A radar signal sorting method and apparatus

Abstract

This invention discloses a radar signal sorting method and apparatus. The method includes: acquiring radar signal information; the radar signal information includes the number of radar targets and the arrival time sequence information of intercepted radar pulses; establishing a radar pulse repetition frequency parameter model using the radar signal information; using the number of radar targets, performing radar tag allocation processing on the arrival time sequence information of the intercepted radar pulses to obtain a radar pulse path set; and using a spurious filtering discrimination model and the radar pulse repetition frequency parameter model to filter the radar pulse path set to obtain radar pulse sequence target information. This invention addresses the phenomena of pulse loss and spurious pulses by inputting non-ideal intercepted pulse trains and, through the design of specific likelihood factors, restoring the pulse characteristics before the occurrence of pulse loss and spurious pulses, thereby improving the accuracy of radar pulse signal sorting.

Description

Technical Field

[0001] This invention relates to the field of radar signal sorting, and more particularly to a radar signal sorting method and apparatus. Background Technology

[0002] In non-cooperative electromagnetic environments, the interleaved pulse streams intercepted by reconnaissance equipment originate from different radar radiation sources. Correctly sorting the pulse sequences corresponding to the radar signals from the obtained pulse streams is a crucial prerequisite for accurate radar signal analysis and identification, ensuring effective subsequent jamming. In recent years, with the development of radar technology, the battlefield electromagnetic environment has become unprecedentedly complex, posing significant challenges to radar pulse signal sorting. On the one hand, with the development of low probability of intercept (LPI) radar technology and the overlap of pulse signal parameters in the spatiotemporal frequency domain, reconnaissance equipment struggles to intercept all pulse signals, resulting in severe pulse loss. On the other hand, influenced by electromagnetic signals radiated by various ships at sea, in high-density signal environments, reconnaissance equipment easily intercepts a large number of stray pulses unrelated to the target, further increasing the sorting difficulty. Furthermore, radar operating modes are rapidly evolving, with complex PRI modulation modes constantly emerging. Traditional sequence search-based algorithms are no longer adequate for complex interleaved pulse sorting tasks, necessitating the research of new sorting methods. Especially for non-ideal scenarios involving pulse loss and stray pulses, existing methods struggle to correctly sort radar signals. Summary of the Invention

[0003] To address the problem that existing radar signal sorting methods struggle to correctly sort radar signals due to the rapid updates and iterations of radar operating modes and the continuous emergence of complex PRI modulation modes, this invention discloses a radar signal sorting method, comprising:

[0004] S1, acquire radar signal information; the radar signal information includes the number of radar targets and the arrival time sequence information of intercepted radar pulses;

[0005] S2, using the radar signal information, establish a radar pulse repetition frequency parameter model;

[0006] S3, using the number of radar targets, perform radar tag allocation processing on the arrival time sequence information of the intercepted radar pulses to obtain a radar pulse path set;

[0007] S4. Using the spurious filtering discrimination model and the radar pulse repetition frequency parameter model, the radar pulse path set is filtered to obtain radar pulse sequence target information; the radar pulse sequence target information is used to characterize the radar target corresponding to the intercepted radar pulse.

[0008] The number of radar targets is denoted by K; the arrival time sequence information of the intercepted radar pulses is denoted as {τ}. i},i=1,...,N, where, τ i Let N represent the arrival time of the i-th intercepted radar pulse, and N represent the number of intercepted radar pulses.

[0009] The radar pulse repetition frequency parameter model is used to characterize the probability density function of the pulse repetition frequency of intercepted radar pulses, including a first PRI model, a second PRI model, a third PRI model, and a fourth PRI model; the pulse repetition frequency is represented by PRI.

[0010] The first PRI model has a parameter space of Θ1 for the PRI that intercepts radar pulses. μ1 is the mean PRI of the intercepted radar pulses in the first PRI model. Let ξ1 be the variance of the PRI of the intercepted radar pulses in the first PRI model, and σ1 be the standard deviation of the PRI of the intercepted radar pulses in the first PRI model. The probability density function ξ1(p) of the PRI of the intercepted radar pulses in the first PRI model is... t The expression for ) is:

[0011]

[0012] Where Φ() is the cumulative normal distribution function, p t PRI is the radar pulse intercepted at time t;

[0013] The second PRI model has a parameter space of Θ2 for the PRI that intercepts radar pulses. μ2 is the mean PRI of the intercepted radar pulses in the second PRI model. Let ξ1 be the variance of the PRI of the intercepted radar pulses in the second PRI model, and σ2 be the standard deviation of the PRI of the intercepted radar pulses in the second PRI model. The probability density function ξ2(p1) of the PRI of the intercepted radar pulses in the second PRI model is... t The expression for ) is:

[0014]

[0015] Where, p t-1 Let PRI be the radar pulse intercepted at time t-1, and α be the sliding step size of the second PRI model;

