Sequential Radar–Acoustic Drone Detection Using A Lightweight Deep Learning Framework for Edge-Based Sensor Networks
Faizal Mohd Amin SharifuldinNur Emileen Abdul RashidMohd Adli Md AliMohd Mustafa Awang KechikKhairul Khaizi Mohd Shariff
Aug 2026· Sensing and Imaging· Vol 27· 0 citations· 45 references
TL;DR
Experimental results demonstrate that the proposed sequential pipeline preserves high specificity while reducing missed detections compared to radar-only processing and indicates that sequential processing can offer a viable alternative to parallel fusion for edge-based drone detection under SNR and computing constraints.
The increasing density of unmanned aerial systems (UAS) in urban low-altitude airspace introduces significant safety and security challenges, particularly for detecting non-cooperative drones in environments where RADAR (Radio Detection and Ranging) deployment is impractical. This paper presents a distributed, artificial intelligence (AI)-enabled multi-sensor surveillance framework integrating visual, acoustic, and radio frequency (RF) sensing through weighted decision-level fusion.Each sensing modality is processed using dedicated deep learning models, while a decision-level fusion mechanism combines predictions based on confidence scores and reliability weights. The modular architecture enables asynchronous communication through a publisher–subscriber paradigm, supporting distributed deployment and resilience under partial sensor degradation.The system is evaluated through both laboratory experiments and simulated urban environments with varying complexity. Results demonstrate that the proposed framework achieves over 90% detection precision, maintains false positive rates below 10%, and supports real-time processing exceeding 50 Hz. Furthermore, the fusion strategy effectively mitigates performance degradation in individual sensing modalities, particularly under noisy or obstructed conditions.These results highlight the potential of AI-based distributed sensor fusion systems as a scalable and cost-effective solution for real-time drone surveillance in smart urban airspace, contributing to resilient monitoring within emerging U-space ecosystems.
Neno Ruseno, Enrique Puertas, Aurilla Aurelie Arntzen Bechina· International Conference on...· 0 citations
Multimodal radar–audio sensing combines the spatial and reflectivity cues provided by millimeter-wave radar with the class-discriminative spectral patterns of acoustic signals for unmanned aerial vehicle classification. Under asymmetric degradation, however, a corrupted acoustic branch may become detrimental because continuously weighted acoustic features can remain in the fused representation. To investigate this failure mode, a controlled benchmark is constructed by applying Gaussian perturbations to stored Log-Mel representations while keeping the paired radar inputs unchanged. RAD-Gating is proposed as a reliability-aware fusion framework that decouples classifier-backbone training, reliability-gate calibration, and threshold-based inference. After the classifier backbone has been trained, its parameters are frozen and a lightweight reliability estimator is calibrated using polarized clean and degraded samples. During inference, the predicted acoustic reliability score is converted into a binary gate; unreliable acoustic features are zero-masked, and activation-scale compensation is applied to the retained radar representation. On the fixed 80/20 MMAUD split, the accuracy at −10 dB controlled feature-level degradation is increased from 29.23% with Softmax fusion to 56.15% with RAD-Gating, while the dedicated radar-only classifier achieves 61.54%. Fixed-checkpoint analyses are further conducted to examine reliability-score separation, threshold selection, feature-scale control, and soft versus hard inference. Within this controlled synthetic feature-level benchmark, the results indicate that reliability-aware hard suppression can mitigate multimodal negative transfer. These findings should not be interpreted as evidence of robustness to waveform-level or real-world acoustic disturbances.
Radar signal processing has become a fundamental technology for autonomous vehicles because of its ability to provide reliable object detection and tracking under diverse environmental conditions, including rain, fog, snow, and low-light scenarios. Conventional radar systems often face challenges such as clutter, noise, multipath interference, and limited target resolution, which can affect the accuracy of perception. This paper presents an advanced radar signal processing framework for autonomous vehicles that integrates adaptive preprocessing, target detection, clutter suppression, feature extraction, and object classification to improve perception performance. The proposed approach employs digital signal processing techniques combined with deep learning-based classification to accurately identify vehicles, pedestrians, cyclists, and other road obstacles from radar data. Multi-target tracking algorithms are incorporated to estimate object position, velocity, and trajectory in real time, enabling safe navigation and collision avoidance. Experimental evaluation demonstrates that the proposed framework achieves higher detection accuracy, improved target localization, and robust performance under challenging weather and traffic conditions while maintaining low computational complexity suitable for real-time deployment. The proposed radar signal processing system enhances the reliability, safety, and efficiency of autonomous driving by providing accurate environmental perception for intelligent decision-making.
