Performance Comparison of Dynamic Time Warping (DTW) and a Maximum Likelihood (ML) Classifier in Measuring Driver Behavior with Smartphones

The ubiquitous presence of smartphones provides a new platform on which to implement sensor networks for Intelligent Transport Systems (ITS) applications. Smartphone-based driving behavior monitoring has applications in the insurance industry, fleet management, driver training, and for law enforcement. In this paper we propose a Maximum Likelihood (ML) classifier to identify and classify the recklessness of driving maneuvers using the embedded sensors and GPS receiver of a smartphone. We compare the developed approach to the commonly used Dynamic Time Warping (DTW) based method. The solutions are both suitable for real-time applications, such as driver assistance and safety systems. An endpoint detection algorithm is used on filtered accelerometer and gyroscope data to find the start- and endpoints of driving events. The events are isolated with the endpoint detection algorithm are then classified using the DTW algorithm and an ML classifier. Results show that the ML classifier outperforms the DTW approach.