Training method, using method and device of unmanned aerial vehicle safety evaluation model

By constructing a dynamic security assessment model, and updating neuron weights using Euclidean distance and near-neighbor discrimination rules, the problem of lack of intelligent analysis capabilities of drone safety assessment methods in the existing technology is solved, and intelligent monitoring and rapid response to the real-time behavior of drones is achieved, and the safety and reliability of drones are improved.

CN119719773APending Publication Date: 2025-03-28ELECTRIC POWER RESEARCH INSTITUTE OF STATE GRID QINGHAI ELECTRIC POWER COMPANY +1
0 Cites 0 Cited by

Patent Information

Application Number
CN202411774926.6
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-12-05
Publication Date
2025-03-28

AI Technical Summary

Technical Problem

The existing drone safety assessment methods rely on fixed thresholds and rules, and lack the intelligent analysis capabilities of drone real-time behavior, making it difficult to accurately identify abnormal behaviors and potential security threats, increasing security risks.

Method used

A training method for the drone safety evaluation model is proposed. By obtaining training sample vectors based on the drone ontology or edge devices, updating neuron weights using Euclidean distance and near-neighbor discrimination rules, building a dynamic security evaluation model, evaluating whether the drone behavior is normal in real time, and dynamically adjusting operation permissions and task requirements.

Benefits of technology

It improves the safety and reliability of the drone during flight, can identify abnormal behaviors and potential threats in real time, adjust permissions and tasks in a timely manner, and reduces safety hazards.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

The invention discloses a training method, a use method and a device of an unmanned aerial vehicle safety evaluation model, and the training method comprises the steps: obtaining a training sample vector X based on an unmanned aerial vehicle body or an edge device, and the training sample vector X comprises the three-dimensional coordinate, the speed, the acceleration, the posture and the track of the unmanned aerial vehicle; determining an Euclidean distance d between an original neuron in a pre-obtained original model and each training sample vector X, and determining the most similar neuron of each training sample vector X; determining a neighborhood value between each most similar neuron and the original neuron according to the Euclidean distance d; updating the weight of the original neuron according to the Euclidean distance d and the neighborhood value based on a preset neighbor discrimination rule; and obtaining the unmanned aerial vehicle safety evaluation model based on the updated weight of the original neuron and the original model. According to the technical scheme, the safety and reliability of the unmanned aerial vehicle in the flight process can be improved.
Need to check novelty before this filing date? Find Prior Art