Magnitude-based pulse width estimation via efficient edge detection

In recent years, researchers have addressed the problem of using noncoherent approaches to estimate pulse width and repetition interval. Since measured transmitter is noncooperative, integration gain can be realized, input signal-to-noise ratio (SNR) for these estimators becomes critical. We examine multiple edge detectors that exploit moving sums calculated as part a Haar filtering received signal magnitudes. Two different tests are considered in addition (or “difference boxes”) approach, binary hypothesis test designed based on “smallest of” constant false alarm rate formulation. Probability arguments then invoked derive readily evaluated expressions detection thresholds. Tests conducted, indicating performance ratio-based comparable terms processed peak-to-background ratio. However, comparisons root mean-squared (RMS) error indicate difference-based (Haar) approach produces lower than both approaches. The filter further demonstrated remain effective (100% detection, 0% alarm, RMS estimation errors <3 % ) at low SNRs ∼0 dB.