Research Of Recognition For Non-cooperative Signal System In Cognitive Radio

The rapid growth in wireless communications has contributed to a huge demand onthe deployment of new wireless services in both the licensed and unlicensed frequencyspectrum. However, recent studies show that the fixed spectrum assignment policyenforced today results in poor spectrum utilization. To address this problem, cognitiveradio (CR) has emerged as a promising technology to enable the access of theintermittent periods of unoccupied frequency bands, called white space or spectrumholes,and there by increase the spectral efficiency. Modulation recognition is one of keytechniques in cognitive radio. Modulation recognition systems have to be able tocorrectly classify the incoming signal’s modulation scheme in the presence of noise.Existing modulation classification algorithms include classic likelihood approaches andfeature-based approaches.The objective of this thesis is focused on the modulation recognition methods ofsingle and mixed signals. According to the signals’ different distribution in the domainof Wigner-Ville transform and wavelet transform, two approaches for modulationrecognition of single and mixed signals are proposed, namely, feature extraction basedmethods respectively. The contributions and main content are summarized as follows:(1)The background of cognitive radio is informed and the basic concepts and keytechnologies of cognitive radio is discussed,then the impact of modulation recognitionon cognitive is analyze.(2)Wavelet transform and Wigner-Ville distribution is introduced,which providetechinical support to the proposed algorithm.(3)Based on the analysis of the feature extraction of Wigner-Ville, an algorithm offeature analysis is proposed for modulation recognition used in non-cooperationcognitive radio. Simulations shows that the algorithm in the AWGN can recognize themodulation types.(4)Based on the feature extraction of wavelet transform, an algorithm which isused in rayleigh fading channel and AWGN channel is proposed, and numerouscomputer simulations show that this method is valid for modulation recognition of rayleigh fading and AWGN channel signals.(5)Based on the feature extraction of mixed signal, a novel modulation recognitionmethod used in Heterogeneous network is proposed. The simulation results show thatthis method is valid for modulation recognition of multiple co-channel signals.