Recognizing driving behaviour using smartphones

In this chapter at first we will present our methodology for recognizing driving patterns using smartphones, and then we will present in detail the android-based application we have developed to this end, which can monitor driving behavior. The latter can be achieved either using data only from the accelerometer sensor or using a sensor fusion method, which combines data from the accelerometer, the gyroscope and the magnetometer. We can recognize events like hard acceleration, safe acceleration, sharp left turn, safe right turn, sharp left lane change, etc. The application to improve the driving behavior of the driver, displays some hint messages to him after each bad-driving event. All the data from the trips (e.g., driving events that take place during a trip), are stored in a database and the driver has the opportunity to review and analyze them whenever he wants. We believe that organizing drivers in some form of a social network and involving them in a game-like procedure for promoting and rewarding the best driver among them, can motivate drivers to more secure driving customs.