一种自适应可变遗忘因子DCD-RLS算法

摘要

RLS最终的问题可以归结为正规方程的求解。解决正规方程可以采用非线性搜索的方法。为了降低复杂度可以采用二分坐标下降法。可变遗忘因子对算法有影响。可变遗忘因子的设置是通过时间平均误差相关来自适应调节。相对于其他可变遗忘因子的设置方法它的复杂度很低。通过仿真结果发现此算法与传统的算法的性能差不多。%RLS ultimate problem can be formulated as solving the normal equations. We can use nonlinear search method to solve the normal equations. In order to reduce the complexity we can use the dichotomous coordinate descent method. And the variable forgetting factor have effect on the algorithm. Variable forgetting factor are set by adaptation according to the time av-erage error correlation. Compared to other methods of variable forgetting factor setting this method has low complexity. The simulation results showed that the performance of this algorithm is similar to the tradi-tional algorithm.