Method and device for identifying electromagnetic interference of a radio telescope

Abstract

This invention relates to the field of radio astronomy, providing a method and apparatus for identifying electromagnetic interference from radio telescopes. The method includes: separating the original monitored spectrum signal into a background signal A and a residual signal B; fitting A to obtain a first residual signal, performing a confidence level test to obtain an abnormal signal set C1; fitting B to obtain a second residual signal, performing a confidence level test to obtain an abnormal signal set C2; and merging C1 and C2 to obtain an interference signal set C. The apparatus includes a signal separation module, a data fitting module, a residual analysis module, and a merging module. This invention fits the background and residual signals separately and obtains the residuals, achieving "stable processing" and avoiding subjective judgment bias. Compared with the traditional method using threshold decision, the algorithm of this invention has more objective criteria, less human intervention and uncertainty factors, and achieves "automation" in actual measurements. The algorithm is simple to implement, has low computational load, and fast program execution speed.

Description

Technical Field

[0001] This invention relates to the field of radio astronomy, and in particular to a method and apparatus for identifying electromagnetic interference from radio telescopes. Background Technology

[0002] Radio astronomy is a discipline that receives radio signals from the universe and combines astronomy with radio science. It plays a significant role in basic astronomical research, major national needs, national economic development, national defense, and social progress.

[0003] Radio astronomy observations are characterized by high receiver sensitivity, the inability to arbitrarily choose observation frequencies, and the absence of transmission (passive). The celestial objects observed are often located billions or even tens of billions of light-years away in the deep universe, where signals are extremely weak. Frequency sharing between radio astronomy observations with high-sensitivity receivers and active services (human radio services with transmission) is extremely difficult. Even very weak radio frequency interference signals can lead to observation failure or degraded efficiency; therefore, radio astronomy observations require sites with extremely low levels of electromagnetic interference. The International Telecommunication Union (ITU) Radio Regulations and my country's "Regulations on Radio Frequency Allocation" define frequency allocations for radio astronomy services as primary and secondary services. For radio astronomy services to achieve scientific observation goals and avoid various electromagnetic interferences, it is essential to reduce conventional strong interference sources, control low-power interference sources, prevent noise background degradation, and prevent sporadic interference.

[0004] In recent years, astronomical endeavors have developed rapidly, with many new observation frequencies falling outside the allocated radio astronomy frequency bands. Continuous advancements in radio astronomy technology have also led to increasingly higher telescope sensitivity. Simultaneously, with societal progress and the rapid development of radio communications, various radio services have emerged, resulting in an explosive increase in radio interference (RFI) sources and a more complex electromagnetic environment. Potential electromagnetic interference from residential life, industrial and medical equipment, and communications will inevitably impact the electromagnetic environment of radio telescopes. Therefore, to ensure the long-term, safe, and effective operation of existing radio telescopes and avoid the impact of surrounding electromagnetic background noise and severe electromagnetic disturbances, it is necessary to conduct large-scale, long-term monitoring and research of the electromagnetic environment around radio telescope sites. This will not only provide a scientific basis for the long-term, safe, and effective operation of the telescopes but also provide important technical specifications and references for the construction planning of the surrounding areas.

[0005] The principle block diagram of the radio telescope electromagnetic environment monitoring system is as follows: Figure 1As shown, the system mainly consists of a set of antennas, a set of low-noise amplifiers (LNAs), a high-performance spectrum analyzer or RF receiver module, and corresponding computer data processing hardware and software. The antennas are used to receive radio signals propagating in the air; the LNA is the core component of the entire test system, requiring sufficiently high gain to meet the requirements of electromagnetic environment measurement (generally, the noise floor of the spectrum analyzer must be at least 10 dB less than the amplitude of the measured signal), while also minimizing the noise figure; the spectrum analyzer or RF receiver module receives and collects data from the antennas and performs spectrum analysis; the computer processes, analyzes, and stores the data, and can comprehensively analyze radio monitoring data in conjunction with different observation targets and scientific needs to obtain the electromagnetic environment variation characteristics of the measured site. Furthermore, a filter can be installed between the antenna and the LNA to filter out strong signals, ensuring the LNA operates in the linear region; if necessary, an attenuator can also be installed to limit excessively strong received signals and prevent damage to the downstream spectrum analyzer or receiver. Figure 2 The image shows actual measured RFI monitoring data from a radio telescope, where the horizontal axis represents frequency and the vertical axis represents the corresponding measured power level.

