Radiation source template library dynamic generation method and device

Abstract

The present disclosure discloses a radiation source template library dynamic generation method and device, a first pulse description word sequence PDW matrix is obtained according to the extracted parameters of the received electromagnetic wave signal, a second pulse description word sequence PDW matrix is generated according to the time of arrival TOA batch; the third pulse description word sequence PDW matrix is clustered by using the adaptive setting DBSCAN algorithm; a plurality of clusters are obtained according to the single batch radiation source clustering result, the incremental clustering result of the plurality of clusters is compared with the clusters in the radiation source template library to obtain the type of the plurality of clusters; the radiation source parameters in the radiation source template library are adjusted according to the type of the plurality of clusters, and a dynamic template library is generated. The present disclosure can more efficiently process electromagnetic wave signals, extract valuable feature information, adjust the radiation source parameters in the radiation source template library according to the type of the plurality of clusters, and can update the template library in time.

Description

Technical Field

[0001] This invention relates to the field of radiation source identification and management technology, specifically to a method and system for generating a dynamic template library of radiation sources. Background Technology

[0002] Existing technologies for generating and updating radiation source template libraries present numerous problems. For example, in clustering, current techniques often employ fixed clustering algorithm parameters, which cannot adapt to different data distributions, resulting in poor clustering performance and inaccurate segmentation of radiation source signal features, thus affecting the identification and analysis of radiation sources. Furthermore, traditional radiation source template libraries are typically static, generated based on offline modeling, making timely updates difficult once generated. This prevents the radiation source parameters in the template library from reflecting the latest signal characteristic changes and adapting to newly emerging radiation source devices or signal types. For instance, when a new radar system is put into use, traditional template libraries cannot quickly identify and incorporate its features, leading to insufficient identification capabilities for new radiation sources. To maintain the effectiveness of the template library, regular manual calibration and updates are required. This process is not only time-consuming and labor-intensive but also prone to human error, increasing the cost and complexity of template library maintenance. Summary of the Invention

[0003] In view of this, the present disclosure proposes a method for dynamically generating a radiation source template library, which solves the technical problem that previous radiation source template libraries were often static and could not reflect new radiation source characteristics in a timely manner. By adjusting the radiation source parameters in the radiation source template library according to the types of several clusters, a dynamic template library is generated, which enables the template library to be updated in a timely manner, thereby improving the accuracy, real-time performance, and adaptability of the radiation source detection system.

[0004] Specifically, the first embodiment of this disclosure discloses a method for dynamically generating a radiation source template library, the method specifically including the following steps: The first pulse descriptor sequence (PDW) matrix is ​​obtained by extracting parameters from the received electromagnetic wave signal, and the second pulse descriptor sequence (PDW) matrix is ​​generated in batches from the first pulse descriptor sequence (PDW) matrix according to the arrival time (TOA). The pulse width PW, radio frequency RF, and direction of arrival (DOA) in the second pulse descriptor sequence (PDW) matrix are preprocessed to obtain the third pulse descriptor sequence (PDW) matrix. Adaptively set the DBSCAN algorithm parameters and use the DBSCAN algorithm to perform single-batch radiation source clustering on the third pulse descriptor sequence PDW matrix; Several clusters are obtained based on the single batch radiation source clustering results. Incremental clustering is performed on the several clusters according to the merging logic of incremental clustering. The similarity judgment between the incremental clustering results of the several clusters and the clusters in the radiation source template library is used to obtain the types of the several clusters. The radiation source parameters in the radiation source template library are adjusted according to the types of the clusters to generate a dynamic template library; wherein the radiation source parameters are represented by the average value of the three-dimensional feature parameters in the clusters.

[0005] This disclosure further discloses that obtaining the first pulse descriptor sequence (PDW) matrix based on the parameters extracted from the received electromagnetic wave signal further includes: A pulse signal is detected from the electromagnetic wave signal by setting a threshold, and the arrival time TOA, pulse width PW, radio frequency RF, and direction of arrival angle DOA are extracted from the pulse signal to form a first pulse descriptor sequence PDW matrix.

[0006] This disclosure further discloses...

[0007] Where N is the number of PDW matrices that divide the first pulse descriptor sequence according to the arrival time TOA.

[0008] This disclosure further discloses that the data preprocessing of the pulse width PW, radio frequency RF, and direction of arrival (DOA) in the second pulse descriptor sequence (PDW) matrix further includes: The pulse width (PW), radio frequency (RF), and direction of arrival (DOA) values ​​are normalized to the range [0, 1] using the following formula:

[0009]

[0010]

[0011] In the formula, , , Represents the normalized three-dimensional feature parameters; , This represents the minimum and maximum values ​​of the pulse width PW. , This represents the minimum and maximum values ​​of the radio frequency (RF). , This indicates the minimum and maximum values ​​of the direction of arrival angle (DOA). , , These are the pulse width (PW), radio frequency (RF), and direction of arrival (DOA) in the second pulse descriptor sequence (PDW) matrix. .

[0012] This disclosure further discloses that the adaptive setting of DBSCAN algorithm parameters further includes: Adaptively set the neighborhood radius Eps and the number of neighborhood points Minpts; Among them, the number of neighborhood points D is the parameter dimension to be clustered; The neighborhood radius Eps is determined by the k-th nearest neighbor method, where k = Minpts-1. The k-th nearest neighbor distance value of each object is taken to form the (k-1)-th nearest neighbor distance vector. The (k-1)-th nearest neighbor distance vector is sorted from largest to smallest to obtain the distance curve. The inflection point of the distance curve is found, and Eps is calculated.

[0013] This disclosure further discloses...

[0014]

[0015] Where, diff() represents the difference between adjacent elements, and y and x represent the vertical and horizontal axis values ​​of the nearest distance curve elements. This approximates the slope of the distance curve at various points. This represents the change in slope, taking abs( The point value at the maximum is Eps.

[0016] This disclosure further discloses that the step of obtaining several clusters based on the clustering results of a single batch of radiation sources, performing incremental clustering on the several clusters according to the merging logic of incremental clustering, and determining the cluster types by comparing the incremental clustering results of the several clusters with the clusters in the radiation source template library further includes: The clustering results in the radiation source template library are denoted as k takes i clusters, 1, 2, ..., i, and the incremental clustering result of these clusters is denoted as k(i, 1, 2, ..., i). h represents j clusters, from 1 to j. The clustering result parameter ranges for pulse width PW, radio frequency RF, and direction of arrival (DOA) of each cluster in the radiation source template library are calculated respectively. , ,

[0017] The clustering result parameter ranges for the pulse width (PW), radio frequency (RF), and direction of arrival (DOA) of the aforementioned clusters are calculated respectively. , , ; calculate ,

[0018] Set threshold ,like Then and They were merged into one category.

