In this paper, a buffered decode and forward (DF) relay based three-node underlay cooperative cognitive relay network (CRN) is considered with a direct path to the destination. The source and the relay use multiple rates, and joint rate and link selection is performed to maximize throughput. Optimum link and rate selection rules are evolved that ensure buffer stability, and expressions are derived for the throughput assuming peak power and peak interference constraints on the transmit power of the secondary nodes. The expressions are written in a manner that yields useful insights on buffer stability and role of the direct link on performance. A scheme in which the direct link signal is combined with the relayed signal is also considered, and it is demonstrated that it offers additional improvement in performance only in some scenarios. Computer simulations have been presented to verify the accuracy of derived expressions. 2 powers, relays are often needed in the secondary network to increase range and reliability [3] [4].Although they incur a loss in performance as compared to full-duplex relays [5], half-duplex relays are preferred in many situations because of their simplicity. One option to overcome the loss due to half-duplex relays is to utilize rate selection, which requires channel knowledge at the transmitter [6]. In addition, when the direct link between the source and the destination is not shadowed, combining the direct and relayed signals improves performance and harnesses diversity gain in cooperative links. Yet another option to harness diversity gain is to use link selection, which requires the incorporation of a data buffer at the relay [7], [8]. Use of data buffers in relays provides some degree of freedom in scheduling links degraded by fading, and increases throughput. For this reason, buffer-aided relaying has been investigated in different scenarios extensively (relay-selection [9], [10], multi-hop [11], [12], two-way relaying [13], [14], MIMO systems [15], energy harvesting [16], physical layer security [17], [18], NOMA [19], [20], full-duplex relays [21], [22] and CRN [23], [24] etc.). Analysis of performance of buffered relays in underlay CRN has been carried out for half and full-duplex relays in [25], [26] and [27] respectively.
Motivation and ContributionsDue to the interference constraints, the link signal to noise ratios (SNRs) in underlay cognitive radio network (CRN) have large variance. For this reason, use of a buffer-aided relay with link adaptation is appealing in CRN [27] [28]- [30]. For the same reason, use of rate selection is well motivated in CRN, and we investigate this aspect here. As noted already, due to power constraints, the nodes in underlay networks are relatively close to each other for acceptable quality of service (QoS). Taking the direct channel into consideration is therefore important in underlay CRNs. In this work, we consider the direct channel, and perform joint rate and link selection with buffered relays in a two-hop underlay cognitive network. The majo...