On the performance of secondary users in a cognitive radio network

Cognitive radio networks employ opportunistic scheduling of secondary (unlicensed) users for the efficient use of the scarce radio spectrum resources. The main idea is that secondary users (SUs) transmit data opportunistically by utilizing idle licensed frequency bands. The transmission of a SU may get interrupted several times due to the arrival of primary (licensed) users; the SU then needs to sense the spectrum to determine another available channel for retransmission. In this paper, we investigate the performance of SUs in a cognitive radio network. To this end, we develop a three-dimensional continuous-time Markov chain (CTMC) model of the system. We present an efficient method to compute steady-state probabilities by exploiting the specific Quasi-Birth-Death structure of the CTMC. Based on this, several SU performance measures are evaluated such as the mean delay of a SU, the SU interruption probability, the probability of a SU getting discarded from the system after an interruption and the SU blocking probability upon arrival. Numerical examples illustrate the influence of system parameters such as the sensing rate on the SU performance