Unmanned aerial vehicle frequency hopping signal detection and sorting method, device, equipment and medium

Abstract

The invention discloses a method, a device, equipment and a medium for detecting and sorting frequency hopping signals of an unmanned aerial vehicle, and relates to the technical field of signal detection. And different clustering centers divide data objects in a certain radius range into different classes by taking the distance as a reference. Compared with a traditional K-means algorithm, when unmanned aerial vehicle frequency hopping signal detection data is processed, an initial k supervision value does not need to be set, self-adaptive classification is directly carried out through elements in a data set, and the influence on a data sorting result by setting an initial manually determined k classification value is avoided; and secondly, the clustering radius weight is adopted to balance the value, so that the clustering result is more reasonable, and non-spherical clusters can be processed.

Description

Technical Field

[0001] This invention relates to the field of signal detection technology, and in particular to a method, apparatus, equipment and medium for detecting and sorting UAV frequency hopping signals based on improved K-means clustering. Background Technology

[0002] With the development of the low-altitude economy, the effective control of drones and other flying equipment has become crucial for low-altitude safety. Currently, drones on the market generally use spread spectrum hopping technology to improve the anti-interference capability between the drone and the remote controller. Therefore, the ability to detect drone frequency hopping signals in real time is fundamental to effective countermeasures against drones. Current methods for detecting drone frequency hopping signals mainly include time continuity detection, connected region detection, bandwidth detection, waveform feature detection, energy detection, dwell time detection, and time sequence sorting detection. The data obtained through these methods needs to be classified and parameter estimated. The results serve as the identification feature data for drone frequency hopping signals. Detection parameters are then compared in subsequent electromagnetic environments. When the parameter estimation is within the allowable error range, the signal in that electromagnetic environment is identified as the drone signal type in the database. Therefore, the effective classification of drone frequency hopping signal detection data is the key and foundation for drone detection. Currently, commonly used algorithms for clustering analysis of data include distance-based clustering (K-means, hierarchical clustering), density-based clustering (DBSCAN), and model-based clustering (Gaussian mixture distribution). This design uses an improved K-means algorithm to perform clustering analysis on data obtained from UAV frequency hopping signal detection.

[0003] like Figure 7 As shown, analysis of UAV signal characteristics reveals that UAV frequency-hopping signals exhibit fixed patterns of variation with other signals, show significant similarity among individual signals, and contain a large number of individual signals within the sampling time. Therefore, clustering algorithms can be used to cluster the characteristics of UAV frequency-hopping signals for detection. Then, based on the classification results, parameter estimation of the UAV frequency-hopping signals can be performed on each category of signals.

[0004] K-means clustering is a classic algorithm for solving general data analysis problems. It is characterized by its simplicity and speed of computation, maintaining good processing efficiency even with large datasets, making it widely used in pattern recognition and machine learning. Its basic idea is as follows: First, determine the number of clusters k. Then, randomly select k cluster centers from the set. Calculate the distance from the remaining elements to the k initial cluster centers using Euclidean distance, and assign the data to the cluster of the nearest cluster center. Next, use the average value of the data elements in the new cluster as the new cluster center, and then classify all data elements again. Compare the classification results of the two clustering steps. If the data's category has not changed, the algorithm terminates; otherwise, it continues.

[0005] However, the traditional K-means clustering algorithm requires specifying the K value in advance. The K value represents the number of clusters and must be determined manually. A poorly chosen K value can severely affect the results; for example, it might forcibly divide data that should be divided into two clusters into four. The K-means clustering algorithm is sensitive to initial cluster centers and outliers. Randomly selecting initial cluster centers may lead to a local optimum, while outliers can skew the cluster centers, resulting in inaccurate classification. Summary of the Invention

[0006] In view of the above problems, the present invention provides a method, apparatus, device and medium for detecting and sorting frequency-hopping signals of unmanned aerial vehicles (UAVs) to overcome the above problems or at least partially solve the above problems.

