Distributed radar multi-radiation source signal sorting method combining signal similarity and positioning aggregation

Abstract

The invention belongs to the field of radar radiation source signal sorting, and relates to a distributed radar multi-radiation source signal sorting method combining signal similarity and positioning aggregation. The method specifically comprises the following steps: step 1, modeling a distributed radar positioning scene, assuming that all interference source radiation signals can cover all distributed radar nodes, and obtaining signal representation received by each distributed radar node; step 2, aiming at a signal received by each distributed radar node, separating interference source radiation signals based on linear constraint beam forming; 3, combining the signal similarity and the AOA positioning aggregation to realize interference radiation source signal pairing; therefore, signal sorting is completed.

Description

Technical Field

[0001] The invention belongs to the field of radar emitter signal sorting, and relates to a distributed radar multi-emitter signal sorting method combining signal similarity and positioning aggregation. Background Art

[0002] Emitter positioning has important strategic significance in the field of electronic warfare. By accurately identifying enemy radars, communication devices, and jammers, it can provide key information for strategic decision-making and significantly enhance battlefield surveillance and strike capabilities. Distributed radars, with their advantages of long baseline array configurations and signal-level cooperative processing, can achieve multi-perspective and multi-dimensional observations, such as measuring parameters like Time Difference of Arrival (TDOA) and Angle of Arrival (AOA), and have been widely used in emitter positioning. However, existing positioning methods mainly focus on single-emitter scenarios, concentrating on aspects such as the combination of measurement parameters, the design of solution algorithms, and strategies for reducing deviations. In multi-emitter scenarios, although the Spatial Matched Filter (SMF) can weaken the mutual interference between signals to a certain extent, in high-radiation power scenarios, it still leads to problems such as decreased measurement accuracy and parameter mismatches, thereby affecting the overall positioning performance. Therefore, in multi-emitter positioning, emitter signal sorting is the prerequisite and foundation for achieving efficient positioning.

[0003] Currently, scholars at home and abroad have carried out extensive research in the field of emitter signal sorting. The main research directions can be summarized into three categories: emitter sorting based on inter-pulse modulation characteristics, emitter sorting based on intra-pulse modulation characteristics, and emitter sorting methods based on machine learning. Among them, the sorting method based on inter-pulse modulation characteristics extracts parameters such as pulse arrival time, carrier frequency, pulse width, and arrival direction, and uses template matching or pulse repetition interval analysis to achieve emitter sorting. However, this type of method has limited performance when dealing with non-cooperative emitters and high-signal duty cycle scenarios. The sorting method based on intra-pulse modulation characteristics relies on techniques such as time-frequency analysis, ambiguity function, and higher-order statistics to extract and analyze the time-frequency distribution, envelope characteristics, phase characteristics, and higher-order statistics of signals to achieve sorting. However, this type of method has insufficient generalization ability and is usually only applicable to signals of specific modulation types. In coherent signal scenarios, its performance is relatively poor. The emitter sorting method based on machine learning usually requires a large number of data samples of the emitters to be identified in advance to train the classifier in order to improve the generalization ability. However, in practical applications, it is often unrealistic to obtain a large number of samples of non-cooperative emitters. In addition, problems such as long training time and high model complexity also limit the engineering application value of this type of method. Summary of the Invention

[0004] In view of this, the present invention combines signal similarity measurement and the AOA positioning aggregation characteristic, and proposes a distributed radar multi-radiation source signal sorting method that combines signal similarity and positioning aggregation, which can effectively sort multi-radiation source signals in both non-coherent and coherent signal scenarios.

[0005] The technical solution of the present invention is implemented as follows:

[0006] A distributed radar multi-radiation source signal sorting method that combines signal similarity and positioning aggregation, and the specific process is as follows:

[0007] Step 1, model the distributed radar positioning scenario, assuming that all interference source radiation signals can cover all distributed radar nodes, and obtain the signal representation received by each distributed radar node;

[0008] Step 2, for the signals received by each distributed radar node, separate the interference source radiation signals based on linearly constrained beamforming;

[0009] Step 3, combine signal similarity and AOA positioning aggregation to achieve interference radiation source signal pairing; thus, the signal sorting is completed.

