基于高斯混合隐马尔科夫模型与人工神经网络的紧急换道行为预测方法

doi:10.3969/j.issn.1004-132X.2020.23.012

中国机械工程

基于高斯混合隐马尔科夫模型与人工神经网络的紧急换道行为预测方法

于扬;梁军;陈龙;陈小波;朱宁;华国栋

1. 江苏大学汽车工程研究院，镇江， 212013

2. 静冈理工科大学机械系，静冈袋井， 437-0032

3. 江苏智行未来汽车研究院，南京， 210000

出版日期:2020-12-10 发布日期:2020-12-18
基金资助:
国家重点研发计划资助项目(2018YFB1600500);
国家自然科学基金资助项目(U1664258,51875255,61601203);
江苏省六大人才高峰项目(2015-DZXX-048);
江苏省高等学校自然科学研究重大项目(18KJA580002)

Vehicle Emergency Lane-changing Behavior Prediction Method Based on GMM-HMM and ANN

YU Yang;LIANG Jun;CHEN Long;CHEN Xiaobo;ZHU Ning;HUA Guodong

1. Automotive Engineering Research Institute, Jiangsu University, Zhenjiang, Jiangsu, 212013

2. Department of Mechanical, Shizuoka Institute of Science and Technology, Fukuroi, Shizuoka, 437-0032，Japan

3. Jiangsu Smart Travel Future Automobile Research Institute, Nanjing, 210000

Online:2020-12-10 Published:2020-12-18

摘要/Abstract

摘要： 为了有效降低因驾驶员紧急换道行为而诱发的交通事故，提高道路交通事故链阻断效率，提出一种基于高斯混合隐马尔科夫模型（GMM-HMM）和人工神经网络（ANN）的紧急换道行为预测方法。首先利用GMM-HMM对车辆行驶状态以及驾驶行为连续观察序列进行换道意图辨识，采用ANN预测下一时段的驾驶行为，再预测换道过程中的横向加速度变化率，从而判断紧急换道的危险程度。驾驶员在环仿真实验及实车实验结果表明，该方法预测避险成功率达92.83%，实验避险成功率达90.32%。该方法能有效地对紧急换道行为进行提前警告与干预。

关键词: 换道行为预测, 高斯混合隐马尔可夫模型, 人工神经网络, 道路交通事故链阻断

Abstract: In order to effectively reduce the traffic incidents caused by drivers’ emergency lane-changing behaviors and improve the blocking efficiency of chains of road traffic incidents, a method of predicting emergency lane-changing behaviors was proposed based on GMM-HMM and ANN. First, GM-HMM was utilized to identify the intention of lane-changing based on the continuous observation sequences of vehicle states and the driving behaviors. ANN was utilized to predict the driving behaviors in the next interval, and then lateral acceleration rate during lane-changing was predicted to judge the dangerous degree of emergency lane-changing. The results of driver in-the-loop simulations and real vehicle experiments show that the predictive risk avoidance success rate is as 92.83% and the experiment risk avoidance success rate is as 90.32%. The method may effectively warn and intervene the emergency lane-changing behaviors.

Key words: lane-changing behavior prediction, Gaussian mixed model-hidden Markov model(GMM-HMM), artificial neural network(ANN), blocking of chains of road traffic incidents

中图分类号:

U580.70

于扬;梁军;陈龙;陈小波;朱宁;华国栋. 基于高斯混合隐马尔科夫模型与人工神经网络的紧急换道行为预测方法[J]. 中国机械工程, DOI: 10.3969/j.issn.1004-132X.2020.23.012.

YU Yang;LIANG Jun;CHEN Long;CHEN Xiaobo;ZHU Ning;HUA Guodong.

Vehicle Emergency Lane-changing Behavior Prediction Method Based on GMM-HMM and ANN

[J]. China Mechanical Engineering, DOI: 10.3969/j.issn.1004-132X.2020.23.012.

导出引用管理器 EndNote|Ris|BibTeX

链接本文: http://www.cmemo.org.cn/CN/10.3969/j.issn.1004-132X.2020.23.012

http://www.cmemo.org.cn/CN/Y2020/V31/I23/2874

参考文献

［1］梁军，张婉婉，何美玲，等. 车联网环境下道路交通事故链阻断效率评价方法研究［J］.中国公路学报, 2018, 31(4):215-225.

LIANG Jun, ZHANG Wanwan,HE Meiling, et al.A Blocking Efficiency Evaluation Method for Chains of Road Traffic Incident in IOV Environment［J］. China Journal of Highway and Transport, 2018, 31(4):215-225.

［2］吕岸, 胡振程, 陈慧. 基于高斯混合隐马尔科夫模型的高速公路超车行为辨识与分析［J］.汽车工程, 2010, 32(7):630-634.

