Video Aware Wireless Networks (VAWN) Research Program

Monday, February 24, 2014

As the next generation of wireless networks develops, how to efficiently deliver video content is at the forefront of inquiry within the research community.

“Video is becoming a huge fraction of Internet traffic, and more and more people are using their smart phones, laptops or tablets to stream video. It is eating up all of the bandwidth,” Dr. Robert Heath, Director of the WNCG, states.

According to Dr. Heath, the problem today is how to send more video in a more efficient way within the framework of current wireless resources. Research is needed to explore how the unique characteristics of video can be used to the advantage of wireless networks as a whole. This is where the research of the Video Aware Wireless Networks Program (VAWN) within the WNCG takes the lead.

The VAWN project was a multi-year effort funded by Intel, Cisco and Verizon that involved several teams from UT Austin, Cornell, University of Southern California, University of California San Diego and Moscow State.  The project focused on the development of new algorithms for processing, transmitting and displaying video, and drew heavily from work in networking theory, image and video processing, perceptual quality assessment of images and video, machine learning and optimization and physical layer techniques.

The UT team included 10 students led by five WNCG professors: Robert Heath, Alan Bovik, Jeffrey Andrews, Gustavo de Veciana and Constantine Caramanis.

Several problems afflict the current usage of data infrastructure for streaming videos that the VAWN program addresses. Primarily, the current available data cannot meet the growth needs of video streaming. According to Dr. Alan Bovik, video traffic is projected to increase by up to 100 times the current rate within the next few years. 

One of the cornerstones of the WNCG team’s results focused on developing Perceptually AwareAlgorithms. The key idea is that the WNCG wants to optimize and operate wireless networks that transmit video to maximize the end user’s Quality of Experience given the availability of network resources.

“In recent years we’ve been learning a lot about the brain, how it processes visual information and specifically, how humans judge the quality of what they observe,” Dr. Alan Bovik says. Wireless networks have yet to adapt to how human brains and eyes receive and process information.

Dr. Alan Bovik focuses on this problem of perception through his research and the VAWN group believes that video networks should adapt to these perceptual distortions at every layer and time-scale of the larger network. 

WNCG faculty research focuses on the development of three main fields of inquiry. The first regards Video Quality Metrics for Adaptation (VQA) that seek to advance video quality assessment, improve distortion models that adapt easily to networks, decrease video distortion and develop models for determining 2D and 3D visual quality. This information would then determine network scheduling and adaptation decisions.

Spatio-temporal Interference Management (STIM) in Heterogeneous Networks serves as the second focus for VAWN. Under this research thrust, WNCG faculty explore how interference can be used efficiently, rather than avoided within wireless networks. The goal is to use the characteristics of real time and streamed videos to deliver videos efficiently via cellular systems and to use video buffering times to strategically schedule cellular transmissions.

The third main category within the VAWN program is Learning for Video Network Adaptation (LV). The LV thrust researches the implementation of a data-driven learning framework used to adapt wireless networks operations, create perceptually efficient video transmissions and predict user demand to smooth traffic over wireless links.

“We want to be relative to human judgments, to evaluate different types of video quality, different types of distortion on video, to see how people respond so that we can have models pertinent to what people will actually experience,” Dr. Gustavo De Veciana, VAWN researcher and WNCG faculty member, states.

The VAWN project produced many cutting-edge research results published in top peer-reviewed venues. More importantly, it produced many of the new ideas that will drive the growth and development of next generation wireless systems. These ideas fueled the Ph.D.’s of many WNCG students who will be the innovators and leaders of wireless networks of the future.