Abstract

AGUIRRE, Miguel; FRANCO, Zulay  and  PATETI, Antonio. Design and simulation of a neural network using the VHDL language and its aplication in the filtering of electrocardiogram signal. uct [online]. 2010, vol.14, n.57, pp.261-268. ISSN 1316-4821.

In this article is presented the design and simulation of a VHDL model of a Backpropagation neural network for the cancellation of interferences of 60 Hz in the electrocardiogram signal (ECG). This is presented in first place with a simulation algorithm of the Neural Network using Matlab, as the base for the development of a VHDL architecture. The process of functional simulation of the VHDL codification was realized in Modelsim. The sampling of the signal was carried out to 250 Hz. using 8 resolution bits for the samples. The discretization of the weights of the neural network and the calculations were carried out in fixed point. In the simulation of the MATLAB model, an error quadratic means of 6,8E-5 was obtained for a signal of electrocardiogram with 1 volt of amplitude. In VHDL modeling and simulation similar results were obtained.

Keywords : Neural Network; VHDL; Simulation; Interference; Filter.