Two-Stage Wireless Device Classification for Drone Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing networks face challenges in efficiently identifying and categorizing aerial wireless devices, such as drones, due to their unique radio propagation characteristics, which can cause interference with terrestrial devices, and current methods result in high computational and signaling overhead.
Innovation Solution
A two-stage detection procedure involving primary and secondary criteria using machine learning to classify wireless devices into drone and non-drone categories, where the first stage uses readily available measurements for a crude classification and the second stage employs more detailed measurements on a subset of devices for accurate classification.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If detailed measurements are collected from all wireless devices for accurate classification, then classification accuracy is improved, but signaling overhead and processing complexity increase significantly
Solution Approach 1:
The patent divides the classification process into two distinct stages: a first stage using readily available measurements for initial classification, and a second stage using detailed measurements only for devices that require more accurate classification. This segmentation reduces the overall signaling overhead while maintaining classification accuracy where needed.
Solution Approach 2:
The patent applies different measurement qualities to different devices based on their classification needs. Readily available measurements are used for most devices, while detailed measurements are collected only for specific devices that require more precise classification, optimizing the balance between accuracy and overhead.
2Measurement precision
If detailed measurements are collected from all wireless devices for accurate classification, then classification accuracy is improved, but processing complexity increases significantly
Solution Approach 1:
The classification process is segmented into two stages with different processing requirements. The first stage processes readily available measurements for all devices with low computational complexity, while the second stage processes detailed measurements only for a subset of devices requiring higher accuracy, thereby reducing overall processing complexity.
Solution Approach 2:
Instead of collecting detailed measurements from all devices (excessive action), the patent collects detailed measurements only from devices that require more accurate classification (partial action), reducing the processing burden while maintaining necessary accuracy.
3Productivity
If aerial wireless devices are not identified and categorized, then network operations continue without interruption, but interference to terrestrial devices increases
Solution Approach 1:
The patent performs preliminary classification of wireless devices into aerial and terrestrial categories using available measurements before interference mitigation actions are taken. This preliminary identification allows the network to apply appropriate interference management strategies to aerial devices while maintaining normal operations for terrestrial devices.
Solution Approach 2:
The patent converts the unique propagation characteristics of aerial devices that cause interference into beneficial classification features. By measuring timing advance, signal strength, and other parameters that differ between aerial and terrestrial devices, the system identifies aerial devices and applies targeted interference mitigation, turning a harmful situation into a manageable one.
Data Source
Figure 1
Figure 2a
Figure 2b
AI summary
Embodiments herein relate to methods and apparatus for determining whether each of a plurality of wireless devices are drones or non-drones. The method comprises determining based on binary classification of the first information which of the wireless devices meet all of at least one primary criterion; transmitting a request for second information to the wireless devices that meet all of the at least one primary criterion; receiving second information from each of the wireless device that meet all of the at least one primary criterion; determining based on binary classification of the second information which of the wireless devices that meet all of the at least one primary criterion also meet all of at least one secondary criterion; and classifying the wireless devices that meet both all of the at least one primary criterion and all of the at least one secondary criterion into the first category.