Low Snr Signal Detection And Parameter Estimation Methods

Ultra-wideband radar is a new type of radar system, and it has high range resolution because of its wide bandwidth. Thus, it has been used for target identification, precise position, target imaging and discrimination. How to detect the UVB radar signal and estimate the signal parameters in low signal to noise ratio condition is very important. The main concerned signals are Linear Frequency Modulated (LFM) signal.The main work can be summarized as:1. Ultra-wideband LFM is non-stationary signals and the conventional Fourier analysis can't describe their character completely. We make use of the typical time-frequency analysis which includes STFT, WVD and Hough_Wigner to analyze the LFM. After that, we proposed a mending Hough_Wigner to detect the strong signals and weak signal at the same time.2. Time series signals can be decomposed to IMFs(Intrinsic mode function) by EMD(Empirical mode decomposition). This paper makes use of EMD algorithm to decompose the mix signals and get the LFM signals. We proposed a new method to eliminate the end effect and got a better decomposition.3. Signal to noise ratio has strong affection for the signal detection. We can detect the signal in high signal to noise ratio easily but difficultly in low signal to noise ratio. This paper made use of wavelet transform to denoise signals for increase the signal to noise ratio. We proposed a new denosing method which denoise signals in frequency time in stead of time domain and got a better result.4. We had a comparison between the phase difference method and time-frequency presentation which can be used for estimating the instantaneous frequency. The dechirp method for estimating the parameter had less computation and lower precision ; the maximum likelihood method had more computation and higher precision ;as a result ,we can estimate the instantaneous frequency by making use of the two method's advantages.