改进变步长自适应分布式滤波算法

为提高传统LMS算法在分布式传感器网络自适应滤波中的收敛速度并保持较小的稳态误差,在扩散LMS算法基础上,提出一种基于Sigmoid函数改进变步长的分布式扩散LMS自适应滤波算法,算法通过Sigmoid函数变换构建步长因子与迭代误差之间非线性模型,并通过设置控制因子合理调整模型的步长曲线波型,使得算法在初期使用较大的步长以加快收敛速度,而在后期则采用平稳变化步长以保持较低的稳态误差,从而在在不失传统算法稳态精度的情况下,实现较快的收敛速度。实验结果表明:相比于已有算法,所提算法的估计性能更优。

Estimation Improvement Variable Step Size Adaptive Distributed Filtering Algorithm

To improve the convergence speed of the traditional LMS algorithm in the adaptive filtering of distributed sensor networks and maintain a small steady-state error,an adaptive variable step-length distributed LMS filtering algorithm based on Sigmoid function is proposed. In the proposed algorithm,a nonlinear model between the step factor and the iterative error is constructed through the Sigmoid function transformation and the control factor is set to adjust the step curve pattern of the model reasonably so that a larger step size is used in the initial stage to accelerate the convergence speed and a smooth variation step is used to maintain a low steady-state error,thereby achieving a faster convergence speed without losing the steady-state accuracy of the conventional algorithm. The simulation results show that the proposed algorithm performs much better than the existing algorithms.