Machine Learning Techniques to Identify Unsafe Driving Behavior by Means of In-Vehicle Sensor Data

作者：

Highlights：

• In-vehicle sensor data are useful to identify unsafe driving behavior.

• A feed-forward network shows an accuracy above 90% in classifying driver’s behavior.

• An objective validation of machine learning for driving behavior is introduced.

• The method doesn’t use any tracking system which compromises the driver’s privacy.

摘要

•In-vehicle sensor data are useful to identify unsafe driving behavior.•A feed-forward network shows an accuracy above 90% in classifying driver’s behavior.•An objective validation of machine learning for driving behavior is introduced.•The method doesn’t use any tracking system which compromises the driver’s privacy.

论文关键词：Road safety,Driving behavior,Machine learning,Neural networks,Support vector machines

论文评审过程：Received 25 May 2020, Revised 15 January 2021, Accepted 28 February 2021, Available online 10 March 2021, Version of Record 31 March 2021.

论文官网地址：https://doi.org/10.1016/j.eswa.2021.114818