Single-frequency pulse signal detection method

Abstract

The present invention relates to the field of signal processing technology, and in particular to a single-frequency pulse signal detection method, comprising receiving a signal; segmenting the received signal according to the pulse width, and detecting each segment of the signal x i The length of the signal is N; Fourier transform is performed on each segment of data; the discrete Fourier transform of the rectangular window signal with a length of N equal to the pulse width is calculated; deconvolution is then performed to obtain the signal spectrum; a detection threshold is set, the signal spectrum is detected, and the peak of the signal spectrum is found. When the peak is greater than the detection threshold, it is considered that a single-frequency pulse signal has been detected. The present invention uses a deconvolution method to deconvolve the signal spectrum, enhance the peak of the single-frequency pulse signal spectrum, eliminate frequency ambiguity and sidelobe leakage, and improve the subsequent detection probability.

Description

Technical Field

[0001] The present invention relates to the technical field of signal processing, and in particular to a single-frequency pulse signal detection method. Background Art

[0002] Single-frequency pulse signals are easy to generate and transmit, have a simple structure, and are easy to analyze and process. They are one of the most intensively studied and widely used signal forms. Single-frequency pulse signal detection is widely used in radar, sonar, communications, medical, and other fields. Due to their narrow bandwidth, single-frequency pulse signals are easily affected by transmission, reception, and medium motion, resulting in frequency deviation, making them difficult to detect using pulse compression technology. Existing conventional engineering methods generally use a Fourier transform of the received signal to search for peaks near a given frequency to detect the signal. The pulse signal has a limited pulse width and can be regarded as an infinite-length single-frequency signal windowed in the time domain. Therefore, a direct Fourier transform will cause the signal bandwidth to be broadened, resulting in frequency ambiguity, reduced resolution, and increased frequency estimation error. At the same time, the conventional method causes the pulse energy to leak to other frequency points, generating sidelobes, which affect other frequencies.

[0003] Based on this, this application is made. Summary of the Invention

[0004] In order to solve the above-mentioned defects existing in the prior art, the present invention provides a single-frequency pulse signal detection method, which uses a deconvolution method to deconvolute the signal spectrum, enhance the peak value of the single-frequency pulse signal spectrum, eliminate frequency ambiguity and sidelobe leakage to a certain extent, and improve the subsequent detection probability.

[0005] In order to achieve the above object, the technical solution adopted by the present invention is as follows:

[0006] A single-frequency pulse signal detection method includes the following steps:

[0007] Step 1, receiving a signal;

[0008] Step 2: Segment the received signal according to the pulse width. Each segment of the signal x i The length of is N;

[0009] Step 3, perform Fourier transform on each segment of data;

[0010] Step 4, calculating the discrete Fourier transform of the rectangular window signal equal to the pulse width, where the Fourier transform length is N;

[0011] Step 5, performing deconvolution using the Fourier transform obtained in steps 3 and 4 to obtain the signal spectrum;

[0012] Step 6: Set the detection threshold, detect the signal spectrum, and find the peak of the signal spectrum. When the peak is greater than the detection threshold, it is considered that a single-frequency pulse signal is detected, otherwise, no single-frequency pulse signal is detected.

[0013] The present invention provides a preferred solution, in step 2, the pulse signal exists completely in a certain segment of data after segmentation; adjacent signal segments partially overlap; the length of each segmented signal is twice that of the single-frequency pulse signal, and 50% of the length of at least one segment of the signal overlaps with the single-frequency pulse signal.

[0014] Compared with the prior art, the present invention can achieve the following beneficial technical effects:

[0015] (1) The present invention provides a single-frequency pulse signal detection method, studies the cause of frequency expansion, decomposes the single-frequency pulse signal into the product of a rectangular window function and a single-frequency signal, performs deconvolution in the frequency domain, increases the output signal-to-noise ratio, and improves the signal detection probability.

[0016] (2) At the same time, deconvolution eliminates the frequency ambiguity caused by the frequency expansion caused by windowing to a certain extent, making the frequency estimation more accurate.

[0017] (3) The present invention estimates and offsets the amplitude and position of the side lobes of the single-frequency pulse signal, reducing the influence of energy leakage and other frequencies.

[0018] (4) Compared with other high-resolution methods, this method basically does not affect the amplitude of other frequencies, does not change the spectral structure of the signal, and can more realistically reflect the energy distribution of each frequency.

[0019] (5) This method is independent of the frequency of the signal to be detected and can be applied to the detection of single-frequency pulse signals with unknown frequency or large frequency offset.

