CN114184848B - Point-by-point scanning real-time processing method of spaceborne VHF transient signals based on Goertzel algorithm - Google Patents

Abstract

The invention relates to the technical field of VHF transient signal detection, in particular to a satellite-borne VHF transient signal point-by-point scanning real-time processing method based on a Goertzel algorithm, which comprises the following steps of: the method comprises the steps of (1) performing broadband sampling on VHF transient signals to obtain original time domain waveforms of a detection frequency band, (2) calculating Fourier components of a plurality of frequency points in real time by utilizing a Goertzel algorithm in a point-by-point sliding window, and (3) judging the signals according to judging conditions; in the method, fourier components of 20 frequency points are calculated in real time in two detection frequency bands, a threshold value passing duration time and the like are used as discrimination conditions, when the frequency points meeting the discrimination conditions reach M, and the method has the characteristics of high frequency first arrival, the signals can be considered to be effective, the complexity of a hardware circuit is reduced, the anti-interference capability of the circuit is enhanced, and the method also has the advantages of real time, high efficiency, adjustable parameters and the like, in addition, the acquisition of the waveform of an original time domain signal can be completed, and information is provided for ionosphere propagation group delay and waveform characteristic analysis.

Description

Point-by-point scanning real-time processing of spaceborne VHF transient signals based on Goertzel algorithm method

Technical field

The invention relates to the technical field of VHF transient signal detection, and in particular to a point-by-point scanning real-time processing method of spaceborne VHF transient signals based on the Goertzel algorithm.

Background technique

The satellite platform is used to detect VHF transient signals, which is mainly used to capture natural or artificially generated electromagnetic radiation signals above the earth's surface. At present, existing signal processing methods at home and abroad mainly use analog technology, which divides the detected time domain signal into multiple channels through multiple analog filters, and then uses mixing down-conversion technology to obtain the intermediate frequency signals of multiple channels, and then passes through After the data is collected, it is sent to the digital processing unit for judgment. For example, the electromagnetic radiation detector carried by my country's XX satellite 3 is designed to use 5 analog channels (choose 3 from 5 as the trigger condition), and the VHF band electromagnetic radiation detector carried by the US FORTE satellite is designed to use 8 narrowband analog channels. (Select 5 out of 8 as the trigger condition). Its shortcomings are as follows:

1) The amount of information collected is small, and there is no original waveform of the time domain signal, which provides insufficient information for the analysis of ionospheric propagation group delay and waveform characteristics;

2) Non-real-time data processing, there is the possibility of loss events;

3) It is mostly implemented by frequency mixing and down-conversion, which is susceptible to interference in complex electromagnetic environments.

In summary, the development of a point-by-point scanning real-time processing method for spaceborne VHF transient signals based on the Goertzel algorithm is still a key issue that urgently needs to be solved in the field of VHF transient signal detection technology.

Contents of the invention

In view of the above shortcomings of the existing technology, the present invention is to provide a point-by-point scanning real-time processing method of spaceborne VHF transient signals based on the Goertzel algorithm. The present invention is based on the Goertzel algorithm and uses digital multi-channel short-time Fourier transform combined with point-by-point sliding windows. Scan mode, fix the calculation window width, select the step size as one sampling point, perform point-by-point multi-channel narrowband channel spectrum analysis on the received signal, and realize the capture of VHF transient events passing through the ionosphere. The Goertzel algorithm can be used in the reception The signal is calculated at the same time without waiting for all the data involved in the calculation to be ready. It has good real-time performance. The use of sliding window recursive calculation can avoid repeated operations and further speed up the calculation.

In order to achieve the purpose of the present invention, the present invention provides the following technical solutions:

Point-by-point scanning real-time processing method of spaceborne VHF transient signals based on Goertzel algorithm includes the following steps:

(1) Perform broadband sampling of the VHF transient signal to obtain the original time domain waveform of the detection frequency band;

(2) Use the Goertzel algorithm to calculate the Fourier components of multiple frequency points in real time using a point-by-point sliding window;

(3) Discriminate the signal according to the discrimination conditions.

The present invention is further configured as follows: in step (1), two band-pass filters are used to filter out signals in two detection frequency bands.

The present invention is further configured such that the two detection frequency bands are low frequency band 25MHz-75MHz and high frequency band 110MHz-150MHz respectively.

