Frequency hopping signal parameter estimation method, device, equipment and medium based on clustering

Abstract

The present invention relates to the field of signal processing technology, and provides a frequency hopping signal parameter estimation method, device, equipment and medium based on clustering, the method comprising: performing short-time Fourier transform on a received frequency hopping signal to obtain a time-frequency diagram, detecting the initial frequency point and initial bandwidth of the frequency hopping signal based on the time-frequency diagram; clustering signals in different time periods in the frequency hopping signal to obtain the time parameters of the frequency hopping signal; extracting a first target signal from the time domain data of the frequency hopping signal based on the time parameters; performing frequency conversion and downsampling processing on the first target signal according to the initial frequency point and the initial bandwidth to obtain a second target signal; estimating the parameters of the frequency hopping signal based on the second target signal. Performing secondary signal extraction on the frequency hopping signal through the time parameters obtained by clustering, and performing frequency conversion and downsampling on the secondary extracted signal, thereby improving the signal resolution and thus improving the accuracy of parameter estimation, and the recognition accuracy of the frequency hopping signal can be improved according to the estimated parameters of the frequency hopping signal.

Description

Technical Field

[0001] The present invention relates to the field of signal processing technology, and in particular to a frequency hopping signal parameter estimation method, device, equipment and medium based on clustering. Background Art

[0002] Frequency hopping communication technology is a technology that transmits information by rapidly switching carrier frequencies. Due to the unpredictability of frequency hopping, it is difficult for eavesdroppers to capture complete signals, which greatly enhances the concealment and anti-interference capabilities of communications. Therefore, it is widely used in mobile communications to improve anti-interference capabilities and confidentiality of communications. With the continuous development of wireless communication technology, frequency hopping communication is increasingly widely used in the efficient use of spectrum resources and frequency diversity technology.

[0003] In the process of detecting and analyzing frequency hopping signals, how to accurately identify the frequency and bandwidth of the signal is the key. The rapid frequency change of the frequency hopping signal makes it difficult for the traditional spectrum analysis method to effectively capture its characteristics, which puts higher requirements on the recognition and extraction of the signal. The existing detection method of frequency hopping signals is easily affected by noise when processing signals in a strong interference environment, resulting in a decrease in the accuracy of signal recognition. In addition, the time resolution of spectrum analysis is insufficient, and it is difficult to accurately capture the instantaneous frequency change of the frequency hopping signal. When the bandwidth of the initial signal is wide, the frequency resolution is also insufficient, and the estimated signal parameters such as frequency and bandwidth are not accurate enough, which also affects the accuracy of the recognition of the frequency hopping signal. Summary of the invention

[0004] The present invention provides a frequency hopping signal parameter estimation method, device, equipment and medium based on clustering, which are used to solve the defect that the resolution of the existing frequency hopping signal detection is insufficient, resulting in insufficient estimation accuracy of the signal parameters and affecting the accuracy of frequency hopping signal recognition.

[0005] The present invention provides a frequency hopping signal parameter estimation method based on clustering, comprising:

[0006] Performing short-time Fourier transform on the received frequency hopping signal to obtain a time-frequency diagram of the frequency hopping signal, and detecting an initial frequency point and an initial bandwidth of the frequency hopping signal based on the time-frequency diagram;

[0007] Clustering signals in different time periods of the frequency hopping signal to determine time parameters of the frequency hopping signal; the time parameters include a start time and an end time;

[0008] extracting a first target signal from the time domain data of the frequency hopping signal based on the time parameter;

[0009] Performing frequency conversion and down-sampling processing on the first target signal according to the initial frequency point and the initial bandwidth to obtain a second target signal;

[0010] Parameters of the frequency hopping signal are estimated based on the second target signal.

[0011] According to the frequency hopping signal parameter estimation method based on clustering provided by the present invention, clustering signals in different time periods of the frequency hopping signal to determine the time parameters of the frequency hopping signal includes:

[0012] Converting signals of different time periods in the frequency hopping signal into a signal matrix;

[0013] Clustering the frequency hopping signal based on the signal matrix to obtain clustering parameters of the frequency hopping signal; the clustering parameters include the number of categories, target data points under each category, the number of data points of the target data points under each category, the index and the center value;

[0014] The clustering of the signal in the second time period is performed on the basis of the clustering result of the signal in the first time period, the signal in the first time period is a signal in any time period of the signals in the different time periods, and the signal in the second time period is a signal in a time period subsequent to the signal in the first time period;

[0015] A time parameter of the frequency hopping signal is determined according to the clustering parameter.

[0016] According to the frequency hopping signal parameter estimation method based on clustering provided by the present invention, clustering the frequency hopping signal based on the signal matrix to obtain the clustering parameters of the frequency hopping signal includes:

[0017] Initialize the first cell array, the second cell array, the first vector, and the first matrix;

[0018] Classifying the frequency hopping signal point by point based on the signal matrix, classifying and saving each data point of the frequency hopping signal into the first cell array, and saving the index of each data point of the frequency hopping signal into the second cell array;

[0019] According to the target data points of each category stored in the first cell array and the indexes of the target data points of each category stored in the second cell array, counting the number and center value of the data points of each category;

[0020] The number of data points is saved in the first vector, and the center value is saved in the first matrix, so as to obtain clustering parameters of frequency hopping signals of each category.

[0021] According to the frequency hopping signal parameter estimation method based on clustering provided by the present invention, the frequency hopping signal is classified point by point based on the signal matrix, each data point of the frequency hopping signal is classified and saved in the first cell array, and the index of each data point of the frequency hopping signal is saved in the second cell array, including:

[0022] Obtain existing categories and the center value of each category in the existing categories;

[0023] Selecting a target data point of the frequency hopping signal from the signal matrix, and calculating a target distance between the target data point and the center value of each category in the existing categories;

[0024] Determine, according to the target distance, a target category among the existing categories that is closest to the target data point;

[0025] If the target distance between the target category and the target data point is greater than a preset threshold, a new category corresponding to the target data point is created in the first cell array, the target data point is saved in the first cell array, and the index of the target data point is saved in the second cell array;

[0026] If the target distance between the target category and the target data point is less than or equal to a preset threshold, saving the target data point as a data point under the target category to the first cell array, and saving the index of the target data point to the second cell array;

[0027] Update the number and center value of the data points of the newly added category or the target category according to the target data point;

[0028] Return to and execute the step of selecting a target data point of the frequency hopping signal from the signal matrix until the target data point is the last data point of the frequency hopping signal.

[0029] According to the clustering-based frequency hopping signal parameter estimation method provided by the present invention, determining the time parameter of the frequency hopping signal according to the clustering parameter includes:

[0030] Determine, according to the clustering parameters, a first category and a second category among the categories corresponding to the frequency hopping signal; the first category is a category in which there are newly added data points after clustering the frequency hopping signal, and the second category is a category in which there are no newly added data points after clustering the frequency hopping signal;

[0031] Determine a start time index and an end time index of the second category according to the index in the clustering parameter of the second category;

[0032] According to the start time index and the end time index, a time parameter of the frequency hopping signal corresponding to the target data point under the second category is determined.

[0033] According to the clustering-based frequency hopping signal parameter estimation method provided by the present invention, the first target signal is subjected to frequency conversion and down-sampling processing according to the initial frequency point and the initial bandwidth to obtain the second target signal, including:

[0034] Performing frequency conversion processing on the first target signal according to the initial frequency point to convert the initial frequency point of the first target signal into a preset frequency to obtain a frequency conversion signal;

[0035] The frequency conversion signal is downsampled according to the initial bandwidth to extract and filter the high bandwidth of the frequency conversion signal to obtain a second target signal.

