eduzhai > Applied Sciences > Engineering >

Data Augmentation of IMU Signals and Evaluation via a Semi-Supervised Classification of Driving Behavior

  • Save

... pages left unread,continue reading

Document pages: 7 pages

Abstract: Over the past years, interest in classifying drivers behavior from data hassurged. Such interest is particularly relevant for car insurance companies who,due to privacy constraints, often only have access to data from InertialMeasurement Units (IMU) or similar. In this paper, we present a semi-supervisedlearning solution to classify portions of trips according to whether driversare driving aggressively or normally based on such IMU data. Since the amountof labeled IMU data is limited and costly to generate, we utilize RecurrentConditional Generative Adversarial Networks (RCGAN) to generate more labeleddata. Our results show that, by utilizing RCGAN-generated labeled data, theclassification of the drivers is improved in 79 of the cases, compared to whenthe drivers are classified with no generated data.

Please select stars to rate!

         

0 comments Sign in to leave a comment.

    Data loading, please wait...
×