[0007] This invention provides the following solution: A method for detecting and sorting frequency-hopping signals from unmanned aerial vehicles (UAVs), comprising: Raw spectrum data is collected, and signal thresholds are obtained through adaptive estimation based on the distribution characteristics of noise and signal data. The original spectrum data is processed based on the signal threshold to obtain a set of frequency point signals; An improved K-means clustering analysis method is used to cluster the frequency signal set to obtain clustering results. The improved K-means clustering analysis method includes dynamically determining the cluster radius based on the minimum distance set between data objects; calculating the density of each object using the cluster radius as a scale, and selecting the one with the largest density as the cluster center for category division; iterating this process until all objects are classified to obtain the clustering results. Based on the clustering results, parameter estimation is performed on each signal cluster to complete the detection and sorting of frequency hopping signals.

[0008] Preferably, the signal threshold estimation method includes: For a frame of spectrum data, divide it into intervals according to its level value and count the number of data points in each interval; Find the peak interval with the most data points; From the peak range towards the direction of increasing level value, find the first trough along the data point distribution curve; The level value corresponding to the first trough is determined as the signal threshold.

[0009] Preferably, the method for determining the cluster radius includes: Calculate the Euclidean distance between all pairs of data objects in the frequency signal set; Obtain the minimum distance from each data object to other objects, and form a set of minimum distances; Take the minimum non-zero value in the set. and maximum value The cluster radius is obtained by calculating the following formula:

[0010] In the formula: This represents the preset weighting coefficient. .

[0011] Preferably: the weighting coefficient .

[0012] Preferably, selecting the cluster center with the highest density includes: For the current subset of data to be clustered, calculate the number of other objects contained in the neighborhood of each data object with the cluster radius as the radius, define the number as the density of the object, and select the object with the highest density as the cluster center of the current round.

[0013] Preferably, during the iteration process, after each successful classification of a category, all data objects contained in that category are removed from the set to be clustered. The distances are recalculated based on the remaining data objects, and new cluster radii and cluster centers are determined for the next round of clustering.

[0014] Preferably, the parameter estimation includes calculating the statistical characteristics of each signal cluster in the time domain, frequency domain, and energy domain, wherein the statistical characteristics include one or more of the following: average center frequency, frequency hopping period, average bandwidth, and average signal energy.

[0015] A UAV frequency hopping signal detection and sorting device, characterized in that it is used to perform the above-described UAV frequency hopping signal detection and sorting method, the device comprising: The signal threshold estimation unit is used to acquire raw spectrum data and perform adaptive estimation to obtain the signal threshold based on the distribution characteristics of noise and signal data. A frequency point signal set calculation unit is used to process the original spectrum data based on the signal threshold to obtain a frequency point signal set; A clustering unit is used to cluster the frequency signal set using an improved K-means clustering analysis method to obtain clustering results. The improved K-means clustering analysis method includes dynamically determining the cluster radius based on the minimum distance set between data objects; calculating the density of each object using the cluster radius as a scale, and selecting the object with the highest density as the cluster center for category division; iterating this process until all objects are classified to obtain the clustering results. The detection and sorting unit is used to perform parameter estimation for each signal cluster based on the clustering results, thereby completing the detection and sorting of frequency hopping signals.

[0016] A frequency-hopping signal detection and sorting device for unmanned aerial vehicles (UAVs), the device comprising a processor and a memory: The memory is used to store program code and transmit the program code to the processor; The processor is used to execute the above-described UAV frequency hopping signal detection and sorting method according to the instructions in the program code.

[0017] A computer-readable storage medium for storing program code for executing the above-described UAV frequency hopping signal detection and sorting method.

