摘要
针对自适应算法收敛速度和计算复杂度之间的矛盾,提出一种基于集员滤波的分割式比例仿射投影算法(SM-SPAPA)。该算法中只有当参数估计误差大于给定的误差门限时滤波器系数才进行迭代更新,从而能有效地减少滤波器系数的迭代次数。仿真结果表明,由于每次迭代将对误差性能贡献最大的输入信号筛选出来作为输入,从而能加快收敛速度,同时还能够减少算法的运算量。
For the contradictions of the adaptive algorithm between convergence speed and computational complexity, proposes a split ratio affine projection algorithm （SM-SPAPA） based on set member- ship filter. The algorithm only when the parameter estimation error is greater than a given error threshold filter coefficients before iterative updates, thus effectively reducing the number of iter- ations of the filter coefficients. The simulation result shows that each iteration will select the in- put signal which is the greatest contribution to the error performance as an input, thus speeding up the convergence rate, while also reducing the computational complexity of the algorithm.
出处
《现代计算机》
2012年第12期14-17,20,共5页
Modern Computer
关键词
SPAPA
回声消除
集员滤波
SPAPA
Echo Cancellation
Set Membership Filter
