High fault-tolerant radar signal sorting method, device and product based on improved plane transformation

Abstract

The application discloses a high-fault-tolerance radar signal sorting method and device based on improved plane transformation, and a product, and relates to the radar signal processing field.The method comprises the following steps: processing a radar electromagnetic pulse signal to obtain a TOA sequence; obtaining a preset PRI array based on a preset PRI dynamic range and a search step, performing plane information entropy calculation on the TOA sequence, determining local maximum values, and forming a preliminary PRI candidate set; establishing a jitter PRI supplementary identification mechanism, storing the PRI determined as jitter modulation in the preset PRI array in a compensation queue; performing two-dimensional plane mapping on the preliminary PRI candidate set, extracting feature curves and corresponding TOA sequences in sequence, and storing the PRI determined as a false PRI in a false PRI queue; performing intersection operation on each TOA sequence index extracted in pairs to complete real PRI judgment and obtain a PRI set; and removing the false PRI queue and supplementing the compensation queue to realize the optimized integration of the PRI set.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of radar signal processing in electronic reconnaissance, and particularly relates to a high-fault-tolerance radar signal sorting method, device and product based on improved plane transformation. BACKGROUND

[0002] As the core perception component of the modern electronic reconnaissance system, the radar reconnaissance system continuously or periodically monitors the radar emitter electromagnetic signals in a passive manner through deployment on multi-dimensional detection platforms such as airborne, shipborne and spaceborne platforms, and analyzes and processes the signals. By obtaining the radiation pulse characteristic parameters of the target signals, the radar emitter signal technical parameters, spatial positioning information, radar model and working mode and other intelligence information are ultimately obtained, which provides reliable information support and decision basis for subsequent electromagnetic situation awareness and radar electronic protection.

[0003] As the core technology and important link of the radar reconnaissance system, the radar signal sorting undertakes the key task of deinterleaving the emitters from the complex overlapped signal stream, is the core step of constituting electromagnetic perception, and is also an important technical basis for subsequent parameter analysis and extraction and radar type identification of the radar reconnaissance system. With the rapid development of new radar technology and the sharp increase in the number of electromagnetic equipment, the traditional sorting algorithm inevitably presents certain limitations in the current complex electromagnetic environment, and phenomena such as batch increase and batch leakage occur, which has a cascading interference on the accuracy of subsequent identification and positioning, which poses a severe challenge to the information processing capability of the traditional radar electronic reconnaissance system and the comprehensive performance of the radar signal sorting algorithm.

[0004] In the sorting algorithm system, the pulse repetition interval (PRI) sorting is one of the methods with the earliest origin and application except the template matching method, and has the advantages of simple principle, high efficiency and high engineering practical value. In comparison, the template matching method has been difficult to meet the current complex radar signal sorting task with the innovation of radar system. The core mechanism of the PRI-based sorting is to reasonably mine the PRI values of the radar pulse sequences hidden in the time of arrival (TOA) information, and complete the radar pulse sequence search and extraction according to the PRI values. Compared with the sorting based on intra-pulse modulation characteristics or the sorting based on machine learning, this method has lower complexity and lower demand for computing resources, can meet the real-time processing requirements, has good system compatibility, and can realize multi-domain joint sorting combined with other sorting methods and features.

[0005] However, traditional PRI methods, such as Cumulative Difference Histogram (CDIF), Sequential Difference Histogram (SDIF), PRI transform method, etc., often lack data fault tolerance (pulse loss, parameter estimation error) and interference fault tolerance (pulse false alarm, noise interference) capabilities, resulting in a significant decrease in robustness in dynamic electromagnetic environments. At the same time, in the face of complex scenarios where multiple PRI modulation modes such as fixed PRI, uneven PRI, and jittered PRI coexist, the algorithm lacks effective adaptation mechanisms, resulting in insufficient comprehensiveness and versatility, and gradual decline in overall processing performance, which to some extent restricts its engineering applicability in complex electromagnetic environments. Therefore, it is urgent to develop high fault tolerance PRI sorting algorithms with stronger environmental adaptability and anti-interference capability to meet the increasingly severe electromagnetic reconnaissance needs and technical challenges. SUMMARY

[0006] To overcome the limitations of the above radar signal sorting methods in complex environments, enhance the data fault tolerance and interference fault tolerance of the sorting algorithm, and improve the adaptability and processing performance of the algorithm to adapt to the increasingly complex practical application environment of the radar reconnaissance system, the present application proposes a high fault tolerance radar signal sorting method based on improved plane transformation, comprising:

[0007] Step 1: Process the received radar electromagnetic pulse signal to obtain a TOA sequence;

[0008]

[0008] Step 2: Based on the preset PRI dynamic range and search step, obtain a preset PRI array, apply entropy theory to complete plane information entropy calculation on the TOA sequence, determine the local maximum value, and construct a preliminary PRI candidate set from all maximum value corresponding PRI values;

[0009] Step 3: Establish a jittered PRI supplementary identification mechanism, and sequentially perform jitter synthesis judgment on each element value in the preset PRI array obtained in step 2, store the PRI value determined as jitter modulation in the compensation queue as jitter modulation;

[0010] Step 4: Perform two-dimensional plane mapping on the preliminary PRI candidate set obtained in step 2, sequentially extract feature curves and their corresponding TOA sequences, and store false PRIs in a false PRI queue through PRI verification;

[0011] Step 5: Perform two-by-two intersection operation on the TOA sequence indexes extracted in step 4, and complete real PRI judgment to obtain a PRI set;

[0012] Step 6: Based on the PRI set obtained in step 5, the sequence data is integrated, the false PRI queue in step 4 is eliminated, the compensation queue in step 3 is supplemented, and the optimized integration of the PRI set is realized.

