Pulse linear frequency modulation signal detection method and system based on FPGA

Abstract

The invention discloses a pulse linear frequency modulation signal detection method and system based on an FPGA, and belongs to the field of radar signal processing. The signal detection method comprises the following steps: setting a receiving window length and a receiving window sliding length of an acquired signal according to the length of a target signal to be detected; conjugate inversion is carried out on the collected target signal to be detected to obtain a corresponding matched filter and an amplitude-frequency response, and the matched filter and the amplitude-frequency response are stored in the FPGA; carrying out data recombination on the acquired to-be-detected target signal, and converting single-path serial data into parallel P-path data; compressing the processed parallel data to obtain compressed pulse signal data; carrying out peak position detection; and if the peak position is in a preset interval [L0-N, L0] in the receiving window and the peak signal-to-noise ratio is greater than a threshold value, determining that the receiving window detects the target signal, and predicting the arrival time of the next signal. Precise detection and prediction of pulse signals are realized, the prediction time is sufficient, and successful in-orbit calibration of the spectrometer is ensured.

Description

Technical Field

[0001] The present invention belongs to the field of radar signal processing, and more specifically, relates to a pulse linear frequency modulation signal detection method and system based on FPGA. Background Art

[0002] The SWIM (Surface Waves Investigation and Monitoring) instrument, carried on the Sino-French ocean satellite, is the world's first satellite-borne spectrometer. It is a small-incidence-angle, multi-beam, rotating, scanning true-aperture radar capable of global ocean wave detection in the 20-500m wavelength range. Its wave directional spectrum data products are of great research and application value. The spectrometer's wave directional spectrum detection principle is to extract the range-dependent high-resolution sea surface backscatter coefficient from the received power based on an accurate calibration factor, thereby inverting the wave directional spectrum. It is not difficult to see that the accuracy of the spectrometer's wave spectrum detection depends on the accuracy of the calibration factor.

[0003] The calibration factors of the satellite-borne spectrometer can be obtained through external calibration using a ground-based calibrator. As the spectrometer passes through the calibrator, it sequentially transmits a series of high-bandwidth, long-pulse-width linear frequency modulation signals at different incident beams (i.e., 0°, 2°, 4°, 6°, 8°, and 10°). Each beam operates for tens of milliseconds, and the signal repetition rate remains constant within a duty cycle, but varies between incident beams. The calibrator receives and detects the spectrometer's transmitted signals and forwards them to the spectrometer within the satellite-borne spectrometer's signal reception window (within tens of microseconds). The spectrometer's radar parameters can be inverted from the received signals from the calibrator and spectrometer, thereby obtaining the calibration factors.

[0004] As shown in the above on-orbit spectrometer calibration process, the calibrator needs to detect signals with a large number of points and wide bandwidth in real time (within tens of microseconds). Existing technologies are mostly used for processing small-bandwidth signals and generally do not require high real-time performance, making them unsuitable for on-orbit spectrometer calibration. Summary of the Invention

[0005] In response to the defects of the related art, the purpose of the present invention is to provide a pulse linear frequency modulation signal detection method and system based on FPGA, aiming to solve the problem that the existing technology solutions are not suitable for application scenarios of spectrometer signal detection and processing, and there are problems such as long signal processing time, inability to timely predict the signal arrival time, and insufficient forwarding time, which leads to failure of on-orbit calibration of the spectrometer.

[0006] To achieve the above object, the present invention provides a pulse linear frequency modulation signal detection method based on FPGA, comprising:

[0007] S1. Use a receiving window to continuously slide and collect the original target signal to be detected; conjugate the collected target signal to be detected to obtain the corresponding matched filter, fill the time domain length of the matched filter with zeros to the receiving window length, and then obtain the amplitude-frequency response H(n) of the matched filter through Fourier transform, and store it in the FPGA; wherein the receiving window length is L0, the sliding step length is L1, the signal length of the original target signal to be detected is N, N <L0≤2N,0<L1≤L0-N;

[0008] S2. Reorganize the collected target signal to be detected, convert the single-channel serial data into P-channel parallel data, and the length of each line of data is Q;

[0009] S3. Performing Fourier transform on the parallel data obtained after processing, and performing point multiplication with the amplitude-frequency response of the matched filter to obtain the frequency domain result of pulse compression, and then performing inverse Fourier transform on the point-multiplied data to obtain the time domain result of pulse compression, thereby obtaining compressed pulse signal data;

[0010] S4. Detecting the peak position of the compressed pulse signal data; if the peak position is within a preset interval [L0-N, L0] within the receiving window, determining that a complete pulse has been received, and calculating the peak signal-to-noise ratio of the pulse compression signal; if the peak position is not within the preset interval within the receiving window, discarding the data within the window;

[0011] S5. Determine whether the peak signal-to-noise ratio is greater than a threshold; if so, determine that the receiving window has detected the target signal, record the arrival time of the target signal each time it is detected, calculate the time interval between it and the last detected target signal, and predict the arrival time of the next signal; if not, determine that the receiving window has not detected the target signal.

