Accelerometer and Gyroscope Synthetic Data Calculation based on Driver Smartphone GPS

Determining driver behavior during road traffic is an actively researched task, with numerous systems aimed at identifying dangerous maneuvers, detecting driver fatigue, and monitoring the driver’s condition. The simplest way to obtain data for studying driver behavior is to collect vehicle movement data using a smartphone that utilizes built-in sensors such as GPS, accelerometers, and gyroscopes. These sensors provide comprehensive data on the car’s speed, location, acceleration, and angles of deviation, enabling in-depth research of driver behavior. However, it is not always possible to obtain accelerometer and gyroscope data. This paper proposes an algorithm for calculating accelerometer and gyroscope synthetic data based on GPS that provide possibilities to classify driver behaviour based on these synthetic data for devices that do not have accelerometer / gyro-scope sensors. The method involves constructing quaternions to describe rotations, followed by conversion to Krylov-Euler angles. We evaluated proposed method on the dataset that has been captured in-the-wild during driving vehicles by several drivers. Developed software that implements the proposed method is available as open-source.