Radar pulse signal detection method and apparatus based on spectral sparsity perception

Abstract

This invention relates to a radar pulse signal detection method and apparatus based on sparse spectrum sensing. The radar pulse signal detection apparatus based on sparse spectrum sensing includes a sparse spectrum sensing module, a dynamically adjusted adaptive filtering module, and a dual-threshold pulse detection module. The sparse spectrum sensing module performs single-bit sampling on the original signal to obtain the signal frequency and bandwidth. The dynamically adjusted adaptive filtering module filters the signal according to the signal frequency and bandwidth. Utilizing the sign bit of the AD data for spectral sensing, the system monitors and tracks the input spectral distribution in real time. Compared to the traditional short-time Fourier transform, this significantly reduces the resources required for signal processing and increases the processing bandwidth. During filtering, based on the pulse signal center frequency and bandwidth obtained from the spectral monitoring module, the dynamically adjusted adaptive filtering module adjusts the filter's center frequency to achieve tracking and filtering of the radar pulse signal.

Description

Technical Field

[0001] This invention relates to the field of radar pulse signal detection and interception technology in electronic reconnaissance, specifically to a radar pulse signal detection method and apparatus based on spectral sparse sensing. Background Technology

[0002] A large number of radar radiation sources exist in the 30MHz–18GHz range, each with different frequencies, bandwidths, and signal patterns, creating a highly complex electromagnetic environment in modern society. Furthermore, the anti-interception capabilities of electromagnetic signals from various military electromagnetic devices used on the battlefield are constantly improving, placing higher demands on the system performance of reconnaissance receivers, such as wider monitoring bands, greater dynamic range, and higher sensitivity. The increasing demand for commercial spectrum, the sharing and congestion of spectrum resources, and other factors further complicate the electromagnetic spectrum environment faced by radar reconnaissance receivers. Wideband open-channel receivers possess high probability of intercept, high sensitivity, high frequency resolution, and adaptability to simultaneously arriving signals, effectively addressing the high-density signal environment of modern times. In recent years, the sampling frequency of analog-to-digital converters (ADCs) has increased rapidly, with the highest sampling rate reaching over 10GHz. While high-speed ADCs significantly expand the instantaneous bandwidth of digital receivers, they also pose greater challenges to digital storage and processing capabilities. Channelization processing in digital receivers is primarily implemented by programmable logic array (FPGA) chips and dedicated programs. At a sampling rate of 10 GHz, the current capacity and speed of FPGAs make it difficult to implement traditional digital channelization processing. Therefore, how to further reduce the resource consumption and power consumption of FPGAs, and quickly and efficiently detect and receive complex radar signals with low signal-to-noise ratios, is a challenging problem.

[0003] Chinese Patent CN113447893B discloses an automatic radar pulse signal spectrum detection method, system, and medium. The method includes: digitally channelizing an intermediate frequency (IF) complex signal to obtain sub-channel signals; performing adaptive threshold signal detection to obtain the effective signals of the sub-channels; performing cross-channel decision-making on each pair of adjacent sub-channels to determine the cross-channel status of adjacent sub-channels; digitally channelizing sub-channels without cross-channel status and performing adaptive threshold signal detection on each sub-channel, then outputting the effective signals of the next-level sub-channels; and calculating the center frequency and bandwidth of the corresponding cross-channel broadband signal for sub-channels with cross-channel status, configuring a corresponding DDS signal generator and variable bandwidth filter to perform signal matching and detection on the IF complex signal. This invention employs two-stage channelization detection, achieving automatic noise threshold updates and adding processing for broadband cross-channel signals, thus enabling adaptive detection of broadband radar signal bandwidth. However, when using the above-mentioned prior art for detection, the problems of high FPGA resource consumption and high processing power consumption still exist, making it difficult to quickly and efficiently detect low signal-to-noise ratio complex radar signals. Summary of the Invention

[0004] The purpose of this invention is to provide a radar pulse signal detection method based on spectral sparse sensing to solve the technical problems of high resource consumption and high power consumption of FPGA in the prior art, and the difficulty in quickly and efficiently detecting radar signals with low signal-to-noise ratio and complex system. The purpose of this invention is also to provide a radar pulse signal detection device based on spectral sparse sensing.