[0012] Optionally, the target signal to be detected is a pulse linear frequency modulation signal, and the expression is:

[0013]

[0014] Where A is the signal amplitude, K is the signal linear modulation frequency, T is the signal pulse width, and t is time.

[0015] Optionally, the expression of the amplitude-frequency response H(t) of the matched filter is:

[0016] H(t)=FFT[x(t0-t)]

[0017] Where t0 is the delay between the impulse response of the matched filter and the signal.

[0018] Optionally, the receiving window length L0 satisfies L0=2 m , m is a positive integer.

[0019] Optionally, P and Q need to satisfy the following formula:

[0020] P=2 p ,Q=2 q

[0021] Wherein, p and q are both positive integers;

[0022] The relationship with the receiving window length L0 is:

[0023] L0=P×Q=2 p+q =2 m

[0024] Among them, p+q=m.

[0025] Optionally, when performing Fourier transform on the data, an adaptive time decimation radix 2FFT algorithm is used; when performing inverse Fourier transform on the data, an adaptive frequency decimation radix 2FFT algorithm is used.

[0026] Optionally, the calculating of the peak signal-to-noise ratio of the pulse compression signal includes:

[0027] Calculate all noise within the entire receiving window and remove the peak value;

[0028] The decrease in the signal-to-noise ratio within the window caused by the aliased signal is eliminated by adjusting the denominator when calculating the signal-to-noise ratio within the window.

[0029] In a second aspect, the present invention further provides a signal detection system based on an FPGA pulse linear frequency modulation signal, comprising:

[0030] The receiving window setting module is used to continuously slide and collect the original target signal to be detected using the receiving window; conjugate the collected target signal to be detected to obtain the corresponding matched filter, fill the time domain length of the matched filter with zeros to the receiving window length, and then obtain the amplitude-frequency response H(n) of the matched filter through Fourier transform, and store it in the FPGA; wherein the receiving window length is L0, the sliding step length is L1, the signal length of the original target signal to be detected is N, N <L0≤2N,0<L1≤L0-N;

[0031] The data reassembly module is used to reassemble the collected target signal to be detected, converting the single-channel serial data into P-channel parallel data, with the length of each line of data being Q;

[0032] The signal compression module is used to perform Fourier transform on the parallel data obtained after processing, and perform point multiplication with the amplitude-frequency response of the matched filter to obtain the frequency domain result of pulse compression, and then perform inverse Fourier transform on the frequency domain result to obtain the time domain result of pulse compression, thereby obtaining compressed pulse signal data;

[0033] a peak position detection module for detecting the peak position of the compressed pulse signal data; if the peak position is within the preset interval [L0-N, L0]] within the receiving window, it is determined that a complete pulse has been received, and the peak signal-to-noise ratio of the pulse compression signal is calculated; if the peak position is not within the preset interval within the receiving window, the data within the window is discarded;

[0034] The pulse signal detection module is used to determine whether the peak signal-to-noise ratio is greater than a threshold; if so, it is determined that the receiving window has detected the target signal, record the arrival time of the target signal each time it is detected, calculate the time interval between it and the last detected target signal, and predict the arrival time of the next signal; if not, it is determined that the receiving window has not detected the target signal.

