Abstract: A novel method was proposed to extract harmonic signals from measured data contaminated with chaotic interference. A new phase space projection method was established to separate harmonic signals from strong chaos by combining phase space reconstruction with wavelet transformation based on the geometry of chaos. Power spectrum density estimation was used to extract the frequency of harmonic signals, and extraction accuracy for harmonic signals was raised by means of wavelet filtering and least-squares estimation. The excellent results were achieved from computer simulation experiments, which verify the effectiveness of the approach. 

Key words: information processing technology, extraction of harmonics, wavelet transform, chaos, least square estimation 