[0019] This disclosure further discloses that, for clusters that meet the merging conditions and Create a new cluster to contain and All data points are included, and the pulse width (PW), radio frequency (RF), and direction of arrival (DOA) parameter ranges of the new clusters are recalculated. The radiation source template library is updated, and the old one is removed. and Replace with the new merged cluster.

[0020] This disclosure further discloses that, if , , Not satisfied The clustering result parameters of the pulse width PW, radio frequency RF, and direction of arrival (DOA) of the aforementioned clusters are then within the range. , , It is added as a new, independent cluster to the radiation source template library.

[0021] A second embodiment of this disclosure discloses a device for dynamically generating a radiation source template library, the device comprising: The first generating device is used to extract parameters based on the received electromagnetic wave signal to obtain a first pulse descriptor sequence (PDW) matrix, and to generate a second pulse descriptor sequence (PDW) matrix in batches based on the arrival time (TOA) of the first pulse descriptor sequence (PDW) matrix. The second generating device is used to perform data preprocessing on the pulse width PW, radio frequency RF, and direction of arrival DOA in the second pulse descriptor sequence PDW matrix to obtain the third pulse descriptor sequence PDW matrix. A computing device is used to adaptively set DBSCAN algorithm parameters and use the DBSCAN algorithm to perform single-batch radiation source clustering on the third pulse descriptor sequence PDW matrix; The determination device is used to obtain several clusters based on the clustering results of a single batch of radiation sources, perform incremental clustering on the several clusters according to the merging logic of incremental clustering, and determine the cluster types by comparing the incremental clustering results of the several clusters with the clusters in the radiation source template library. An updating device is used to adjust the radiation source parameters in the radiation source template library according to the types of the plurality of clusters to generate a dynamic template library; wherein the radiation source parameters are represented by the average value of the three-dimensional feature parameters in the plurality of clusters.

[0022] This disclosure obtains a first pulse descriptor sequence (PDW) matrix by extracting parameters from received electromagnetic wave signals; generates a second pulse descriptor sequence (PDW) matrix in batches based on the time of arrival (TOA); performs single-batch radiation source clustering on the third pulse descriptor sequence (PDW) matrix using an adaptive setting DBSCAN algorithm; obtains several clusters based on the single-batch radiation source clustering results; performs incremental clustering on the several clusters based on the merging logic of incremental clustering; determines the cluster types by comparing the incremental clustering results of the several clusters with the clusters in the radiation source template library based on similarity; and adjusts the radiation source parameters in the radiation source template library according to the types of the several clusters to generate a dynamic template library. It can process electromagnetic wave signals more efficiently, extract valuable feature information, and provide a more accurate data foundation for subsequent clustering and template library generation; it can better discover hidden patterns and structures in radiation source signal data, improve the accuracy and reliability of clustering, adjust radiation source parameters in the radiation source template library according to the type of several clusters, and update the template library in a timely manner so that it can adapt to the constantly changing radiation source environment, thereby improving the practicality and effectiveness of the template library and providing a more accurate reference for the identification and analysis of radiation sources.

[0023] The above description is merely an overview of the technical solution disclosed herein. In order to better understand the technical means of this disclosure and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this disclosure more apparent and understandable, preferred embodiments are described below in detail with reference to the accompanying drawings. Attached Figure Description

[0024] To more clearly illustrate the technical solutions of the embodiments of this disclosure, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this disclosure. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0025] Figure 1 A flowchart of a method for dynamically generating a radiation source template library provided in an embodiment of this disclosure.

[0026] Figure 2 A flowchart illustrating a method for generating a second pulse descriptor sequence (PDW) matrix based on a received electromagnetic wave signal, as provided in an embodiment of this disclosure.

[0027] Figure 3This is a flowchart illustrating a method for single-batch radiation source clustering using the DBSCAN algorithm on a third pulse descriptor sequence (PDW) matrix, as provided in an embodiment of this disclosure.

[0028] Figure 4 This is a schematic diagram of the structure of the dynamic generation device for the radiation source template library provided in the embodiments of this disclosure.

[0029] Figure 5 This is a schematic diagram illustrating the incremental clustering merging effect provided in an embodiment of this disclosure.

[0030] Figure 6 This is a schematic diagram of the structure of a computer device provided in an embodiment of the present disclosure. Detailed Implementation

[0031] The embodiments of this disclosure will now be described in detail with reference to the accompanying drawings.

[0032] It should be understood that the following specific examples illustrate the implementation of this disclosure, and those skilled in the art can easily understand other advantages and effects of this disclosure from the content disclosed in this specification. Obviously, the described embodiments are only a part of the embodiments of this disclosure, and not all of them. This disclosure can also be implemented or applied through other different specific implementation methods, and the details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of this disclosure. It should be noted that, in the absence of conflict, the following embodiments and features in the embodiments can be combined with each other. Based on the embodiments in this disclosure, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this disclosure.

[0033] It should be noted that the following description covers various aspects of embodiments within the scope of the appended claims. It will be apparent that the aspects described herein can be embodied in a wide variety of forms, and any particular structure and / or function described herein is merely illustrative. Based on this disclosure, those skilled in the art will understand that one aspect described herein can be implemented independently of any other aspect, and two or more of these aspects can be combined in various ways. For example, any number of aspects set forth herein can be used to implement the device and / or practice the method. Additionally, this device and / or method can be implemented using other structures and / or functionalities besides one or more of the aspects set forth herein.

[0034] It should also be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of this disclosure. The drawings only show the components related to this disclosure and are not drawn according to the number, shape and size of the components in actual implementation. In actual implementation, the form, quantity and proportion of each component can be arbitrarily changed, and the layout of the components may also be more complex.

[0035] Furthermore, specific details are provided in the following description to facilitate a thorough understanding of the examples. However, those skilled in the art will understand that the described aspects can be practiced without these specific details.

[0036] For detailed implementation procedures, please refer to the appendix. Figure 1 , attached Figure 1 A method for dynamically generating a radiation source template library is disclosed, which specifically includes the following steps.

[0037] Step S100: Listen to the radar signal, extract parameters based on the received electromagnetic wave signal to obtain the first pulse descriptor sequence (PDW) matrix, and generate the second pulse descriptor sequence (PDW) matrix in batches based on the arrival time (TOA).

[0038] Appendix Figure 2 The implementation process of step S100 is described in detail below: Step S1001: The reconnaissance receiver receives electromagnetic wave signals in real time and extracts parameters from the electromagnetic wave signals.

[0039] Reconnaissance receivers monitor airborne radar signals in real time via their antenna systems, receiving the electromagnetic wave signals emitted by the radar. These signals typically exist in pulse form, containing information such as the pulse's amplitude, frequency, and phase. The received analog signals are converted into digital signals by an analog-to-digital converter (ADC) for subsequent digital signal processing. The ADC's sampling rate needs to be high enough to ensure accurate reconstruction of the signal's details.

