检索结果-内蒙古大学图书馆

Driving Behavior Prediction Based on Combined Neural Network Model

IEEE TRANSACTIONS ON COMPUTATIONAL SOCIAL SYSTEMS 2024年第3期11卷 4488-4496页

作者： Li, Runmei Shu, Xiaoting Li, Chen Beijing Jiaotong Univ Sch Elect & Informat Engn Beijing 100044 Peoples R China

Accurate behavior prediction of surrounding vehicles can greatly improve the operating safety of autonomous vehicles. However, in real traffic scence, the complexity and uncertainties of traffic flow bring great challenges to driving behavior prediction. This article proposes a driving behavior prediction model using a wide-deep framework that combines gradient boosting decision tree (GBDT), convolutional neural network (CNN), and long short-term memory network (LSTM) algorithm to fully mine driving behavior characteristics while improve interpretability of the CNN-LSTM model. The GBDT algorithm can quantitatively describe the interaction between the autonomous vehicle and its surrounding vehicles during the driving process, obtaining a series of driving behavior rules, and integrating the driving behavior rule features into the CNN-LSTM neural network. The CNN-LSTM neural network model is constructed to find the spatial features in driving trajectory by CNNs and the temporal features by LSTM networks. The accuracy of the driving behavior prediction model is further improved. Simulation experiments show the rationality and validity of themodel and algorithm.

关键词： Behavioral sciences Predictive models Convolutional neural networks Hidden Markov models Feature extraction Neural networks Prediction algorithms driving behavior prediction gradient boosting tree algorithm long short-term memory neural network wide-deep

来源：评论

学校读者我要写书评

暂无评论

Driver Fatigue Detection Control System

Driver Fatigue Detection Control System

引用

37th Chinese Control Conference (CCC)

作者： Fan Dachuan Tang Xinxing Changchun Univ Technol Sch Mechatron Engineer Changchun 130012 Jilin Peoples R China

ISBN: (纸本)9789881563958

In order to realize automatic on-line monitoring of driver fatigue state, four modules are mainly included in this system, which are image capture module, image preprocessing module, feature detection and extraction module and feature classification recognition module. Firstly, a detection system platform is built by using computer. Visual Studio and some other software and hardware equipment, and image pre-processing is carried out to detect and locate the driver face region in real-time, aiming to the shortcoming of Camshift tracking algorithm, an algorithm combining the Camshift tracking algorithm with Kalman filter is proposed to realize the real-time tracking of:human face region. And then the face model is obtained by training the sample images calibrated the facial feature points by using the gradient regression tree algorithm. The regions of eyes and mouth can be located by using this face model on the detected face. to verify the accuracy of the proposed driver fatigue detection algorithm, a fatigue driving detection experiment is carried out in the Honda's car. The driver's face images are captured by installing the COMS camera with infrared function on the front windshield, and the data are calculated and analyzed by computer. Experiment contents include the face region detection and tracking, facial features detection and state recognition, as well as fatigue recognition based on facial features and analysis. The experiment results show that the system has good accuracy, real-time and robustness, and the established driver fatigue warning can meet the real-time requirement of the driver fatigue state detection.

关键词： Fatigue detection Kalman Filter Face tracking gradient boosting tree algorithm Ellipse fitting

来源：评论

学校读者我要写书评

暂无评论

3D sensor network location spatial positioning technology based on machine learning

引用

INTERNATIONAL JOURNAL OF EMERGING ELECTRIC POWER SYSTEMS 2023年第1期24卷 13-23页

作者： Lu, Zhiyong Tan, Xiaodan Southern Med Univ Sch Biomed Engn Guangzhou 510515 Guangdong Peoples R China Guangdong Vocat Coll Elect Technol Guangzhou 510515 Guangdong Peoples R China

The purpose of this paper is to combine machine learning to locate the 3D sensor network space. Real life is mostly a three-dimensional environment. Whether it is a factory in manufacturing or a vegetation base in agriculture, it needs to be monitored and positioned. In this paper, the localization algorithm is discussed to a certain extent. This paper firstly introduces the relevant background and organizes related work. It also wrote related algorithms, such as ranging-based positioning algorithms in the free space of wireless sensors. It shows the positioning link by introducing the wireless sensor network structure system and node structure. And this paper summarizes the Bounding-box Method positioning principle, TDOA algorithm principle, and TDOA positioning principle. It then describes the gradient boosting tree classification algorithm based on machine learning, and focuses on the admiral boosting tree classification algorithm related to the experiment. This paper also describes the ranging technology combining RSSI algorithm and DV-Hop algorithm in three-dimensional space, and mentions two algorithms of RSSI and DV-Hop. In the fourth part, the machine learning coordinate prediction accuracy improvement experiment and the three-dimensional space positioning algorithm optimization experiment and result analysis are carried out. It is proved by experiments that the model evaluation effect of the gradient boosting tree classification algorithm in machine learning is the best. It can be applied to the calculation of relative position coordinates of label nodes. It then carried out the three-dimensional positioning effect test experiment of IDV-Hop algorithm. This shows that when the network density in the experimental environment reaches more than 12, the localization coverage of IDV-Hop algorithm and DV-Hop algorithm are both higher than 91%. Finally, the hybrid algorithm of RSSI and DV-Hop algorithm is used to compare the positioning accuracy, positioning cover

关键词： 3D sensor network DV-Hop algorithm gradient boosting tree algorithm hybrid algorithm localization algorithm machine learning

来源：评论

学校读者我要写书评

暂无评论

Driver Fatigue Detection Control System

Driver Fatigue Detection Control System

引用

第37届中国控制会议

作者： Fan Dachuan Tang Xinxing School of Mechatronic Engineer Changchun University of Technology

In order to realize automatic on-line monitoring of driver fatigue state,four modules are mainly included in this system,which are image capture module,image preprocessing module,feature detection and extraction module and feature classification recognition ***,a detection system platform is built by using computer,Visual Studio and some other software and hardware equipment,and image pre-processing is carried out to detect and locate the driver face region in real-time,aiming to the shortcoming of Camshift tracking algorithm,an algorithm combining the Camshift tracking algorithm with Kalman filter is proposed to realize the real-time tracking of human face *** then the face model is obtained by training the sample images calibrated the facial feature points by using the gradient regression tree *** regions of eyes and mouth can be located by using this face model on the detected face,to verify the accuracy of the proposed driver fatigue detection algorithm,a fatigue driving detection experiment is carried out in the Honda’s *** driver’s face images are captured by installing the COMS camera with infrared function on the front windshield,and the data are calculated and analyzed by *** contents include the face region detection and tracking,facial features detection and state recognition,as well as fatigue recognition based on facial features and *** experiment results show that the system has good accuracy,real-time and robustness,and the established driver fatigue warning can meet the real-time requirement of the driver fatigue state detection.

关键词： Fatigue detection Kalman Filter Face tracking gradient boosting tree algorithm Ellipse fitting

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：