Learning Data Generation for Unknown Wireless Device Recognition
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing technologies face challenges in improving the accuracy of recognizing unknown transmission apparatuses based on wireless signals, as supervised learning methods struggle to distinguish and accurately identify unregistered or unlearned devices.
Innovation Solution
A learning data generation system that includes an input unit, extraction unit, first clustering unit, estimation information acquisition unit, and generation unit to generate a relationship matrix between clustered sample features and estimation information, enhancing recognition accuracy by integrating unknown signals or radio features with known data and correct answer labels.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If supervised learning is used to recognize transmission apparatus, then recognition accuracy for known devices is improved, but recognition accuracy for unknown transmission apparatus deteriorates
Solution Approach 1:
The system performs preliminary clustering of sample features from unknown transmission apparatus before supervised learning classification. By pre-grouping unknown samples into clusters based on feature similarity, the system prepares structured data that can be later associated with estimation information, enabling unknown devices to be recognized through their cluster characteristics rather than requiring pre-existing knowledge of each specific unknown device.
Solution Approach 2:
The system introduces clustering results as an intermediary between raw unknown signals and final recognition. The clustering process creates intermediate groupings that capture patterns in unknown transmission apparatus, which then serve as basis for association with estimation information from external systems, bridging the gap between unknown inputs and meaningful identification.
2Measurement precision
If more learning data is collected from unknown transmission apparatus, then recognition accuracy improves, but system complexity and data processing burden increase
Solution Approach 1:
The system segments the complex task of unknown device recognition into distinct stages: feature extraction, clustering of sample features, association with estimation information, and final recognition. By dividing the processing pipeline into modular segments, each handling a specific aspect of the problem, the system manages complexity while improving accuracy through systematic data organization and multiple processing steps.
Data Source
AI summary
It is possible to improve recognition accuracy of an unknown transmission apparatus. A system inputs n pieces of first information (an unknown signal or a radio feature thereof), inputs the n pieces of first information to a supervised learning model that has been learned with respect to a known signal, and extracts n sample features. The system performs clustering processing on the n sample features. The system inputs m pieces of estimation target information to be estimated by an estimation apparatus that performs processing different from the clustering processing on the n unknown signals or radio features thereof, and acquires M pieces of estimation information. The system generates a relationship matrix indicating a relationship between a result classified into K groups by the clustering processing and the M pieces of estimation information.