آشكارسازي حفره هاي طيفي در سيستم هاي راديوي هوشمند با استفاده از فيلتر ذره اي در تخمين كانال

چكيده انگليسي :

Spectrum Hole Detection in Cognitive Radio Systems Using Particle Filter in Channel Estimation Seyed Mojtaba Alavi sm alavi@ec iut ac ir Date of Submission 2010 06 7 Department of Electrical and Computer Engineering Isfahan University of Technology Isfahan 84156 83111 Iran Degree M Sc Language FarsiSupervisor Mehdi Mahdavi m mahdavi@cc iut ac irAdvisor Ali Mohammad Doost Hosseini alimdh@cc iut ac irAbstract Underutilization of many parts of radio frequency spectrum has increased the interest in dynamicspectrum allocation Cognitive radios CR have been suggested as an enabling technology for dynamicallocation of spectrum resources Spectrum sensing used for finding free spectrum is a key task in cognitiveradio systems It is an important requirement for the realization of cognitive radio networks Spectrumsensing enables unlicensed users referred to as cognitive radio users to adapt to the environment bydetecting unused spectrum bands without causing interference to the licensed network referred to as theprimary network By detecting particular spectral holes where the licensed primary radio systems are idle and exploiting them rapidly the cognitive radio can improve the spectrum utilization significantly Therefore in order to recognize a spectral hole one must study the channel variation and use the estimate ofthe received signal power in the next instant Whereas most of the existing literature on spectrum sensingconsiders impairment by additive white Gaussian noise AWGN only in practice CRs also have to copewith various types of non Gaussian noise such as man made impulsive noise co channel interference andultra wideband interference Several measurement studies have shown that in many outdoor and indoorfrequency bands the noise distribution has heavier tails than Normal distribution This study presents a new method spectrum sensing based on particle filter theory in the presenceof non gaussian noise of environment It has been shown that a broad and increasingly important class ofnon Gaussian phenomena encountered in practice can be characterized as impulsive noise Herein alpha stable distribution is proposed for such a noise For the proposed noise model we apply particle filter toestimate the complex channel impulse response CIR which is rooted in Bayesian estimation and MonteCarlo simulation To our knowledge the implementation of the Particle filter is novel for such a system Furthermore we compared performance of Kalman filter and Particle filter in the presence of non gaussiannoise environment Our results reveals that filter predictor has better results than Kalman filter for a non Gaussian noise environment When CIR was estimated in order to primary user signal detection we userobust Lp norm detector which do not require any a priori knowledge about the primary user signal andperform well for a wide range of non Gaussian noises Furthermore we analyze the probabilities of falsealarm and missed detection of the proposed detectors in the low signal to noise ratio regime Foroptimization of Lp norm detection we propose a direct approach based on minimization of the probabilityof false alarm for a given probability of missed detection and a simpler approach based on maximization ofthe deflection coefficient of the detector Analytical and simulation results show that the proposed Lp normdetectors achieve significant performance gains over conventional energy detection in non Gaussian noise Keywords Cognitive Radio Spectrum Sensing Impulsive Noise Symmetric alpha stable ParticleFilter Kalman Filter Lp norm Detection

علوي، مجتبي

آشكارسازي حفره هاي طيفي در سيستم هاي راديوي هوشمند با استفاده از فيلتر ذره اي در تخمين كانال

حس كردن طيف , نويز ضربه اي , آلفا پايدار متقارن , فيلتر كالمن , آشكار ساز LP-norm