[0006] Observe the raw data from the monitoring antenna, from Figure 2 It is easy to see that the signal characteristics of the monitoring data in different frequency bands vary greatly, the signal amplitude fluctuates significantly and changes irregularly, and the size of the interference peaks is also irregular. In practical applications, there is no unified quantitative standard for identifying RFI.

[0007] Currently, the method for monitoring RFI (Radio Frequency Intrusion) in antenna data involves displaying the data graphically on a computer and visually identifying frequencies and bands with electromagnetic interference. This method is inefficient and cannot process large amounts of data in a short time. Because signal amplitude fluctuates, simply setting a threshold to identify RFI is unlikely to be effective for all test data. Another commonly used method is the "three sigma criterion" monitoring method, which calculates the standard deviation of the monitoring data. If a single data point exceeds three times the standard deviation (or, depending on the actual situation, increases the probability criterion to five or seven times the standard deviation), it can be considered an outlier, and the frequency point corresponding to the outlier is considered to have RFI. However, the measured signal clearly does not possess the characteristics of "white noise" nor does it conform to a normal distribution; it exhibits typical "non-stationarity" characteristics. Therefore, the conventional "three sigma criterion" method for detecting outliers is not applicable here, often resulting in "false positives" and "false negatives."

[0008] Currently, researchers have attempted to apply artificial intelligence algorithms such as neural networks and deep learning to the monitoring and identification of RFIs. However, these algorithms are complex to design, require a significant amount of work for initial parameter tuning, and involve enormous computational demands, placing high demands on computer hardware, requiring fast processor speeds and massive storage space. For long-term RFI monitoring, the amount of monitoring data generated is enormous; therefore, the hardware cost of using artificial intelligence methods such as neural networks and deep learning for RFI identification is very high, making it not highly feasible.

[0009] In summary, there is an urgent need for intelligent automatic RFI monitoring and identification methods in practical work, so as to efficiently and accurately process the data of the monitoring antenna and quickly find the frequency points or frequency bands where RFI exists. Summary of the Invention

[0010] The purpose of this invention is to overcome the shortcomings of the prior art and provide a method and device for identifying electromagnetic interference (RFI) in radio telescopes. This method and device can efficiently and accurately process data from monitoring antennas and quickly identify frequency points or bands where RFI exists.

[0011] The present invention adopts the following technical solution:

[0012] On one hand, the present invention provides a method for identifying electromagnetic interference from radio telescopes, comprising the following steps:

[0013] S1: Separate the original monitoring spectrum signal into a background signal A and a residual signal B;

[0014] S2: Perform data fitting on the background signal A to obtain the first residual signal of the background signal A, and perform confidence level test on the first residual signal to obtain the set of abnormal signals C1 of the first residual signal;

[0015] S3: Perform data fitting on the remaining signal B to obtain the second residual signal of the remaining signal B, and perform confidence level test on the second residual signal to obtain the abnormal signal set C2 of the second residual signal;

[0016] S4: Merge set C1 and set C2 to obtain the final set of electromagnetic interference signals for the radio telescope, C.

[0017] In addition to the aspects and any possible implementations described above, a further implementation is provided in which, in step S1, the separation of the original monitoring spectrum signal employs a moving average filter; wherein, the length of the moving average window is determined by analyzing the Fourier transform of the original spectrum signal and the frequency response function of the moving average filter. Specifically, the length is determined by calculating the Fourier transform F of the original spectrum signal and the frequency response function H of the moving average filter, ensuring that the low-pass range of H covers the low-frequency band where the spectral peak of F is significant.

