Momma Hiroshi, Suzurikawa Jun, Takeshima Rie, Sawada Yuki, Harada Yusuke, Kondo Tomoko
Annu Int Conf IEEE Eng Med Biol Soc. 2020 Jul;2020:3212-3215. doi: 10.1109/EMBC44109.2020.9176296.
A mobility scooter is a major assistive technology that replaces human ambulatory functions for people with disabilities. A license is often not required for driving a mobility scooter; therefore, less skilled drivers might create safety concerns. An effective way of reducing these safety risks is by assessing the driving skills of users. The existing assessment measures mostly score the task performance using manual observations. In this study, we have developed a novel assessment system that logs the driving operations by using add-on sensors. This system can monitor the operations of a mobility scooter including the angles of the throttle lever and the steering tiller. The subjects were seven older adults who participated in the driving test involving six tasks; the driver performances were video recorded, and the vehicle operation data were logged. The video analysis showed that two subjects crashed their scooters into objects or made contact with objects during the test course. To extract the characteristic patterns of the operations from the logs, 2D histograms of the operational status durations were investigated for each subject and task. Subsequent analysis of the operation logs identified the two subjects who had crashed their vehicles during the test drive. Our results proved that the driving operation logs could be used complementarily as a simple and low-cost tool for assessing a person's driving skills.
电动代步车是一种主要的辅助技术,可替代残疾人的行走功能。驾驶电动代步车通常不需要驾照;因此,驾驶技术欠佳的人可能会引发安全问题。降低这些安全风险的有效方法是评估使用者的驾驶技能。现有的评估措施大多通过人工观察对任务表现进行评分。在本研究中,我们开发了一种新颖的评估系统,该系统通过附加传感器记录驾驶操作。该系统可以监测电动代步车的操作,包括油门杆和转向柄的角度。受试者为七名老年人,他们参加了包含六项任务的驾驶测试;对驾驶员的表现进行了视频记录,并记录了车辆运行数据。视频分析显示,两名受试者在测试过程中将他们的代步车撞到物体上或与物体发生了接触。为了从日志中提取操作的特征模式,针对每个受试者和任务研究了操作状态持续时间的二维直方图。随后对操作日志的分析确定了在试驾过程中撞车的两名受试者。我们的结果证明,驾驶操作日志可以作为一种简单且低成本的工具,用于辅助评估一个人的驾驶技能。