Unknown radar signal sorting method, system, device and medium based on sequence recovery

Abstract

The present invention relates to the field of digital signal processing and discloses a method, system, device, and medium for sorting unknown radar signals based on sequence recovery. The method comprises: short-time slot information extraction based on cluster analysis, parameter correlation analysis based on multi-time window accumulation, sequence information recovery processing based on matching processing, and sorting processing based on time series information correlation. The present invention primarily reconstructs radar signal pulse sequences through a sequence recovery method, accumulates parameter information in multiple time windows, improves signal significance, utilizes the repetitiveness between multiple intercepted signal sequences to restore signal sequence information, and sorts unknown signals based on the sequence information, thereby improving the confidence of signal sorting results in complex environments and reducing false alarms caused by noisy data in the environment.

Description

Technical Field

[0001] The present invention relates to the field of digital signal processing, and in particular to a method, system, device and medium for sorting unknown radar signals based on sequence recovery. Background Art

[0002] With the continuous development of electronic technology and digital signal processing technology, the radiation source signals in the current electromagnetic environment are dense and complex, the target signals and background signals are highly intertwined, and there are a large number of unknown targets in the environment for which there is a lack of prior knowledge. This poses a great challenge to the sorting of radar radiation source signals.

[0003] Traditional histogram-based methods use statistical analysis of conventional parameters to isolate statistically significant signals. This method, with its simplicity and efficiency, is widely used and effective in simple electromagnetic environments with high target parameter differentiation. However, in today's complex and dense electromagnetic environments, conventional parameter differentiation is low, and the complex and agile nature of advanced radar systems and the low probability of signal interception make it difficult to generate statistical information with high significance and differentiation, making this technique less adaptable.

[0004] The sequence matching-based method enriches the discernible parameter dimensions by introducing radar sequence template information, which can effectively improve the adaptability to complex environments and objects. However, this method relies on prior knowledge and cannot adapt to the processing of unknown targets.

[0005] With the continuous development of artificial intelligence technology in recent years, technologies such as machine learning have gradually become a hot research direction in the field of signal sorting. Most of these methods use network models for sample training, replacing manual design with data-driven methods to extract signal features, which has improved the degree of automation and object adaptability of signal sorting to a certain extent. However, these methods rely on manually labeled prior information and have limited adaptability to unknown targets. Summary of the Invention

[0006] To address the above-mentioned problems, the present invention proposes a method, system, device, and medium for sorting unknown radar signals based on sequence recovery. The method mainly reconstructs radar signal pulse sequences based on a sequence recovery method, accumulates parameter information in multiple time windows, improves signal significance, and utilizes the repetitiveness between multiple intercepted signal sequences to restore signal sequence information. The method then sorts unknown signals based on the sequence information, thereby improving the confidence level of signal sorting results in complex environments and reducing false alarms caused by noise data in the environment.

[0007] The technical solution adopted in the present invention is as follows:

[0008] A method for sorting unknown radar signals based on sequence recovery, comprising:

[0009] Short-time-slot information extraction based on cluster analysis: sequentially cluster the data samples intercepted in the current time window to obtain the signal clustering results within the short-time segment; extract cluster information from each signal clustering result in the current time window and record it;

[0010] Parameter association analysis based on multi-time window accumulation: Based on the cluster information recorded in multiple time windows, the consistency of conventional parameters is compared between each two, clusters with the same conventional parameters are merged, and whether the statistical significance conditions are met is determined;

[0011] Sequence information recovery based on matching: For signals that meet the statistical significance criteria, sequence recovery is performed using multiple sequences stored in the signal to obtain updated sequence descriptions and cluster information, i.e., sorting results.

[0012] Sorting based on temporal information association: Based on the updated sequence description and cluster information, the signal is further sorted and tracked in subsequent time windows.

[0013] Furthermore, sequentially clustering the data samples intercepted in the current time window to obtain the signal clustering results within the short time segment includes:

[0014] According to the preset conventional parameter threshold, the data samples that meet the judgment conditions are Clustered into one class, where n is the number of pulses in the sample, and the i-th sample in the data sample X Represents the conventional parameter description of the i-th pulse signal; the judgment conditions include: corresponding to any two samples and , the difference of the conventional parameters in any dimension is not greater than the corresponding conventional parameter threshold;

[0015] After clustering, the data samples The clustering result is , where m represents the number of clusters formed by the data sample X, and any clustering result It is composed of the elements in the data sample X, that is, , where k is the clustering result The number of pulses contained in .

