摘要
Buffer-aided relaying can fully utilize the available selection gain of relay channels by allowing relays to store the received packets in the data buffers when the first-hop and second-hop channels are mismatched in quality. For energy-constrained Internet-of-Things (IoT) networks, some relays may lack sufficient energy to forward the buffered packets even their relay-to-destination channels are strong enough, leading to significant performance loss. This work studies buffer-aided relaying for relays that accumulate the energy harvested from source signal using finite-size energy buffers. A relay selection scheme considering both data buffer and energy buffer status is proposed and its performance is theoretically analyzed through Markov-chain modeling. Due to the complex behavior of dual buffers, the power consumption for reporting buffer status is first ignored in the theoretical analysis, which is then extended to, including feedback energy consumption. We validate the accuracy of the theoretical analysis and extensively discuss the performance of the proposed relay selection scheme subject to numerous key parameters. Numerical results demonstrate that the proposed relay selection scheme can fully exploit the diversity gain of multiple relays when ignoring energy consumption of feedback. With nonnegligible feedback energy consumption, the proposed relay selection scheme still significant outperforms some existing buffer-aided relay selection schemes.