论文部分内容阅读
经典最小均方(LMS)算法收敛性能与步长成正比,但仍然有很大局限性,难以实现快速收敛。针对这个问题,文章重点研究步长对算法收敛速度的影响,介绍了经典最小均方(LMS)算法、稳健变步长最小均方(RVSSLMS)算法以及最优自适应步长最小均方(OASSLMS)算法。在迭代次数相同的情况下,对三种算法的仿真进行分析比较,结果表明:三种算法都能在期望信号波达方向上形成峰值,在干扰方向上形成零陷。其中,最优自适应步长(OASSLMS)算法的权值,平方误差收敛速度最快,对期望信号的跟踪效果最好。步长优化后,权值收敛需要的迭代次数也明显减少。
The convergence performance of the classical Least Mean Square (LMS) algorithm is proportional to the step size, but it still has a lot of limitations and it is difficult to achieve fast convergence. In order to solve the problem, this paper focuses on the influence of the step size on the convergence rate of the algorithm. The algorithms of LMS, RVSSLMS and OASSLMS )algorithm. The simulation results of the three algorithms are compared and analyzed under the same number of iterations. The results show that all the three algorithms can form the peak in the direction of arrival of the desired signal and form the zero-depression in the direction of the interference. Among them, the weight of the optimal adaptive step (OASSLMS) algorithm, the square error of the fastest convergence, the best tracking of the desired signal. After the step size is optimized, the number of iterations required for weight convergence is also significantly reduced.