Cross-layer performance analysis and adaptation for real-time wireless video streaming

Abstract

OF THE DISSERTATION Cross-Layer Performance Analysis and Adaptation for Real-Time Wireless Video Streaming by Michael T. Loiacono Dissertation Director: Dr. Wade Trappe The proliferation of wireless technology, mobile computing, and increasingly sophisticated video codecs have fueled the sharp increase in demand for user and machinecentric applications (such as IPTV, telemedicine, cyber-physical control, and surveillance) involving wireless video streaming. Unfortunately, today’s state of the art local wireless technologies, such as IEEE 802.11, can be easily demonstrated to fail when subjected to the conditions associated with many of these real-world applications. This is because wireless video streaming involves complex relationships between the the video codec, the wireless PHY and MAC, and the application’s (or user’s) sensitivity to distortions in the video signal, and these relationships are not well understood or exploited by today’s video streaming systems. In this dissertation, we reveal and analyze several key obstacles to wireless video streaming, and propose a set of adaptive real-time cross-layer approaches to deal with them. The dissertation begins with an overview of the major hurdles to wireless video streaming. A particular focus is on the instability of current link adaptation algorithms in CSMA/CA based wireless systems under the congested scenarios associated with wireless transmission of multiple simultaneous uplink video streams. Additionally, we focus on congestion control with respect to the fairness policies employed by the IEEE