Spectral Interference Classification with Wi-Fi Pulse Removal

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing wireless networks face interference from non-Wi-Fi devices such as microwave ovens and wireless cameras, which are difficult to distinguish and classify using conventional methods.

Innovation Solution

A computing device processes spectrogram data to identify time and frequency domain edges, removes Wi-Fi pulses, extracts metrics, and uses a Machine Learning model to classify the type of interferer based on these metrics.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If conventional interference detection methods are used, then the detection process is simple, but the ability to distinguish and classify non-Wi-Fi interference sources is insufficient

Engineering Contradiction:

Improveinterference classification accuracyVSAvoidspectral analysis system complexity

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The patent segments the interference detection process into distinct stages: receiving spectrogram data, identifying time and frequency domain edges, removing Wi-Fi pulses, extracting metrics, and classifying interferer types. This segmentation allows complex spectral analysis to be broken down into manageable steps, improving classification accuracy while making the system more tractable

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent introduces an intermediary processing stage that extracts metrics from the spectrogram data before classification. This intermediary step transforms raw spectral data into meaningful features that bridge the gap between raw data and final classification, enabling accurate distinction between different interferer types without requiring direct complex analysis of all spectral components

Inventive Principle:

Principle #24Intermediary (Mediator)

2Measurement precision

If spectral analysis with pulse removal is performed, then interference classification accuracy improves, but processing time increases

Engineering Contradiction:

Improveinterferer type classification accuracyVSAvoidspectrogram processing time

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

The patent performs preliminary action by removing Wi-Fi pulses from the spectrogram data before extracting metrics and performing classification. This preliminary removal of known interference patterns (Wi-Fi pulses) simplifies the subsequent analysis by eliminating a major source of confusion, allowing the system to focus computational resources on identifying and classifying non-Wi-Fi interferers more efficiently

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The patent extracts and removes Wi-Fi pulses from the mixed spectrogram data, separating the known interference pattern from the unknown non-Wi-Fi interferers. This extraction process isolates the problematic signals, enabling more accurate classification of remaining interference sources without being confounded by Wi-Fi pulse patterns

Inventive Principle:

Principle #2Taking out (Extraction)

Data Source

PatentUS12567913B1Spectral analysis for non-Wi-Fi interference classification

Publication Date: 2026.03.03 CISCO TECHNOLOGY INC

AI summary

Classification of non-Wi-Fi interference may be provided. A computing device may receive spectrogram data. Next, time domain and frequency domain edges of a plurality of pulses in the spectrogram data may be identified. Wireless pulses may be removed from the spectrogram data. Metrics may then be extracted from the plurality of pulses in the spectrogram data with the wireless pulses removed. A Machine Learning (ML) model may be used to determine an interferer type based upon the metrics.