Electromagnetic Signal Classification Using Neural Network Preprocessing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for classifying electromagnetic signals are hindered by statistical and methodical uncertainties, as well as perturbations like background noise and distortions, which complicate accurate signal classification.
Innovation Solution
The use of an artificial neural network system to remove perturbations and complement missing information in electromagnetic signals, employing pattern recognition techniques to generate a modified measurement signal that enhances classification reliability.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional classification methods are used, then the classification process is simple, but the classification accuracy is reduced due to statistical uncertainties, methodical uncertainties, and perturbations
Solution Approach 1:
The patent applies preliminary action by performing signal preprocessing operations before classification. Specifically, missing signal portions are reconstructed and perturbations are removed in advance using signal processing techniques, so that the subsequent classification operates on cleaned and completed signal data, thereby improving accuracy without requiring the classification algorithm itself to be overly complex
Solution Approach 2:
The patent introduces an intermediary signal processing stage between signal acquisition and classification. This intermediary component processes the raw measurement signal by reconstructing missing parts and removing perturbations, acting as a mediator that prepares the signal for accurate classification while keeping the classification module itself relatively simple
2Reliability
If signal preprocessing is performed to remove perturbations and complement missing information, then classification reliability is improved, but processing time increases
Solution Approach 1:
The patent applies partial action by selectively processing only the problematic portions of the signal - specifically reconstructing only the missing signal segments and removing only the identified perturbations - rather than performing exhaustive processing on the entire signal, thus achieving improved reliability with minimal additional processing time
Data Source
Figure 1~2
Figure 3~4
AI summary
A method for classifying electromagnetic signals is disclosed. The method comprises the following steps: an electromagnetic signal is received; a measurement signal is generated from the received electromagnetic signal; the measurement signal is complemented with information missing in the measurement signal and/or perturbations are removed from the measurement signal to provide a modified measurement signal by using at least one machine learning module (16); and the received electromagnetic signal is classified based on the modified measurement signal. Moreover, an analysing system (10) for classifying electromagnetic signals is disclosed.