[0005] To achieve the above objectives, the radar pulse signal detection device based on spectral sparse sensing of the present invention adopts the following technical solution:

[0006] A radar pulse signal detection device based on sparse spectrum sensing includes a sparse spectrum sensing module, which includes a single-bit analog-to-digital converter and a processor. The single-bit analog-to-digital converter is used to perform single-bit sampling on the original signal to obtain a discrete sample set, and the processor is used to process the obtained discrete sample set to obtain the signal frequency and signal bandwidth.

[0007] The dynamic adjustment adaptive filtering module is electrically connected to the spectrum sparse sensing module to receive the discrete sample set, signal frequency and signal bandwidth transmitted by it. Its function is to down-convert the discrete sample set according to the signal frequency and to perform low-pass filtering on the down-converted signal according to the signal bandwidth.

[0008] The dual-threshold pulse detection module is electrically connected to the dynamically adjusted adaptive filter module to receive the filtered signal transmitted by it, and is used to detect the filtered signal.

[0009] Furthermore, it also includes a signal acquisition module for acquiring raw full-band signals, which is electrically connected to the spectrum sparse sensing module to transmit the acquired raw full-band signals to it.

[0010] Furthermore, it also includes a signal output module, which is electrically connected to the dual-threshold pulse detection module to receive the detection results transmitted by it and output the detection results.

[0011] The radar pulse signal detection method based on spectral sparse sensing of the present invention adopts the following technical solution:

[0012] A radar pulse signal detection device based on sparse spectrum sensing includes the following steps: First, acquiring the original full-band signal;

[0013] The second step is to perform single-bit sampling on the original full-band signal to obtain a discrete sample set, and process the discrete sample set to obtain the signal frequency and signal bandwidth.

[0014] The third step is to downconvert the discrete sample set to obtain the zero intermediate frequency signal, and then filter the zero intermediate frequency signal according to the obtained signal bandwidth.

[0015] The fourth step is to perform dual-threshold pulse detection on the filtered signal.

[0016] Furthermore, in the third step, the discrete sample set is down-converted using a mixer, and the center frequency of the mixer is adjusted in real time according to the acquired signal frequency.

[0017] Furthermore, in the second step, a DFT transformation is performed on the acquired discrete sample set to obtain its signal frequency and signal bandwidth.

[0018] Furthermore, in the second step, the discrete sample set is divided into time slices and DFT transformation is performed on each slice. Then, the DFT data of multiple time slices within the pulse duration are merged and spectral peak analysis and spectral parameter extraction are performed. Based on this, the signal frequency and signal bandwidth are calculated. When dividing the discrete sample set into time slices, it should be ensured that there is overlap between adjacent time slices.

[0019] Furthermore, the signal after the third filtering step is In the formula, x(n) is the input data at the nth sampling point, M is the order of the filter, and e -jωkn It is a mixer, k is the center frequency of the pulse signal, k is a real number, the frequency k of the mixer is the same as the center frequency k of the pulse signal, W(m) is the prototype low-pass filter, y k (n) is the filtered pulse signal.

[0020] The beneficial effects of this invention are as follows: using a sparse spectrum sensing module to perform broadband sparse real-time sensing of the input radar signal, and using the sign bits of the AD data for spectrum sensing, the spectral distribution of the input can be monitored and tracked in real time. Compared with the traditional short-time Fourier transform, this can greatly reduce the resources required for signal processing and increase the processing bandwidth of the signal processing.

[0021] Furthermore, during filtering, based on the center frequency and bandwidth of the pulse signal obtained from the spectrum monitoring module, a dynamically adjusted adaptive filtering module is used to adjust the center frequency of the filter, achieving tracking filtering of the radar pulse signal. A dual-threshold pulse detection module is used to detect the radar pulse signal, improving signal detection sensitivity and ensuring the detection of complete pulse signals. Since the mixing frequency of the mixer module is adjusted in real time for broadband signals, a single pulse may have multiple mixing frequencies. This approach, compared to a single mixing frequency, can further improve the pulse signal-to-noise ratio and increase detection accuracy.

[0022] Experiments have shown that this invention can simultaneously detect 200M broadband linear frequency modulated radar pulse signals with a signal-to-noise ratio of -10dB and ordinary radar pulse signals. Attached Figure Description

[0023] Figure 1 This is a structural block diagram of a radar pulse signal detection device based on spectral sparse sensing according to the present invention.

[0024] Figure 2 This is a flowchart of a radar pulse signal detection method based on spectral sparse sensing according to the present invention.

