A Metaheuristic Based Approach for Threshold Optimization for Spectrum Sensing in Cognitive Radio Networks

Abstract

Background: The mounting growth of wireless technology is attracting a high demand for the frequency spectrum. The measurements of spectrum usage depict that a significant portion of the spectrum lays unoccupied or overcrowded. The main cause of the glitch is the existing inefficient and fixed scheme of spectral allocation. Cognitive radio is one such technology that permits wireless devices to detect the unused frequency band and reconfigure its operating parameters to attain the required quality of service. Objective: To permit the dynamic allocation of the frequency band, spectrum sensing is performed which is an essential function of Cognitive radio and involves detection of an unused spectrum space to set up a communication link. Method: This paper presents a meta-heuristic approach for selection of a decision threshold for energy detection based spectrum sensing. At low SNR and in the presence of noise uncertainty, the performance of energy detection method fails. A novel adaptive double threshold based spectrumsensing method is proposed to avoid such a sensing failure. Further, the metaheuristic approach employs Particle Swarm Optimization (PSO) algorithm to compute an optimal value of the threshold to attain robustness against noise uncertainty at low SNR. Results: The simulation results of the proposed metaheuristic double threshold based spectrum sensing method demonstrate enhanced performance in comparison to the existing methods in terms of reduced error rate and increased detection probability. Some of the existing methods have been analyzed and compared from a survey of recent patents on spectrum sensing methods to support the new findings The concept of adaptive thresholding improves the detection probability by 39 % and 27 % at noise uncertainty of 1.02 and 1.04, respectively at a signal to noise ratio of -10 dB. Furthermore, the error probability reduces to 58% at the optimal threshold using Particle Swarm Optimization (PSO) algorithm for the signal to noise ratio of -9 dB. Conclusions: The main outcome of this work is the reduction in the probability of sensing failure and improvement in the detection probability using adaptive double thresholds at low SNR. Further, particle swarm optimization helps in obtaining the minimum probability of error under noise uncertainty with an optimal threshold.