Optimum transmit range and capacity of mobile infostation networks

A mobile infostation network stipulates all transmissions to occur when nodes are in proximity. In this paper, the effect of transmit range on the capacity of four transmission strategies is studied. We show that a stipulated transmit range improves the capacity compared to the Grossglauser-Tse strategy with an unconstrained transmit range by 25%, and outperforms the non rate-adaptive strategy by 68%. This indicates an optimal trade-off exists between spatial transmission concurrency and spectral efficiency on individual links. The optimal number of neighbors is invariant to node density, and is between 0.6 to 1.2 for our transmission strategies. This should be contrasted to a magic number of 6 to 8 neighbors for multihop networks, where the expected forward progress per hop is maximized. This reflects the different optimization criteria of mobile infostation and multihop ad hoc networks. In addition, the capacity per unit area increases linearly with node density. This is counter-intuitive but can be explained using a rescaling argument drawn from percolation theory.