A deep learning-based unmanned aerial vehicle intrusion prevention system

By combining improved RTMDet and Mamba models with multi-sensor data fusion, the drone defense system solves the problems of inaccurate sensor accuracy and behavior prediction in traditional systems, achieving high-precision, real-time drone intrusion defense with adaptive capabilities and rapid countermeasures.

CN122416375APending Publication Date: 2026-07-17GUANGDONG POLYTECHNIC OF IND & COMMERCE
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
GUANGDONG POLYTECHNIC OF IND & COMMERCE
Filing Date
2026-05-22
Publication Date
2026-07-17

AI Technical Summary

Technical Problem

Existing drone defense systems suffer from problems such as poor sensor accuracy and reliability, inaccurate target behavior prediction, insufficient data fusion technology, and delayed implementation of countermeasures, making it difficult to provide high-precision, real-time, and adaptive defense capabilities.

Method used

By employing an improved RTMDet target detection algorithm, the Mamba model, and an improved DETR model, combined with multi-sensor data fusion, an intelligent process for UAV intrusion detection, threat assessment, and countermeasures is realized, including modules such as data acquisition, target detection and localization, behavior prediction, intrusion determination, multi-modal data fusion, and threat assessment.

Benefits of technology

It improves the accuracy and real-time performance of drone intrusion detection, enhances the system's adaptability, enables dynamic adjustment of countermeasures, adapts to complex and ever-changing intrusion scenarios, and has the advantages of high precision, high response speed, and intelligent response.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122416375A_ABST
    Figure CN122416375A_ABST
Patent Text Reader

Abstract

This invention discloses a deep learning-based drone intrusion prevention system, comprising the following modules: a data acquisition module for acquiring and preprocessing real-time monitoring data; a target detection and localization module for identifying and localizing drone targets using an improved RTMDet target detection algorithm; a drone behavior prediction module for time-series modeling using the Mamba model; an intrusion behavior determination module for determining whether intrusion criteria are met using preset rules; a multimodal data fusion module for fusing data from different sensors; a threat assessment module for threat assessment using an improved DETR model; a countermeasure selection and execution module for selecting and executing matching countermeasures based on threat levels; and a feedback and monitoring module for monitoring the effectiveness of countermeasures and adjusting strategies in real time. This invention combines deep learning models and multi-sensor data fusion to effectively improve the accuracy and robustness of the drone intrusion prevention system.
Need to check novelty before this filing date? Find Prior Art