[0016] Furthermore, the conventional parameter thresholds include: ,in is the pulse arrival angle threshold, is the pulse frequency threshold, is the pulse width threshold, is the pulse arrival time threshold.

[0017] Furthermore, extracting cluster information from each signal clustering result in the current time window and recording the information includes:

[0018] For any clustering result , extract cluster information :

[0019]

[0020]

[0021]

[0022]

[0023]

[0024] in, is the pulse arrival angle, is the pulse frequency, is the pulse width; sequence information The jth element in It is calculated by the arrival time difference of adjacent elements, that is, ; Clustering results The number of pulses in , here the value is k.

[0025] Furthermore, the cluster information recorded in multiple time windows is compared with each other for consistency of conventional parameters, clusters with the same conventional parameters are merged, and whether statistical significance conditions are met is determined, including:

[0026] For the jth cluster information in the i-th time window , and the nth cluster information in the mth time window If the difference between the two conventional parameters in any dimension is not greater than the preset conventional parameter threshold, the two are merged and the number of elements is accumulated;

[0027] For N results to be merged in different time windows, the cumulative number of pulses ; If the number of pulses Greater than the preset threshold , then the corresponding signal is considered to be statistically significant in the conventional parameter dimension and can recover the sequence information.

[0028] Furthermore, for the signal determined to meet the statistical significance condition, sequence recovery is performed using the multiple sequence information stored in the signal to obtain updated sequence description and cluster information, i.e., sorting results, including:

[0029] For the sequence with sequence , compare whether the two have a valid common subsequence; if so, reconstruct the common sequence of the two by filling the head and tail;

[0030] After forming the common sequence, the sequences that have not been compared are selected one by one and compared with the common sequence repeatedly. If a valid common subsequence can still be obtained, it is further reconstructed to form a new common sequence;

[0031] After all sequences are compared, the final common sequence is obtained:

[0032] SEQ com

[0033] in, The public sequence SEQ com Time series data;

[0034] Based on the public sequence SEQ com , get the final cluster information:

[0035] cls_info com =

[0036] in, Class cluster information cls_info com General parameter statistics of .

[0037] Furthermore, the signal is continuously sorted and tracked in subsequent time windows based on the updated sequence description and cluster information, including:

[0038] Based on cluster information cls_info com Compare cluster information cls_info in subsequent time windows new , if the difference between the two general parameters is less than the preset general parameter threshold, and cls_info new The sequence is cls_info com A subset of the sequence, then cls_info new The corresponding class cluster and cls_info com The corresponding signals are the same signal, which can be sorted and tracked, and cls_info can be updated. com The relevant information is collected to form the sorting results.

[0039] A system for sorting unknown radar signals based on sequence recovery, comprising:

[0040] The short-time-slot information extraction module based on cluster analysis is configured to sequentially cluster the data samples intercepted in the current time window to obtain the signal clustering results within the short-time segment; extract cluster information from each signal clustering result in the current time window and record it;

[0041] The parameter association analysis module based on multi-time window accumulation is configured to combine clusters with the same conventional parameters by pairwise comparison based on cluster information recorded in multiple time windows, and determine whether the statistical significance conditions are met;

[0042] A sequence information recovery processing module based on matching processing is configured to perform sequence recovery on a signal determined to meet the statistical significance condition using multiple sequence information stored in the signal to obtain an updated sequence description and cluster information, i.e., a sorting result;

[0043] The sorting and processing module based on the association of time series information is configured to continue sorting and tracking the signals in subsequent time windows based on the updated sequence description and cluster information.

[0044] A computer device includes a memory and a processor, wherein the memory stores a computer program, and the processor implements the above-mentioned unknown radar signal sorting method based on sequence recovery when executing the computer program.

[0045] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the above-mentioned unknown radar signal sorting method based on sequence recovery.

[0046] The beneficial effects of the present invention are:

[0047] (1) The present invention can effectively separate true radar signals from complex noisy environments, reducing the false alarm rate of true signals while suppressing false alarms caused by noise signals. Unlike traditional signal sorting methods, the present invention is based on the cumulative processing of dynamically intercepted data in multiple observation windows. Through layer-by-layer data processing, it forms a calculation of the statistical significance of conventional parameters of radar signals and extracts the sequence information of regularly transmitted signals, thereby enhancing the correlation of the same signal between multiple observation time slots and improving the discrimination between true signals and noise signals.

