MRI: Acquisition of Connected Autonomous Vehicles (CAV) Infrastructure to Support Cooperative Human-Robot Driving and Pedestrian Safety

该项目通过购买和部署先进的环境传感器和车辆对一切(V2X)通信设备，为创建内华达大学拉斯维加斯分校(UNLV)混合驾驶(混合)“活实验室”提供资金。该设备将增强UNLV的基本自动化车辆演示平台，并为校园附近的至少三个十字路口配备高分辨率摄像头、雷达和激光雷达以及V2X通信，以实现车辆基础设施协作传感和安全。该设备将使互联和自动驾驶汽车(CAV)的新研究成为可能，重点是能够支持从人类驾驶员向全自动驾驶汽车过渡的人工智能(AI)，并为研究提供现实世界的试验台。虽然从车辆或基础设施进行环境传感方面取得了重大进展，但关注它们之间的合作和协作的工作要少得多。该项目将考虑V2X连接如何实现车辆和基础设施之间的信息共享，以及如何开发利用各自优势的人工智能算法-例如，从基础设施可观察到的高可用性和行为多样性的车辆的高分辨率测量。这项研究将考虑i)车辆-基础设施协作传感和控制，ii)道路行为建模，以便更好地理解人类行为，以便在混合交通情况下更少地保守和更自然地进行CAV操作，iii)基于计算机视觉的人工智能算法开发，用于持续适应，以及iv)通过轨迹分析实时、连续地表征交叉口的安全。混合项目将促进未来自动化车辆和人类之间的合作，并实现更安全、更自然的互动。这项研究将明确考虑人类的行为和意图，而不是道路规则，以确保安全，以改善行人和其他易受伤害的道路使用者周围的骑兵的安全。作为一个为少数族裔和西班牙裔服务的机构，该项目将为UNLV的不同学生群体提供技术技能和在自动驾驶汽车、人工智能和计算机视觉等不断增长的领域的独特学习和培训经验。完整的项目详细信息，包括出版物、数据集和代码，请访问https://MixeD.sites.unlv.edu.该网站将维护至少五年。这一奖项反映了NSF的法定使命，并通过使用基金会的智力优势和更广泛的影响审查标准进行评估，被认为值得支持。

This project funds the creation of the University of Nevada, Las Vegas (UNLV) Mixed Driving (MixeD) “living laboratory” through the purchase and deployment of advanced environmental sensors and vehicle-to-everything (V2X) communication equipment. The equipment will augment UNLV’s basic automated vehicle demonstration platform and equip at least three intersections adjacent to campus with high-resolution cameras, radars, and lidars and V2X communication for collaborative vehicle-infrastructure sensing and safety. The equipment will enable new research in connected and autonomous vehicles (CAVs), with emphasis on artificial intelligence (AI) that can support the transition from human drivers to fully self-driving cars, and provides a real-world testbed for the research.While significant advances have been made in environmental sensing from a vehicle or infrastructure, much less work has focused on their cooperation and collaboration. This project will consider how V2X connectivity can enable sharing of information between vehicles and infrastructure and how to develop AI algorithms that take advantage of the strengths of each – e.g., high resolution measurements from a vehicle with the high availability and behavior diversity observable from infrastructure. The research will consider i) collaborative vehicle-infrastructure sensing and control, ii) on-road behavior modeling for better understanding of human actions for less conservative and more natural CAV operation in mixed traffic situations, iii) computer vision-based AI algorithm development for continual adaption, and iv) real-time, continuous characterization of safety at intersections through trajectory analysis. The MixeD project will promote collaboration between future automated vehicles and humans and enable safer and more natural interactions. The research will explicitly consider human behavior and intentions rather than rules-of-the-road to ensure safety to improve the safety of CAVs around pedestrians and other vulnerable road users. As a Minority- and Hispanic-Serving Institution, project will provide UNLV’s diverse student population with technical skills and unique learning and training experience in the growing areas of self-driving cars, AI, and computer vision. Full project details, including publications, datasets, and code, are available at https://MixeD.sites.unlv.edu. The website will be maintained for at least five years.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.