[0035] The above technical solutions conceived by the present invention can achieve the following beneficial effects compared with the prior art: the embodiment of the present invention provides a pulse linear frequency modulation signal detection method based on FPGA, which shortens the width of the receiving window, shortens the processing data, and speeds up the processing flow; by determining the position of the peak of the processing result within the window, a peak position constraint is added. When the acquisition window is shortened, this constraint can ensure that the acquired signal is a complete pulse, so that the calculated peak value should be the maximum value of the pulse pressure theory; at the same time, because the peak position is constrained, the range of the signal used for judgment is narrowed, thereby reducing the false alarm range caused by noise; by reorganizing the signal data obtained in the receiving window, converting serial data into parallel data for processing, multiple channels of data are processed synchronously, greatly improving the processing speed. The prior art solutions are not applicable to the current application scenarios of spectrometer signal detection and processing, and there are technical problems such as the inability to timely predict and insufficient signal forwarding time, which leads to the failure of spectrometer on-orbit calibration. The method achieves the beneficial effect of accurately detecting each pulse signal and accurately predicting the next pulse signal, and the prediction time is sufficient to ensure that the time of signal transmission is within the receiving window of the spectrometer, thereby ensuring the successful on-orbit calibration of the spectrometer. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 This is a schematic diagram of signal reception and receiving window sliding in a pulse linear frequency modulation signal detection method based on FPGA provided by the present invention;

[0037] Figure 2 This is a schematic diagram of reorganizing receiving window data in a pulse linear frequency modulation signal detection method based on FPGA provided by the present invention;

[0038] Figure 3 This is a flow chart of pulse compression in a pulse linear frequency modulation signal detection method based on FPGA provided by the present invention;

[0039] Figure 4 It is a schematic diagram of frequency domain multiplication of a pulse signal and an amplitude-frequency response of a matched filter in a pulse linear frequency modulation signal detection method based on FPGA provided by the present invention;

[0040] Figure 5 This is a simulation diagram of a pulse linear frequency modulation signal detection method based on FPGA provided by the present invention, in which the signal is located at the head of the receiving window;

[0041] Figure 6 This is a simulation diagram of a pulse linear frequency modulation signal detection method based on FPGA provided by the present invention, in which the signal is located at the tail end of the receiving window;

[0042] Figure 7 This is a schematic diagram of pulse compression of a spectrometer's actual predicted signal result in a pulse linear frequency modulation signal detection method based on FPGA provided by the present invention;

[0043] Figure 8 The present invention provides a schematic diagram of the power of the transmitted and received signals in a pulse linear frequency modulation signal detection method based on FPGA. DETAILED DESCRIPTION

[0044] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely for the purpose of explaining the present invention and are not intended to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below may be combined with each other as long as they do not conflict with each other.

[0045] The contents involved in the above embodiment are described below in conjunction with a preferred embodiment.

[0046] Example 1

[0047] An embodiment of the present invention provides a pulse linear frequency modulation signal detection method based on FPGA, comprising:

[0048] S1. Use a receiving window to continuously slide and collect the original target signal to be detected; conjugate the collected target signal to be detected to obtain the corresponding matched filter, fill the time domain length of the matched filter with zeros to the receiving window length, and then obtain the amplitude-frequency response H(n) of the matched filter through Fourier transform, and store it in the FPGA; wherein the receiving window length is L0, the sliding step length is L1, the signal length of the original target signal to be detected is N, N <L0≤2N,0<L1≤L0-N;

[0049] S2. Reorganize the collected target signal to be detected, convert the single-channel serial data into P-channel parallel data, and the length of each line of data is Q;

[0050] S3. Performing Fourier transform on the parallel data obtained after processing, and performing point multiplication with the amplitude-frequency response of the matched filter to obtain the frequency domain result of pulse compression, and then performing inverse Fourier transform on the frequency domain result to obtain the time domain result of pulse compression, thereby obtaining compressed pulse signal data;

[0051] S4. Detecting the peak position of the compressed pulse signal data; if the peak position is within a preset interval [L0-N, L0] within the receiving window, determining that a complete pulse has been received, and calculating the peak signal-to-noise ratio of the pulse compression signal; if the peak position is not within the preset interval within the receiving window, discarding the data within the window;

[0052] S5. Determine whether the peak signal-to-noise ratio is greater than a threshold; if so, determine that the receiving window has detected the target signal, record the arrival time of the target signal each time it is detected, calculate the time interval between it and the last detected target signal, and predict the arrival time of the next signal; if not, determine that the receiving window has not detected the target signal.

