Extracting Instantaneous Frequency Features

MATLAB 9K 263 views 0 downloads 1 credits
Tags:

Resource Overview

Extracting seven instantaneous information-based features from signals: maximum value of zero-centered normalized instantaneous amplitude power spectral density, standard deviation of zero-centered normalized instantaneous amplitude absolute value, standard deviation of absolute value of instantaneous phase nonlinear component in zero-centered non-weak signal segments, standard deviation of instantaneous phase nonlinear component in zero-centered non-weak signal segments, standard deviation of absolute value of zero-centered normalized instantaneous frequency in non-weak signal segments, maximum value of normalized instantaneous frequency power spectral density in a signal segment, and a novel feature parameter based on different projection characteristics of QPSK and 16QAM signals on the XI axis.

Detailed Documentation

During signal extraction, we select seven feature parameters based on instantaneous information: maximum value of zero-centered normalized instantaneous amplitude power spectral density, standard deviation of zero-centered normalized instantaneous amplitude absolute value, standard deviation of absolute value of instantaneous phase nonlinear component in zero-centered non-weak signal segments, standard deviation of instantaneous phase nonlinear component in zero-centered non-weak signal segments, standard deviation of absolute value of zero-centered normalized instantaneous frequency in non-weak signal segments, and maximum value of normalized instantaneous frequency power spectral density in a signal segment. Additionally, based on the distinct projection characteristics of QPSK and 16QAM signals on the XI axis, we propose new feature parameters to better characterize signal properties.

Implementation approach involves calculating these features through signal processing algorithms where key functions would include Hilbert transform for instantaneous parameter extraction, spectral density estimation using Welch's method, and statistical computations for standard deviations. The XI axis projection analysis requires constellation diagram processing and differential feature calculation between modulation schemes.

You May Also Like