Automatic Modulation Classification (AMC) is a key component of spectrum sharing techniques that incorporate signal identification schemes and require knowledge of spec- tral occupancy. Traditional likelihood-based and feature-based techniques have limited learning capability and suffer from performance in the presence of unpredictable interference. In such scenarios, Deep Neural Network (DNN) inspired techniques can enable better performance by continuous learning of channel usage statistics and identifying interference. The main goal of this study is to evaluate the performance of state-of-the-art convolutional and hybrid DNN models on a baseline dataset containing a wide range of signal impairments. It is envisaged that the outcome can be useful in targeting different computing platforms with varying power constraints.
Analysis of deep neural networks for automatic modulation classification
01.01.2023
915002 byte
Aufsatz (Konferenz)
Elektronische Ressource
Englisch
Depth Analysis in Deep Learning-Based Automatic Modulation Classification
DOAJ | 2024
|Interference Classification Using Deep Neural Networks
IEEE | 2020
|