Signal TOA estimation method and device based on frequency domain weighted inverse Fourier transform

Abstract

The invention discloses a signal TOA estimation method and device based on frequency domain weighted inverse Fourier transform, and belongs to the field of radar signal processing, and the method comprises the steps: carrying out the frequency domain weighted parameter analysis and parameter estimation real-time processing, and determining a frequency domain weighting mode in the frequency domain weighted parameter analysis; in the parameter estimation real-time processing, a frequency domain weighting mode determined in frequency domain weighting parameter analysis is utilized to carry out frequency domain weighting calculation on a frequency spectrum data frame where a pulse front edge is located, and then a TOA estimation value is calculated by utilizing a TOA detection threshold determined in the frequency domain weighting parameter analysis. According to the method, bandwidth adaptive filtering requirements of signals with different pulse widths can be met, so that the signal-to-noise ratio of TOA estimation is improved, and particularly, the TOA estimation precision under the condition of low signal-to-noise ratio is improved.

Description

Technical Field

[0001] The present invention relates to the field of radar signal processing, and more specifically, to a signal TOA estimation method and device based on frequency domain weighted inverse Fourier transform. Background Art

[0002] Signal time of arrival (TOA) estimation is an important topic in passive radar signal processing and is the basis for measuring pulse width (PW) and pulse repetition interval (PRI) parameters. As part of the parameter measurement function of passive radar signal receivers, TOA estimation is inseparable from the signal detection process. Correct detection of the radiator signal is the prerequisite for TOA estimation. However, under low signal-to-noise ratio conditions, if only the signal is detected but TOA cannot be accurately measured, the radiator pulse signal lacks the correct TOA, PRI and PW parameter description, and signal detection loses its practical significance. The FFT processing receiver detects and estimates TOA frame by frame for broadband high-speed sampling signals: first, a long-length FFT spectrum is generated for high-sensitivity signal detection, and then the FFT spectrum outside a certain bandwidth is set to zero based on the pilot signal frequency (i.e., narrowband frequency domain filtering), and then IFFT is performed to transform back to the time domain to extract the envelope and perform TOA estimation. The bandwidth retained by the FFT spectrum needs to be considered to match the minimum measurable pulse width and TOA time domain resolution requirements. Generally, several times the frequency domain resolution will be taken. This results in the time domain signal-to-noise ratio after IFFT being lower than the FFT frequency domain detection signal-to-noise ratio in most pulse width situations. However, existing literature has shown that the signal-to-noise ratio required for accurate TOA estimation should be higher than the receiver sensitivity. Therefore, when the signal-to-noise ratio is low, the problem of detectable signals in the frequency domain but TOA measurement errors often occurs.

[0003] Among the traditional TOA estimation algorithms, the adaptive threshold method proposed by TORRIERID J is a commonly used method in engineering. It has a simple algorithm and high computational efficiency, but it does not meet the requirements for estimating TOA under low signal-to-noise ratio conditions. The segmented FFT method proposed by Chan YT et al. and the reverse correlation accumulation method proposed by Hu Guobing et al. directly process the original broadband sampling data superimposed with Gaussian white noise, which is not suitable for out-of-band noise filtering of FFT processing receivers. In addition, the calculation process is complicated and the engineering application scenarios are limited. The original broadband sampling data of the FFT processing receiver has been filtered to set the out-of-band spectrum to zero. The Gaussian white noise contained in it still has a certain amount of narrowband color noise after processing. Under low signal-to-noise ratio conditions, it will cause large fluctuations in the signal envelope and easily cause TOA estimation errors. Therefore, under low signal-to-noise ratio conditions, it is still necessary to consider improving the filtering processing method to further reduce the impact of noise on TOA estimation and improve the accuracy of TOA estimation. Summary of the invention

[0004] The purpose of the present invention is to overcome the shortcomings of the prior art and provide a signal TOA estimation method and device based on frequency domain weighted inverse Fourier transform. Specifically, according to the principle of TOA estimation based on inverse Fourier transform after narrowband filtering in an FFT processing receiver, a frequency domain weighted filtering and TOA correction estimation method is proposed to meet the bandwidth adaptive filtering requirements of signals with different pulse widths, thereby improving the signal-to-noise ratio of TOA estimation, especially improving the TOA estimation accuracy under low signal-to-noise ratio conditions.

