Detection and arrival time estimation of a pulsed sinusoid

A novel double-FFT method is presented for the computationally efficient detection and time-of-arrival (TOA) estimation of a pulsed sinusoid. Starting from the time samples x(n), the first FFT produces X(k), whose magnitude is summed for detection. If a pulse is detected, a second complex FFT, of the X(k)/sup 2/, produces a peak whose location corresponds to the TOA. Simulation results have been used to generate the ROC curves for the detector, and an evaluation of the TOA estimation accuracy. For SNR /spl ges/ 5 db, the TOA root mean square errors are less than ten time samples.

[1]  Don Torrieri,et al.  Adaptive Thresholding Systems , 1977, IEEE Transactions on Aerospace and Electronic Systems.

[2]  K. C. Ho,et al.  Pulse arrival time estimation based on pulse sample ratios , 1995 .

[3]  B. H. Maranda On the false alarm probability for an overlapped FFT processor , 1996 .

[4]  Yiu-Tong Chan,et al.  Comparison of various periodograms for sinusoid detection and frequency estimation , 1999 .

[5]  R. Inkol,et al.  A comparative study of FFT-summation and polyphase-FFT CFAR detectors , 2004, Canadian Conference on Electrical and Computer Engineering 2004 (IEEE Cat. No.04CH37513).