行人过马路安全中的人因及安全辅助设计策略研究
文献类型：学位论文
| 作者 | 庄想灵
|
| 学位类别 | 博士 |
| 答辩日期 | 2015-05 |
| 授予单位 | 中国科学院研究生院 |
| 授予地点 | 北京 |
| 导师 | 吴昌旭 ; 张侃 |
| 关键词 | 行人安全 人因 情境意识 违规行为 安全辅助设计 |
| 其他题名 | Pedestrian Safety in Crossing Roads: Human Factors and Safety Assistance Design Strategies |
| 学位专业 | 心理学 |
| 中文摘要 | 行人安全，尤其是行人在过马路中的安全是道路交通安全中亟待解决的问题。在多种影响行人安全的因素中，驾驶员与行人的人因（Human Factors）是最重要的因素，因此，本研究将分别从驾驶员和行人的角度分析保护行人过马路安全要考虑的人因，并进一步探索可能的安全辅助设计策略，提升行人过马路安全。 根据对以往文献和事故类型的分析，辅助驾驶员决策的重点是利用行人的行为特征减少驾驶员与行人交互中的情境意识（对行人行为的知觉、解释和预测）错误，减少违规行为；而辅助行人做出正确决策的关键在于根据行人违规的机制减少行人闯红灯和在清空时期违规。本论文中，研究一和研究二针对前者，研究三和研究四针对后者。 为了辅助驾驶员对行人行为的预测，研究一基于作者早期的观察和对行人的访谈结果，将行人过马路的过程描述为不断权衡知觉到的风险和过马路的效率进行动态决策的过程，建立了行人在多车道路段过马路的路径产生模型，并从真实环境中采集数据对模型进行了验证。为辅助驾驶员对行人过马路意图的理解，研究二探索了行人使用手势与驾驶员沟通的可能性。根据驾驶员对11个手势在可视性、含义、含义明确度、熟悉性和礼貌度的评价结果选出四个手势，并在现实场景中测试行人对驾驶员使用不同手势时，驾驶员的让行行为及其它反应。结果发现左手臂弯曲平掌可将驾驶员让行比例从3.5%增加到12.9%，将减速通过的驾驶员从32.9%增加到48.2%，而且没有明显的副作用。 辅助行人做出正确决策，减少违规的前提是明确影响行人违规的因素。在闯红灯违规中，以往研究发现行人知觉到的等待时长是较为重要的预测因子，因此研究三主要探索了如何通过行人信号灯的红灯时期的设计减少行人知觉到的等待时长。其中，研究三A基于时长估计的注意闸门模型，利用问卷调查了行人在等待红灯时缓解焦虑的策略及信息需求。调查发现了可在行人等红灯时显示的24类碎片化信息，并进一步将其聚为四类，按照偏好从高到低为：与过马路相关的信息、情境特征信息、学习欣赏类信息和娱乐广告信息。其中，行人对娱乐广告信息普遍持负性态度，而对过马路有关信息较为欣赏。同样为了减少行人知觉到的等待时长，研究三B根据现实场景中的时长估计特点和以往人机交互的研究成果，在模拟实验中比较了倒计时信号灯的红灯倒计时方式对行人知觉时长的影响，发现由于等红灯时注意分配的特征，信号灯早期变化的速度往往决定了行人对信号灯时长的最终估计，先均匀分布然后指数递增的计时方式被知觉到的时长最短。 与行人闯红灯违规相比，行人在清空时期违规（在我国，指绿闪时期进入人行横道）只在一些研究中提到，但违规行为的机制尚不清楚，更没有针对此类违规行为的干预措施。因此，研究四首先分析了影响行人在清空时期违规的因素，然后依据可操纵的影响因素探索清空时期行人信号灯的设计，以辅助行人做出正确的判断。其中，研究四A观察了486名在绿闪时期到达的行人的过马路时机选择，发现85.2%的行人属于绿闪时期违规，且绿闪违规的行人与机动车辆的冲突次数显著高于绿闪不违规的行人。进一步的分析发现，年龄越低、路面上已经在过马路的人数越多，已经过马路的行人离目标越远则行人越容易在绿闪阶段过马路，有次任务或者距离停车较近（在路口同一侧）的行人更可能等待。虽然剩余时间和道路宽度对行人过马路的速度要求越高，行人越可能等待，但即使是速度要求达到7m/s时，78%的行人还是会选择立即通过马路。由于行人这种冒险行为是根据绿闪倒计时的剩余时间线索做出的决策，研究四B在现实场景中模拟了行人在绿闪时期对不同过马路难度线索的判断过程，相对于以剩余时间为线索，在以速度需求为难度线索时，不仅能够让行人有更保守的判断标准，与其真正的过马路能力相匹配，而且对不同场景下过马路难度的分辨力也更高。 本文研究结果不仅在理论上促进了对行人行为的理解，也为车载驾驶辅助系统设计和以行人为中心的信号灯安全辅助设计提供依据，因此，对最终提高行人安全具有重要现实意义。 |
| 英文摘要 | Pedestrian safety, especially safety in crossing roads has been a challenge for many developing countries, including China. Among contributing factors like roadway environment and vehicle design, human factors in pedestrian and drivers played a vital role in pedestrian crashes. Therefore, this thesis surveyed important human factors that should be concerned in promoting pedestrian safety when crossing roads, and further explored design strategies for safety assistance. Human factors in pedestrian safety on the drivers' aspect mostly concerned with drivers' situation awareness, i.e. their perception, interpretation and prediction of pedestrian behaviors, thus it is crucial to understand pedestrian behaviors to facilitate drivers' situation assessment. In line with this goal, Study 1 modeled pedestrian paths in crossing unmarked roadway as a result of dynamic decision making in balancing perceived risk and crossing efficiency. The model was calibrated and validated with real data, and has good fitness. Study 2 explored how pedestrians could use gestures to communicate their intention to cross road at uncontrolled crosswalks to facilitate driver interpretation. Eleven gestures were evaluated by drivers in terms of visibility, meaning, clarity, familiarity and courtesy, of which four were tested in real scene. Experiment shows that the gesture with left arm bent and hands level were effective in promoting driver yielding behavior, while cause no significant increase in side effects. On the pedestrians' aspect, one critical countermeasure to increase pedestrian safety is to reduce pedestrian violations in signalized crosswalks, including running against red light and violation at the clearance phase. Since pedestrian perceived waiting time is an important determinants of their red light running behaviors, Study 