CAREER: Toward a Specialized Edge for Augmented Reality

点击翻译按钮获取中文摘要

Augmented Reality (AR) applications are a popular media that benefits millions of users. Domain experts and the public use AR applications for daily tasks, including safety training, tourism, and online shopping. However, existing edge-assisted AR systems often fall short in handling noisy sensor data, managing edge resources under dynamic workloads, and mitigating visual privacy risks. The proposed work will lay the scientific foundations for a specialized edge by designing a generic device-to-edge pipeline that AR developers can use to implement and deploy efficient and privacy-preserving AR tasks. The proposed techniques will complement and extend existing approaches for mobile AR frameworks and edge resource management. The proposed research thrusts include: (1) techniques for high-quality data capture, adaptive and scalable data sharing, and resource-efficient execution; (2) methods and datasets for AR researchers to gain new insights, and resource management techniques for developers to improve the performance of their AR applications; and (3) privacy-preserving approaches that help AR developers mitigate privacy risks associated with AR development and deployment. The proposed techniques will span development through deployment and improve AR applications' end-to-end performance. The PI regularly open-sources systems and datasets, promoting adoption and benefiting the research community. The research program will support the inclusion of undergraduate researchers by providing opportunities to students with diverse technical backgrounds to engage in mobile and edge computing research. The PI will create and pilot interactive course materials on mobile AR in the PI's existing courses and make them publicly available. Collaboration with AR technology leaders and domain experts will ease technology transfer and expose students to real-world AR products.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.