Advanced search
Start date

Extreme learning machines applied to the equalization problem

Grant number: 13/15024-2
Support Opportunities:Scholarships in Brazil - Scientific Initiation
Effective date (Start): September 01, 2013
Effective date (End): June 30, 2014
Field of knowledge:Engineering - Electrical Engineering - Telecommunications
Principal Investigator:Romis Ribeiro de Faissol Attux
Grantee:Pedro Augusto Santos de Castro
Host Institution: Faculdade de Engenharia Elétrica e de Computação (FEEC). Universidade Estadual de Campinas (UNICAMP). Campinas , SP, Brazil


The problem of deconvolution (or equalization) is central to modern signal processing theory. Classically, the problem is solved with aid of a linear filter in the role of equalizer i.e. of a filter designed to invert the noxious channel effects. However, in many scenarios, the possibility of properly recovering the signal of interest depends on the possibility that the processing structure be capable of generating a non-linear response. An illustrative example of this point is the Bayesian equalizer for a linear channel with additive noise. Having these facts in view, extreme learning machines can be considered interesting options to deal with deconvolution, as they can generate non-linear projections of the input data while keeping a training cost equivalent to that of a linear combiner. In this work, we will study these neural networks and, subsequently, will systematically analyze their applicability to several instances of the focused problem.

News published in Agência FAPESP Newsletter about the scholarship:
Articles published in other media outlets (0 total):
More itemsLess items

Please report errors in scientific publications list using this form.