[0013] Preferably, the TOA sequence obtained in step 1 comprises the arrival times of n pulse signals {toa1, toa2, …, toan}, wherein toan represents the arrival time of the nth pulse signal. n}, wherein toan represents the arrival time of the nth pulse signal. n}, wherein toan represents the arrival time of the nth pulse signal.

[0014] Preferably, the step 2 comprises:

[0015] Step 2-1: presetting the PRI dynamic range as [PRImin, PRImax], wherein PRImin is the PRI preset minimum value, PRImax is the PRI preset maximum value, setting the search step as step, and obtaining the preset PRI array PRI min max min max preset ;

[0016] Step 2-2: taking the number of elements contained in PRI preset as the histogram bin number M;

[0017] Step 2-3: taking each element in PRI preset as a two-dimensional plane mapping transformation width W; counting the number of pulses in the ith bin under the current W as n i , and obtaining the probability density p i of the ith bin as:

[0018]

[0019] In the formula, n is the total number of received pulses, and the current W plane information entropy H(W) is obtained according to the following formula:

[0020]

[0021] Step 2-4: after traversing each element in the PRI array, performing a negative operation on the information entropy data, and converting the optimization problem of local minimum value detection of the information entropy into the problem of detecting local maximum value in the negative entropy function;

[0022] Step 2-5: based on the change trend of the plane information entropy value, adopting an adaptive strategy to determine the potential local maximum value point, and the observation window width corresponding to the local maximum value point constitutes the preliminary PRI candidate set PRI cand .

[0023] Preferably, the step 3 comprises:

[0024] ​​​​Step 3-1: Using PRI respectively preset Each element value in the matrix is ​​used as the width W of the two-dimensional plane mapping transformation and the number of histogram bins MM;

[0025] Step 3-2: Let the bin counting result N = {n1, n2, ..., n} i ,...,n M}, n i This represents the number of pulses in the i-th sub-bin, when the following formula is satisfied:

[0026] n i >mean(N)+std(N)

[0027] Preliminary determination indicates that this binning corresponds to one of the characteristic curves; where mean(N) and std(N) represent the mean and standard deviation, respectively;

[0028] Step 3-3: For all characteristic curves obtained in Step 3-2, perform jitter comprehensive judgment on the data corresponding to the bin index. If the judgment condition is met, determine that the PRI value corresponding to W and M is jitter modulation and store it in the compensation queue.

[0029] Preferably, the jitter comprehensive determination in step 3-3 includes: simultaneously satisfying the criteria of continuity, range, and chi-square test, wherein,

[0030] The continuity determination includes: first, using a set relatively low threshold to preliminarily determine whether there is a continuous array in the set index; if a continuous array exists, it is corrected for continuity and the binning index is completed; then, using a set relatively high threshold to determine the continuous array, if both thresholds meet the continuity condition, the index is determined to meet the continuity condition.

[0031] The range is determined as follows: the range of N is greater than the mean of N;

[0032] The chi-square test determines that: if the chi-square test is performed on N, the corresponding value rejects the uniform distribution hypothesis.

[0033] Preferably, step 4 includes:

[0034] Step 4-1: Preliminary PRI candidate set obtained from Step 2 cand Extract the histogram binning index set and the corresponding binning count result N = {n1, n2, ..., n}. M}, characteristic curves and corresponding TOA sequences; perform continuity correction on the histogram binning index set. If the set before and after correction is inconsistent and the corrected set does not have continuity characteristics, it indicates that the PRI is a false value and is stored in the false PRI queue; if the set before and after correction is consistent, proceed to step 4-2:

[0035] Step 4-2: Traverse each bin count result, and judge according to the following judgment criteria:

[0036] min{n i :n i >μ+σ}-max{n i :n i <μ+σ}>max{μ,σ}

[0037] Wherein, μ is the mean of the current N, and σ is the standard deviation of the current N.

[0038] If not, the corresponding PRI is a false PRI value, which is stored in the false PRI queue.

[0039] Preferably, the step 5 comprises:

[0040] The TOA sequence indexes extracted in step 4 are subjected to two-by-two intersection operation, and the PRI values with similar TOA sequences are merged into a unified subset, that is, the real PRI value and its harmonics are classified into one set.

[0041] It is determined that there is only one real PRI value in each subset, and the PRI with the least number of characteristic curves and the strongest vertical coordinate continuity in the subset is selected as the real PRI value, thereby obtaining the current PRI set.

