Parameter Estimation And Analysis Of Underwater Acoustic Signal

With the national development and utilize of the marine resources, Underwater AcousticTechnology is now becoming one of the dominant technology, and then the Underwater AcousticSignal processing becomes a very important issue. In this paper takes the Underwater Acoustic Signalas the research object and takes a deep research into the estimation of the basic parameters of theUnderwater Acoustic Signal. Targeted raised the parameter estimation algorithm and makes analysisand comparison. The main work and contributions of this paper are as follows:1. Discussion of the two more mature classic SNR estimation algorithm: Singular valuedecomposition fourth-order and second-order moment method (M2M4), and makes analysis andcomparison of their estimation performance, provides a theoretical basis for engineering applications.according to the estimated shortcomings of the criteria using the minimum description length(MDL)for the singular value decomposition method to estimate the dimension of the signal subspace, thecombined information Criterion (CIC) to estimate the signal subspace dimension was put forward. Weget the improved algorithm CIC-SVD, the experiment proved that the improved algorithm estimatesthe performance is better than the original algorithm. The selection of two singular valuedecomposition method to estimate the signal-to-noise ratio of the measured underwater acousticsignals, and get better results, but due to the limitations of the existing conditions as well as mypersonal level, measured underwater acoustic signal SNR estimation results temporarily unable toverify.2. Firstly, the FFT-based interpolation algorithm detailed study using the respective merits of theRife algorithm and Quinn algorithm, the combined spectrum refine technology and frequency shiftingtechnique, the improved algorithm, namely Q-R comprehensive algorithm. Simulation experimentsprove stable Q-R algorithm full-band frequency shifting technique, the improved algorithm, namelyQ-R comprehensive algorithm. Simulation experiments prove stable Q-R algorithm full-bandfrequency estimation performance. Secondly, for the multi-frequency signal frequency estimationproblem of subspace-based frequency estimation algorithm, namely MUSIC algorithm and ESPRITalgorithm. Again, frequency estimation for non-stationary signals, leads to the concept ofinstantaneous frequency, frequency analysis method applied to analyze the instantaneous frequency ofthe chirp signal. Finally, the choice of Rife algorithm and its improved algorithm estimated frequency underwater acoustic signals measured shallow water vertical receiver. Shallow water noise complexityis far greater than the white Gaussian noise, the estimated performance was significantly lower thanthe theoretical simulation, but still have real-time frequency measurement capabilily verified Rifealgorithm based on FFT and its improved algorithm, the estimated performance to QR algorithm foroptimal.3. Based on the Hilbert transform analysis of transient characteristics does not have the inherentshortcomings of the adaptive analysis capabilities, proposed using a complex analytic wavelettransform to extract the instantaneous signal characteristics, improved algorithm based on theinstantaneous characteristics, improved algorithm based on the instantaneous characteristicparameters signal modulation recognition. Improve recognition performance under conditions of lowsignal-to-noise ratio, and only five characteristic parameters applied to simplify the identificationprocess.