Comparative sensitivity analysis of energy detection techniques for cognitive radio application
With sensitivity being an important factor in spectrum sensing based Cognitive Radio (CR) application; it remains unclear which out of the many existing Energy Detector (ED) techniques provides the best sensitivity performance for CR application. Consequently, this paper reports a study of some k...
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Austrian E-Journals of Universal Scientific Organization
2014
|
| Subjects: | |
| Online Access: | http://irep.iium.edu.my/46813/1/Vol4_Iss12_762-785_Comparative_Sensitivity_Analysis_of.pdf http://irep.iium.edu.my/46813/ http://aeuso.org/includes/files/articles/Vol4_Iss12_762-785_Comparative_Sensitivity_Analysis_of.pdf |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | With sensitivity being an important factor in spectrum sensing based Cognitive Radio (CR)
application; it remains unclear which out of the many existing Energy Detector (ED) techniques
provides the best sensitivity performance for CR application. Consequently, this paper reports a
study of some known parametric and non-parametric Energy Detector (ED) schemes for
Cognitive Radio (CR) application towards providing relevant information. The models studied
are the Simple Periodogram (SP), Welch Periodogram (WP), Multi-Taper (MT), Yule-Walker
(YW), Burg (BG), and Covariance (CV). Each technique was developed using known
mathematical models and appropriate signals were simulated for comparative analysis. However,
owing to the limitation of the typical Receiver Operating Characteristic (ROC) curve to infer
comparative information, our study proposes a decomposition of the ROCs of each technique into
respective detection and false alarm probability curves in comparison with estimated threshold
levels to enhance comparative inference. From our findings, it was observed that a detection
performance gain of about 50% can be achieved when using parametric techniques over nonparametric
methods especially in low SNR conditions. Furthermore, a possible 15dB increase in
sensitivity performance can be achieved in narrow than wideband sensing for all techniques.
Finally, an increase in sensing time might not necessarily improve detection performance in low
SNR conditions provided a low false alarm performance must be maintained.
|
|---|