Link quality estimators for multi-hop mesh network

In this paper, the problem of estimating the link quality in mesh networks has been considered. Such a problem is a major task to develop an efficient network layer, since an accurate knowledge of the link qualities allows routing protocols to efficiently use neighbors as relays for multi-hop communications. In the last years, a number of link-quality aware routing metrics based on the packet delivery ratio have been proposed and analyzed. However, very few works have addressed the problem of correctly estimating the delivery ratio of a wireless link. Therefore, in this paper we resort to a 802.11b mesh network to carry on a performance comparison between two widely adopted delivery ratio estimators, namely the Simple Moving Average and the Exponentially Weighted Moving Average, and a recently proposed one based on the neural network paradigm.