Enhanced Radar Signal Classification Using AMP and Visibility Graph for Multi-Signal Environments

Electronic warfare today often takes place in crowded electromagnetic environments. Separating and identifying radar signals that overlap each other is hard, especially when the signal‑to‑noise ratio (SNR) is low. Standard spectrogram methods usually cannot tell apart radars that use similar scan patterns. This project tackles the problem with a two‑step classification approach that combines amplitude‑pattern (AMP) analysis and visibility‑graph methods. First, AMP analysis groups the received radar signals into broad classes. This step reduces noise and pulls out the main features of each signal. Next, a visibility‑graph algorithm turns the amplitude time series into a network graph. The graph makes it possible to separate overlapping signals in detail, for example

the Palmer series scan. The resulting graphs are then fed to deep‑learning models—GoogLeNet and ResNet—to assign the correct class. The overall method gives a strong way to classify signals when many are present at once. The work includes building a realistic signal model, creating the visibility graphs, and measuring how accurate the classification is at different SNR levels.

Message: Hi FE, I need help with “Enhanced Radar Signal Classification Using AMP and Visibility Graph for Multi-Signal Environments” in “Electronics & Communication Engineering”