In recent years, widespread Internet use and digitalization have greatly changed people’s behavior as consumers. People can now access products and services from all over the world without leaving their homes, making their lives significantly more convenient. Moreover, the coronavirus outbreak further accelerated the digitalization of the society. On the other hand, this momentum of digitalization has created a new challenge through commercial areas in cities losing their liveliness as a result of its effect on the consumption activities and the movement of people in physical spaces.

In this context, major changes are required in the way cities are built. Cities are increasingly reliant on commercial activities to support the local economy, and it is becoming more important to develop the areas around train stations as a city hub and commercial center. Effective city management can have a positive impact on the flow of people and their behavior as consumers, which can directly lead to the revitalization of the city.

Following various research and initiatives undertaken by both academia and business communities to create a lively and a bustling city, our research group focuses on the flow of people. We use Digital Twin technology to analyze actual data such as the flow of people, transportation, sales, and energy, and visualize the city management in detail as well as through human behavior. Based on the results of the data analysis, we will develop a solution for optimizing city management by simulating the behavioral paths of people and other factors that affect city management and managing the flow of people. Specifically, our research group aims to build a simulator and an application that can predict and navigate human behavior as well as their purchasing activities. These can be used to identify the challenges facing commercial areas and commercial facilities for businesses such as railway companies and real estate developers. In the end, we will strive to maximize the potential of cities by recommending strategies for implementing optimal measures to attract people to the city.

Conducted demonstration tests by installing beacons mainly on the commercial facilities around Odakyu Ebina Station. This demonstration test has made it possible to understand the behavioral characteristics of human flow and visualize clusters based on similarities in behaviors by analyzing actual human flow data in the city. Our goal is to use this research as a foundation for the social implementation of the human flow navigation system, through understanding the characteristics of human flow behavior in cities, verifying the effects of measures by cluster, and predicting the effects of event measures, etc.
Also, captured and visualized macro-level data on behavioral characteristics of inter-regional migration by combining beacons data with human flow sensing data such as ticket gate data and GPS data.

Classified visitors into clusters based on their behavioral similarities by analyzing human flow data and built a behavior reproduction model targeting each cluster.
Tried to develop a model that can reproduce more realistic behaviors such as “migratory behavior in commercial areas” or “strolling behavior”, etc. by considering the situation and status, such as time of day and traffic mode (rail, car, bicycle, walking), and resident/non-resident on the actual data of human flow acquired through the demonstration test for “1. Understanding the current situation”.
Conducted a growth potential assessment on each visitor cluster by setting characteristic indicators (e.g., migration indicators, frequency of visits, etc.) depending on the level of contribution to the “Increase in human flow that contributes to add value to a city”.
Conducted a growth potential assessment on each visitor cluster by setting characteristic indicators (e.g., migration indicators, frequency of visits, etc.) depending on the level of contribution to the “Increase in human flow that contributes to add value to a city”.

Conducted a behavior reproduction simulation by generating agents given the “behavior and decision-making model according to the situation and status” and “behavior scheduling model” with parameters for each cluster as per the scenario.

Reproduced incentive measures on the simulator, such as holding customer-attracting events and disseminating information by utilizing signages or beacons, which are expected to induce the flow of people and estimated the effectiveness of each incentive measure. The estimated results are expected to be utilized for designing and suggesting effective incentive measures in detail based on matching for optimal city management.
In order to test the long-term transition of the city, we defined scenarios with varying parameters such as human flow and cluster compositions on a given day in the future to reproduce the human flow in a given city. We aim to use this information to formulate a long-term growth strategy for cities by simulating the change in flow of people and the liveliness of the commercial area according to changes in resident-visitor ratio, the transformation of commercial facilities, and the city’s infrastructures, etc.

The value this research is expected to create is to provide new perspectives in city planning and city management. The Next-Generation AI City Simulator can predict the effectiveness of each incentive measure in detail for inducing flow of people, aiming to suggest an optimal measure for city management based on the simulation results.
By reconstructing incentives that can influence flow of people, such as holding events and disseminating information by signages or beacons, as well as estimating their effectiveness, it becomes possible to suggest concrete incentives. Through the induction of human flow, it is expected to contribute to the revitalization of the city, promote commercial activities, and enhance the value of the city.

The simulator can simulate the long-term change of the city. Specifically, a scenario is set up in which parameters such as human flow and cluster composition are varied on a given day in the future. The results of the simulation are then used to predict how the human flow in the city will change. This information is valuable for cities to formulate long-term growth strategies and as city plans that can adapt to the long-term changes in the city.

This research project is expected to change conventional city management into data-driven city management by visualizing and modeling human flow and simulating human flow throughout a city. This will lead to the revitalization and sustainable development of the entire city.