[0018] According to specific embodiments provided by the present invention, the present invention discloses the following technical effects: This application provides a method, apparatus, device, and medium for detecting and sorting UAV frequency-hopping signals. During clustering, a density-based approach is used to select cluster centers. Different cluster centers are then used to classify data objects within a certain radius into different classes based on distance. Compared to the traditional K-means algorithm, this approach eliminates the need for setting an initial k-supervised value when processing UAV frequency-hopping signal detection data. It directly performs adaptive classification using elements in the dataset, avoiding the impact of manually determining an initial k-classification value on the data sorting results. Secondly, it employs a cluster radius weighting method, resulting in more reasonable clustering results and the ability to handle non-spherical clusters. Of course, any product implementing this invention does not necessarily need to achieve all of the advantages described above at the same time. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly described below. Obviously, the drawings described below are merely some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without any creative effort.

[0020] Figure 1This is a flowchart of a method for detecting and sorting frequency-hopping signals of unmanned aerial vehicles (UAVs) according to an embodiment of the present invention; Figure 2 This is a flowchart of the UAV frequency hopping signal detection process provided in an embodiment of the present invention; Figure 3 This is a schematic diagram of the spectrum data distribution characteristics provided in an embodiment of the present invention; Figure 4 This is a flowchart of the improved clustering algorithm provided in an embodiment of the present invention; Figure 5 This is a schematic diagram of a UAV frequency hopping signal detection and sorting device provided in an embodiment of the present invention; Figure 6 This is a schematic diagram of a UAV frequency hopping signal detection and sorting device provided in an embodiment of the present invention; Figure 7 This is a flowchart of the existing technology for UAV frequency hopping signal detection and parameter estimation based on cluster analysis. Detailed Implementation

[0021] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention are within the scope of protection of the present invention.

[0022] See Figure 1 This invention provides a method for detecting and sorting frequency-hopping signals from unmanned aerial vehicles (UAVs), such as... Figure 1 As shown, the method may include: S101: Acquire raw spectrum data, and obtain the signal threshold through adaptive estimation based on the distribution characteristics of noise and signal data; in specific implementation, the signal threshold estimation method provided in this application embodiment may include: For a frame of spectrum data, divide it into intervals according to its level value and count the number of data points in each interval; Find the peak interval with the most data points; From the peak range towards the direction of increasing level value, find the first trough along the data point distribution curve; The level value corresponding to the first trough is determined as the signal threshold.

[0023] S102: Process the original spectrum data based on the signal threshold to obtain a set of frequency point signals; S103: An improved K-means clustering analysis method is used to cluster the frequency signal set to obtain clustering results; the improved K-means clustering analysis method includes dynamically determining the cluster radius based on the minimum distance set between data objects; calculating the density of each object using the cluster radius as a scale, and selecting the one with the largest density as the cluster center for category division; iterating this process until all objects are classified to obtain the clustering results; in specific implementation, the method for determining the cluster radius in this application embodiment may include: Calculate the Euclidean distance between all pairs of data objects in the frequency signal set; Obtain the minimum distance from each data object to other objects, and form a set of minimum distances; Take the minimum non-zero value in the set. and maximum value The cluster radius is obtained by calculating the following formula:

[0024] In the formula: This represents the preset weighting coefficient. .

[0025] Furthermore, the weighting coefficients .

[0026] The cluster centers selected from those with the highest density include: For the current subset of data to be clustered, calculate the number of other objects contained in the neighborhood of each data object with the cluster radius as the radius, define the number as the density of the object, and select the object with the highest density as the cluster center of the current round.

[0027] During the iteration process, after each successful classification of a category, all data objects contained in that category are removed from the set to be clustered. The distances are recalculated based on the remaining data objects, and new cluster radii and cluster centers are determined for the next round of clustering.

[0028] S104: Based on the clustering results, perform parameter estimation for each signal cluster to complete the detection and sorting of frequency hopping signals.

[0029] The UAV frequency hopping signal detection and sorting method provided in this application selects cluster centers from a dataset and calculates the cluster radius. The dataset is then clustered using this radius. Compared to the traditional K-means method, its innovation lies in selecting cluster centers from data objects without needing to set a k value, thus avoiding the influence of setting an initial k-classification value on the data sorting results. Furthermore, for frequency hopping signal data, this method can quickly and effectively classify the processed data to a certain extent. Through analysis of the key points of UAV frequency hopping signal detection, clustering analysis can be performed based on the similar residence time, bandwidth, energy characteristics, and waveform morphology characteristics of each frequency hopping signal generated during UAV communication.