[0040] Step S1002: Extract parameters based on the received electromagnetic wave signal to obtain the first pulse descriptor sequence (PDW) matrix.

[0041] Detecting a single pulse signal from a digital signal. This is typically achieved by setting an amplitude threshold; when the signal amplitude exceeds this threshold, a pulse is considered detected. For each detected pulse, its key parameters are extracted, including Time of Arrival (TOA), Pulse Width (PW), Radio Frequency (RF), and Direction of Arrival (DOA). These parameters together constitute a First Pulse Descriptor Word (PDW). The extracted PDWs are stored in chronological order in a data structure (such as a list or array) to form a PDW matrix. Each PDW contains information such as TOA, PW, RF, and DOA.

[0042] Step S1003: Generate the second pulse descriptor sequence PDW matrix in batches from the first pulse descriptor sequence PDW matrix according to the arrival time TOA.

[0043] Time of Arrival (TOA) is a key parameter in the first pulse descriptor (PDW) matrix. Since the signal is received by the reconnaissance receiver in chronological order, batches can be divided according to rules such as time intervals or fixed quantities. For example, a time threshold can be set; if the TOA difference between two adjacent PDWs exceeds this threshold, they are considered to belong to different batches. Alternatively, each batch can be pre-defined to contain a fixed number of PDWs, such as 100 PDWs per batch. For time interval division, the determination of this time threshold can be based on experience, system real-time requirements, and signal environment factors. PDW data can be processed in batches according to the TOA order. Several methods can be used: Time interval division: Set a time threshold (e.g., 10 milliseconds). When the TOA difference between two adjacent PDWs exceeds this threshold, they are divided into different batches.

[0044] Fixed quantity division: Pre-set each batch to contain a fixed number of PDWs (e.g., 100), and then group the PDWs sequentially.

[0045] Dynamically adjust the batch size based on the signal density. For example, decrease the batch size in areas with dense signals and increase the batch size in areas with sparse signals.

[0046] The stored PDW sequences undergo preliminary preprocessing, including removing erroneous data and standardizing parameters. For example, PDWs with clearly reversed TOA order are removed, and the units of PW, RF, and DOA are standardized. The PDW sequences are then divided into multiple batches according to a defined batching method. Simultaneously, necessary standardization processing is performed on parameters such as PW, RF, and DOA, including unit unification and normalization, to ensure better processing of the data by the subsequent DBSCAN algorithm.

[0047] Based on the previously determined batching criteria, the stored PDW sequence is traversed and divided into different batches. For example, if a fixed number of PDWs is used, they are divided into batches of N sequentially; if the batching is based on a time interval, the TOA of each PDW is checked against the TOA of the previous PDW, and if the difference exceeds a set threshold, it is assigned to a new batch.

[0048] For each pre-defined batch, the DBSCAN algorithm is applied separately. This avoids the problems of excessive computational complexity and long computation time caused by processing a large amount of data at once. For example, for each batch, the PW, RF, and DOA parameters are extracted as input feature vectors, appropriate DBSCAN algorithm parameters are set (how to adaptively set these parameters will be discussed later), and then cluster analysis is performed to identify clusters formed by pulse signals emitted from the same radiation source.

[0049] The pulse descriptor sequence is represented as:

[0050] In the formula, N is the pulse descriptor sequence, and N is the number of PDWs. When N is too large, the DBSCAN algorithm becomes too complex and takes too long to compute. Therefore, the detected PDWs are processed in batches according to the TOA order.

[0051] Step S1004: Integrate the clustering results from different batches into the template library to construct a complete description of the radiation source signal features.

[0052] Pulse signals from the same radiation source may exist across different batches. Whether they belong to the same radiation source can be determined by comparing the characteristics of clusters in adjacent batches (such as average RF, average PW, DOA range, etc.). Integrating the clustering results from different batches into a template library constructs a complete description of the radiation source signal characteristics. This facilitates dynamic updates to the template library, improving the accuracy and real-time performance of the identification.

[0053] For each batch of clustering results, key features for comparison are extracted, including average RF, average PW, and DOA range. The features of clusters in adjacent batches are compared, and feature similarity indices, such as Euclidean distance or crossover ratio, are calculated to determine whether they belong to the same radiation source.

[0054] Clusters belonging to the same radiation source in adjacent batches are identified to ensure that these related clusters can be identified during subsequent integration. Clustering results of the same radiation source from different batches are integrated into a template library. Feature-weighted averaging or other fusion strategies can be used to construct a more comprehensive description of the radiation source signal characteristics.

[0055] After integrating clustering results from different batches, the template library is updated. Newly identified radiation sources are added, the feature parameters of existing radiation sources are updated, and sources that no longer appear or have been merged are deleted. Outdated or redundant information is cleaned up, and the data structure is optimized to improve query and update efficiency. A data decay mechanism can be introduced to reduce the weight of old data, making the template library more closely reflect the current signal environment.

[0056] From extracting the first pulse descriptor word sequence (PDW) matrix to batch generating the second pulse descriptor word sequence (PDW) matrix, and subsequent data preprocessing, electromagnetic wave signals can be processed more efficiently, extracting valuable feature information and providing a more accurate data foundation for subsequent clustering and template library generation. Through these steps, the detected PDW data can be processed in batches according to the TOA order, reducing computational complexity and improving the system's real-time performance and efficiency.

[0057] Step S200: Perform data preprocessing on the pulse width (PW), radio frequency (RF), and direction of arrival (DOA) in the second pulse descriptor sequence PDW matrix to obtain the third pulse descriptor sequence PDW matrix.

[0058] Extract the values ​​of three parameters—pulse width (PW), radio frequency (RF), and direction of arrival (DOA)—from the second pulse descriptor sequence (PDW) matrix. Each pulse corresponds to one PW, RF, and DOA value. Determine the range of each parameter and find the minimum (PW_min) and maximum (PW_max) values ​​among all PW values. Find the minimum (RF_min) and maximum (RF_max) values ​​among all RF values. Find the minimum (DOA_min) and maximum (DOA_max) values ​​among all DOA values.

[0059] Use the following formula to normalize the value of each parameter to the range [0, 1]:

[0060]

[0061]

[0062] In the formula, , , This represents the normalized three-dimensional feature parameters. , Indicates the maximum and minimum values ​​of the pulse width. , Indicates the maximum and minimum values ​​of the carrier frequency. , This represents the maximum and minimum values ​​of the angle of arrival.

[0063] The normalized PW, RF, and DOA values ​​are stored in a new data structure to obtain the third pulse descriptor sequence PDW matrix for subsequent cluster analysis. In real-time data processing scenarios, dynamic updates are required. , , , , The value is adjusted to adapt to changes in the data.

