CCRI: Planning: Collaborative Research: Low-Latency for Augmented Reality Interactive Systems (LLARIS)

增强现实和虚拟现实(AR/VR)技术可以极大地重塑人类交互、工业应用、城市规划、教育等许多领域。AR应用程序通常从物理设备捕获数据，将其发送到远程服务器(越来越多地通过云平台托管)进行处理，然后服务器将结果返回给应用程序。这些互动需要迅速发生。AR应用程序可以容忍的阈值网络延迟通常在10到50毫秒之间。但是，当涉及云托管服务器时，实际延迟可能远远超过50毫秒。由于光速限制了网络延迟，仅仅创建具有更多带宽的更快的网络并不能解决这个问题。相反，我们必须将计算移动到更接近网络边缘的位置，同时利用现有云平台的计算能力。该规划方案通过(1)利用协作规划(更多的想法、观点)来为人机交互研究人员开发更高质量、广泛使用的基础设施；(2)在规划阶段动员大量的人机交互和网络研究人员，从而鼓励和增加基础设施的采用；(3)提供可应用于多个研究领域(如机器学习)的有价值的、全面的发现，这将为研究人员提供新的信息来推进他们的工作；(4)为未来的AR基础设施制作概念验证原型，制造业、航空、教育、娱乐、国防等领域对AR的需求迅速增长。该项目探索如何为下一代AR系统构建下一代基础设施，这些系统将普及到人们的日常生活中。这个项目不仅对人机交互和网络社区做出了贡献，也对人类技术前沿(国家科学基金会的十大想法之一)的工作的未来做出了贡献，启动了计算机科学家和认知/学习科学家之间关于如何使AR应用学习更容易的合作探索，并制作了新的跨学科课程。这项工作奠定的基础为下一代研究基础设施和变革性的计算和通信系统铺平了道路。该奖项反映了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。

Augmented Reality and Virtual Reality (AR/VR) technologies can massively reshape human interactions, industrial applications, urban planning, education, and many other fields. An AR application typically captures data from a physical device, sends them to a remote server (increasingly hosted through cloud platforms) for processing, and the server then returns the results to the application. These interactions need to happen quickly. The threshold network delay that an AR application can tolerate is typically between 10 to 50 milliseconds. However, when cloud-hosted servers are involved, the actual latency can be much more than 50 milliseconds. Since the speed of light limits network latency, merely creating faster networks with more bandwidth will not solve this problem. Instead, we must move the computation closer to the network's edge while exploiting the computational capacity of the existing cloud platforms. This planning proposal advances state of the art by (1) leveraging collaborative planning (more ideas, perspectives) to develop higher quality, widely usable infrastructure for human-computer interaction researchers; (2) engaging a large group of human-computer interaction and networking researchers in the planning phase, which encourages and increases infrastructure adoption; (3) providing valuable, comprehensive findings applicable to multiple research areas, like machine learning, that will offer researchers novel information to advance their work; and (4) producing a proof-of-concept prototype for future AR infrastructure.The demand is rapidly growing for AR in domains such as manufacturing industry, aviation, education, entertainment, and defense. This project explores how to build the next-generation infrastructure for next-generation AR systems that will become pervasive in people's everyday life. This project not only contributes to the human-computer interaction and networking community but also contributes to the Future of the Work at the Human-Technology Frontier (one of National Science Foundation's 10 big ideas), launches collaborative explorations between computer scientists and cognitive/learning scientists on how to make learning AR applications easier, and produces new interdisciplinary courses. The groundwork laid by this work paves the path for the next-generation research infrastructure and transformative computing and communication systems.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.