A novel approach for reconstruction of high frequency signal

This paper proposes a new method for reconstructing a signal from low-frequency sampling data based on wavelet multiresolution analysis. Firstly, the low sample-rate data with unknown noise is filtered in the time-frequency domain. Then the proposed algorithm is used to reconstruct the high frequency signal. A hierarchical approximating algorithm is adopted to reduce the reconstruction error, and its accuracy analysis is given. This algorithm has the advantages of noise-free, high reconstructing accuracy and explicit physical nature.