[0048] (2) During data processing, by correlating the representation of time series information, the same signal can be stably tracked on the basis of forming an effective sorting result, thereby improving the stability of the sorting result. The present invention has good scalability and robustness and can adapt to sorting and recognition processing in complex environments.

[0049] (3) Compared with traditional algorithms based on histogram statistics, the present invention can enhance the correlation between the same signals by introducing sequence information, reduce false alarms caused by noise, and make the sorting results more robust.

[0050] (4) Compared with the algorithm using sequence matching recognition, the present invention introduces a method based on the combination of multiple time windows to automatically restore sequence information. It can sort signals without relying on prior knowledge, thereby enhancing the ability to process unknown signals.

[0051] (5) The present invention combines signal sorting technology with time series data processing technology to adapt to the sorting of unknown signals in complex noise environments. It reconstructs the sequence information of the signal by means of sequence recovery, and obtains signal time series characteristics that are difficult to obtain based on methods such as histogram statistics. It also combines with the statistical significance of multi-time window parameters to improve the correlation between signals and the distinction between signals and noise, and can achieve a better sorting effect on unknown signals in complex environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 -Flowchart of the unknown radar signal sorting method based on sequence recovery of the present invention.

[0053] Figure 2 -Generate a common sequence map by position shift comparison.

[0054] Figure 3 - Aliasing of target 1 and target 2 signal data and environmental noise data.

[0055] Figure 4 -The reconstructed sequence of target 1 signal data in the first two interception time windows.

[0056] Figure 5 -Reconstructed sequence of target 1 signal data formed in the first 5 interception time windows.

[0057] Figure 6 -The final sequence of target 1 signal data recovered by multi-time window processing.

[0058] Figure 7 -The reconstructed sequence of target 2 signal data formed in the first two interception time windows.

[0059] Figure 8 -The reconstructed sequence of target 2 signal data formed in the first five interception time windows.

[0060] Figure 9 -The final sequence of target 2 signal data recovered by multi-time window processing.

[0061] Figure 10 - "Time-frequency" distribution diagram of the sorting results of the newly intercepted data of Target 1 and Target 2.

[0062] Figure 11 - "Time-weighted cycle" distribution chart of the sorting results of the newly intercepted data for Target 1 and Target 2.

[0063] Figure 12 - “Time-Pulse Width” distribution diagram of the sorting results of the newly intercepted data for Target 1 and Target 2.

[0064] Figure 13 - "Time-amplitude" distribution diagram of the sorting results of the newly intercepted data for Target 1 and Target 2. DETAILED DESCRIPTION

[0065] In order to have a clearer understanding of the technical features, purposes and effects of the present invention, the specific embodiments of the present invention are now described. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention. That is, the embodiments described are only part of the embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making creative work are within the scope of protection of the present invention.

[0066] Example 1

[0067] like Figure 1 As shown, this embodiment provides a method for sorting unknown radar signals based on sequence recovery, including:

[0068] Short-time-slot information extraction based on cluster analysis: sequentially cluster the data samples intercepted in the current time window to obtain the signal clustering results within the short-time segment; extract cluster information from each signal clustering result in the current time window and record it;

[0069] Parameter association analysis based on multi-time window accumulation: Based on the cluster information recorded in multiple time windows, the consistency of conventional parameters is compared between each two, clusters with the same conventional parameters are merged, and whether the statistical significance conditions are met is determined;

[0070] Sequence information recovery based on matching: For signals that meet the statistical significance criteria, sequence recovery is performed using multiple sequences stored in the signal to obtain updated sequence descriptions and cluster information, i.e., sorting results.

[0071] Sorting based on temporal information association: Based on the updated sequence description and cluster information, the signal is further sorted and tracked in subsequent time windows.

[0072] It should be noted that in engineering practice, due to the interweaving of radiation source signals with noise and spurious signals in the environmental background, the two are highly overlapped in the conventional parameter dimension, making it difficult to effectively distinguish between true and false, resulting in a high false alarm rate caused by background noise, while the real target signal has a stable emission timing.

[0073] Therefore, this method forms information accumulation and restores the signal pulse sequence through multiple interceptions, which can be used to improve the distinction between target signals and background signals, enhance the correlation between target signals, and thus improve the ability to sort unknown signals in complex environments.

[0074] Furthermore, considering that radar pulse sequences, in addition to conventional parameters such as frequency, pulse width and repetition period, often have unique timing patterns and are periodic, repeating in different time windows, they can provide an important basis for signal sorting and identification.

