自适应滤波器的性能曲面具有某些有益的几何特性 ,根据这些特性 ,本文提出FIR自适应几何中心法(FIRAGCM) ,然后通过转化IIR滤波器的性能曲面 ,得到全局收敛的IIRAGCM算法 .与梯度法不同的是 :FIRAGCM算法无需步长选择 ,需要的存储量极少 ,并且收敛迅速 ,IIRAGCM继承了FIRAGCM的全部优点 ,并且大幅度提高了收敛精度 .仿真结果表明 ,两种算法性能相当优越 According to these characteristics, the FIR adaptive geometric center method (FIRAGCM) is proposed in this paper, and then the IIRAGCM algorithm with global convergence is obtained by transforming the performance surface of the IIR filter.Compared with the gradient The difference is: FIRAGCM algorithm requires no step size, requires very little storage, and converges quickly. IIRAGCM inherits all the advantages of FIRAGCM and greatly improves the convergence precision. The simulation results show that the performance of the two algorithms is quite superior