用模糊自适应滤波算法估计时延

为了对系统中的时延进行估计，基于MRAS结构、最小均方算法和模糊理论，利用系统的输入和输出数据，提出了自适应模糊滤波AFLMS算法。该算法的运算量比通常具有IIR形式的系统辨识算法小得多，且收敛快，便于实时应用。在各种最小均方算法中，μ因子决定算法的收敛速度和稳定性，然而它是难以选择的。采用模糊技术来确定μ因子，克服了其困难。采用PRBS作为输入测试信号，利用计算机和采样电路，对带有时延的系统的仿真结果说明了该算法的有效性。

Time Delay Estimation with Adaptive Fuzzy Filtering Algorithm

In order to estimate ti me delay of the system by using input andoutput data of the system, the adap tive fuzzy filtering least mean square （AFLMS）algorithm is proposed based MRAS （model reference adaptive system）construction, LMS algorithm and fuzzy theory. The algorithm in IIR form needs much less multiplication than conventional sy stem identification methods. It also has the advantage of faster convergence. S o it is suitable for real-time use. It overcomes the difficulty of μ facto r selecting in LMS and ELMS algorithms because of using the fuzzy rules. With PR BS as test signals, time delay of the systems is estimated by a computer an d a sampled-circuit. The results of the simulations in the examples show the v alidity of this algorithm.