[0075] Therefore, this method combines multiple time windows for processing, leveraging the cumulative data of signals with identical parameters observed over multiple time slots to enhance the salience of repetitive signals and extract their temporal features. This salience enhances the distinction between true and noisy signals, while time-series restoration strengthens the correlation between identical signals, ultimately removing noise and robustly separating true signals.

[0076] Specifically, the unknown radar signal sorting method of this embodiment can be implemented by the following steps:

[0077] (1) Short-slot information extraction based on cluster analysis

[0078] Short-time slot clustering: data samples intercepted in the current window Perform sequential clustering to obtain the signal clustering results within a short time segment, where n is the number of pulses in the sample and the i-th element Represents the general parameter description of the i-th pulse signal.

[0079] Specifically, according to the pre-set parameter threshold, the qualified data samples can be clustered into one category, and the parameter threshold of each dimension is expressed as Any two samples and The criteria for being clustered into the same category are: the difference in any dimension parameter is not greater than the parameter threshold, such as the frequency dimension must satisfy delta_rf ij ,in:

[0080] delta_rf ij

[0081] After clustering, the data samples The clustering result is , where m represents the number of clusters formed by the data sample X, and any clustering result It is composed of the elements in X, that is, , where k is The number of pulses contained in the class.

[0082] Cluster information extraction: Extract cluster information for each clustering result in the current window and record it, including the statistical values ​​of each dimension parameter and time series data.

[0083] Specifically, for clusters , its cluster information is recorded as follows:

[0084]

[0085] in, is the number of pulses in the cluster, which is k. The other parameters are calculated as follows:

[0086]

[0087]

[0088]

[0089]

[0090] Among them, sequence information The jth element in is calculated by the arrival time difference of adjacent elements:

[0091]

[0092] (2) Parameter correlation analysis based on multi-time window accumulation

[0093] According to the cluster information recorded in multiple observation windows, the consistency of parameters is compared between two of the three-dimensional conventional parameters such as azimuth, frequency and pulse width. Clusters with the same conventional parameters are merged and it is determined whether they meet the statistical significance conditions.

[0094] For example, for the jth cluster information in the i-th time window and the nth cluster information in the m-th time window:

[0095]

[0096]

[0097] If the differences between the two in the three conventional parameter dimensions of DOA, RF, and PW are all less than the preset thresholds TLR_DOA, TLR_RF, and TLR_PW, the two are merged and the number of their elements is accumulated. Similarly, for the N results to be merged in different time windows, the number of their pulses is accumulated:

[0098]

[0099] If the value NUM is greater than the preset threshold , then the signal is considered to be statistically significant in the conventional parameter dimension and the sequence information can be recovered.

[0100] (3) Sequence information recovery based on matching

[0101] For signals that are determined to be statistically significant after accumulation of multiple time windows, the multiple sequence information stored in the signal is used Perform sequence recovery to obtain a more complete sequence description.

[0102] Specifically, for two sequences and , compare whether it has a valid common subsequence, if so, reconstruct the common sequence of the two by filling the head and tail, and the two sequences form a common sequence through sequence matching and reconstruction. Figure 2 shown.

[0103] After forming the common sequence, select the sequences that have not been compared and repeat the above process with the common sequence one by one. If a valid common subsequence can still be obtained, further reconstruct it to form a new common sequence. After all sequences are compared, the final common sequence is obtained:

[0104] SEQ com

[0105] in, The public sequence SEQ com Time series data;

[0106] Based on the public sequence SEQ com , get the final cluster information:

[0107] cls_info com =

[0108] in, Class cluster information cls_info com General parameter statistics of .

[0109] (4) Sorting based on temporal information association

[0110] The above cluster information cls_info com That is, it is an effective sorting result after multi-time window processing. The cluster information can be used to sort and track the signal in subsequent observation windows.

[0111] Specifically, based on the cluster information cls_info com Compare cluster information cls_info in subsequent time windows new , if the difference between the two general parameters is less than the preset general parameter threshold, and cls_info new The sequence is cls_info com A subset of the sequence, then cls_info new The corresponding class cluster and cls_info com The corresponding signals are the same signal, which can be sorted and tracked, and cls_info can be updated. com Update relevant information in the process to form reportable sorting conclusions.

