Noise Reduction Method for Chaotic Signal Based on Phase Space Reconstruction

Based on the chaotic theory of reconstruction phase space, a novel noise reduction method is presented. In the proposed approach, the attractor set is reconstructed by embedding the time domain signal onto a d-dimensional space through the method of delay reconstruction. Based on the reconstruction phase space, matching pursuit approach is used to decompose the signal into linear expansion of waveforms to best match the signal structures. In order to show the validity of the approach in noise reduction, the chaotic signals generated by Lorenz model are applied for simulation analysis and the experimental results show that this method performs well in reduction noise and the algorithm converges.