[0016] The third PRI model has a parameter space of Θ3 for intercepting radar pulse sequences, where Θ3 = (A, B, π), and A = [a ij ] M×M A is the PRI state transition matrix, a ij Let π represent the probability of PRI transitioning from state i to state j, where M is the number of states in PRI, and π = (π1, π2, ..., π). M), where π is the initial state distribution of the PRI for the intercepted radar pulse sequence in the third PRI model, π l This represents the l-th initial state of the PRI in the third PRI model for intercepting radar pulse sequences, where l = 1, 2, ..., M. B represents the set of Gaussian distribution models for the PRI state. This represents the Gaussian distribution model of the m-th state of PRI. μ 3,m Let represent the mean of the Gaussian distribution of the m-th state of PRI. The variance of the Gaussian distribution of the m-th state of the PRI is q = 1, 2, ..., M. The PRI sequence value P of the intercepted radar pulse of the third PRI model is given. k Represented as P k ={p1,p2,...,p k}, P k In this context, p1 represents the value of the first PRI in the PRI sequence, and so on. k p in k Let ξ3 represent the value of the k-th PRI in the PRI sequence, where k represents the number of PRIs contained in the PRI sequence, and let ξ3(P) represent the probability distribution of the PRI sequence. k The calculation expression for ) is:

[0017]

[0018] α1(m)=π m f m (p1), m = 1, 2, ..., M,

[0019]

[0020] In the formula, f m for The corresponding Gaussian distribution probability density function, α1(m), represents the probability value of the first PRI in the PRI sequence being in the m-th state, and so on, where α... t+1 (m) represents the probability value of the (t+1)th PRI in the PRI sequence being in the m-th state, where α t (h) represents the probability value of the t-th PRI in the PRI sequence being in the h-th state, a hm p represents the element in the h-th row and m-th column of matrix A. t+1 This represents the value of the (t+1)th PRI in the PRI sequence, where t < k-1;

[0021] The fourth PRI model has the following expression for the PRI of intercepted radar pulses:

[0022] pt =Asin(2πft+φ)+c+ω t ,

[0023] Where A is the amplitude of the fourth PRI model, f is the frequency of the fourth PRI model, c is the amplitude offset of the fourth PRI model, t is the pulse number of the intercepted radar pulse, and ω t The Gaussian noise in the amplitude of the fourth PRI model. φ is the initial phase of the amplitude of the fourth PRI model, c > A; the parameter space of the PRI of the intercepted radar pulse of the fourth PRI model is Θ4. The probability density function ξ4(p) of the intercepted radar pulse in the fourth PRI model t The expression for ) is:

[0024]

[0025] in, The variance of the Gaussian noise in the fourth PRI model amplitude is given.

[0026] The step of using the number of radar targets to perform radar tag allocation processing on the arrival time sequence information of the intercepted radar pulses to obtain a radar pulse path set includes:

[0027] S31, The arrival time sequence information of the intercepted radar pulses is sorted according to time sequence to obtain the pulse time sequence;

[0028] S32, for the first arrival time information in the pulse time sequence, each radar tag in the radar tag set [1,...,K,spurious] is assigned sequentially to obtain K+1 initialized radar pulse paths; the radar pulse path is a sequence composed of several radar tags; the first K radar tags in the radar tag set [1,...,K,spurious] are used to characterize the 1st to the Kth radar to which the intercepted radar pulse belongs, respectively, and spurious is used to characterize the intercepted radar pulse as a spurious pulse;

[0029] S33, using the K+1 initialized radar pulse paths, a radar pulse path set is constructed; the storage limit of the radar pulse path set is (K+1). e There are 1 radar pulse path, where e is a preset exponent value and e is a positive integer greater than or equal to 4.

[0030] The method of using a spurious filtering discrimination model and a radar pulse repetition frequency parameter model to filter the radar pulse path set and obtain radar pulse sequence target information includes:

[0031] S401, initialize the iteration number t, and get t = 2;

[0032] S402, for the t-th arrival time information in the pulse time sequence, assign each radar tag in the radar tag set [1,...,K,spurious] in sequence to obtain the path tail node set; the path tail node set includes K+1 radar tags;

[0033] S403, for each radar pulse path, each radar tag in the path tail node set is added to the end of the radar pulse path, and the radar pulse path and the radar tags in the path tail node set are merged to obtain K+1 extended radar pulse paths.

[0034] S404, determine whether the iteration number t is greater than e-1, and obtain the storage discrimination result;

[0035] If the storage determination result is negative, proceed to step S405;

[0036] If the storage determination result is yes, proceed to step S409;

[0037] S405, store the extended radar pulse path obtained in S403 into the radar pulse path set;

[0038] S406, using the radar target similarity function, calculates the similarity value of each radar pulse path in the radar pulse path set;

[0039] S407, The calculated similarity value of each radar pulse path is stored in a similarity value set; the similarity value set includes the similarity value and its corresponding radar pulse path information;

[0040] S408, Set the iteration number t to increment by 1, and execute step S402;

[0041] S409, using the radar target similarity function, calculate the similarity value of the extended radar pulse path obtained in S403;

[0042] S410, sort the similarity values ​​of the similarity value set and the similarity values ​​of the extended radar pulse path in descending order of value, and search to obtain the first (K+1) values ​​before sorting. e The similarity value of the names and the corresponding radar pulse path information;

[0043] S411, Clear the set of similarity values ​​and sort the values ​​before (K+1). e The similarity values ​​of the names are stored in a similarity value set; the radar pulse path set is cleared, and the values ​​are sorted into (K+1) values.e The radar pulse paths corresponding to the similarity values ​​of the names are saved into the radar pulse path set.