[0005] The object of the present invention is achieved through the following solutions:

[0006] A signal TOA estimation method based on frequency domain weighted inverse Fourier transform comprises the following steps:

[0007] Frequency domain weighted parameter analysis and parameter estimation real-time processing, determining the frequency domain weighting method in the frequency domain weighted parameter analysis; in the parameter estimation real-time processing, using the frequency domain weighting method determined in the frequency domain weighted parameter analysis, performing frequency domain weighted calculation on the spectrum data frame where the pulse leading edge is located, and then using the TOA detection threshold determined in the frequency domain weighted parameter analysis to calculate the TOA estimation value.

[0008] Furthermore, the frequency domain weighted parameter analysis also includes the following sub-steps:

[0009] Step 1, determining a frequency domain weighting method;

[0010] Step 2, time domain waveform analysis of signals with different duty cycles based on frequency domain weighting;

[0011] Step 3: TOA detection threshold fitting.

[0012] Furthermore, the parameter estimation real-time processing further includes the following sub-steps:

[0013] Step S1, FFT frequency domain detection: generate an FFT spectrum for the sampled broadband digital signal, perform signal detection processing, and then output the signal frequency and determine the frame number where the pulse leading edge is located;

[0014] Step S2, spectrum zeroing and weighted calculation: based on the frequency domain weighting method determined in step 2 during the spectrum weighted parameter analysis process, perform frequency domain weighted calculation on the spectrum data frame where the pulse front edge of step S1 is located;

[0015] Step S3, IFFT transformation and envelope extraction: performing IFFT operation based on the weighted FFT spectrum in step S2, converting it into time domain waveform data, and calculating the amplitude envelope;

[0016] Step S4, TOA time point search: Based on the amplitude envelope data of step S3 and the TOA detection threshold fitting formula determined in step 3 during the frequency domain weighted parameter analysis, calculate the TOA amplitude detection threshold that matches the current time domain envelope waveform, then search for the TOA time point in the amplitude envelope data to give a TOA estimation value.

[0017] Furthermore, in step 1, the method for determining frequency domain weighting specifically includes sub-steps: frequency domain weighting specifically includes two parts: out-of-band filtering and in-band weighting. A certain filtering bandwidth is set, with the signal frequency as the center, and the spectrum weighting coefficient outside the bandwidth is set to zero. The spectrum weighting coefficient within the bandwidth should be positively correlated with the spectrum amplitude, so as to achieve the purpose of filtering out noise by adaptively matching the actual signal bandwidth.

[0018] Further, in step 2, the different duty cycle signals are analyzed based on the time domain waveform of frequency domain weighting, specifically including the sub-steps of:

[0019] The FFT spectrum is processed by the frequency domain weighting determined in step 1, and then the time domain waveform envelope is extracted by inverse Fourier transform. The TOA amplitude threshold is adjusted. The influence on the optimal TOA amplitude threshold is comprehensively analyzed by analyzing the time domain waveform of signals with different duty cycles based on the frequency domain weighting.

[0020] Furthermore, in step 3, the TOA detection threshold fitting specifically includes the sub-steps of: in the case of frequency domain weighted inverse Fourier transform processing, fitting and generating a new TOA amplitude threshold formula according to the filtered time domain waveform data corresponding to different duty cycle signals.

[0021] Further, in step S1, generating the FFT spectrum specifically includes generating the FFT spectrum frame by frame.

[0022] Further, in step S3, searching for the TOA time point specifically includes searching for the TOA time point starting from a maximum value.

[0023] Further, searching for the TOA time point from the maximum value specifically includes searching for the TOA time point in reverse order from the maximum value.

[0024] A signal TOA estimation device based on frequency domain weighted inverse Fourier transform comprises a processor and a memory, wherein a computer program is stored in the memory, and when the computer program is loaded by the processor, any of the above methods is executed.

[0025] The beneficial effects of the present invention include:

[0026] Compared with the unweighted spectrum IFFT processing method, the frequency domain weighting method of the present invention realizes adaptive matched filtering processing with the actual signal bandwidth, improves the TOA estimation signal-to-noise ratio, and especially improves the TOA estimation accuracy of the original broadband data under low signal-to-noise ratio conditions.

