【作者】 杨蕾；

【导师】 郭军；

【作者基本信息】 西北大学 ， 软件工程（专业学位）， 2018， 硕士

【摘要】 随着社会和经济的不断发展,人们开始了自驾的全民热潮,几乎所有成年之后的学生都将考驾照列入了自己的必备技能之中,各个地方的驾校天天人满为患,道路上“新手”司机与日益增。虽然这些“新手”司机们可以正常驾驶车辆,但是经验的缺乏导致他们并不具备良好的驾驶习惯和处理突发问题的能力,造成巨大的安全隐患。而现有的驾驶员行车预警系统基本上都需要专业的设备,并且只有高端车辆才安装了这些注意力检测或者疲劳检测的“高配”设备。因此,研究普适的危险驾驶检测方法,对降低道路交通安全隐患以及提醒驾驶员错误的驾驶习惯,促进“新手”司机形成科学的驾驶行为具有重要意义。本文针对驾驶员行车途中的固定场景,提出了利用WiFi和手机内置传感器相结合的危险驾驶动作的检测方法,由利用WiFi信号进行驾驶员危险动作识别和利用智能手机内置传感器进行车辆状态检测两部分组成。该方法通过驾驶员动作与车辆状态之间的强相关关系将人、车两方面的信息进行结合,利用手机内置传感器数据的检测结果来校准WiFi信号对危险驾驶动作的识别,一定程度上弥补了周围环境和设备对无线信号影响的缺陷,实现高精度的危险驾驶动作检测。具体的研究内容如下:(1)结合现有研究成果以及日常行车途中的状况分析,规定了15种驾驶过程中的“危险驾驶动作”,包括疲劳驾驶、注意力分散、对拥挤路段不合适的应对操作等具有安全隐患的驾驶动作。在车载Wi-Fi的无线网络环境中,通过检测WiFi信号的CSI(Channel State Information)信道状态信息识别危险驾驶动作,对驾驶员行车途中的驾驶动作进行实时“监控”。(2)针对无线信号的固有缺陷,使用智能手机中的内置传感器对危险驾驶动作识别进行辅助检测以降低误检率,提高检测精度。将智能手机固定在车上,结合多个手机内置传感器的信息分析车辆状况,检测车辆在行驶过程中的真实状态。(3)行车途中检测到危险驾驶动作时对驾驶员进行预警提醒,并且对驾驶员进行驾驶行为评估,促使“新手”司机形成良好的驾驶习惯,有效降低交通安全隐患。更多还原

【Abstract】 With the continuous development of society and economy,people have begun to drive the entire people’s upsurge of self-driving.Almost all adult students have included their driving licenses as their essential skills.The driving schools in various places are overcrowded every day and the number of "novice" drivers on the road continues to increase.Although these "novice" drivers can normally drive vehicles,the lack of experience leads them to not have good driving habits and the ability to deal with unexpected problems,causing huge security risks.However,existing driver warning systems basically require professional equipment,and only high-end vehicles have installed these "high-profile" devices for attention detection or fatigue detection.Therefore,it is of great significance to study universal dangerous driving detection methods to reduce the hidden dangers of road traffic safety and to promote the formation of scientific driving behavior for "novice" drivers.This paper proposes a method of detecting dangerous driving behavior using WiFi and built-in sensors in a fixed scene where the driver is driving..It uses the WiFi signal to identify the driver’s dangerous movement and uses the built-in sensor of the smart phone to detect the vehicle status.The method combines the information of the driver and the vehicle by the strong correlation between the driver’s actions and the vehicle’s status,and uses the detection results of the built-in sensor data of the smart phone to calibrate the recognition of the dangerous driving action by the WiFi signal,this way compensates for the defects of the surrounding environment and the vehicle’s influence on the wireless signal to some extent,and realizes high-precision detection of dangerous driving actions.The specific research content is as follows:1)Combining with the existing research results and the analysis of the conditions during daily driving,we have defined 15 kinds of "dangerous driving actions" in the driving process,including driving actions with safety hazards such as fatigued driving,distracted attention,and inappropriate handling of congested road sections.In the WiFi wireless network environment of the vehicle,the danger driving operation is recognized by detecting the CSI(Channel State Information)channel state information of the WiFi signal,thereby real-time "monitoring" of the driving action while the driver is driving.2)For the intrinsic defects of wireless signals,we use built-in sensors in smart phones to calibrate dangerous driving motion recognition,which reduces false detection rate and improve detection accuracy.The smart phone is fixed on the vehicle,and the information of the vehicle is combined with the information of the built-in sensors of a plurality of smart phones to analyze the state of the vehicle.3)When the dangerous driving action is detected during driving,the driver is given an early warning reminder,and the driver is evaluated for driving behavior,so that the "novice" driver can form a good driving habit and effectively reduce the hidden danger of traffic safety.更多还原