[0018] In addition to the aspects and any possible implementations described above, a further implementation is provided in which, in step S1, the original monitoring spectrum signal is separated by filtering the original signal using the following filter:

[0019] (a) Take the impulse response function (IRF) of the filter coefficients as: And satisfy:

[0020] ;

[0021] (b) The corresponding normalized frequency response function (FRF) is:

[0022] , ;

[0023] and ;

[0024] (c) Regarding the error and cutoff frequency Seeking the optimal To satisfy:

[0025] ;

[0026] in It is the impulse response function of a finite number of terms that satisfy conditions (a) and (b) above; the optimal low-pass filter is obtained as:

[0027] ;

[0028] in, ;

[0029] The filter order N is selected by solving the following formula:

[0030] .

[0031] In addition to the aspects and any possible implementations described above, a further implementation is provided in which, in step S1, the original monitoring spectrum signal is separated using Wiener filtering.

[0032] In addition to the aspects and any possible implementations described above, a further implementation is provided in which, in step S1, the original monitoring spectrum signal is separated using Kalman filtering.

[0033] In addition to the aspects and any possible implementations described above, a further implementation is provided in which, in step S1, the original monitoring spectrum signal is separated using adaptive filtering.

[0034] In addition to the aspects and any possible implementations described above, a further implementation is provided in which, in step S2, the fitting is achieved using a polynomial fitting method.

[0035] In addition to the aspects described above and any possible implementations, a further implementation is provided in which, in step S2, the fitting is achieved using piecewise linear interpolation.

[0036] In addition to the aspects described above and any possible implementations, a further implementation is provided in which, in step S2, the fitting is achieved using polynomial interpolation.

[0037] In addition to the aspects described above and any possible implementations, a further implementation is provided in which, in step S2, the fitting is achieved using spline interpolation.

[0038] In addition to the aspects described above and any possible implementations, a further implementation is provided in which, in step S3, the fitting is achieved using a polynomial fitting method.

[0039] In addition to the aspects described above and any possible implementations, a further implementation is provided in which, in step S3, the fitting is achieved using piecewise linear interpolation.

[0040] In addition to the aspects described above and any possible implementations, a further implementation is provided in which, in step S3, the fitting is achieved using polynomial interpolation.

[0041] In addition to the aspects described above and any possible implementations, a further implementation is provided in which, in step S3, the fitting is achieved using spline interpolation.

[0042] On the other hand, the present invention also provides a device for identifying electromagnetic interference from radio telescopes, comprising:

[0043] The signal separation module separates the original signal into a background signal A and a residual signal B.

[0044] The data fitting module performs data fitting on the background signal A to obtain the first residual signal of A, and performs data fitting on the remaining signal B to obtain the second residual signal of B.

[0045] The residual analysis module performs a confidence level test on the first residual signal to obtain the outlier set C1 of the first residual signal; and performs a confidence level test on the second residual signal to obtain the outlier set C2 of the second residual signal.

[0046] The merging module merges the anomaly sets C1 and C2 to obtain the electromagnetic interference signal set C of the radio telescope.

[0047] On the other hand, the present invention also provides an information processing terminal using the above-described method for identifying electromagnetic interference from radio telescopes.

[0048] On the other hand, the present invention also provides a computer-readable storage medium including instructions that, when executed on a computer, cause the computer to perform the above-described method for identifying electromagnetic interference from radio telescopes.

[0049] The beneficial effects of this invention are as follows:

[0050] 1. The technical solution of this invention utilizes the basic principles of digital signal processing and mathematical statistics to separate the original signal into two parts: a "background" signal and a "residual term" signal. Then, based on the characteristics of each part, targeted processing is performed to monitor RFI separately. The algorithm fits the two parts of data and obtains the residuals, achieving "stationarization processing." By analyzing the statistical characteristics of the residuals, the RFI signal is monitored. Therefore, the identification and judgment method has probabilistic criteria, avoiding bias caused by subjective judgment.

[0051] 2. The method of the present invention decomposes the signal and then processes it separately, thus avoiding the problem of neglecting some aspects due to the "one-size-fits-all" processing of the original signal. The effect is obvious and more accurate results are obtained.

[0052] 3. Compared with the method of simply using threshold decision, the criterion of the method of the present invention is more objective, with less human intervention and uncertainty, and achieves "automation" in actual test.

