![]() This leveraging of existing neural networks is called transfer learning. You use the representations, called scalograms, and leverage an existing CNN by retraining the network to classify the signals. You do not need to separate the signal into I and Q channels. You use the CWT to create time-frequency representations of complex-valued signals. This example explores a framework to automatically extract time-frequency features from signals and perform signal classification using a deep learning network. ![]() While effective, this procedure can require extensive effort and domain knowledge to yield an accurate classification. Typically, to identify these waveforms and classify them by modulation type it is necessary to define meaningful features and input them into a classifier. Modulation classification has numerous applications, such as cognitive radar and software-defined radio. Modulation classification is an important function for an intelligent receiver. This example shows how to generate and deploy a CUDA® executable that performs modulation classification using features extracted by the continuous wavelet transform (CWT), and a pretrained convolutional neural network (CNN).
0 Comments
Leave a Reply. |