[0044] S412, determine whether the iteration number t is greater than the number of intercepted radar pulses N, and obtain the result of stopping the iteration;

[0045] If the result of stopping iteration is yes, select the maximum similarity value from the set of similarity values, determine the radar pulse path corresponding to the maximum similarity value, and use the radar pulse path corresponding to the maximum similarity value as the target information of the radar pulse sequence;

[0046] If the result of stopping the iteration is negative, the iteration number t is incremented by 1, and step S402 is executed.

[0047] The radar target similarity function is expressed as follows:

[0048]

[0049]

[0050] In the formula, L is the calculated similarity value. For radar pulse path, ND represents the number of radar tags contained in the radar pulse path. Let Λ be the i-th radar tag in the radar pulse path. i S is the spurious pulse indication factor for the i-th radar tag; k (ii) is the radar pulse path The radar tag k that appears in the radar pulse path When it appears for the iith time, in the radar pulse path The position index in S, s(k) is the position index of S. k The number of elements contained in (), where β is the preset stray pulse filtering coefficient; For the Sth k (ii) the time interval between intercepted radar pulses, For the Sth k (ii) the arrival time of the intercepted radar pulse.

[0051] The present invention also discloses a data processing system for radar signal sorting, comprising:

[0052] A first memory, a first processor, and a computer program stored in the first memory and executable on the first processor, wherein the radar signal sorting method is implemented when the first processor executes the program.

[0053] The present invention also discloses a data processing apparatus for radar signal sorting, the apparatus comprising:

[0054] Memory containing executable program code;

[0055] A processor coupled to the memory;

[0056] The processor calls the executable program code stored in the memory to execute the data processing method for radar signal sorting.

[0057] The present invention also discloses a computer-storable medium storing computer instructions, which, when invoked, are used to execute the data processing method for radar signal sorting.

[0058] The beneficial effects of this invention are as follows:

[0059] 1. This invention discloses a radar signal sorting method. Addressing pulse loss and spurious pulse phenomena, it inputs a non-ideal intercepted pulse train and, by designing a specific likelihood factor, restores the pulse characteristics before the occurrence of pulse loss and spurious pulses, thereby improving the accuracy of radar pulse signal sorting. Furthermore, the proposed model has a solution space complexity of O(K). N This is an NP-hard problem. This invention assigns all possible labels to each pulse, calculates and retains paths with high similarity values, and sets a limited buffer size to avoid exponential explosion caused by path iteration, thus reducing the solution complexity of this model to O(K). 5 N).

[0060] 2. This invention takes into account the pulse loss and stray pulse phenomena in the environment, thereby improving the radar signal sorting accuracy in non-ideal environments; this invention reduces the complexity of model solving and saves computation time.

[0061] 4. The method disclosed in this invention is applicable to situations where some prior knowledge of radar is known. It extracts known radar pulses through a model in order to analyze and identify the parameters of unknown radar pulses. Attached Figure Description

[0062] Figure 1 This is a flowchart illustrating the implementation of the radar signal sorting method disclosed in this invention.

[0063] Figure 2 This is a schematic diagram of missing pulses during the implementation of the method of the present invention;

[0064] Figure 3 This is a schematic diagram of stray pulses during the implementation of the method of the present invention;

[0065] Figure 4This is a comparison chart of the sorting performance of the method of the present invention under different β and stray pulse rates;

[0066] Figure 5 This is a comparison chart of the sorting results of the method of this invention and the traditional sorting method. Detailed Implementation

[0067] To better understand the content of this invention, two embodiments are provided here.

[0068] Example 1

[0069] During radar signal reconnaissance, the pulse stream intercepted by the reconnaissance equipment is the Pulse Description Word (PDW) sequence, which contains six parameters: Time of Arrival (TOA), Pulse Repetition Interval (PRI), Pulse Width (PW), Pulse Amplitude (PA), Radio Frequency (RF), and Direction of Arrival (DOA). Among them, PRI represents the temporal pattern of the pulse sequence and is an important parameter required for sorting.

[0070] This invention discloses a radar signal sorting method, such as... Figure 1 As shown, it includes:

[0071] S1, acquire radar signal information; the radar signal information includes the number of radar targets and the arrival time sequence information of intercepted radar pulses;

[0072] S2, using radar signal information, establish a radar pulse repetition frequency parameter model;

[0073] S3, using the number of radar targets, performs radar tag allocation processing on the arrival time sequence information of the intercepted radar pulses to obtain a set of radar pulse paths;

[0074] S4. Using the spurious filtering discrimination model and the radar pulse repetition frequency parameter model, the radar pulse path set is filtered to obtain radar pulse sequence target information; the radar pulse sequence target information is used to characterize the radar target corresponding to the intercepted radar pulse.

