Rethinking the achievable throughput formulation of cognitive radio ad hoc networks

In this paper, we investigate the achievable throughput of cognitive radio ad hoc networks. The current research on throughput evaluations focus on the average value over spatial distribution of random nodes. However, these average models neglect performance discrepancies resulting from different spatial nodes. In order to overcome this limitation, we derive the achievable rate coverage probability (equivalent to complementary cumulative distribution function (CCDF)) of the secondary network according to the random locations of secondary users. Based on the achievable rate coverage probability, the achievable throughput is derived for the secondary network under the primary outage constraint. Through stochastic geometric analysis, it is shown that the achievable throughput is mainly dominated by the densities of primary and secondary nodes, as well as the quality of service (QoS) of the primary network and the rate coverage probability of the secondary network. Besides, the achievable rate coverage probability is optimized to maximize the achievable throughput of the secondary network.