Achievable throughput using cognitive principles on a frequency band with data traffic

Cognitive radio is often considered as means to achieve more efficient use of radio spectrum. The achievable gain, how-ever, is highly dependent on the characteristics of the traffic existing on the band cognitive radio uses for transmission as well as on the method that cognitive radio uses to obtain the information on the availability of the spectrum. In this paper, studies are conducted on the achievable throughput using cognitive radio principles on a spectrum band occupied by data traffic. Data traffic with two different set of parameters is used in order to study the impact of how frequently primary user's channel state changes to the performance. In the investigations, cognitive radio uses energy detection to obtain knowledge about the spectrum opportunities. It performs a simple loop with defined sensing, transmission and waiting times. The waiting time refers to a time period which the cognitive radio system waits before sensing the spectrum again in the case of an occupied channel. The influence of the waiting time has not been studied before since traditionally sensing has been assumed to take place in the beginning of each frame. Sensing time is influenced by target probabilities of false alarm and detection. The effect of these probabilities and the waiting time on the performance are investigated in a scenario considering one cognitive radio in a single spectrum band. The results are given as the percentage of spectrum opportunities found, the amount of interference caused to the primary user as well as the amount of time used for spectrum sensing. The achievable throughput is dependent on the probability of false alarm, whereas probability of detection affects the interference caused to the primary user. The results show that an alternative way to significantly reduce the interference to the primary user is to increase the waiting time. By doing this also the amount of sensing time required can be considerably reduced.