[0010] Optionally, in the distributed radar positioning scenario of the present invention, it is assumed that there are N radar nodes and K interference radiation sources, and the signal received by the i-th distributed radar node can be expressed as:

[0011]

[0012] Where: a i (u k ) represents the array response vector of the i-th distributed radar node to the k-th radiation source signal, s k (t) represents the complex envelope of the k-th radiation source signal; τ i (u k ) represents the time delay of the k-th radiation source signal arriving at the i-th distributed radar node; n i (t) represents the Gaussian white noise vector at the i-th distributed radar node.

[0013] Optionally, in the present invention, it is assumed that the receiving weight vector related to the k-th beamformer at the i-th distributed radar node based on linear constraints is Then the corresponding separated signal representation is:

[0014]

[0015] Optionally, in the present invention, the w i,k is:

[0016]

[0017] Wherein: is a constraint matrix, represents AOA measurement information, is the corresponding constraint response vector.

[0018] Optionally, in step three of the present invention, for successful pairing based on signal similarity, the following two conditions need to be satisfied: (i) The minimum correlation coefficient of the two separated signals reaches or exceeds the set similarity threshold δ η ; (ii) The TDOA estimation result of the k0th separated signal of the reference node and the k i th separated signal of the ith radar node in the pairing combination falls within the time difference window where and are respectively the minimum and maximum values of all possible time differences that can be measured by the reference node and the ith radar node in the reconnaissance area.

[0019] Optionally, the minimum correlation coefficient in the present invention is:

[0020]

[0021]

[0022] Wherein, represents the correlation coefficient of the separated signals and .

[0023] Optionally, when the radiation source emits coherent signals in the present invention, pairing is further achieved based on the AOA positioning aggregation index.

[0024] Optionally, the pairing based on the AOA positioning aggregation index in the present invention is specifically:

[0025] Calculate the root mean square deviation as the evaluation index, and its calculation formula is:

[0026]

[0027] Wherein: represents the centroid of the AOA positioning results of all p nodes in the gth candidate pairing combination, represents the AOA positioning estimation result, N combs represents the total number of positioning results;

[0028] The RMSD of each candidate pairing group g is sorted in ascending order, and finally the top K pairings are selected as the effective output.

[0029] Beneficial effects:

[0030] The present invention provides a distributed radar multi-radiation source signal sorting method that combines signal similarity and positioning aggregation. First, based on the given AOA measurement information, a linear constraint beamformer is designed within the node to minimize the energy reception from other interfering directions while maximizing the energy reception in the expected direction, thereby effectively separating the radiation source signals from different directions. Subsequently, a time difference window constraint is introduced to define the signal similarity metric, and the AOA positioning aggregation is further combined to achieve effective pairing of cross-node signals. The present invention can efficiently sort the radiation source signals from different directions in both non-coherent and coherent signal scenarios, thus effectively avoiding the mutual interference between radiation source signals during the positioning process. Description of the Drawings

[0031] To more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the drawings required in the embodiments. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0032] Figure 1 It is a flowchart of the distributed radar multi-radiation source signal sorting method;

[0033] Figure 2 It is a schematic diagram of the distributed radar multi-radiation source positioning scenario;

[0034] Figure 3 It is a schematic diagram of the AOA positioning result distribution of the p nodes in the candidate pairing group;

[0035] Figure 4 It is the correlation coefficient of the distributed radar separated signals in the simulation example, (a) non-coherent signal scenario, (b) coherent signal scenario;

[0036] Figure 5 It is the preliminary pairing result of the distributed radar in the non-coherent signal scenario in the simulation example;

[0037] Figure 6 It is the pairing result of the distributed radar in the coherent signal scenario in the simulation example, (a) preliminary pairing result, (b) secondary pairing result. Detailed Embodiment

[0038] The following will describe the embodiments of the present invention in detail with reference to the drawings.

[0039] It should be noted that, without conflict, the following embodiments and the features in the embodiments can be combined with each other; and all other embodiments obtained by those of ordinary skill in the art based on the embodiments in the present disclosure without creative efforts belong to the scope of protection of the present disclosure.

