Blind Source-Separation in Mixed-Signal VLSI Using the InfoMax Algorithm

Authors:
Waldo Valenzuela;Gonzalo Carvajal;Miguel Figueroa
Affiliations:
Department of Electrical Engineering, Universidad de Concepción, Concepción, Chile;Department of Electrical Engineering, Universidad de Concepción, Concepción, Chile;Department of Electrical Engineering, Universidad de Concepción, Concepción, Chile
Venue:
ICANN '08 Proceedings of the 18th international conference on Artificial Neural Networks, Part II
Year:
2008

Citing 5
Cited 1

Independent component analysis, a new concept?

Signal Processing - Special issue on higher order statistics
An information-maximization approach to blind separation and blind deconvolution

Neural Computation
Independent component analysis using an extended infomax algorithm for mixed subgaussian and supergaussian sources

Neural Computation
Hardware Implementation of a PCA Learning Network by an Asynchronous PDM Digital Circuit

IJCNN '00 Proceedings of the IEEE-INNS-ENNS International Joint Conference on Neural Networks (IJCNN'00)-Volume 2 - Volume 2
Effects of Analog-VLSI hardware on the performance of the LMS algorithm

ICANN'06 Proceedings of the 16th international conference on Artificial Neural Networks - Volume Part I

A Case Study of ICA with Multi-scale PCA of Simulated Traffic Data

ICANN '09 Proceedings of the 19th International Conference on Artificial Neural Networks: Part II

Quantified Score

Hi-index	0.00

Visualization

Abstract

This paper describes a VLSI implementation of the InfoMax algorithm for Independent Component Analysis in mixed-signal CMOS. Our design uses on-chip calibration techniques and local adaptation to compensate for the effect of device mismatch in the arithmetic modules and analog memory cells. We use our design to perform two-input blind source-separation on mixtures of audio signals, and on mixtures of EEG signals. Our hardware version of the algorithm successfully separates the signals with a resolution within less than 10% of a software implementation of the algorithm. The die area of the circuit is 0.016mm2and its power consumption is 15μW in a 0.35μm CMOS process.