[0042] Preferably, the step 6 comprises:

[0043] The intersection of the false PRI queue identified in step 4 and the current PRI set is calculated, and the PRI in the intersection and the corresponding TOA sequence are deleted.

[0044] The compensation queue in step 3 is extracted and supplemented to the real PRI value set and the corresponding TOA sequence set, thereby obtaining the real PRI value set and the corresponding TOA sequence set.

[0045] In a second aspect, the present application provides a computer device comprising a memory, a processor and a computer program stored in the memory, wherein the processor executes the computer program to realize the steps of the above method.

[0046] In another aspect, the present application provides a computer program product comprising a computer program which, when executed by a processor, realizes the steps of the above method.

[0047] Compared with the prior art, the present application has the following advantages:

[0048] 1. In view of the limitations of traditional PRI sorting algorithms in mixed modulation scenarios, the present application maps radar pulse time of arrival (TOA) to a specific two-dimensional feature plane, making full use of the planar feature separability of different modulation types such as fixed PRI, staggered PRI, and dithered PRI, as well as the independence of the two-dimensional mapping parameters of each radar emitter. The present application can effectively distinguish TOA sequences of multiple radar emitters and multiple modulation types, accurately determine the PRI value, and accurately identify the PRI modulation type.

[0049] 2. The present application introduces a two-dimensional plane information entropy quantization model, providing a quantitative judgment basis for candidate PRI values. Without relying on additional pulse search algorithms, the present application can achieve effective deinterleaving of radar pulse signals by extracting the constituent points of the PRI corresponding feature curve, reducing the algorithmic complexity while avoiding system sorting errors that may be introduced by sequence search algorithms, and improving the accuracy and reliability of the sorting results.

[0050] 3. The present application has strong data fault tolerance and interference fault tolerance, and the random distribution of stray pulses does not have a substantial impact on the overall representation of the two-dimensional plane. Even in situations where pulse loss and pulse false alarm are high, the present application can still achieve high-quality signal sorting. BRIEF DESCRIPTION OF DRAWINGS

[0051] Figure 1 is a two-dimensional plane information entropy curve and extreme point distribution diagram;

[0052] Figure 2 is a dithered PRI modulation plane transformation schematic diagram;

[0053] Figure 3 is a dithered PRI modulation histogram bin frequency statistics diagram;

[0054] Figure 4 is a fixed PRI modulation plane transformation schematic diagram;

[0055] Figure 5 is a fixed PRI modulation histogram bin frequency statistics diagram;

[0056] Figure 6 is a staggered PRI modulation plane transformation schematic diagram;

[0057] Figure 7 is a staggered PRI modulation histogram bin frequency statistics diagram;

[0058] Figure 8 is a high fault tolerance radar signal sorting algorithm flowchart based on improved plane transformation. DETAILED DESCRIPTION

[0059] The radar emitter reconnaissance system based on radar emitter reconnaissance system acquires the radar emitter pulse signals in the designated reconnaissance area, and processes the received radar electromagnetic pulse signals to obtain TOA data;

[0060] The method comprises:

[0061] Step 1) Based on the preset PRI dynamic range and search step, a preset PRI array is obtained, each element value in the PRI array is sequentially used as the mapping width W of a two-dimensional plane, and the number of elements contained in the PRI array is used as the number of histogram bins. Histogram statistics are performed on the PRI array elements, and the plane information entropy is calculated by applying the entropy theory. The obtained information entropy data is subjected to a negative operation and analyzed for the value change trend to determine the local maximum value. The PRI value corresponding to the obtained maximum value constitutes an initial candidate PRI set.

[0062] Step 2) In view of the problem that the plane information entropy feature of the dithered PRI modulated pulse is not significant, a dithered PRI supplementary identification mechanism is established. Each element value in the PRI array is sequentially used as the mapping width W of a two-dimensional plane and the number of histogram bins M, and a comprehensive judgment criterion based on histogram statistical sequence continuity, range value and chi-square test is constructed. When the dithering condition is met, it is determined that the PRI value corresponding to W is dithered modulation and is stored in the compensation queue.

[0063] Step 3) The candidate PRI set in step 1) is subjected to two-dimensional plane mapping. The values in the histogram statistics that exceed the sum of the mean value and the standard deviation are defined as the corresponding PRI feature curve, and the longitudinal coordinate sequence of all histogram bins that meet the condition and the corresponding TOA sequence index are extracted. The histogram bin sequence that meets the threshold condition is subjected to continuity correction, and whether the PRI is a dithered modulation type or a false PRI is determined according to the continuity processing result and the corresponding histogram range value. If it is a false PRI, it is stored in the false PRI queue.

[0064] Step 4) In order to eliminate the influence of harmonic interference, the extracted TOA sequence index is subjected to two-by-two intersection operation, and the PRI values with similar TOA sequences are merged into a unified subset. At this time, it can be determined that there is and only one real PRI value in each subset. The degree of continuity of the longitudinal coordinate sequence extracted in step 3) and the number of corresponding sequence features are used to determine the real PRI.

