LFM signal detection using LPP-Hough transform

This paper presents a new method to detect linear frequency modulated (LFM) signals by jointly using the local polynomial periodogram (LPP) and the Hough transform. Theoretical comparison is made on the 3 dB signal-to-noise ratios (SNRs), achieved by the pseudo-Wigner–Ville distribution (PWVD) and the LPP, to show that the latter is important to achieve significant increase of noise margins in the Hough transform domain. The results of computer simulations are presented for the detection of mono- and multi-component LFM signals corrupted by additive white Gaussian noise and impulsive noise. It is also found that by using the time–frequency filtering, the computational complexity of the detection can be substantially reduced. Both the theoretical analysis and the simulation results show that the proposed method achieves significant performance improvement on detecting the LFM signals in very low signal-to-noise ratio environments.

Research highlights
► Increasing the 3-dB signal to noise ratio with time-frequency analysis for chirp-like signals.
► Application of local polynomial periodogram and Hough transform (LHT).
► Detection of frequency modulated signals in low signal to noise ratio with reduced computational complexity.