Energy states aided relay selection and optimal power allocation for cognitive relaying networks

Energy harvesting (EH) is a promising technique for cognitive relaying transmission (CRT) where secondary users (SUs) and relaying nodes do not have a constant power supply each. Unlike conventional CRT where the end-to-end data rate is usually maximised without taking into account the energy consumption at the source and relay, in this study, the energy consumption is characterised by means of energy efficiency, defined as the achievable data rate per Joule. In particular, the energy states at each node (either at a SU or a relay) is modelled as a finite-state Markov chain and the transmit power at a node is optimally allocated by jointly accounting for the interference threshold prescribed by primary users (PUs), the maximum allowable transmit power and the harvested energy at the node. To maximise the energy efficiency, a best relay selection criterion is proposed and the subsequent optimal transmit power allocation is initially formulated as a non-linear fractional programming problem and, then, equivalently transformed into a parametric programming problem and, finally, solved analytically by using the classic Karush-Kuhn-Tucker conditions. With extensive Monte-Carlo simulation results, the effectiveness of the proposed relay selection algorithm and corresponding optimal power allocation strategy are corroborated.