Performance access of various ann layers in cognitive radio network

Abstract

In this section, the taxonomy outcomes attained by the signal classification system defined in the preceding section are deliberated. To examine the performance, 400 signals were produced and modulated by means of three modulation systems: BPSK, QPSK and QAM. They are transmitted via AWGN channel with 0dB, 1dB, 5 dB and 10dB noise density. By means of temporal and spectral features with ANN classifier, a sum of 5, 10, 15, 20 and 25 hidden layers are taken for the modulation of the signals. 300 signals are employed for training, and 100 signals are tested with ANN classifier with dissimilar distance measures, Euclidean, city block, cosine and correlation.