第11卷第3期 智能系统学报 Vol.11 No.3 2016年6月 CAAI Transactions on Intelligent Systems Jun.2016 D0I:10.11992/is.201504022 网络出版地址:http://www.cmki.net/kcms/detail/23.1538.tp.201509030.1456.002.html 智能手机车辆异常驾驶行为检测方法 周后飞123,刘华平23,石红星4 (1.重庆交通大学土木工程学院,重庆400074:2.清华大学计算机科学与技术系,北京100084:3.清华大学智能技 术与系统国家重点实验室,北京100084:4.北京城建道桥建设集团有限公司,北京100080) 摘要:将智能手机作为车辆异常驾驶行为检测工具,设计了一种车辆异常驾驶行为检测方法和系统。系统通过获 取车载智能手机内部的加速度传感器数据、陀螺仪传感器数据以及磁场传感器数据,经坐标旋转和特征提取,并利 用基于核方法极限学习机(核ELM)得到的驾驶行为在线分析算法,以实现能实时识别包括频繁变道、频繁变速及急 刹车在内的多种车辆异常驾驶行为,并在车辆出现异常驾驶行为时开启报警语音。测试结果表明,基于核ELM算法 的驾驶行为分类器性能比基于支持向量机(SVM)算法更好,提出的异常驾驶行为检测系统能有效识别各种驾驶 行为。 关键词:智能手机:异常驾驶行为检测:传感器:核方法:极限学习机:支持向量机 中图分类号:TP29:U49文献标志码:A文章编号:1673-4785(2016)01-0410-08 中文引用格式:周后飞,刘华平,石红星,等.智能手机车辆异常驾驶行为检测方法[J].智能系统学报,2016,02(1):410417. 英文引用格式:ZHOU Houfei,LIU Huaping,SHI Hongxing,etal.Abnormal driving behavior detection based on the smart phone [J].CAAI Transactions on Intelligent Systems,2016,02(1):410-417. Abnormal driving behavior detection based on the smart phone ZHOU Houfei,LIU Huaping,SHI Hongxing' (1.School of Civil Architecture Engineering,Chongqing Jiaotong University,Chongqing 400074,China;2.Department of Computer Science and Technology,Tsinghua University,Beijing 100084,China;3.State Key Laboratory of Intelligent Technology and Systems, Tsinghua University,Beijing 100084,China;4.Road and Bridge Construction Group Co.,Ltd.of Beijing,Beijing 100080,China) Abstract:Using the smart phone as a tool for detecting abnormal driving behavior,this paper designs an abnormal driving behavior detection method and a practical system.First,the system obtains data from the acceleration,mag- netic,and gyroscope sensors of an on-board smart phone.Then,through coordinate rotation,feature extraction, and an online driving behavior analysis algorithm,which is based on the kernel extreme learning machine (ELM) algorithm,the system identifies real-time abnormal driving behavior,including frequent lane-changing,frequent speed-changing,and emergency braking.It then sets off an alarm when abnormal driving behavior has been identi- fied.Test results indicate that the driving behavior classifier,which is based on the kernel ELM algorithm,performs better than the support vector machine algorithm.In addition,the proposed abnormal driving behavior detection sys- tem can effectively identify various driving behaviors. Keywords:smart phone;abnormal driving behavior detection;sensor;kernel method;extreme learning machine (ELM);support vector machine 近年来,国内外研究人员相继研究利用智能手 研究主要集中在人体行为识别领域[],但也有部 机识别各种行为。目前,基于智能手机的模式识别 分学者尝试将智能手机应用到车辆驾驶行为识别方 收稿日期:2015-04-09.网络出版日期:2015-09-30. 面。其中,Dai等)将车载智能手机的加速度和方 基金项目:国家重点基础研究与发展计划项目(2013CB329403). 向传感器数据作为车辆的运行参数来检测醉驾行 通信作者:刘华平hpliu@tsinghua..edu.cm.