基于虚拟现实和近红外脑氧信号的驾驶员脑功能评估技术研究

发布时间：2018-12-16 09:46

【摘要】：自从人类发明交通工具以来,交通事故就如影随形,给人民的生产生活和生命安全造成了妨碍。如何减少交通事故的发生,提高交通运输的安全性,是学者们长久以来一直关注的课题。该问题解决的关键,在于如何探究交通事故的产生机理,只有明确了交通事故的成因,才能预先发现事故的可能性,并制定对策以防止交通事故的产生。驾驶员是"人——车——路"系统的主要构成因素之一,在相当的程度上决定了交通事故是否会发生。通过评估驾驶员的驾驶能力,可以对驾驶员发生交通事故的可能进行预测。排除身体躯干运动能力的影响,驾驶能力是脑功能的直接反映。本文在使用近红外光谱技术对驾驶员脑功能评估方面进行了初步尝试。本文的主要工作分为三部分。其一,提出一种基于小"光源——探头"间距的功能性近红外光谱表层生理噪声抵消方法。从近红外光谱血氧仪的检测原理以及人体头部的解剖学结构入手分析,汇总现有的小间距探头布置测量规律,并根据具体使用的近红外光谱仪参数规格,设计了具体的小间距探头布置方案。之后,针对基于小波相位相干性的功能连接分析和基于相位耦合函数的效应连接分析两种脑功能数据分析方法的要求,提出了一种基于连续小波变换的自适应噪声抵消方法,对其中的技术难点进行了阐述。其二,基于Unity 3D的虚拟驾驶器的设计和开发。基于作者的本科毕业设计,改进机械结构设计,选用并安装改进的传感器,设计传感器的单片机控制电路。基于Unity 3D,根据不同的实验设计要求,开发了多种针对性的虚拟驾驶任务,可以对驾驶员的驾驶表现进行记录和评估。其三,使用虚拟驾驶器进行了驾驶任务实验,使用近红外光谱血氧仪同步测量被试者静息态和任务态情况下的脑活动情况。使用基于连续小波变换的自适应噪声抵消方法对脑功能数据进行了预处理,并对进行过滤波去噪的数据和未进行滤波去噪的数据都进行了小波相位相干性功能连接分析,并对结果进行了进一步的统计分析。证实本文中自适应滤波去噪方法的有效性。综上,本文提出了一种用于功能性近红外光谱法测量脑功能的表层生理信号滤除方法,并设计虚拟驾驶任务实验对方法的有效性进行了证实。该方法进一步提高了功能性近红外光谱法测量脑组织血氧变化的准确性,这对于实时、动态、准确地评估驾驶员的脑功能具有重要意义。
[Abstract]:Ever since the invention of transportation, traffic accidents have been accompanied by traffic accidents, which have hindered the production, life and safety of people. How to reduce the occurrence of traffic accidents and improve the safety of transportation has long been a topic of concern to scholars. The key to solve this problem lies in how to probe into the mechanism of traffic accidents. Only when the causes of traffic accidents are clear can the possibility of accidents be detected in advance and countermeasures be formulated to prevent the occurrence of traffic accidents. Driver is one of the main components of the "man-car-road" system, which to a certain extent determines whether the traffic accident will occur or not. By evaluating driver's driving ability, we can predict the possibility of driver's traffic accident. Excluding the influence of body trunk motor ability, driving ability is a direct reflection of brain function. In this paper, a preliminary attempt was made to evaluate the brain function of drivers by near-infrared spectroscopy (NIR). The main work of this paper is divided into three parts. Firstly, a method of canceling the surface physiological noise of functional near infrared spectroscopy based on the small distance between light source and probe is proposed. Based on the principle of near infrared spectrometer (NIR) and the anatomical structure of human head, the present measurement rules of small spaced probes are summarized, and the parameters of NIR spectrometer are specified. A small spacing probe arrangement scheme is designed. Then, an adaptive noise cancellation method based on continuous wavelet transform is proposed to solve the problem of functional connection analysis based on wavelet phase coherence and effect connection analysis based on phase coupling function. The technical difficulties are expounded. Secondly, the design and development of virtual driver based on Unity 3D. Based on the author's undergraduate design, improving the mechanical structure design, selecting and installing the improved sensor, designing the single chip microcomputer control circuit of the sensor. Based on Unity 3D, according to different experimental design requirements, a variety of targeted virtual driving tasks are developed, which can record and evaluate driver's driving performance. Thirdly, the virtual driver was used to carry out the driving task experiment, and the brain activity of the subjects was measured synchronously in the resting state and the task state by using the near infrared spectroscopic oximeter. The adaptive noise cancellation method based on continuous wavelet transform is used to preprocess the brain functional data, and the wavelet phase coherence functional connection analysis is carried out for both the filtered wave de-noising data and the unfiltered de-noised data. Further statistical analysis is made on the results. The effectiveness of the adaptive filter denoising method in this paper is verified. In summary, a method of surface physiological signal filtering for measuring brain function by functional near infrared spectroscopy is proposed, and the validity of the method is verified by designing virtual driving task experiment. This method further improves the accuracy of functional near infrared spectroscopy (FNIR) in measuring the changes of blood oxygen in brain tissue, which is of great significance for the real-time, dynamic and accurate evaluation of the brain function of the driver.
【学位授予单位】：山东大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：U491.254

【相似文献】