摘要
随着车辆保有量不断上升,人们对驾驶安全性行为的关注越来越多。行车数据隐藏了驾驶员的驾驶行为特征,所以对行车数据进行挖掘并分析驾驶员的驾驶行为成为研究热点之一。论文通过对车载设备收集到的行车数据进行分析和处理,提取驾驶员的驾驶特征,提出一种多层有监督训练与学习微调的DBN分类模型,以用于研究驾驶行为。最后,论文与基础DBN和BP神经网络进行实验对比,结果表明改进后的DBN分类性能较好。
With the increasing number of vehicles,people pay more and more attention to driving safety.Driving data hides the characteristics of drivers'driving behavior,so mining and analyzing driving data has become one of the research hotspots.Based on the analysis and processing of vehicle data collected by vehicle equipment,this paper extracts driver's driving characteristics,and proposes a multi-layer DBN classification model with supervised training and fine-tuning learning,which can be used to study driving behavior.Finally,the experimental comparison with the basic DBN and BP neural network shows that the improved DBN has better classification performance.
作者
黄丽蓉
潘雨青
HUANG Lirong;PAN Yuqing(School of Computer Science and Communication Engineering,Jiangsu University,Zhenjiang 212013)
出处
《计算机与数字工程》
2020年第8期1958-1964,共7页
Computer & Digital Engineering
关键词
驾驶行为分析
数据挖掘
DBN
特征提取
driving behavior analysis
data mining
DBN
feature extraction