[0065] Step 5) Based on the PRI set determined in step 4), the sequence data is integrated, the false PRIs identified in step 3) are removed, the corresponding sequences are deleted, the compensation dithered PRIs filtered in step 2) are supplemented, the corresponding sequence extraction is completed, and the PRI modulation type is determined by using the number of corresponding sequence features and the histogram bin continuity. Finally, a complete output containing a real PRI value set, a modulation type set and a corresponding TOA sequence set is formed.

[0066] Step 6) Using the accuracy rate (the degree of matching between the sorted radar radiation source and the real signal), false detection rate (the proportion of pulses misidentified as a certain cluster), and false detection rate (the proportion of pulses not sorted out) as the core indicators, the data fault tolerance and interference fault tolerance performance of the present invention is evaluated, and the actual sorting performance of the present invention under different datasets and different pulse false alarm rates and pulse loss rates is verified.

[0067] The technical solution of the present invention will be described in detail below with reference to the accompanying drawings and embodiments.

[0068] Example 1

[0069] like Figure 8 As shown, Embodiment 1 of the present invention provides a highly fault-tolerant radar signal sorting method based on improved plane transformation, comprising the following steps:

[0070] 1. Initial variable setting and determination of candidate PRI set

[0071] Let the TOA sequence contain the arrival times of n pulse signals {toa1, toa2, ..., toa...} n}, The dynamic range of PRI is [PRI min ,PRI max ], of which PRI min Set a minimum value for PRI. max If a maximum value is preset for PRI and the search step size is set to step, then the preset PRI array can be represented as:

[0072]

[0073] in For the set of non-negative integers, PRI preset The element iteration in the transformation is used as the width W of the two-dimensional plane mapping transformation, and the transformation formula can be expressed as:

[0074]

[0075] In the formula, mod is the modulo function. It is the floor function, i.e., x i for toa i The remainder when divided by W, y i for toa i The quotient divided by W.

[0076] Taking fixed PRI modulation as an example, let toa q with toa p Let be the pulse arrival time from the same radar, with an interval of k·PRI, where k is any positive integer and PRI is the pulse repetition interval of that radar. This applies when the transform width W = PRI.

[0077]

[0078] From the above formula, when the correct PRI value is selected as the observation window width W, the pulse signals from the same radar radiation source exhibit significant geometric characteristics in the two-dimensional plane. Specifically, the abscissa tends to be equal, while the ordinate differs due to the pulse arrival time, and under ideal conditions, the ordinate difference exhibits regular characteristics of integer continuity.

[0079] Thus, for radar signals with fixed PRI modulation, their two-dimensional feature mapping exhibits a single line segment approximately perpendicular to the x-axis; for radar signals with staggered PRI modulation, their two-dimensional feature mapping exhibits a plurality of parallel approximately vertical line segments, and the number of feature line segments corresponds to the number of radar sub-periods; and for radar signals with jittered PRI modulation, their two-dimensional feature mapping exhibits a series of continuously distributed vertical line segments, which can also be visually regarded as a continuous and relatively large band region relative to the x-axis.

[0080] When the real PRI is selected as the observation window width, compared with other candidate values, the corresponding two-dimensional plane feature point distribution exhibits higher compactness and effectiveness, and a quantitative indicator of this distribution characteristic can be represented by the plane information entropy. The plane information entropy corresponding to the real PRI exhibits a local minimum value on the entire entropy value change curve, and the essence is that the correct PRI parameter enables the internal periodicity of the pulse time sequence to be best represented, and the distribution of the two-dimensional mapping points reaches the maximum regularity and the minimum randomness. Based on the above theoretical analysis, the application applies the plane information entropy to preliminarily determine the PRI candidate set PRI cand .

[0081] The number of elements contained in the PRI array is taken as the number of histogram bins M, and the number of pulses in the i-th bin under the current W is taken as n i , and the probability density of each bin is:

[0082]

[0083] In the formula, n is the total number of received pulses, and the plane information entropy under the current W is calculated using the probability density of each bin under the current W:

[0084]

[0085] The formula avoids the occurrence of log(0), when the plane feature point approaches to the ideal state of random distribution, the plane entropy approaches to log(n). In turn, the element value in the PRI array is taken as W, the overall calculation of H(W) is completed, in order to facilitate the application of mature peak detection algorithm, the negative operation is implemented on the original information entropy data, and the optimization problem of local minimum value detection of information entropy is converted into the problem of detecting local maximum value in the negative entropy function.

[0086] W * = argmax W (-H) (7)

[0087] Based on the change trend of the plane information entropy value, an adaptive strategy is used to determine the potential local maximum value point. The observation window width corresponding to the local maximum value point identified by the algorithm constitutes a preliminary PRI candidate set PRI cand . As Figure 1 shown in the figure, the relationship curve between the plane information entropy and the observation window width is shown, wherein the triangular marks represent the local minimum value points of the information entropy.