The present invention is further configured as follows: in step (1), select the sampling frequency fS=160MHz, and use two A/D chips to simultaneously convert signals in two frequency bands, directly sample the low frequency band to obtain a digital signal of 25MHz-75MHz, and obtain a digital signal of 25MHz-75MHz for the high frequency band. The frequency band is undersampled to obtain a digital signal of 10MHz-50MHz.

The present invention is further configured as follows: in step (1), the sampling duration is selected to be 256us.

The present invention is further configured as follows: in step (2), the following steps are included:

1) Set the number of channels to 20;

2) Select the calculation window width N = 101, 101 time domain signal sampling points to obtain a Fourier component frequency point amplitude, and calculate a Fourier component frequency point amplitude for each moving sampling point;

3) Select the Fourier component frequency;

4) Optimize the iterative formula of Goertzel algorithm;

5) Point-by-point sliding window calculates the Fourier components of multiple frequency points in real time.

The present invention is further set as follows: the iterative formula of the optimized Goertzel algorithm is: In the formula, N is the window width, and X(N) is the sampling value.

The present invention is further configured as follows: in step (3), the threshold and the duration of crossing the threshold are used as the discrimination conditions. When the number of frequency points that meet the discrimination conditions reaches M and has the characteristic of high frequency arriving first, the signal is considered valid.

beneficial effects

The technical solution provided by the present invention has the following beneficial effects compared with the known public technology:

This invention calculates the Fourier components of 20 frequency points in real time in two detection frequency bands of 25MHz-75MHz and 110MHz-150MHz, and uses thresholds, threshold-passing duration, etc. as discrimination conditions. When the number of frequency points that meet the discrimination conditions reaches M (M≤20, can be set), and has the characteristics of high frequency arriving first, so the signal can be considered effective. This method reduces the complexity of the hardware circuit, enhances the anti-interference ability of the circuit, and also has the advantages of real-time, high efficiency, and adjustable parameters. , In addition, it can also complete the collection of original time domain signal waveforms, providing information for ionospheric propagation group delay and waveform characteristic analysis.

Description of the drawings

Figure 1 is a diagram of the point-by-point sliding window scanning detection results in Embodiment 1;

Figure 2 is a schematic diagram of signal triggering judgment in Embodiment 1;

Figure 3 is a flow chart of signal trigger determination in Embodiment 1.

Detailed ways

In order to make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be described clearly and completely below. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without making creative efforts fall within the scope of protection of the present invention.

In order to achieve the effects of the present invention, the present invention will be further described below in conjunction with examples.

Example 1:

Referring to Figure 1-3, the point-by-point scanning real-time processing method of spaceborne VHF transient signals based on the Goertzel algorithm includes the following steps:

(1) Perform broadband sampling on the VHF transient signal to obtain the original time domain waveform of the detection frequency band.

In this embodiment, AD9642 (14bits) is selected for sampling.

1) Select the sampling frequency band: The target VHF transient signal frequency is mainly distributed in tens of kHz to hundreds of MHz. Due to the ionospheric propagation effect, signals greater than 20MHz can pass through the ionosphere. Affected by the ionospheric TEC, the group delays of electromagnetic waves with different frequencies propagating in the ionosphere are significantly different. The lower the frequency, the greater the delay.

In order to minimize the influence of ground radio stations or other natural signals, the detection frequency band is set as the low-frequency band 25MHz-75MHz and the high-frequency band 110MHz-150MHz. Before sampling, two band-pass filters are used to filter out the signals in the above two frequency bands.

2) Select the sampling frequency fS: For the detection frequency band, the Nyquist Frequency should be selected above the highest detection frequency in the detection frequency band or between the high frequency band and the low frequency band in order to cover the detection frequency band; at the same time, the Nyquist frequency It is necessary to try to avoid selecting fN at the edge of the detection frequency band to reduce clutter interference, so the sampling frequency fS=160MHz is selected. Use two A/D chips to simultaneously convert signals in two frequency bands, directly sample the low-frequency band to obtain a 25MHz-75MHz digital signal, and under-sample the high-frequency band to obtain a 10MH-50MHz digital signal.

