Delaunay Triangulation Based Cognitive Information Compression for Cognitive Radio Networks

Geo-location database is an efficient way to delivery the spectrum occupancy information, which is defined as "cognitive information", to unlicensed users in cognitive radio networks (CRNs). In this paper, we study the compression of cognitive information in geo-location database supported CRNs. We propose a delaunay triangulation based triangle mesh division algorithm. It is proved that the RPE of triangle meshes is smaller than that of rectangle meshes when the number of meshes are the same. Then the triangle meshes are clustered to reduce the redundancy of cognitive information. Finally, the boundaries of the clusters are extracted, which form polygons and the cognitive information is further compressed. The numerical results are provided to verify the correctness of the theoretical analysis and the efficiency of the cognitive information compression algorithms.

I. Introduction

Cognitive radio system (CRS) was proposed to improve spectrum utilization and solve spectrum scarcity problem [2]. CRS is a system which can obtain the cognitive information that contains the knowledge of the environment and reconfigure itself to adapt to the environment [3]. The information on the available channels is defined as cognitive information in this paper [4]. The cognitive information can be obtained by spectrum sensing and geo-location database. As geo-location database alleviate computational complexity of unlicensed users, the geo-location database assisted CRNs have been widely studied. In geo-location database, the cognitive information delivery between the database and unlicensed users is critical to enable the spectrum sharing in CRNs.