[0064] Step S300: Adaptively set the DBSCAN algorithm parameters and use the DBSCAN algorithm to perform single-batch radiation source clustering on the third pulse descriptor sequence PDW matrix. See Appendix for specific implementation methods. Figure 3 .

[0065] Step S3001: Adaptively set the DBSCAN algorithm parameters according to the actual situation.

[0066] DBSCAN is a density-based clustering algorithm capable of identifying clusters of arbitrary shapes and robust to noise. It primarily relies on two parameters: the neighborhood radius Eps and the number of neighboring points Minpts. The neighborhood radius Eps determines the reachable neighborhood from a given point; the number of neighboring points Minpts determines whether a point is a core point. If the number of points in a point's Eps neighborhood is greater than or equal to Minpts, then that point is considered a core point.

[0067] The value of Minpts is related to the dimension of the data. The empirical formula is Minpts = 2D, where D is the dimension of the parameters to be clustered. Here, the parameters to be clustered are pulse width (PW), radio frequency (RF), and direction of arrival (DOA), i.e., D = 3, so Minpts = 2³ = 6.

[0068] The Eps parameter is determined using the k-nearest neighbor method. The main steps are as follows: Traverse each object in the dataset, calculate the distance from each object to all other objects, and construct a distance matrix. Let k = Minpts - 1 = 5. Sort the distances of each object in ascending order, and take the 5th distance value to form the 5th nearest neighbor distance vector. Sort the 5th nearest neighbor distance vectors in descending order to form a distance curve. The inflection point of the distance curve reflects the change in data density; the distance value at this inflection point is a reasonable Eps value. Before the inflection point, the distances are larger, and the data points are sparser; after the inflection point, the distances are smaller, and the data points are denser. Selecting the distance at the inflection point as Eps can better balance the density of data points within clusters and the number of clusters during clustering. Excluding extreme points on the distance curve, the inflection point is found by calculating the rate of change of the slope of the distance curve. Specifically, we first calculate the slope k at each point on the curve, then calculate the change in slope dk, and take the point with the largest absolute value of the rate of change of slope, i.e., the point where abs(dk) is the largest, as Eps. This is because the point where the rate of change of slope is the largest means that the curve is the most curved, which is the location of the inflection point.

[0069] To find the inflection point of the distance curve, we calculate the rate of change of the slope at each point. First, we calculate the slope k at each point, approximating it using the difference method. Specifically, for two adjacent points (x1, y1) and (x2, y2), where x is the index of the data point and y is the corresponding 5th nearest neighbor distance value, the slope k ≈ (y2 - y1) / (x2 - x1). Since x is a consecutive integer, x2 - x1 = 1, therefore the slope k is equal to y2 - y1. By calculating the slope between adjacent points sequentially, we obtain a slope vector. Next, calculate the change in slope. Similarly, the difference method is used, that is This is the difference between two adjacent slope values. Thus, we obtain a vector representing the change in slope. Then, calculate this. Find the absolute value of each element in the vector, and then find abs( The position with the largest value. This position corresponds to the inflection point of the distance curve, because at this position, the curve's curvature is greatest and the slope change is most drastic. Finally, the 5th nearest neighbor distance value corresponding to this position is taken as the adaptively set Eps parameter. The adaptive setting of the Eps parameter, excluding the extreme points of the distance curve, selects the value at the point where the slope of the distance curve changes the most as Eps. Details are as follows:

[0070]

[0071] Step S3002: Since clustering is required for the three parameters PW, RF, and DOA, the parameter dimension D=3.

[0072] Before performing clustering operations, the raw data must be preprocessed to ensure that the pulse width (PW), radio frequency (RF), and direction of arrival (DOA) parameters are normalized. The goal of normalization is to transform parameters with different dimensions and numerical ranges into a unified scale, such as the [0, 1] interval. This eliminates the influence of differences in dimensions and numerical ranges between parameters, ensuring that each parameter plays a fair role in subsequent clustering analysis and preventing any single parameter from dominating operations such as distance calculations due to its larger value. Linear normalization can be used, where for each parameter, the original value is subtracted from its minimum value, and then divided by the difference between the maximum and minimum values ​​to obtain the normalized value.

[0073] According to empirical formulas We can calculate Minpts = 6. This formula is based on general cluster analysis experience. The value of Minpts is related to the dimensionality of the data, which can ensure that there are enough points in the neighborhood to form clusters with a certain density during the clustering process.

[0074] Step S3003: Calculate the 5th nearest neighbor distance vector. For each object in the dataset (i.e., each PDW data point), it is necessary to calculate its Euclidean distance to all other objects in the dataset. The Euclidean distance is calculated by taking the normalized values ​​of PW, RF, and DOA for objects i and j respectively, and then substituting them into the formula. , Calculate the distance between the two objects. After obtaining the distances from this object to all other objects, sort these distances in ascending order and take the 5th distance value (note that the index starts counting from 0, so it corresponds to the position with index 4). This value is the 5th nearest neighbor distance of this object. After traversing all objects, collect these 5th nearest neighbor distance values ​​to form the 5th nearest neighbor distance vector.

[0075] The elements in the obtained 5th nearest neighbor distance vector are sorted in descending order. The sorted result forms a distance curve, where the horizontal axis represents the index of the data points (integers starting from 0 and incrementing sequentially to identify the position of each data point in the sorted order), and the vertical axis represents the corresponding 5th nearest neighbor distance value. This curve visually shows the distribution of each object at its 5th nearest neighbor distance, providing a basis for subsequently determining the Eps parameter.

[0076] In this disclosure, after determining the two key parameters, Eps and Minpts, they can be substituted into the DBSCAN algorithm to perform cluster analysis on the current batch of radiation source data. The DBSCAN algorithm automatically divides the data into different clusters based on neighborhood relationships (the neighborhood range is determined by Eps) and density information (the lower density limit is determined by Minpts). Specifically, the algorithm starts by randomly selecting an unvisited point in the dataset and finds all points within the neighborhood of that point with Eps as the radius. If the number of points in the neighborhood is greater than or equal to Minpts, the point is marked as a core point, and a cluster is grown around it, incorporating all points in the neighborhood into this cluster. This process continues recursively, checking the neighborhoods of these neighboring points and continuously expanding the cluster until it can no longer be expanded. If the number of points in the neighborhood is less than Minpts, the point is marked as a boundary point or a noise point, depending on the specific situation. Ultimately, the points in the dataset are divided into different clusters, each corresponding to a set of pulse signals from the same radiation source, thus achieving effective clustering of a single batch of radiation sources. Throughout the implementation, the key is to accurately calculate the 5th nearest neighbor distance vector and reasonably determine the Eps parameter. This ensures that the DBSCAN clustering algorithm can correctly identify different radiation source clusters, improving the accuracy and efficiency of radiation source identification and analysis.

