Neural network and DSP based decoder for DTMF signals

ANN based demodulator for UMTS signal measurements

“…After numerous tests, the best configuration of the ANN based demodulator consists of the multiplayer perceptron architecture with 20 neurons in the input layer, 210 neurons in the hidden and 2 neurons in the output layer. The ANN is trained by using modulated signals generated according to the UMTS standard [3], [5] and [7] in the following operating conditions: (i) assigned scrambling code, and (ii) assigned channel set [0, 1,5,7]. Therefore, the trained ANN can be used in the following The trend of the percentage error, defined as difference between the assigned symbol and the detected one, is shown in Fig.2 versus the epochs of the training phase.…”

“…The advantageous use of the Artificial Neural Networks (ANNs) in measurement instrumentation and, in particular, to process the modulated signals for telecommunication has been shown in previous papers [1]- [2]. In [1] the ANN based decoder for Dual Tone Multi Frequency signals was presented, and the results of tests performed on the DSP implementation are shown.…”

“…In [1] the ANN based decoder for Dual Tone Multi Frequency signals was presented, and the results of tests performed on the DSP implementation are shown. The comparison with the traditional ones permits to highlight the advantageous performances of the ANN based demodulator beyond the limits given by the international recommendations.…”

Discrete time recurrent neural network observer

“…One of the advantages to develop a discrete algorithm is the direct implementation in an electronic device like a DSP, FPGA, etc., it can see in [14]. In the other side, the application given in this paper is about a model of HIV, and usually the illness monitoring implies the analysis of a sample in a period of time, this is naturally linked with discrete-time systems.…”

Validating a method for classifying digitally modulated signals

“…The most part of these, however, identifies only few modulation schemes as M-PSK or M-QAM [1][2][3][4], or requires some knowledge on the signal parameters [5][6][7][8].…”