[0075] The number of radar targets is denoted by K; the arrival time sequence information of the intercepted radar pulses is denoted as {τ}. i},i=1,...,N, where, τ i Let N represent the arrival time of the i-th intercepted radar pulse, and N represent the number of intercepted radar pulses.

[0076] The radar pulse repetition frequency parameter model is used to characterize the probability density function of the pulse repetition frequency of intercepted radar pulses, including a first PRI model, a second PRI model, a third PRI model, and a fourth PRI model. The pulse repetition frequency is denoted by PRI.

[0077] The first PRI model has a parameter space of Θ1 for the PRI that intercepts radar pulses. μ1 is the mean PRI of the intercepted radar pulses in the first PRI model. Let ξ1 be the variance of the PRI of the intercepted radar pulses in the first PRI model, and σ1 be the standard deviation of the PRI of the intercepted radar pulses in the first PRI model. The probability density function ξ1(p) of the PRI of the intercepted radar pulses in the first PRI model is... t The expression for ) is:

[0078]

[0079] Where Φ(·) is the cumulative normal distribution function, p t PRI is the radar pulse intercepted at time t;

[0080] The second PRI model has a parameter space of Θ2 for the PRI that intercepts radar pulses. μ2 is the mean PRI of the intercepted radar pulses in the second PRI model. Let ξ1 be the variance of the PRI of the intercepted radar pulses in the second PRI model, and σ2 be the standard deviation of the PRI of the intercepted radar pulses in the second PRI model. The probability density function ξ2(p1) of the PRI of the intercepted radar pulses in the second PRI model is... t The expression for ) is:

[0081]

[0082] Where, p t-1 Let PRI be the radar pulse intercepted at time t-1, and α be the sliding step size of the second PRI model;

[0083] The third PRI model has a parameter space of Θ3 for intercepting radar pulse sequences, where Θ3 = (A, B, π), and A = [a ij ] M×M A is the PRI state transition matrix, a ij Let π represent the probability of PRI transitioning from state i to state j, where M is the number of states in PRI, and π = (π1, π2, ..., π). M ), where π is the initial state distribution of the PRI for the intercepted radar pulse sequence in the third PRI model, π lThis represents the l-th initial state of the PRI in the third PRI model for intercepting radar pulse sequences, where l = 1, 2, ..., M. B represents the set of Gaussian distribution models for the PRI state. This represents the Gaussian distribution model of the m-th state of PRI. μ 3,m Let represent the mean of the Gaussian distribution of the m-th state of PRI. The variance of the Gaussian distribution of the m-th state of the PRI is q = 1, 2, ..., M. The PRI sequence value P of the intercepted radar pulse of the third PRI model is given. k Represented as P k ={p1,p2,...,p k}, P k In this context, p1 represents the value of the first PRI in the PRI sequence, and so on. k p in k Let ξ3 represent the value of the k-th PRI in the PRI sequence, where k represents the number of PRIs contained in the PRI sequence, and let ξ3(P) represent the probability distribution of the PRI sequence. k The calculation expression for ) is:

[0084]

[0085] α1(m)=π m f m (p1), m = 1, 2, ..., M,

[0086]

[0087] In the formula, f m for The corresponding Gaussian distribution probability density function, α1(m), represents the probability value of the first PRI in the PRI sequence being in the m-th state, and so on, where α... t+1 (m) represents the probability value of the (t+1)th PRI in the PRI sequence being in the m-th state, where α t (h) represents the probability value of the t-th PRI in the PRI sequence being in the h-th state, a hm p represents the element in the h-th row and m-th column of matrix A. t+1 This represents the value of the (t+1)th PRI in the PRI sequence, where t < k-1;

[0088] The fourth PRI model has the following expression for the PRI of intercepted radar pulses:

[0089] p t =Asin(2πft+φ)+c+ω t ,

[0090] Where A is the amplitude of the fourth PRI model, f is the frequency of the fourth PRI model, c is the amplitude offset of the fourth PRI model, t is the pulse number of the intercepted radar pulse, and ω t The Gaussian noise in the amplitude of the fourth PRI model. φ is the initial phase of the amplitude of the fourth PRI model. Since PRI is positive, c > A; the parameter space of the PRI of the intercepted radar pulse of the fourth PRI model is Θ4. The probability density function ξ4(p) of the intercepted radar pulse in the fourth PRI model t The expression for ) is:

[0091]

[0092] in, The variance of the Gaussian noise in the fourth PRI model amplitude is given.

[0093] The process of using the number of radar targets to perform radar tag allocation on the arrival time sequence information of intercepted radar pulses to obtain a radar pulse path set includes:

[0094] S31, The arrival time sequence information of the intercepted radar pulses is sorted according to time sequence to obtain the pulse time sequence;

[0095] S32, for the first arrival time information in the pulse time sequence, each radar tag in the radar tag set [1,...,K,spurious] is sequentially assigned to obtain K+1 initialized radar pulse paths; the radar pulse path is a sequence composed of several radar tags; the first K radar tags in the radar tag set [1,...,K,spurious] are used to represent the 1st to the Kth radar to which the intercepted radar pulse belongs, respectively, and spurious is used to represent that the intercepted radar pulse is a spurious pulse; specifically, the intercepted radar pulse represented by the radar tag is the intercepted radar pulse corresponding to the arrival time sequence information of the intercepted radar pulse. A radar tag in the radar pulse path serves as a node in the radar pulse path.

