Loading...

Improvement of Resource Management Algorithms in Cognitive Radio Networks

Ramezani, Yosef | 2010

564 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 41187 (19)
  4. University: Sharif University of Technology
  5. Department: Computer Engineering
  6. Advisor(s): Hemmatyar, Ali Mohammad Afshin
  7. Abstract:
  8. Recent researches show inefficient use of frequency spectrum such that there is shortage of frequency in operation. In order to overcome this problem cognitive radio are introduced that opportunist usage of frequency band is their prominent characteristic. Problems and challenges caused by using these networks are wide and increasing. In this thesis we focus on improving resource management algorithms in cognitive radio. In this study in order to have a dynamic and efficient management in choosing reliable and quality channels, reinforcement learning algorithms are used based upon existing data and experiences. Since this tool has the learning capability and analysis in dynamic situation of the network and considering mutual effects of the users, it is suitable for this usage. Moreover, to improve the algorithm efficiency, learning classifier systems are utilized. Using this classifier, appropriate rules proportional to channels situation and management policy of the channels are produced to determine importance of the channel quality parameters. Finally to reduce the collision caused by presence of the primary and secondary users at the same time in the reserved channel, presence period of the primary user is forecasted. Radial basis function networks that are a type of artificial neural networks are utilized for forecasting. In the first section of this thesis, using learning classifier system, importance of each one of the effective parameters in the channel quality is estimated. Results show that in common situations, activity rate of the primary user and its absence time in the past have the most influence in decision making. In the next step, utilizing reinforcement learning algorithm, channels are valued and by learning from the past experiences and received rewards and penalties, reliable channels are selected. Results from simulation of a typical cognitive radio show that using this method channel switch rate by the secondary users and collision caused by simultaneously presence of the primary and secondary users are reduced. In the last step, using radial basis function networks, presence time of the primary users is forecasted and by that efficiency spectrum is increased noticeably
  9. Keywords:
  10. Cognitive Radio ; Reinforcement Learning ; Learning Classifier System ; Neural Network ; Radial Basis Function

 Digital Object List

 Bookmark

No TOC