Optimized Topology Control for Large-Scale IoT Networks using Graph-based Localization

Internet of Things (IoT) will continue to pervade every walk of life over the next few years with the aim of improving quality of life and enhancing surrounding living conditions, while balancing available resources, like energy and computational power. As we deal with massive number of heterogeneous devices contributing to each IoT network, it is of paramount importance that the IoT network topology can be designed and controlled in such a way that coverage and throughput can be maximized using a minimum number of devices, while tackling challenges like poor link quality and interference. In this paper, we address the critical challenge of optimizing network topology in large-scale, resource-constrained IoT network environments. Unlike conventional methods that treat localization and topology design as separate problems, we propose a unified, graph-based framework that embeds both end-node and gateway positions into a cohesive spatial structure using partial and noisy distance measurements. We introduce IoTNTop, a novel iterative algorithm that performs topology control by jointly optimizing transmit power, error probability, and data transmission code rate. The algorithm segments the network into overlapping sub-graphs, aligns and stitches them using eigenvector synchronization and landmark matching, and refines the global topology using semi-definite programming and majorization techniques. Numerical and simulation results demonstrate that IoTNTop consistently outperforms baseline methods, by achieving lower transmission error, reduced energy consumption, with lower computational complexity.