[0112] To demonstrate the effectiveness of this method, a test was conducted on a simulated data sample in a dense and complex X-band environment. The data sample contained signal samples of two radars at the same direction: target 1 (frequency 9760MHz, pulse width 0.8us) and target 2 (frequency 9765MHz, pulse width 0.55us), and also contained random noise signals in the same frequency band and direction at 9740MHz~9780MHz. The aliasing data of the target signal and the noise signal are shown in Figure 2. Figure 3 shown.

[0113] Figures 4 to 6 The recovery process of the target 1 signal sequence under the simulation data is given. Figures 7 to 9 The recovery process of the target 2 signal sequence under the simulation data is given. It can be seen that with the accumulation of time windows, the recovered sequence gradually tends to be perfect.

[0114] Figures 10 to 13 The results of the sorting of the newly captured data samples by using sequence recovery information to perform correlation tracking are given. As can be seen from the figure, both target signals can be effectively sorted from the background signal and can be distinguished from each other.

[0115] In summary, the method of the present invention can effectively sort radar signals in complex environments through multi-time window sequence recovery, and has certain scalability and practicality.

[0116] Example 2

[0117] This embodiment provides an unknown radar signal sorting system based on sequence recovery, including:

[0118] The short-time-slot information extraction module based on cluster analysis is configured to sequentially cluster the data samples intercepted in the current time window to obtain the signal clustering results within the short-time segment; extract cluster information from each signal clustering result in the current time window and record it;

[0119] The parameter association analysis module based on multi-time window accumulation is configured to combine clusters with the same conventional parameters by pairwise comparison based on cluster information recorded in multiple time windows, and determine whether the statistical significance conditions are met;

[0120] A sequence information recovery processing module based on matching processing is configured to perform sequence recovery on a signal determined to meet the statistical significance condition using multiple sequence information stored in the signal to obtain an updated sequence description and cluster information, i.e., a sorting result;

[0121] The sorting and processing module based on the association of time series information is configured to continue sorting and tracking the signals in subsequent time windows based on the updated sequence description and cluster information.

[0122] Example 3

[0123] This embodiment is based on embodiment 1:

[0124] This embodiment provides a computer device including a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, it implements the unknown radar signal sorting method based on sequence recovery of Embodiment 1. The computer program may be in source code form, object code form, an executable file, or some intermediate form.

[0125] Example 4

[0126] This embodiment is based on embodiment 1:

[0127] This embodiment provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the unknown radar signal sorting method based on sequence recovery of embodiment 1. The computer program may be in source code form, object code form, executable file, or some intermediate form. The storage medium includes: any entity or device capable of carrying computer program code, recording medium, computer memory, read-only memory (ROM), random access memory (RAM), electric carrier signal, telecommunication signal, and software distribution medium. It should be noted that the content contained in the storage medium may be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, the storage medium does not include electric carrier signal and telecommunication signal.

[0128] It should be noted that, for the sake of simplicity, the aforementioned method embodiments are described as a series of action combinations. However, those skilled in the art should be aware that this application is not limited by the order of the actions described, because according to this application, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in this specification are all preferred embodiments, and the actions and modules involved are not necessarily required by this application.

Claims

1. A method for sorting unknown radar signals based on sequence recovery, characterized in that: include: Short-time-slot information extraction based on cluster analysis: sequentially cluster the data samples intercepted in the current time window to obtain the signal clustering results within the short-time segment; Extract cluster information from each signal clustering result in the current time window and record it; Parameter association analysis based on multi-time window accumulation: Based on the cluster information recorded in multiple time windows, the consistency of conventional parameters is compared between each two, clusters with the same conventional parameters are merged, and whether the statistical significance conditions are met is determined; Sequence information recovery based on matching: For signals that meet the statistical significance criteria, sequence recovery is performed using multiple sequences stored in the signal to obtain updated sequence descriptions and cluster information, i.e., sorting results. Sorting based on temporal information association: Based on the updated sequence description and cluster information, the signal is sorted and tracked in subsequent time windows. The cluster information recorded in multiple time windows is compared with each other for consistency of conventional parameters, clusters with the same conventional parameters are merged, and whether statistical significance conditions are met are determined, including: For the jth cluster information in the i-th time window , and the nth cluster information in the mth time window If the difference between the two conventional parameters in any dimension is not greater than the preset conventional parameter threshold, the two are merged and the number of elements is accumulated; For N results to be merged in different time windows, the cumulative number of pulses If the number of pulses Greater than the preset quantity threshold , then the corresponding signal is considered to be statistically significant in the conventional parameter dimension and can recover the sequence information.

