Study On Parameters Estimation Of The Complex Dynamic Signals

With the development of communication technology, the direction of data transmission is reliable and fast. The traditional narrow band transmission mode is unsuitable in anti-interference, security and reliability. The modern radar communication requires the transmission of the signal with complex waveform and large bandwidth to improve security. In order to meet the need, the signal model used in this kind of wireless communication is complex and dynamic. Many more parameters of complex dynamic signals need to be estimated, but they are hard to acquire. Existing methods is difficult to obtain the parameters estimation directly. How to obtain the parameters estimation of complex dynamic signal in non-cooperative communication environment is urgently to be solved.Aiming to solve the problems, the thesis studies on the complex dynamic signals such as sinusoidal frequency modulation(SFM) signal and reconnaissance signal combined with quadratic frequency modulation and pseudo random binary phase code(QFM-PRBC) signal, reconnaissance signal combined with linear frequency modulation and pseudo-random binary phase code(LFM-PRBC) signal. The work of the thesis is shown below:(1) Models of several complex and dynamic signals are established, including SFM signal and QFM-PRBC signal and LFM-PRBC signal. The characteristics such as the time domain wave forms and the frequence domain wave forms and the time-frequence domain wave forms of the above signals are analyzed. The traditional methods are introduced particularly.(2) To solve the problem of single channel source separation and parameters estimation of SFM signals under the white Gaussian noise environment, the thesis proposes a method based on pulse repetition internal(PRI) transform. Firstly, the modulation frequency of each component is obtained according to PRI transform. Then the signal is decomposed on the modified discrete sinusoidal frequency modulation basis. After doing the Fourier transform, the carrier frequency and corresponding modulation coefficient of each component is obtained by searching of the maximum peak. The amplitude is acquired multiplied by the signal which is reconstructed by the estimated coefficients.(3) To solve the problem of single channel source separation and parameter estimation of SFM signals under the white Gaussian noise environment and the disturbing of LFM, the thesis proposes a method based on PRI transform and median filtering. Firstly, the spectrum of the SFM signals is obtained according to median filtering. Then the signal is decomposed on the modified discrete sinusoidal frequency modulation basis. After doing the Fourier transform, the carrier frequency and corresponding modulation coefficient of each component is obtained by searching of the maximum peak. The amplitude is acquired multiplied by the signal which is reconstructed by the estimated coefficients.(4) To estimate the pseudo noise(PN) code of QFM-PRBC, the thesis presents an algorithm based on the fractional ambiguity function and a modified method that using a reduced interference distribution kernel based on the triangular window. Firstly, square method is computed to eliminate the phase mutation, and the accumulated method is used to reduce the bad effect of the square method. Secondly, fractional ambiguity is adopted to estimate second and third coefficients. Thirdly, the original signal is multiplied by the conjugated signal which is reconstructed by the estimated coefficients to get a new compound signal. Lastly, the new compound signal consists of sine carrier and the PRBC signal, and the original PN code can be restored by the way of RIDT which is modified by the singular value decomposition(SVD).(5) To estimate the PN code of LFM-PRBC signals, the thesis proposes a method based on the linear canonical transform(LCT). Firstly, square method is computed to eliminate the phase mutation, and the accumulated method is used to reduce the bad effect of the square method. Secondly, the modulation rate and the carrier frequency are estimated by using LCT method. The pseudo random sequence and the signal amplitude are estimated by using the character of that the pseudo random sequence is a real signal. Thirdly, after cleaning the impact of the estimated signal, the parameters of all component signals are obtained following the above steps.In the thesis, the signal model and the related algorithm are analyzed and explained by the corresponding computer simulations. The result shows that the proposed method is effective under the condition of different signal to noise ratio(SNR).