Communication Channel Equalization-Pattern Recognition or Neural Networks?

Author

Singh, Satnam ; Blanding, Wayne ; Ravindra, Vishal ; Pattipati, Krishna

Author_Institution

Dept. of Electr. & Comput. Eng., Univ. of Connecticut, Storrs, CT

fYear

2006

fDate

27-30 Nov. 2006

Firstpage

1

Lastpage

5

Abstract

The communication channel equalization is a difficult problem, especially when the channel is nonlinear and complex. Numerous algorithms are presented in the neural networks literature to solve this problem. In this paper, a comparison is made among the latest neural network techniques (Complex Minimal Resource Allocation Networks (CMRAN) [D. Jianping et al., 2002]), a classical communication technique (Viterbi algorithm), and two pattern recognition techniques (Support Vector Machine (SVM), Learning Vector Quantization (LVQ)) to solve this problem. The simulation results show that Viterbi (MLSE decoding technique), and SVM methods outperform the CMRAN method.

Keywords

channel allocation; equalisers; neural nets; pattern recognition; telecommunication computing; Viterbi algorithm; communication channel equalization; complex minimal resource allocation networks; learning vector quantization; neural networks; pattern recognition; support vector machine; Communication channels; Decoding; Machine learning; Maximum likelihood estimation; Neural networks; Pattern recognition; Resource management; Support vector machines; Vector quantization; Viterbi algorithm;

fLanguage

English

Publisher

ieee

Conference_Titel

Communication Technology, 2006. ICCT '06. International Conference on

Conference_Location

Guilin

Print_ISBN

1-4244-0800-8

Electronic_ISBN

1-4244-0801-6

Type

conf

DOI

10.1109/ICCT.2006.341737

Filename

4146301