Modeling and Performance Analysis of IoT-over-LEO Satellite Systems under Realistic Operational Constraints: A Stochastic Geometry Approach

Abstract

Current theoretical studies on IoT-over-LEO satellite systems often rely on unrealistic assumptions, such as infinite terrestrial areas and omnidirectional satellite coverage, leaving significant gaps in theoretical analysis for more realistic operational constraints. These constraints involve finite terrestrial area, limited satellite coverage, Earth curvature effect, integral uplink and downlink analysis, and link-dependent interference. To address these gaps, this paper proposes a novel stochastic geometry based model to rigorously analyze the performance of IoT-over-LEO satellite systems. By adopting a binomial point process (BPP) instead of the conventional Poisson point process (PPP), our model accurately characterizes the geographical distribution of a fixed number of IoT devices in a finite terrestrial region. This modeling framework enables the derivation of distance distribution functions for both the links from the terrestrial IoT devices to the satellites (T-S) and from the satellites to the Earth station (S-ES), while also accounting for limited satellite coverage and Earth curvature effects. To realistically represent channel conditions, the Nakagami fading model is employed for the T-S links to characterize diverse small-scale fading environments, while the shadowed-Rician fading model is used for the S-ES links to capture the combined effects of shadowing and dominant line-of-sight paths. Furthermore, the analysis incorporates uplink and downlink interference, ensuring a comprehensive evaluation of system performance. The accuracy and effectiveness of our theoretical framework are validated through extensive Monte Carlo simulations. These results provide insights into key performance metrics, such as coverage probability and average ergodic rate, for both individual links and the overall system.