计算机与数字工程Issue(9)：2180-2184,2200,6.DOI:10.3969/j.issn.1672-9722.2019.09.015
基于径向基神经网络的驾驶人无意识车道偏离识别模型∗
Identification Model of Driver's Unconscious Lane Deviation Based on RBF Neural Network
摘要
Abstract
For the vehicle active safety system in the identification of vehicle movement is difficult to distinguish between the road and the driver's unconscious lane deviation. In the real environment to collect the steering wheel angle,yaw rate and lane dis?tance when lane change and driver's unconscious lane deviation are collected. Based on RBF neural network,a model of distinguish?ing lane change and driver's unconscious lane departure is established. In order to further improve the overall recognition rate of the model,the weights and threshold parameters of the neural network are optimized by normalization,principal component analysis and genetic algorithm. Through the training and testing of the optimized neural network model,the results show that when the time window is 1.8s. The overall recognition rate of the optimized neural network is 90%,which is 92% for lane change recognition rate, 88% for the unconscious lane departure recognition rate,to meet the requirements of the vehicle active safety system.关键词
lane departure/neural network/principal component analysis/genetic algorithmKey words
lane departure/neural network/principal component analysis/genetic algorithm分类
信息技术与安全科学引用本文复制引用
孟妮,山岩..基于径向基神经网络的驾驶人无意识车道偏离识别模型∗[J].计算机与数字工程,2019,(9):2180-2184,2200,6.基金项目
陕西省自然科学基金项目(编号:2016JQ5096)资助. （编号:2016JQ5096）