[0077] Normalizing the data before performing DBSCAN clustering is a crucial step. Because the numerical ranges of the parameters (PW, RF, DOA) can vary significantly, non-normalization may cause certain parameters to dominate distance calculations, affecting the accuracy of the clustering results. Euclidean distance is used here, but in practical applications, other distance metrics, such as Manhattan distance, can be chosen based on data characteristics and specific needs. Different distance metrics may have a certain impact on the clustering results, requiring selection and adjustment as needed. The DBSCAN algorithm is quite sensitive to the selection of parameters Eps and Minpts; different parameter combinations may lead to different clustering results. Adaptive parameter setting can improve the robustness and adaptability of the algorithm to some extent, but for complex data distributions, fine-tuning of parameters or other optimization strategies may still be necessary. When the data scale is large, calculating the k-th nearest neighbor distance matrix can be time-consuming. In this case, approximate nearest neighbor algorithms (such as tree-based or hash-based methods) can be considered to accelerate the calculation, or preprocessing operations such as data sampling can be performed to improve the computational efficiency of the algorithm.

[0078] Adaptive parameter settings for the DBSCAN algorithm can better uncover hidden patterns and structures in radiation source signal data, improving the accuracy and reliability of clustering. Single-batch radiation source clustering results can more realistically reflect the characteristic differences of radiation sources, providing a more reliable basis for subsequent template library generation.

[0079] Step S400: Analyze the clustering of a single batch of radiation sources to obtain several clusters, process the several clusters according to the merging logic of incremental clustering, and determine the type of the several clusters based on the similarity between the processing results of the several clusters and the clusters in the radiation source template library.

[0080] The first batch of detected radiation source data underwent preprocessing (such as normalization) and DBSCAN cluster analysis to obtain several clusters, each representing a group of radiation source signals with similar characteristics. For each cluster, the minimum and maximum values ​​of its three characteristic parameters (PW, RF, and DOA) were calculated to determine the three-dimensional characteristic parameter range of the cluster. These ranges describe the characteristic boundaries of the signals within the cluster. Each cluster and its corresponding three-dimensional characteristic parameter range were stored as initial entries in a template library. These entries constitute the basic data of the template library, and subsequent batches of clustering results will be compared and incrementally clustered based on this.

[0081] Based on the incremental clustering merging logic, in addition to batch processing, each batch of clustering results needs to be compared with the results in the radiation source template library. Incremental clustering is then implemented according to the merging logic, which is based on the intersection-union ratio (IUGR). The clustering results in the radiation source template library are denoted as... Let k be i clusters of type 1, 2, ..., i. Calculate and record the minimum and maximum values ​​of their three characteristic parameters PW, RF, and DOA to form the characteristic parameter range. , , The clustering results for the current batch are denoted as... h selects j clusters (1, 2, ..., j), calculates and records the minimum and maximum values ​​of their three characteristic parameters PW, RF, and DOA, forming the characteristic parameter range. , , Calculate the range of three-dimensional feature parameters for each cluster, i.e. , , , respectively denoted as , , , , , Calculate the intersection-union ratio, such as:

[0082]

[0083]

[0084] In the formula, The threshold value is 0.3, which we set here. , , These represent the similarity of the clustering results across each parameter dimension. If... Then and Merge them into one category. If , , All are greater than or equal to the threshold =0.3, then it is considered and If they are similar enough across all three feature dimensions, they can be merged into one class. Then, each cluster in the current batch of clustering results is processed sequentially. With each cluster in the radiation source template library Perform intersection, union, and comparison calculations and make judgments.

[0085] Step S500: Adjust the radiation source parameters in the radiation source template library according to the several cluster types to generate a dynamic template library; wherein the radiation source parameters are represented by the mean of the three-dimensional feature parameters in the cluster.

[0086] For clusters that meet the merging conditions and Create a new cluster to contain and All data points in the dataset are analyzed, and the characteristic parameter range of the new cluster is recalculated. The radiation source template library is updated, replacing the old one. and Replace with the new merged cluster. If If the conditions for merging with any cluster in the template library are not met, it is added as a new independent cluster to the radiation source template library.

[0087] After processing each new batch of radiation source data, new clustering results are obtained. Based on the incremental clustering merging logic, the clustering results of the new batch are compared and merged with the existing clustering results in the template library to form new clusters. For each updated cluster, the minimum and maximum values ​​of its three-dimensional feature parameters (PW, RF, DOA) are recalculated to determine the new feature parameter range. This step ensures that the feature parameter range of each cluster in the template library reflects the distribution of all relevant data.

[0088] For all PDW data points in each cluster, calculate the mean values ​​of PW, RF, and DOA parameters. Iterate through each PDW in the cluster, summing its PW value, and finally divide by the total number of PDWs in that cluster, N, to obtain the mean value of the PW parameter. Similarly, sum the RF and DOA values ​​separately and divide by N to obtain the mean values ​​of the RF and DOA parameters.

[0089] The mean value of the feature parameters for each cluster is updated in the template library, replacing the old feature parameter representation (both the feature parameter range and the mean value can be retained, depending on the specific needs). At this point, each cluster in the template library not only has a feature parameter range but also a description of the mean value, or the mean value can be directly used as the feature representative of the cluster.

[0090] For newly processed radiation sources, after incremental clustering, the radiation source parameters in the library are updated. The mean of the three-dimensional feature parameters in the cluster is used to represent the cluster. The template library outputs the parameters of each cluster, as follows:

[0091]

[0092]

[0093] In the formula, , , This represents the mean of the three-dimensional feature parameters of each cluster. This indicates the number of PDWs in a cluster.

[0094] The mean feature parameters (PW, RF, DOA) of each cluster in the template library are organized into a structured data format, such as a list or dictionary. For example, the mean parameters of each cluster can be combined into a tuple or dictionary, and then the data of all clusters can be collected into a list for easy output and use. A query interface is designed to allow external systems to query the corresponding mean feature parameters based on the cluster number or other identifier. In this way, other systems or modules can easily obtain information from the radiation source template library for further analysis, identification, or decision-making.

[0095] The template library is a large dictionary where the keys are unique identifiers for clusters (e.g., "cluster_1", "cluster_2", etc.), and the values ​​are another dictionary storing the detailed characteristic parameter ranges for that cluster. Each cluster's dictionary contains six key-value pairs, representing the minimum and maximum values ​​of the three characteristic parameters: PW, RF, and DOA. The keys are named "PW_min", "PW_max", "RF_min", "RF_max", "DOA_min", and "DOA_max", and the corresponding values ​​are the specific numerical values ​​for the respective parameters. For example, {"PW_min":0.2, "PW_max": 0.5, "RF_min": 1000, "RF_max": 1005, "DOA_min": 30, "DOA_max":35} represents the characteristic parameter range of a cluster, where the pulse width (PW) is between 0.2 and 0.5, the radio frequency (RF) is between 1000 and 1005, and the direction of arrival (DOA) is between 30 and 35 degrees.