[0088] 2, jitter PRI supplementary mechanism

[0089] The two-dimensional plane mapping representation of the jitter PRI modulated pulse is represented as a plurality of continuous line segments, when the jitter range is too large, its tightness and effectiveness are not significant under the information entropy feature, thereby affecting the detection result. In order to solve the problem, the jitter PRI detection mechanism is designed, the element iteration in the PRI preset is taken as the two-dimensional plane mapping transformation width W and the histogram bin number M, which is different from the information entropy calculation strategy, this setting can ensure that the histogram bin corresponds to the PRI unit time. Let N={n1, n2,..., n M}, n i represents the pulse number in the i-th bin. When the i-th bin satisfies the condition:

[0090] n i > mean(N) + std(N) (8)

[0091] , it is preliminarily determined that the bin corresponds to one of the characteristic curves.

[0092] The continuity of the histogram binning index set that meets the threshold condition can be used as a criterion for the significance of jitter PRI. First, a preliminary determination is made as to whether a continuous array exists in the set of indices; in this case, a relatively low threshold is applied. If a continuous array exists, continuity correction is performed. This is because the distribution of pulse numbers within the jitter PRI is not uniform, and a small number of binning indices may be within the PRI jitter range but do not meet the threshold condition. Continuity correction will complete the determination of these binning indices. Then, a continuous array determination is performed, this time with a relatively high threshold. If both thresholds meet the continuity condition, the index is determined to satisfy the continuity condition.

[0093] Since this threshold is the sum of the mean and one standard deviation, almost all bins in N will contain bins that exceed this threshold. Therefore, relying solely on continuity judgment may not be feasible in N. i When the distribution is uniform, it is mistakenly identified as jitter. Therefore, if the continuity condition is met, the following conditions need to be further verified:

[0094] (1) Range verification: The range of N should be greater than the mean of N.

[0095] (2) Chi-square test: Perform a chi-square test on N. The corresponding value should be able to reject the uniform distribution hypothesis.

[0096] If the data corresponding to the binning index meets the judgment criteria in terms of histogram statistical sequence continuity, range, and chi-square test, then the PRI value corresponding to W and M is determined to be jitter modulation, and it is stored in the candidate queue. Figure 2 The diagram shown is a schematic of the plane transformation of jittered PRI modulation. Figure 3 This is a histogram of frequency distributions of the jittered PRI modulation.

[0097] 3. Candidate PRI Validation and Feature Curve Extraction

[0098] This step requires processing the candidate PRI set determined in step 1). cand A two-dimensional plane mapping analysis is performed, and since step 2) already covers the corresponding two-dimensional plane mapping calculation, PRI can be directly extracted from the calculation results of step 2). cand Histogram binning index set and corresponding binning count results N = {n1, n2, ..., n} M It is important to note that although both involve two-dimensional mapping, their target orientation and the objects processed by PRI differ, and the subsequent judgment criteria and specific operations are also different.

[0099] The core goal of this step is to verify the authenticity of the candidate PRI and extract the TOA index corresponding to the feature curve point. To this end, the present application first implements continuity correction on the histogram bin index set, and preliminarily distinguishes by comparing the consistency of the set before and after processing. If the set before and after correction is inconsistent, it indicates that the PRI may be a false value or a dithering PRI modulation. Further, the continuity of the set after continuity correction is evaluated: if it has continuity characteristics, it is determined that the PRI is a dithering signal; if it lacks continuity, it is identified as a false PRI.

[0100] Considering the threshold setting characteristics, there is still a problem that simply judging according to continuity may be n i In the case of relatively uniform distribution, false judgments may occur. To solve this problem, the present application introduces the following judgment criteria:

[0101] min{n i :n i >μ+σ}-max{n i :n i <μ+σ}>max{μ,σ} (9)

[0102] Where μ is the mean of the current N, and σ is the standard deviation of the current N. The criterion requires that the difference between the minimum value of the histogram bin index set corresponding to the count and the maximum value of the count corresponding to the remaining histogram bin index set should be greater than the mean and standard deviation of the current N.

[0103] If the above condition is not met, the PRI is also identified as a false PRI value. The false PRI is stored in the corresponding false queue, and the Figures 4-7 The two-dimensional mapping diagram of the verified real PRI (including fixed PRI and uneven PRI modulation type) is shown. The present application extracts all histogram bin ordinate sequences that meet the conditions as the basis for subsequent harmonic analysis, and obtains the corresponding TOA sequence index to form a candidate radar pulse set.

[0104] 4. Harmonic interference suppression

[0105] The generation of the harmonic phenomenon essentially depends on the multiple relationship between the two-dimensional mapping width W and the real pulse repetition interval. Here, the fixed PRI modulation type is still taken as an example for detailed analysis:

[0106] (1) If the two-dimensional plane mapping width is equal to the real PRI, i.e. W = PRI, according to the description of step 1), the two-dimensional mapping plane can effectively represent the feature curve corresponding to the PRI, thereby providing a reliable basis for subsequent pulse sorting, PRI value determination and modulation type identification.

