Driving style recognition using machine learning and smartphones

Background: The lack of real-time monitoring is one the reasons for awareness among drivers their dangerous driving behavior. This work aims to develop a driver profiling system where smartphone’s built-in sensors are used alongside machine learning algorithms classify different behaviors. Methods: We attempt determine optimal combination smartphone such as accelerometer, gyroscope, and GPS in order an accurate algorithm capable identifying events (e.g. turning, accelerating, or braking). Results: In our preliminary studies, we encountered some difficulties obtaining consistent events, which had potential add “noise” observations, thus reducing accuracy classification. However, after pre-processing, included manual elimination extraneous erroneous with use Convolutional Neural Networks (CNN), have been able distinguish about 95%. Conclusions: Based on results determined that proposed approach effective classifying turn will allow us driver’s behavior.