Combining Sparse Modeling and Data Assimilation for Human Flow Prediction to Capture Behavioral Irregularities

2019 Fiscal Year Final Research Report

• Project/Area Number: 17K12979
• Research Category: Grant-in-Aid for Young Scientists (B)
• Allocation Type: Multi-year Fund
• Research Field: Social systems engineering/Safety system
• Research Institution: Shizuoka University
• Principal Investigator: Sudo Akihito 静岡大学, 情報学部, 講師 (80588369)
• Project Period (FY): 2017-04-01 – 2019-03-31
• Keywords: 機械学習

Outline of Final Research Achievements

In this study, we developed a machine learning model as a basis for a highly accurate behavioral prediction model that reflects the irregular movement of people in the city. The technical key to solving this problem is a nonlinear model that can discover a small number of feature values that can be predicted with high accuracy, while our previous studies have not been able to predict individual people's behavior to a practical level. In this study, we aim to construct a method of neural networks that extends sparse modeling in a nonlinear manner.

Free Research Field

情報通信

Academic Significance and Societal Importance of the Research Achievements

本研究で構築した手法により、交通量のリアルタイム推定とプライバシー保護を両立できる技術が確立できたと考える。実応用の際には、少数の観測地点で交通量を取得しておき、その交通量データを用いて都市全体の交通量を推定すればよい。本研究の結果は、GPSデータの収集が行われている国であれば適用でき、交通混雑の予測、消費者行動の把握、インフルエンザやエボラ出血熱等の感染経路の特定、大災害時の救援部隊や物資の配分といった実応用のほか、複雑な人の動きの理解という学術的な関心にも寄与する。