[0053] In current applications of spectrometer signal detection and processing, the spectrometer cannot receive signals during transmission. Therefore, transmission must occur within the spectrometer's reception window. To ensure successful on-orbit calibration of the spectrometer, the signal arrival time and the spectrometer's receiving window must be predicted based on the pulse compression time. To address the challenges of existing technologies, a comprehensive pulse compression signal detection scheme is needed for the calibrator. This scheme ensures that every signal reaching the threshold is detected while also ensuring high real-time pulse compression. This ensures that the signal prediction time is sufficiently advanced, ensuring that the calibrator's signal is emitted within the spectrometer's receiving window. This paper proposes an FPGA-based pulse linear frequency modulation signal detection method that constrains the peak of the detection signal to the non-aliasing region while rapidly implementing pulse compression to ensure real-time signal prediction. The method focuses on reducing the amount of data processing required for pulse compression detection, increasing the speed and accuracy of signal detection, and reducing false alarm rates due to noise.

[0054] The signal to be processed x(n) is a pulse linear frequency modulation signal, and its expression is:

[0055]

[0056] Where A is the signal amplitude, K is the signal linear modulation frequency, T is the signal pulse width, and t is time.

[0057] The known pulse width of the chirp signal to be detected is \(T\), and the sampling rate of the signal is \(f\). After collecting the pulse signal, its width is \(N = Tf\). If aliasing is not to occur in the result of frequency-domain pulse compression processing, the data length to be processed needs to be at least \(2N\). Therefore, in order to reduce the total amount of data within a receiving window, the width \(L_0\) of the receiving window ranges from \(N\lt L_0\lt2N\). Based on determining the value of \(L_0\), to enable fast processing on the digital board later, the requirements of the data format need to be met, that is, both the data length and the arrangement format need to satisfy powers of 2. Therefore, the receiving window length \(L_0\) satisfies \(L_0 = 2\) m , where both \(N\) and \(m\) are positive integers; among them, \(m=[\log_2N]+1\), and the symbol “[]” in the formula represents the integer function. Performing anti-folding conjugation on the target signal to be detected from the spectrometer can obtain the time-domain expression of the matched filter: \(h(t)=x(t_0 - t)\), where \(t_0\) is the time delay between the impulse response of the matched filter and the signal, and its length is \(N\) at this time; padding zeros at the end of the frequency-domain filter to extend its length from \(N\) to \(L_0\) to obtain \(h\) ′ (n), and performing Fourier transform to obtain the spectrum of the matched filter, with the expression:

[0058] H(t)=FFT[x(t_0 - t)]

[0059] Its length is also \(L_0\). To speed up the data processing rate later, it is stored in the FPGA in advance.

[0060] The sliding length of the receiving window needs to ensure that for any pulse signal, there must be a window within which the signal can be completely collected. Therefore, the range of the sliding length of the receiving window is \(0\lt L_1\leq L_0 - L\). At the same time, to reduce the total amount of data processing, in this embodiment, the sliding length of the receiving window is taken as \(L_1 = L_0 - L\). The receiving window slides continuously to collect signals, and each time it slides, it outputs the data of one window, realizing continuous collection and output of signals. As Figure 1 shown, it shows how the receiving window collects signals and how the receiving window slides. In the figure, \(L_0\) is greater than the signal length and less than 2 times the signal length.

[0061] Recombine the data within the receiving window to obtain multi-channel parallel data, and based on this recombination method, pre-store the frequency-domain response of the matched filter. At the same time, merge the data streams output after recombination into one data stream for pulse compression. Specifically, it includes: dividing the data within one receiving window into \(P\) channels for parallel output after recombination, and the length of each channel of data is \(Q\). At this time, the length of the data stream generated by the acquisition window will also decrease from the original \(L_0\) to \(Q\). For the values of \(P\) and \(Q\), they are limited to powers of 2, and take \(P = 2\) p , \(Q = 2\) q . Among them, \(P\) and \(Q\) need to satisfy the following formula:

[0062] P=2 p ,Q=2 q

[0063] Wherein, p and q are both positive integers;

[0064] The relationship with the receiving window length L0 is:

[0065] L0=P×Q=2 p+q =2 m

[0066] Among them, p+q=m.

[0067] For example, the received data is expressed as x(0), x(1), x(2) ... x(L0-2), x(L0-1), and the reorganization method is as follows: Figure 2 shown.

[0068] The frequency domain response of the matched filter is reassembled in the same manner as the receiving window data reassembly. Before signal detection begins, the frequency domain response of the matched filter is stored in the FPGA.

