Distributed Beamforming for Cognitive Radio Networks

Abstract

Cognitive radio (CR) is an approach with high potentials in the effort of battling spectrum scarcity, introduced by J. Mitola in the late 90’s. It is capable of sensing the communication environment and adapting to it by changing its radio parameters. Beamforming technique allows the cognitive users to opportunistically access the licensed spectrum without interfering the licensed users by exploiting the spatial domain in the radio transmission. Distributed beamforming is a new concept to form communication beams by utilizing the distributed wireless nodes in order to transmit signals towards a distant destination. Applying distributed beamforming to the cognitive radio system gives the flexibility to the system, since the CR nodes are treated as virtual antenna arrays. CR also improves the lifespan of each node because the communication signal is generated in a distributed manner. This advantage of CR aligns with the principal of green communication which has the objective of less energy consumption. Power efficiency area is a particular field of interest in the cognitive radio domain, where cognitive nodes are small battery-powered devices, distributed randomly in an area. Optimal number of nodes selection is proposed in this thesis to optimize the power usage in the implementation of distributed beamforming in cognitive radio networks. The selection is based on the requirements of the primary licensed user in order to guarantee the quality of service (QoS) of the licensed user’s link. Recent works on the distributed beamforming field describe that the average beampattern shows a deterministic result and the mainlobe is independent of the particular node locations when the number of utilized nodes is considered as very large. However, for finite number of nodes, the beampattern does depend on the number of nodes and the way they are selected. In this thesis research, an approach is presented for deciding the number of nodes for distributed beamforming. Furthermore, several novel node selection algorithms such as Euclidean based, sector based, ring range, and circle range are proposed and their performances are studied in detail in this thesis work. In these entire schemes, the precise locations of distributed nodes of CR are needed. Further in this thesis, location errors of CR nodes and destination are taken into consideration. Its impact on the generated beampattern of the selected nodes is reported. The results show which of the proposed methods are more robust to the location information error of the CR nodes.