[0027] The present invention aims at the problem that the signal-to-noise ratio required for effective TOA estimation in a traditional FFT processing receiver is lower than the frequency domain detection capability, reduces the signal-to-noise ratio requirement for effective TOA estimation, makes it closer to the frequency domain detection capability, and improves the overall working capability of the receiver. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0029] Figure 1 It is a schematic diagram of the composition of the FFT processing receiver;

[0030] Figure 2 It is a schematic diagram of the process of the present invention;

[0031] Figure 3 The TOA detection threshold fitting analysis of different duty cycle signals after frequency domain weighting and IFFT processing in a specific embodiment;

[0032] Figure 4 The Monte Carlo algorithm verification effect analysis of two different TOA under different SNR conditions in the specific embodiment;

[0033] Figure 5 This is an analysis of the Monte Carlo algorithm verification effect for two different TOA under different SNR conditions in a specific embodiment. DETAILED DESCRIPTION

[0034] All features disclosed in all embodiments in this specification, or steps in all methods or processes implicitly disclosed, except for mutually exclusive features and / or steps, can be combined and / or expanded or replaced in any manner.

[0035] In a preferred embodiment, the present invention relates particularly to the field of passive radar signal processing, and in particular to a method for estimating the arrival time of a signal in an FFT processing receiver, such as Figure 1 and Figure 2 As shown, the processing process is divided into two parts: frequency domain weighted parameter analysis and parameter estimation real-time processing.

[0036] The steps of frequency domain weighted parameter analysis are as follows:

[0037] Step 1, determine the frequency domain weighting method: frequency domain weighting includes out-of-band filtering and in-band weighting. Set a certain filtering bandwidth, take the signal frequency as the center, set the spectrum weighting coefficient outside the bandwidth to zero, and the spectrum weighting coefficient within the bandwidth should be positively correlated with the spectrum amplitude, so as to achieve the purpose of adaptively matching the actual signal bandwidth to filter out noise;

[0038] Step 2, time domain waveform analysis based on frequency domain weighting of signals with different duty cycles: The frequency domain weighting determined in step 1 is applied to the FFT spectrum, and then IFFT (inverse Fourier transform) is performed to extract the time domain waveform envelope. Compared with unweighted IFFT processing in the frequency domain, the pulse waveform will be broadened and the amplitude value corresponding to the real TOA will be changed. Therefore, the TOA amplitude threshold needs to be adjusted. By analyzing the time domain waveforms of signals with different duty cycles based on frequency domain weighting, the impact on the optimal TOA amplitude threshold is comprehensively analyzed.

[0039] Step 3, TOA detection threshold fitting: In the case of frequency domain weighted IFFT processing, a new TOA amplitude threshold formula is generated by fitting according to the filtered time domain waveform data corresponding to signals with different duty cycles.

[0040] The real-time processing steps for parameter estimation are as follows:

[0041] Step S1, FFT frequency domain detection: Generate a long-length FFT spectrum frame by frame for the sampled broadband digital signal, perform signal detection processing, and then output the signal frequency and determine the frame number where the pulse leading edge is located;

[0042] Step S2, spectrum zeroing and weighted calculation: based on the frequency domain weighting method determined in step 2 of the spectrum weighted parameter analysis process, perform frequency domain weighted calculation on the spectrum data frame where the pulse leading edge of step S1 is located;

[0043] Step S3, IFFT transformation and envelope extraction: performing IFFT operation based on the weighted FFT spectrum in step S2, converting it into time domain waveform data, and calculating the amplitude envelope;

[0044] Step S4, TOA time point search: Based on the amplitude envelope data of step S3 and the TOA detection threshold fitting formula of step 3 of the frequency domain weighted parameter analysis process, calculate the TOA amplitude detection threshold that matches the current time domain envelope waveform, and then search the TOA time point in reverse order starting from the maximum value in the amplitude envelope data to give a TOA estimation value.

[0045] Example 1

[0046] The FFT processing receiver of this embodiment adopts 1 GHz broadband digital sampling, and processes the original broadband sampling data frame by frame according to 1024-point FFT to generate spectrum and perform frequency domain signal detection. TOA estimation is completed by time domain processing after FFT spectrum weighting and IFFT transformation.

