DRIVER BEHAVIOUR ANALYSIS USING ANDROID APP

Abstract

Texting-while-driving (TD) is one of the fatal accident resulting behaviours of drivers. Till date many interesting systems and mobile phone applications have been designed to help to detect or combat TD. However, for a practical TD detection system, one of the key property is its capability to distinguish between driver’s mobile phone usage pattern and behaviour from passengers. As of now all the existing solutions are dependent on user’s manual input. Also they utilize specific GPS devices to determine whether a mobile phone is at driver’s location and in a particular position. In this paper, we propose a method which is able to detect TD automatically without using any extra devices thus a cost efficient method. The idea is very simple: when a user; (specifically Driver) is composing messages or misusing Smartphone while driving apart from regular usage, the Smartphone embedded sensors (i.e. gyroscopes, accelerometers, and GPS) collect the associated information such as touch strokes, holding orientation and vehicle speed. This information will b recorded and then be analyzed and categorised to see whether there exist some specific patterns similar to current recorded one. Extensive experiments were conducted by different persons and in different driving scenarios. The results show that our approach can achieve good detection accuracy with low false positive rate and thus focusing more on passenger’s safety on road. This system is infrastructure free and provides high accuracy. Also this method does not access the content of messages and therefore is privacy-preserving