PSO based Cross Layer Optimization for Primary User selection in Cognitive Radios

Abstract

In Cognitive Radio Networks, selection of primary users (PUs) is an important process for achieving optimal performance during a session. The selection process of PU is limited due to the strict boundaries between layers that are enforced in Open System Interconnection (OSI) model which prevent the coordination, interaction and data transfer between the layers. To overcome such limitations, cross layer optimization is proposed where different operating parameters such as transmission power, packet length, bandwidth etc. across the OSI layers of a device are optimized. In this paper Particle Swarm Optimization (PSO) algorithm is proposed to optimize different operating parameters with the objective to optimize throughput, power consumption, interference, Bit Error Rate and spectral efficiency for a set of PUs across the physical, network and Media Access Control (MAC) layer in OSI model. The fitness values of these objective functions in different modes and channels are investigated using MATLAB and the results shows that PSO is 70% faster than the Genetic Algorithm in terms of convergence rate. Finally paper proposes that PSO based algorithm is an efficient, reliable and fast technique for primary user selection in cognitive radios.