This paper presents a theoretical analysis for estimating the coverage probability in two-dimensional (2D) and three-dimensional (3D) peer-to-peer (P2P) millimeter-wave (mmWave) wireless networks. The analysis is carried out adopting suitable link state models and realistic propagation conditions, involving path-loss attenuation, angular dispersion, mid- and small-scale fading, which comply with recent channel measurements. The presented framework accounts in detail for the actual shape of the transmitting/receiving antenna patterns and for the spatial statistic that describes the node location, by considering the widely adopted Poisson point process, the uniform distribution, and the random waypoint mobility model. Analytical expressions for the statistic of the received power and simple integral formulas for the coverage probability in the presence of interference and noise are derived. The accuracy of the obtained estimations and of the introduced approximations is checked by independent Monte Carlo validations. As possible applications in the 3D mmWave context, the conceived mathematical theory is used to discuss the impact of the interference model on the reliability of the noise-limited approximation, and to estimate the average link capacity of an interfered P2P communication.
