GraphSAGE for Real-Time Target Tracking in UAV Surveillance
Bibliographic record
Abstract
The increase in the need to have intelligent aerial surveillance has required the creation of real-time, correct, and scalable target tracking systems for Unmanned Aerial Vehicles (UAVs). The older methods of doing so generally include Kalman filters, LSTM-based trackers, and convolutional graph networks; however, they tend to face problems with scalability, generalizability to new environments, and fast real-time operation under dynamic missions. To address these gaps, we propose a GraphSAGE-based inductive learning framework customized for real-time UAV surveillance and multi-target tracking. The system can be understood as a dynamic graph with UAVs, targets, and communication areas corresponding to nodes and their relationships to each other as edges. Then, the node embeddings are updated continuously using GraphSAGE neighborhood aggregation. The framework facilitates building online graphs and making decisions so that the drone does not need retraining to adjust to new arrangements. On experimental analysis, our model displays 94.4% tracking accuracy with minimum average error (12.5 m) and a complete 100 percent interruption recovery rate, which surpasses baseline advanced baselines, such as GAT and DGCNN, by a significant margin. Besides high accuracy, the system is high-performance, with inference occurring in real-time and at low energy consumption, making it ideal for embedded applications on edge UAV systems. The given GraphSAGE solution is not only state-of-the-art in air/aural surveillance but also creates a scalable and robust architecture that can learn and adapt in real-time, multi-agent, and mission-critical processes. The study presents a new and practical perspective on AI systems utilizing UAVs, serving as a starting point for implementing intelligent graph-based control in next-generation aerial surveillance operations.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.002 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.001 |
| Bibliometrics | 0.001 | 0.000 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.001 | 0.002 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.003 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".