A Performance Study of Genetic Algorithm-Assisted Beamforming in Distributed Cognitive Radio Networks

Cognitive Radio (CR) is a technology that has gained much interest recently due to the increasing scarcity of the radio frequency spectrum. Large portions of the radio frequency spectrum are licensed to users who then have exclusive access to the bandwidth, and unlicensed bands can be a challenge to use due to interference from unlicensed users. Despite the seeming scarcity, tests of bands allocated by the Federal Communications Committee (FCC) to licensed and unlicensed user have shown that many are underutilized and often unoccupied by the user to whom they are licensed. CR aims to exploit this unused spectrum and thus use it more efficiently. This must be accomplished in a way that does not obstruct licensed communications. To achieve this, CR must sense the availability of the spectrum in one or more dimensions such as space, time, frequency, etc. and adjust its parameters to communicate in the unoccupied spectrum. In this thesis, a CR network uses beamforming to communicate in the presence of a primary network such that a beam is steered towards cognitive receivers while a null is steered toward primary receivers to prevent interference. As the number of users in the primary and secondary networks increase and the constraints become more varied, a closed-form solution becomes more difficult to find. Due to the potentially dynamic nature of the network parameters and constraints, it would be of interest to study the application of a general algorithm that can be used to solve the problem. The Genetic Algorithm (GA) is a broadly applicable algorithm inspired by evolutionary biology in which solutions are encoded onto “chromosomes” and go through a process of natural selection to optimize some function. For this work, the GA was used to optimize the CR beamforming problem under various networks configurations and the effect of the GA parameters on its performance were studied. Adviser: Yaoqing (Lamar) Yan