第 11 卷第 3 期 智 能 系 统 学 报 Vol.11 №.3 2016 年 6 月 CAAI Transactions on Intelligent Systems Jun. 2016 DOI:10.11992 / tis.201504022 网络出版地址:http: / / www.cnki.net / kcms/ detail / 23.1538.tp.201509030.1456.002.html 智能手机车辆异常驾驶行为检测方法 周后飞1,2,3 ,刘华平2,3 ,石红星4 (1.重庆交通大学 土木工程学院,重庆 400074; 2.清华大学 计算机科学与技术系,北京 100084; 3.清华大学 智能技 术与系统国家重点实验室,北京 100084; 4.北京城建道桥建设集团有限公司,北京 100080) 摘 要:将智能手机作为车辆异常驾驶行为检测工具,设计了一种车辆异常驾驶行为检测方法和系统。 系统通过获 取车载智能手机内部的加速度传感器数据、陀螺仪传感器数据以及磁场传感器数据,经坐标旋转和特征提取,并利 用基于核方法极限学习机(核 ELM)得到的驾驶行为在线分析算法,以实现能实时识别包括频繁变道、频繁变速及急 刹车在内的多种车辆异常驾驶行为,并在车辆出现异常驾驶行为时开启报警语音。 测试结果表明,基于核 ELM 算法 的驾驶行为分类器性能比基于支持向量机( SVM) 算法更好,提出的异常驾驶行为检测系统能有效识别各种驾驶 行为。 关键词:智能手机;异常驾驶行为检测;传感器;核方法;极限学习机;支持向量机 中图分类号:TP29;U49 文献标志码:A 文章编号:1673⁃4785(2016)01⁃0410⁃08 中文引用格式:周后飞,刘华平,石红星,等.智能手机车辆异常驾驶行为检测方法[J]. 智能系统学报, 2016, 02(1): 410⁃417. 英文引用格式:ZHOU Houfei, LIU Huaping, SHI Hongxing,et al. Abnormal driving behavior detection based on the smart phone [J]. CAAI Transactions on Intelligent Systems, 2016, 02(1): 410⁃417. Abnormal driving behavior detection based on the smart phone ZHOU Houfei 1,2,3 , LIU Huaping 2,3 , SHI Hongxing 4 (1.School of Civil & Architecture Engineering, Chongqing Jiaotong University, Chongqing 400074, China; 2. Department of Computer Science and Technology, Tsinghua University, Beijing 100084, China; 3. State Key Laboratory of Intelligent Technology and Systems, Tsinghua University, Beijing 100084, China; 4. Road and Bridge Construction Group Co., Ltd. of Beijing, Beijing 100080, China) Abstract:Using the smart phone as a tool for detecting abnormal driving behavior, this paper designs an abnormal driving behavior detection method and a practical system. First, the system obtains data from the acceleration, mag⁃ netic, and gyroscope sensors of an on⁃board smart phone. Then, through coordinate rotation, feature extraction, and an online driving behavior analysis algorithm, which is based on the kernel extreme learning machine (ELM) algorithm, the system identifies real⁃time abnormal driving behavior, including frequent lane⁃changing, frequent speed⁃changing, and emergency braking. It then sets off an alarm when abnormal driving behavior has been identi⁃ fied. Test results indicate that the driving behavior classifier, which is based on the kernel ELM algorithm, performs better than the support vector machine algorithm. In addition, the proposed abnormal driving behavior detection sys⁃ tem can effectively identify various driving behaviors. Keywords:smart phone; abnormal driving behavior detection; sensor; kernel method; extreme learning machine (ELM); support vector machine 收稿日期:2015⁃04⁃09. 网络出版日期:2015⁃09⁃30. 基金项目:国家重点基础研究与发展计划项目(2013CB329403). 通信作者:刘华平.hpliu@ tsinghua.edu.cn. 近年来,国内外研究人员相继研究利用智能手 机识别各种行为。 目前,基于智能手机的模式识别 研究主要集中在人体行为识别领域[1⁃6] ,但也有部 分学者尝试将智能手机应用到车辆驾驶行为识别方 面。 其中,Dai 等[7] 将车载智能手机的加速度和方 向传感器数据作为车辆的运行参数来检测醉驾行