3) Select the sampling duration of an event: According to the characteristics of the ionospheric group delay, when the TEC is 1×10 ¹⁷ m ^-2 , the delay difference between the electromagnetic waves of 30MHz and 31MHz is 1.62×10 ³ ns, and the time delay of the electromagnetic waves of 150MHz and 151MHz is 1.62×10 3 ns. The delay difference is 7.9ns, and the total delay difference between 30MHz and 150MHz electromagnetic waves is 1.43×10 ⁴ ns. Considering that the ionospheric concentration changes greatly, the sampling duration of one event in the present invention is 256 us, that is, 40960 sampling points.

(2) Using the Goertzel algorithm, the point-by-point sliding window is used to calculate the Fourier components of multiple frequency points in real time.

In this embodiment, the following steps are included:

1) Set the number of channels: In the past, the number of channels set by the satellite load was 5 channels (such as my country's XX Satellite 3) and 8 channels (such as the FORTE satellite of the United States). Taking into account the reliability of rapid identification of target signals, the present invention The number of channels in is set to 20, that is, the Fourier components of 20 frequency points need to be calculated in real time when processing the data acquisition signal.

2) Window width N: The window width N is to allocate N Fourier transform frequency points within the acquisition frequency range (0-π), and perform N iterative operations to obtain the Fourier transform spectrum within the window width range. Fourier transform frequency point f _k =kf _s /N, frequency point interval f _s /N, where f _S is the sampling frequency of the system, N is the window width, and f _S /N is the frequency resolution. The larger N is selected, the higher the spectrum resolution is and the greater the amount of calculation required. Therefore, N is related to the calculation cumulative error and system resources. For real-time judgment of signal properties, it is necessary to ensure considerable isolation and appropriate frequency bandwidth between selected frequency points to avoid frequency interference and detection reliability between selected frequency points; it is also necessary to ensure calculation accuracy and effective use of resources. . After weighing, the calculation window width N=101 is selected, that is, 101 time domain signal sampling points are used to obtain a Fourier component frequency point amplitude, and a Fourier component frequency point amplitude is calculated for each moving sampling point.

3) Select the Fourier component frequency: the sampling frequency f _S is 160MHz, so the Fourier component frequency resolution is 1.58416MHz; in this embodiment, the Fourier component frequency interval is selected to be 3.16832MHz, and 2 groups of Fourier components can be artificially set Leaf weight. In this embodiment, the Fourier component frequencies are selected as shown in Table 1.

Table 1 Fourier component frequency table

4) Optimize the iterative formula of Goertzel algorithm:

According to the Goertzel recursive algorithm iteration formula:

V _k (n) = 2cos (2πk/N) V _k (n-1)-V _k (n-2) + X (n), each time a sample value x (n) is input, iterative calculation requires a real number Multiplication and two real number additions, calculating V _k (N) requires N times of calculation.

When the sampling rate is low, iterative calculations can be performed in the sampling gap. However, when facing high-speed sampling, the sampling gap is not enough to complete iterative calculations of this scale, so the iterative formula of the Goertzel algorithm needs to be optimized.

After derivation, the above iteration formula can be transformed into:

_In the formula, N is the _{window width} , and X(n)(0≤n≤N+1), directly calculate V _k (N+1), and substitute N=101 into the above converted formula, we can get:

In the formula, for each given k value, w _k ,/> is a constant.

5) Calculate the Fourier components of multiple frequency points in real time using a point-by-point sliding window:

In this embodiment, the Zynq-7000 series chip XC7z045ffg900-2 is used. In order to implement pipeline operations, 4 multipliers and 5 adders (subtractors) are required, of which 2 multipliers and 2 adders are implemented with 2 DSP48 multipliers.

The calculation time of the lookup table multiplier in FPGA is 4 times the operation speed (4clk), the calculation time of the fabric adder (subtractor) is 2clk, and the calculation time of the two multipliers and adders of DSP48s is 1clk for the multiplier and 1clk for the adder. 3clk, multiplier 1clk, and adder 2clk.

In the case of FPGA computing speed fC=2fS=320MHz, the pipeline method can be used to complete the equation Iterative calculation (the calculation result is only delayed by a dozen clocks relative to the input data). Make full use of the parallel processing advantages of FPGA to complete frequency domain analysis of multiple frequency points within one sampling clock cycle, that is, each sampling clock simultaneously generates a set of 20-channel narrow-band multi-frequency frequency domain results to achieve real-time capture of transient signals , where the point-by-point sliding window scanning detection results are shown in Figure 1

(3) Discriminate the signal according to the discrimination conditions.

