Robust Relay Selection for Large-Scale Energy-Harvesting IoT Networks

We consider the relay selection problem in large-scale energy harvesting (EH) networks. It is known that if channel state information (CSI) is available at EH relays, a diversity order equal to the number of relays can be obtained, however, at the penalty of a feedback overhead (necessary to obtain accurate CSI) which is not suitable for energy-limited devices intended, e.g., for Internet-of-Things applications. In this paper, we therefore propose a new EH relay selection scheme which is based on the residual energy at each relay's battery, and on information on the distribution of the channels between relays and the destination. The method thus minimizes both the outage probability and the feedback cost. Where previous work relay selection based on channel distribution information consider only small-scale fading distribution, we employ a stochastic geometry approach to consider jointly the geometrical distribution (i.e., large-scale fading) and small-scale fading yielding a simple relay selection criterion that furthermore utilizes only rough information on the relay's location, i.e., an ordinal number from the destination. The outage probability of the proposed relay selection scheme is analytically derived, and the achievable diversity order of the proposed approach is investigated. Computer simulations confirm our theoretical analyses and show that our approach is robust against errors in the estimation of the distances between nodes.