Analysis of Delay Distribution and Rate Control over Burst-Error Wireless Channels Analysis of Delay Distribution and Rate Control over Burst-Error Wireless Channels
In real-time communication services, delay constraints are among the most important QoS (Quality of Service)factors. In particular, it is difficult to guarantee the delay requirement over wireless channels, since they exhibitdynamic time-varying behavior and even severe burst-errors during periods of deep fading. Channel throughputmay be increased, but at the cost of the additional delays when ARQ (Automatic Repeat Request) schemes areused. For real-time communication services, it is very essential to predict data deliverability. This paper derivesthe delay distribution and the successful delivery probability within a given delay budget using a priori channelmodel and a posteriori information from the perspective of queueing theory. The Gilbert-Elliot burst-noise channelis employed as an a priori channel model, where a two-state Markov-modulated Bernoulli process (MMBP2) isused. For a posteriori information, the channel parameters, the queue-length and the initial channel state areassumed to be given. The numerical derivation is verified and analyzed via Monte Carlo simulations. Thisnumerical derivation is then applied to a rate control scheme for real-time video transmission, where an optimalencoding rate is determined based on the future channel capacity and the distortion of the reconstructed pictures.
In real-time communication services, delay constraints are among the most important QoS (Quality of Service)factors. In particular, it is difficult to guarantee the delay requirement over wireless channels, since they exhibitdynamic time-varying behavior and even severe burst-errors during periods of deep fading. Channel throughputmay be increased, but at the cost of the additional delays when ARQ (Automatic Repeat Request) schemes areused. For real-time communication services, it is very essential to predict data deliverability. This paper derivesthe delay distribution and the successful delivery probability within a given delay budget using a priori channelmodel and a posteriori information from the perspective of queueing theory. The Gilbert-Elliot burst-noise channelis employed as an a priori channel model, where a two-state Markov-modulated Bernoulli process (MMBP2) isused. For a posteriori information, the channel parameters, the queue-length and the initial channel state areassumed to be given. The numerical derivation is verified and analyzed via Monte Carlo simulations. Thisnumerical derivation is then applied to a rate control scheme for real-time video transmission, where an optimalencoding rate is determined based on the future channel capacity and the distortion of the reconstructed pictures.
※ AI-Helper는 부적절한 답변을 할 수 있습니다.