[0069] Each receiving window needs to record the start time when collecting signals, and perform pulse compression on the received data. The whole process is divided into Fourier transform (FFT), frequency domain multiplication and inverse Fourier transform (IFFT). The flowchart of the above content is as follows Figure 3 Specifically including:

[0070] The parallel data obtained after processing is subjected to a Fourier transform of length Q. At this time, P paths are transformed simultaneously, and the output result of the parallel Q-point Fourier transform is P points each time. The output result is subjected to a P-point Fourier transform to complete the complete Fourier transform. The time-selected radix 2FFT transform is used to speed up the data processing rate.

[0071] Twitch factors used in data processing After transformation, the expression is as follows:

[0072]

[0073] In the formula, P is the number of points, The value of k needs to be changed to get k ′ , so that it complies with a continuous transformation of length Q.

[0074]

[0075] Where l = 0, 1, 2…Q-1;

[0076] After the complete Fourier transform is completed, the output data is multiplied with the frequency domain response of the frequency domain filter stored in advance. When the received data completes FFT, the signal spectrum is obtained, which is a parallel sequence X(ω), ω=0, 1, 2…L0. Frequency domain multiplication is required, and the multiplication output is a parallel sequence Y(n). The multiplication method is as follows Figure 4 shown.

[0077] The multiplied data is first subjected to an inverse Fourier transform of length P, using a frequency decimation radix 2FFT transform, and the rotation factors used in the transform are The data is then transformed in the same way as the forward Fourier transform of length P. Then, a Fourier transform of length Q is performed on the data, with P paths being transformed simultaneously using a frequency decimation radix 2 FFT transform.

[0078] Among them, the rotation factor The expression is:

[0079] In the formula, P is the number of points, The value of k needs to be changed to make it conform to the continuous transformation of length Q.

[0080]

[0081] Where l = 0, 1, 2…Q-1;

[0082] When the inverse Fourier transform of length P is completed, P parallel data channels are output and an IFFT of length Q is performed. When performing the IFFT of length Q, a frequency decimation radix 2-IFFT algorithm is used. At this time, the output data sequence y(n) is the data after pulse compression is completed.

[0083] After the pulse compression is completed, the results need to be tested in two aspects: one is the peak position within the receiving window, and the other is the amplitude signal-to-noise ratio within the receiving window.

[0084] Find the maximum value position from the result after pulse compression, that is, find the maximum value from L0 data and calculate its position L in the original receiving window max , and record the peak value y max ; Sum all the results to get y sum :

[0085]

[0086] Optionally, the calculating of the peak signal-to-noise ratio of the pulse compression signal includes:

[0087] Calculate all noise within the entire receiving window and remove the peak value;

[0088] The decrease in the signal-to-noise ratio within the window caused by the aliased signal is eliminated by adjusting the denominator when calculating the signal-to-noise ratio within the window.

[0089] Due to the existence of aliasing in the signal, the value used for noise averaging should be 2N, so as to eliminate the decrease in the signal-to-noise ratio within the window caused by the aliasing signal when calculating the signal-to-noise ratio.

[0090] y sum De-peak y max , and get the noise sum y noise_sum =y sum -y max .

[0091] You need to change y noise_sum Divide by 2N and calculate the average value of the noise result in the receiving window Calculating the signal-to-noise ratio

[0092] If L max >N, and at the same time, the SNR is greater than the set threshold, indicating that the complete signal is acquired and detected.

[0093] When a complete signal is detected, the signal arrival time can be calculated based on the peak position and the receiving window time. When the signal is detected continuously, the arrival time of the next pulse can be calculated based on these two times, and the calculation and correction are continuously made to achieve real-time prediction effect and receive at the predicted time.

[0094] When a signal is detected, the current time t1 is recorded, and when the next signal arrives, the second time t2 is recorded.

[0095] Calculate the pulse repetition period T between the two current signals pulse =t2-t1.

[0096] Predict the third arrival time t3 = T pulse +t2.

[0097] Because the repetition period of the signal will change in steps, the above process needs to be repeated continuously until the task is completed.

[0098] The present invention can be illustrated by the following simulation and measurement results.

[0099] For example, a signal with a frequency of 0±160 MHz, a bandwidth of 320 MHz, a pulse width of 50 μs, an amplitude of 1, and a pulse repetition period of 200 μs is detected. The sampling rate is 400 MHz, so the number of acquisition points for a complete pulse is 20,000. Combined with the above method, the length of the receiving window is 32,768, so real-time pulse compression is required for data with a length of 32,768. Simulation results of the pulse compression scheme of the present invention on the above signal are presented here, and two extreme cases are discussed.

