A fast algorithm for multi-component LFM signal analysis exploiting segmented DPT and SDFrFT

Abstract

This paper addresses the problem of estimating the chirp rates of multi-component linear frequency modulated signals, which is important in radar, sonar and navigation signal processing. The main difficulties in the estimation procedure lie in the cross-terms between multi-components and the high computation burden. To solve these problems, a novel algorithm that combines segmented discrete polynomial-phase transform and sparse discrete fractional Fourier transform is proposed to yield a significant reduction of the computational load with a satisfactory estimation performance. Simulation results are provided to demonstrate the effectiveness of the proposed approach.

Segmented discrete polynomial‐phase transform with coprime sampling

“…In this paper, we developed a modified segmented DPT in which multiple segmentation sets with coprime segmentation rates are exploited to achieve unambiguous chirp estimation of high chirp rates. As such, it yields a high accuracy and a low complexity but significantly extends the detectable chirp rates from the previous work [6]. For the convenience of readers who wish to implement the proposed algorithm and reproduce the numerical simulations, a MATLAB demo is available at https://dx.doi.org/10.13140/RG.2.2.24899.76326.…”

“…A problem with the DPT is that, in the scenarios where the chirps to be processed is with low signal‐to‐noise ratio (SNR), the above signal product is very noisy, thus compromising the accuracy of chirp parameter estimation. In [6], a segment‐integration‐based approach is proposed to effectively implement the DPT algorithm for weak LFM signals. More specifically, the entire data is first split into multiple non‐overlapping segments, and the coherent integration of the signal over a segment enhances the LFM signal.…”

Modulation Parameter Estimation of LFM Interference for Direct Sequence Spread Spectrum Communication System in Alpha-Stable Noise

“…According to the LFM interference model and the alphastable noise model described in (1) Let the estimated vector be θ = [f 0 , k]. Then, the loglikelihood function of the estimated vector θ is expressed as follows [24]…”

“…L INEAR frequency modulation (LFM) occupies a wide bandwidth to offer a high system processing gain. Due to this attribute, LFM have been widely used in wireless communications, radar, and sonar systems [1]. Due to the wide applications of LFM, wireless communication systems, such as direct sequence spread spectrum (DSSS) communication systems, may suffer from the interference caused by LFM interference, which can be characterized as a common broadband non-stationary interference.…”

An Improved FPGA Implementation of Sparse Fast Fourier Transform

“…If two or more black balls are divided into the same bucket, that is, if multiple large frequency points are mixed together, then the subsequent FFT/inverse FFT (IFFT) calculation results will inevitably lead to errors. This phenomenon is termed collision, in order to avoid which, one must ensure that the large frequency coefficients follow a uniform random distribution [5,6]. Nevertheless, the commonly encountered signals seldom satisfy this condition.…”