[0053] 4. The algorithm used in this invention is simple and easy to implement, requires little computation, and has a fast execution speed. Attached Figure Description

[0054] Figure 1 The diagram shown is a schematic of an electromagnetic environment monitoring system.

[0055] Figure 2 The image shows the measured electromagnetic interference of the radio telescope; where (a) frequency range: 50M-1GHz, (b) frequency range: 1-2GHz, and (c) frequency range: 2-3GHz.

[0056] Figure 3 The image shows the original spectral signal characteristics; where dark color represents the original signal and light color represents the background signal.

[0057] Figure 4 The diagram shows a flowchart of a method for identifying electromagnetic interference in a radio telescope according to an embodiment of the present invention.

[0058] Figure 5 The image shows the frequency response (normalized frequency) of the moving average filters of different lengths in the embodiment.

[0059] Figure 6 The figure shows the spectral amplitude (normalized frequency) obtained by second Fourier transform of the measured radio telescope RFI monitoring spectrum signal in the embodiment.

[0060] Figure 7 The diagram shown is a structural schematic of a radio telescope electromagnetic interference identification device according to an embodiment of the present invention. Detailed Implementation

[0061] The specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. It should be noted that the technical features or combinations of technical features described in the following embodiments should not be considered in isolation, but can be combined with each other to achieve better technical effects.

[0062] Careful examination Figure 3 The electromagnetic interference monitoring spectrum data shown in the image reveals that, across the entire frequency band, the spectrum roughly follows a baseline (i.e., the background). Figure 3-2 (As shown by the light-colored line in the image) It fluctuates randomly, with "abrupt changes" in some areas where the RFI is stronger.

[0063] This invention decomposes monitored spectrum data into two parts: a "baseline signal" and a "residual signal." Then, based on the characteristics of each part, it processes the data separately and designs corresponding algorithms to identify the frequency points and bands where RFI (Radio Frequency Identification) exists. The "baseline signal" is relatively "smooth" and "robust," while the "residual signal" exhibits more "spiking" and "protrusions." Therefore, it is necessary to further design effective algorithms based on their respective characteristics to complete the RFI identification.

[0064] like Figure 4 As shown in the figure, an embodiment of the present invention provides a method for identifying electromagnetic interference in radio telescopes, comprising the following steps:

[0065] S1: Separate the original monitoring spectrum signal into a background signal A and a residual signal B;

[0066] S2: Perform data fitting on the background signal A to obtain the first residual signal of the background signal A, and perform confidence level test on the first residual signal to obtain the set of abnormal signals C1 of the first residual signal;

[0067] S3: Perform data fitting on the remaining signal B to obtain the second residual signal of the remaining signal B, and perform confidence level test on the second residual signal to obtain the abnormal signal set C2 of the second residual signal;

[0068] S4: Merge set C1 and set C2 to obtain the final set of electromagnetic interference signals for the radio telescope, C.

[0069] Regarding the separation of the original signal:

[0070] In step S1, the separation of the original signal begins with the extraction of the "baseline signal". The baseline signal can be seen as the entire "trend" of the original signal, which is a relatively smooth curve. From the perspective of digital signal processing, the extraction of the baseline is a filtering process, that is, using "low-pass filtering" to remove "high-frequency noise" (which can also be seen as the randomness "spicules" of the data).

[0071] Considering that in an ideal scenario without RFI, the data received by the spectrum analyzer or receiver from each channel is white noise, the "weight" of the signal at each frequency point can be assumed to be the same. Therefore, the simplest moving average method can be used to extract the signal background. "Data smoothing" involves performing a moving average on the data within a "window" of a specific length. However, when using the moving average method to extract the signal background, the size of the moving window needs to be considered, as its length directly affects the final processing result. If the moving average window is too short, the "low-pass" effect is not ideal, and the smoothness of the filtered curve is insufficient; if the length is too long, it cannot accurately reflect the changes in the background and also increases the amount of data storage and computation. Therefore, a trade-off must be made based on the actual situation during data processing.

