Non-Contact Assessment of a Driver's Vital Signs - Driving Awareness

功能齐全、价格合理、大规模购买的自动驾驶汽车（AV）将需要很多年的时间。目前的半自动驾驶汽车（semi-autonomous vehicles，简称SA）在无人驾驶汽车无法处理的情况下，需要驾驶员监督并立即接管，这种汽车将得到广泛应用。监督自动驾驶系统的驾驶员必须保持对道路和车辆的态势感知。如果需要的话，监控司机注意力水平的方法有限。这种系统的一个应用是监测注意力不集中的司机，包括分心驾驶或使用麻醉剂，还有可能因衰老或疾病而驾驶能力下降的老年司机。该项目将研究使用环境传感技术评估驾驶员生命体征的技术，该技术可以集成到普通车辆功能中。它将把申请人的研究结合并扩展到驾驶员健康和驾驶行为之间的联系，并将环境传感器的新应用用于评估福祉。通过非接触视频放大可见光和热视频环境评估将探索评估心脏和呼吸速率。尽管两个不同身体位置的心率可以确定脉搏传递时间（PTT），但环境感知血压仍然是一个重大挑战。PTT的变化与相对血压的变化相关。手和头是评估心率的两个潜在位置，因为血压变化可能表明压力、焦虑或困倦。通过将头部跟踪、眼动跟踪和眼睑分析相结合的面向驾驶员的热视频和可见光视频，可以前瞻性地进行注意力分析。在该项目中，广泛的指标将使用数据融合技术进行整合，以获得驾驶员注意力以及控制车辆的一般能力的测量。一个关键的焦点是对车内条件具有鲁棒性的分析技术，这需要创建针对环境光照水平、运动、摄像机位置和环境温度变化的算法解决方案。申请人在复杂算法和分析方面的背景将有助于解决这一系列复杂的问题，包括从大型老年驾驶数据集中获取见解的众多出版物。申请人在SAM3创新中心担任领导职务，与国际临床研究界建立了工作关系，开发了用于检测和测量健康的遥感应用。申请人目前正在使用驾驶模拟器，该模拟器允许在受控条件下进行研究实验，验证和测试场景，同时确保安全。此外，申请人还可以使用渥太华地区X.0自动驾驶汽车测试轨道，这是一个独特的、高度仪器化的设施，用于使用实际车辆进行测试，包括自动驾驶汽车和自动驾驶汽车驾驶应用。

Fully functional, affordable, and mass purchased Autonomous Vehicles (AV) will take many years. Current semi-autonomous vehicles (SA) that require drivers to supervise and immediately take over in situations the SA vehicle can't handle will be used extensively. Drivers supervising the SA system must maintain situational awareness of the road and vehicle. There are limited ways to monitor for a driver's attention levels to take over, if needed. One application for such a system is to monitor inattentive drivers including distracted driving or the use of intoxicants and there is also aging drivers who may be declining in driving competencies because of aging or illness. This program will research techniques for assessing driver vital signs using ambient sensing techniques that could be integrated into ordinary vehicle functionality. It will combine and expand research by the applicant into the connection between drivers' health and driving behaviours with work on novel applications of ambient sensors for the assessment of well-being. Ambient assessment through non-contact video magnification of visible light and thermal video will be explored to assess heart and respiration rate. The ambient sensing of blood pressure remains a significant challenge, though heart rate from two distinct body locations allows Pulse Transit Time (PTT) to be determined. Changes in PTT correlate with changes in relative blood pressure. Hands and head provide two potential locations for assessment of heart rate as blood pressure change can indicate stress, anxiety or drowsiness. Analysis of attention can prospectively be conducted through driver-facing thermal and visible light video that combines head tracking, eye tracking and eye lid analysis. In the program, a wide range of indicators will be integrated using data fusion techniques to obtain measures of driver attentiveness as well as general capability to take control of the vehicle. A key focus is analysis techniques that are robust to in-vehicle conditions, requiring the creation of algorithms solutions for variations in ambient light level, motion, camera location, and ambient temperature. The applicant's background in complex algorithms and analytics will help address this complex set of problems, including numerous publications capturing insights from a large geriatric driving dataset. From his leadership position within the SAM3 Innovation Hub, the applicant has working relationships with an international clinical research community, developing remote sensor applications to detect and measure well-being. The applicant is currently working with driving simulators, which allow for research experimentation, validation and testing scenarios in controlled conditions while also ensuring safety. In addition, the applicant has access to the Ottawa Area X.0 Autonomous Vehicle Test track, a unique, highly-instrumented facility for testing using actual vehicles, including AV and SA vehicle driving applications.