The Design And Implementation Of Novel Early Warning System For Aggressive Driving Detection Based On Smartphone

The increasing amount of vehicles on road lead more traffic accidents.According to the statistic survey from the foreign researchers,the main cause of traffic accidents is aggressive driving,such as sharp turn,sharp snap,which will endanger the drivers.In order to address this problem,some researchers are working on Steering Pattern Detection System which can notify the users of aggressive driving themselves.But there is a drawback that this kind of system needs additional device to detect the driving activity.Current mobile phones always contain many built-in sensors.In the meanwhile,some mobile Apps such as AutoNavi and Baidu Maps could provide much traffic information,but they are always lack of alert of surrounding aggressive driving.In this thesis,we present an early warning system for aggressive driving activities which combines sensors with urban computing framework,this system can indicate whether the users are driving in the sections where the accidents likely occur.The main contributions of our work is as below:1.We present an aggressive driving detection algorithm using mobile phone sensors.The calculation combines the angular velocity of the vehicle,front steering angle and the voice of turning.2.We implement the early warning of aggressive driving based on urban computing.By analyzing the sensor data provided by users,the system can notify the aggressive driving in the current section of road.3.We conduct the experimental evaluation.The evaluation demonstrates the efficiency and accuracy of our system.Compared with previous systems,we can achieve a high accuracy and robustness.