Cognitive link layer for wireless local area networks

Several attempts have been made to optimize a link layer of wireless local area networks. Most of them focus on one parameter at a time and propose optimizing approaches which rely on network state information usually not directly available at network nodes. In this paper, we propose a cognitive link layer optimization capable of taking large number of parameters into account and perform optimization with no explicit knowledge on the causes of the performance degradation at the end-to-end paths. Performance evaluation results demonstrate advantages of the proposed cognitive link layer over the case when link layer parameters are fixed during runtime.