Neural Network RF Interference Detection With Rules-Based Characterization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional rules-based systems for RF interference detection struggle in environments with broad frequency ranges, multiple local minima, or background noise, leading to false positives and incorrect interference characterization, which can degrade the signal of interest.
Innovation Solution
A neural network is used to identify potential interference windows in a received RF signal, followed by a rules-based system to accurately characterize interference, reducing incorrect identifications and improving filter precision.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a rules-based system is used to identify and characterize interference in broad frequency ranges with multiple local minima and noise, then the system can operate with simple logic and structure, but the measurement precision and reliability of interference identification deteriorate due to false positives and incorrect characterization
Solution Approach 1:
The frequency spectrum is divided into multiple windows, and the neural network evaluates each window independently to identify potential interference regions. This segmentation allows the system to handle broad frequency ranges by processing them in manageable segments, improving detection accuracy without overwhelming the system structure.
Solution Approach 2:
A neural network is introduced as an intermediary between the received signal and the rules-based system. The neural network pre-processes the signal to identify and flag potential interference windows, which then are passed to the rules-based system for detailed characterization. This intermediary layer improves measurement precision while keeping the overall system structure manageable.
2Speed
If a rules-based system operates on the entire received signal, then the processing speed is maintained, but the reliability of interference detection deteriorates in environments with noise and multiple local minima
Solution Approach 1:
The neural network performs preliminary action by pre-identifying potential interference windows before the rules-based system processes them. This preliminary filtering step reduces the amount of data the rules-based system must analyze, maintaining processing speed while improving detection reliability by focusing only on suspicious regions.
Solution Approach 2:
Different processing approaches are applied to different parts of the signal: the neural network evaluates the entire signal to identify potential interference regions, while the rules-based system applies detailed characterization only to those specific local regions identified as problematic. This local quality approach improves reliability without sacrificing overall processing efficiency.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A system and method for utilizing a neural network or other artificial intelligence to identify windows of potential interference in a radio frequency signal is provided. The identified windows are then utilized in a rules-based interference detection system to identify a center, upper, and lower frequencies of the interference. The identified interference may then be remediated using conventional techniques.