从ＡμＬＭＳ自适应滤波算法的几何涵义出发，提出了Ｓｇｎ－ＡμＬＭＳ１和Ｓｇｎ－ＡμＬＭＳ２两种小运算量的ＡμＬＭＳ简化算法。在此基础上，利用条件判别，演化出ＭＳｇｎ－ＡμＬＭＳ１和ＭＳｇｎ－ＡμＬＭＳ２两种改进算法。计算机模拟实验结果表明，简化算法的稳态性能与ＡμＬＭＳ算法相近；其改进算法的收敛和稳态性能优于ＡμＬＭＳ算法，具有运算量小，性能良好的特点 Based on the geometric meaning of AμLMS adaptive filtering algorithm, two small computational AμLMS algorithms, Sgn-AμLMS1 and Sgn-AμLMS2, are proposed. On this basis, the two improved algorithms of MSgn-AμLMS1 and MSgn-AμLMS2 are evolved using conditional discriminant analysis. The results of computer simulation show that the steady-state performance of the simplified algorithm is similar to that of AμLMS algorithm. The improved algorithm has better convergence and steady-state performance than the AμLMS algorithm, and has the advantages of low computational complexity and good performance