| A Cognitive Radio (CR) Network is a mobile multi-hop wireless network consisting of CR nodes. A CR node has the ability to dynamically adapt to local spectrum availability and it can operate on any one channel from a set of (wireless) channels that are part of the frequency spectrum. The set of channels available to a CR node may vary with time, location and the actions of other entities such as licensed users. Layer-2 configuration for a CR node involves determining its neighborhood, a set of locally common channels that it can use to communicate with its neighbors and, if required, determining globally common channels for communication over the entire network. Due to node mobility and dynamism of the channel availability set, this problem of determining a node's neighborhood and the channels the node can communicate on is not trivial. In this work, we propose several time-efficient distributed layer-2 configuration algorithms for various CR node models. The algorithms are symmetric (no leader is needed and all nodes execute the same algorithm) and deterministic. Time is slotted and nodes have access to synchronized clocks. We also present lower bounds on the number of timeslots required for neighborhood discovery that indicate that our algorithms are either optimal or close to being optimal. |