[0020] In summary, time domain windowing is equivalent to the convolution of the spectrum of the window function and the impulse function in the frequency domain. The spectrum of the window function is known. Therefore, the present invention uses a deconvolution method to perform deconvolution operations on the spectrum of the signal, enhance the peak value of the single-frequency pulse signal spectrum, eliminate frequency ambiguity and sidelobe leakage, and improve the subsequent detection probability. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 This is a flow chart of the single-frequency pulse signal detection method of this embodiment;

[0022] Figure 2 is a spectrum diagram of a typical rectangular window signal in this embodiment;

[0023] Figure 3 This is a schematic diagram of a segmented method for receiving signals in this embodiment;

[0024] Figure 4 This is the simulation result diagram using the conventional method;

[0025] Figure 5 Graph showing simulation results using the method of this embodiment. DETAILED DESCRIPTION

[0026] In order to make the technical means of the present invention and the technical effects that can be achieved more clearly and comprehensively disclosed, the following embodiments are provided and described in detail with reference to the accompanying drawings:

[0027] See also Figure 1 , a single-frequency pulse signal detection method of this embodiment includes the following:

[0028] Step 1, receiving a signal;

[0029] Step 2: Segment the received signal according to the pulse width. Each segment of the signal x i The length of N is to ensure that the pulse signal exists completely in a certain signal segment, while the data cannot be too long to ensure the signal-to-noise ratio of the signal; there must be a certain overlap between adjacent signal segments to prevent the pulse signal from being truncated and to retain certain historical information. In engineering, the signal segment length is usually twice that of the single-frequency pulse signal, with 50% overlap. Figure 3 The green part in the middle is a single-frequency pulse signal, which is completely retained in the data segment x i middle.

[0030] Step 3: Perform Fourier transform on each segment of data; to deal with the boundary problem, take the definition in The value on the left is X, retaining its amplitude and ignoring the phase part. i (k).

[0031] Step 4, calculate the discrete Fourier transform of the rectangular window signal equal to the pulse width, the Fourier transform length is N; the discrete Fourier transform is defined in The value on the G is obtained by retaining its amplitude and ignoring its phase part. T (k). The Fourier transform of a typical rectangular window signal is as follows Figure 2 shown.

[0032] Step 5: Use the Fourier transform obtained in steps 3 and 4 to perform Lucy-Richardson iterative deconvolution. According to the specific scenario, select the appropriate number of iterations. Take the middle N points of data from the result, which is the new definition in The signal spectrum X on i (k).

[0033] Step 6: Set the detection threshold, detect the signal spectrum, and i (k) Find the peak value. Find the peak value of the signal spectrum. When the peak value is greater than the detection threshold, it is considered that a single-frequency pulse signal is detected. Otherwise, no single-frequency pulse signal is detected.

[0034] The single-frequency pulse signal can be regarded as the product of the single-frequency signal and the rectangular window function, that is,

[0035] cw T (t) = G T (t)·exp(jω0t) (1)

[0036] Where cw T (t) is a single-frequency pulse signal, G T (t) is the rectangular window function of pulse width T, then cw T Spectrum of (t)

[0037]

[0038] Equation 2 represents the spectrum CW T (ω) is the rectangular window function G T Spectrum of (t) The convolution of the domain impulse function δ(ω-ω0) is exactly The function causes frequency ambiguity and side lobes. Therefore, this method considers using appropriate methods such as the Lucy-Richardson algorithm to perform deconvolution and then approximate the impulse function, reduce frequency ambiguity, eliminate side lobes, and concentrate energy on the main lobe.

[0039] like Figure 4 and Figure 5 , is a comparison of the normalized results of the conventional method and the proposed method for simulation data. It can be seen that the peak of the proposed method is sharper and the noise is smaller.

[0040] The above content is a further detailed description of the technical solution provided in combination with the preferred implementation methods of the present invention. It cannot be determined that the specific implementation of the present invention is limited to the above descriptions. For ordinary technicians in the technical field to which the present invention belongs, they can make several simple deductions or replacements without departing from the concept of the present invention, which should be regarded as falling within the scope of protection of the present invention.

Claims

1. A single-frequency pulse signal detection method, characterized in that: These include: Step 1, receiving a signal; Step 2: Segment the received signal according to the pulse width, and the length of each segment signal xi is N; In step 2, the pulse signal exists completely in a certain segment of data after segmentation, and adjacent signal segments partially overlap; Step 3, perform Fourier transform on each segment of data; In step 3, the Fourier transform is defined as The value on , retaining its magnitude but ignoring its phase part, is obtained Step 4, calculating the discrete Fourier transform of the rectangular window signal equal to the pulse width, where the Fourier transform length is N; In step 4, the discrete Fourier transform is defined as The value on the left retains its amplitude but ignores its phase part, and we get G T (k); Step 5, performing deconvolution using the Fourier transform obtained in steps 3 and 4 to obtain the signal spectrum; In step 5, Lucy-Richardson iterative deconvolution is used, and the appropriate number of iterations is selected. Take the middle N points of data from the result and define them as the new The signal spectrum X on i (k); Step 6: Set the detection threshold, detect the signal spectrum, and find the peak of the signal spectrum. When the peak is greater than the detection threshold, it is considered that a single-frequency pulse signal is detected, otherwise, no single-frequency pulse signal is detected.

2. A single-frequency pulse signal detection method as claimed in claim 1, characterized in that: In step 2, the length of each segmented signal is twice that of the single-frequency pulse signal, and 50% of the length of at least one segment of the signal overlaps with the single-frequency pulse signal.

Images