Date of Award
Master of Science (MS)
Wireless networks and information traffic have grown exponentially over the last decade. Consequently, an increase in demand for radio spectrum frequency bandwidth has resulted. Recent studies have shown that with the current fixed spectrum allocation (FSA), radio frequency band utilization ranges from 15% to 85%. Therefore, there are spectrum holes that are not utilized all the time by the licensed users, and, thus the radio spectrum is inefficiently exploited. To solve the problem of scarcity and inefficient utilization of the spectrum resources, dynamic spectrum access has been proposed as a solution to enable sharing and using available frequency channels. With dynamic spectrum allocation (DSA), unlicensed users can access and use licensed, available channels when primary users are not transmitting. Cognitive Radio technology is one of the next generation technologies that will allow efficient utilization of spectrum resources by enabling DSA. However, dynamic spectrum allocation by a cognitive radio system comes with the challenges of accurately detecting and selecting the best channel based on the channelâs availability and quality of service. Therefore, the spectrum sensing and analysis processes of a cognitive radio system are essential to make accurate decisions. Different spectrum sensing techniques and channel selection schemes have been proposed. However, these techniques only consider the spectrum occupancy rate for selecting the best channel, which can lead to erroneous decisions. Other communication parameters, such as the Signal-to-Noise Ratio (SNR) should also be taken into account. Therefore, the spectrum decision-making process of a cognitive radio system must use
techniques that consider spectrum occupancy and channel quality metrics to rank channels and select the best option. This thesis aims to develop a utility function based on spectrum occupancy and SNR measurements to model and rank the sensed channels.
An evolutionary algorithm-based SNR estimation technique was developed, which enables adaptively varying key parameters of the existing Eigenvalue-based blind SNR estimation technique. The performance of the improved technique is compared to the existing technique. Results show the evolutionary algorithm-based estimation performing better than the existing technique. The utility-based channel ranking technique was developed by first defining channel utility function that takes into account SNR and spectrum occupancy. Different mathematical functions were investigated to appropriately model the utility of SNR and spectrum occupancy rate. A ranking table is provided with the utility values of the sensed channels and compared with the usual occupancy rate based channel ranking. According to the results, utility-based channel ranking provides a better scope of making an informed decision by considering both channel occupancy rate and SNR. In addition, the efficiency of several noise cancellation techniques was investigated. These techniques can be employed to get rid of the impact of noise on the received or sensed signals during spectrum sensing process of a cognitive radio system. Performance evaluation of these techniques was done using simulations and the results show that the evolutionary algorithm-based noise cancellation techniques, particle swarm optimization and genetic algorithm perform better than the regular gradient descent based technique, which is the least-mean-square algorithm.
Quadri, Adnan, "A Channel Ranking And Selection Scheme Based On Channel Occupancy And SNR For Cognitive Radio Systems" (2018). Theses and Dissertations. 2319.