[0040] It should be noted that the following description relates to various aspects of embodiments within the scope of the appended claims. It will be apparent that the aspects described herein can be embodied in a wide variety of forms, and any specific structure and / or function described herein is illustrative only. Based on this disclosure, those skilled in the art should understand that one aspect described herein can be implemented independently of any other aspect, and two or more of these aspects can be combined in various ways. For example, any number of the aspects set forth herein can be used to implement an apparatus and / or practice a method. Additionally, this apparatus can be implemented and this method can be practiced using other structures and / or functionality in addition to one or more of the aspects set forth herein.

[0041] An embodiment of the present application provides a distributed radar multi-radiation source signal sorting method that combines signal similarity and positioning clustering. In the present invention, the sorting of signals includes two steps: signal separation and signal pairing, as Figure 1 shown, and specifically includes the following steps:

[0042] Step 1: Modeling the distributed radar positioning scenario

[0043] Figure 2 Describes the distributed radar multi-radiation source positioning scenario in the northeast celestial coordinate system. The distributed radar system consists of N + 1 radar nodes located at known coordinates i = 0, 1,..., N, which are used to locate K radiation sources from unknown coordinates k = 1, 2,..., K. Within the node, the system can independently determine the AOA measurement information of the signal from the unknown radiation source, including the azimuth angle α i,k ∈(-π, π] and the elevation angle β i,k ∈[-π / 2, π / 2]; between nodes, a centralized processing method is adopted to transmit the received echoes and AOA measurement information of all radar nodes to the reference node s0 for signal-level collaborative processing.

[0044] The true value of the AOA measurement information of the distributed radar is closely related to the positions of the distributed radar nodes and the radiation sources, and shows a non-linear relationship, which can be expressed as follows:

[0045]

[0046] The measured value of the k-th radiation source can be expressed in the following vector form:

[0047]

[0048] Where: Represents the measurement vector of the k-th radiation source under noise perturbation; Represents the true measurement value corresponding to the k-th radiation source, Measurement error vector obeys a zero-mean Gaussian independent distribution, and its covariance matrix can be expressed as:

[0049]

[0050] Assume that the radiation signals of all interference sources can cover all distributed radar nodes. Then, the signal received by the i-th distributed radar node can be expressed as:

[0051]

[0052] Where: represents the array response vector of the i-th distributed radar node to the signal of the k-th radiation source, and M i is the number of array elements, and ||a i (u k )|| = 1, where ||·|| represents the 2-norm; s k (t) represents the complex envelope of the signal of the k-th radiation source; τ i (u k ) = r i,k / c represents the time delay of the signal of the k-th radiation source arriving at the i-th distributed radar node, and r i,k = ||u k - s i || represents the distance between the k-th radiation source and the i-th distributed radar node, and c is the speed of light; n i (t) represents the Gaussian white noise vector at the i-th distributed radar node.

[0053] Step 2: Separation of radiation source signals based on linearly constrained beamforming

[0054] Since the spatial positions of the jammers are different, and their spatial angles relative to each radar node are also different. Therefore, within the node, the separation of radiation source signals with different directions of arrival can be achieved through adaptive beamforming. Inside the radar node, the design problem of the receiving weight vector of the beamformer can be described as: while ensuring maximum energy reception in the expected direction, minimizing the energy reception in other interference directions of arrival. This means that the main lobe of the antenna array beam points to the expected direction of arrival and forms energy nulls in the other K - 1 interference directions of arrival.

[0055] Define as the receiving weight vector related to the k-th beamformer at the i-th distributed radar node. Then, the corresponding separated signal can be expressed as

[0056]

[0057] In the ideal case, the separated signal in Equation (5) only contains the signal of the k-th radiation source.

[0058] The distributed nodes can provide the AOA measurement information of the radiation sources To ensure that the gain of the beamforming is maximum for the direction and nulls are formed in the other K - 1 interference directions, the receiving weight vector needs to satisfy the following linear constraints

[0059]

[0060] where: is the constraint matrix, is the corresponding constraint response vector. In addition, to avoid the influence of the beamformer energy, the receiving weight vector should also satisfy the unit energy constraint, i.e., w i,k = 1.

