TOA Estimation of Unknown Chirp Signal Based on Short Time FRFT and Hough Transform

Abstract

This paper studies the time-of-arrival (TOA) estimation problem for unknown chirp signals. The signal envelope is assumed to be trapezoidal, and TOA is defined as the arrival time for envelope to rise to the half-platform value. The current state-of-the-art technique is based on short time fractional Fourier transform (STFRFT). It utilizes STFRFT to obtain the time-profile of time-frequency spectrum, and proves that the profile can be used to estimate TOA. However, this method uses only the peak of profile, which induces information losses and thus degrades the estimation accuracy. To alleviate this, in this paper, Hough Transform is combined with STFRFT to make better use of the time-frequency spectrum. On the other hand, the chirp rate of signal is conventionally assumed to be accurately detected by a previous FRFT step. But the accuracy is actually limited by the computational complexity of FRFT. Hence, a dechirp technique is also involved to make improvements on this aspect. Numerical results show that the dechirp technique brings advantages in detecting chirp rate; and combining Hough Transform with STFRFT is beneficial for TOA estimation.