[0036] According to the clustering-based frequency hopping signal parameter estimation method provided by the present invention, the parameters of the frequency hopping signal include a target bandwidth, a center frequency point, and a signal duration; and estimating the parameters of the frequency hopping signal based on the second target signal includes:

[0037] Extracting time domain data of a single-hop signal in the second target signal;

[0038] Based on a preset time domain threshold, starting from the beginning of the time domain data, point-by-point detection is performed backward to obtain a first target point whose signal amplitude is greater than the time domain threshold, and starting from the end of the time domain data, point-by-point detection is performed forward to obtain a first second target point whose signal amplitude is less than the time domain threshold;

[0039] Taking the time index corresponding to the first target point as the target start time, taking the time index corresponding to the second target point as the target end time, calculating the difference between the target start time and the target end time, and obtaining the signal duration of the frequency hopping signal;

[0040] Performing Fourier transform on the time domain data to obtain a target spectrum of the single-hop signal;

[0041] Based on a preset frequency domain threshold, point-by-point detection is performed starting from the low-frequency end of the target spectrum to obtain a third target point whose first spectrum amplitude is greater than the frequency domain threshold, and point-by-point detection is performed starting from the high-frequency end of the target spectrum to obtain a fourth target point whose first spectrum amplitude is less than the frequency domain threshold;

[0042] Taking the frequency of the third target point as the starting frequency, taking the frequency of the fourth target point as the ending frequency, and determining the target bandwidth of the frequency hopping signal according to the difference between the starting frequency and the ending frequency;

[0043] The center frequency of the frequency hopping signal is determined according to an average value of the starting frequency and the ending frequency.

[0044] The present invention also provides a frequency hopping signal parameter estimation device based on clustering, comprising the following modules:

[0045] A coarse estimation module, used for performing a short-time Fourier transform on the received frequency hopping signal to obtain a time-frequency diagram of the frequency hopping signal, and detecting an initial frequency point and an initial bandwidth of the frequency hopping signal based on the time-frequency diagram;

[0046] A clustering module, used to cluster signals of different time periods in the frequency hopping signal to determine time parameters of the frequency hopping signal; the time parameters include a start time and an end time;

[0047] A secondary extraction module, configured to extract a first target signal from the time domain data of the frequency hopping signal based on the time parameter;

[0048] A frequency conversion sampling module, used for performing frequency conversion and down-sampling processing on the first target signal according to the initial frequency point and the initial bandwidth to obtain a second target signal;

[0049] A precise estimation module is used to estimate the parameters of the frequency hopping signal based on the second target signal.

[0050] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the steps of the clustering-based frequency hopping signal parameter estimation method as described above are implemented.

[0051] The present invention also provides a non-transitory computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the steps of any of the above-mentioned clustering-based frequency hopping signal parameter estimation methods are implemented.

[0052] The present invention also provides a computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the steps of any of the above-mentioned clustering-based frequency hopping signal parameter estimation methods are implemented.

[0053] The clustering-based frequency hopping signal parameter estimation method, device, equipment and medium provided by the present invention detects the initial bandwidth and initial frequency of the frequency hopping signal by performing time-frequency analysis on the frequency hopping signal, and further clusters the signals in different time periods to obtain the start and end time of the signal. The start and end time are used to perform secondary signal extraction on the time domain data of the frequency hopping signal, and the first target signal extracted from the secondary extraction is subjected to frequency conversion and downsampling processing according to the detected initial frequency and initial bandwidth, thereby improving the time resolution and frequency resolution of the signal processing, and estimating the parameters of the frequency hopping signal according to the data after the frequency conversion and downsampling processing, thereby improving the estimation accuracy of the parameters, being able to more accurately detect the parameters of the frequency hopping signal, and thus improving the recognition accuracy of the frequency hopping signal. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0055] Figure 1 It is a flow chart of the frequency hopping signal parameter estimation method based on clustering provided by the present invention.

[0056] Figure 2 It is a schematic diagram of the frequency hopping signal parameter estimation process provided by the present invention.

[0057] Figure 3 It is a schematic diagram of time domain data of the frequency hopping signal provided by the present invention.

[0058] Figure 4 It is a frequency domain data schematic diagram of the frequency hopping signal provided by the present invention.

[0059] Figure 5 It is a schematic diagram of a time-frequency diagram of a frequency hopping signal provided by the present invention.

[0060] Figure 6 It is a schematic diagram of a time-frequency diagram of a single-hop signal of the frequency-hopping signal provided by the present invention.

[0061] Figure 7 It is a schematic diagram of the time domain waveform of a single-hop signal of the frequency hopping signal provided by the present invention.

[0062] Figure 8 It is a schematic diagram of marking the time parameters of the frequency hopping signal provided by the present invention in the time-frequency diagram.

[0063] Fig. 9 It is a structural schematic diagram of a frequency hopping signal parameter estimation device based on clustering provided by the present invention.

[0064] Fig.10 It is a structural schematic diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0065] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0066] Figure 1 FIG. 1 is a flow chart of a frequency hopping signal parameter estimation method based on clustering provided by the present invention, such as Figure 1 As shown, the method comprises the following steps:

[0067] Step 100, performing short-time Fourier transform on the received frequency hopping signal to obtain a time-frequency diagram of the frequency hopping signal, and detecting an initial frequency point and an initial bandwidth of the frequency hopping signal based on the time-frequency diagram;

[0068] Step 200, clustering signals in different time periods of the frequency hopping signal to determine time parameters of the frequency hopping signal; the time parameters include a start time and an end time;

[0069] Step 300, extracting a first target signal from the time domain data of the frequency hopping signal based on the time parameter;

[0070] Step 400, performing frequency conversion and down-sampling processing on the first target signal according to the initial frequency point and the initial bandwidth to obtain a second target signal;

[0071] Step 500: Estimate parameters of the frequency hopping signal based on the second target signal.

[0072] Perform short-time Fourier transform on the received frequency hopping signal to obtain the time-frequency diagram of the frequency hopping signal, and perform time-frequency analysis on the frequency hopping signal based on the time-frequency diagram to detect the initial frequency point and initial bandwidth of the frequency hopping signal. The received frequency hopping signal includes different hopping signals of multiple stations, and the frequency hopping signal is short-time Fourier transform (STFT) is performed on the frequency hopping signal to convert the time domain IQ data of the frequency hopping signal into frequency domain IQ data. The coordinate axis is constructed with time and frequency, and the data points in the same frequency domain at different time domains are accumulated and averaged to construct the time-frequency diagram.

[0073] Based on the generated time-frequency diagram, the parameters of the frequency hopping signal are roughly estimated initially, and the initial frequency point and initial bandwidth of the frequency hopping signal are detected.

[0074] In one embodiment, the time domain IQ data of the frequency hopping signal is sampled and saved at a sampling rate of fs. If the total length of the time domain IQ data is , where L is the length of data processed each time the short-time Fourier transform is performed, and Q is the number of short-time Fourier transform processing times, that is, the frequency hopping signal is divided into Q segments of data, and the Q segments of data are numbered, and each segment of data is subjected to short-time Fourier transform processing and time-frequency analysis. When performing short-time Fourier transform on the frequency hopping signal, L data points are taken in sequence each time for transformation, and denoising and threshold judgment are performed on the basis of the time-frequency diagram to find the frequency domain position index of the effective signal in the frequency hopping signal, and then the initial frequency point and initial bandwidth of the effective signal are calculated to roughly estimate the parameters of the frequency hopping signal. At this time, the initial frequency point and initial bandwidth of the effective signal and the number i (i=1,2,…,Q) of the data segment containing the effective signal are recorded.

