.png){kind=icon}
.png){kind=icon}
Return
Aircraft Individual GRU Recognition Algorithm Based on Interpretable Track Features
.png){kind=icon}
OA
AI
DOI：10.1155/ijae/6632363.png)
.png){kind=icon}
.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Aiming at the technical bottlenecks existing in aircraft individual identification in battlefield situation awareness-including problems such as the prominent availability of single-dimensional data due to environmental constraints of multisource features and the insufficient mining of deep features of measured track data-this study proposes an aircraft individual Gate Recurrent Unit (GRU) recognition algorithm based on interpretable track features. Firstly, the scientific nature of the six-dimensional track parameters of longitude, latitude, altitude, speed, climb rate, and heading angle as individual identification features was demonstrated from the dimensions of flight dynamics and spatial geography. Then, a track feature screening and evaluation system based on the area under the curve (AUC) index was constructed. On this basis, a GRU deep-learning framework integrating the attention mechanism is designed to achieve the spatiotemporal correlation modeling of individual aircraft features. The experiment adopted the measured track dataset containing eight individual aircraft. The effectiveness of the feature selection scheme was verified through feature visualization analysis and AUC quantitative evaluation. The final test results showed that the individual recognition accuracy of the proposed GRU-attention model for the measured unknown tracks reached 68.31%, which was superior to other classification algorithms. The feasibility of this technology in achieving individual aircraft identification in a complex battlefield environment has been verified from the technical research level. This research provides a new technical path for the precise identification of aircraft, especially having significant application value in combat scenarios dominated by single-source features.
Keywords:
attention mechanism
feature selection model evaluation
gate recurrent unit
individual recognition
interpretable track features
AI Summary
Key information extracted from the uploaded paper, including a brief overview, abstract, background, key highlights, visual analysis, and future outlook.
Journal
IF:
1.2
Papers:
31
Citations:
1.9K