Study On The Method Of Learning And Decision-making For Cognitive Radio

Nowadays, with the fast development of wireless communication and the enhances of people's demands, the spectrum resources seems deficient day by day, the fixed spectrum assignment system results in the idle state of spectrum in many time. Cognitive radio (CR) is considered an effective technique to solve the conflict between lower spectrum efficiency and spectrum resources shortage. CR renovates the traditional spectrum management system, using a brand-new viewpoint of utilizing spectrum resource to achieve the goal of improving spectrum efficiency. Cognitive radio technology is a hot research topic in wireless communication.Cognitive radio is a highly intelligent communication system, in which artificial intelligence technology plays an important role. This thesis summarizes the applications of knowledge representation, expert system, neural network, genetic algorithm, fuzzy logic, game theory and other artificial intelligence technologies in cognitive radio. Cognitive engine is the best platform for applying artificial intelligence. The working process of cognitive radio forms cognitive cycle. According to the key operations in cognitive cycle, the structure and function of cognitive engine model are improved in this thesis.Sensing electromagnetism environment and detecting spectrum hole are the fundaments of cognitive communication. Learning and reasoning the sensed information and making decision of cognitive device parameter optimization are the key in cognitive communication, which expresses the intelligent of cognitive radio. Based on neural network and genetic algorithm, this thesis provides a learning and decision-making method. Neural network is used to learn parameters and genetic algorithm is used to achieve multi-objectives optimizations of adjustable parameters, which develops the best decision-making for idle spectrum utilization.GSM system is widely used today, which belongs to the fixed spectrum communication system. At leisure of night or on the low population density place, the idle spectrum of GSM system may be found and used by cognitive radio. Cognitive device utilizes the idle spectrum through learning and making decision. According to the characteristics of GSM cellular system and the distance between cognitive devices, different communication types in GSM system are classified in this thesis. Using the cognitive learning and decision-making method studied in this thesis, we simulated the parameters learning and adaptive adjustment process in cognitive engine.