[0075] Based on the generated time-frequency diagram, the signals in different time periods in the time-frequency diagram are clustered to obtain the time parameters of the frequency hopping signal, which include the start time and end time of the frequency hopping signal, specifically including the start time and end time of each single hopping signal in the frequency hopping signal. Based on the time parameters, the first target signal is extracted from the acquired time domain data of the original frequency hopping signal, specifically, the start and end times of the frequency hopping signal are determined according to the time parameters of the frequency hopping signal, so as to obtain the signal corresponding to the start and end times from the time domain data of the frequency hopping signal, and obtain the first target signal.

[0076] Optionally, the signals in different time periods in the time-frequency diagram are clustered. Specifically, based on the detected initial frequency point and initial bandwidth, the signals in different time periods are clustered according to the frequency point and bandwidth, etc., to obtain signals in one or more categories, and the time index corresponding to the signal in each category, that is, the start time and end time of the frequency hopping signal.

[0077] Furthermore, based on the detected initial frequency and initial bandwidth, the first target signal is frequency-converted and down-sampled to obtain the second target signal. The data after frequency conversion and down-sampling only contains single-hop data and no longer contains different hop data of multiple stations, thereby improving the time resolution and frequency resolution of the signal. Estimating the signal parameters such as the target bandwidth and center frequency of the frequency-hopping signal based on the second target signal is to accurately estimate the parameters of the frequency-hopping signal, thereby ensuring the parameter estimation accuracy of the frequency-hopping signal.

[0078] In this embodiment, by performing time-frequency analysis on the frequency hopping signal, detecting the initial bandwidth and initial frequency of the frequency hopping signal, and further clustering the signals in different time periods, the start and end time of the signal is obtained. The start and end time is used to perform secondary signal extraction on the time domain data of the frequency hopping signal, and the first target signal extracted from the secondary signal is subjected to frequency conversion and down-sampling processing according to the detected initial frequency and initial bandwidth, thereby improving the time resolution and frequency resolution of signal processing, and estimating the parameters of the frequency hopping signal according to the data after frequency conversion and down-sampling processing, thereby improving the estimation accuracy of the parameters, being able to more accurately detect the parameters of the frequency hopping signal, and thus improving the recognition accuracy of the frequency hopping signal.

[0079] In one embodiment, step 200 specifically includes:

[0080] Step 201, converting the signals of different time periods in the frequency hopping signal into a signal matrix;

[0081] Step 202, clustering the frequency hopping signal based on the signal matrix to obtain clustering parameters of the frequency hopping signal; the clustering parameters include the number of categories, target data points under each category, the number of data points of the target data points under each category, the index and the center value;

[0082] The clustering of the signal in the second time period is performed on the basis of the clustering result of the signal in the first time period, the signal in the first time period is a signal in any time period of the signals in the different time periods, and the signal in the second time period is a signal in a time period subsequent to the signal in the first time period;

[0083] Step 203: Determine the time parameter of the frequency hopping signal according to the clustering parameter.

[0084] When clustering frequency hopping signals, the signals in different time periods of the frequency hopping signals are first converted into a signal matrix. Based on the time-frequency data of the frequency hopping signals, the time domain data of different time periods are numbered, and the detected initial frequency point, initial bandwidth and data number are converted into a signal matrix. The rows of the matrix represent the number of data points, and the columns of the matrix represent the data dimension, which is the initial frequency point, initial bandwidth and data number.

[0085] Furthermore, the frequency hopping signal is clustered based on the signal matrix to obtain clustering parameters of the frequency hopping signal, which include the number of categories, target data points under each category, the number of data points of the target data points under each category, index and center value, etc.

[0086] Optionally, since the frequency hopping signal is divided into multiple signal segments, when the frequency hopping signal is clustered, the clustering of the signal in the second time period is performed on the basis of the clustering result of the signal in the first time period, and the signal in the first time period and the signal in the second time period are signals of any two time periods adjacent to each other. Optionally, the signal in the first time period is a signal in any time period of the frequency hopping signal, and the signal in the second time period is a signal in a subsequent time period adjacent to the signal in the first time period.

[0087] The time parameters of the frequency hopping signal are determined according to the clustering parameters, specifically, the time parameters of the category are determined according to the clustering parameters of the data points of each category obtained by clustering. Specifically, for the target data points under any category, the clustering parameters of the category include the index of each target data point, and the index includes the time index and the frequency index. Based on the index of each target data point, the first target data point and the last data point in the target data points of the category in the time domain can be determined, and the time index of the first target data point is used as the start time of the category, and the time index of the last target data point is used as the end time of the category, so as to obtain the time parameters of the target data points under the category.

[0088] In some embodiments, clustering of the frequency hopping signals based on the signal matrix of the frequency hopping signals is achieved by classifying the frequency hopping signals point by point. Based on this, step 202 further includes:

[0089] Step 212, initializing the first cell array, the second cell array, the first vector and the first matrix;

[0090] Step 222, classifying the frequency hopping signal point by point based on the signal matrix, classifying and saving each data point of the frequency hopping signal into the first cell array, and saving the index of each data point of the frequency hopping signal into the second cell array;

[0091] Step 232, counting the number and center value of the data points of each category according to the target data points of each category stored in the first cell array and the indexes of the target data points of each category stored in the second cell array;

[0092] Step 242: Save the number of data points to the first vector, save the center value to the first matrix, and obtain clustering parameters of each category of frequency hopping signals.

[0093] When clustering frequency hopping signals based on a signal matrix, first initialize a first cell array, a second cell array, a first vector and a first matrix, wherein the first cell array C is initialized to store the values ​​of data points under each category of clustering; a second cell array Cidx is initialized to store the index of the data points in each category; the first vector cCnt is initialized to store the number of target data points in each category; and the first matrix CC is initialized to store the center value of each category.

[0094] The frequency hopping signal is classified point by point based on the signal matrix, and each data point of the frequency hopping signal is classified and saved in the first cell array C. At the same time, according to the classification result of each data point, the index of each data point is saved in the second cell array Cidx. Further, according to the target data points of each category saved in the first cell array C and the index of the target data points of each category saved in the second cell array, the number of data points and the central value of each category are counted, and the number of data points of the target data points under each category is saved in the first vector cCnt, and the central value of the target data points under each category is saved in the first matrix CC, so as to obtain the clustering parameters of the frequency hopping signal of each category.

[0095] In one embodiment, when clustering the first signal segment in the frequency hopping signal, initialize a first cell array C to store the values ​​of the data points in each category of the cluster, and optionally, preset the first data point to the data point of the first category. Initialize a second cell array Cidx to store the index of the data points in each category, and optionally, preset the index of the first data point to the index of the data point of the first category. Initialize a first vector cCnt to store the number of target data points in each category, and optionally, preset the number of data points in the first category to 1. Initialize a first matrix CC to store the center value of each category, and optionally, preset the value of the first data point to the center value of the first category.

[0096] In another embodiment, when clustering non-first segment signals in a frequency hopping signal, after initializing the first cell array C, the second cell array Cidx, the first vector cCnt and the first matrix CC, it is also necessary to obtain an existing category, which is obtained by clustering the signal segment before the current signal segment of the frequency hopping signal. Based on this, step 232 may also include:

[0097] Step 2321, obtaining existing categories and the central value of each category in the existing categories;

[0098] Step 2322, selecting a target data point of the frequency hopping signal from the signal matrix, and calculating a target distance between the target data point and the center value of each category in the existing categories;

[0099] Step 2323, determining the target category closest to the target data point in the existing categories according to the target distance;

[0100] Step 2324: if the target distance between the target category and the target data point is greater than a preset threshold, then creating a new category corresponding to the target data point in the first cell array, saving the target data point to the first cell array, and saving the index of the target data point to the second cell array;

[0101] Step 2325, if the target distance between the target category and the target data point is less than or equal to a preset threshold, save the target data point as a data point under the target category to the first cell array, and save the index of the target data point to the second cell array;

[0102] Step 2326, updating the number and center value of the data points of the newly added category or the target category according to the target data point;

[0103] Step 2327, return and execute the step of selecting a target data point of the frequency hopping signal from the signal matrix until the target data point is the last data point of the frequency hopping signal.

