On The Convergence of Blind DD-LMS Decision-Feedback Equalizers Using a Joint Equalization & Decoding Procedure

Cristiano Magalhães Panazio, Aline De Oliveira Neves, João Marcos Travassos Romano

DOI: 10.14209/sbrt.2001.18600282

Evento: XIX Simpósio Brasileiro de Telecomunicações (SBrT2001)

Keywords:

Abstract

"This paper shows that it is possible to achieve convergence to the optimum solution in a blind equalization framework with the use of least mean square algorithm in decision-directed mode (DD-LMS). In linear equalizers structures, the attainment of the optimum solution strongly depends on the filter weights initialization. We also show that decision-feedback equalizers in DD mode (DD-DFE) converges to undesired local minima, when all its weights are initialized with zeros, for a certain class of channels. However, it is possible to improve convergence and remove such local minimum by making use of joint equalization and convolutional codes. The major contribution of the present work is the convergence study of the mentioned configuration by means of a comparative analysis of the performance surface and the dynamical adaptation of both conventional DFE and joint DFE & decoding structures."

Download