[0096] Alternatively, an object-oriented data structure can be used, for example: Define the Cluster class: property: id: Represents a unique identifier for a cluster, such as "cluster_1", used to distinguish different clusters in the template library.

[0097] pw_min and pw_max represent the minimum and maximum pulse widths, respectively, defining the feature range of this cluster in the PW dimension.

[0098] rf_min and rf_max represent the minimum and maximum values ​​of the radio frequency, which determine the characteristic range of the cluster in the RF dimension.

[0099] doa_min and doa_max represent the minimum and maximum values ​​of the direction of arrival angle, which limit the feature range of the cluster in the DOA dimension.

[0100] `update_range`: Used to update the range of feature parameters of a cluster. When new data is merged into the cluster, the minimum and maximum values ​​of each feature parameter are recalculated and updated based on the new data.

[0101] get_range: Returns the range of feature parameters for this cluster, making it convenient to quickly obtain relevant parameters during comparison.

[0102] The template library is a list or dictionary used to store multiple Cluster object instances. When you need to find or manipulate a cluster, you can locate it in the list or dictionary by its cluster number or other identifier.

[0103] Both data structures have their advantages. The dictionary structure is simple and intuitive, facilitating quick lookups and updates; the class object structure, on the other hand, is more extensible and encapsulated, making it suitable for use as system functionality becomes increasingly complex. In practical applications, the appropriate data structure can be chosen based on factors such as specific project requirements, the characteristics of the programming language, and the habits of the development team.

[0104] In this way, the dynamic template library can reflect the characteristics of newly processed radiation source data in real time and provide the characteristic information of radiation sources in a concise and clear mean form, which facilitates subsequent radiation source identification and analysis.

[0105] Adjusting the radiation source parameters in the radiation source template library according to the types of several clusters, and representing the radiation source parameters with the mean of the three-dimensional feature parameters in the clusters, enables timely updates to the template library, allowing it to adapt to the constantly changing radiation source environment. This improves the practicality and effectiveness of the template library and provides a more accurate reference for the identification and analysis of radiation sources.

[0106] The second embodiment of this disclosure discloses a device for dynamically generating a radiation source template library, such as... Figure 4 As shown, the device includes: The first generating device 601 is used to extract parameters based on the received electromagnetic wave signal to obtain a first pulse descriptor sequence (PDW) matrix, and to generate a second pulse descriptor sequence (PDW) matrix in batches based on the arrival time (TOA) of the first pulse descriptor sequence (PDW) matrix. The second generating device 602 is used to perform data preprocessing on the pulse width PW, radio frequency RF, and direction of arrival DOA in the second pulse descriptor sequence PDW matrix to obtain the third pulse descriptor sequence PDW matrix. Computing device 603 is used to adaptively set DBSCAN algorithm parameters and use the DBSCAN algorithm to perform single-batch radiation source clustering on the third pulse descriptor sequence PDW matrix; The judgment device 604 is used to obtain several clusters based on the clustering results of a single batch of radiation sources, perform incremental clustering on the several clusters according to the merging logic of incremental clustering, and judge the similarity between the incremental clustering results of the several clusters and the clusters in the radiation source template library to obtain the type of the several clusters. The updating device 605 is used to adjust the radiation source parameters in the radiation source template library according to the types of the plurality of clusters to generate a dynamic template library; wherein the radiation source parameters are represented by the average value of the three-dimensional feature parameters in the plurality of clusters.

[0107] The following is in conjunction with the appendix Figure 5 The invention will be further described in detail below with specific embodiments. In this embodiment, the pulse descriptor word (PDW) generated by the simulated reconnaissance receiver is generated using the dynamic generation system of the radiation source template library of the present invention. The specific steps are as follows: Step 1: The radiation source pulse descriptor generated by the reconnaissance receiver is represented as:

[0108] In the formula, N is the pulse descriptor sequence, and N is the number of PDWs. When N is too large, the DBSCAN algorithm becomes too complex and takes too long to compute. Therefore, the detected PDWs are processed in batches according to the TOA order.

[0109] The characteristic parameters of the radiation sources are shown in Table 1. The simulated reconnaissance receiver received seven radiation sources, and the PDW data densities generated by different radiation sources were different. To simulate the actual situation, errors should be added to the parameters. The radiation sources were set to have a certain degree of aliasing in the three clustering and sorting parameters of RF, PW, and DOA. The PDW of all pulses was calculated, and the data was normalized. Then, the normalized frequency, pulse width, and angle of arrival were used as the feature values ​​of the clustering part.

[0110] To simulate batch processing, the radiation sources processed in the first batch were numbered 1, 2, 3, and 4; the radiation sources processed in the second batch were numbered 1, 5, 6, and 7.

[0111] Table 1 Radiation source parameter settings

[0112] Step 2: Select pulse width, frequency, and angle of arrival as feature parameters, normalize the data, and adjust the range of feature parameter values ​​to ensure a clear two-dimensional display effect.

[0113] Step 3: Adaptively determine the DBSCAN algorithm parameters and perform cluster analysis on the preprocessed radiation source data to identify radiation sources with similar characteristics. The DBSCAN algorithm can effectively handle noisy data and automatically determine the number of clusters. Considering that the feature parameter dimension of the radiation source dataset is 3, the number of neighborhood points (Minpts) is set to 6. Each object in the dataset is traversed, and the distance from each object to all other objects is calculated to obtain a distance matrix. k = Minpts - 1 is taken, and the k-th nearest neighbor distance value of each object is used to form a k-nearest neighbor distance vector. The k-nearest neighbor distance vectors are sorted from largest to smallest to obtain a distance curve. The inflection point of the distance curve is found, Eps is calculated, and it is substituted into the DBSCAN algorithm to cluster the radiation source data. The clustering results of the first batch of radiation sources are as follows: Figure 3 As shown, the clustering effect of the second batch of radiation sources is as follows: Figure 4 As shown; Step 4: Based on the incremental clustering merging logic, in addition to batch processing, the clustering results of each batch need to be compared with the results in the radiation source template library. Incremental clustering is then implemented according to the merging logic. The merging logic is based on intersection-union ratio (IUU), as follows: The clustering results in the radiation source template library are denoted as Let k be 1, 2, ..., i, representing i clusters in total. The clustering result of the current batch is denoted as... h represents j clusters, 1, 2, ..., j; calculate the range of three-dimensional feature parameters for each cluster, i.e. , , , respectively denoted as , , , , , Calculate the intersection-union ratio, such as:

[0114]

[0115]

[0116] In the formula, The threshold value is 0.3, which we set here. , , These represent the similarity of the clustering results across each parameter dimension. If... Then and They were merged into one category.

