In this paper, two-hop transmissions in channel-hopping (CH) cognitive-radio wireless networks (CRWNs) with buffer-aided relays are studied. The two-hop relay (rendezvous) model is improved using symmetric CH sequences, which are more suitable for practical CRWNs than the old model using asymmetric ones. A proper buffer storage size in the relays is formulated to minimize the two-hop data-transfer latency. The lower and upper bounds of the steady-state maximum-time-to-rendezvous (MTTR) and throughput with relays are derived with formal proofs. Using the proper buffer size, both metrics can be better controlled and found no worse than twice the original MTTR and throughput without relays. To explore the impacts of utilization probability of Internet-of-Things (IoT) devices to their performances in a small-scale CRWN, the single-region rendezvous-success (RS) rate and throughput are formulated. For a large-scale CRWN, the model is extended to formulate the multi-region RS rate and throughput with and without relays for the first time. Numerical studies of these analytical models are performed and validated with computer simulations. In summary, the transmission range and number of rendezvousing IoT devices are improved in the large-scale CRWN with relays, and the multi-region RS rate and throughput with relays are found better than those without relays.
Read full abstract