AI-enabled Real-time Spectrum Awareness ​and Interference Detection​

Avatar for Yiyang PEI
Yiyang PEI    
Associate Professor

Read More 

Avatar for Neelakantam V. VENKATARAYALU
Neelakantam V. VENKATARAYALU    
Associate Professor

Read More 

Develop real-time spectrum awareness and interference detection capabilities in the 2.4 GHz band, where the training signal operates.

 

Problem Statement:

Train a Deep Learning (DL) model using RF data to classify known signals and detect unknown signals as anomalies.
 

 

A line graph showing a synthetic RF spectrum of the 2.4 GHz ISM band. Color-coded regions identify signal power for CBTC (yellow), Bluetooth (blue), and WiFi (green) across the frequency range.

 

A confusion matrix for CNN-based signal classification performance. It compares true versus predicted classes for Noise, WiFi, Bluetooth, CBTC, and various combinations of overlapping signals, showing high diagonal accuracy percentages.

 

A confusion matrix for CNN-based signal classification performance. It compares true versus predicted classes for Noise, WiFi, Bluetooth, CBTC, and various combinations of overlapping signals, showing high diagonal accuracy percentages.