Research Article
Development of Radio Environment Map Enabled Case- and Knowledge-Based Learning Algorithms for IEEE 802.22 WRAN Cognitive Engines
@INPROCEEDINGS{10.1109/CROWNCOM.2007.4549770, author={Youping Zhao and Joseph Gaeddert and Lizdabel Morales and Kyung Bae and Jung-Sun Um and Jeffrey H. Reed}, title={Development of Radio Environment Map Enabled Case- and Knowledge-Based Learning Algorithms for IEEE 802.22 WRAN Cognitive Engines}, proceedings={2nd International ICST Conference on Cognitive Radio Oriented Wireless Networks and Communications}, publisher={IEEE}, proceedings_a={CROWNCOM}, year={2008}, month={6}, keywords={Algorithm design and analysis Chromium Cognitive radio Digital TV Engines Interference Mobile computing TV receivers Testing USA Councils}, doi={10.1109/CROWNCOM.2007.4549770} }- Youping Zhao
Joseph Gaeddert
Lizdabel Morales
Kyung Bae
Jung-Sun Um
Jeffrey H. Reed
Year: 2008
Development of Radio Environment Map Enabled Case- and Knowledge-Based Learning Algorithms for IEEE 802.22 WRAN Cognitive Engines
CROWNCOM
IEEE
DOI: 10.1109/CROWNCOM.2007.4549770
Abstract
The IEEE 802.22 wireless regional area network (WRAN) is the first worldwide commercial application of cognitive radio (CR) networks in unlicensed television broadcast bands. With the intent of efficiently occupying under-utilized spectrum, the network must be cognizant of spectrum available for secondary use and vacate channels as primary users are present. According to FCC’s recent public notice, WRAN products are anticipated to be available for the market by February, 2009. This paper first presents a generic architecture for WRAN cognitive engines (CE), and details the design of a CE leveraging the radio environment map database and case- and knowledge-based learning algorithms (REM-CKL). Furthermore, the performance of REM-CKL CE has been evaluated under various radio scenarios and compared to search-based optimizers, including a genetic algorithm (GA). The simulated results show that the WRAN CE can make significantly faster adaptations and achieve near-optimal utility by synergistically leveraging REMCKL and a local search (LS). Insights into REM-CKL, GA, and LS CE have been gained through the WRAN CE testbed development and preliminary testing.