3A and 3B aims to decrease their perceived waiting time via display design of signal lights. More specifically, Study 3A explored pedestrian information needs in wafting red light, as the attention-gate model of duration estimation suggest that when people are distracted on non-temporal information, their estimated duration will be shorter. Survey results shows that among the four types of information pedestrian mentioned, pedestrian preferred the information relate with road crossing task most, followed by context information, learning and appreciation information. Pedestrians generally pose a negative preference towards public entertainment information. Further analysis shows that females and highly educated people have higher preference over context information than their counterparts. Study 3B reduced pedestrian perceived waiting time from a different approach, given that time estimation in daily life has greatly relied on external temporal clues, and researchers in HCI has found effects of different progress bars on perceived duration. The experiment compared pedestrians estimated duration of red light strips with five types of counting down styles. Pedestrians seem to report waiting time based on the time distribution feature at the early phase of red light, and has a lower estimation when stripes at early phase has shorter durations. Pedestrian violations at clearance phase have been observed in some previous work, but predictors and countermeasures of this type of violation haven’t been explored. Therefore, before developing countermeasures in Study 4B, an observational study of pedestrians' choices after their arrival at the clearance phase (in China, it means flashing green phase, FG) was conducted in Study 4A. Among the 486 pedestrians observed, 85.2% committed a violation, and have significantly more conflicts with vehicles. Binary logit model suggest that a pedestrian is more likely to violate during FG when they are younger, has no secondary tasks, at the different side of the intersection with stopped vehicles, there has been more people crossing the street and those crossers were farther from destinations when the pedestrian arrived, and speed requirement for them to cross successfully (time remained/road width) was lower. Although pedestrian are more likely to wait when speed requirement is higher, 78% of the pedestrians choose to cross the road even when speed requirement is as high as 7m/s. Since pedestrian made their crossing decisions based on the remained time on the counting down lights, Study 4B explored whether this clue has induced risky decisions in pedestrians, and compared it with an alternative clue: speed requirement. Experiment in simulated crossing at real sites suggests that "remained time" do induce more complacent judgments of one's ability to finish crossing during FG than “speed requirement”. Compared with "remained time", "Speed requirement" resulted in much more conservative judgment criterion, which is compatible with pedestrians' real capability, and assisted pedestrians to have better ability to discriminate the crossing difficulty in various scenarios. Findings in this thesis not only improved understanding of pedestrian behaviors in crossing roads, but also have important implications for designing in-vehicle pedestrian protection systems to assist drivers, and user-centered design of traffic lights, which will finally improve pedestrian safety |
| 学科主题 | 应用心理学 |
| 语种 | 中文 |
| 源URL | [http://ir.psych.ac.cn/handle/311026/19793]  |
| 专题 | 心理研究所_社会与工程心理学研究室 |
| 作者单位 | 中国科学院心理研究所 |
| 推荐引用方式 GB/T 7714 | 庄想灵. 行人过马路安全中的人因及安全辅助设计策略研究[D]. 北京. 中国科学院研究生院. 2015. |
入库方式： OAI收割
来源：心理研究所
浏览0
下载0
收藏0
其他版本
除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。