Identification the Modulation Type in Cognitive Radio Network Based on Alexnet Architecture
Naseer A. Mousa, Sattar B. Sadkhan · 2022 5th International Conference on Engineering Technology and its Applications (IICETA) · 2022
Automatic Modulation Recognition (AMR) plays a major role in telecommunications and in Software Define Radio (SDR), due to its importance in the field of signal classification, as it is involved in many civil, cognitive Radio (CR) and military applications, including in the intelligence field. Many algorithms were used to classify the signals and then developed to enter deep learning (DL) applications on a large scale in this field to detect and classify radio frequency (RF) signals. The reason for adopting these techniques is that it determines the presence of signals without the need for complete information, as well as, it can classify non-contact waveforms, such as radar signals. Alexnet system which is part of Deep Neural Network (DNN) was adopted in our work to distinguish between signals generated in Radio ML2016a, and the goal is to get good accuracy if we compare it with other algorithms. We have implemented the work on an emulator (Google colaboratory), which simulates the Python program. We obtained an accuracy ranging from 79% to 91.3% for positive SNR values ranging from 2 dB to 18 dB. The above ratio is considered good with the proposed system.