| Summary: | 碩士 === 國立清華大學 === 通訊工程研究所 === 95 === Abstract
With the increasing usage of wireless networks and the demand for multimedia
applications, such as video streaming, videophone, high definition television (HDTV) etc.,
it is very important to provide quality of service (QoS) for real-time applications over
Wireless Local Network (WLANs). To support QoS in WLAN, the 802.11 Working
Group has developed a new protocol, IEEE 802.11e, to provide service differentiation at
the medium access control (MAC) layer. Because of wireless channel characteristics,
such as shadowing, multi-path, fading, and interferences, the allocated bandwidth is
usually not fixed. It results in ineluctable degradation in QoS. Solutions to deal with
wireless bandwidth fluctuations consist of link adaptation (LA) and rate adaptation (RA).
However, these approaches rely on the slow feedback of physical layer to adapt
transmission rate and do not consider characteristics of traffic and codec. Thus, they may
not satisfy requirements of strict bandwidth guarantee and latency limitation for video
streaming. In this thesis, we propose a cross layer mechanism to improve H.264 video
streaming transmission over IEEE 802.11e wireless networks, when encountering
short-term bandwidth fluctuations. The proposed cross layer mechanism consists of slice
classification in the application layer, a dynamic packet selective transmission (DPST)
scheme in the MAC layer, and channel condition prediction using two-state Markov
chain. Performance of the proposed mechanism is evaluated through extensive
simulations. Results show that the proposed cross layer mechanism provides better video
quality and lower packet delay.
|
|---|
