基于移动智能终端的高速公路异常驾驶行为检测技术

发布时间：2018-01-06 06:26

本文关键词：基于移动智能终端的高速公路异常驾驶行为检测技术　出处：《浙江大学》2015年硕士论文　论文类型：学位论文

【摘要】：交通安全问题是关系到国家经济、社会民生的重大问题,而高速公路作为交通路网中的“大动脉”,保障其通行安全更是重中之重。随意变道、掉头行驶等驾驶行为是高速公路交通事故的重要诱因,而紧急制动、避障等驾驶行为则往往预示着已经发生的交通意外事件。及时发现这些异常的驾驶行为并通知相关人员做出响应,能够尽可能地减轻意外交通事件引起的负面影响,甚至直接避免交通事故的发生。本文提出的异常驾驶行为检测技术以目前十分普及的移动智能终端为基础,通过对加速度计、陀螺仪、旋转向量、GPS传感器数据进行融合、坐标系变换、滤波等预处理,再对数据进行区间划分形成候选行为事件,并利用基于分类规则的机器学习方法对候选行为事件上提取的行为特征进行学习和分类,从而实现对紧急制动和转向等基本行为的实时检测。在此基础上,两个和三个转向行为分别被组合形成变道、避障等组合行为的候选行为事件,并利用与基本行为相同的检测方法完成组合行为的检测。本文还基于以上技术设计实现了DriveDroid驾驶行为检测原型系统,并在实际道路中进行了验证实验。实验结果显示本文提出的检测技术具有较高准确率,具有一定的实用价值。
[Abstract]:Traffic safety is a major problem related to the national economy and the people's livelihood. As the "artery" in the traffic network, it is the most important to ensure the safety of the highway. Driving behavior such as turning around is an important cause of highway traffic accidents, while emergency braking. Driving behavior such as obstacle avoidance often foreshadowed traffic accidents that had already occurred. Timely detection of these abnormal driving behavior and notification of relevant personnel to respond. It can minimize the negative impact caused by traffic accidents and even avoid traffic accidents directly. The technology of abnormal driving behavior detection is based on the mobile intelligent terminal which is very popular at present. Through the data fusion of accelerometer, gyroscope and rotation vector GPS sensor, coordinate system transformation, filtering and other preprocessing, the data are divided into interval to form candidate behavior events. And the machine learning method based on classification rules is used to learn and classify the behavior features extracted from candidate behavior events, so as to achieve real-time detection of basic behaviors such as emergency braking and steering. Two and three steering behaviors were combined to form candidate behavior events for combinatorial behaviors such as path change, obstacle avoidance and so on. And using the same detection method as the basic behavior to complete the combined behavior detection. This paper also designed and implemented the DriveDroid driving behavior detection prototype system based on the above technology. The experimental results show that the proposed detection technology has high accuracy and has certain practical value.
【学位授予单位】：浙江大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：U495;TP212

【相似文献】