基于隐马尔可夫模型的半挂车制动侧滑状态辨识  

Recognition of the state of semi-trailer braking slid based on hidden Markov model

作　　者：步玫 宇仁德[1] 吕安涛 马晨浩 胡婧晖 BU Mei;YU Rende;LYU Antao;MA Chenhao;HU Jinghui(School of Transportation and Vehicle Engineering,Shandong University of Technology,Zibo 255049,China;Shandong Transportation Research Institute,Jinan 250100,China)

机构地区：[1]山东理工大学交通与车辆工程学院,山东淄博255049 [2]山东省交通科学研究院,山东济南250100

出　　处：《山东理工大学学报（自然科学版）》2021年第3期60-64,69,共6页Journal of Shandong University of Technology:Natural Science Edition

摘　　要：为了识别半挂车在制动过程中是否有侧滑的危险,及时将车辆状态反馈给驾驶员以提高车辆主动监控预警,从车辆重心偏移、车速和道路三方面选出对制动侧滑影响较明显的4个特征参数,输入到ADAMS仿真软件进行直线制动仿真实验。建立基于隐马尔可夫模型的制动侧滑辨识模型,采用Baum-Welch算法对模型进行训练,得到似然模型的最优参数λ。优化后的HMM模型能有效识别半挂车的制动侧滑状态且辨识准确率接近90%。In order to identify whether the semi-trailer is in danger of slipping during braking,timely feedback the vehicle status to the driver and to improve vehicle active monitoring and early warning,speed and road,four characteristic parameters with obvious effects on braking sideslip are selected from the three aspects of vehicle center of gravity deviation,and they are input to ADAMS simulation software for linear braking simulation experiments.A braking slip identification model based on Hidden Markov model is established and it is trained using Baum-Welch algorithm to obtain the optimal parameterλof the likelihood model.The optimized HMM model effectively recognized the braking slip of the semi-trailer and the recognition accuracy was close to 90%.

关 键 词：制动侧滑 ADAMS 隐马尔可夫模型 Baum-Welch算法 安全状态辨识 

分 类 号：U491[交通运输工程—交通运输规划与管理]