Recent Advances in Deep Hyperbolic Graph Attention Network-Based Collaborative Routing Algorithm for Clustered IoT-MANETs: A Systematic Review

Abstract

The rapid expansion of Internet of Things (IoT) technologies and mobile wireless communication has enabled the emergence of dynamic and decentralized networking systems such as Mobile Ad Hoc Networks (MANETs), where numerous heterogeneous devices communicate without relying on fixed infrastructure. However, IoT-MANET environments face significant challenges including frequent topology changes, limited energy resources, scalability issues, and unreliable routing performance. Conventional routing protocols like AODV, DSR, and OLSR are often inadequate in handling such highly dynamic and complex network conditions. To overcome these limitations, recent research has increasingly focused on integrating artificial intelligence and advanced graph-based learning techniques to enhance routing efficiency and adaptability. In particular, Graph Neural Networks (GNNs) and Graph Attention Networks (GATs) have shown strong potential in capturing network topology and supporting intelligent routing decisions. Additionally, hyperbolic graph learning has emerged as an effective approach for modeling hierarchical and scale-free network structures, enabling more efficient representation of complex IoT-MANET systems. Deep hyperbolic graph attention models further improve routing performance by embedding network data into hyperbolic space and optimizing communication in clustered environments. This review analyzes recent advancements in deep hyperbolic graph attention-based collaborative routing techniques for IoT-MANETs, comparing methods based on packet delivery ratio, energy consumption, throughput, and latency. It also highlights key challenges and outlines future research directions for developing scalable, energy-efficient, and intelligent routing frameworks for next-generation IoT networks.