| 重型载货汽车气压制动系统危险状态识别 |
| |
| 引用本文： | 史培龙,余强,赵轩,袁晓磊,刘攀. 重型载货汽车气压制动系统危险状态识别[J]. 中国公路学报, 2019, 32(7): 182-190. DOI: 10.19721/j.cnki.1001-7372.2019.07.019 |
| |
| 作者姓名： | 史培龙 余强 赵轩 袁晓磊 刘攀 |
| |
| 作者单位： | 1. 长安大学 汽车学院, 陕西 西安 710064;2. 长安大学 汽车运输安全保障技术交通行业重点试验室, 陕西 西安 710064 |
| |
| 基金项目： | 国家重点研发计划项目（2017YFC0803904）;中央高校基本科研业务费专项资金项目（300102228108）;陕西省重点产业创新链（群）项目（2018ZDCXL-GY-05-03-01）;陕西省重点研发计划重点项目（2018ZDXM-GY-082）;陕西省博士后基金项目（2017BSHEDZZ36）;榆林市科技计划项目（211822190134） |
| |
| 摘 要： | 针对重型载货汽车因气压制动系统发生管路破裂、机械故障或热衰退导致制动效能下降且不易察觉从而引发严重交通事故的问题,提出基于主成分分析降维（PCA降维）和马尔可夫模型的气压制动系统危险状态识别方法。考虑到三轴载货汽车双回路制动系统的结构复杂性以及制动过程制动踏板动作、系统压力建立和实现车辆减速具有明显的时序性特点,首先采用PCA降维的方法对系统状态进行辨识;然后运用驾驶人制动意图与制动系统响应的双层隐形马尔可夫模型对系统状态进行识别。受驾驶人习惯影响制动踏板作用瞬间辨识度低,采用混合高斯聚类法提取不同制动意图时制动保持阶段数据建立制动意图识别模型和系统响应识别模型,通过二者匹配程度判定系统状态。最后,分别依据实车试验数据对模型进行离线训练和在线辨识验证。试验结果表明：系统正常状态下,基于PCA降维和马尔可夫模型相结合的识别方法能够准确、有效地识别制动系统状态;制动管路断开压力降低状态下,PCA降维方法能够及时有效识别其危险状态。
|
| 关 键 词： | 汽车工程 气压制动系统 马尔可夫模型 重型载货汽车 状态识别 |
| 收稿时间： | 2018-07-09 |
Risk State Identification of Pneumatic Braking System for Heavy Duty Truck |
| |
| SHI Pei-long,YU Qiang,ZHAO Xuan,YUAN Xiao-lei,LIU Pan. Risk State Identification of Pneumatic Braking System for Heavy Duty Truck[J]. China Journal of Highway and Transport, 2019, 32(7): 182-190. DOI: 10.19721/j.cnki.1001-7372.2019.07.019 |
| |
| Authors: | SHI Pei-long YU Qiang ZHAO Xuan YUAN Xiao-lei LIU Pan |
| |
| Affiliation: | 1. School of Automobile, Chang'an University, Xi'an 710064, Shaanxi, China;2. Key Laboratory of Automotive Transportation Safety Techniques of Ministry of Transport, Chang'an University, Xi'an 710064, Shaanxi, China |
| |
| Abstract: | To resolve the problem that potential factors easily to ignore for pneumatic braking system cause serious traffic accidents, including pipeline breakdown, mechanical faults, or thermal recession, a method was developed for identifying the risks of pneumatic braking systems based on principal component analysis (PCA) dimensionality reduction and a Markov model. For the dual-circuit pneumatic brake system of a three-axle heavy duty truck, in order to consider the structural complexity and obvious time-series characteristics during the braking process, brake pedal operation, establish system pressure, and achieve vehicle deceleration, the system state was first identified by characteristic value using PCA dimension reduction. Then, a double-layer hidden Markov model with driver braking intention and braking system response was used to recognize the system state. Affected by driver's habits,the recognizability of braking pedal action moment was low. In this paper, Mixture Gauss Clustering Method was used to extract the data of braking holding stage to build the braking intention recognition model and system response recognition model. The system state was represented by matching degree between the two models. Finally, braking tests were performed to train the identification of the mode and validate the accuracy of the model online. The test results show that the identification method based on PCA dimensionality reduction and Markov model can identify the status of the braking system accurately and effectively in the normal state. The PCA dimensionality reduction method can also identify the dangerous state promptly and effectively when brake line breakdown and pressure reduction occur. |
| |
| Keywords: | automotive engineering pneumatic braking system Markov model heavy-duty truck state identification |
|
| 点击此处可从《中国公路学报》浏览原始摘要信息 |
|
点击此处可从《中国公路学报》下载免费的PDF全文 |
|
