Wavelet-based Single-channel Frequency Overlapped Signal Parameters And Realization Of Research

It's a hotspot and difficulty to analyze the time-frequency overlapped both wireless communication and Radar signals in single channel. The technique is very useful and needful in both commercial communication and military applications.A novel algorithm for parameter estimation of digital modulated signals is proposed, which convert the point of view from time-frequency or modulation field to the singular-points field through multi-resolution analysis. The algorithm extracts the singular-points of mixed signals through Haar transform using the optimal scales firstly, and then separates them through LS polynomial curve fitting or SSWIA method, finally, estimates the parameters for each signal, respectively. This method is able to estimating the number of mixed signal components, calculating their symbol rates. A creative method of how to choose the scales of Haar transform is proposed in the paper, which could improve the SNR of signal being processed. The precision is improved more notably compared with conventional methods form the simulation.Modulation recognition is also an important research of the paper. Based on previous works, the paper improves some algorithm for the processing of single signal on the same condition, which improves the precision and robustness defending to noise. Then separate the singular-points of multi-signals on the domain, which extend the algorithm to the processing for multi-signals.The paper analysis the complexity of each algorithm, and implement them on Virtex5-SX95T FPGA. Then list all the precision, timing and resource information according to the results of synthesis and placement & route, which provide reference for the designing of optimization algorithm in practice.Based on the algorithm proposed in chapter 3, this paper implement some of them on FPGA, such as Haar transform for multi-signals based on PDDA-FIR architecture, accurate frequency estimation based on high-resolution energy center, the SNR estimation based on 16-orders matrix SVD for single signal, the estimation of number of signals and modulation symbol rate based on SSWIA and so on. All of them are imperative in electronic reconnaissance. All the algorithm could be implemented on one FPGA chip from the results of synthesis and placement & route, which proved the great foreground.