Kuruba Theja Kuruba Theja, Dharavath Sunil Dharavath Sunil, Dr B Ramprasad Dr B Ramprasad· International Journal of Sci...· 0 citations
Radar-based target classification and intent recognition play crucial roles in autonomous driving, surveillance, and aerospace applications. Traditional methods relying on handcrafted features and conventional deep learning architectures often fail to capture long-range dependencies in radar sequences. This paper proposes a novel Transformer-based neural network architecture specifically designed for radar signal processing. Our approach integrates multi-head self-attention mechanisms with temporal convolutional networks to effectively model both local patterns and global dependencies in radar data. We design a hierarchical feature extraction module that processes range-Doppler maps and micro-Doppler signatures for robust target classification. Furthermore, we introduce an intent recognition module that leverages trajectory prediction and behavioral pattern analysis. Extensive experiments on simulated and real radar datasets demonstrate that our method achieves 94.7% classification accuracy and 89.3% intent recognition accuracy, outperforming state-of-the-art methods including CNNbased and LSTM-based approaches. The proposed architecture also shows superior computational efficiency with 35% fewer parameters compared to baseline models.
Jun Hu, Cheng Yu, Yahui Hu et al.· International Conference on...· 0 citations
Continuous traffic monitoring is critical for accurate structural load assessment and fatigue life estimation of highway bridges. However, conventional vision-based methods suffer from limitations such as line-of-sight restrictions, susceptibility to adverse weather and lighting conditions, and limited spatial coverage. Distributed acoustic sensing (DAS) offers a robust alternative by repurposing existing telecommunications dark fiber into dense, kilometer-scale sensor arrays. Nevertheless, interpreting the complex DAS signals generated by vehicular traffic remains challenging due to overlapping dynamic signatures and the scarcity of ground-truth data for model training. To overcome this, we present a cross-modal (vision-to-optic) supervision framework that transforms existing fiber infrastructure into a traffic monitoring system. We deployed a synchronized camera–DAS testbed along a roadway segment served by dark fiber. Video data is processed using modern computer vision foundation models SAM3 to automatically extract vehicle trajectories and classifications. These camera-derived labels supervise a deep sequence learning model trained on the corresponding DAS strain data. Once trained, the fiber-optic system independently achieves accurate vehicle detection, classification, localization, and speed estimation. Finally, we demonstrate how these continuous, DAS-derived traffic metrics can be directly translated into dynamic load profiles, providing a scalable, continuous monitoring solution for bridge fatigue and structural health assessment.
Cong Chen, Shenghan Zhang· e-Journal of Nondestructive...· 0 citations
This study presents an optimized, cost-effective radar sensor system for real-time, and contactless driver drowsiness detection. The system consist on a 24 GHz frequency-modulated continuous Wave radar module integrated behind the central rearview mirror for highly sensitive analysis of driver micromovements, including head tilts, and facial dynamics. The core innovation lies in the edge-deployed sensor-specific adaptation, where a custom CNN-LSTM pipeline processes complex radar returns to estimate eye opening and mouth opening metrics without visual data. This pipeline enhances the system's ability to isolate subtle motion characteristics—such as range and velocity—improving the sensing fidelity crucial for fatigue monitoring. Deployed on a resource-constrained Raspberry Pi platform, the model demonstrates robust integration and efficiency. Experimental validation on 20 subjects shows excellent performance, achieving up to 98.1% accuracy for eye blinking, and 97.9% for spontaneous yawning. This radar-based approach advances automotive sensing by providing a privacy-preserving, high-performance solution for safety-critical applications.
María-José López, César Palacios-Arias, B. Ordoñez et al.· IEEE Sensors Letters· 0 citations
Known for his clear and elegant writing style, Bertsekas shaped fields from control and optimization to large-scale computation and artificial intelligence.
