congestion records.
7. "Yorutoku" coupons are Kisarazu Outlet and Kisarazu outlets on Saturdays, Sundays, and holidays from 15:00 to 20:00, which is the time of aqualine traffic jam Suisui project implemented by NEXCO EAST, Umihotaru, Kisarazu City, and Mitsui Outlet Park. It is a great deal coupon that can be used at shops in the city.

• corevo is a registered trademark of Nippon Telegraph and Telephone Corporation. (http://www.ntt.co.jp/corevo/)

Click here for the special site for this demonstration test