C. Wang and T. Lin (Taiwan)
Wireless Technologies, 802.11 WLANs, QoS, Neural Networks
In error-prone IEEE 802.11 WLAN environments, link qualities can significantly affect the channel allocation among hosts and consequently the user perceived QoS of multimedia services. In this paper we propose an adaptive mechanism which on-line adjusts the initial backoff window size depending on the current channel state in order to provide fair multimedia QoS. It is a table-driven optimization approach which off-line pre-establishes the table of best initial window size associated with channel conditions based on a cost-reward function. Neural networks are utilized to learn the nonlinear mapping function and to generalize that to other channel conditions. A video streaming transmission scenario is used to evaluate the performance of our scheme. The results demonstrate the effectiveness of the proposed mechanism.
Important Links:
Go Back