[0107] (2) If the width of the two-dimensional plane mapping is an integer multiple of the real PRI, i.e., W = m·PRI, where m is a positive integer, let toa q with toa p Let toa be the arrival time of pulses from the same radar, with an interval of k·PRI, where k is any positive integer. q =q·PRI+r, According to the definition of PRI, r is a constant and 0 ≤ r < PRI, so the mapped coordinates are:

[0108]

[0109] It should be noted that since r is a constant, while y p x p Since it is obtained from the modulo or integer operation, the existence of the extra constant r only affects y. p x p The absolute numerical offset does not affect its relative distribution relationship.

[0110] From the above derivation, it can be seen that the horizontal coordinates tend to appear at m equally spaced positions, with the position distribution exhibiting an equidistant PRI interval pattern. The vertical coordinates, however, show continuity at their respective horizontal coordinate positions. Therefore, the two-dimensional mapping characteristics under this condition are: m characteristic curves appear within the observation window, and the vertical coordinates of each curve exhibit good continuity.

[0111] (3) If the actual PRI is an integer multiple of the width W of the two-dimensional plane mapping, i.e., PRI = m·W. Similarly, let toa q with toa p Given the arrival times of pulses from the same radar, the mapped coordinates can be obtained as follows:

[0112]

[0113] The derivation shows that, in this case, the abscissa remains nearly equal, while the ordinate values ​​no longer exhibit continuity, with the ordinate interval between adjacent pulses from the same radar being m. Therefore, the two-dimensional mapping characteristic under this condition is: a single characteristic curve appears within the observation window, but the ordinate of the curve loses its continuity.

[0114] Extending the above analysis to other PRI modulation schemes, their characteristic features follow the same pattern: when W = m·PRI, the number of characteristic curves becomes m times the mapping width of the true PRI, while the continuity of the ordinate remains unchanged; while when PRI = m·W, the number of characteristic curves remains unchanged, but the continuity of the ordinate decreases, and the interval between adjacent ordinates increases to m times that of W = PRI. Therefore, under a given PRI, the mapping condition of W = PRI has fewer characteristic curves and stronger continuity of the ordinate compared to W = m·PRI and PRI = m·W.

[0115] When W is a harmonic value, the correct pulse sequence can still be obtained and pulse sorting can be completed by histogram binning extraction. That is, ideally, the pulse extracted under PRI harmonics is consistent with the pulse extracted under true PRI, but errors may occur in determining the true PRI value and identifying the modulation type.

[0116] Therefore, to ensure the completeness and accuracy of the method, this invention first performs pairwise intersection operations on the extracted TOA sequence indices based on the characteristic that the pulse sequences extracted from the harmonics and their corresponding true PRI values ​​are identical. This merges PRI values ​​with similar TOA sequences into a unified subset, essentially grouping the true PRI values ​​and their harmonics into one set. Each subset is defined as containing one and only one true PRI value. The PRI value with the fewest characteristic curves and the strongest continuity on the ordinate is selected as the true PRI value, effectively suppressing harmonic interference and ensuring accurate identification of the PRI value and modulation type.

[0117] 5. PRI set optimization and integration

[0118] Based on the PRI set determined in step 4), the sequence data integration process is performed. First, the intersection of the false PRI set identified in step 3) and the current PRI set is calculated, and the PRIs and their corresponding sequences in the intersection are deleted. Simultaneously, the compensated jitter PRI queues and their corresponding sequences selected in step 2) are extracted and added to the true PRI value set and its corresponding TOA sequence. For each jitter PRI, only the corresponding TOA sequence within the maximum continuity range of the histogram binning is extracted. Through the above processing, the true PRI value set and its corresponding TOA sequence set are obtained.

[0119] For PRI modulation type identification, if the number of feature curves is 1, it is determined to be fixed PRI modulation; if the number of feature curves is greater than 1 and does not have histogram binning continuity, it is determined to be staggered PRI modulation; if the number of feature curves is much greater than 1 and clearly has histogram binning continuity, it is determined to be jittery PRI modulation; if the PRI originally belongs to the compensated jittery PRI queue, no discrimination is needed, and it is directly identified as jittery PRI modulation. This results in a complete output containing the set of true PRI values, the set of PRI modulation types, and the corresponding set of TOA sequences.

[0120] 6. Algorithm sorting performance evaluation and verification

[0121] The TOA sequence extracted by the algorithm is compared with the corresponding real radar TOA sequence. The data fault tolerance and interference fault tolerance capabilities of the method of the present invention are evaluated by the accuracy rate (ACC), false positive rate (FPR), and false negative rate (LPR) of radar radiation source signal pulse sorting, and the actual sorting effect of the method is verified. The formula is as follows:

[0122]

[0123] Example 2

[0124] Embodiment 2 of the present invention provides a computer device comprising: at least one processor, a memory, at least one network interface, and a user interface. The various components of the device are coupled together via a bus system. It is understood that the bus system is used to enable communication between these components. In addition to a data bus, the bus system also includes a power bus, a control bus, and a status signal bus.

[0125] The user interface may include a display, keyboard, or clicking device (e.g., mouse, trackball, touchpad, or touchscreen).

