Enhanced Pre-Whitening TDLMS Adaptive Noise Canceller

No Thumbnail Available
Journal Title
Journal ISSN
Volume Title
International Conference on Technological Advances in Electrical Engineering (ICTAEE 2018)
In this paper, we propose a new algorithm for adaptive noise cancellation. The main idea behind the proposed algorithm is to introduce a first-order adaptive decorrelation filter to decorrelate the error signal in the existing pre-whitening transform domain least mean square (PW-TDLMS) algorithm. The obtained decorrelated error signal is then used in the adaptation equation of the PW-TDLMS algorithm. The proposed adaptive noise canceller is applied to speech denoising and the obtained results are compared to those of the PW-TDLMS and discrete cosine transform-based algorithms. The comparison is performed in terms of the computational complexity, mean square error (MSE) convergence speed, reached steady state level, steady state excess MSE, misadjustment and output signal to noise ratio
Laid CHERGUI, Saad BOUGUEZEL, Enhanced Pre-Whitening TDLMS Adaptive Noise Canceller, ICTAEE'18 SKIKDA ALGERIA, 2018.