[0025] Figure 3 It is the DFT output spectrum of a single-tone signal.

[0026] Figure 4 It is the single-bit DFT output spectrum of a single-tone signal.

[0027] Figure 5 It is the DFT output spectrum of the two-tone signal.

[0028] Figure 6 It is the single-bit DFT output spectrum of a two-tone signal.

[0029] Figure 7 It is a time-domain diagram of the superposition of a linear frequency modulated signal and a conventional signal arriving simultaneously.

[0030] Figure 8 It is a time-frequency diagram of the superposition of a linear frequency modulated signal and a conventional signal arriving simultaneously.

[0031] Figure 9 It is a time-domain graph of a conventional signal after dynamic adjustment and adaptive filtering.

[0032] Figure 10 It is the time-domain graph of a linear frequency modulated signal after dynamic adjustment and adaptive filtering.

[0033] Figure 11 This is a schematic diagram of dual-threshold pulse detection. Detailed Implementation

[0034] An embodiment of the radar pulse signal detection device based on spectral sparse sensing according to the present invention:

[0035] The specific structure of a radar pulse signal detection device based on spectral sparse sensing is as follows: Figure 1 As shown, it includes a signal acquisition module, a spectrum sparse sensing module, a dynamic adjustment adaptive filtering module, a dual-threshold pulse detection module, and an output module.

[0036] The signal acquisition module and the sparse spectrum sensing module are electrically connected to transmit the acquired raw signal. The sparse spectrum sensing module includes a single-bit analog-to-digital converter (ADC) and a processor. The ADC performs single-bit sampling of the raw signal, and the processor processes the discrete sample set obtained from the single-bit sampling to obtain the signal frequency and bandwidth. The sparse spectrum sensing module is electrically connected to the dynamically adjusted adaptive filtering module, transmitting the discrete sample set obtained from the single-bit sampling, as well as the acquired signal frequency and bandwidth. The dynamically adjusted adaptive filtering module down-converts the discrete sample set obtained from the single-bit sampling based on the acquired signal center frequency and performs low-pass filtering on the zero-IF signal obtained after down-conversion based on the signal bandwidth. The dynamically adjusted adaptive filtering module is electrically connected to the dual-threshold pulse detection module, which detects the filtered signal. The dual-threshold pulse detection module is electrically connected to the output module to transmit the detection results, and the output module outputs the received detection results.

[0037] An embodiment of the radar pulse signal detection method based on spectral sparse sensing according to the present invention:

[0038] The specific process of a radar pulse signal detection method based on spectral sparse sensing is as follows: Figure 2 As shown, it includes the following steps:

[0039] The first step is to collect radar signals;

[0040] The collected radar signal is a complete raw full-band signal, which includes noise signals and may also include radar signals.

[0041] The second step is to calculate the signal spectrum using the sparse spectrum sensing module to obtain the signal frequency and signal bandwidth.

[0042] In this step, a single-bit analog-to-digital converter (ADC) is first used to convert the signal input from the signal acquisition module to digital. Then, the signal, in the form of a discrete sample set after conversion, is divided into time slices, and the Discrete Fourier Transform (DFT) is calculated separately for each slice. There is some overlap between adjacent time slices to ensure that the DFT processing gain of the pulse signal across time slices is not reduced. After merging the DFT data from multiple time slices within the pulse duration, spectral peak analysis and spectral parameter extraction are performed, and the signal frequency and bandwidth are calculated. Calculating the DFT separately for each time slice of the signal in the form of a discrete sample set after conversion reduces signal processing resources and improves processing efficiency.

[0043] The DFT transform formula for a single-bit sampled signal based on the ADC sign bit is:

[0044]

[0045] In the formula, x(n) is the input data of the nth sampling point, N is the number of points of the DFT, and X(ω) is the output data of the DFT. X(ω) actually reflects the spectrum data of the signal.

[0046] In single-bit sampling mode, x(n) only takes two values: 0 and 1. Therefore, multiplication is not required when calculating X(ω). The DFT algorithm can be implemented simply by accumulating the frequency rotation factor e-iωm using an accumulator with addition and subtraction control. Compared to the traditional short-time DFT, single-bit sampling based on the ADC sign bit can significantly reduce signal processing resources and increase signal processing bandwidth. The implementation structure of the DFT algorithm in this invention is as follows:

[0047]

[0048] The DFT output result of a single-tone signal (i.e., a single-frequency radio frequency signal) under a 3dB signal-to-noise ratio condition is as follows: Figure 3 As shown, the single-bit DFT output result under a 3dB signal-to-noise ratio condition for a single-tone signal is as follows: Figure 4 As shown. Comparison Figure 3 and Figure 4 It can be seen that the single-bit DFT of the single-tone signal is very close to the DFT of the original signal, but the signal-to-noise ratio of the single-bit DFT is slightly lower.

