Title :
FPGA based crosstalk-resistant adaptive decorrelator
Author_Institution :
Sch. of Eng., AUT Univ., Auckland, New Zealand
Abstract :
Field Programmable Gate-Arrays (FPGAs) are increasingly becoming widespread in the area of communications and signal processing. This paper investigates the implementation of an adaptive decorrelator based on two cross-coupled least-mean squares (LMS) algorithms. The device was designed using the rapid prototyping software developed by National Instruments using a CompactRio. It enables two sound sources which are mixed in a real room environment to be separated by decorrelating the cross-coupling transfer functions that link them together. The work has application in such areas as requires blind-source separation of random signals. Typically this includes hearing aids, smart microphones and speech recognition systems.
Keywords :
adaptive filters; blind source separation; crosstalk; decorrelation; field programmable gate arrays; hearing aids; least mean squares methods; microphones; rapid prototyping (industrial); speech processing; transfer functions; CompactRio; FPGA; blind source separation; cross-coupling transfer functions; crosstalk-resistant adaptive decorrelator; field programmable gate arrays; hearing aids; least mean squares algorithms; rapid prototyping software; separated by decorrelating; signal processing; smart microphones; speech recognition systems; Decorrelation; Field programmable gate arrays; Gain control; Least squares approximations; Noise; Pipeline processing; Random access memory; FPGA; Lab View; adaptivve filtering; decorrelation;
Conference_Titel :
Innovative Computing Technology (INTECH), 2013 Third International Conference on
Conference_Location :
London
Print_ISBN :
978-1-4799-0047-3
DOI :
10.1109/INTECH.2013.6653694