Abnormal Driving Behavior And Collision Accident Identification Technology Research Based On Mobile Terminals

With the explosive growth in the number of used vehicles,all kinds of frequent traffic accident caused heavy casualties and property loss,the road traffic safety is more and more serious,and the whole society is concerned about the importance of traffic safety.Accident prevention and rescue in time after the accident has great significance.According to the survey,drivers themselves are the main factors of road traffic accidents.besides,Studies have shown that the driver is safer driving behavior under supervision;In addition,a large number of survey show that the number of casualties due to the relief is not timely far more than the number of casualties due to the accident itself.therefore,from the perspective of rescue in time after the accident,timely learned the vehicle collision accident is particularly important.As a kind of widespread used equipment,Smart phones have powerful data network transmission function,and also integrates equipments such as accelerometer sensor,gyroscope sensor and the gyroscope sensor which can acquire the outside world state parameters,So,the development of Smart phones’ sensors become a hot spot.In this paper,the vehicle state is divided into three categories: normal driving,abnormal driving and collision happened.and based on the data from on-board smart phones’ accelerometer sensor,gyroscope sensor and magnetic sensor,This paper proposes a abnormal driving detection algorithm and develops a system.Firstly,System collects accelerometer sensors’ data,gyroscope sensors’ data and magnetic sensors’ data of the on-board smart phones,and processes the collected data by rotating and filtering.Then,extracts the features of the data unit.and based on the driving behavior online analysis algorithm which is obtained by support vector machine(SVM),The system can identify a variety of driving behavior including right and left lane-changing,acceleration,deceleration,braking and normal driving.And by setting the driving frequency threshold in the time window to determine whether abnormal driving behavior happened.In the vehicle collision accident,the system according to the change range of mobile acceleration sensor data to identify the collision.Experimental results show that the proposed abnormal vehicle driving behavior and collision detection method can effectively identify the driving behavior which are defined in this paper,and open alarm when abnormal driving and collision happened.The paper analyzes the various factors may affect the result of the experiment,and aiming at the shortcomings of the proposed method,the research train of thought in the future is put forward.