傳統(tǒng)的回聲消除系統(tǒng)無法在嘈雜的雙工環(huán)境下工作。由于雙工信號的干擾，誤差信號
不能正確地指導(dǎo)濾波器系數(shù)的更新。本文中提出了一種基于相關(guān)函數(shù)的自適應(yīng)濾波算法，擴(kuò)展型相關(guān)函數(shù)LMS（ECLMS）算法。計算機(jī)仿真結(jié)果顯示ECLMS 算法的表現(xiàn)要優(yōu)于傳統(tǒng)LMS 算法。然而，ECLMS 的收斂速度仍然不能令人滿意。為了提高ECLMS 算法的收斂速度，本文推出了一種新的正比例步長調(diào)整算法，針對自適應(yīng)濾波器系數(shù)的每一個抽頭，提供一個合理的步長調(diào)整參數(shù)。計算機(jī)仿真結(jié)果證實(shí)了正比例擴(kuò)展相關(guān)LMS 算法理論的正確性。
關(guān)鍵詞：回聲消除，自適應(yīng)濾波器，雙工環(huán)境，相關(guān)函數(shù)，正比例步長調(diào)整算法
Proportionate Extended Correlation LMS Algorithm CHEN Rui1，E Zhi-feng1，YU Shao-jun1, Mohammad Reza Alsharif2 (1.Central South University of Forestry & Technology, Changsha, China,410004 2.University of the Ryukyus, Okinawa, JAPAN) Abstract: The conventional algorithms in the echo canceling system have drawback when they are faced with double-talk condition in noisy environment. Since the double-talk and noise signal are exist, then the error signal is contaminated to estimate the gradient correctly. In this paper, a new class of adaptive algorithm based on the correlation function processing is presented，which is called Extended Correlation LMS (ECLMS) algorithm. The computer simulation results show that
the ECLMS algorithm has better performance than conventional LMS algorithm. However the
convergence speed is not sufficient. In order to improve the convergence speed, we proposed a new proportionate step-size adjustment algorithm. The computer simulation results support the theoretical findings.
Key words: Echo canceling, LMS algorithm, Double-talk, Correlation function, Proportionate
step-size adjustment algorithm