[0104] When the frequency hopping signal is classified point by point based on the signal matrix, the existing category and the center value of the existing category are obtained. Optionally, if the signal segment currently classified is the first signal segment of the frequency hopping signal, the existing category is empty. The acquired data points of the existing category are saved in the first cell array C, and the center value of the existing category is saved in the first matrix CC.

[0105] Furthermore, a target data point of the frequency hopping signal is selected from the signal matrix, and a target distance between the target data point and the center value of each category in the existing categories is calculated. According to the target distance between the target data point and the center value of each category in the existing categories, a target category closest to the target data point is selected, and according to the target distance between the target data point and the center value of the target category, it is determined whether the target data point belongs to the target category.

[0106] Optionally, the target distance between the target data point and the center value of the target category is compared with a preset threshold; if the target distance between the target data point and the center value of the target category is less than or equal to the preset threshold, the target data point belongs to the target category, and the target data point is saved to the target area corresponding to the target category in the first cell array C, and the index of the target data point is saved in the second cell array Cidx, specifically, saved to the target index area corresponding to the target category in the second cell array Cidx.

[0107] If the target distance between the target data point and the center value of the target category is greater than the preset threshold, the target data point does not belong to the target category, a new category is created in the first cell array C, and the target data point is saved as a data point under the new category in the first cell array C, specifically, it is saved to the area corresponding to the new category in the first cell array C, and an index area corresponding to the new category is created in the second cell array Cidx, and the index of the target data point is saved in the second cell array Cidx, specifically, it is saved to the index area corresponding to the new category in the second cell array Cidx.

[0108] Update the number and center value of data points of the newly added category or target category according to the target data point. Specifically, if the target data point belongs to the target category, update the number and center value of data points of the target category according to the target data point; if the target data point belongs to the newly added category, update the number and center value of data points of the newly added category according to the target data point.

[0109] Among them, if the target data point belongs to a newly added category, the target data point is the first data point of the newly added category. At this time, the number of updated data points of the newly added category is 1, and the updated center value is the value of the target data point.

[0110] After the target data point is classified, the next data point of the frequency hopping signal is obtained from the signal matrix as the target data point, and the above classification operation is repeated until the currently classified target data point is the last data point of the frequency hopping signal.

[0111] Optionally, when classifying the frequency hopping signal point by point, each data point in the frequency hopping signal is sorted. In one embodiment, each data point in the frequency hopping signal is sorted in time order, and each data point of the frequency hopping signal is classified point by point in the sorted order. The number of categories cNum of the existing categories is obtained, and the first cell array C is initialized to contain data point values ​​of the existing categories and the newly added categories. The second cell array Cidx and the first vector cCnt are initialized, and the data points and center values ​​of the existing categories are stored in the first cell array C and the first matrix CC, respectively. For each existing category, the number of data points therein is counted and saved in the first vector cCnt to update the first vector cCnt.

[0112] For each data point after the frequency hopping signal is sorted, the first data point is initialized as the first category as an existing category. The initialization of the first category mainly includes saving the first data point as the first category in the first cell array C, saving the index of the first data point in the second cell array Cidx, saving the value of the first data point as the center value of the first category in the first matrix CC, initializing the number of data points of the first category in the first vector cCnt to 1, and performing classification processing point by point starting from the second data point. If there are data points of existing categories, classification processing is performed point by point starting from the first data point of the frequency hopping signal.

[0113] During the point-by-point classification process, calculate the distance between the current data point and the center value of each existing category, and determine whether the current data point belongs to an existing category. If the current data point does not belong to any existing category, create a new category, add the current data point to the new category, and update the number of data points, center value, index, and category number cNum of the new category. If the current data point belongs to an existing category, classify the current data point into the target category with the smallest distance, and update the number of data points, index, and center value of the target category.

[0114] In one embodiment, the center value of each category includes the center value of each data dimension, specifically including the frequency center value and the bandwidth center value, and the target distance between the target data point and the center value of each category includes the distance of each data dimension, that is, the frequency distance and the bandwidth distance. When determining whether the target data point belongs to any category in the existing categories, the distance between the target data point and the center value of all dimensions of the category is less than the corresponding preset threshold. Optionally, the distance between the frequency of the target data point and the center value of the frequency of the category is less than the preset frequency threshold, and the distance between the bandwidth of the target data point and the center value of the bandwidth of the category is less than the preset bandwidth threshold.

[0115] Furthermore, after completing the classification of each data point in the frequency hopping signal, the classified data points, indexes and center values ​​are obtained. Initialize the cell arrays dataGroup and idxGroup to store the classified data points and their indexes respectively, store the data points of each category in the cell array dataGroup, and store the index of the data points of each category in the cell array idxGroup. Store the center value of each category in the matrix ctGroup, and count the number of data points in each category numPerGroup and the total number of categories numGroup.

[0116] By clustering the frequency hopping signal point by point, the data points of the input frequency hopping signal can be dynamically classified, and the cluster center is updated after each classification, thereby continuously optimizing the classification results and improving the detection accuracy and efficiency of the frequency hopping signal.

[0117] In one embodiment, the time parameter of the frequency hopping signal is determined according to the clustering parameter, specifically, the start time and the end time of the frequency hopping signal are determined according to the time index in the clustering parameter. Therefore, step 203 further includes:

[0118] Step 213, determining a first category and a second category among the categories corresponding to the frequency hopping signal according to the clustering parameters; the first category is a category in which there are newly added data points after clustering the frequency hopping signal, and the second category is a category in which there are no newly added data points after clustering the frequency hopping signal;

[0119] Step 223, determining a start time index and an end time index of the second category according to the index in the clustering parameter of the second category;

[0120] Step 233: Determine the time parameter of the frequency hopping signal corresponding to the target data point in the second category according to the start time index and the end time index.

[0121] Optionally, the index of the target data point under each category in the clustering parameter includes a time index, and the start time and the end time of the frequency hopping signal are determined according to the time index, thereby obtaining the time parameter of the frequency hopping signal.

[0122] Specifically, after the frequency hopping signal is clustered, by detecting whether there are newly added elements in each category, it is determined which categories of frequency hopping signals have ended. First, according to the clustering parameters obtained after clustering, the first category and the second category in each category corresponding to the frequency hopping signal are determined, wherein the first category is a category with newly added data points after the frequency hopping signal is clustered, and the second category is a category with no newly added data points after the frequency hopping signal is clustered. Optionally, the second category is a category other than the first category in each category corresponding to the frequency hopping signal, and the frequency hopping signal of the first category with newly added data points is an unended frequency hopping signal.

[0123] According to the index in the clustering parameters of the second category, the start time index and the end time index of the frequency hopping signal of the second category are determined, the start time of the frequency hopping signal of the second category is determined according to the start time index, and the end time of the frequency hopping signal of the second category is determined according to the end time index, so as to obtain the time parameters of the frequency hopping signal corresponding to the target data point under the second category.

[0124] Optionally, the target data points under each category can also be arranged in chronological order. When determining the starting time index of the frequency hopping signal of the second category, the time index of the first data point under the second category is used as the starting time index of the frequency hopping signal of the second category, and the time index of the last data point under the second category is used as the ending time index of the frequency hopping signal of the second category.