[0061] In summary, the optimal solution of the receiving weight vector can be expressed as:

[0062]

[0063] Through the above receiving weight vector, the radiation signals from different interference sources can be effectively separated.

[0064] Step 3: Pairing of radiation source signals by combining signal similarity and AOA positioning aggregation

[0065] The separated signals and the AOA measurement information sequence (hereinafter referred to as the "signal - parameter sequence") obtained according to Equation (5) are denoted as

[0066]

[0067] For the given N + 1 radar nodes and K(N + 1) groups of signal - parameter sequences, there are a total of K N+1 possible pairing combinations, as shown below:

[0068]

[0069] Among them, only K groups are correct. At this time, the N + 1 groups of signal - parameter sequences in the pairing combination come from the same interference source.

[0070] To achieve effective pairing, the present invention proposes a two - stage pairing strategy that combines signal similarity and AOA positioning aggregation to identify the effective pairing from all possible combinations. In the first stage, based on signal similarity, a preliminary pairing is performed to narrow down all possible combinations to a candidate pairing set. The signals of the same radiation source usually show a high similarity due to the shared modulation characteristics, while the signals of different radiation sources have a low similarity due to the differences in modulation methods. To quantify the similarity of each pairing combination, the minimum correlation coefficient is used as a metric:

[0071]

[0072] Among them

[0073]

[0074] represents the separation signal and The correlation coefficient, whose value range is [0, 1]. The equal signs at both ends hold when the signals are orthogonal or linearly correlated, and the closer the correlation coefficient is to 1, the more similar the two signals are.

[0075] According to Equation (10), a similarity threshold δ η is introduced to evaluate the signal similarity of the paired combinations. When the minimum correlation coefficient this paired combination is regarded as a high - similarity combination; otherwise, it is classified as a low - similarity combination.

[0076] To further improve the reliability of the evaluation, geometric constraints are integrated into the pairing strategy. Define as the TDOA estimation result of the k0 - th separated signal of the reference node and the k - th separated signal of the i - th radar node in the paired combination. A successful pairing needs to meet the following two conditions: (i) The minimum correlation coefficient reaches or exceeds the similarity threshold δ i ; (ii) η ; (ii) falls within the time - difference window where and are the minimum and maximum values of all possible time differences that can be measured by the reference node and the i - th radar node in the reconnaissance area respectively. It is determined by the reconnaissance area and the geometric configuration of the distributed radars and has nothing to do with the nature of the radiation source itself. By combining the two, a more robust signal similarity metric can be constructed to more effectively evaluate the pairing:

[0077]

[0078] where: ε(·) and δ(·) are the Heaviside step function and the Dirac impulse function respectively; ∩ represents the intersection operation. According to Equation (12), for any paired combination, when meets the condition and the pairing is considered successful, and at this time Conversely, if or the similarity condition is not met, the pairing fails, and at this time The pairing process can be summarized as follows:

[0079]

[0080] According to Equation (13), from all possible K N+1G is selected from the paired combinations K There are candidate paired groups, shown as follows

[0081]

[0082] Wherein: represents the set of signal parameters corresponding to the i-th distributed radar node in the candidate pairing represents the index of the signal parameter sequence in the candidate paired group at the i-th distributed radar node

[0083] When the radiation source emits incoherent signals, due to the significant difference in similarity between different signals, effective pairing can be achieved through Equation (13). At this time, all candidate pairings are regarded as valid, that is, G K =K. On the contrary, for coherent signals, this similarity difference is weakened, resulting in an increased risk of pairing failure. In this case, there may be mispairings in the candidate pairings obtained by Equation (14), causing the number of candidate pairings to exceed the number of actual radiation sources, that is, G K >K

[0084] To solve the above problems, the second stage of the pairing strategy adopts the AOA positioning clustering method to optimize the candidate pairings and ensure effective pairing in the coherent signal scenario. Considering that at least two nodes are required for three-dimensional AOA positioning, a system with N+1 distributed nodes can generate N combs possible node pairing combinations, shown as follows

[0085]

[0086] Wherein: C represents the combination operation; p, 2≤p≤N+1 represents the number of nodes participating in AOA positioning;! represents the factorial operation. For each candidate paired group, N combs kinds of node combination methods will generate N combs AOA positioning estimation results where nn = 1,..., N combs .

