基于情绪计算的驾驶行为研究

发布时间：2018-05-25 19:14

本文选题：驾驶行为 + 情绪计算　；参考：《广东工业大学》2014年硕士论文

【摘要】：人工心理和情绪计算作为新兴科技中具有代表性的一门学科,本文提出将其融进人机交互技术应用中,开展关于人工心理和情绪计算在汽车驾驶行为中的应用研究,从而预防攻击性驾驶等不良驾驶行为,将促进安全驾驶的发展。全书共分五章,第1章主要介绍了人工情绪的概念,分析了国内外对于情绪计算课题的研究现状和发展趋势,探讨了关于情绪识别的几个方法,包括从生理方面的识别、语音识别和人脸表情识别。第2章分析了驾驶行为,尤其是愤怒情绪下的攻击性驾驶行为,并分别从驾驶员的生理、心理、性格因素和社会大环境因素着手去分析攻击性驾驶行为的成因,并提出了基于车辆运行状态和基于驾驶员特征的驾驶情绪检测方法。第3章将各类驾驶因子融合起来,包括油门与制动踏板的操作、方向盘转动操作、车灯切换和鸣笛操作等多因子的指标评价,运用Matlab建立了BP神经网络模型,并设计参数对其进行训练。第4章用机器视觉技术识别驾驶员的驾驶表情,通过中值滤波、直方图均衡、图像的腐蚀与膨胀等图像算法定位跟踪人脸,用Adaboost级联分类器训练定位人眼,并大胆提出了驾驶紧张度和驾驶平和度的概念。第5章,设计了实验研究方案,对驾驶紧张度和驾驶平和度分别建模,并运用matlab得出驾驶紧张度和驾驶平和度的曲线拟合和归一化模型,并最终得到驾驶情绪的表达式。
[Abstract]:Artificial psychology and emotion calculation is a representative subject in the emerging science and technology. This paper proposes to integrate it into the application of human-computer interaction technology, and to carry out the research on the application of artificial psychology and emotional calculation in the driving behavior of automobile. So as to prevent aggressive driving and other bad driving behavior, will promote the development of safe driving. The first chapter mainly introduces the concept of artificial emotion, analyzes the research status and development trend of emotion calculation at home and abroad, and discusses several methods of emotion recognition, including physiological recognition. Speech recognition and facial expression recognition. Chapter 2 analyzes driving behavior, especially aggressive driving behavior under anger, and analyzes the causes of aggressive driving behavior from driver's physiological, psychological, personality and social environment factors, respectively. A driving emotion detection method based on vehicle running state and driver characteristics is proposed. Chapter 3 combines all kinds of driving factors, including the operation of throttle and brake pedal, steering wheel rotation, lamp switching and siren operation, and establishes BP neural network model by using Matlab. The parameters are designed to train it. Chapter 4 uses machine vision technology to recognize driver's driving expression, uses median filter, histogram equalization, image corrosion and expansion to locate and track human face, and uses Adaboost cascade classifier to train and locate human eyes. And boldly put forward the concept of driving tension and driving peace. In chapter 5, the experimental research scheme is designed, the driving tension and driving peace are modeled separately, and the curve fitting and normalization model of driving tension and driving peace are obtained by using matlab, and the expression of driving emotion is obtained finally.
【学位授予单位】：广东工业大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：U491.254

【参考文献】