A kent chaos artificial bee colony algorithm based wavelet thresholding method for signal denoising

The selection of wavelet threshold and the determination of thresholding function would directly affect the quality of the signal denoising using wavelet thresholding method. In the conventional thresholding denoising approaches, some aspects require improvement, such as the fixed threshold and the inflexible thresholding rules. To address these problems, a Kent chaos artificial bee colony (KCABC) based wavelet thresholding denoising approach is proposed in this paper. Firstly, a sine function based parametric wavelet thresholding function is put forward to devise the flexibility of the classical thresholding methods. Then, three strategies are employed to improve the performance of the basic ABC algorithm. The threshold and shape tuning parameter are initialized as the position of the individual, and the mean square error between the original and the thresholded signals is taken as the fitness function. Finally, the performances of the proposed algorithm and the existing methods are tested by denoising four benchmark signals with different noise cases. The simulation results indicate the proposed approach outperforms the existing methods in the capability of noise reduction.