Smartphone Sensor Based Human And Vehicle Status Continuous Perception And Dangerous Driving Identification

Dangerous driving behavior seriously endangers road traffic safety.The current monitoring methods are mainly based on road equipment monitoring and traffic police spot check,which are difficult to identify dangerous driving accurately and timely due to their poor implementation and timeliness and low popularity rate.As a portable de-vice for users,smart phones have a very high penetration rate.Moerover,smart phones are rich in embedded sensors,which have been widely studied and applied to the iden-tification of dangerous driving behavior patterns.Based on the smart phone sensor to identify dangerous driving behavior,this paper has done the following three aspects of research:Firstly,multi-dimensional data based on the inertial sensor of the mobile phone was fused with Ground truth.The data determines the upper limit of the algorithm.The error of the inertial sensor itself and the dependence on the external Ground truth information are always the bottleneck for improving the accuracy of the recognition algorithm.This paper proposes a strategy that does not rely on other devices,and only carries out multi-dimensional data fusion based on the mobile phone inertial sensor to obtain the Ground truth information between the mobile phone,vehicle and inertial reference frame,and at the same time improve the accuracy and reliability of the mobile phone inertial sensor data.Secondly,the continuous perception of human and vehicle status based on smart-phone sensors.Based on the sensor data of the user's mobile phone and Ground truth information,we designed the human-vehicle state perception algorithm.On the one hand,according to the different motion modes of the user's mobile phone,this paper divides and identifies the mobile phone's time sequence state to perceive the user's dangerous driving state.On the other hand,the multi-dimensional information such as driving mode,speed and direction,etc.of the vehicle are continuously detected based on the perception of the vehicle status by smart phones.Thirdly,multi-task dangerous driving behavior recognition based on smart phone sensor.Dangerous Driving behavior is closely related to the state of the driver and the vehicle.Then we design a multitask identification algorithm based on the human-vehicle state information as input,then identifying the Distracted Driving,Speeding and Reckless Driving.The dangerous driving behavior recognition scheme in this project is based on the user's smart phone.The low-privacy inertial sensor is used to continuously perceive the human-vehicle state,which is not dependent on the vehicle equipment and insensitive to the individual characteristics of the user's mobile phone.In addition,this paper classifies.and defines dangerous driving behaviors and designs a multi-task recognition algorithm accordingly,which solves the problems existing in relevant researches and provides a novel strategy for the recognition of dangerous driving behaviors.