Open Access

REVIEW

Machine Learning for NTN-Assisted IoT: A Bibliometric-Assisted Survey of Optimization across Trajectory, Resource, Energy, and Security Aspects

Oluwatosin Ahmed Amodu¹, Zurina Mohd Hanapi^1,*, Chedia Jarray², Huda Althumali³, Faten A. Saif ⁴, Raja Azlina Raja Mahmood¹, Mohammed Sani Adam⁵, Nor Fadzilah Abdullah⁵

1 Department of Communication Technology and Network, Faculty of Computer Science and Information Technology, Universiti Putra Malaysia, UPM Serdang, Selangor, Malaysia
2 Systems Modeling, Analysis, and Control Research Laboratory (MACS), University of Gabes, Avenue Omar Ibn El Khattab, Zrig Eddakhlani, Tunisia
3 Computer Science Department, Faculty of Science and Humanities, Imam Abdulrahman bin Faisal University, Jubail, Saudi Arabia
4 Department of Information Technology, Gulf Colleges, Hafar Al Batin, Saudi Arabia
5 Department of Electrical, Electronics & Systems Engineering, Faculty of Engineering & Built Environment, Universiti Kebangsaan Malaysia, UKM Bangi, Selangor, Malaysia

* Corresponding Author: Zurina Mohd Hanapi. Email:

(This article belongs to the Special Issue: Artificial Intelligence for 6G Wireless Networks)

Computer Modeling in Engineering & Sciences 2026, 147(2), 6 https://doi.org/10.32604/cmes.2026.077054

Received 01 December 2025; Accepted 04 February 2026; Issue published 27 May 2026

Abstract

Non-terrestrial networks (NTNs)—including UAVs, HAPs, and satellite systems—are rapidly becoming key enablers of wide-area, resilient connectivity for large-scale IoT applications. As these platforms integrate with terrestrial networks to form space–air–ground architectures, optimization challenges related to trajectory, resource management, energy efficiency, and security become increasingly complex. Machine learning (ML) has emerged as a central tool for addressing these challenges by enabling adaptive, data-driven decision-making under uncertainty. This survey presents an optimization-centric review of ML-based NTN-assisted IoT systems focusing on aspect-specific datasets. Using a structured methodology involving dataset curation, keyword filtering, metadata analysis, and citation-based paper selection, we analyze representative and influential works across four core optimization themes: trajectory planning, resource allocation, energy utilization, and security. We develop a taxonomy that captures problem types, learning approaches, architectural configurations, and cross-layer constraints, and discuss insights, complemented by a focused review of top-cited contributions in each theme as well as discussions relating to their complexities and practicality. Our analysis reveals clear methodological trends, including the growing use of deep and multi-agent reinforcement learning, the emergence of distributed intelligence through federated learning, and the increasing interplay among mobility, computation, communication, resource allocation, energy optimization and security. Finally, we highlight key lessons and future research opportunities related to scalable cooperative learning, energy-efficient operation, secure distributed intelligence, and multi-tier optimization across space–air–ground integrated networks, offering a roadmap toward resilient and intelligent 6G-era connectivity.

---

Keywords

---