Worcester Polytechnic Institute (WPI) researcher Tian Guo has been awarded a prestigious $657,776 CAREER Award from the National Science Foundation to develop novel software techniques that will improve the performance and privacy of mobile augmented reality (AR) systems, an increasingly popular technology that superimposes computer-generated images on a user's view of the real world.
Guo, an assistant professor in the Department of Computer Science, will focus her five-year project on edge computing, which involves processing data close to its physical source. She will develop techniques to efficiently manage edge servers, that are close to AR users whose mobile devices are interacting with the servers. The proposed techniques will be prototyped with commercially available edge servers, and the resulting software and hardware bundles will be deployed to support indoor AR use cases.
“Augmented reality is an emerging interactive communication mechanism with the power to create experiences for users that go beyond pre-recorded media like traditional 2D video streaming, but new mechanisms are needed to enable such interactive experiences, as well as new policies to deliver good user experiences,” Guo said. “One goal of this project is to design a generic pipeline that connects sensors, either on AR devices or embedded within the environment, with AR applications, providing the capability to capture high-quality data at the right time and make that data available to AR developers.”
Guo will develop noise-tolerant and mobility-aware techniques that will better capture high-quality data from AR devices, such as handheld smartphones or headsets with cameras. She also will determine how to efficiently coordinate and schedule AR tasks on edge computing clusters to produce faster, improved visual experiences for AR users. During later stages of the project, Guo will build and deploy edge cluster testbeds that could be used to evaluate AR applications such as visualizing the molecular structure of proteins and touring an art museum.
Part of the project will focus on developing techniques to preserve the privacy of AR users, whose cameras can collect sensitive data, such as images and videos, at a scene. Guo will develop lightweight yet effective techniques to remove sensitive information from scene data, such as family photos or credit cards, so that private data is not shared with other AR users, without impacting the visual experiences.
Two graduate students and several undergraduate students will assist Guo with her project, which builds on her previous NSF-funded research that pinpointed performance bottlenecks in mobile deep-learning applications and developed improvements.
“AR is an emerging interactive communication tool that creates powerful storytelling experiences, and it’s important to me to contribute to this field by doing fundamental computer science research on the infrastructure behind augmented reality,” Guo said. “My hope is that as this technology moves forward, it will attract artists and others from diverse backgrounds who will similarly contribute their talents to this field to make AR even more compelling for end-users.”