[0030] The following is a detailed description of the UAV frequency hopping signal detection and sorting method provided in this application.

[0031] Because UAV frequency-hopping signals are characterized by short, frequent bursts and numerous commonalities among themselves, the detected signals can be clustered and sorted based on these shared characteristics. By filtering signals layer by layer using the features of the frequency-hopping signals, the final extracted signals are solely those of frequency-hopping signals. This algorithm employs density-based characterization to select cluster centers, and different cluster centers further divide data objects within a certain radius into different classes based on distance.

[0032] Detecting frequency-hopping signals requires step-by-step detection based on the characteristics of UAV frequency-hopping signals. The specific process for detecting UAV frequency-hopping signals is as follows: Figure 2 As shown.

[0033] (1) Analysis of the distribution of spectral data points.

[0034] For example, in the 2.4GHz-2.48GHz frequency band, the distribution characteristics of spectrum data levels are quite obvious: the noise region is basically concentrated between -15dB and 20dB, and the signal level is significantly higher than the level values ​​of most data points in a frame of spectrum data. According to the definition in ITU-RP372 Recommendation, radio noise follows a normal distribution under certain conditions. That is, the noise distribution is most concentrated in the densest areas, and the number of noise points gradually decreases on both sides of the densest areas. Downwards, there is always noise, while upwards, the data gradually transitions from noise to signal data. Therefore, the data distribution of a frame of spectrum data has the following characteristics, such as... Figure 3 As shown.

[0035] By statistically analyzing the spectral data points along the level value axis, we obtain the noise data point distribution curve in a frame of spectral data, as shown in the figure. Based on this, the noise data point distribution shows a region where the noise reaches a peak (a densely populated noise area), then gradually decreases towards both sides. The direction of increasing level value corresponds to the signal direction. Furthermore, since spectral data typically reaches a relatively stable range within a short sampling time, if a suitable noise threshold (the first upward trough position) can be found as shown in the figure, the signal's data characteristics can be quickly identified without repeatedly calculating the threshold, significantly improving data processing efficiency.

[0036] (2) Threshold estimation of spectrum data signals.

[0037] Based on the above characteristics of noise and signal data distribution, the following threshold calculation steps are given: ① During the sampling time, collect spectrum data of the specified frequency band and process the spectrum data of each frame in sequence; ② Traverse and find the minimum level min and maximum level max of the frequency data. Starting from the minimum min, divide the interval [min, max] into segments with a certain step size. ③ Count the number of data points in the current frame data in each segment interval, and obtain the upper boundary level value of the interval with the maximum number of data points. Indicates which spectral data interval li is currently being processed; ④ Locate the spectrum data range The maximum value interval lmax is determined by the direction of increase in the level value within the lmax interval, as follows: Figure 3 The spectral data distribution characteristics shown indicate that the first trough is located in the interval above the current spectral data. ⑤ The upper boundary value of the interval where the trough is located is the signal threshold value of the sampled data; ⑥ Based on this threshold value, the original spectrum data is processed to obtain a signal list, and then cluster analysis is performed.

[0038] In specific clustering, this application embodiment employs an improved K-means clustering analysis method, the flowchart of which is shown below. Figure 4 As shown. Specifically: Step 1: Given Data objects ,in , , For the attribute dimensions of the data, Any two data objects in a set of data objects and European distance The calculation is expressed as follows:

[0039] Based on the Euclidean distance between two objects, let:

[0040]

[0041]

[0042]

[0043]

[0044] In the formula: The minimum distance between two elements. The maximum distance between two elements. Let be the minimum value of the set of minimum distances between two elements. Let R be the maximum value of the set of minimum distances between two elements, and R be the cluster radius. The radius weight is typically a given empirical value of 0.5.