[0100] Figure 5 The results shown are for signals located at the head of the receiving window. Graph a shows the imaginary part of the pulse compression result. It can be seen that aliasing occurs at the beginning of the data. Comparing the results of the real part, it can be seen that the peak position of the pulse compression is at 20,000.

[0101] Figure 6 The results shown show that the signal is located at the end of the receiving window. Figure a is the imaginary part of the pulse compression result. It can be seen that aliasing occurs in the middle part of the data. Comparing the results of the real part, it can be seen that the peak position of the pulse compression is at 32768.

[0102] In summary, the peak position of the pulse compression result falls in the interval [20000,32768], which meets the theoretical requirements.

[0103] Figure 7 For the data successfully predicted and received by the spectrometer, it can be seen from the pulse compression results that the abscissa of the pulse peak is greater than 20000, and the signal is complete within the receiving window, so the arrival time of the signal is successfully predicted.

[0104] like Figure 8 As shown, the transmitted signal is a step-by-step transmission, with each step signal varying in pulse repetition period and power. The collected signal successfully predicts and receives all signals. Therefore, the pulse compression and detection scheme described in this invention meets real-time requirements, successfully predicting and receiving every signal with good timeliness and accuracy. Therefore, the signal detection scheme proposed on this basis can achieve in-situ calibration of spectrometers.

[0105] In the embodiment of the present invention, the width of the receiving window is shortened, the processing data is shortened, and the processing flow is accelerated; by judging the position of the peak value of the processing result within the window, a constraint on the peak position is added. When the acquisition window is shortened, this constraint can ensure that the acquired signal is a complete pulse, so that the calculated peak value should be the maximum value of the pulse pressure theory; at the same time, because the peak position is constrained, the range of the signal used for judgment is narrowed, thereby reducing the false alarm range caused by noise; by reorganizing the signal data acquired by the receiving window, the serial data is converted into parallel data for processing, so that multiple channels of data are processed synchronously, and the processing speed is greatly improved.

[0106] Example 2

[0107] The present invention also provides a pulse linear frequency modulation signal detection system based on FPGA, comprising:

[0108] The receiving window setting module is used to continuously slide and collect the original target signal to be detected using the receiving window; conjugate the collected target signal to be detected to obtain the corresponding matched filter, fill the time domain length of the matched filter with zeros to the receiving window length, and then obtain the amplitude-frequency response H(n) of the matched filter through Fourier transform, and store it in the FPGA; wherein the receiving window length is L0, the sliding step length is L1, the signal length of the original target signal to be detected is N, N <L0≤2N,0<L1≤L0-N;

[0109] The data reassembly module is used to reassemble the collected target signal to be detected, converting the single-channel serial data into P-channel parallel data, with the length of each line of data being Q;

[0110] The signal compression module is used to perform Fourier transform on the parallel data obtained after processing, and perform point multiplication with the amplitude-frequency response of the matched filter to obtain the frequency domain result of pulse compression. The data after point multiplication is then inverse Fourier transformed to obtain the time domain result of pulse compression, thereby obtaining compressed pulse signal data.

[0111] a peak position detection module for detecting the peak position of the compressed pulse signal data; if the peak position is within the preset interval [L0-N, L0]] within the receiving window, it is determined that a complete pulse has been received, and the peak signal-to-noise ratio of the pulse compression signal is calculated; if the peak position is not within the preset interval within the receiving window, the data within the window is discarded;

[0112] The pulse signal detection module is used to determine whether the peak signal-to-noise ratio is greater than a threshold; if so, it is determined that the receiving window has detected the target signal, record the arrival time of the target signal each time it is detected, calculate the time interval between it and the last detected target signal, and predict the arrival time of the next signal; if not, it is determined that the receiving window has not detected the target signal.

[0113] An embodiment of the present invention provides an FPGA-based pulse linear frequency modulation signal detection system for executing an FPGA-based pulse linear frequency modulation signal detection method of embodiment 1, which has the same or similar beneficial effects and will not be described in detail here.

