Aggressive Driving Detection Using Deep Learning-based Time Series Classification
| Autor: | Mounir Ghogho, Hakim Hafidi, Youness Moukafih |
|---|---|
| Rok vydání: | 2019 |
| Předmět: |
050210 logistics & transportation
business.industry Computer science Deep learning 05 social sciences Feature extraction 02 engineering and technology Machine learning computer.software_genre Session (web analytics) Field (computer science) Data modeling Aggressive driving 0502 economics and business 0202 electrical engineering electronic engineering information engineering 020201 artificial intelligence & image processing Artificial intelligence Time series business computer Intelligent transportation system |
| Zdroj: | INISTA |
| DOI: | 10.1109/inista.2019.8778416 |
| Popis: | Driver aggressiveness is a major cause of traffic accidents. Aggressive driving detection is an important application in the field of intelligent transportation systems (ITS). Developing systems capable of automatically detecting aggressive driving behavior should help improve traffic safety. In this paper we propose a novel solution to the problem of drivers' behavior classification based on a Long Short Term Memory Fully Convolutional Network (LTSM-FCN) to detect if a driving session involves aggressive behavior. We formulate the problem as a time series classification and test the validity of our approach on the UAH-DriveSet, a public dataset that provides a large amount of naturalistic driving data obtained from smartphones via a driving monitoring application. The proposed solution is compared to other deep learning and classical machine learning models for different processing time window sizes. It is shown that the proposed system outperforms the other methods in terms of the F-measure score, which reaches 95.88% for a 5 minutes window length. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|