[0045] ①If Then the given All data in a data object belongs to one category, and there is only one category.

[0046] ②If Then given In each data object, each object is divided into a separate class, and each class contains only one element: the cluster center.

[0047] ③If Then given When a data object is exactly the maximum value among the minimum distances, the classification has one and only one element: the cluster center.

[0048] Therefore, the clustering radius directly affects the final result of data classification. A larger clustering radius R results in fewer clusters and more elements per cluster, which may not meet the specified classification requirements for data requiring more detailed classification. Conversely, a smaller clustering radius R results in fewer elements per cluster and more clusters, leading to either coarse or over-clustering, which are both unreasonable results. Therefore, it is necessary to analyze the actual data characteristics and understand the data's distribution before proceeding. and We set weights on the cluster radius to make the clustering results more reasonable.

[0049] Step 2: Using R as the radius, for the given... The data density of each data element in a data object is calculated as follows: Initially, the density of all data element objects is set to 0. All elements in the given data set are sequentially retrieved. When the distance between other data elements and the current data element object is less than the radius R, the density of the current data element is incremented by 1. The data element with the highest density among the data element objects is taken as the cluster center. Using this cluster center as the data processing object and R as the radius, the data element objects in the set are divided into classes. Data elements that do not belong to the class of the current cluster center are added to a new set of objects. The radius R is recalculated, and a new cluster center is found until all data has been classified.

[0050] Step 3: Detection, sorting, and parameter estimation of frequency-hopping signals. The following steps are further provided to determine the data classification for a given dataset: ① Given Data objects Take in sequence Calculate the Euclidean distance between each of the objects and the rest of the objects; ②Let Indicates the first The minimum non-zero value in the row; ③ Set radius ,right ,use Indicates that the cluster center is Cluster radius is ; ④ Take the following values ​​in sequence with radius R. For each of the objects, calculate the Euclidean distance between it and the remaining objects, and statistically determine the distances to the current object that are less than [a certain value]. Number of data elements Find the maximum value ,in At the same time, the maximum value is obtained. Cluster center element at time ; ⑤ The elements obtained in step ④ As cluster center, For radius pair Divide the data objects into sections. They were classified into one category at the time; ⑥ Remove all element objects that already belong to a category in step ⑤, and repeat step ① for the remaining data. Finally, all data element objects are classified.

[0051] The calculation steps are as follows:

[0052] Understandably, many improved methods have been developed based on the K-means algorithm, such as K-means++ and the KHM algorithm. The improvement of K-means++ over the traditional K-means algorithm lies in the selection principle for the remaining k-1 cluster centers (excluding the first cluster center). The selection principle is to choose the point furthest from the currently selected cluster center as the proposed cluster center, continuing until k cluster centers have been selected. The remaining steps are consistent with the K-means algorithm.

[0053] The KHM algorithm uses the Harmonic Average (HA) instead of the Euclidean distance from data points to centers in the K-Means algorithm, which to some extent reduces the impact of K-means' sensitivity to initial centers on data classification performance.

[0054] In summary, the UAV frequency hopping signal detection and sorting method provided in this application uses density-based characterization to select cluster centers during clustering. Different cluster centers are then used to classify data objects within a certain radius into different classes based on distance. Compared to the traditional K-means algorithm, when processing UAV frequency hopping signal detection data, it eliminates the need to set an initial k-supervised value, directly performing adaptive classification based on the elements in the dataset, thus avoiding the impact of manually determining an initial k-classification value on the data sorting results. Secondly, it uses a cluster radius weighting method, resulting in more reasonable clustering results and the ability to handle non-spherical clusters.

