A New Game Theory Approach for Noise Reduction in Cognitive Radio Network

Noise reduction is one of the fundamental signal processing functions of the cognitive radio system, as noise affects data integrity and obscures the using of the unused channels in the licensed spectrum of Primary Users (PU). The majority of traditional filter methods handles only specific types of noises in a certain range of frequencies. Despite the intensive research in this area, building accurate noise reduction engines that equilibria between signal-to-noise ratio and computational cost remain a holy grail. In this paper, we made our initial effort towards this issue by building a novel game theory engine for effective noise reduction that incorporates the advantage of both adaptive filtering and wavelet denoising strategies. Adaptive filter solves the convergence problem and low mean square error rate. Wavelet transform as a nonstationary signal analysis can describe signal's local features either spatially or spectrally and obtains the asymptotically optimal estimation of original signals. Game theory (GT) is adapted for optimal decision making under competition; so that it chooses the best strategies to achieve the best noise reduction equilibrium when others are affected by the decision of the individual player, as the experiments indicate that the proposed model on its ability to reduce noise significantly.