Please use this identifier to cite or link to this item: http://hdl.handle.net/10266/4093
Title: Parameter Estimation and Tracking of Sinusoidal Signal using Variable-Step-Size LMS Algorithms
Authors: Garg, Rohit
Supervisor: Kohli, Amit Kumar
Keywords: Delay-estimation;Amplitude-estimation
Issue Date: 16-Aug-2016
Abstract: The current advancements in technologies like voice prediction, noise filtering, and system identification are possible only because of variety of tools present in discrete-time-signal- processing (DSP). But if the system is designed with some particular conditions, then it is not always possible that the system will respond in the same way under different conditions. So, we need to make it adaptive using some factor so that it can adapt itself accordingly. As in the case of filtering, if the coefficients of the filter are predefined then the system response remains constant in particular conditions and we cannot directly implement it in the real world scenario as the conditions change on regular basis. Taking into account these restrictions, it is recommended to use some adaptable system that can adapt itself according to the real world conditions. In these systems, the coefficients are made adaptable. In this research work, system identification by using signal received at two wireless sensors is carried out. The problem of adaptively estimating the time delay of sinusoidal signals received at two spatially separated sensors is a bit different from real scenario. Thus, relative amplitude and delay of the sinusoidal signal are needed to be estimated vis-a-vis a reference sinusoidal signal using finite-impulse-response (FIR) filter. There is always a tradeoff between convergence rate and misadjustment. To eliminate this dilemma, variable-step-size (VSS) methods are proposed. The VSS algorithm has a big step-size at the beginning for a maximum convergence speed and a much smaller step-size after the convergence for a minimum residual error. In this thesis, I present the performance evaluation of different VSS algorithms, which are used for parameter estimation of sinusoidal signal. The VSS in least-mean-square (LMS) algorithm is not only varied based on the error and input signal correlations, but also it can be varied by using sigmoid function as an alternative approach. The performance of VSS-LMS algorithm is appraised on the basis of convergence and tracking characteristics, in terms of mean-squared-error (MSE), when delay, as well as amplitude, need to be estimated. Simulation results are presented to demonstrate the efficiency and efficacy of both types of VSS criteria in combination with LMS algorithm.
Description: Master of Engineering-Wireless Communication
URI: http://hdl.handle.net/10266/4093
Appears in Collections:Masters Theses@ECED

Files in This Item:
File Description SizeFormat 
4093.pdf927.68 kBAdobe PDFThumbnail
View/Open


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.