[0087] Figure 3 shows the schematic diagram of the p-node AOA positioning result distribution corresponding to G K candidate paired groups. For the true pairing, its AOA positioning results will gather in the area close to the true position of the signal source, while the AOA positioning results of the invalid pairing are relatively scattered. To quantify the clustering degree, the Root Mean Square Deviation (RMSD) is introduced as an evaluation index, and its calculation formula is:

[0088]

[0089] Wherein: Denote the centroid of the AOA positioning results of all p nodes in the g-th candidate pairing combination. As can be seen from Equation (16), a smaller RMSD g indicates a higher clustering degree of the AOA positioning results, meaning a greater probability of effective pairing; on the contrary, a larger RMSD g reflects a more discrete distribution, indicating a lower probability of correct pairing. The RMSD of each candidate pairing group g is sorted in ascending order, and finally the top K pairings are selected as the effective output, denoted as:

[0090]

[0091] The following gives a simulation example of applying the present invention and explains its specific process.

[0092] Step 1: Modeling the distributed radar multi-radiation source positioning scenario.

[0093] In the simulation example, an S-band distributed radar system composed of 5 phased array antennas is considered, where the antenna element spacing is half a wavelength. The positions of the radar system and the antenna parameters are shown in the following table. Set the standard deviation of the AOA measurement error to 0.1°.

[0094]

[0095]

[0096] The range of the key reconnaissance area is from (50 km, -45°, 0°) to (225 km, 45°, 8°), with an interval of (1 km, 0.1°, 0.1°). The values in the brackets represent the slant range, azimuth angle, and elevation angle in the northeast celestial coordinate system, respectively. There are 2 radiation sources to be located in the reconnaissance area, and the position parameters are as follows

[0097]

[0098] Step 2: Separating the radiation source signals based on linear constrained beamforming.

[0099] Assume that the radiation sources emit uncorrelated high-power noise signals or signals with the same modulation parameters to construct an incoherent signal scenario and a coherent signal scenario, respectively. In each scenario, set the signal-to-noise ratio (SNR) of radiation source 1 arriving at the reference node to 20 dB, and the SNR of radiation source 2 arriving at the reference node to 40 dB. We can evaluate the signal separation performance by calculating the correlation coefficient between the separated signal and the pure interference signal.

[0100] Figure 4The signal separation performance of distributed radar in non - coherent signal and coherent signal scenarios is shown. It can be observed that, due to the significantly higher transmission power of radiation source 2 than that of radiation source 1, in the separation method based on spatial matched filter, the energy of radiation source 2 leaks to the separated signal of radiation source 1 through sidelobes, resulting in the correlation coefficient dropping to as low as 0.59. While the designed linearly constrained beamformer forms an energy null in the direction of radiation source 2, effectively suppressing the influence of interference signals, enabling the correlation coefficient of the separated signal to always remain above 0.92, demonstrating excellent signal separation performance. In the coherent signal scenario, the lowest correlation coefficient of the separated signal based on the linearly constrained beamformer reaches 0.95, significantly higher than 0.65 of the spatial matched filter, which further proves the superiority of the designed method in complex signal scenarios, capable of effectively dealing with non - coherent and coherent signal scenarios and ensuring the accurate separation of radiation source signals.

[0101] Step 3: Pair the radiation source signals by jointly considering signal similarity and AOA positioning clustering.

[0102] Set the similarity threshold to 0.5. The preliminary pairing results of signals in the non - coherent scenario are as Figure 5 shown, where the true pairing numbers are 11 and 13. Only when the similarity of the pairing exceeds the detection threshold can it be recognized as a valid pairing and marked by a red dashed line. This is because of the significant similarity differences between different radiation source signals in the non - coherent signal scenario, making the similarity values of mis - paired pairs relatively low, so that valid pairings can be reliably identified only through the preliminary pairing process.