[0125] In one embodiment, when determining whether there are new data points in each category, a vector is initialized, which contains the number of data points in each category before and supplements the number of newly added categories for comparison. The difference between the number of data points in each category and the number of data points before is calculated to determine which categories have new data points. For the second category without new data points, the index of the second category is recorded. For the second category without new data points, it is indicated that the frequency hopping signals of these first categories have ended. At the same time, the first category with new data points is obtained, and the index of these first categories is recorded. The first category with new data points indicates that the frequency hopping signals of these first categories are still continuing.

[0126] Furthermore, for the second category without new data points, the corresponding frequency hopping signal has ended, and the clustering parameters of these second categories are output for subsequent signal detection, recorded, and the number of categories of these second categories is determined for further processing.

[0127] In the further processing, the temporary time index is initialized, including the historical time index of each category, and the number of categories of the newly added category is supplemented to record the time index of the newly added category. For the second category without new data points, the start time index and end time index of the category are recorded, and the time index of each category is used to identify the duration of the signal. For the first category with new data points, its start time index is updated to continue to track the changes of the frequency hopping signal of this category.

[0128] After the frequency hopping signal is clustered, it is dynamically determined whether each category of frequency hopping signal is finished and relevant time parameters are extracted for further processing, thereby improving the accuracy and efficiency of frequency hopping signal detection.

[0129] The start time index and end time index of each category obtained correspond to a start time and an end time of the initial time domain data of the frequency hopping signal. According to the start time and the end time, the initial time domain data of the frequency hopping signal is extracted twice to obtain the time domain data of the time range of the single hopping signal corresponding to the category, that is, the first target signal. According to the detected initial frequency point and initial bandwidth, the first target signal is frequency-converted and down-sampled to obtain the second target signal.

[0130] Furthermore, step 400 may also include:

[0131] Step 401, performing frequency conversion processing on the first target signal according to the initial frequency point, so as to convert the initial frequency point of the first target signal into a preset frequency, thereby obtaining a frequency conversion signal;

[0132] Step 402: down-sample the frequency conversion signal according to the initial bandwidth to extract and filter the high bandwidth of the frequency conversion signal to obtain a second target signal.

[0133] According to the detected initial frequency point and initial bandwidth, the extracted first target signal is subjected to frequency conversion and down-sampling processing, wherein the frequency conversion processing is based on the single-hop signal corresponding to the second category, changing from the initially estimated frequency point fc to a preset frequency value, such as zero frequency, and the down-sampling processing is to extract and filter the first target signal from a high bandwidth to a low bandwidth, thereby reducing the amount of data that needs to be processed.

[0134] Optionally, the data points after frequency conversion and downsampling only contain single-hop data of the frequency-hopping signal and no longer contain data of different hops of multiple stations. Therefore, the specific start time and end time of the frequency-hopping signal can be detected directly in the time domain to calculate the duration of the frequency-hopping signal. The start frequency and end frequency of the frequency-hopping signal can also be calculated from the spectrum to calculate the target bandwidth and center frequency of the frequency-hopping signal.

[0135] The parameters of the frequency hopping signal to be estimated include the target bandwidth, the center frequency and the signal duration. Step 500 specifically includes:

[0136] Step 501, extracting time domain data of a single-hop signal in the second target signal;

[0137] Step 502, based on a preset time domain threshold, starting from the beginning of the time domain data, performing point-by-point detection backwards, obtaining a first target point whose signal amplitude is greater than the time domain threshold, and starting from the end of the time domain data, performing point-by-point detection forwards, obtaining a first second target point whose signal amplitude is less than the time domain threshold;

[0138] Step 503, taking the time index corresponding to the first target point as the target start time, taking the time index corresponding to the second target point as the target end time, calculating the difference between the target start time and the target end time, and obtaining the duration of the frequency hopping signal;

[0139] Step 504, performing Fourier transform on the time domain data to obtain a target spectrum of the single-hop signal;

[0140] Step 505: Based on a preset frequency domain threshold, point-by-point detection is performed starting from the low frequency end of the target spectrum to obtain a third target point whose first spectrum amplitude is greater than the frequency domain threshold, and point-by-point detection is performed starting from the high frequency end of the target spectrum to obtain a fourth target point whose first spectrum amplitude is less than the frequency domain threshold;

[0141] Step 506, taking the frequency of the third target point as the starting frequency, taking the frequency of the fourth target point as the ending frequency, and determining the target bandwidth of the frequency hopping signal according to the difference between the starting frequency and the ending frequency;

[0142] Step 507: Determine the center frequency of the frequency hopping signal according to the average value of the starting frequency and the ending frequency.

[0143] For the second target signal, extract the time domain data of the single-hop signal therein, and based on the preset time domain threshold, perform point-by-point detection from the beginning of the time domain data backward to obtain the first target point whose signal amplitude is greater than the time domain threshold, and perform point-by-point detection from the end of the time domain data forward to obtain the first second target point whose signal amplitude is less than the time domain threshold. Use the time index corresponding to the first target point as the target start time, and the time index corresponding to the second target point as the target end time, calculate the difference between the target start time and the target end time, and obtain the signal duration of the frequency hopping signal.

[0144] Perform Fourier transform on the extracted time domain data to obtain the target spectrum of the single-hop signal in the second target signal. Based on the preset frequency domain threshold, perform point-by-point detection from the low-frequency end of the target spectrum to obtain the third target point whose first spectrum amplitude is greater than the frequency domain threshold, and perform point-by-point detection from the high-frequency end of the target spectrum to obtain the fourth target point whose first spectrum amplitude is less than the frequency domain threshold. Take the frequency of the third target point as the starting frequency, take the frequency of the fourth target point as the ending frequency, calculate the difference between the starting frequency and the ending frequency, obtain the target bandwidth of the frequency hopping signal, and calculate the average value of the starting frequency and the ending frequency to obtain the center frequency of the frequency hopping signal.

[0145] In one embodiment, the specific start time and end time of the frequency hopping signal are detected in the time domain. Specifically, the time domain data of the single hopping signal is extracted from the second target signal after frequency conversion and downsampling. Threshold detection: set a suitable time domain threshold to detect the energy change of the frequency hopping signal. When the amplitude of the frequency hopping signal exceeds the time domain threshold, the frequency hopping signal is considered to have started. When the amplitude of the frequency hopping signal is lower than the time domain threshold and lasts for a certain time, the signal is considered to have ended. Among them, start time detection: perform point-by-point detection backward from the beginning of the time domain data, find the first point that exceeds the time domain threshold, and record the time index of the point as the start time of the frequency hopping signal. End time detection: perform point-by-point detection forward from the end of the time domain data, find the first point whose amplitude is lower than the time domain threshold and lasts for a certain time, and record the time index of the point as the end time of the frequency hopping signal. Signal duration calculation: calculate the signal duration by the index difference between the start time and the end time.

[0146] Furthermore, the starting frequency and ending frequency of the frequency hopping signal are calculated on the spectrum. First, the time domain data of the extracted single-hop signal is Fourier transformed to obtain the target spectrum of the single-hop signal. Spectrum threshold detection: A frequency domain threshold is set to detect the frequency component of the frequency hopping signal. When the spectrum amplitude exceeds the frequency domain threshold, the frequency point is considered to belong to the frequency hopping signal. Starting frequency detection: Point-by-point detection is performed starting from the low-frequency end of the target spectrum to find the first frequency point whose spectrum amplitude exceeds the frequency domain threshold, and the frequency point is recorded as the starting frequency of the frequency hopping signal. End frequency detection: Point-by-point detection is performed starting from the high-frequency end of the target spectrum to find the first frequency point whose spectrum amplitude is lower than the frequency domain threshold, and the frequency point is recorded as the ending frequency of the frequency hopping signal. Target bandwidth and center frequency calculation: The target bandwidth of the frequency hopping signal is calculated by the difference between the starting frequency and the ending frequency, and the center frequency of the frequency hopping signal is calculated by the average value of the starting frequency and the ending frequency.

