Predict if AI is at home To reduce the burden of redelivery and carry out proof experiment July 9 17:35
A demonstration experiment of a system in which AI = artificial intelligence predicts whether or not residents are at home and reduces the burden of redelivery of courier drivers from the collected data of electricity meters in Yokosuka City, Kanagawa Prefecture I will be told.
This demonstration experiment was conducted by the University of Tokyo, Sagawa Express, and the Japan Data Science Research Institute, which handles data analysis, etc., and the mechanism was introduced at a press conference held in Tokyo on the 9th.
As a result, AI will analyze in real time the data collected from the electricity meters installed in approximately 200 households in Yokosuka City that agreed to cooperate with the experiment, and then determine whether the residents are at home.
It also predicts whether you will be at home for a few hours from the accumulated data and displays the most efficient delivery route on the driver's smartphone.
According to a simulation conducted by Professor Noboru Koshizuka of the University of Tokyo at the university, we were able to reduce the proportion of deliveries in the absence of people by about 90%.
The demonstration test is scheduled to start in September this year, and research groups and others aim to put it into practical use in two years, based on the results.
Professor Koshizuka said, "I would like to advance the research so that it can be put to practical use nationwide by taking advantage of Japan's major feature that electricity meters are installed in almost all households."