[0055] See Figure 5 The present invention can also provide a multimodal noise tag correction device, such as... Figure 5 As shown, the apparatus for performing the above-described multimodal noise label correction method may include: The signal threshold estimation unit 501 is used to acquire raw spectrum data and perform adaptive estimation to obtain the signal threshold based on the distribution characteristics of noise and signal data. The frequency point signal set calculation unit 502 is used to process the original spectrum data based on the signal threshold to obtain the frequency point signal set; Clustering unit 503 is used to cluster the frequency signal set using an improved K-means clustering analysis method to obtain clustering results; the improved K-means clustering analysis method includes dynamically determining the cluster radius based on the minimum distance set between data objects; calculating the density of each object using the cluster radius as a scale, and selecting the one with the largest density as the cluster center for category division; iterating this process until all objects are classified to obtain the clustering results; The detection and sorting unit 504 is used to perform parameter estimation for each signal cluster based on the clustering results, thereby completing the detection and sorting of frequency hopping signals.

[0056] This invention can also provide a multimodal noise tag correction device, the device including a processor and a memory: The memory is used to store program code and transmit the program code to the processor; The processor is used to execute the steps of the multimodal noise label correction method described above according to the instructions in the program code.

[0057] like Figure 6 As shown in the figure, an embodiment of the present invention provides a UAV frequency hopping signal detection and sorting device, which may include: a processor 10, a memory 11, a communication interface 12, and a communication bus 13. The processor 10, the memory 11, and the communication interface 12 all communicate with each other through the communication bus 13.

[0058] In this embodiment of the invention, the processor 10 may be a central processing unit (CPU), a graphics processing unit (GPU), an application-specific integrated circuit, a digital signal processor, a field-programmable gate array, or other programmable logic devices.

[0059] The processor 10 can call the program stored in the memory 11. Specifically, the processor 10 can execute the operations in the embodiments of the UAV frequency hopping signal detection and sorting method.

[0060] The memory 11 is used to store one or more programs. The programs may include program code, which includes computer operation instructions. In this embodiment of the invention, the memory 11 stores at least a program for implementing the following functions: Raw spectrum data is collected, and signal thresholds are obtained through adaptive estimation based on the distribution characteristics of noise and signal data. The original spectrum data is processed based on the signal threshold to obtain a set of frequency point signals; An improved K-means clustering analysis method is used to cluster the frequency signal set to obtain clustering results. The improved K-means clustering analysis method includes dynamically determining the cluster radius based on the minimum distance set between data objects; calculating the density of each object using the cluster radius as a scale, and selecting the one with the largest density as the cluster center for category division; iterating this process until all objects are classified to obtain the clustering results. Based on the clustering results, parameter estimation is performed on each signal cluster to complete the detection and sorting of frequency hopping signals.

[0061] In one possible implementation, the memory 11 may include a program storage area and a data storage area. The program storage area may store the operating system and applications required for at least one function (such as file creation or data read / write). The data storage area may store data created during use, such as initialization data.

[0062] In addition, memory 11 may include high-speed random access memory, and may also include non-volatile memory, such as at least one disk storage device or other volatile solid-state storage device.

[0063] Communication interface 12 can be an interface for the communication module, used to connect with other devices or systems.

[0064] Of course, it should be noted that, Figure 6 The structure shown does not constitute a limitation on the UAV frequency hopping signal detection and sorting device in the embodiments of the present invention. In practical applications, the UAV frequency hopping signal detection and sorting device may include more than Figure 6 More or fewer components as shown, or combinations of certain components.

[0065] This invention can also provide a computer-readable storage medium for storing program code for executing the steps of the above-described UAV frequency hopping signal detection and sorting method.

[0066] It should be noted that, in this document, relational terms such as "first" and "second" are used only 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. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0067] As can be seen from the above description of the embodiments, those skilled in the art can clearly understand that this application can be implemented by means of software plus necessary general-purpose hardware platforms. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in various embodiments or some parts of the embodiments of this application.

[0068] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, for system or system embodiments, since they are basically similar to method embodiments, the description is relatively simple, and relevant parts can be referred to the descriptions in the method embodiments. The systems and system embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without creative effort.

[0069] The above description is merely a preferred embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention are included within the scope of protection of the present invention.

