凸组合μ-律比例自适应非线性回声消除方法OA
Convex Combination μ-law Proportional Adaptive Nonlinear Echo Cancellation Method
在非线性自适应回声消除中,非线性滤波器系数经常存在着一定的冗余,只有部分系数能对信号建模发挥明显作用,即非线性滤波器存在稀疏行为.为了提高对非线性回声信号的消除效果,论文提出了基于凸组合μ-律比例函数链接自适应滤波器的非线性回声消除方法.将一个线性自适应滤波器和一个 μ-律比例非线性自适应滤波器输出进行凸组合,并用一个线段函数替代μ-律比例算法中的对数函数,以降低计算负担.最后的实验结果显示所提出方法在非线性回声消除中明显优于其他方法,具有更好的回声消除效果和更快的收敛速度.
In nonlinear adaptive echo cancellation,the nonlinear filter coefficients often have some redundancy,and only some coefficients can play an obvious role in signals modeling,that is,the nonlinear filter has sparse behavior.In order to improve the effect of eliminating nonlinear echo signal,a nonlinear echo cancellation method based on convex combination μ-law propor-tionate functional link adaptive filter is proposed.The output of a linear adaptive filter and a μ-law proportionate functional link adaptive filter is convexly combined,and the logarithm function in μ-law proportionate algorithm is replaced by a line segment function to reduce the computational burden.The experimental results show that the proposed method is better than other methods in nonlinear echo cancellation and has better echo cancellation effect and faster convergence rate.
赵益波;李业宁;刘明华
南京信息工程大学电子与信息工程学院 南京 210044||南京信息工程大学江苏省大气环境与装备技术协同创新中心 南京 210044南京信息工程大学电子与信息工程学院 南京 210044||南京信息工程大学江苏省大气环境与装备技术协同创新中心 南京 210044井冈山大学电子与信息工程学院 吉安 343009
信息技术与安全科学
自适应滤波器凸组合μ-律比例算法非线性回声消除
adaptive filterconvex combinationμ-law proportionate algorithmnonlinear echo cancellation
《计算机与数字工程》 2026 (1)
1-6,27,7
国家自然科学基金项目(编号:61871230)资助.
评论