LYU An, HU Zhencheng, CHEN Hui. Identification and Analysis of Overtaking Behavior of Expressway Based on Gaussian Mixed Hidden Markov Model［J］. Automotive Engineering, 2010, 32(7):630-634.

［3］叶创鑫, 谭满春. 基于SVM与人工神经网络组合模型的物流规划车辆行程时间预测［J］.暨南大学学报：自然科学与医学版, 2011, 31(5):451-456.

YE Chuangxin, TAN Manchun. Prediction of Vehicle Travel Time Based on SVM and Artificial Neural Network Combined Model［J］. Journal of Jinan University:Natural Science and Medicine, 2011, 31(5):451-456.

［4］LI G, LI S E, LIAO Y, et al. Lane Change Maneuver Recognition via Vehicle State and Driver Operation Signals：Results from Naturalistic Driving Data［C］// 2015 Intelligent Vehicles Symposium(Ⅳ). Seoul:IEEE, 2015:865-870.

［5］JIANG B, FEI Y. Traffic and Vehicle Speed Prediction with Neural Network and Hidden Markov Model in Vehicular Networks［C］// 2015 Intelligent Vehicles Symposium(Ⅳ). Seoul:IEEE, 2015:1082-1087.

［6］JIANG B, FEI Y. Vehicle Speed Prediction by Two-level Data Driven Models in Vehicular Networks［J］. IEEE Transactions on Intelligent Transportation Systems, 2017，18(7):1793-1801.

［7］JULA H, KOSMATOPOULOS E B. Collision Avoidance Analysis for Lane Changing and Merging［J］. IEEE Transactions on Vehicular Technology, 2000, 49(6):2295-2308.

［8］PENG W Y. Research on Optimization and Implementation of BP Neural Network algorithm［C］// IEEE 2014 7th International Conference on Intelligent Computation Technology and Automation. Changsha, 2014:104-107.

［9］熊晓夏, 陈龙, 梁军,等. 危险换道驾驶行为预测方法研究［J］. 汽车工程, 2017, 39(9):1040-1046.

XIONG Xiaoxia, CHEN Long, LIANG Jun, et al. Research on Prediction Methods of Dangerous Lane Driving Behavior［J］. Automotive Engineering, 2017, 39(9):1040-1046.

［10］PONGSATHORN R, TAKUYA M, MASAO N. Direct Yaw Moment Control System Based on Driver Behaviour Recognition［J］. Vehicle System Dynamics, 2008, 46(S1):911-921.

［11］SALVUCCI D D. Inferring Driver Intent:a Case Study in Lane-change Detection［J］. Human Factors & Ergonomics Society Annual Meeting Proceedings, 2004, 48(19):2228-2231.

[1]	赵丽娟, 刘旭南, 孙强. 基于VPANNs的掘进机回转台可靠性分析[J]. 中国机械工程, 2014, 25(5): 602-607.
[2]	宋春生, 张锦光, 张建国. 基于混合算法的磁悬浮隔振器模型辨识与验证[J]. 中国机械工程, 2014, 25(14): 1929-1934.
[3]	袁长峰, 王万雷, 李剑锋, 陈燕. 基于粗糙集和多人工神经网络模型的产品敏捷定制设计方法 [J]. 中国机械工程, 2011, 22(8): 926-931.
[4]	王正如, 梁晋, 王立忠, 梁新合. 基于模糊综合评价的车身曲面品质分析 [J]. 中国机械工程, 2011, 22(6): 748-751.
[5]	马瑞;赵军;屈晓阳;. 盒形件智能化拉深压边力规律的实时预测[J]. J4, 2009, 20(17): 0-2021.
[6]	王永靖, 刘飞, 王琦峰, . 汽车制造企业绿色制造模式评价研究[J]. 中国机械工程, 2008, 19(24): 2934-2943.
[7]	陶冶, 黄洪钟, 刘志杰, . 在Multi-Agent环境下机械产品协同优化[J]. 中国机械工程, 2008, 19(1): 52-57.
[8]	李威;仝军令. 基于遗传算法和神经网络的油气弹簧非参数化建模[J]. J4, 2007, 18(23): 0-2897.
[9]	陈拂晓;李贺军;郭俊卿;杨永顺. 基于人工神经网络的轴承保持架超塑性成形工艺参数优化[J]. J4, 2007, 18(23): 0-2859.

基于高斯混合隐马尔科夫模型与人工神经网络的紧急换道行为预测方法

Vehicle Emergency Lane-changing Behavior Prediction Method Based on GMM-HMM and ANN

PDF

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 9

编辑推荐

Metrics