[0072] To select an appropriate moving average window length that strikes a balance between smoothness and rate of change in the filtering result, this invention considers using second-order Fourier analysis to perform secondary spectral analysis on the original spectral signal. Specifically, this method involves calculating the Fourier transform of the original spectral signal and the frequency response function of the moving average filter. The length of the moving average window is determined by analyzing the relationship between the two. Specifically, the Fourier transform F of the original spectral signal and the frequency response function H of the moving average filter are calculated, ensuring that the low-pass range of H covers the low-frequency band where the spectral peaks of F are significant, thus determining the length of the moving average window. It is important to note that this further Fourier transform of the spectral data is only used to analyze the digital characteristics of the spectral peaks and does not delve into their actual physical meaning.

[0073] According to the theory of digital signal processing, the mathematical expression for digital filtering is:

[0074]

[0075] in This represents the original input signal (i.e., the original FAST monitoring antenna spectrum data). This represents the filtered signal (i.e., the extracted background signal). These represent the coefficients of the filter.

[0076] For the moving average method, the above equation becomes:

[0077]

[0078] For ease of analysis, the length of the moving average window is set to... And the impulse response function (IRF) is:

[0079] ,

[0080] The corresponding frequency response function (FRF) is:

[0081]

[0082] Expanding and simplifying the above expression, we get:

[0083]

[0084] Figure 5 The graph presents the frequency response curves of moving average filters with different lengths at the normalized frequency. As can be seen from the graph, when the moving window is small, i.e., the number of averaging terms is small, the "low-pass" filter's ability to remove "high-frequency noise" is poor. For example, with a length of 10, there are wide side lobes outside the main lobe—indicating that high-frequency components can still pass through, so the filtered output is not very "smooth." Increasing the moving window length to 100 significantly improves its high-frequency filtering performance. However, as the moving window length increases, the improvement in high-frequency filtering performance gradually slows down, indicating that a longer moving window is not always better for achieving satisfactory high-frequency filtering results. In practical applications, it is necessary to select an appropriate moving filter length based on the characteristics of the measured signal.

[0085] In one specific embodiment, a method for selecting the length of the sliding filter window is given. Figure 6 The figure shows the amplitude of the Fourier transform of the spectral peak of a telescope's measured data at the normalized frequency. As can be seen from the figure, the significant low-frequency band of the spectral peak is "narrow," and at the normalized frequency, it is basically concentrated in the (0, 0.03) interval (the horizontal axis needs to be magnified locally). In this case, considering the frequency response curve of the moving average filter mentioned earlier, a window length of 120-150 is sufficient.

[0086] The above-described method of separating the original signal using a moving average is suitable for most radio telescope testing environments. However, when the background signal exhibits particularly strong fluctuations and jitter, a filter with a very small low-pass frequency band is required to achieve the desired effect. In this case, a very long moving average filter is needed, which leads to a large computational load and increased computation time, thus affecting the efficiency of interference identification. An improved method is to use an optimal low-pass filter to separate the original signal.

[0087] The design method for the optimal low-pass filter is as follows:

[0088] (a) The impulse response function (IRF) of the filter coefficients is taken as and satisfy

[0089]

[0090] (b) The corresponding frequency response function (FRF) is

[0091] ,

[0092] and,

[0093] (c) Regarding the error and cutoff frequency Seeking the optimal To satisfy

[0094]

[0095] in, It is a finite impulse response function that satisfies all conditions (a) and (b) above. Based on the optimal criterion given in (c) above, the optimal low-pass filter can be obtained as:

[0096]

[0097] in, .

[0098] Similar to the moving average processing described above, the optimal low-pass filter also faces the problem of order (or length) selection. The larger the order, the better the filtering effect, but the computational load will also increase accordingly. At the same time, the "tailing" effect of the filter will also cause more data to be discarded.

[0099] In practical applications, the filter order N can be selected according to the following formula:

[0100]

[0101] The length of the filter order N can be estimated and selected by solving the above inequalities.

[0102] For certain special measurement environments, such as when the interference of spectral signals around a radio telescope site is particularly strong in a certain frequency band, in order to deal with this situation, a special digital filter can be designed according to actual needs to separate the original signal. This can be achieved with the help of a dedicated digital filter design tool.

