Performance Analysis of Recurrence Matrix Statistics for the Detection of Deterministic Signals in Noise

摘要

Understanding the limitations to detecting deterministic signals in the presence of noise, especially additive, white Gaussian noise, is of importance for the design of LPI systems and anti-LPI signal defense. In this report, we investigate the use of recurrence plots for such detection, and we compare the performance of a recurrence-plot-based detector to a standard power- detection approach. Performance is evaluated using receiver-operator characteristic (ROC) curves.