Voice Capacity of Cognitive Radio Networks

Providing multimedia services over cognitive radio (CR) networks has become an interesting research topic during past few years. As multimedia applications require specific quality of service (QoS) guarantees, supporting multimedia applications to secondary users over a CR network is a challenging task due to the random nature of resource availability. In this paper, we consider a secondary system operating over a timeslotted primary system with multiple channels and secondary users accessing the channels at the spectrum holes without interfering with primary users. We derive the voice capacity of the CR system based on the theories of effective bandwidth (EB) and effective capacity (EC). The capacity is represented in terms of the number of simultaneous independent voice calls that the system can support, providing stochastic delay guarantee. It is shown that (i) the analytical results match well with simulation results and stays slightly lower than the simulation results due to the conservative nature of the EB and EC theories, and (ii) the mean duration of channel being unavailable to secondary users has a significant impact on the system capacity. With proper medium access control, this analysis can help to develop a call admission control policy for QoS provisioning in CR networks.