Performance comparison of device-to-device mode selection schemes

In this paper, we build a unified analytical framework that allows for analysis and comparison of three device-to-device (D2D) mode selection schemes proposed in the literature to date, namely the distance cut-off scheme, the link gain scheme and the guard zone scheme. In the framework we adopt Poisson point process (PPP) assumptions to model the cellular and D2D interference, respectively. Using stochastic geometry, we derive easy to implement expressions for the success probability at a typical base station (BS) and a typical D2D receiver (RX) in an underlay in-band D2D-enabled single tier cellular network. Comparing the derived analytical results with simulations, we show that the PPP assumptions are accurate for the success probability at the BS. Moreover, they provide a good approximation for the success probability at the D2D RX when the D2D RX is located close to the cell edge. Furthermore, the distance cut-off scheme generally outperforms other mode selection schemes.