Available bandwidth estimation in Wireless Mesh Networks using neural networks

In the next-generation home networks, Wireless Mesh Networks will be the solution to support future services and devices with an extended coverage of the home area. The increasing use of multimedia applications with high data rates is triggering the need for a better Quality of Service support in these IEEE 802.11-based networks; this relies on the ability to estimate the available bandwidth on each link. In this paper, we propose to use a neural network for this purpose. The proposed approach does not generate additional traffic and uses information a node can gather by monitoring its environment. Furthermore, our approach adapts to the particular conditions of the network thanks to the adaptive learning ability of the neural network.