Font Size: a A A

Driving Anger Identification And The Risk Analysis

Posted on:2020-09-07Degree:MasterType:Thesis
Country:ChinaCandidate:R Y WangFull Text:PDF
GTID:2392330626453435Subject:Traffic and Transportation Engineering
Abstract/Summary:PDF Full Text Request
Driving anger is an important driver factors for traffic accidents.In order to reduce the occurrence of driving anger,improve traffic safety,study driving anger characteristics and identification methods,and its impact on traffic safety,this paper conducts driving simulation experiments to select driving anger indicators include eye movements and physiological parameters,establishes driving anger identification model,and analyzes mandatory lane changing under driving anger.(1)Driving simulation experiments are conducted.Driving anger experiment plan and process are designed,using eye tracker and physiological detector to collect drivers' eye movement and physiological data.Anger emotion are inspired by videos before experiments,this paper designs four kinds of traffic scenes to inspire driving anger,assesses the drivers' anger level during the experiments,collects eye movement,ECG,GSR,vehicle data.(2)Driving anger indicators are selected.Using wavelet transform denoising to denoise the raw data collected by device,this paper analyzes the significance of the eye movements and physiological indicators under driving anger and stable emotion,extracts four eye movement indicators,six ECG indicators and two GSR indicators related to driving anger.Principal component analysis is used to reduce the dimension of the 12 indicators and extracts 5 main components.(3)Driving anger recognition model based on BP neural network is constructed.In the experiment,451 effective samples are obtained,and 371 experiment samples are selected for neural network model training including 196 stable emotion samples and 175 angry emotion samples.The remain 40 stable emotion and 40 angry emotion samples are used for model verification,the accuracy of stable emotion recognition is 82.5%,the accuracy of driving anger emotion recognition is 80%,the accuracy of overall emotion recognition is 81.25%,which proved that the neural network driving anger recognition model has high accuracy.(4)Drivers' behavior under driving anger are analyzed.The paper analyzes vehicle operating parameters under driving anger,compares vehicle running speed,acceleration,brake,throttle and steering wheel angle in anger and stable emotion,and proposes a mandatory lane changing risk evaluation method considering velocity,lane change position and traversable space,analyzes mandatory lane changing risk under driving anger emotion.
Keywords/Search Tags:Traffic Safety, Driving Anger, Driving Simulation, Physiological data, Driving Anger Identification, Lane Changing Risk
PDF Full Text Request
Related items