Research And Implementation Of Abnormal Driving Behavior Detection System Based On Intelligent Terminal

The problem of road traffic safety has always been the focus of people's concern.However,the existence of massive abnormal driving behavior has brought serious security risks to road traffic.The relevant data show that the abnormal driving behavior is generally caused by the motorist's bad driving habits,while effective supervision and reminding mechanism plays a significant role for improvement of motorist's bad driving habits.Smartphones own rich sensor resources and greater computing power,which are widely used in various fields.Based on the above research background,an abnormal driving behavior detection system of Android smartphone is designed to remind,supervise and improve motorist's bad driving habits.Firstly,this paper studies the various doped noise in the driving data obtained by smartphone,then analyzes the sources of data error and designs a five steps IIR low-pass filtering algorithm based on elliptic filter to pre-process the original data of motion sensor.The simulation results show that the filtering algorithm can eliminate the noise effectively.Secondly,this paper analyzes the features of the typical abnormal driving behavior in accelerometer and gyroscope data on each axis.The 20 features which can best characterize the typical abnormal driving behavior are extracted by experiment.The features are taken as the input vector of the BP neural network to implement the detection and classification of abnormal driving behavior.In addition,an average energy endpoint detection algorithm based on resultant acceleration of y-axis and z-axis is proposed to calculate the starting and ending points of candidate driving behavior events specific to the problem that duration of various abnormal driving is not fixed.The validity of endpoint detection algorithm is verified by simulation experiment.Finally,this paper designs and implements an abnormal driving behavior system based on Android smartphone.The system is mainly composed of detection module,recording module,summary module and display module.The performance of the system was tested in the experiment,the testing results show that the system has higher detection accuracy rate for five typical abnormal driving behaviors,which include rapid acceleration,rapid deceleration,sharp turn,emergency lane-change and weaving.