2021 Fiscal Year Final Research Report
Deep learning model for accident risk prediction
Project/Area Number |
20K14851
|
Research Category |
Grant-in-Aid for Early-Career Scientists
|
Allocation Type | Multi-year Fund |
Review Section |
Basic Section 22050:Civil engineering plan and transportation engineering-related
|
Research Institution | Ehime University |
Principal Investigator |
Tsubota Takahiro 愛媛大学, 理工学研究科(工学系), 講師 (00780066)
|
Project Period (FY) |
2020-04-01 – 2022-03-31
|
Keywords | Traffic accident / Traffic engineering / Deep learning / 畳込みニューラルネットワーク |
Outline of Final Research Achievements |
This study aims to build a traffic accident risk prediction model using a deep learning model and to identify input data that can contribute to improving the prediction accuracy of the model. A near-future traffic accident risk prediction model was built for an intercity expressway. The model is based on a convolutional neural network, and speed, traffic volume, and time occupancy (OCC) of vehicle detectors were used as inputs. In order to examine the impact of the quality of the input data, an analysis was conducted considering the time-varying components of the traffic conditions. The results showed that the time-varying components of traffic volume and OCC were effective in improving forecast accuracy. It was also confirmed that the output of the constructed model was generally consistent with the actual accident probability.
|
Free Research Field |
交通工学
|
Academic Significance and Societal Importance of the Research Achievements |
交通事故は物的・人的に多大な損失をもたらす為、事故の起こりやすさ、すなわち事故リスクを予測し、ドライバーへの情報提供や交通対策により事故の発生を未然に防ぐことが重要となる。本研究で構築した事故リスク予測モデルは近未来の事故リスクを高精度に予測する為、事故リスク低減を目指した交通管制に活用可能であると期待される。また、深層学習モデルの精度向上において、交通工学分野の知見が有用であることを示した点においても、同分野におけるAI技術の活用において重要な示唆を与えたといえる。
|