[0103] The electromagnetic environment around some sites is complex and variable, with signals changing rapidly and fluctuating greatly. Using filters with fixed parameters to separate the signals may not be very effective. In such cases, statistical signal processing methods or random signal processing methods that can adjust filter parameters in real time, such as Wiener filtering, Kalman filtering, and adaptive filtering, can be considered to separate the original signals.

[0104] Identification of Radio Interference Signals (RFI)

[0105] After separating the original signal into the "ground signal" and the "residual signal," effective algorithms can be designed based on their respective characteristics to identify RFI. Generally, RFI signals manifest as outliers at a specific frequency point or frequency band. The idea behind this invention is to use signal fitting to obtain residual information, and then monitor RFI by analyzing the statistical characteristics of the residuals. The data fitting, residual extraction, and analysis processes are described in detail below.

[0106] For a general multinomial data fitting model, random variables and multiple ordinary variables Related, can be expressed as:

[0107] ,

[0108] The residuals follow a normal distribution, i.e. .in , All are with Irrelevant unknown parameters. The parameters are estimated using maximum likelihood estimation, i.e., taking... , so that , , ..., hour:

[0109]

[0110] To reach the minimum.

[0111] For ease of explanation, a matrix is ​​introduced:

[0112] , ,

[0113] Therefore, the data fitting model can be written as:

[0114]

[0115] In the sense of maximum likelihood, The optimal solution is:

[0116]

[0117] equation

[0118]

[0119] This is called the p-data polynomial fitting equation.

[0120] According to statistical theory, residuals It approximately follows a normal distribution. Therefore, for the electromagnetic interference monitoring spectrum data of radio telescopes, the separated "background" and "random term" signals can be modeled separately using polynomial models:

[0121]

[0122] The data is fitted and the residuals are derived accordingly. Since the residuals approximately conform to a normal distribution, outliers are detected using conventional confidence level monitoring methods. As long as the corresponding confidence level threshold is set to monitor each residual value, the outlier values ​​of the residuals in a probabilistic statistical sense can be obtained. The frequency points corresponding to the outlier values ​​of the residuals are the frequency points where RFI exists.

[0123] Data fitting can be achieved using various methods. In practical applications, other methods can also be adopted based on the characteristics of the interference signal, such as spline fitting, polynomial interpolation fitting, piecewise linear interpolation fitting, etc.

[0124] The technical solution of this invention can be implemented in software on a general-purpose computer, or it can be implemented in hardware by designing a dedicated signal processing chip, such as DSP, FPGA, GPU, etc.

[0125] like Figure 7 As shown, in one specific embodiment, a device for identifying electromagnetic interference from a radio telescope is provided, characterized in that the device comprises:

[0126] The signal separation module separates the original signal into a background signal A and a residual signal B.

[0127] The data fitting module performs data fitting on the background signal A to obtain the first residual signal of A, and performs data fitting on the remaining signal B to obtain the second residual signal of B.

[0128] The residual analysis module performs a confidence level test on the first residual signal to obtain the outlier set C1 of the first residual signal; and performs a confidence level test on the second residual signal to obtain the outlier set C2 of the second residual signal.

[0129] The merging module merges the anomaly sets C1 and C2 to obtain the electromagnetic interference signal set C of the radio telescope.

[0130] In one specific embodiment, an information processing terminal is provided to implement the above-described method for identifying electromagnetic interference from radio telescopes.

[0131] In one embodiment, a computer-readable storage medium is provided, including instructions that, when executed on a computer, cause the computer to perform the above-described method for identifying electromagnetic interference from radio telescopes.

[0132] This invention utilizes the fundamental principles of digital signal processing and mathematical statistics to separate the original signal into two parts: a "background" signal and a "residual term." Then, based on the characteristics of each part, it processes them separately for RFI monitoring. The algorithm fits the two data parts and obtains the residuals, achieving "stationarization." By analyzing the statistical characteristics of the residuals, the RFI signal is monitored. Therefore, the identification and judgment method has probabilistic criteria, avoiding bias caused by subjective judgment.

