UAV Detection Using RF Spectrograms and Remote ID Fusion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current technologies face challenges in accurately detecting, classifying, and assessing threats from unmanned aerial vehicles (UAVs) due to their small size and low signal-to-noise ratios, making it difficult to differentiate between friendly and hostile drones, especially in noisy environments.
Innovation Solution
A system employing convolutional neural networks (CNNs) to process radio frequency (RF) signals and remote identification data, using spectrograms to enhance classification accuracy by denoising and filtering out noise, allowing for the detection and classification of UAVs even at low signal-to-noise ratios, and integrating with sensor fusion for threat assessment.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional detection methods are used to detect UAVs, then the system structure remains simple, but the detection precision and classification accuracy deteriorate due to small UAV size and low signal-to-noise ratios
Solution Approach 1:
The patent combines multiple detection methods including RF signal detection, acoustic detection, and visual detection into a unified sensor fusion system. This merging of multiple sensing modalities enables high-precision detection and classification of UAVs while maintaining manageable system complexity through integrated processing architecture
Solution Approach 2:
The patent introduces spectrogram analysis as an intermediary processing step between raw signal acquisition and UAV classification. By transforming time-domain signals into frequency-time representations, the intermediary spectrogram processing enhances feature extraction capability and improves detection precision in noisy environments
2Measurement precision
If signal processing is performed without denoising, then the processing speed remains fast, but the classification accuracy deteriorates in noisy environments
Solution Approach 1:
The patent applies denoising filters as a preliminary processing step before classification. By pre-processing signals to remove noise and enhance relevant features beforehand, the system achieves high classification accuracy while minimizing the time required during the actual classification phase
Solution Approach 2:
The patent transforms signals from the time domain to the frequency-time domain using spectrogram analysis. This dimensional transformation allows the system to process signals in a representation where noise and useful information are more easily separated, improving classification accuracy without proportionally increasing processing time
3Reliability
If remote identification data is integrated with field sensor data, then the threat assessment accuracy improves, but the device complexity increases
Solution Approach 1:
The patent merges remote identification data from regulatory databases with field sensor data in a unified threat assessment framework. This combination of disparate data sources improves threat assessment accuracy by cross-validating information and reducing false positives
Solution Approach 2:
The patent creates a multi-functional processing architecture that handles both remote ID verification and field sensor data analysis through a single integrated system. This universal approach improves reliability while managing complexity by avoiding separate dedicated systems for each function
Data Source
AI summary
The present disclosure describes various embodiments of systems and methods of detecting, classifying, and making a threat assessment of an unmanned aerial vehicle (UAV). One such method comprises detecting a radio frequency (RF) signal; determining that the RF signal is generated from an unmanned aerial vehicle (UAV) based on the detected RF signal; classifying at least a make and model of the UAV based on the detected RF signal; sensing for a remote identification field data broadcasted by the UAV; receiving remote identification information of the UAV from a network database, if the network database is available; assessing a threat likelihood of the UAV based on joint processing of at least the RF signal based classification of the UAV and the received remote identification information of the UAV; and signaling an alert containing a description of the UAV and the threat if the UAV is assessed as a harmful threat.