Spectrum Utilization based Performance Evaluation for Cognitive Radio Networks

Abstract

Abstract The rapid growth of wireless communication devices and applications demand new spectrum band, but the existing radio spectrum is too much congested by the existing services. So a new technology Cognitive Radio (CR) is proposed by the researchers to fulll the scarcity of band allocation for new wireless applications and ecient utilization of existing spectrum. As the basic objective of cognitive radio is to deal with the scarcity of spectrum in the wireless communication area, so spectrum sensing is most important part of the cognitive radio. The decision about whether to use specic spectrum band or not is based on the spectrum sensing results. So to utilize spectrum efficiently without interfering to primary users (PUs), spectrum sensing should be done properly. There are some spectrum sensing techniques in the literature which dier to each other by performance, complexity and processing time etc. Spectrum sensing can be aected by channel impairments like fading and shadowing, so non-cooperative spectrum sensing among the secondary users (SUs), sometimes can create harmful interference to PUs. Cooperative spectrum sensing is considered as the solution of fading and shadowing like issues. All SUs share their sensed information to each other or to a central entity called as a base station or fusion center to make the decision. Once the decision about spectrum availability is made, ecient utilization of spectrum opportunity without interfering the PUs is the key to cognitive radio. Spectrum utilization is known as a successful data transmission by SUs without interfering to the PUs in a given time frame. So optimal sensing and transmission time slots in a time frame of SUs are very important for maximizing the spectrum utilization. In practical systems, however, frame structure needs to be standardized and xed and so is the length of the time-slot. So increasing sensing time duration will reduce transmission time duration or vice versa. Increasing sensing time duration improves the probability of successful detection of the presence or the absence of PU but reduces transmission time duration due to a xed length of the time-slot, which results in lower spectrum utilization. On the contrary, reducing sensing time duration allows a large transmission time duration. This may cause interference to PUs due to a lower probability of successful detection of the presence or the absence of PUs. Thus again it does not improve spectrum utilization. Motivated by the fact that there is an interesting trade-o between spectrum utilization and sensing time duration. The main objective of this dissertation is to determine the optimal sensing time duration and the optimal transmission time in a xed time frame of SUs to maximize the spectrum utilization in a given mean idle time of PU. Besides this, an optimal time frame of SUs is also determined in a given mean idle time of PU to achieve the maximum spectrum utilization by the SUs. It is also shown that when transmission time duration decreases relative to mean idle time of PU, the probability of reappearance of PU in the spectrum becomes low and the probability of no interference from SUs to PUs becomes high, which is helpful to achieve high spectrum utilization. Optimization of spectrum utilization given minimum sensing error rate in Rayleigh fading channel is also done by calculating the optimal detection threshold to minimize the sensing error. Furthermore, a cooperative spectrum sensing using soft data fusion scheme is investigated to analyzes the impact of sensing duration and reporting duration due to increasing number of SUs and reporting bits on spectrum utilization.