[0147] By accurately detecting the start and end time of the frequency hopping signal in the time domain and accurately calculating the start and end frequencies of the frequency hopping signal in the spectrum, the frequency hopping signal parameters can be accurately estimated. After the frequency conversion and downsampling, the time resolution and frequency resolution of the data are improved, ensuring the accuracy and reliability of the frequency hopping signal detection results.

[0148] In one embodiment, referring to Figure 2 The parameter estimation process of the frequency hopping signal shown in the figure mainly includes several stages such as time-frequency diagram generation and analysis, signal clustering, time parameter extraction, signal secondary extraction, frequency conversion and down sampling, and signal parameter estimation.

[0149] In the time-frequency diagram generation and analysis stage, the received frequency hopping signal is subjected to short-time Fourier transform to generate a time-frequency diagram of the frequency hopping signal, and the frequency hopping signal is subjected to time-frequency analysis based on the time-frequency diagram to detect the initial frequency point and initial bandwidth of the frequency hopping signal. Optionally, the received frequency hopping signal includes frequency hopping signals of different hopping stations on multiple networks, such as Figure 3 and Figure 4 As shown, Figure 3 It is a frequency hopping signal with 129 hops in 4 networks in the time domain. Figure 4 is a frequency hopping signal with 129 hops in 4 networks in the frequency domain. After performing short-time Fourier transform (STFT) on the frequency hopping signal, the following is generated: Figure 5 The time-frequency diagram of a data segment where a single-hop signal is located is shown in Figure 6 As shown in the figure, there are 4 DDCs (Digital Down Converters) in the data segment, and the time domain time corresponding to the data segment where the single-hop signal is located is 1.92ms. The time domain waveform detection result of the first DDC is as follows: Figure 7 As shown, in Figure 7 In the example, the sampling rate fs of the frequency hopping signal is 32000K, the time accuracy is 0.0003125ms, the frequency is 12.49MHZ, the bandwidth is 1287.5KHZ, the signal duration is 0.67ms, and the number of possible frequency hopping signals is 1.

[0150] Further, in Figure 7 In the above, the time domain waveform detection of the frequency hopping signal is performed based on the configured threshold of 49.5, which is used to characterize the maximum detectable amplitude. The time domain waveform detection of the frequency hopping signal includes the modulus square after filtering and extraction of the DDC, that is, the square value of the modulus after filtering and extraction of the DDC, and the morphological closing operation of the frequency hopping signal.

[0151] In the signal clustering stage, the frequency hopping signals in different time periods are clustered to obtain the start time and end time of the frequency hopping signals. Specifically, the density clustering algorithm is used to cluster the signals according to the detected initial frequency point and initial bandwidth according to time and frequency, and the start time and end time of each single hop signal in the frequency hopping signal are determined to obtain the time parameters of each single hop signal. The start and end time marks of the time parameters in the time-frequency diagram are as follows: Figure 8 shown.

[0152] In the secondary signal extraction stage, according to the acquired time parameters of each hopping signal, the time domain data in the time period corresponding to the start time and the end time are obtained from the initial time domain data of the frequency hopping signal, so as to extract the first target signal of the frequency hopping signal.

[0153] Furthermore, in the frequency conversion and downsampling processing stage, the second target signal extracted is subjected to frequency conversion and downsampling processing according to the detected initial frequency point and initial bandwidth to obtain a second target signal. Through the frequency conversion and downsampling processing, the amount of data to be processed is reduced, while the time resolution and frequency resolution of the signal are improved.

[0154] In the signal parameter estimation stage, the specific start time and end time of the frequency hopping signal are detected in the time domain, and the energy change of the frequency hopping signal is determined through time domain threshold detection, the precise start time and end time of each hopping signal in the frequency hopping signal is determined, and the time domain characteristics such as the signal duration of the frequency hopping signal are extracted; the start frequency and end frequency of the frequency hopping signal are detected in the frequency domain, and the start frequency and end frequency of each hopping signal in the frequency hopping signal are calculated through Fourier transform (FFT) and frequency domain threshold detection, so as to accurately estimate the target bandwidth and center frequency of each hopping signal in the frequency hopping signal.

[0155] In this embodiment, the time-frequency diagram of the frequency hopping signal is analyzed to detect the initial frequency point and initial bandwidth of the frequency hopping signal, and the frequency hopping signals detected in different time periods are further clustered to obtain the start time and end time of the frequency hopping signal. The time parameters are extracted to perform secondary signal extraction on the time domain data of the frequency hopping signal, and the extracted signal is frequency converted and down-sampled according to the detected initial frequency point and initial bandwidth to improve the time resolution and frequency resolution of the signal. The parameters of the frequency hopping signal are estimated based on the signal after the frequency conversion and down-sampling processing, thereby improving the accuracy of signal estimation, so that the center frequency point and bandwidth of the frequency hopping signal can be detected more accurately.

[0156] Furthermore, a dynamic clustering algorithm is used to perform cluster analysis on the time and frequency characteristics of the signal, which can effectively distinguish different frequency hopping signals and extract accurate time parameters, solving the problems of high computational complexity and poor real-time performance of traditional clustering methods. At the same time, the amount of data to be processed is reduced through frequency conversion and downsampling processing, which improves processing efficiency and ensures the accuracy and resolution of the signal under low bandwidth.

[0157] Furthermore, on the basis of frequency conversion and down-sampling processing of the secondary extracted signal, point-by-point detection is performed in the time domain and frequency domain respectively to obtain more accurate start and end times and start and end frequencies, so that the parameters such as signal duration, bandwidth and center frequency of the frequency hopping signal can be estimated more accurately, and the accuracy of parameter estimation is improved through point-by-point detection.

[0158] The frequency hopping signal parameter estimation device based on clustering provided by the present invention is described below. The frequency hopping signal parameter estimation device based on clustering described below and the frequency hopping signal parameter estimation method based on clustering described above can be referred to each other.

[0159] Reference Fig. 9 , a frequency hopping signal parameter estimation device based on clustering provided by an embodiment of the present invention comprises:

[0160] A coarse estimation module 10 is used to perform a short-time Fourier transform on the received frequency hopping signal to obtain a time-frequency diagram of the frequency hopping signal, and detect an initial frequency point and an initial bandwidth of the frequency hopping signal based on the time-frequency diagram;

[0161] A clustering module 20 is used to cluster signals of different time periods in the frequency hopping signal to determine time parameters of the frequency hopping signal; the time parameters include a start time and an end time;

[0162] A secondary extraction module 30, configured to extract a first target signal from the time domain data of the frequency hopping signal based on the time parameter;

[0163] A frequency conversion sampling module 40, configured to perform frequency conversion and down-sampling processing on the first target signal according to the initial frequency point and the initial bandwidth to obtain a second target signal;

[0164] The precise estimation module 50 is configured to estimate parameters of the frequency hopping signal based on the second target signal.

[0165] In one embodiment, the clustering module 20 is further configured to:

[0166] Converting signals of different time periods in the frequency hopping signal into a signal matrix;

[0167] Clustering the frequency hopping signal based on the signal matrix to obtain clustering parameters of the frequency hopping signal; the clustering parameters include the number of categories, target data points under each category, the number of data points of the target data points under each category, the index and the center value;

[0168] The clustering of the signal in the second time period is performed on the basis of the clustering result of the signal in the first time period, the signal in the first time period is a signal in any time period of the signals in the different time periods, and the signal in the second time period is a signal in a time period subsequent to the signal in the first time period;

[0169] A time parameter of the frequency hopping signal is determined according to the clustering parameter.

