Parameter Estimation Of Sinusoidal Signal Under Strong Noise

The problem of sinusoidal signal parameters estimation plays an important part in fieldsof biomedicine, measuring technology, vibration signal analysis and sonar. Sinusoidal signalis a common signal in the process of calculating physical quantities. In the process ofcalculating physical quantities, signals are disturbed by noise in varying degrees. The signalswhich are sent out by sensor include not only the information of the signals but also noise.Hence, for physical sense and measurement, the feature extraction of sinusoidal signalinfluences the accuracy of estimation. It is a crucial technology of measurement. This paperstudies the parameter estimation of sinusoidal signal under strong noise. The main researchworks are as follows:1. The sinusoidal signal amplification algorithm is proposed. The correlation ofsinusoidal signal can be enhanced by using the algorithm and the noise can be decreased. SoSNR can be improved.2. The sinusoidal signal auto-correlation algorithm based on wavelet decomposition isproposed. The estimation accuracy can be improved by using this algorithm. Therefore theamplitude estimation of sinusoidal signal can be got under low SNR.3. The sinusoidal signal bispectral analysis algorithm based on wavelet decomposition isproposed. The method of wavelet decomposition is applied to construct a new signal. In orderto deal with the problem of pseudo peak applied when using bispectral analysis, we amplifythe sampling signal and make correlation analysis between the amplified sampling signal andthe estimated signal. Finally, the accurate estimation of initial phase can be got under strongnoise.