Low Earth Orbit (LEO) satellite constellations are crucial for connecting Internet of Things (IoT) networks globally. They facilitate Direct-to-Satellite (DtS) communications, especially in remote areas lacking terrestrial infrastructure. Although Low-Power Wireless Area technologies, such as LoRa, provide long-range connectivity in DtS-IoT networks, scalability remains a significant challenge due to heavy channel congestion associated with large-scale deployments. The LoRa Optimistic Collision Information-based (L-OCI) estimator was introduced to address a satellite's insufficient awareness of possible congestion in its service area by estimating network size. However, this mechanism was designed for a single LEO satellite.


This thesis introduces Constellation L-OCI (CL-OCI), a novel mechanism for estimating network size in multi-satellite constellations. CL-OCI integrates estimations from multiple satellites, accounting for their ground track deviations relative to the node deployment region, and groups the node population for estimation by each satellite. We thoroughly tested CL-OCI using FloRaSat, a discrete-event satellite constellation simulator in realistic orbital DtS-IoT scenarios. The results show significant scalability improvement, doubling the number of nodes it can estimate and reducing node-energy costs by seven-fold compared to LoRa/LoRaWAN OCI.

Leveraging OCI benefits, we apply the network size estimator in two novel window selection mechanisms for Aloha-based MAC protocols, namely Frame Selection Game and Intelligent Probabilistic, tailored for device sets of unknown size. These mechanisms enable nodes to select advantageous time frames for channel contention through network size estimation. Extensive simulations in Matlab demonstrate that these mechanisms outperform traditional Frame Slotted Aloha protocols, improving scalability to accommodate up to 1,500 nodes.