FBEGS-PAP 알고리즘과 Volterra 필터링을 이용한 비선형 반향신호 제거기법

Abstract

본 논문에서는 Volterra filtering과 fast block exact Gauss-Seidel pseudo affine projection(FBEGS-PAP) algorithm을 적용한 비선형 반향신호제거 기법을 제안한다. 특히 계수 update를 block 별로 시행하는 FBEGS-PAP 알고리즘은 기존의 GS-PAP 알고리즘에 비해 연산량은 줄이면서 유사한 수렴 성능을 보이며 대부분 선형 반향신호 제거에 응용되어 왔는데, 본 논문에서는 비선형 반향신호 제거를 위해 볼테라 필터링에 기반한 적응 알고리즘으로 FBEGS-PAP algorithm이 적용되었으며 시뮬레이션을 통해 제안한 기법의 반향신호 제거 성능을 확인하였다.; Echo signals may be undesirable in communication systems since it causes degradation of system quality and can make the communication system unstable. This problem can be solved by employing an echo cancellation approach based on linear adaptive filtering. The well-known normalized least mean square(NLMS), recursive least square(RLS) and affine projection(AP) algorithms have been widely used for that purpose. However, in real systems, the nonlinear distortion may occur and it is difficult to cancel these nonlinear echo signal by applying these linear adaptive filtering techniques. For this reason, an efficient nonlinear echo cancellation method is proposed, in this thesis, whereby the fast block exact Gauss-Seidel pseudo affine projection(FBEGS-PAP) is further utilized along with adaptive Volterra filtering. In particular, the proposed nonlinear echo cancellation approach leads to lower computational complexity as in the conventional adaptive linear echo cancellation methods(e.g., NLMS and GS-PAP), and still provides echo cancellation performance similar to that of the GS-PAP method. Finally, echo cancellation performance of the proposed approach is demonstrated by providing some simulation results.