The rapid adoption of Unmanned Aerial Vehicles (UAVs) has introduced critical security risks to sensitive infrastructures such as airports and power plants. This paper presents a real-time system for UAV detection and identification by leveraging radio frequency (RF) spectrogram analysis and deep learning. We propose a Software-Defined Radio (SDR)based framework using the USRP X310 to capture RF signals from drone remote control (RC) link, WiFi, and Bluetooth, generating a custom dataset of spectrogram images. The You Only Look Once (YOLO) architectures, YOLOv8n and YOLOv9t, are trained and benchmarked for drone detection.A comprehensive evaluation reveals the following trade-offs: YOLOv8n achieves superior inference speed ($\mathbf{1. 6 ~ m s}$ vs. $\mathbf{3. 0}$ ms), making it suitable for latency-sensitive applications, while YOLOv9t offers a lighter parameter footprint despite comparable computational complexity (7-8 GFLOPs). Both models exhibit similar accuracy, with mean Average Precision (mAP5095) stabilizing at $\mathbf{0. 8}-\mathbf{0. 8 5}$ and matching classification/box loss trajectories. However, YOLOv8n slightly outperforms in bounding box refinement (22% better DFL loss) and precision.Our findings underscore YOLOv8n’s dominance in real-time anti-drone systems, balancing speed and accuracy, whereas YOLOv9t’s compact design suits resource-constrained environments. This work bridges RF signal processing and AI-driven threat mitigation, offering actionable insights for deploying scalable drone surveillance solutions.

This work presents the design of a novel hybrid controller topology for the biasing and protection of GaN HEMT transistors in radio frequency (RF) power amplifiers. These high-electron-mobility transistors require dual power supplies with opposite polarities, making their management and protection challenging, particularly when relying on manual sequential biasing. The proposed controller, built on the STM32 platform, optimizes the regulation of transistor temperature, drain-source voltage Vds, and drain current Id. It ensures precise biasing control and proper sequencing during the power-on and power-off cycles of the RF power amplifier. At startup, an integrated dichotomy algorithm dynamically adjusts the gate-source voltage Vgs, enhancing operational stability and efficiency. By improving transistor control and reliability, this approach mitigates rapid degradation, thereby reducing manufacturing costs and accelerating the commercial adoption of GaN HEMT power amplifiers.

The use of commercial off-the-shelf (COTS) drones has been exponential growth, whether for leisure or professional purposes. In addition to the beneficial applications, they recently raised security threats for critical and strategic sites such as nuclear power plants, airports, and stadiums. Therefore, the need to detect these targets becomes essential for national security. The chosen approach is based on the detection of the RF signal related to the remote control (RC) link of the drone system using software-defined radio (SDR). The RF fingerprints are then extracted, we have considered the standard deviation (std), kurtosis, skewness, and Peak-to-Average Power Ratio (PAPR). These features and used for classification based on a customized Multi layer perceptron (MLP) model to identify the detected target.

This paper introduces a method for classifying and recognizing flying objects using trajectory features and artificial neural networks (ANN). Initially, the video sequence undergoes processing through a Gaussian mixture model (GMM) to detect and track the flying objects. Then, we will extract trajectory features from our dataset and use them to feed ANN for flying object classification. These features include turning angle, speed, acceleration and centroid distance function. The classical ANN is applied for feature vector classification to discriminate between birds and drones. Experimental results are conducted to showcase the effectiveness of the proposed method in classifying drones and birds. Moreover, the automated approach can be valuable in aiding military services to differentiate drones from other objects.

Mini unmanned aerial vehicles (UAVs), commonly known as small drones, have witnessed significant advancements in recent years and found applications in diverse fields such as search and rescue, wildlife monitoring, and border patrol. However, their misuse for illegal surveillance, smuggling, and potential involvement in terrorist activities poses serious concerns. Additionally, these mini-UAVs can disrupt the operations of commercial and military aircraft and infringe upon personal privacy. Hence, the development of an effective anti-UAV monitoring system is imperative to safeguard sensitive locations.In this paper, we propose a layered dashboard architecture specifically designed for an anti-drone system. Our software solution presents an intuitive interface that efficiently processes and visually represents video data captured by surveillance sensors. It incorporates AI-powered functionalities, including object detection, target tracking, and classification of small drones. To evaluate the effectiveness of the multi-layered solution, we deployed it within a no-fly zone, where it successfully identified and tracked drones in near real-time. The resulting dashboard offers a user-friendly interface and exhibits significant potential in enhancing the monitoring capabilities and prompt response to potential drone threats.