[0170] In one embodiment, the clustering module 20 is further configured to:

[0171] Initialize the first cell array, the second cell array, the first vector, and the first matrix;

[0172] Classifying the frequency hopping signal point by point based on the signal matrix, classifying and saving each data point of the frequency hopping signal into the first cell array, and saving the index of each data point of the frequency hopping signal into the second cell array;

[0173] According to the target data points of each category stored in the first cell array and the indexes of the target data points of each category stored in the second cell array, counting the number and center value of the data points of each category;

[0174] The number of data points is saved in the first vector, and the center value is saved in the first matrix, so as to obtain clustering parameters of frequency hopping signals of each category.

[0175] In one embodiment, the clustering module 20 is further configured to:

[0176] Obtain existing categories and the center value of each category in the existing categories;

[0177] Selecting a target data point of the frequency hopping signal from the signal matrix, and calculating a target distance between the target data point and the center value of each category in the existing categories;

[0178] Determine, according to the target distance, a target category among the existing categories that is closest to the target data point;

[0179] If the target distance between the target category and the target data point is greater than a preset threshold, a new category corresponding to the target data point is created in the first cell array, the target data point is saved in the first cell array, and the index of the target data point is saved in the second cell array;

[0180] If the target distance between the target category and the target data point is less than or equal to a preset threshold, saving the target data point as a data point under the target category to the first cell array, and saving the index of the target data point to the second cell array;

[0181] Update the number and center value of the data points of the newly added category or the target category according to the target data point;

[0182] Return to and execute the step of selecting a target data point of the frequency hopping signal from the signal matrix until the target data point is the last data point of the frequency hopping signal.

[0183] In one embodiment, the clustering module 20 is further configured to:

[0184] Determine, according to the clustering parameters, a first category and a second category among the categories corresponding to the frequency hopping signal; the first category is a category in which there are newly added data points after clustering the frequency hopping signal, and the second category is a category in which there are no newly added data points after clustering the frequency hopping signal;

[0185] Determine a start time index and an end time index of the second category according to the index in the clustering parameter of the second category;

[0186] According to the start time index and the end time index, a time parameter of the frequency hopping signal corresponding to the target data point under the second category is determined.

[0187] In one embodiment, the frequency conversion sampling module 40 is further used for:

[0188] Performing frequency conversion processing on the first target signal according to the initial frequency point to convert the initial frequency point of the first target signal into a preset frequency to obtain a frequency conversion signal;

[0189] The frequency conversion signal is downsampled according to the initial bandwidth to extract and filter the high bandwidth of the frequency conversion signal to obtain a second target signal.

[0190] In one embodiment, the parameters of the frequency hopping signal include a target bandwidth, a center frequency, and a signal duration; the precise estimation module 50 is further used to:

[0191] Extracting time domain data of a single-hop signal in the second target signal;

[0192] Based on a preset time domain threshold, starting from the beginning of the time domain data, point-by-point detection is performed backward to obtain a first target point whose signal amplitude is greater than the time domain threshold, and starting from the end of the time domain data, point-by-point detection is performed forward to obtain a first second target point whose signal amplitude is less than the time domain threshold;

[0193] Taking the time index corresponding to the first target point as the target start time, taking the time index corresponding to the second target point as the target end time, calculating the difference between the target start time and the target end time, and obtaining the signal duration of the frequency hopping signal;

[0194] Performing Fourier transform on the time domain data to obtain a target spectrum of the single-hop signal;

[0195] Based on a preset frequency domain threshold, point-by-point detection is performed starting from the low-frequency end of the target spectrum to obtain a third target point whose first spectrum amplitude is greater than the frequency domain threshold, and point-by-point detection is performed starting from the high-frequency end of the target spectrum to obtain a fourth target point whose first spectrum amplitude is less than the frequency domain threshold;

[0196] Taking the frequency of the third target point as the starting frequency, taking the frequency of the fourth target point as the ending frequency, and determining the target bandwidth of the frequency hopping signal according to the difference between the starting frequency and the ending frequency;

[0197] The center frequency of the frequency hopping signal is determined according to an average value of the starting frequency and the ending frequency.

[0198] Fig.10 An example of a physical structure diagram of an electronic device is shown in FIG. Fig.10As shown, the electronic device may include: a processor 1010, a communication interface 1020, a memory 1030 and a communication bus 1040, wherein the processor 1010, the communication interface 1020, and the memory 1030 communicate with each other through the communication bus 1040. The processor 1010 may call the logic instructions in the memory 1030 to execute the steps of the clustering-based frequency hopping signal parameter estimation method, for example, including:

[0199] Performing short-time Fourier transform on the received frequency hopping signal to obtain a time-frequency diagram of the frequency hopping signal, and detecting an initial frequency point and an initial bandwidth of the frequency hopping signal based on the time-frequency diagram;

[0200] Clustering signals in different time periods of the frequency hopping signal to determine time parameters of the frequency hopping signal; the time parameters include a start time and an end time;

[0201] extracting a first target signal from the time domain data of the frequency hopping signal based on the time parameter;

[0202] Performing frequency conversion and down-sampling processing on the first target signal according to the initial frequency point and the initial bandwidth to obtain a second target signal;

[0203] Parameters of the frequency hopping signal are estimated based on the second target signal.

[0204] In addition, the logic instructions in the above-mentioned memory 1030 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.

[0205] On the other hand, the present invention further provides a computer program product, the computer program product includes a computer program, the computer program can be stored in a non-transitory computer-readable storage medium, and when the computer program is executed by a processor, the computer can execute the steps of the clustering-based frequency hopping signal parameter estimation method provided in the above embodiments, for example, including:

[0206] Performing short-time Fourier transform on the received frequency hopping signal to obtain a time-frequency diagram of the frequency hopping signal, and detecting an initial frequency point and an initial bandwidth of the frequency hopping signal based on the time-frequency diagram;

[0207] Clustering signals in different time periods of the frequency hopping signal to determine time parameters of the frequency hopping signal; the time parameters include a start time and an end time;

[0208] extracting a first target signal from the time domain data of the frequency hopping signal based on the time parameter;

[0209] Performing frequency conversion and down-sampling processing on the first target signal according to the initial frequency point and the initial bandwidth to obtain a second target signal;

[0210] Parameters of the frequency hopping signal are estimated based on the second target signal.

[0211] In another aspect, the present invention further provides a non-transitory computer-readable storage medium having a computer program stored thereon, which is implemented when the computer program is executed by a processor to perform the steps of the clustering-based frequency hopping signal parameter estimation method provided in the above embodiments, for example, including:

[0212] Performing short-time Fourier transform on the received frequency hopping signal to obtain a time-frequency diagram of the frequency hopping signal, and detecting an initial frequency point and an initial bandwidth of the frequency hopping signal based on the time-frequency diagram;

[0213] Clustering signals in different time periods of the frequency hopping signal to determine time parameters of the frequency hopping signal; the time parameters include a start time and an end time;

[0214] extracting a first target signal from the time domain data of the frequency hopping signal based on the time parameter;

[0215] Performing frequency conversion and down-sampling processing on the first target signal according to the initial frequency point and the initial bandwidth to obtain a second target signal;

[0216] Parameters of the frequency hopping signal are estimated based on the second target signal.