[0096] S33, using the K+1 initialized radar pulse paths, a radar pulse path set is constructed; the storage limit of the radar pulse path set is (K+1). e A radar pulse path, where e is a preset exponent value and e is a positive integer greater than or equal to 4;

[0097] The method of using a spurious filtering discrimination model and a radar pulse repetition frequency parameter model to filter the radar pulse path set and obtain radar pulse sequence target information includes:

[0098] S401, initialize the iteration number t, and get t = 2;

[0099] S402, for the t-th arrival time information in the pulse time sequence, assign each radar tag in the radar tag set [1,...,K,spurious] in sequence to obtain the path tail node set; the path tail node set includes K+1 radar tags;

[0100] S403, for each radar pulse path, each radar tag in the path tail node set is added to the end of the radar pulse path, and the radar pulse path and the radar tags in the path tail node set are merged to obtain K+1 extended radar pulse paths; the path merging process is to also use the radar tags in the path tail node set as the last node of the original radar pulse path.

[0101] For K+1 initialized radar pulse paths, after the operation described in S403, (K+1) is obtained. 2 An extended radar pulse path; for (K+1) 2 The radar pulse path, after being processed by S403, yields (K+1). 3 An extended radar pulse path;

[0102] S404, determine whether the iteration number t is greater than e-1, and obtain the storage discrimination result;

[0103] If the storage determination result is negative, proceed to step S405;

[0104] If the storage determination result is yes, proceed to step S409;

[0105] S405, store the extended radar pulse path obtained in S403 into the radar pulse path set;

[0106] S406, using the radar target similarity function, calculates the similarity value of each radar pulse path in the radar pulse path set;

[0107] S407, The calculated similarity value of each radar pulse path is stored in a similarity value set; the similarity value set includes the similarity value and its corresponding radar pulse path information;

[0108] S408, Set the iteration number t to increment by 1, and execute step S402;

[0109] S409, using the radar target similarity function, calculate the similarity value of the extended radar pulse path obtained in S403;

[0110] S410, sort the similarity values ​​of the similarity value set and the similarity values ​​of the extended radar pulse path in descending order of value, and search to obtain the first (K+1) values ​​before sorting. e The similarity value of the names and the corresponding radar pulse path information;

[0111] S411, Clear the set of similarity values ​​and sort the values ​​before (K+1). e The similarity values ​​of the names are stored in a similarity value set; the radar pulse path set is cleared, and the values ​​are sorted into (K+1) values. e The radar pulse paths corresponding to the similarity values ​​of the names are saved into the radar pulse path set.

[0112] S412, determine whether the iteration number t is greater than the number of intercepted radar pulses N, and obtain the result of stopping the iteration;

[0113] If the result of stopping iteration is yes, the maximum similarity value is selected from the set of similarity values, and the radar pulse path corresponding to the maximum similarity value is determined; the radar pulse path corresponding to the maximum similarity value is used as the radar pulse sequence target information; the radar tag in the radar pulse path is used to characterize the radar target corresponding to the intercepted radar pulse.

[0114] If the result of stopping the iteration is negative, the iteration number t is incremented by 1, and step S402 is executed.

[0115] The radar target similarity function is expressed as follows:

[0116]

[0117]

[0118] In the formula, L is the calculated similarity value. For radar pulse path, ND represents the number of radar tags contained in the radar pulse path. Let Λ be the i-th radar tag in the radar pulse path. i S is the spurious pulse indication factor for the i-th radar tag; k (ii) is the radar pulse path The radar tag k that appears in the radar pulse path When it appears for the iith time, in the radar pulse path The position index in S, s(k) is the position index of S.k The number of elements contained in () is β, which is a preset spurious pulse filtering coefficient. The position number is in the radar pulse path. The radar pulse path is obtained by changing the order in which the radar tags appear. The first radar tag in the array has a location number of 1, and the last radar tag has a location number of ND. k () represents the radar pulse path The set of location indices of all radar tags k appearing in the array. For example, when... When, S1=(1,3,4), S2=(2,5,6). Δτ Sk(ii) =τ Sk(ii) -τ Sk(ii-1) ,Δτ Sk(ii) For the Sth k (ii) the time interval between intercepted radar pulses, τ Sk(ii) For the Sth k (ii) The arrival time of each intercepted radar pulse. The radar target similarity function uses the function max(·) to determine whether pulse loss has occurred. If no pulse loss has occurred, then... At the location where one pulse is missing, there is At the location where two pulses are missing, there is spurious is used to indicate receiving spurious pulses. This is a stray pulse constraint term, where β is a preset stray pulse filtering coefficient. By selecting a reasonable value for β, stray pulses can be filtered.

[0119] This invention discloses a data processing system for radar signal sorting, comprising:

[0120] A first memory, a first processor, and a computer program stored in the first memory and executable on the first processor, wherein the radar signal sorting method is implemented when the first processor executes the program.

