Neural network design and implementation for time series signal processing

Abstract EN

Time series function is defined as a sequence of vector or scalar values which depend on time.

NN, have been widely used as time series forecasting like weather forecast and load forecast.

In this paper we used neural networks to simulate the behavior of the electronic circuits that process time signals.

In fact, electronic networks are very often used to model the computations performed by NNs.

Signal processing is used to analyze the signals in discrete and continuous time.

The analog signal processor involves linear and non linear electronic circuit such as passive filter , active filter , adaptive mixers integrator , differentiator , delay lines , amplifiers , voltage_ controllers , voltage _ control filters and so on.

Here the designed NN is assumed to act as the analog signal processor.

To start, RC circuit is considered such that the NN simulates its behaviour .

Elman neural network ,which is a type of RNNs , is used for, the training of the input/output waveforms of NN.

The obtained results show that the behaviour of the NN is the same as the behaviour of the electronic circuit under test .