[0126] It is understood that the memory in the embodiments disclosed in this application may be volatile memory or non-volatile memory, or may include both volatile and non-volatile memory. The non-volatile memory may be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. The volatile memory may be random access memory (RAM), which is used as an external cache. By way of example, but not limitation, many forms of RAM are available, such as Static Random Access Memory (SRAM), Dynamic Random Access Memory (DRAM), Synchronous DRAM (SDRAM), Double Data Rate Synchronous DRAM (DDRSDRAM), Enhanced Synchronous DRAM (ESDRAM), Synchlink DRAM (SLDRAM), and Direct Rambus RAM (DRRAM). The memories described herein are intended to include, but are not limited to, these and any other suitable types of memory.

[0127] In some implementations, the memory stores elements such as executable modules or data structures, or subsets thereof, or extended sets thereof: operating systems and applications.

[0128] The operating system includes various system programs, such as the framework layer, core library layer, and driver layer, used to implement various basic business functions and handle hardware-based tasks. The application programs include various applications, such as media players and browsers, used to implement various application functions. Programs implementing the methods of the embodiments of this disclosure can be included in the application programs.

[0129] In the above embodiments, the processor can also invoke programs or instructions stored in memory, specifically programs or instructions stored in an application program, for the following purposes:

[0130] The steps of performing the method of Example 1.

[0131] The method of Embodiment 1 can be applied to a processor or implemented by a processor. The processor may be an integrated circuit chip with signal processing capabilities. During implementation, each step of the above method can be completed by integrated logic circuits in the processor's hardware or by instructions in software form. The processor can be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. It can implement or execute the methods, steps, and logic block diagrams disclosed in Embodiment 1. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the method disclosed in Embodiment 1 can be directly implemented by a hardware decoding processor, or by a combination of hardware and software modules in the decoding processor. The software modules can reside in random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, or other mature storage media in the art. This storage medium is located in memory; the processor reads information from the memory and, in conjunction with its hardware, completes the steps of the above method.

[0132] It is understood that the embodiments described in this invention can be implemented in hardware, software, firmware, middleware, microcode, or a combination thereof. For hardware implementation, the processing unit can be implemented in one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field-programmable gate arrays (FPGAs), general-purpose processors, controllers, microcontrollers, microprocessors, other electronic units for performing the functions described in this application, or combinations thereof.

[0133] For software implementation, the technology of this invention can be achieved by executing the functional modules (e.g., procedures, functions, etc.) of this invention. The software code can be stored in memory and executed by a processor. The memory can be implemented in the processor or externally.

[0134] Example 3

[0135] Embodiment 3 of the present invention may also provide a computer program product, including a computer program / instructions. When the computer program / instructions are executed by a processor, the various steps in the above method embodiments can be implemented.

[0136] Simulation example:

[0137] The effectiveness of this invention can be verified using the following simulation data:

[0138] Seven radar radiation sources were set up, and their parameters are shown in Table 1. To simulate a real-world scenario, the radar clusters were similar in certain dimensions and the number of pulses was uneven. This invention considers the coexistence of multiple PRI modulation types, including staggered, fixed, and jitter modulation methods. Based on the parameter information provided in Table 1, radar pulse descriptor word (PDW) data received by the receiving station was generated. Taking a 1000ms pulse sequence as an example, 5-40% of the pulses in this sequence were randomly lost, and a 10% false alarm rate was also set. The number of pulses for each radar radiation source in Table 1 represents the original sequence without pulse loss or false alarms. The actual number of pulses detected by each radar is not constant under different false alarm and loss rates.

[0139] Table 1 Radar radiation source parameters

[0140]

[0141] Using the accuracy (ACC), false positive rate (FPR), and false negative rate (LPR) evaluation indicators from step 6), the specific sorting results of the example are shown in Table 2.

[0142] Table 2 Sorting Results of Simulation Examples

[0143]

[0144]

[0145] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to the embodiments, those skilled in the art should understand that modifications or equivalent substitutions to the technical solutions of the present invention do not depart from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A highly fault-tolerant radar signal sorting method based on improved plane transformation, comprising: Step 1: Process the received radar electromagnetic pulse signal to obtain the TOA sequence; Step 2: Based on the preset dynamic range and search step size of PRI, obtain the preset PRI array, apply entropy theory to complete the planar information entropy calculation of the TOA sequence, determine the local maxima, and form a preliminary PRI candidate set by constructing the PRI values ​​corresponding to all maxima. Step 3: Establish a jitter PRI supplementary identification mechanism. Perform jitter comprehensive judgment on each element value in the preset PRI array obtained in Step 2 in turn, and store the PRI value that is judged as jitter modulation into the compensation queue as jitter modulation. Step 4: Perform two-dimensional plane mapping on the preliminary PRI candidate set obtained in Step 2, extract the feature curves and their corresponding TOA sequences in sequence, and store the false PRI values ​​into the false PRI queue through PRI verification. Step 5: Perform pairwise intersection operations on the TOA sequence indices extracted in Step 4, and complete the true PRI determination to obtain the PRI set; Step 6: Based on the PRI set obtained in Step 5, integrate the sequence data, remove the false PRI queue from Step 4, and supplement the compensation queue from Step 3 to achieve optimized integration of the PRI set.

