To search, Click below search items. |
| All Published Papers Search Service |
|
|
|
Title |
Deep Learning based Automated Modulation Recognition for Cognitive Radio Networks |
|
Author |
Syed Muhammad Shehram Shah, Sajjad Ali Memon, Erum Saba, Abdul Latif, and Aqeel Ahmed |
| Citation |
Vol. 22 No. 5 pp. 393-400 |
|
Abstract |
Cognitive radio communication systems are regarded as the future of wireless communication systems since they can cater to demands in terms of data rate, latency, and quality-of-service. One of the challenges for cognitive radios is the automated recognition of modulation schemes, which is required for identifying unknown secondary transmissions. Automated modulation recognition (AMR) is the task of recognizing the type of modulation scheme used by the unknown secondary transmission sources from amongst a pool of possible modulation schemes. In this paper, we benchmark the performance of two types of deep neural network (DNN) architectures against a feature-engineered machine learning model based on Cumulant features and Logistic Regression. We show that DNN architectures outperform the handcrafted features models, thereby, highlighting the well-known learning ability of DNNs. We also show that the ability of DNNs to recognize modulation schemes is limited by the SNR of the received modulation symbols. |
|
Keywords |
Automated Modulation Recognition; Cognitive Radio, Wireless Communication; Deep Learning |
|
URL |