The threshold and the duration of crossing the threshold are used as the discrimination conditions. When the number of frequency points that meet the discrimination conditions reaches M (M ≤ 20, can be set), and has the characteristic of high frequency arriving first, the signal is considered valid.

In this embodiment, according to the analysis of the data received by the XX Satellite No. 3 electromagnetic radiation detector, some specific areas often receive a certain single-frequency signal (such as radar, etc.). This signal lasts for a long time and is consistent with the target signal. A big difference. In order to filter out such signals, the energy peak sustaining time judgment is added to the electromagnetic pulse receiver integrated processing unit. If the narrowband signal continuously exceeds the threshold value t _P , the signal will not be accepted, thus filtering out most non-target signals. At the same time, the target source signal is a pulse signal. After being scattered by the ionosphere, its spectrum will cover different detection frequency bands. Judging from 20 narrow-band channels, within a certain time window t _W , only M (≤20, configurable) channels can satisfy The acquisition is carried out only when the conditions are met to reduce the impact of the single-frequency pulse signal on the load detection signal. The signal trigger judgment diagram is shown in Figure 2.

The signal triggering process is as follows: write the digital threshold in the latch, and compare the digital acquisition signal with the set threshold. When the acquisition signal is greater than the set threshold, an over-threshold signal is generated; when the signal passes the peak maintenance time, it enters 20 Select the M module for judgment. When the threshold signal ≥ M is exceeded within the time window t _W , the event waveform is stored and the event trigger moment is latched. The signal trigger judgment process is shown in Figure 3.

Example 2:

The method of the present invention is compared with the existing technology, and the comparison data are shown in Table 2.

Table 2 Comparison data table

As can be seen from Table 2, the present invention reduces the complexity of the hardware circuit, enhances the anti-interference ability of the circuit, and also has the advantages of real-time, high efficiency, and adjustable parameters; in addition, it can complete the collection of the original time domain signal waveform, which is the ionization The analysis of layer propagation group delay and waveform characteristics provides information.

Claims (5)

1. Point-by-point scanning real-time processing method of spaceborne VHF transient signals based on Goertzel algorithm, which is characterized by including the following steps: (1) Perform broadband sampling of VHF transient signals to obtain the original time domain waveform of the detection frequency band; (2) Use the Goertzel algorithm to calculate the Fourier components of multiple frequency points in real time using a point-by-point sliding window; Includes the following steps: 1) Set the number of channels to 20; 2) Select the calculation window width N=101, 101 time domain signal sampling points to obtain a Fourier component frequency point amplitude, and calculate a Fourier component frequency point amplitude for each moving sampling point; 3) Select the Fourier component frequency; 4) Optimize the iterative formula of Goertzel algorithm; Among them, the iterative formula of the optimized Goertzel algorithm is: , where N is the window width, X (N) is the sampling value, for each given k value, /> ,/> is a constant; 5) Point-by-point sliding window calculates the Fourier components of multiple frequency points in real time; Discriminate signals according to discrimination conditions; The threshold and the duration of crossing the threshold are used as the discrimination conditions. When the number of frequency points that meet the discrimination conditions reaches M and has the characteristic of high frequency arriving first, the signal is considered valid. 2. The point-by-point scanning real-time processing method of spaceborne VHF transient signals based on the Goertzel algorithm according to claim 1, characterized in that in step (1), two band-pass filters are used to filter out two detection frequency bands signal of. 3. The point-by-point scanning real-time processing method of spaceborne VHF transient signals based on Goertzel algorithm according to claim 2, characterized in that the two detection frequency bands are low frequency band 25MHz-75MHz and high frequency band 110MHz-150MHz respectively. . 4. The point-by-point scanning real-time processing method of spaceborne VHF transient signals based on Goertzel algorithm according to claim 1, characterized in that, in step (1), select the sampling frequency fS=160MHz, and use two A/Ds The chip converts signals in two frequency bands at the same time, directly sampling the low-frequency band to obtain a digital signal of 25MHz-75MHz, and under-sampling the high-frequency band to obtain a digital signal of 10MHz-50MHz. 5. The point-by-point scanning real-time processing method of spaceborne VHF transient signals based on Goertzel algorithm according to claim 1, characterized in that, in step (1), the sampling duration is selected to be 256us.