抗串扰QR格型自适应滤波算法

A Crosstalk-Resistant QR Lattice Adaptive Filtering Algorithm

本文就抗串扰自适应噪声抵消器，研究了一种递归最小二乘（ＲＬＳ）格型自适应算法。算法的导出仅利用数值稳健的Ｇｉｖｅｎｓ旋转方法，因此具有优越的数值性能和较好的模块化结构性能。算法不含平方根运算，十分适合用通用数字信号处理器实现。文中还就算法的阵列实现进行了探讨，只需增加少量的时间，即可实现高性能的自适应噪声抵消器。模拟结果表明，算法的收敛速度明显优于现有的ＬＭＳ算法。

A receive least squares (RLS) lattice adaptive algorithm is developed for a cmsstalk-resistant adaptive noise canceller (CTRANC). The derivation of the algorithm is based solely on numerical robust Givens rotations, which have superior numerical property and better modular structure performance. Becaase square-root operations are not included, the proposed algorithm is very suitable for the implementation ed a conventiona digital signal procossor(DSP). An array implementation scheme is discussed and a high performance ANC can be realixzed by increasing the computing tiee slightly. Simulation results show that the convergence speed is much faster than that of the ekishng least mean square (LMS) algorithm.