用于声学回声抵消的基于小波变换和快速QR分解递归最小M-估计的新型鲁棒自适应算法

• 作 者: ZhouYi[1]周翊[2]ZhengChengshi[1]郑成诗[2]BaoMing[1]鲍明[2]LiXiaodong[1]李晓东[2]
• 机 构: Key Laboratory of Noise and Vibration Research, Institute of Acoustics,Chinese Academy of Sciences,
• 出 处: 中国电子学会第十五届青年学术年会
• 发表日期: 2009年期

　　This paper proposes a new robust wavelet-based fast adaptive filtering algorithm for acoustie echo cancellation.Using partial Haar wavelet transform to explore the sparsity of echa path and updating the system with a shorter length fast QRD Recursive Least Squares filter,the new algorithm achieves fast convergence rate and low computational complexity.Moreover,the new algorithm also employs robust M-estimation to improve its robustness to impulsive noise.Experiments are also conducted to verify the efficiency of the new algorithm.