Weak Signal Detection Method

This paper researches and presents a series of new methods for signal detectionin low SNR, which is based on the theories of modern signal process. And at thesame time, lots of simulation works are done.Chapter 1 introduces meaning and contents about this paper.Chapter 2 discusses the methods for signal detection based on high-orderstatistics. Besides a method using bispectrum is introduced simply, the method andperformance for signal detection using the third-order cumulant is studied mostly.Chapter 3 studies the adaptive filter in order to cancel noise. With purpose ofimproving the standard adaptive algorithm, a new adaptive algorithm is presented.Besides, the adaptive algorithm based on the third-order cumulant is researched andimproved. And the performance is prior to the algorithm above.Chapter 4 discusses the methods for signal detection using the theories ofWavelet Transform, Wigner-Ville Distribution, Radon-Wigner Transform andFractional Fourier Transform. This part mostly presents a detection method byaccumulating Wavelet coefficients.Chapter 5 studies the detection methods using Neural Networks. Firstly, amethod based on Back-propagation Neural Network is presented. Secondly, a newmethod based on Auto-Regressive Model and Learning Vector Quantization NeuralNetwork is presented.Chapter 6 discusses the difference of the detection methods and compares theperformance of the detection methods.Chapter 7 summarizes the whole contents in this paper.