[0049] The DFT output result of a dual-tone signal (i.e., a dual-frequency radio frequency signal) under a 3dB signal-to-noise ratio condition is as follows: Figure 5 As shown, the single-bit DFT output result of the two-tone signal under a 3dB signal-to-noise ratio condition is as follows: Figure 6 As shown. Comparison Figure 5 and Figure 6It can be seen that some spurious spectral components appear in the single-bit DFT calculation result compared with the original signal. Therefore, the dynamic range of the single-bit spectrum of the signal will be lost under multi-tone conditions.

[0050] The third step is to dynamically adjust the bandwidth of the adaptive filtering module and implement filtering based on the signal frequency and bandwidth obtained in the second step.

[0051] In this step, the input signal is down-converted using the signal center frequency obtained from spectral sparse sensing, and the zero-IF signal is low-pass filtered based on the signal bandwidth. Specifically:

[0052]

[0053] The filtered signal is:

[0054]

[0055] In the formula, x(n) has the meaning as above, representing the input data at the nth sampling point; M is the order of the filter, typically ranging from [256, 384]; e-jωkn is the mixer; k is the center frequency of the pulse signal, a real number; the mixer frequency k is consistent with the center frequency k of the pulse signal; W(m) is the prototype low-pass filter; y k (n) is the filtered pulse signal. The mixing frequency e -jωkn The values ​​of both the prototype low-pass filter w(n) and the signal frequency and bandwidth obtained from the spectral sparse sensing module are dynamically adjusted and adaptively filtered.

[0056] When dynamically adjusting adaptive filtering to follow changes in frequency parameters, the mixer's center frequency should be updated in real time based on the signal frequency obtained from the spectral sparsity sensing module to maintain waveform output continuity. Whether the mixer's center frequency is updated is determined by the amount of signal frequency change obtained from the spectral sparsity sensing module; generally, the mixer's center frequency is updated when the signal frequency change exceeds 20 MHz.

[0057] To verify the effectiveness of this invention in separating signals through adaptive filtering after spectral sparsity sensing, the following example is provided:

[0058] A set of signals is generated, which is a synthesis of simultaneously arriving linear frequency modulated radar signals and conventional radar signals. Its frequency domain diagram is shown below. Figure 7 As shown, the time-frequency graph is as follows: Figure 8 As shown. By Figure 7 and Figure 8As can be seen, due to the wide bandwidth of linear frequency modulated signals, their spectra tend to overlap with those of conventional signals. However, in the time-frequency plane, the two signal components do not overlap or only partially overlap, so they can be separated by dynamically adjusting an adaptive filter.

[0059] The above signals are fed into a sparse spectrum sensing module. The module detects the spectral components of the two signals and then uses two adaptive filters for separation. For conventional signals, the center frequency of the adaptive filters remains essentially constant. For linear frequency modulated (LFM) signals, the frequency measured by the sparse spectrum sensing module continuously changes, and the mixing frequency of the adaptive filters is dynamically adjusted accordingly, achieving time-varying tracking filtering of the LFM signal. After adaptive filtering, the time-domain signals of the two signal components are obtained as follows: Figure 9 , Figure 10 As shown. By Figure 9 and Figure 10 As can be seen, after down-conversion, the signal becomes a baseband signal; after adaptive filtering, the signal-to-noise ratio is significantly improved.

[0060] The fourth step involves performing dual-threshold pulse detection on the output signal after adaptive filtering in the third step.

[0061] In this step, a dual-threshold pulse detection module is used to set two thresholds to ensure the integrity of pulse detection. The detection objective is to retain signals within the set thresholds as detection values. The specific detection principle is as follows:

[0062] The pulse start and end thresholds are:

[0063] The pulse duration threshold is:

[0064] in the formula This represents the noise mean of n sample points (for each time slice, the sample points are sorted according to their amplitude; the median of the sorted channels is the noise in the corresponding channel, and the average of the noise in all channels is obtained). This represents the noise variance of n sample points, where a is a pre-defined constant (the value of a is generally determined by the false alarm rate, and the default value is set to 4).

