Joint Spatial and Propagation Models for Cellular Networks

In all current models for cellular networks, independent randomness in the positions of the base stations (BSs) and the propagation conditions is assumed. In practice, however, in coverage-oriented deployments, where the goal is to achieve good baseline coverage, cellular operators place the base stations further apart if propagation is favorable, and vice versa. We thus propose a new class of cellular model, where BSs are deployed such that all cell edge users achieve a minimum target signal power level from the serving BS. The spatial structure of the BSs is a result of the propagation environment and the target signal power. We call such network models joint spatial and propagation (JSP) models. To formulate such models, we assume the path loss follows a power law with a variable path loss exponent, so that the target signal power is achieved at the cell edges, given the distribution of the BSs. The coverage probability, defined as the probability that the signal-to- interference-plus-noise-ratio (SINR) exceeds a threshold, is evaluated and compared with the standard Poisson and lattice models. Our results show that networks with Poisson distributed BSs appear to the user like lattice networks if the dependence between BS placement and propagation is accounted for.