Synthetic Human Flow Generation Based on Large-scale Movement History for Urban Areas

Bibliographic Information

Other Title
  • 都市を対象とした大規模移動履歴に基づく疑似人流データ生成手法

Search this article



There is a growing demand for traffic and infection simulations for use in urban planning and research. On the other hand, the widespread use of wearable devices has made it possible to collect a large amount of user location history with high accuracy, and it is expected that this data will be used for simulation. However, it is difficult to collect location histories for the entire population of a city, and detailed data that can reproduce trajectories is expensive. In addition, such personal location histories contain private information such as addresses and workplaces, which restricts the use of raw data. Therefore, methods to generate synthetic movement trajectories have been actively studied, but they require a large amount of teacher data and a detailed movement history from which movement trajectories can be obtained. Using real-world movement history data, this method models users' movement and stay tendencies with unsupervised learning, and generates synthetic human flow data at the city level. The key points of this method are that it uses unsupervised learning for modeling, does not require special labeling of the movement history data, and uses only frequently collected GPS data. Using this method, we generate synthetic human flow data without private information. Evaluation experiments confirmed that this method can approximate urban human flows with spatially more granular generated data than conventional methods.


Details 詳細情報について

Report a problem

Back to top