For Am, Psk, And Fm Signal Detection In Cognitive Radio Spectrum

With the growing demand for radio communications, wireless spectrum resource has become increasingly. So people propose Cognitive Radio (CR) technology, from the time and space to use those idle spectrum resources, in order to effectively address the problem of insufficient spectrum resources. Comparing with traditional static spectrum management, CR communication is based on dynamic spectrum access. Its key technologies are the perception of its surrounding environment, testing and finding unused blank band (spectrum holes) for communication, namely the spectrum sensing technology. However, if the primary signal of the band exists but the CR user does not detect and access to communications, leading to much interference to the primary signal. The difficulty of spectrum sensing technology in CR communication, is sensing and classifying the primary signals accurately, where signal to noise ratio is very low, even at the SNR = ?2 0 dB ~ ?4 0dB.This thesis focuses on the signal detection part of the spectrum sensing technology in CR communication. The current main methods are analyzed and classified, also a new improved time-frequency algorithm is proposed. The work presented in this thesis is as following aspects:(1) The first step in spectrum sensing of CR communications is proposed. The nature of the CR communication determines that CR users need to process received broadband signals, and an effective division of the spectrum is the first step in CR spectrum sensing. This chapter is based on wavelet transform detection methods, to avoid the inherent defects of traditional filters. Together with the idea of wavelet de-noising, it can also apply to big noisy environment. Simulations based on signal power spectral density compare the within and without wavelet de-noising methods, analyze impact in high and low SNR environments, and draw a positive conclusion.(2) The second step of the spectrum sensing is proposed. It is described the characteristics of cyclostationarity-based feature detection method, derived spectral correlation function (SCF), given out simulations of AM and PSK signals which have cyclostationary characteristic and analyzes them. Simulation results show that the method can be applied in very low signal to noise ratio environment. This is the main advantage of the feature detection method(3) Introduces the Local Polynomial Time-Frequency Transform (LPTFT) method, using the algorithm of the method based on time-frequency analysis, simulates three FM signals (LFM, WM and FSK signals, for example) and analyzes, obtains the corresponding time-frequency map, fills the vacancies that spectral correlation function does not apply to FM signals in last method.(4) Since LPTFT performs not well in noise in, a new improved algorithm -- related to Correlative-LPTFT (C-LPTFT) is proposed. It is carried out on C-LPTFT theoretical derivation, and also simulations on the LFM, WM and FSK signals are analyzed to show that the algorithm can be applied in lower SNR environment than LPTFT.Although there are many signal detection methods can be applied to CR communication, including the wavelet transform, cyclostationary approach, LPTFT and its improved C-LPTFT in this paper, in fact, each method has its own advantages and shortcomings. Following works should continue to enhance the anti-noise ability, find new detection methods, and combine them together effectively.