[0121] This invention discloses a data processing apparatus for radar signal sorting, the apparatus comprising:

[0122] Memory containing executable program code;

[0123] A processor coupled to the memory;

[0124] The processor calls the executable program code stored in the memory to execute the data processing method for radar signal sorting.

[0125] The present invention discloses a computer-storable medium storing computer instructions, which, when invoked, are used to execute the data processing method for radar signal sorting.

[0126] Example 2

[0127] In this embodiment, as Figure 1 As shown, the present invention provides a radar signal sorting method, comprising the following steps:

[0128] S1, Input the number of radar targets K, and the pulse arrival time {τ} i}, i=1,...,N, PRI parameter model. In this embodiment, it is assumed that the number of radar targets K is 3, the PRI modulation type is the first PRI, and the parameter space is Θ=(31,1 2 ), Θ = (67, 1 2 ), Θ = (55, 1 2 );

[0129] PRI parameter models can be classified into the following categories:

[0130] First PRI Model: The parameter space of the first PRI model is Θ = (μ, σ 2 ), using the mean μ and variance σ 2 The truncated Gaussian distribution indicates that its PRIp t The probability density function ξ(p) t )as follows:

[0131]

[0132] In the formula, Φ(·) is the cumulative normal distribution function. Since PRI is a positive value, the cutoff interval of this distribution is (0,+∞).

[0133] The second PRI model: The parameter space of the second PRI model is Θ = (α, σ). 2 Considering only the positive slip case, PRI can be described as p within one slip cycle. t =α+p t-1 +ω t Where the sliding step size α > 0, ω t ~N(0,σ 2 The noise is Gaussian. Since the PRI modulation type exhibits Markov property, the conditional probability density function ξ(p) is used. t |p t-1 The distribution is represented by the following formula:

[0134]

[0135] The third PRI model: The parameter space for this modulation type is Θ = (A, B, π), where A = [a ij ] M×M It is a transition matrix for M states, π = (π1, π2, ..., π). M ) is the initial state distribution. The Gaussian model representing each state, i.e. PRI sequence P k ={p1,p2,...,p k The probability distribution ξ(P) of} k )as follows:

[0136]

[0137] In the formula, Let be the probability density function of a Gaussian distribution.

[0138] The fourth PRI model: The fourth PRI model can be represented as p t =Asin(2πft+φ)+c+ω t Its parameter space Θ=(A,f,c,σ) 2 ,φ), where A is the amplitude, f is the frequency, c is the offset, and ω t ~N(0,σ 2 Let φ be Gaussian noise and φ be the initial phase. Since PRI is positive, c > A. Its probability density function is ξ(p t )as follows:

[0139]

[0140] In the formula, t is the pulse number.

[0141] S2. Assign labels 1, ..., K, spurious to the first arriving pulse τ1, where spurious indicates that the pulse is a spurious pulse. Set i = 2 and the buffer size to K. 4 The initialization path is placed into the buffer.

[0142] S3. For each path in the buffer, τ i Assign labels 1,...,K, and calculate the similarity function value. The formula for calculating the similarity function is:

[0143]

[0144]

[0145] In the formula, For pulse train label, S k (i) is the sequence number of the i-th pulse matched to radar k in the interleaved string. For example, when At that time, S1 = (1,3,4) and S2 = (2,5,6). The pulse interval is used to determine whether a pulse is missing using the function max(·). If no pulse is missing, then... At the location where one pulse is missing, there is At the location where two pulses are missing, there is spurious is used to receive stray pulses. For stray pulse constraint terms, β is a coefficient, and Λ i This is a stray pulse indicator factor; by selecting an appropriate β value, stray pulses can be filtered out.

[0146] The likelihood function design is based on the following two considerations:

[0147] 1. Pulse absence diagram as shown below Figure 2 As shown, the interval between two adjacent pulses of radiation source 2 is p. t and p t+1 Due to the influence of missing pulses, the new pulse interval becomes p. t +p t+1 Clearly, the latter does not satisfy the original PRI distribution, therefore it needs to be corrected for missing impulses. Since the difference between adjacent PRIs is relatively small, assume p... t ≈p t+1 Therefore, under the influence of a missing pulse from radiation source 2, the new pulse interval p new =p t +p t+1 It can be approximated as twice the value before the deletion, i.e., p new ≈2p t Use p new / 2 restores the pulse interval before the missing pulse, then ξ(p new / 2)>ξ(p new Therefore, the function max(·) can be used to determine whether a pulse is missing.

[0148] 2. The stray pulse diagram is shown below. Figure 3 As shown, if the stray pulse is identified as belonging to radiation source 1, the original pulse interval p t It will become p t_1 and p t_2 This will inevitably disrupt its timing correlation characteristics, therefore the spurious pulse scenario needs to be considered separately. Utilizing Filtering stray pulses, when β>logξ(p) t_1 )+logξ(p t_2 )-logξ(p t If the purple pulse is spurious, it is considered to be more likely to belong to radiation source 1; otherwise, the pulse is considered to belong to radiation source 1.