[0047] The frequency domain weighted parameter analysis process is processed as follows:

[0048] Step 1, determine the frequency domain weighting method: take the signal detection frequency as the center, set the weighting coefficient of the out-of-band spectrum beyond the difference of 10MHz to 0, and set the weighting coefficient of the in-band frequency point within the difference of 10MHz to the normalized amplitude coefficient of the corresponding frequency point (that is, the frequency point weighting coefficient is completely proportional to the amplitude);

[0049] Step 2, fitting the TOA amplitude detection threshold after frequency domain weighting: Generate random samples for typical frequency pulse signals with different duty cycles, extract the time domain amplitude envelope after frequency domain weighting and IFFT, and then normalize the amplitude envelope data relative to the maximum amplitude, and observe the relationship between the normalized envelope data and the amplitude detection threshold corresponding to the known TOA, such as Figure 3 As shown, the minimum value of the normalized amplitude envelope is selected as the feature, and a broken line is used to approximate the TOA amplitude detection threshold;

[0050] th(x)=max{(x-0.35)·0.8+0.45,0.45}

[0051] Where x represents the minimum value of the normalized amplitude envelope, and th(x) is the normalized TOA amplitude detection threshold calculated by fitting.

[0052] The real-time processing of parameter estimation is processed as follows:

[0053] Step S1, FFT frequency domain detection: perform 1024-point FFT spectrum calculation frame by frame for the input broadband sampling data, and perform signal detection according to a certain frequency domain detection threshold or based on the CFAR threshold, output the frequency point of the detected signal, and determine the frame number where the pulse leading edge is located, and then perform TOA estimation on the FFT spectrum data of this frame;

[0054]

[0055] Wherein, X(k) is the FFT calculation result of the frame where the pulse front is located, and X(k0) is the FFT calculation result of the detection signal frequency point.

[0056] Step S2, frequency domain zeroing and weighted calculation: based on the spectrum weighting method determined by the frequency domain weighted parameter analysis process, the FFT spectrum is corrected;

[0057]

[0058] Among them, X M (k), k = 0, 1, 2, ..., 511 is the weighted spectrum data.

[0059] Step S3, IFFT transformation and envelope extraction: performing IFFT operation based on the weighted FFT spectrum in step 2;

[0060]

[0061] Then extract the envelope amplitude data, i.e. |x M (n)|,n=0,1,2,…,511. It should be noted that the sampling time interval here becomes twice the original AD sampling interval.

[0062] Step S4: Based on the amplitude envelope data, the TOA detection threshold is generated according to the fitting formula of step 2 of the frequency domain weighted parameter analysis process:

[0063]

[0064] Search for TOA time points in the amplitude envelope data: Based on the amplitude detection threshold, start searching for TOA time points in reverse order from the maximum value point of the amplitude envelope data, and select the time points that meet the conditions. Make TOA estimates;

[0065]

[0066] In order to illustrate the effect of the algorithm of the present invention, a single frame of 1 GHz sampling data (1024 points) containing the pulse front edge was simulated, two TOA parameters (0.4us and 0.8us) were selected for simulation, and Gaussian white noise with different signal-to-noise ratios was superimposed. The effect of the method of the present invention and other three algorithms (the adaptive threshold method proposed by TORRIERID J, the segmented FFT method proposed by Chan YT et al., and the reverse correlation accumulation method proposed by Hu Guobing et al.) were verified. Figure 4 is the RMS error statistics of the four algorithms at different SNRs (TOA = 0.4us), Figure 5 The RMS error statistics of the four algorithms at different SNRs (TOA = 0.8us). Figure 4 and Figure 5 It can be seen that in the low signal-to-noise ratio range (typical range -12db to -5db), the method of the present invention is superior to the other three methods, and the signal-to-noise ratio required for effective TOA estimation is closer to the frequency domain detection capability. In the high signal-to-noise ratio range (above -5db), although the estimation accuracy of the method of the present invention is slightly lower than that of other methods, the root mean square error of 30ns can already meet the needs of most scenarios.

[0067] The units involved in the embodiments of the present invention may be implemented by software or hardware, and the units described may also be arranged in a processor. The names of these units do not, in some cases, limit the units themselves.

[0068] According to one aspect of an embodiment of the present invention, a computer program product or a computer program is provided, the computer program product or the computer program includes a computer instruction, and the computer instruction is stored in a computer-readable storage medium. A processor of a computer device reads the computer instruction from the computer-readable storage medium, and the processor executes the computer instruction, so that the computer device executes the method provided in the above various optional implementations.

[0069] As another aspect, an embodiment of the present invention further provides a computer-readable medium, which may be included in the electronic device described in the above embodiment; or may exist independently without being assembled into the electronic device. The above computer-readable medium carries one or more programs, and when the above one or more programs are executed by an electronic device, the electronic device implements the method described in the above embodiment.

