Multi-modal cloud based system for detecting early signs of driver cognitive decline and/or high risk driving behaviors

项目摘要/摘要 物理科学公司(PSI)与密歇根州立大学(MSU)和密歇根大学 密歇根大学(密歇根大学)计划开发一种新技术，该技术将非侵入性地监测司机的认知 拒绝和/或特别适合老年人(65岁)的高风险驾驶行为。与上下文相关 将获得生理生物标记以及基于计算机视觉的驾驶质量测量 通过基于云的采集系统。作为认知状态衡量标准的实时风险评估 将在云上执行，为老年人提供全面的驱动器质量指标分析， 不同程度的认知障碍。 评估驾驶员的认知状态需要使用复杂的监控平台，该平台可以 无缝安装在任何汽车上，不会干扰或增加驾驶的复杂性。因此，我们的目标是 是开发基于云的、多模式的常规认知/驾驶评估系统，该系统结合了Smart 转向套(S3)，监测生理生物标志物，以及计算机视觉(CV)系统， 为生物标志物的变化提供背景洞察，并监控司机的习惯。基于云的计算机 视觉智能转向套(C-CVS3)传感器系统将能够监测心脏等生物标志物 速度、握力和皮肤表面出汗过多表明压力水平 影响潜在的驾驶质量下降，这也是预计与认知相关的因素 关键事件期间的减损。仪表盘摄像头提供包括反应时间在内的时间背景 除了司机的安全措施外，还有意外危险，如留在车道内，维护 与周围车辆的距离。从这两个通道传输到云的数据将用于推导 驾驶风险的认知/驾驶质量指数(QI)，可与增加的压力水平、注意力 个体的缺陷和/或认知能力下降。这种模式将产生一个附加自适应平台 衡量认知状态/衰退的汽车也为汽车保险业提供了价值，RMV 作为认知障碍患者的照顾者，除了个人本身，还可以确定一个可以量化的 在放弃驾驶或寻求额外的认知诊断之前，请检查参数。 该综合驱动器监控系统的基本功能将在第一阶段进行评估 驾驶模拟器中的个体数量。该技术将在第二阶段得到改进，并将进一步 在真实的驾驶场景中进行了测试。

Project Summary/Abstract Physical Sciences Inc. (PSI), in collaboration with Michigan State University (MSU) and the University of Michigan (U-M), proposes to develop a novel technology that will non-intrusively monitor driver cognitive decline and/or high-risk driving behavior that is especially suited for seniors (aged >65). Contextual physiological biomarkers along with computer vision based drive quality measurements will be obtained through a cloud-based acquisition system. Real-time risk assessments that are measures of cognitive status will be performed on the cloud to provide comprehensive analysis on drive quality metrics for seniors with varying levels of cognitive impairment. Assessment of a driver’s cognitive status requires the use of a complex monitoring platform that can be seamlessly installed into any automobile without interfering or adding complexity to driving. Therefore, our goal is to develop a cloud-based, multi-modal routine cognitive/drive assessment system that combines a Smart Steering Sleeve (S3), which monitors physiological biomarkers, along with a Computer Vision (CV) system that provides contextual insight for biomarker changes and monitors a driver's habits. The Cloud-based Computer Vision Smart Steering Sleeve (C-CVS3) sensor system will be capable of monitoring biomarkers such as heart rate, hand gripping force, and the presence of excessive sweat from the skin surface indicating stress levels affecting potential drive quality degradation, which are also factors expected to be associated with cognitive impairment during critical events. Dash cameras provide temporal context including reaction times to unexpected hazards in addition to driver’s safety measures, such as staying within the lane, and maintaining distance to surrounding vehicles. The transmitted data to the cloud from both modalities will be used to derive cognitive/drive quality indices (QI) for driving risks that can be associated with increased stress levels, attention deficiency and/or cognitive decline in individuals. This modality will result in an add-on adaptive platform in automobiles that gauges cognitive status/decline providing value to the auto insurance industry, RMV, as well as to the caregivers of the cognitively impaired patient apart from the individual himself to identify a quantifiable parameter before giving up driving or to seek additional cognitive diagnosis. The basic functions of this comprehensive drive monitoring system will be evaluated in Phase I on a number of individuals in a driving simulator. The technology will be improved in Phase II and will be further tested during real-world driving scenarios.