汽车危险驾驶行为检测系统的设计与实现

Abstract

Road traffic safety has always been one of the major social security issues, relating to the lives and property of people. Driver's dangerous driving behavior is an important incentive for road traffic accidents. Timely detection of these dangerous driving behavior and inform the driver to respond, can effectively reduce the occurrence of road traffic accidents. Now with the rapid development of science and technology, various sensors are integrated in smartphones that are widely available in people's lives. Therefore, the design and implementation of dangerous driving behavior detection system based on the sensor data on smartphone is of great significance.

Multiple sensors such as GPS, accelerometer and gyroscope are used in the dangerous driving behavior detection technology. The rules for fatigue driving, distraction driving, and speeding determination are relatively simple. We have designed decision rules based on relevant traffic laws and implemented them on the client side. For rapid acceleration, rapid deceleration, sharp turn, people's subjective feelings is the basis for the judgment. Based on the characteristics of the driving behavior, we used forward and backward filtering to divide the driving behavior. The duration, speed variation and maximum acceleration are taken as the eigenvalues of the events. Duration, steering angle, maximum normal acceleration and maximum angular velocity are extracted as feature values for turning events. Eigenvalue imbalance, we have normalized the various eigenvalues, and finally used kNN algorithm to classify the candidate driving behavior. After detecting dangerous driving we used text and voice to alert the driver to help them respond promptly.

Based on the above techniques, we design and implementation of a dangerous driving behavior detection system on iOS, and tested on the actual road, the results show that the dangerous driving behavior detection system with high accuracy, with some practical value .

Key words：Detection of dangerous driving Mobile phone sensor

kNN algorithm Componentization

目录

摘要.............................................................................................................. I Abstract.......................................................................................................... II 1 绪论

1.1 研究背景与意义 (1)

1.2 国内外研究概况 (3)

1.3 主要研究内容与章节安排 (5)

2 危险驾驶行为检测技术基础研究

2.1 移动智能终端传感器概况 (8)

2.2 常用的分类器 (12)

2.3 本章小结 (13)

3 危险驾驶行为检测系统需求分析

3.1 危险驾驶行为检测系统功能性需求 (14)

3.2 危险驾驶行为检测系统非功能性需求 (16)

3.3 本章小结 (17)

4 危险驾驶行为检测系统设计

4.1 危险驾驶行为检测系统总体架构 (19)

4.2 危险驾驶行为检测规则和算法设计 (20)

4.3 系统功能组件设计 (26)

4.4 数据库设计 (32)

4.5 本章小结 (33)

5 危险驾驶行为检测系统实现

5.1 开发环境介绍 (34)