[0149] S4. The buffer retains the top K values ​​with the highest similarity function values. 4 There are several paths, and we set i = i + 1.

[0150] S5. When i = N, the iteration ends. Extract the path with the highest similarity function value from the buffer and output the label corresponding to the pulse train. Otherwise, return to step 3.

[0151] Figure 4 This is a comparison of sorting performance under different β values ​​and stray pulse rates. It can be seen that when β is in the range [-10, -6], Acc is at a relatively high level; when β is less than -10 and greater than -6, Acc decreases significantly. This is because if β is too small, even with a pulse interval p... t_1 and p t_2 The distribution no longer conforms to the PRI distribution of radiation source 1, but the above inequality still does not hold, failing to effectively filter spurious pulses, resulting in a large number of spurious terms in the target signal; if the value of β is too large, such that β > logξ(p t If the pulse from radiation source 1 is identified as a stray term, then the constraint term is determined to be invalid. The dominant signal causes a large number of target pulse signals to be misclassified as spurious terms. Therefore, in this embodiment, β = -8 is chosen.

[0152] Figure 5 This figure compares the sorting results of the proposed method and the traditional sorting method. In the figure, blue labels represent correctly sorted pulses, and red labels represent incorrectly sorted pulses. It can be seen that the traditional method exhibits continuous incorrect sorting near the location of missing pulses, as indicated by the dotted line in the figure. The proposed method, after correcting for missing pulses, avoids this situation.

[0153] The above description is merely an embodiment of this application and is not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A radar signal sorting method, characterized in that, include: S1, acquire radar signal information; The radar signal information includes the number of radar targets and the arrival time sequence information of the intercepted radar pulses; S2, using the radar signal information, establish a radar pulse repetition frequency parameter model; S3, using the number of radar targets, perform radar tag allocation processing on the arrival time sequence information of the intercepted radar pulses to obtain a radar pulse path set; S4. Using the spurious filtering discrimination model and the radar pulse repetition frequency parameter model, the radar pulse path set is filtered to obtain radar pulse sequence target information; The radar pulse sequence target information is used to characterize the radar target corresponding to the intercepted radar pulse.

2. The radar signal sorting method as described in claim 1, characterized in that, The number of radar targets is denoted by K; the arrival time sequence information of the intercepted radar pulses is denoted as... ,in, This represents the arrival time of the i-th intercepted radar pulse. This indicates the number of radar pulses intercepted.

3. The radar signal sorting method as described in claim 2, characterized in that, The radar pulse repetition frequency parameter model is used to characterize the probability density function of the pulse repetition frequency of intercepted radar pulses, including a first PRI model, a second PRI model, a third PRI model, and a fourth PRI model; the pulse repetition frequency is represented by PRI.

4. The radar signal sorting method as described in claim 3, characterized in that, The parameter space of the first PRI model, which intercepts radar pulses, is: , , The mean PRI of the intercepted radar pulses in the first PRI model. Let be the variance of the PRI that intercepts radar pulses in the first PRI model. Let be the standard deviation of the PRI of intercepted radar pulses in the first PRI model, and let be the probability density function of the PRI of intercepted radar pulses in the first PRI model. The expression is: in, The cumulative normal distribution function is... PRI is the radar pulse intercepted at time t; The parameter space of the second PRI model, which intercepts radar pulses, is: , , The mean PRI of the intercepted radar pulses in the second PRI model. Let be the variance of the PRI that intercepts radar pulses in the second PRI model. Let be the standard deviation of the PRI of intercepted radar pulses in the second PRI model, and let be the probability density function of the PRI of intercepted radar pulses in the second PRI model. The expression is: ， in, PRI is the radar pulse intercepted at time t-1. This represents the sliding step size of the second PRI model; The parameter space of the third PRI model that intercepts radar pulse sequences is: , ,in, , It is the PRI state transition matrix. Let represent the probability that PRI transitions from state i to state j, and M be the number of states in PRI. , This is the initial state distribution of the PRI in the intercepted radar pulse sequence in the third PRI model. The third PRI in the third PRI model represents the PRI that intercepts the radar pulse sequence. An initial state, , , This represents the set of Gaussian distribution models for the PRI state. This represents the Gaussian distribution model of the m-th state of PRI. , Let represent the mean of the Gaussian distribution of the m-th state of PRI. This represents the variance of the Gaussian distribution of the m-th state of PRI. The PRI sequence values ​​of the intercepted radar pulses in the third PRI model Represented as, , In This represents the value of the first PRI in the PRI sequence, and so on. In This represents the value of the k-th PRI in the PRI sequence, where k represents the number of PRIs contained in the PRI sequence, and the probability distribution of the PRI sequence. The calculation expression is: In the formula, for The corresponding probability density function of the Gaussian distribution, This represents the probability value of the first PRI in the PRI sequence being in the m-th state, and so on. This represents the probability value of the (t+1)th PRI in the PRI sequence being in the m-th state. This represents the probability value of the t-th PRI in the PRI sequence being in the h-th state. Represents the matrix The element in the h-th row and m-th column, This represents the value of the (t+1)th PRI in the PRI sequence. ; The fourth PRI model has the following expression for the PRI of intercepted radar pulses: ， in, For the magnitude of the fourth PRI model, This represents the fourth PRI model frequency. This represents the amplitude offset of the fourth PRI model. The pulse number of the intercepted radar pulse. The Gaussian noise in the amplitude of the fourth PRI model. , The initial phase of the fourth PRI model amplitude. The parameter space of the PRI model for intercepting radar pulses in the fourth PRI model is: , The probability density function of the PRI for intercepting radar pulses in the fourth PRI model. The expression is: in, The variance of the Gaussian noise in the fourth PRI model amplitude is given.