[0103] Correspondingly, the preliminary pairing results in the coherent signal scenario are as Figure 6 (a) shown. Due to the high correlation between radiation source signals, the similarity of all possible pairings exceeds the similarity detection threshold, resulting in the number of candidate pairings exceeding the actual number of radiation sources. Therefore, effective pairing cannot be achieved only relying on the preliminary pairing process. For this reason, a secondary pairing strategy is adopted. Figure 6 (b) shows the distribution of AOA positioning results of all nodes related to candidate pairings. As can be seen from the red dashed line markings, the AOA positioning results of true pairings are more "concentrated". Finally, by selecting the two candidate pairings with the smallest RMSD as the effective output, the accurate identification of candidate pairings is achieved.

[0104] In summary, the above are only the implementation cases of the present invention and are not used to limit the protection scope of the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A distributed radar multi-radiation source signal sorting method that combines signal similarity and positioning aggregation, characterized in that The specific process is as follows: Step 1, model the distributed radar positioning scenario. Assume that the radiation signals of all interference sources can cover all distributed radar nodes, and obtain the signal representation received by each distributed radar node; Step 2, for the signals received by each distributed radar node, separate the radiation signals of the interference sources based on linearly constrained beamforming; Step 3, jointly implement the pairing of the interference radiation source signals by using signal similarity and AOA positioning aggregation; thus, the signal sorting is completed.

2. The distributed radar multi-radiation source signal sorting method according to claim 1 for combining signal similarity and positioning aggregation, characterized in that, Suppose there are N radar nodes and K interference radiation sources in the distributed radar positioning scenario. The signal received by the i-th distributed radar node can be expressed as: where: a i (u k ) represents the array response vector of the i-th distributed radar node to the k-th radiation source signal, s k (t) represents the complex envelope of the k-th radiation source signal; τ i (u k ) represents the time delay of the k-th radiation source signal arriving at the i-th distributed radar node; n i (t) represents the Gaussian white noise vector at the i-th distributed radar node.

3. The distributed radar multi-radiation source signal sorting method according to claim 1 for combining signal similarity and positioning aggregation, wherein, Let the received weight vector related to the k-th beamformer at the i-th distributed radar node based on linear constraints be Then the corresponding separated signal is expressed as:

4. The distributed radar multi-radiation source signal sorting method according to claim 3, characterized in that The said w i,k is as follows: Wherein: is a constraint matrix, represents AOA measurement information, is the corresponding constraint response vector.

5. The distributed radar multi-radiation source signal sorting method according to claim 1, characterized in that In step 3, for successful pairing based on signal similarity, the following two conditions need to be met: (i) The minimum correlation coefficient of the two separated signals reaches or exceeds the set similarity threshold δ η ; (ii) The TDOA estimation result of the k0-th separated signal of the reference node and the k i -th separated signal of the i-th radar node in the pairing combination falls within the time difference window , where and are the minimum and maximum values of all possible time differences that can be measured in the reconnaissance area by the reference node and the i-th radar node, respectively.

6. The distributed radar multi-radiation source signal sorting method combining signal similarity and positioning aggregation according to claim 5, characterized in that, The minimum correlation coefficient is: Among them, represents the separation signal and the correlation coefficient.

7. The distributed radar multi-radiation source signal sorting method combining signal similarity and positioning aggregation according to claim 1, characterized in that When the radiation source emits coherent signals, pairing is further implemented based on the AOA positioning aggregation index.

8. The distributed radar multi-radiation source signal sorting method combining signal similarity and positioning aggregation according to claim 7, characterized in that, The implementation of pairing based on the AOA positioning aggregation index is specifically as follows: Calculate the root mean square deviation as the evaluation index, and its calculation formula is: Wherein: represents the centroid of the AOA positioning results of all p nodes in the g-th candidate pairing combination, represents the AOA positioning estimation result, N combs represents the total number of positioning results; RMSD of each candidate pairing group g Sorted in ascending order, and finally the top K pairings are selected as the valid output.

Images