[0065] A schematic diagram illustrating the principle of dual-threshold pulse detection is shown below. Figure 11As shown, the upper line represents the pulse start and end thresholds, the lower line represents the pulse duration threshold, the middle line represents the threshold average, and the signal between the pulse start and end thresholds and the pulse duration threshold (i.e., the signal set in the middle) is the signal retained after detection. Using dual-threshold pulse detection can more accurately measure the bandwidth, frequency, and pulse width of the pulse signal, greatly improving the sensitivity of pulse signal detection.

[0066] Fifth step, output the detection results.

[0067] Although the present invention has been described in detail above with general descriptions and specific embodiments, modifications or improvements can be made to it, which will be obvious to those skilled in the art. Therefore, all such modifications or improvements made without departing from the spirit of the present invention fall within the scope of protection claimed by the present invention.

Claims

1. A radar pulse signal detection device based on spectral sparse sensing, characterized in that: The system includes a sparse spectrum sensing module, which comprises a single-bit analog-to-digital converter (ADC) and a processor. The ADC performs single-bit sampling on the original full-band signal to obtain a discrete sample set. The processor processes the obtained discrete sample set to obtain the signal frequency and bandwidth. The digital signal in the discrete sample set takes only two values: 0 and 1. The discrete sample set processing includes: dividing the discrete sample set into time slices and performing DFT transformation on each slice; then merging the DFT data from multiple time slices within the pulse duration and performing spectral peak analysis and spectral parameter extraction; and calculating the signal frequency and bandwidth accordingly. When dividing the discrete sample set into time slices, it is necessary to ensure that there is overlap between adjacent time slices. The dynamic adjustment adaptive filtering module is electrically connected to the spectrum sparse sensing module to receive the discrete sample set, signal frequency and signal bandwidth transmitted by it. Its function is to down-convert the discrete sample set according to the signal frequency and to perform low-pass filtering on the down-converted signal according to the signal bandwidth. The dual-threshold pulse detection module is electrically connected to the dynamically adjusted adaptive filter module to receive the filtered signal transmitted by it, and is used to detect the filtered signal.

2. The radar pulse signal detection device based on spectral sparse sensing according to claim 1, characterized in that: It also includes a signal acquisition module for acquiring raw full-band signals, which is electrically connected to the spectrum sparse sensing module to transmit the acquired raw full-band signals to it.

3. The radar pulse signal detection device based on spectral sparse sensing according to claim 1, characterized in that: It also includes a signal output module, which is electrically connected to the dual-threshold pulse detection module to receive the detection results transmitted by it and output the detection results.

4. A radar pulse signal detection method based on spectral sparse sensing, characterized in that: The process includes the following steps: First, acquire the raw full-band signal; The second step is to perform single-bit sampling on the original full-band signal to obtain a discrete sample set, and process the discrete sample set to obtain the signal frequency and signal bandwidth. The digital signal in the discrete sample set takes only two values: 0 and 1. The discrete sample set processing includes: dividing the discrete sample set into time slices and performing DFT transformation on each slice; then merging the DFT data of multiple time slices within the pulse duration and performing spectral peak analysis and spectral parameter extraction; and calculating the signal frequency and signal bandwidth accordingly. When dividing the discrete sample set into time slices, it should be ensured that there is overlap between adjacent time slices. The third step is to downconvert the discrete sample set to obtain the zero intermediate frequency signal, and then filter the zero intermediate frequency signal according to the obtained signal bandwidth. The fourth step is to perform dual-threshold pulse detection on the filtered signal.

5. The radar pulse signal detection method based on spectral sparse sensing according to claim 4, characterized in that: In the third step, the discrete sample set is down-converted using a mixer, and the center frequency of the mixer is adjusted in real time according to the acquired signal frequency.

6. The radar pulse signal detection method based on spectral sparse sensing according to claim 5, characterized in that: The signal after the third filtering step is In the formula, x(n) is the input data at the nth sampling point, M is the order of the filter, and e - jωkn is a mixer, k is the center frequency of the pulse signal, k is a real number, the frequency k of the mixer is the same as the center frequency k of the pulse signal, w(m) is a prototype low-pass filter, y k (n) is the filtered pulse signal.

Images