Title: A variable tap-length DILMS algorithm with variable parameters for wireless sensor networks
Authors: Ghanbar Azarnia; Mostafa Hassanlou
Addresses: Engineering Faculty of Khoy, Urmia University of Technology, Urmia, 5815914853, Iran ' Engineering Faculty of Khoy, Urmia University of Technology, Urmia, 5815914853, Iran
Abstract: In most applications, the estimation of the filter length, along with the adoption of the filter coefficients, is an exact requirement. Accordingly, followed the stand-alone adaptive filters, the variable tap-length adaptive estimation has recently been considered in wireless sensor networks (WSNs). Among these, the fractional tap-length (FT) distributed incremental least mean square (DILMS) algorithm is of particular interest due to its superior performance. However, the main drawback of this method is that its performance, especially under high noise conditions, depends on the suitable selection of the length adaption parameters. On this basis, in this paper, a variable tap-length algorithm with variable parameters is proposed where the parameters of the FT algorithm is adapted according to the error correlation. Such an adaptation will lead to good performance, even under high noise conditions. The performed simulations also confirm such improved performance.
Keywords: WSNs; wireless sensor networks; variable tap-length; incremental topology; adaptive parameters; DILMS algorithm.
International Journal of Sensor Networks, 2021 Vol.36 No.2, pp.97 - 107
Received: 13 Nov 2020
Accepted: 28 Dec 2020
Published online: 05 Jul 2021 *