Claims

1. A method for detecting and sorting frequency-hopping signals from unmanned aerial vehicles (UAVs), characterized in that, include: Raw spectrum data is collected, and signal thresholds are obtained through adaptive estimation based on the distribution characteristics of noise and signal data. The original spectrum data is processed based on the signal threshold to obtain a set of frequency point signals; An improved K-means clustering analysis method is used to cluster the frequency signal set to obtain clustering results. The improved K-means clustering analysis method includes dynamically determining the cluster radius based on the minimum distance set between data objects; calculating the density of each object using the cluster radius as a scale, and selecting the one with the largest density as the cluster center for category division; iterating this process until all objects are classified to obtain the clustering results. Based on the clustering results, parameter estimation is performed on each signal cluster to complete the detection and sorting of frequency hopping signals.

2. The UAV frequency hopping signal detection and sorting method according to claim 1, characterized in that, The signal threshold estimation method includes: For a frame of spectrum data, divide it into intervals according to its level value and count the number of data points in each interval; Find the peak interval with the most data points; From the peak range towards the direction of increasing level value, find the first trough along the data point distribution curve; The level value corresponding to the first trough is determined as the signal threshold.

3. The UAV frequency hopping signal detection and sorting method according to claim 1, characterized in that, The method for determining the cluster radius includes: Calculate the Euclidean distance between all pairs of data objects in the frequency signal set; Obtain the minimum distance from each data object to other objects, and form a set of minimum distances; Take the minimum non-zero value in the set. and maximum value The cluster radius is obtained by calculating the following formula: In the formula: This represents the preset weighting coefficient. .

4. The UAV frequency hopping signal detection and sorting method according to claim 3, characterized in that, The weighting coefficient .

5. The UAV frequency hopping signal detection and sorting method according to claim 1, characterized in that, The cluster centers selected from those with the highest density include: For the current subset of data to be clustered, calculate the number of other objects contained in the neighborhood of each data object with the cluster radius as the radius, define the number as the density of the object, and select the object with the highest density as the cluster center of the current round.

6. The UAV frequency hopping signal detection and sorting method according to claim 1, characterized in that, During the iteration process, after each successful classification of a category, all data objects contained in that category are removed from the set to be clustered. The distances are recalculated based on the remaining data objects, and new cluster radii and cluster centers are determined for the next round of clustering.

7. The UAV frequency hopping signal detection and sorting method according to claim 1, characterized in that, The parameter estimation includes calculating the statistical characteristics of each signal cluster in the time domain, frequency domain, and energy domain, wherein the statistical characteristics include one or more of the following: average center frequency, frequency hopping period, average bandwidth, and average signal energy.

8. A device for detecting and sorting frequency-hopping signals from unmanned aerial vehicles (UAVs), characterized in that, The apparatus for performing the UAV frequency hopping signal detection and sorting method according to any one of claims 1-7, the apparatus comprising: The signal threshold estimation unit is used to acquire raw spectrum data and perform adaptive estimation to obtain the signal threshold based on the distribution characteristics of noise and signal data. A frequency point signal set calculation unit is used to process the original spectrum data based on the signal threshold to obtain a frequency point signal set; A clustering unit is used to cluster the frequency signal set using an improved K-means clustering analysis method to obtain clustering results. The improved K-means clustering analysis method includes dynamically determining the cluster radius based on the minimum distance set between data objects; calculating the density of each object using the cluster radius as a scale, and selecting the object with the highest density as the cluster center for category division; iterating this process until all objects are classified to obtain the clustering results. The detection and sorting unit is used to perform parameter estimation for each signal cluster based on the clustering results, thereby completing the detection and sorting of frequency hopping signals.

9. A frequency-hopping signal detection and sorting device for unmanned aerial vehicles (UAVs), characterized in that, The device includes a processor and a memory: The memory is used to store program code and transmit the program code to the processor; The processor is used to execute the UAV frequency hopping signal detection and sorting method according to any one of claims 1-7 according to the instructions in the program code.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium is used to store program code for executing the UAV frequency hopping signal detection and sorting method according to any one of claims 1-7.

Images