Date of Award

2017

Document Type

Open Access Master's Report

Degree Name

Master of Science in Computer Science (MS)

Administrative Home Department

Department of Computer Science

Advisor 1

Min Song

Advisor 2

Xiaohua Xu

Committee Member 1

Bo Chen

Abstract

Rendezvous is a fundamental process establishing a communication link on common channel between a pair of nodes in the cognitive radio networks. How to reach rendezvous efficiently and effectively is still an open problem. In this work, we propose a guaranteed cycle lengths based rendezvous (CLR) algorithm for cognitive radio networks. When the cycle lengths of the two nodes are coprime, the rendezvous is guaranteed within one rendezvous period considering the time skew between the two nodes. When Ti and Tj are not coprime, i.e., Ti=Tj, the deadlock checking and node IDs are combined to decide the time point and the way to independently change the cycle length on each node to guarantee rendezvous. In detail, as long as the deadlock situation is detected based on the threshold, each node can independently change its cycle length be based on the current checking bit of the node ID. The threshold used for deadlock checking is defined as the length of the maximum possible rendezvous period between the two nodes. As long as the current checking bits between the two nodes are different, the rendezvous will be reached in the following rendezvous period, The theoretical analysis also proves the guarantee of the CLR algorithm under both the two cases. We use three metrics: success rate of rendezvous, expected time to rendezvous and channel load to conduct simulation studies. The simulation results show that the CLR algorithm always has higher successful rendezvous rate of 100%, and stable and low expected time to rendezvous compared to the HH algorithm. In addition, the channel loads are smoothly distributed on all channels with CLR, while HH algorithm depends on the channels with smaller IDs.

Share

COinS