5. The radar signal sorting method as described in claim 4, characterized in that, The step of using the number of radar targets to perform radar tag allocation processing on the arrival time sequence information of the intercepted radar pulses to obtain a radar pulse path set includes: S31, The arrival time sequence information of the intercepted radar pulses is sorted according to time sequence to obtain the pulse time sequence; S32, for the first arrival time information in the pulse time sequence, sequentially assign radar tag sets [ For each radar tag in the set, K+1 initialized radar pulse paths are obtained; the radar pulse path is a sequence composed of several radar tags; the radar tag set [ The first K radar tags in the image are used to identify the radar from the 1st to the Kth radar to which the intercepted radar pulse belongs, respectively, and spurious is used to indicate that the intercepted radar pulse is a spurious pulse; S33, using the K+1 initialized radar pulse paths, a radar pulse path set is constructed; the storage limit of the radar pulse path set is... A radar pulse path, where e is a preset exponential value. It is a positive integer greater than or equal to 4.

6. The radar signal sorting method as described in claim 5, characterized in that, The method of using a spurious filtering discrimination model and a radar pulse repetition frequency parameter model to filter the radar pulse path set and obtain radar pulse sequence target information includes: S401, initialize the iteration number t, and get t=2; S402, for the t-th arrival time information in the pulse time sequence, sequentially assign radar tag sets [ For each radar tag in the [], a path tail node set is obtained; the path tail node set includes K+1 radar tags; S403, for each radar pulse path, each radar tag in the path tail node set is added to the end of the radar pulse path, and the radar pulse path and the radar tags in the path tail node set are merged to obtain K+1 extended radar pulse paths. S404, determine whether the iteration number t is greater than e-1, and obtain the storage discrimination result; If the storage determination result is negative, proceed to step S405; If the storage determination result is yes, proceed to step S409; S405, store the extended radar pulse path obtained in S403 into the radar pulse path set; S406, using the radar target similarity function, calculates the similarity value of each radar pulse path in the radar pulse path set; S407, The calculated similarity value of each radar pulse path is stored in a similarity value set; the similarity value set includes the similarity value and its corresponding radar pulse path information; S408, Set the iteration number t to increment by 1, and execute step S402; S409, using the radar target similarity function, calculate the similarity value of the extended radar pulse path obtained in S403; S410, sort the similarity values ​​in the similarity value set and the similarity values ​​of the extended radar pulse path in descending order of value, and search to obtain the values ​​before sorting. The similarity value of the names and the corresponding radar pulse path information; S411, Clear the set of similarity values ​​and sort the values ​​before... The similarity values ​​of the names are stored in a similarity value set; the radar pulse path set is cleared, and the values ​​are sorted before... The radar pulse paths corresponding to the similarity values ​​of the names are saved into the radar pulse path set. S412, determine whether the iteration number t is greater than the number of intercepted radar pulses N, and obtain the result of stopping the iteration; If the result of stopping iteration is yes, select the maximum similarity value from the set of similarity values, determine the radar pulse path corresponding to the maximum similarity value, and use the radar pulse path corresponding to the maximum similarity value as the target information of the radar pulse sequence; If the result of stopping the iteration is negative, the iteration number t is incremented by 1, and step S402 is executed.

7. The radar signal sorting method as described in claim 6, characterized in that, The radar target similarity function is expressed as follows: In the formula, The calculated similarity value, For radar pulse path, , This represents the number of radar tags contained in the radar pulse path. , Let i be the i-th radar tag in the radar pulse path. is the spurious pulse indication factor for the i-th radar tag; Radar pulse path The radar tag k that appears in the radar pulse path When it appears for the iith time, in the radar pulse path Position number in the middle, for The number of elements contained. This is the preset stray pulse filtering coefficient; , For the first The time interval between intercepted radar pulses, For the first The arrival time of each intercepted radar pulse.

8. A data processing system for radar signal sorting, characterized in that, include: A first memory, a first processor, and a computer program stored in the first memory and executable on the first processor, wherein the first processor, when executing the program, implements the radar signal sorting method as described in any one of claims 1-7.

9. A data processing device for radar signal sorting, characterized in that, The device includes: Memory containing executable program code; A processor coupled to the memory; The processor calls the executable program code stored in the memory to execute the radar signal sorting method as described in any one of claims 1-7.

10. A computer-storable medium, characterized in that, The computer storage medium stores computer instructions, which, when invoked, are used to execute the radar signal sorting method as described in any one of claims 1-7.

Images