[0114] It will be easily understood by those skilled in the art that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A pulse linear frequency modulation signal detection method based on FPGA, characterized in that: include: S1, using the receiving window to continuously slide and collect the original target signal to be detected; The collected target signal to be detected is conjugate deconvoluted to obtain the corresponding matched filter, the time domain length of the matched filter is padded to the receiving window length, and then the amplitude-frequency response H(n) of the matched filter is obtained by Fourier transform and stored in the FPGA; wherein the receiving window length is L0, the sliding step length is L1, the signal length of the original target signal to be detected is N, N <L0≤2N,0<L1≤L0-N; S2. Reorganize the collected target signal to be detected, convert the single-channel serial data into P-channel parallel data, and the length of each line of data is Q; S3. Performing Fourier transform on the parallel data obtained after processing, and performing point multiplication with the amplitude-frequency response of the matched filter to obtain the frequency domain result of pulse compression, and then performing inverse Fourier transform on the frequency domain result to obtain the time domain result of pulse compression, thereby obtaining compressed pulse signal data; S4. Detecting the peak position of the compressed pulse signal data; if the peak position is within a preset interval [L0-N, L0] within the receiving window, determining that a complete pulse has been received, and calculating the peak signal-to-noise ratio of the pulse compression signal; if the peak position is not within the preset interval within the receiving window, discarding the data within the window; S5. Determine whether the peak signal-to-noise ratio is greater than a threshold; if so, determine that the receiving window has detected the target signal, record the arrival time of the target signal each time it is detected, calculate the time interval between it and the last detected target signal, and predict the arrival time of the next signal; if not, determine that the receiving window has not detected the target signal.

2. The signal detection method according to claim 1, wherein: The target signal to be detected is a pulse linear frequency modulation signal, and the expression is: Where A is the signal amplitude, K is the signal linear modulation frequency, T is the signal pulse width, and t is time.

3. The signal detection method according to claim 1, wherein: The expression of the amplitude-frequency response H(t) of the matched filter is: H(t)=FFT[x(t0-t)] Where t0 is the delay between the impulse response of the matched filter and the signal.

4. The signal detection method according to claim 1, wherein: The receiving window length L0 satisfies L0=2 m , m is a positive integer.

5. The signal detection method according to claim 4, wherein: P and Q need to satisfy the following formula: P=2 p ,Q=2 q Wherein, p and q are both positive integers; The relationship with the receiving window length L0 is: L0=P×Q=2 p+q =2 m Among them, p+q=m.

6. The signal detection method according to claim 1, wherein: When performing Fourier transform on data, an adaptive time decimation radix 2FFT algorithm is used; when performing inverse Fourier transform on data, an adaptive frequency decimation radix 2FFT algorithm is used.

7. The signal detection method according to claim 1, wherein: The calculation of the peak signal-to-noise ratio of the pulse compression signal includes: Calculate all noise within the entire receiving window and remove the peak value; The decrease in the signal-to-noise ratio within the window caused by the aliased signal is eliminated by adjusting the denominator when calculating the signal-to-noise ratio within the window.

8. A pulse linear frequency modulation signal detection system based on FPGA, characterized in that: include: The receiving window setting module is used to set the receiving window length of the acquisition signal to L0 and the receiving window sliding length to L1 according to the signal length N of the target signal to be detected x(n). The receiving window slides continuously to collect signals; conjugate deconvolution is performed on the collected target signal to be detected to obtain a matched filter, the time domain length of the matched filter is padded with zeros to the receiving window length, and then the amplitude-frequency response H(n) of the matched filter is obtained by Fourier transform and stored in the FPGA; wherein, N <L0≤2N,0<L1≤L0-N; The data reassembly module is used to reassemble the collected target signal to be detected, converting the single-channel serial data into P-channel parallel data, with the length of each line of data being Q; The signal compression module is used to perform Fourier transform on the parallel data obtained after processing, and perform point multiplication with the amplitude-frequency response of the matched filter to obtain the frequency domain result of pulse compression, and then perform inverse Fourier transform on the frequency domain result to obtain the time domain result of pulse compression, thereby obtaining compressed pulse signal data; a peak position detection module, configured to detect the peak position of the compressed pulse signal data; if the peak position is within a preset interval [L0-N, L0] within a receiving window, it is determined that a complete pulse has been received, and the peak signal-to-noise ratio of the pulse compression signal is calculated; if the peak position is not within the preset interval within the receiving window, the data within the window is discarded; The pulse signal detection module is used to determine whether the peak signal-to-noise ratio is greater than a threshold; if so, it is determined that the receiving window has detected the target signal, record the arrival time of the target signal each time it is detected, calculate the time interval between it and the last detected target signal, and predict the arrival time of the next signal; if not, it is determined that the receiving window has not detected the target signal.

Images