arXiv:2512.08983v1 Announce Type: new 
Abstract: With the rapid development of Unmanned Aerial Vehicles (UAVs) and the increasing complexity of low-altitude security threats, traditional UAV identification methods struggle to extract reliable signal features and meet real-time requirements in complex environments. Recently, deep learning based Radio Frequency Fingerprint Identification (RFFI) approaches have greatly improved recognition accuracy. However, their large model sizes and high computational demands hinder deployment on resource-constrained edge devices. While model pruning offers a general solution for complexity reduction, existing weight, channel, and layer pruning techniques struggle to concurrently optimize compression rate, hardware acceleration, and recognition accuracy. To this end, in this paper, we introduce HSCP, a Hierarchical Spectral Clustering Pruning framework that combines layer pruning with channel pruning to achieve extreme compression, high performance, and efficient inference. In the first stage, HSCP employs spectral clustering guided by Centered Kernel Alignment (CKA) to identify and remove redundant layers. Subsequently, the same strategy is applied to the channel dimension to eliminate a finer redundancy. To ensure robustness, we further employ a noise-robust fine-tuning strategy. Experiments on the UAV-M100 benchmark demonstrate that HSCP outperforms existing channel and layer pruning methods. Specifically, HSCP achieves $86.39\%$ parameter reduction and $84.44\%$ FLOPs reduction on ResNet18 while improving accuracy by $1.49\%$ compared to the unpruned baseline, and maintains superior robustness even in low signal-to-noise ratio environments.

تم تقديم إطار عمل جديد يسمى HSCP لتحسين تحديد الطائرات بدون طيار (UAV) في البيئات ذات الموارد المحدودة. يجمع هذا النهج القائم على التجميع الطيفي في مرحلتين بين تقنيات تقليم الطبقات والقنوات لتقليل تعقيد النموذج بشكل كبير مع الحفاظ على دقة التعرف العالية، مما يعالج قيود الأساليب التقليدية في السيناريوهات المعقدة.

Se ha introducido un nuevo marco llamado HSCP para mejorar la identificación de vehículos aéreos no tripulados (UAV) en entornos con recursos limitados. Este enfoque de agrupamiento espectral en dos etapas combina técnicas de poda de capas y canales para reducir significativamente la complejidad del modelo mientras mantiene una alta precisión de reconocimiento, abordando las limitaciones de los métodos tradicionales en escenarios complejos.

Un nouveau cadre appelé HSCP a été introduit pour améliorer l'identification des véhicules aériens sans pilote (UAV) dans des environnements à ressources limitées. Cette approche de regroupement spectral en deux étapes combine des techniques de taille de couche et de canal pour réduire considérablement la complexité du modèle tout en maintenant une haute précision de reconnaissance, répondant ainsi aux limites des méthodes traditionnelles dans des scénarios complexes.

A new framework called HSCP has been introduced to enhance the identification of Unmanned Aerial Vehicles (UAVs) in resource-constrained environments. This two-stage spectral clustering approach combines layer and channel pruning techniques to significantly reduce model complexity while maintaining high recognition accuracy, addressing the limitations of traditional methods in complex scenarios.

HSCP: A Two-Stage Spectral Clustering Framework for Resource-Constrained UAV Identification

Was this article worth reading? Share it

Airparser

Attentive AI

Research AI

AIPortalX

Octofy

SnapChip

Ready to build your own newsroom?