Department of Electrical & Electronics Engineering
Permanent URI for this collection
Browse
Browsing Department of Electrical & Electronics Engineering by Author "AL-SUDANI, Haidder Jalil Sahib"
Now showing 1 - 1 of 1
Results Per Page
Sort Options
Item HIGH-ORDER CUMULANTS BASED CLASSIFICATION FOR COGNITIVE RADIO APPLICATIONS(2023-01-24) AL-SUDANI, Haidder Jalil Sahib; DALVEREN, YaserModern communication systems have witnessed dramatic changes due to the huge development of wireless technologies applications. These developments caused the scarceness of the spectrum and its inefficiency. However, cognitive radio (CR) is proposed as one of the best solutions to maintain high spectral efficiency (SE), and to treat spectrum scarcity. CR delivers the service to the unauthorized user to utilize the spectrum channel when the channel is out of the needs of the authorized user. However, spectrum sharing needs to be completed without signal interference. Thus, CR has many detection techniques for proper management of the frequency spectrum and interference avoidance. The major detection techniques of CR are Energy Detection (ED), Matched-filter Detection (MFD), and Feature-based Detection (FBD). In practice, each one of them has its own advantages and limitations. In this thesis, the use of statistical features in Machine Learning (ML) is proposed for a FBD. A MATLAB simulation is conducted to evaluate the effectiveness of the proposed detector. To this end, firstly, various modulation schemes with various noisy channels are generated. Then, the high-order moments and cumulants are extracted from the corrupted signals in noisy channels. In fact, these features are selected according to their strength in distinguishing between the signal and noise. The detection outcomes are employed in the support vector machine (SVM) classifier and the probability of detection (Pd) is obtained. The highest Pd value is achieved with 3 high-order cumulants in statistical detection. The same Pd value is obtained using SVM classifier by 1 high-order cumulant which reduces the amount of the processed data and simplify the detector complexity.