Learning Risky Driver Behaviours from Multi-Channel Data Streams Using Genetic Programming

Feng Xie, Andy Song, Flora Dilys Salim, Athman Bouguettaya, Timos K. Sellis, Doug BradbrookAustralasian Conference on Artificial Intelligence 2013 · OpenReview

Abstract

Risky driver behaviours such as sudden braking, swerving, and excessive acceleration are a major risk to road safety. In this study, we present a learning method to recognize such behaviours from smartphone sensor input which can be considered as a type of multi-channel time series. Unlike other learning methods, this Genetic Programming (GP) based method does not require pre-processing and manually designed features. Hence domain knowledge and manual coding can be significantly reduced by this approach. This method can achieve accurate real-time recognition of risky driver behaviours on raw input and can outperform classic learning methods operating on features. In addition this GP-based method is general and suitable for detecting multiple types of driver behaviours.
BibTeX
@inproceedings{xie2013learning,
  author = {Feng Xie and Andy Song and Flora Dilys Salim and Athman Bouguettaya and Timos K. Sellis and Doug Bradbrook},
  title = {{Learning Risky Driver Behaviours from Multi-Channel Data Streams Using Genetic Programming}},
  booktitle = {Australasian Conference on Artificial Intelligence},
  year = {2013},
  url = {https://openreview.net/forum?id=avfRJbegYv},
}

Loading related content...

Figure