ANSIN: Adaptive Neural Symbolic Network for Secure Terrestrial and Non-Terrestrial Integration
- Kamran Ahmad Awan(corresponding author),
- ,
- Sonia Khan,
- Abdullah M. Alqahtani,
- Nikola Ivković,
- Abdul Rehman
- The University of Haripur,
- Biruni Universitesi,
- Youth Office Haripur,
- Jazan University,
- University of Zagreb
Abstract
The integration of terrestrial and nonterrestrial networks (TN-NTN) faces significant challenges in ensuring secure, adaptive, and cost-efficient communication. Current methodologies often fail to address dynamic security threats and optimize resource utilization in heterogeneous environments. This study aims to bridge these gaps by proposing a novel Contextual Symbolic Neural Reasoning (CSNR) framework designed for robust and scalable TN-NTN integration. The CSNR framework introduces a hybrid architecture that combines contextual information processing via a neural network component and logical inference through symbolic reasoning. A novel deep learning-based algorithm with attention mechanisms extracts dynamic features. Simulations were carried out over 450 minutes using the CICIDS2017 dataset and the results demonstrated superior performance, achieving 98.6% detection accuracy and 1.5% privacy leakage, with a computational cost of 0.025/min and an energy consumption of 135.2 kWh.
Research Output Information
Output type
Original language
EnglishPublication milestones
- Published - 2025
Publication status
Publisher
Institute of Electrical and Electronics Engineers Inc.Publication series
- Publication series name: IEEE Conference on Computer Communications Workshops, INFOCOM WKSHPS 2025
ISBN (Electronic)
9798331543709Publication IDs
- Scopus: 105017962993