[0217] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0218] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0219] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A frequency hopping signal parameter estimation method based on clustering, characterized in that: include: Performing short-time Fourier transform on the received frequency hopping signal to obtain a time-frequency diagram of the frequency hopping signal, and detecting an initial frequency point and an initial bandwidth of the frequency hopping signal based on the time-frequency diagram; Clustering signals in different time periods of the frequency hopping signal to determine time parameters of the frequency hopping signal; the time parameters include a start time and an end time; extracting a first target signal from the time domain data of the frequency hopping signal based on the time parameter; Performing frequency conversion and down-sampling processing on the first target signal according to the initial frequency point and the initial bandwidth to obtain a second target signal; estimating parameters of the frequency hopping signal based on the second target signal; The parameters of the frequency hopping signal include target bandwidth, center frequency and signal duration; The estimating the parameters of the frequency hopping signal based on the second target signal includes: Extracting time domain data of a single-hop signal in the second target signal; Based on a preset time domain threshold, starting from the beginning of the time domain data, point-by-point detection is performed backward to obtain a first target point whose signal amplitude is greater than the time domain threshold, and starting from the end of the time domain data, point-by-point detection is performed forward to obtain a first second target point whose signal amplitude is less than the time domain threshold; Taking the time index corresponding to the first target point as the target start time, taking the time index corresponding to the second target point as the target end time, calculating the difference between the target start time and the target end time, and obtaining the signal duration of the frequency hopping signal; Performing Fourier transform on the time domain data to obtain a target spectrum of the single-hop signal; Based on a preset frequency domain threshold, point-by-point detection is performed starting from the low frequency end of the target spectrum to obtain a third target point whose first spectrum amplitude is greater than the frequency domain threshold, and point-by-point detection is performed starting from the high frequency end of the target spectrum to obtain a fourth target point whose first spectrum amplitude is less than the frequency domain threshold; Taking the frequency of the third target point as the starting frequency, taking the frequency of the fourth target point as the ending frequency, and determining the target bandwidth of the frequency hopping signal according to the difference between the starting frequency and the ending frequency; The center frequency of the frequency hopping signal is determined according to an average value of the starting frequency and the ending frequency.

2. The method for frequency hopping signal parameter estimation based on clustering according to claim 1, characterized in that: The clustering of signals in different time periods in the frequency hopping signal to determine the time parameter of the frequency hopping signal includes: Converting signals of different time periods in the frequency hopping signal into a signal matrix; Clustering the frequency hopping signal based on the signal matrix to obtain clustering parameters of the frequency hopping signal; the clustering parameters include the number of categories, target data points under each category, the number of data points of the target data points under each category, the index and the center value; The clustering of the signal in the second time period is performed on the basis of the clustering result of the signal in the first time period, the signal in the first time period is a signal in any time period of the signals in the different time periods, and the signal in the second time period is a signal in a time period subsequent to the signal in the first time period; A time parameter of the frequency hopping signal is determined according to the clustering parameter.

3. The method for frequency hopping signal parameter estimation based on clustering according to claim 2, characterized in that: The clustering of the frequency hopping signal based on the signal matrix to obtain the clustering parameters of the frequency hopping signal includes: Initialize the first cell array, the second cell array, the first vector, and the first matrix; Classifying the frequency hopping signal point by point based on the signal matrix, classifying and saving each data point of the frequency hopping signal into the first cell array, and saving the index of each data point of the frequency hopping signal into the second cell array; According to the target data points of each category stored in the first cell array and the indexes of the target data points of each category stored in the second cell array, counting the number and center value of the data points of each category; The number of data points is saved in the first vector, and the center value is saved in the first matrix, so as to obtain clustering parameters of frequency hopping signals of each category.

4. The method for frequency hopping signal parameter estimation based on clustering according to claim 3, characterized in that: The step of classifying the frequency hopping signal point by point based on the signal matrix, classifying and saving each data point of the frequency hopping signal into the first cell array, and saving the index of each data point of the frequency hopping signal into the second cell array includes: Obtain existing categories and the center value of each category in the existing categories; Selecting a target data point of the frequency hopping signal from the signal matrix, and calculating a target distance between the target data point and the center value of each category in the existing categories; Determine, according to the target distance, a target category among the existing categories that is closest to the target data point; If the target distance between the target category and the target data point is greater than a preset threshold, a new category corresponding to the target data point is created in the first cell array, the target data point is saved to the first cell array, and the index of the target data point is saved to the second cell array; If the target distance between the target category and the target data point is less than or equal to a preset threshold, saving the target data point as a data point under the target category to the first cell array, and saving the index of the target data point to the second cell array; Update the number and center value of the data points of the newly added category or the target category according to the target data point; Return to and execute the step of selecting a target data point of the frequency hopping signal from the signal matrix until the target data point is the last data point of the frequency hopping signal.

5. The frequency hopping signal parameter estimation method based on clustering according to claim 2, characterized in that: The determining the time parameter of the frequency hopping signal according to the clustering parameter comprises: Determine, according to the clustering parameters, a first category and a second category among the categories corresponding to the frequency hopping signal; the first category is a category in which there are newly added data points after clustering the frequency hopping signal, and the second category is a category in which there are no newly added data points after clustering the frequency hopping signal; Determine a start time index and an end time index of the second category according to an index in the clustering parameter of the second category; According to the start time index and the end time index, a time parameter of the frequency hopping signal corresponding to the target data point under the second category is determined.

6. The frequency hopping signal parameter estimation method based on clustering according to claim 1, characterized in that: The performing frequency conversion and down-sampling processing on the first target signal according to the initial frequency point and the initial bandwidth to obtain the second target signal includes: Performing frequency conversion processing on the first target signal according to the initial frequency point to convert the initial frequency point of the first target signal into a preset frequency to obtain a frequency conversion signal; The frequency conversion signal is downsampled according to the initial bandwidth to extract and filter the high bandwidth of the frequency conversion signal to obtain a second target signal.

7. A frequency hopping signal parameter estimation device based on clustering, characterized in that: include: A coarse estimation module, used for performing a short-time Fourier transform on the received frequency hopping signal to obtain a time-frequency diagram of the frequency hopping signal, and detecting an initial frequency point and an initial bandwidth of the frequency hopping signal based on the time-frequency diagram; A clustering module, used to cluster signals of different time periods in the frequency hopping signal to determine time parameters of the frequency hopping signal; the time parameters include a start time and an end time; A secondary extraction module, configured to extract a first target signal from the time domain data of the frequency hopping signal based on the time parameter; A frequency conversion sampling module, used for performing frequency conversion and down-sampling processing on the first target signal according to the initial frequency point and the initial bandwidth to obtain a second target signal; A precise estimation module, configured to estimate parameters of the frequency hopping signal based on the second target signal; The parameters of the frequency hopping signal include target bandwidth, center frequency and signal duration; The precise estimation module is further used for: Extracting time domain data of a single-hop signal in the second target signal; Based on a preset time domain threshold, starting from the beginning of the time domain data, point-by-point detection is performed backward to obtain a first target point whose signal amplitude is greater than the time domain threshold, and starting from the end of the time domain data, point-by-point detection is performed forward to obtain a first second target point whose signal amplitude is less than the time domain threshold; Taking the time index corresponding to the first target point as the target start time, taking the time index corresponding to the second target point as the target end time, calculating the difference between the target start time and the target end time, and obtaining the signal duration of the frequency hopping signal; Performing Fourier transform on the time domain data to obtain a target spectrum of the single-hop signal; Based on a preset frequency domain threshold, point-by-point detection is performed starting from the low frequency end of the target spectrum to obtain a third target point whose first spectrum amplitude is greater than the frequency domain threshold, and point-by-point detection is performed starting from the high frequency end of the target spectrum to obtain a fourth target point whose first spectrum amplitude is less than the frequency domain threshold; Taking the frequency of the third target point as the starting frequency, taking the frequency of the fourth target point as the ending frequency, and determining the target bandwidth of the frequency hopping signal according to the difference between the starting frequency and the ending frequency; The center frequency of the frequency hopping signal is determined according to an average value of the starting frequency and the ending frequency.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the steps of the frequency hopping signal parameter estimation method based on clustering are implemented as described in any one of claims 1 to 6.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the frequency hopping signal parameter estimation method based on clustering are implemented.

Images