Claims

1. A signal TOA estimation method based on frequency domain weighted inverse Fourier transform, characterized in that: The following steps are involved: Frequency domain weighted parameter analysis and parameter estimation real-time processing, determining the frequency domain weighting method in the frequency domain weighted parameter analysis; in the parameter estimation real-time processing, using the frequency domain weighting method determined in the frequency domain weighted parameter analysis, performing frequency domain weighted calculation on the spectrum data frame where the pulse leading edge is located, and then using the TOA detection threshold determined in the frequency domain weighted parameter analysis to calculate the TOA estimation value.

2. The signal TOA estimation method based on frequency domain weighted inverse Fourier transform according to claim 1 is characterized in that: The frequency domain weighted parameter analysis also includes the following sub-steps: Step 1, determining a frequency domain weighting method; Step 2, time domain waveform analysis of signals with different duty cycles based on frequency domain weighting; Step 3: TOA detection threshold fitting.

3. The signal TOA estimation method based on frequency domain weighted inverse Fourier transform according to claim 2 is characterized in that: The parameter estimation real-time processing further includes the following sub-steps: Step S1, FFT frequency domain detection: generate an FFT spectrum for the sampled broadband digital signal, perform signal detection processing, and then output the signal frequency and determine the frame number where the pulse leading edge is located; Step S2, spectrum zeroing and weighted calculation: based on the frequency domain weighting method determined in step 2 during the spectrum weighted parameter analysis process, perform frequency domain weighted calculation on the spectrum data frame where the pulse front edge of step S1 is located; Step S3, IFFT transformation and envelope extraction: performing IFFT operation based on the weighted FFT spectrum in step S2, converting it into time domain waveform data, and calculating the amplitude envelope; Step S4, TOA time point search: Based on the amplitude envelope data of step S3 and the TOA detection threshold fitting formula determined in step 3 during the frequency domain weighted parameter analysis, calculate the TOA amplitude detection threshold that matches the current time domain envelope waveform, then search for the TOA time point in the amplitude envelope data to give a TOA estimation value.

4. The signal TOA estimation method based on frequency domain weighted inverse Fourier transform according to claim 2, characterized in that: In step 1, the method for determining frequency domain weighting specifically includes sub-steps: frequency domain weighting specifically includes two parts: out-of-band filtering and in-band weighting. A certain filtering bandwidth is set, with the signal frequency as the center, and the spectrum weighting coefficient outside the bandwidth is set to zero. The spectrum weighting coefficient within the bandwidth should be positively correlated with the spectrum amplitude, so as to achieve the purpose of filtering out noise by adaptively matching the actual signal bandwidth.

5. The signal TOA estimation method based on frequency domain weighted inverse Fourier transform according to claim 2, characterized in that: In step 2, the different duty cycle signals are analyzed based on the time domain waveforms weighted by the frequency domain, specifically including the following sub-steps: The FFT spectrum is processed by the frequency domain weighting determined in step 1, and then the time domain waveform envelope is extracted by inverse Fourier transform. The TOA amplitude threshold is adjusted. The influence on the optimal TOA amplitude threshold is comprehensively analyzed by analyzing the time domain waveform of signals with different duty cycles based on the frequency domain weighting.

6. The signal TOA estimation method based on frequency domain weighted inverse Fourier transform according to claim 2, characterized in that: In step 3, the TOA detection threshold fitting specifically includes the sub-steps of: in the case of frequency domain weighted inverse Fourier transform processing, according to the filtered time domain waveform data corresponding to different duty cycle signals, fitting and generating a new TOA amplitude threshold formula.

7. The signal TOA estimation method based on frequency domain weighted inverse Fourier transform according to claim 3 is characterized in that: In step S1, generating the FFT spectrum specifically includes generating the FFT spectrum frame by frame.

8. The signal TOA estimation method based on frequency domain weighted inverse Fourier transform according to claim 3 is characterized in that: In step S3, searching for the TOA time point specifically includes searching for the TOA time point starting from a maximum value.

9. The signal TOA estimation method based on frequency domain weighted inverse Fourier transform according to claim 8, characterized in that: The searching for the TOA time point from the maximum value specifically includes searching for the TOA time point in reverse order from the maximum value.

10. A signal TOA estimation device based on frequency domain weighted inverse Fourier transform, characterized in that: The method comprises a processor and a memory, wherein a computer program is stored in the memory, and when the computer program is loaded by the processor, the method according to any one of claims 1 to 9 is executed.

Images