Title :
Combined projection and correlation least mean square algorithm for channel estimation with uncoded and chaotic coded speech
Author :
Müller, A. ; Elmirghani, J.M.H.
Author_Institution :
Dept. of Electr. & Electron. Eng., Wales Univ., Swansea, UK
Abstract :
Channel estimation is conventionally practically implemented using the least mean squares (LMS) algorithm due to its simplicity. This approach can significantly be impaired by channel noise and adaptation can be slow if the drive signal is correlated. We present a combined projection and correlation least mean squares (CPCLS) algorithm based on our former proposed finite correlation algorithm. These attributes speed up adaptation and result in robust estimation in the presence of noise respectively
Keywords :
FIR filters; adaptive filters; adaptive signal processing; chaos; correlation methods; filtering theory; least mean squares methods; speech coding; CPCLS algorithm; FIR adaptation algorithm; LMS algorithm; affine projection algorithm; channel estimation; channel noise; chaotic coded speech; combined projection and correlation LMS algorithm; correlated drive signal; correlation; filter length; finite correlation algorithm; least mean square algorithm; robust estimation; uncoded speech; Additive noise; Channel estimation; Chaos; Finite impulse response filter; Least mean square algorithms; Least squares approximation; Noise level; Noise robustness; Page description languages; Speech;
Conference_Titel :
Communications, 2001. ICC 2001. IEEE International Conference on
Conference_Location :
Helsinki
Print_ISBN :
0-7803-7097-1
DOI :
10.1109/ICC.2001.937369
