Performance study of Spectrum Estimation and sensing Methods in Cognitive Radios

Abstract

Spectrum estimation and spectrum sensing are primarychore of cognitive radio (CR), where in dynamically it explores the radio spectrum and reliably determine portion of the frequency bands that has been used and unused by the primary users. In this paper we propose spectrum estimation using Discrete Fourier transform (DFT) which has Non parametric estimators like Periodogram, Welch and Multitaper methods. Next Discrete cosine transform (DCT) whose performance is compared with respect to fast Fourier transform (FFT) using variance as the parameter. Further spectral sensing using wavelet packet transform for various decomposition level and different types of wavelets is being compared based on its performance parameter Receiver operating characteristics (ROC).In addition this paper also discuss about spectrum sensing using orthogonal frequency division multiplexing (OFDM).Simulation results show DCT provides better spectral estimation than DFT and has less variance. Further for wavelets Daubechies wavelet has better performance in ROC with increase in depth andorder of 10.