[0133] This method of decomposing the signal and then processing it separately, "tasking it one by one," avoids the "over-processing" that leads to problems caused by treating the original signal in a "one-size-fits-all" manner. Its effect is significant, yielding more accurate results. Compared to methods that simply use threshold decisions, this algorithm's criteria are more objective, with less human intervention and uncertainty, achieving "automation" in practical tests. Furthermore, the algorithm is simple to implement, computationally inexpensive, and executes quickly.

[0134] While several embodiments of the present invention have been provided herein, those skilled in the art should understand that modifications can be made to these embodiments without departing from the spirit of the invention. The above embodiments are merely exemplary and should not be construed as limiting the scope of the invention.

Claims

1. A method for identifying electromagnetic interference from radio telescopes, characterized in that, The identification method includes the following steps: S1: Separate the spectrum of the original monitoring signal into a background signal A and a residual signal B; S2: Perform data fitting on the background signal A to obtain the first residual signal of the background signal A, and perform a confidence level test on the first residual signal to obtain the set of abnormal signals C1 of the first residual signal; S3: Perform data fitting on the remaining signal B to obtain the second residual signal of the remaining signal B, and perform confidence level test on the second residual signal to obtain the abnormal signal set C2 of the second residual signal; S4: Merge set C1 and set C2 to obtain the final set of electromagnetic interference signals for the radio telescope, C.

2. The method for identifying electromagnetic interference from radio telescopes as described in claim 1, characterized in that, In step S1, the original monitoring signal spectrum separation adopts a moving average filter; wherein, the length of the moving average window is determined by calculating the Fourier transform F of the original spectrum signal and the frequency response function H of the moving average filter, wherein, it is ensured that the low-pass range of H covers the low-frequency band where the spectral peak of F is significant.

3. The method for identifying electromagnetic interference from radio telescopes as described in claim 1, characterized in that, In step S1, the original monitoring signal spectrum is separated by filtering the original signal using the following filter design: (a) The filter coefficients are taken as follows: The impulse response function IRF is... And satisfy: ； (b) The corresponding normalized frequency response function (FRF) is: , ； and ; (c) Regarding the error and cutoff frequency Seeking the optimal To make it meet the conditions: ； in It is the impulse response function of a finite number of terms that satisfy conditions (a) and (b) above; find the optimal low-pass filter. for: ； in, ; The order N of the filter is selected by solving the following formula: 。 4. The method for identifying electromagnetic interference from radio telescopes as described in claim 1, characterized in that, In step S1, the original monitoring signal spectrum is separated by Wiener filtering, Kalman filtering, or adaptive filtering.

5. The method for identifying electromagnetic interference from radio telescopes as described in any one of claims 1-4, characterized in that, In step S2, the fitting is achieved using a polynomial fitting method, or piecewise linear interpolation, or polynomial interpolation, or spline interpolation.

6. The method for identifying electromagnetic interference from radio telescopes as described in any one of claims 1-4, characterized in that, In step S3, the fitting is achieved using a polynomial fitting method, or piecewise linear interpolation, or polynomial interpolation, or spline interpolation.

7. A device for identifying electromagnetic interference from a radio telescope, characterized in that, The device includes: The signal separation module separates the spectrum of the original monitoring signal into a background signal A and a residual signal B. The data fitting module performs data fitting on the background signal A to obtain the first residual signal of A, and performs data fitting on the remaining signal B to obtain the second residual signal of B. The residual analysis module performs a confidence level test on the first residual signal to obtain the outlier set C1 of the first residual signal; and performs a confidence level test on the second residual signal to obtain the outlier set C2 of the second residual signal. The merging module merges the anomaly sets C1 and C2 to obtain the electromagnetic interference signal set C of the radio telescope.

8. An information processing terminal that implements the method for identifying electromagnetic interference from radio telescopes as described in any one of claims 1-6.

9. A computer-readable storage medium comprising instructions, when executed on a computer, causing the computer to perform the method for identifying electromagnetic interference from a radio telescope as described in any one of claims 1-6.

Images