Apuntamiento automático de antenas para redes WiLD

Abstract

WiLD (WiFi for Long Distances) networks make use of the WiFi technology, primary meant to be used in short distances, to build point-to-point links in the range of some dozens of kilometers. In order to achieve this, the MAC layer of the IP/TCP stack should be modified to take in consideration some factors that are not that important in typical WiFi scenarios e.g. propagation delays. In such deployments, antennas are typically mounted on fixed masts that may not be easy to reach. For this reason, once the link is set up, antennas are perhaps not so easy to re-place or re-steer. This is just a small limitation, mostly solved with a remote controller for the position of the antenna. What a remote controller cannot solve is the limitation in these networks to have mobile endpoints. Since the use of directional antennas is a must in a typical WiLD link, these antennas must be perfectly steered in order for the link to work properly, thus a mobile node is just unfeasible. The only way to give directional antennas the capability to maintain a link where one or even two of the nodes are moving, is the development of a systems that could keep track of the link parameters and rotate the antennas so they can keep steered and the link up. Given that the possibilities could be wide opened, it is from our interest to research how the performance of WiLD links is when using such a mechanism. Of course the idea of auto-steering antennas is not a new, but since WiLD networks have such particular characteristics, factors like required computational power, speed response of the sensor and motors, etc. could vary from other environments.