基于实时车流信息的电动汽车未来行驶工况预测  

Prediction of future driving conditions of electric vehicles based on real-time traffic flow information

作　　者：张毅 黄韬一 刘寅童 ZHANG Yi;HUANG Taoyi;LIU Yintong(School of Vehicle Engineering,Chongqing University of Technology,Chongqing 400054,China;Chongqing Changan Automobile Co.,Ltd.,Chongqing,400020,China)

机构地区：[1]重庆理工大学车辆工程学院,重庆400054 [2]重庆长安汽车股份有限公司,重庆400020

出　　处：《重庆理工大学学报（自然科学）》2023年第11期93-102,共10页Journal of Chongqing University of Technology：Natural Science

基　　金：重庆市教委青年基金项目(KJQN202001105)。

摘　　要：在全球能源短缺和污染加重的背景下,新能源汽车领域关键技术成为研究热点。然而纯电动汽车剩余续航里程的不确定性,严重地影响了纯电动汽车的进一步推广。因此,精确地预测车辆未来能耗以确定其剩余行驶里程,具有重大意义。本研究基于实车在环实验平台,将测试车辆的历史行车数据OBD和同时收集的百度API实时车流信息经过预处理和特征参数的提取,再使用EM聚类分析算法将数据分为8个典型子工况,然后将聚类后的数据用于训练RBF神经网络分类器。根据百度API提供的实时车流信息,采用RBF分类器预测车辆在预定行驶路线上的工况类型和平均能耗,从而精确地预测汽车的剩余电量,即剩余续航里程。With global energy shortages and deteriorating pollutions,key technologies for new energy vehicles have gained keen interest among researchers.Currently,the uncertain remaining cruising range of electric vehicles poses a huge barrier for their continued growth.Therefore,it is crucial to accurately predict the ensuing energy consumption of vehicles to determine their remaining mileage.Through the real vehicle-in-the-loop experimental platform,this research extracts the characteristic parameters after pre-processing the available driving data OBD of the test vehicle and the real-time traffic flow information of Baidu API.Then,the EM cluster analysis algorithm is employed to divide the data into eight typical sub-conditions.The clustered data is used to train the RBF neural network classifier.Finally,according to the real-time traffic information provided by Baidu API,the RBF classifier is employed to accurately predict the remaining power or the remaining crusing range of the electric vehicle through predictions of its working conditions and its average energy consumption on the predetermined driving route.

关 键 词：车辆SOC预测 API EM聚类分析算法 RBF神经网络分类器 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置] U461[自动化与计算机技术—控制科学与工程]