IRES Track II: US-Korea Advanced Transportation Infrastructure Informatics Institutes (ATI3)

交通运输的流动性和安全问题在国际上是一个极为重要的问题。因此，迫切需要开发一种校正和改进城市交通机动性和安全建模的方法，将这两个独立的研究领域联系起来。未能满足这一需求是一个紧迫的国家和国际问题，因为在建模方面缺乏进步的情况下，安全风险和流动性中断将继续导致浪费时间和不必要的生命损失，并继续成为经济负担。鉴于此，主要研究目标是协同美韩合作，通过利用人工智能（AI）创建统一的数据驱动算法框架，自主预测高速公路修复的移动性和安全影响。为了满足这一及时的需求，研究小组提出了一项新的倡议：每年举办三届先进交通基础设施信息研究所（ATI3），与韩国同行合作，培训美国学生的尖端技能。这些研究所将促进国际合作，将最佳的基础设施移动性和安全性分析实践协同集成到统一的人工智能数据驱动算法框架中，该框架在国家科学基金会“利用数据革命”和“融合研究”的大构想方面具有智力价值。该IRES将从参与的美国大学派遣15名研究生；每位学生每年将在韩国大田KAIST学习三周，为期三年。韩国独特的交通挑战和KAIST的解决方案将为美国学生在这一重要研究领域提供真正卓越的、丰富的、具有国际影响力的学习和研究经验。所提出的方法是独特的，因为它通过考虑大数据分析和人工智能技术（ATI3 I）、混合模拟融合的驾驶员行为建模（ATI3 II）和高保真度预测（将驾驶员行为作为主要影响因素，并使用强化人工智能深度学习算法（ATI3 III））来模拟移动性和安全中断的水平。ATI3的拟议研究活动将与一项全面的教育计划紧密相连，其总体目标是通过以项目为中心的研究环境促进学生的转型学习，同时广泛吸引实践者和社区参与以项目为基础的研究项目。中心假设是，建立一个为期三年的年度ATI项目，学习新的方法和技术，然后利用这些技术来模拟由于公路修复而导致的安全和流动性中断水平，这将使IRES研究员能够研究新的发现，这些发现可能会纠正和改进工作区安全和流动性建模的结果。新的安全-出行一体化系统将为比较分析康复方案提供严格的理论基础，从而以一种全新的方式评估驾驶员的不便和安全风险。它将为驾驶员的随机路线选择行为与其在交通队列延迟和碰撞风险中的后果之间的新相互作用提供见解。一旦成功完成，IRES将使研究界和实践者第一次看到系统估计方法，以确定最安全和最经济的交通计划，这些计划将比现有的更智能（更好的机动性，更少的旅行时间，更低的道路使用者成本）和更环保（减少车辆运营成本和环境成本）。该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

Transportation mobility and safety problems are of extreme importance internationally. Accordingly, there is an urgent need to develop a means of correcting and improving urban transportation mobility and safety modeling that connects these two independent research domains. The failure to meet this need is a pressing national and international problem because in the absence of advancements in modeling, safety risk and mobility disruption will continue to result in wasted time and unnecessary loss of life and remain a burden on the economy. In light of these, the key research objective is a synergistic US-Korea collaboration to create a unified data-driven algorithmic framework for autonomously predicting mobility and safety impacts of highway rehabilitation by harnessing artificial intelligence (AI). To meet this timely need, the research team proposes a new initiative: three annual Advanced Transportation Infrastructure Informatics Institutes (ATI3) to train U.S. students in cutting edge skills in collaboration with counterparts in Korea. The institutes will catalyze an international collaboration where the best infrastructure mobility and safety analysis practices are synergistically integrated into a unified AI data-driven algorithmic framework that has intellectual merit with regards to the NSF Big Ideas of ‘Harnessing the Data Revolution’ and ‘Convergence Research’. This IRES will deploy 15 graduate students from participating U.S. universities; each student will spend three weeks annually over the three-year duration at KAIST in Daejeon, Korea. Korea’s unique transportation challenges and KAIST’s solutions will offer U.S. students truly exceptional, enriching international high-impact learning and research experiences in this important research area. The proposed approach is unique because it models the level of mobility and safety disruption by accounting for big data analytics and AI techniques (ATI3 I), drivers’ behavior modeling fused from hybrid simulations (ATI3 II), and high-fidelity prediction that accounts for drivers’ behavior as a major influence with a reinforcement AI deep-learning algorithm (ATI3 III). The proposed research activities at ATI3 will be tightly interwoven with a comprehensive education plan, with the overall goal of promoting students’ transformational learning through a project-based research-centric environment while widely engaging practitioners and communities in the project-based research projects. The central hypothesis is that instituting a three-year annual ATI program for learning new methods and techniques, then leveraging the techniques to model the level of safety and mobility disruption due to highway rehabilitation, will allow the IRES fellows to research new discoveries that may correct and improve the results of work zone safety and mobility modeling. The new safety-mobility integration system will provide a rigorous theoretical basis for comparatively analyzing rehabilitation alternatives so that motorist inconvenience and safety risk can be assessed in a fundamentally new way. It will provide insights into new interactions between drivers’ stochastic route choice behaviors and their consequences in traffic queue delays and crash risks. Once successfully completed, the IRES will result in the research community and practitioners with the first view of a systematic estimation method to determine the safest and most economical transportation plans that would be smarter (better mobility, less travel time, and lower road user cost) and greener (reduced vehicle operating costs and environmental costs) than those in existence today.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.