# DroneShield-AI: A Multi-Modal Sensor Fusion Framework for Real-Time Autonomous Drone Threat Detection, Behavioral Intent Classification, and Swarm Intelligence in Contested Airspace

arXiv: https://arxiv.org/abs/2606.11687 · HTML: https://mlsanity.com/paper/2606.11687/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 9 of its 48 references that are in the map.
- First posted on arXiv 10 Jun 2026.

## Builds on (most cited first, 9 of 9)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [Graph Attention Networks](https://mlsanity.com/paper/1710.10903/) (2017)
- [Neural Relational Inference for Interacting Systems](https://mlsanity.com/paper/1802.04687/) (2018)
- [Detection and Tracking Meet Drones Challenge](https://mlsanity.com/paper/2001.06303/) (2020)
- [Human Motion Trajectory Prediction: A Survey](https://mlsanity.com/paper/1905.06113/) (2019)
- [MMAUD: A Comprehensive Multi-Modal Anti-UAV Dataset for Modern Miniature Drone Threats](https://mlsanity.com/paper/2402.03706/) (2024)
- [Investigation of UAV Detection in Images with Complex Backgrounds and Rainy Artifacts](https://mlsanity.com/paper/2305.16450/) (2023)
- [A Multi-task Learning Framework for Drone State Identification and Trajectory Prediction](https://mlsanity.com/paper/2309.06741/) (2023)
- [Biologically Inspired Swarm Dynamic Target Tracking and Obstacle Avoidance](https://mlsanity.com/paper/2410.11237/) (2024)

## Abstract (arXiv)

Unmanned Aerial Vehicle (UAV) threats have emerged as a defining security challenge of the 21st century. This paper presents DroneShield-AI, a unified open framework integrating six processing layers: RF signal classification, acoustic motor-signature detection, YOLOv8-based visual detection, evidence-weighted sensor fusion, a Behavioral Intent Classification Engine (BICE), and a Graph Neural Network Swarm Intelligence Module (GNN-SIM). This v2 revision reports measured results on the completed implementation (495 automated tests), superseding the v1 simulation-only preprint. On real public data: RF presence detection reaches F1 0.9924; acoustic detection reaches 98.12% accuracy, though a 4-feature statistical baseline reaches 93.25% on the same split, so the model's drone-specific contribution is approximately 4.9 points, a property of dataset provenance rather than a leakage bug; visual detection (YOLOv8) reaches AP50 0.8782 on a held-out test split, from a model still improving when training stopped. No fused-accuracy figure is reportable, since the three real-data layers use independently collected, unsynchronized datasets with no shared ground truth. BICE and GNN-SIM remain evaluated on physics-calibrated simulation data pending real adversarial and swarm field data: BICE shows a consistent robustness advantage over a trivial baseline under distribution shift; GNN-SIM underperforms its own trivial baseline in-distribution but independently replicates BICE's noise-robustness advantage, regarded as this work's most durable finding. All code, trained checkpoints, and the full test suite are publicly released.