Extracting Seven Instantaneous Information-Based Signal Features

This paper implements algorithms for extracting seven key features derived from instantaneous signal information. The feature extraction workflow involves MATLAB-based computation of: 1) Maximum value of zero-centered normalized instantaneous amplitude power spectral density (calculated using FFT and peak detection), 2) Standard deviation of zero-centered normalized instantaneous amplitude absolute values, 3) Standard deviation of absolute values for instantaneous phase nonlinear components in zero-centered non-weak signal segments (requiring SNR thresholding), 4) Standard deviation of instantaneous phase nonlinear components in zero-centered non-weak signal segments, 5) Standard deviation of absolute instantaneous frequency values for zero-centered normalized non-weak signal segments, 6) Maximum value of normalized instantaneous frequency power spectral density for signal segments, and 7) Novel feature parameters distinguishing QPSK and 16QAM modulation schemes based on their differential projections on the XI-axis, implemented through constellation point analysis algorithms.