[0117] When processing the first batch of radiation source PDWs, there were no historical clustering results in the template library, so the first batch of clustering results were directly added to the template library. When processing the second batch of radiation source PDWs, the results need to be compared with the template library, i.e., the first batch of clustering results, and batch merging and batch addition processing is performed.

[0118] Step 5: Adjust radiation source parameters in real time to form a dynamic template library. For newly processed radiation sources, perform incremental clustering and update the radiation source parameters in the library. The mean of the three-dimensional feature parameters in the cluster is used to represent the cluster. The template library outputs the parameters of each cluster, as follows:

[0119]

[0120]

[0121] In the formula, , , This represents the mean of the three-dimensional feature parameters of each cluster. This indicates the number of PDWs in a cluster. The output dynamic template library parameters are shown in Table 2.

[0122] Table 2 Radiation Source Parameters of Dynamic Template Library

[0123] A computer device according to embodiments of the present disclosure includes a memory and a processor. The memory is used to store non-transitory computer-readable instructions. Specifically, the memory may include one or more computer program products, which may include various forms of computer-readable storage media, such as volatile memory and / or non-volatile memory. The volatile memory may, for example, include random access memory (RAM) and / or cache memory. The non-volatile memory may, for example, include read-only memory (ROM), hard disk, flash memory, etc.

[0124] The processor may be a central processing unit (CPU) or other form of processing unit with data processing capabilities and / or instruction execution capabilities, and may control other components in the computer device to perform desired functions. In one embodiment of this disclosure, the processor is used to execute computer-readable instructions stored in the memory, causing the computer device to perform all or part of the steps of the learning outcome prediction method based on learning behavior data mining of the foregoing embodiments of this disclosure.

[0125] Those skilled in the art will understand that, in order to solve the technical problem of how to achieve a good user experience, this embodiment may also include well-known structures such as communication buses and interfaces, and these well-known structures should also be included within the protection scope of this disclosure.

[0126] like Figure 6 This is a schematic diagram of a computer device provided for an embodiment of the present disclosure. It illustrates a structural schematic diagram suitable for implementing the computer device in the embodiments of the present disclosure.

[0127] like Figure 6 As shown, a computer device may include a processor (such as a central processing unit, graphics processing unit, etc.), which can perform various appropriate actions and processes based on programs stored in read-only memory (ROM) or programs loaded from storage devices into random access memory (RAM). The RAM also stores various programs and data required for the operation of the computer device. The processor, ROM, and RAM are interconnected via a bus. Input / output (I / O) interfaces are also connected to the bus.

[0128] Typically, the following devices can be connected to the I / O interface: input devices, such as sensors or visual information acquisition devices; output devices, such as displays; storage devices, such as magnetic tapes or hard drives; and communication devices. Communication devices allow the computer device to communicate wirelessly or wiredly with other devices (such as edge computing devices) to exchange data. Although a computer device with various devices is illustrated, it should be understood that it is not required to implement or possess all of the devices shown. More or fewer devices may be implemented or possessed alternatively.

[0129] In particular, according to embodiments of this disclosure, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of this disclosure include a computer program product comprising a computer program carried on a non-transitory computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device, or installed from a storage device, or installed from a ROM. When the computer program is executed by a processor, all or part of the steps of the learning outcome prediction method based on learning behavior data mining according to embodiments of this disclosure are performed.

[0130] For a detailed description of this embodiment, please refer to the corresponding descriptions in the foregoing embodiments, which will not be repeated here.

[0131] A computer-readable storage medium according to embodiments of the present disclosure stores non-transitory computer-readable instructions. When the non-transitory computer-readable instructions are executed by a processor, all or part of the steps of the learning outcome prediction method based on learning behavior data mining according to the foregoing embodiments of the present disclosure are performed.

[0132] The aforementioned computer-readable storage media include, but are not limited to: optical storage media (e.g., CD-ROM and DVD), magneto-optical storage media (e.g., MO), magnetic storage media (e.g., magnetic tape or portable hard drive), media with built-in rewritable non-volatile memory (e.g., memory card), and media with built-in ROM (e.g., ROM cartridge).

[0133] For a detailed description of this embodiment, please refer to the corresponding descriptions in the foregoing embodiments, which will not be repeated here.

[0134] The basic principles of this disclosure have been described above with reference to specific embodiments. However, it should be noted that the advantages, benefits, and effects mentioned in this disclosure are merely examples and not limitations, and should not be considered as essential features of each embodiment of this disclosure. Furthermore, the specific details disclosed above are for illustrative and facilitative purposes only, and are not limitations. These details do not limit the scope of this disclosure to the necessity of employing the aforementioned specific details for implementation.

[0135] In this disclosure, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. The block diagrams of devices, apparatuses, devices, and systems involved in this disclosure are merely illustrative examples and are not intended to require or imply that they must be connected, arranged, or configured in the manner shown in the block diagrams. As those skilled in the art will recognize, these devices, apparatuses, devices, and systems can be connected, arranged, and configured in any manner. Words such as "comprising," "including," "having," etc., are open-ended terms meaning "including but not limited to," and are used interchangeably with them. The terms "or" and "and" as used herein refer to the terms "and / or," and are used interchangeably with them unless the context clearly indicates otherwise. The term "such as" as used herein refers to the phrase "such as but not limited to," and is used interchangeably with it.

[0136] Additionally, as used herein, the "or" used in a list of items beginning with "at least one" indicates a separate list, such that a list of, for example, "at least one of A, B, or C" means A or B or C, or AB or AC or BC, or ABC (i.e., A and B and C). Furthermore, the word "exemplary" does not imply that the described example is preferred or better than other examples.

[0137] It should also be noted that in the systems and methods of this disclosure, the components or steps can be decomposed and / or recombined. These decompositions and / or recombinations should be considered as equivalent solutions to this disclosure.

[0138] Various changes, substitutions, and modifications can be made to the technology described herein without departing from the teachings defined by the appended claims. Furthermore, the scope of the claims of this disclosure is not limited to the specific aspects of the processes, machines, manufactures, events, means, methods, and actions described above. Currently existing or later-developed processes, machines, manufactures, events, means, methods, or actions that perform substantially the same function or achieve substantially the same result as the corresponding aspects described herein can be utilized. Therefore, the appended claims include such processes, machines, manufactures, events, means, methods, or actions within their scope.