2. The unknown radar signal sorting method based on sequence recovery according to claim 1 is characterized in that: The sequential clustering of the data samples intercepted in the current time window to obtain the signal clustering results within the short time segment includes: According to the preset conventional parameter threshold, the data samples that meet the judgment conditions are Clustered into one class, where n is the number of pulses in the sample, and the i-th sample in the data sample X Represents the conventional parameter description of the i-th pulse signal; the judgment conditions include: corresponding to any two samples and , the difference of the conventional parameters in any dimension is not greater than the corresponding conventional parameter threshold; After clustering, the data samples The clustering result is , where m represents the number of clusters formed by the data sample X, and any clustering result It is composed of the elements in the data sample X, that is, , where k is the clustering result The number of pulses contained in .

3. The unknown radar signal sorting method based on sequence recovery according to claim 2 is characterized in that: The conventional parameter thresholds include: ,in is the pulse arrival angle threshold, is the pulse frequency threshold, is the pulse width threshold, is the pulse arrival time threshold.

4. The unknown radar signal sorting method based on sequence recovery according to claim 3 is characterized in that: The step of extracting cluster information from each signal clustering result in the current time window and recording the cluster information includes: For any clustering result , extract cluster information : in, is the pulse arrival angle, is the pulse frequency, is the pulse width; sequence information The jth element in It is calculated by the arrival time difference of adjacent elements, that is, ; Clustering results The number of pulses in , here the value is k.

5. The unknown radar signal sorting method based on sequence recovery according to claim 1 is characterized in that: For the signal determined to meet the statistical significance condition, sequence recovery is performed using multiple sequence information stored in the signal to obtain updated sequence description and cluster information, i.e., sorting results, including: For the sequence with sequence , compare whether the two have a valid common subsequence; if so, reconstruct the common sequence of the two by filling the head and tail; After forming the common sequence, the sequences that have not been compared are selected one by one and compared with the common sequence repeatedly. If a valid common subsequence can still be obtained, it is further reconstructed to form a new common sequence; After all sequences are compared, the final common sequence is obtained: SEQ com in, The public sequence SEQ com Time series data; Based on the public sequence SEQ com , get the final cluster information: cls_info com = in, Class cluster information cls_info com General parameter statistics of .

6. The unknown radar signal sorting method based on sequence recovery according to claim 1 is characterized in that: The signal is continuously sorted and tracked in a subsequent time window based on the updated sequence description and cluster information, including: Based on cluster information cls_info com Compare cluster information cls_info in subsequent time windows new , if the difference between the two general parameters is less than the preset general parameter threshold, and cls_info new The sequence is cls_info com A subset of the sequence, then cls_info new The corresponding class cluster and cls_info com The corresponding signals are the same signal, which can be sorted and tracked, and cls_info can be updated. com The relevant information is collected to form the sorting results.

7. An unknown radar signal sorting system based on sequence recovery, characterized in that: include: The short-time-slot information extraction module based on cluster analysis is configured to sequentially cluster the data samples intercepted in the current time window to obtain the signal clustering results within the short-time segment; Extract cluster information from each signal clustering result in the current time window and record it; The parameter association analysis module based on multi-time window accumulation is configured to merge clusters with the same conventional parameters based on the cluster information recorded in multiple time windows by comparing the consistency of conventional parameters between two of them, and determine whether the statistical significance conditions are met, including: for the jth cluster information in the i-th time window , and the nth cluster information in the mth time window If the difference between the two conventional parameters in any dimension is not greater than the preset conventional parameter threshold, the two will be merged and the number of elements will be accumulated; for N results to be merged in different time windows, the cumulative number of pulses ; If the number of pulses Greater than the preset threshold , then the corresponding signal is considered to be statistically significant in the conventional parameter dimension and can be used to recover sequence information; A sequence information recovery processing module based on matching processing is configured to perform sequence recovery on a signal determined to meet the statistical significance condition using multiple sequence information stored in the signal to obtain an updated sequence description and cluster information, i.e., a sorting result; The sorting and processing module based on the association of time series information is configured to continue sorting and tracking the signals in subsequent time windows based on the updated sequence description and cluster information.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the unknown radar signal sorting method based on sequence recovery according to any one of claims 1 to 6 is implemented.

9. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the unknown radar signal sorting method based on sequence recovery according to any one of claims 1 to 6 is implemented.

Images