2. The high-fault-tolerant radar signal sorting method based on improved plane transformation according to claim 1, characterized in that, The TOA sequence obtained in step 1 contains the arrival times of n pulse signals {toa1, toa2, ..., toa...} n }, where toa n This indicates the arrival time of the nth pulse signal.

3. The high-fault-tolerant radar signal sorting method based on improved plane transformation according to claim 1, characterized in that, Step 2 includes: Step 2-1: Preset the dynamic range of PRI to [PRI min ,PRI max ], of which PRI min Set a minimum value for PRI. max Set a preset maximum value for PRI, set the search step size to step, and obtain the preset PRI array PRI. preset ; Step 2-2: Using PRI preset The number of elements contained is the number of bins M in the histogram; Steps 2-3: Using PRI respectively preset Each element in the matrix is ​​used as the width W of the two-dimensional plane mapping transformation; the number of pulses in the i-th sub-bin under the current W is counted as n. i The probability density p of the i-th bin is obtained. i for: In the formula, n is the total number of received pulses. The current W-plane information entropy H(W) is obtained according to the following formula: Steps 2-4: After traversing each element in the PRI array, perform a negative operation on the planar information entropy data to transform the optimization problem of detecting local minima of information entropy into the problem of detecting local maxima in the negative entropy function; Steps 2-5: Based on the changing trend of the planar information entropy value, an adaptive strategy is used to determine potential local maxima. The observation window width corresponding to the local maxima constitutes the preliminary PRI candidate set. cand .

4. The high-fault-tolerant radar signal sorting method based on improved plane transformation according to claim 3, characterized in that, Step 3 includes: Step 3-1: Using PRI respectively preset Each element value in the matrix serves as the width W of the two-dimensional plane mapping transformation and the number of histogram bins M; Step 3-2: Let the bin counting result N = {n1, n2, ..., n} i ,...,n M }, n i This represents the number of pulses in the i-th sub-bin, when the following formula is satisfied: n i >mean(N)+std(N) Preliminary determination indicates that this binning corresponds to one of the characteristic curves; where mean(N) and std(N) represent the mean and standard deviation, respectively; Step 3-3: For all characteristic curves obtained in Step 3-2, perform jitter comprehensive judgment on the data corresponding to the bin index. If the judgment condition is met, determine that the PRI value corresponding to W and M is jitter modulation and store it in the compensation queue.

5. The high-fault-tolerant radar signal sorting method based on improved plane transformation according to claim 4, characterized in that, The jitter comprehensive judgment in step 3-3 includes: simultaneously satisfying the continuity, range, and chi-square test criteria, wherein... The continuity determination includes: first, using a set relatively low threshold to preliminarily determine whether there is a continuous array in the set index; if a continuous array exists, it is corrected for continuity and the binning index is completed; then, using a set relatively high threshold to determine the continuous array, if both thresholds meet the continuity condition, the index is determined to meet the continuity condition. The range is determined as follows: the range of N is greater than the mean of N; The chi-square test determines that: if the chi-square test is performed on N, the corresponding value rejects the uniform distribution hypothesis.

6. The high-fault-tolerant radar signal sorting method based on improved plane transformation according to claim 1, characterized in that, Step 4 includes: Step 4-1: Preliminary PRI candidate set obtained from Step 2 cand Extract the histogram binning index set and the corresponding binning count result N = {n1, n2, ..., n}. M }, characteristic curves and corresponding TOA sequences; perform continuity correction on the histogram binning index set. If the set before and after correction is inconsistent and the corrected set does not have continuity characteristics, it indicates that the PRI is a false value and is stored in the false PRI queue; if the set before and after correction is consistent, proceed to step 4-2: Step 4-2: Iterate through the count results of each sub-box and make a judgment based on the following criteria: min{n i :n i >μ+σ}-max{n i :n i <μ+σ}>max{μ,σ} Where μ is the mean of N and σ is the standard deviation of N; If the condition is not met, the corresponding PRI value is considered a false PRI value and is stored in the false PRI queue.

7. The high-fault-tolerant radar signal sorting method based on improved plane transformation according to claim 1, characterized in that, Step 5 includes: Perform pairwise intersection operations on the TOA sequence indices extracted in step 4 to merge PRI values ​​with similar TOA sequences into a unified subset, that is, group the real PRI values ​​and their harmonics into one set; Each subset is identified as having one and only one true PRI value. The PRI value with the fewest feature curves and the strongest continuity of the ordinate in the subset is selected as the true PRI value, thus obtaining the current PRI set.

8. The high-fault-tolerant radar signal sorting method based on improved plane transformation according to claim 7, characterized in that, Step 6 includes: Calculate the intersection of the false PRI queues identified in step 4 and the current PRI set, and delete the PRIs and their corresponding TOA sequences from the intersection; The compensation queue from step 3 is extracted and added to the true PRI value set and its corresponding TOA sequence to obtain the true PRI value set and its corresponding TOA sequence set.

9. A computer device, comprising a memory, a processor, and a computer program stored in the memory, characterized in that, The processor executes the computer program to implement the steps of the method of claim 1.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method of claim 1.

Images