[0139] The above description of the disclosed aspects is provided to enable any person skilled in the art to make or use this disclosure. Various modifications to these aspects will be readily apparent to those skilled in the art, and the general principles defined herein may be applied to other aspects without departing from the scope of this disclosure. Therefore, this disclosure is not intended to be limited to the aspects shown herein, but rather to be carried out within the widest scope consistent with the principles and novel features disclosed herein.

[0140] The above description has been given for purposes of illustration and description. Furthermore, this description is not intended to limit the embodiments of this disclosure to the forms disclosed herein. Although numerous exemplary aspects and embodiments have been discussed above, those skilled in the art will recognize certain variations, modifications, alterations, additions, and sub-combinations therein.

Claims

1. A method for dynamically generating a radiation source template library, characterized in that, The method for dynamically generating the radiation source template library includes: The first pulse descriptor sequence (PDW) matrix is ​​obtained by extracting parameters from the received electromagnetic wave signal, and the second pulse descriptor sequence (PDW) matrix is ​​generated in batches from the first pulse descriptor sequence (PDW) matrix according to the arrival time (TOA). The pulse width PW, radio frequency RF, and direction of arrival (DOA) in the second pulse descriptor sequence (PDW) matrix are preprocessed to obtain the third pulse descriptor sequence (PDW) matrix. Adaptively set the DBSCAN algorithm parameters and use the DBSCAN algorithm to perform single-batch radiation source clustering on the third pulse descriptor sequence PDW matrix; Several clusters are obtained based on the single batch radiation source clustering results. Incremental clustering is performed on the several clusters according to the merging logic of incremental clustering. The similarity judgment between the incremental clustering results of the several clusters and the clusters in the radiation source template library is used to obtain the types of the several clusters. The radiation source parameters in the radiation source template library are adjusted according to the types of the clusters to generate a dynamic template library; wherein the radiation source parameters are represented by the average value of the three-dimensional feature parameters in the clusters.

2. The method according to claim 1, characterized in that, The step of extracting parameters based on the received electromagnetic wave signal to obtain the first pulse descriptor sequence (PDW) matrix includes: A pulse signal is detected from the electromagnetic wave signal by setting a threshold, and the arrival time TOA, pulse width PW, radio frequency RF, and direction of arrival angle DOA are extracted from the pulse signal to form a first pulse descriptor sequence PDW matrix.

3. The method according to claim 2, characterized in that: Where N is the number of PDW matrices that divide the first pulse descriptor sequence according to the arrival time TOA.

4. The method according to claim 1, characterized in that, The data preprocessing of the pulse width PW, radio frequency RF, and direction of arrival (DOA) in the second pulse descriptor word sequence (PDW) matrix further includes: The pulse width (PW), radio frequency (RF), and direction of arrival (DOA) values ​​are normalized to the range [0, 1] using the following formula: In the formula, , , Represents the normalized three-dimensional feature parameters; , This represents the minimum and maximum values ​​of the pulse width PW. , This represents the minimum and maximum values ​​of the radio frequency (RF). , This indicates the minimum and maximum values ​​of the direction of arrival angle (DOA). , , These are the pulse width (PW), radio frequency (RF), and direction of arrival (DOA) in the second pulse descriptor sequence (PDW) matrix. .

5. The method according to any one of claims 1-4, characterized in that, The adaptive setting of DBSCAN algorithm parameters further includes: Adaptively set the neighborhood radius Eps and the number of neighborhood points Minpts; Among them, the number of neighborhood points D is the parameter dimension to be clustered; The neighborhood radius Eps is determined by the k-th nearest neighbor method, where k = Minpts-1. The k-th nearest neighbor distance value of each object is taken to form the (k-1)-th nearest neighbor distance vector. The (k-1)-th nearest neighbor distance vector is sorted from largest to smallest to obtain the distance curve. The inflection point of the distance curve is found, and Eps is calculated.

6. The method according to claim 5, characterized in that: Where, diff() represents the difference between adjacent elements, and y and x represent the vertical and horizontal axis values ​​of the nearest distance curve elements. This approximates the slope of the distance curve at various points. This represents the change in slope, taking abs( The point value at the maximum is Eps.

7. The method according to claim 6, characterized in that, The step of obtaining several clusters based on the single-batch radiation source clustering results, performing incremental clustering on the several clusters according to the incremental clustering merging logic, and determining the cluster types by comparing the incremental clustering results of the several clusters with the clusters in the radiation source template library based on similarity judgment further includes: The clustering results in the radiation source template library are denoted as k takes i clusters, 1, 2, ..., i, and the incremental clustering result of these clusters is denoted as k(i, 1, 2, ..., i). h represents j clusters, from 1 to j. The clustering result parameter ranges for pulse width PW, radio frequency RF, and direction of arrival (DOA) of each cluster in the radiation source template library are calculated respectively. , , ; The clustering result parameter ranges for the pulse width (PW), radio frequency (RF), and direction of arrival (DOA) of the aforementioned clusters are calculated respectively. , , ; calculate , Set threshold ,like Then and They were merged into one category.

8. The method according to claim 7, characterized in that: For clusters that meet the merging conditions and Create a new cluster to contain and All data points are included, and the pulse width (PW), radio frequency (RF), and direction of arrival (DOA) parameter ranges of the new clusters are recalculated. The radiation source template library is updated, and the old one is removed. and Replace with the new merged cluster.

9. The method according to claim 7, characterized in that: if , , Not satisfied The clustering result parameters of the pulse width PW, radio frequency RF, and direction of arrival (DOA) of the aforementioned clusters are then within the range. , , It is added as a new, independent cluster to the radiation source template library.

10. A device for dynamically generating a radiation source template library, the device comprising: The first generating device is used to extract parameters based on the received electromagnetic wave signal to obtain a first pulse descriptor sequence (PDW) matrix, and to generate a second pulse descriptor sequence (PDW) matrix in batches based on the arrival time (TOA) of the first pulse descriptor sequence (PDW) matrix. The second generating device is used to perform data preprocessing on the pulse width PW, radio frequency RF, and direction of arrival DOA in the second pulse descriptor sequence PDW matrix to obtain the third pulse descriptor sequence PDW matrix. A computing device is used to adaptively set DBSCAN algorithm parameters and use the DBSCAN algorithm to perform single-batch radiation source clustering on the third pulse descriptor sequence PDW matrix; The determination device is used to obtain several clusters based on the clustering results of a single batch of radiation sources, perform incremental clustering on the several clusters according to the merging logic of incremental clustering, and determine the cluster types by comparing the incremental clustering results of the several clusters with the clusters in the radiation source template library. An updating device is used to adjust the radiation source parameters in the radiation source template library according to the types of the plurality of clusters, and generate a dynamic template library; wherein the radiation source parameters are represented by the average value of the three-dimensional feature parameters in the plurality of clusters.

Images