Rotary interferometer antenna signal sorting method based on iteration peak taking

Abstract

The invention discloses a rotating interferometer antenna signal sorting method based on iteration peak taking, and aims to solve the problem that a pre-sorting process based on a direction finding angle cannot be introduced into a signal sorting process of a rotating interferometer positioning system antenna. Therefore, the problem of signal sorting with high-density signals and low geometric constraints in the main signal sorting process is solved. According to the method, a signal sorting method based on voting positioning and iterative peak taking is adopted, a process of sorting known target signals and then sorting non-unknown target signals is utilized, and sorting logics of sorting strong parameter condition constraint radiation sources, sorting weak parameter condition constraint radiation sources and iterative peak taking grid position condition constraint are combined. Accurate signal sorting of the rotary interferometer positioning system antenna is realized.

Description

Technical Field

[0001] The invention belongs to the field of signal processing, and in particular relates to a rotating interferometer antenna signal sorting method based on iterative peak extraction. Background Art

[0002] Rotating interferometer positioning is a common radar signal reception and positioning system. Its positioning principle is to construct countless virtual baselines at various moments through the rotation of a long baseline. Each virtual long baseline is then used to measure and locate the phase difference of ground emitters. However, because long baselines introduce ambiguity in phase difference measurement, phase difference ambiguity estimation is typically performed by constructing a geographic grid and performing voting positioning. This process requires high accuracy in sorting the associated pulse sequences of emitters. The introduction of interfering pulses from different emitters directly reduces the accuracy of the voting positioning cost function, ultimately resulting in ambiguous and false positioning of the emitters. Therefore, in a rotating interferometer positioning system, accurate sorting of emitter pulses not only affects the accuracy of emitter parameter estimation but also the accuracy of emitter position estimation. However, unlike fixed interferometer positioning systems, the direction-finding information of each measured pulse in a rotating interferometer positioning system cannot be directly calculated. Therefore, a pre-sorting process based on direction-finding angles cannot be introduced into the emitter sorting process. This leads to the signal sorting challenge of high-density signals with low geometric constraints. Summary of the Invention

[0003] The present invention proposes a method for signal sorting of a rotating interferometer antenna based on iterative peak selection. This method aims to solve the problem that the rotating interferometer positioning system antenna cannot introduce a pre-sorting process based on direction finding angles during the signal sorting process, thereby facing the signal sorting problem of high-density signals and low geometric constraints during the main signal sorting process. The present invention adopts a signal sorting method based on voting positioning and iterative peak selection. It utilizes the process of sorting known target signals first and then sorting non-unknown target signals. It combines the sorting logic of first sorting strong parameter condition constraint radiation sources and then sorting weak parameter condition constraint radiation sources, and iterative peak selection grid position condition constraint to achieve accurate signal sorting of the rotating interferometer positioning system antenna.

[0004] The technical solution to realize the present invention is: a method for sorting antenna signals of a rotating interferometer based on iterative peak extraction, characterized by the following steps:

[0005] Step 1: Classify the known target knowledge base information target according to the characteristics of the radiation source:

[0006] The characteristics of the radiation source include the frequency characteristics of the radiation source, the pulse width characteristics of the radiation source and the repetition characteristics of the radiation source;

[0007] The frequency characteristics of the radiation source include two types: fixed frequency range and non-agile type, and agile frequency type;

[0008] The pulse width characteristics of the radiation source include two types: fixed pulse width and pulse width interval.

[0009] The repetition period characteristics of the radiation source include two types: fixed and optional repetition period, and repetition period interval variation.

[0010] Step 2: According to the known target knowledge base information target, classify and filter the known target-related pulses; according to the constraint ability of the known target knowledge base information target, the idea of ​​first filtering the pulse sequence of the strong constraint type radiation source signal and then filtering the pulse sequence of the weak constraint type radiation source signal is adopted. By continuously diluting the pulse sequence, the initial pulse sequence {pdw} is screened and sorted in sequence, and the pulse sequence corresponding to the successfully sorted radiation source is gradually extracted from the initial pulse sequence {pdw}, thereby improving the reliability of the known target pulse screening and obtaining the initial screening pulse sequence {pdw_filter pdw_id}, where pdw_id represents the number of the pulse sequence.

[0011] Step 3: Use the initial filtering pulse sequence {pdw_filter pdw_id} Carry out voting positioning iterative peak signal sorting to obtain the pulse sequence {pdw_filter pdw_id} vote_id，peak .

[0012] Step 4: Perform signal sorting on the remaining unknown target-related pulses to complete the signal sorting of all pulse sequences:

[0013] Based on the known target knowledge base information target, the signal is sorted and the associated pulse sequence of the radiation source is removed from the original pulse sequence {pdw} of the processing cycle, and the remaining unknown target pulse sequence {pdw} is used. non-key , carry out signal sorting of the remaining unknown target associated pulses.

[0014] Compared with the prior art, the present invention has the following significant advantages:

[0015] (1) Parameter condition pre-sorting: Classify the known target knowledge base and, based on the strength of the parameter constraints of the radiation source’s frequency, pulse width, and repetition rate, first screen the pulse sequences of the radiation source signal with strong constraints and then screen the pulse sequences of the radiation source signal with weak constraints. By gradually diluting the pulse sequences to be sorted, the reliability of the signal parameter condition pre-sorting is improved.

[0016] (2) Position condition pre-sorting: The geographical grid constraint of phase difference voting positioning is used to solve the disadvantage that single pulse cannot find the direction and perform signal position condition pre-sorting under the rotating system. This realizes signal position condition pre-sorting and reduces the difficulty of subsequent main sorting.

[0017] (3) Iterative peak pre-sorting: The idea of ​​iteratively extracting the pulse sequence associated with the sorted radiation source is adopted. The voting positioning grid histogram is updated after each main sorting. This can gradually dilute the pulse sequence to be sorted, improve the accuracy of position condition pre-sorting, and solve the problem of multiple geographical distribution of the same type of radiation source;

[0018] (4) Unknown signal sorting: For unknown target signals, the method of first sorting fixed frequency pulses, then merging the radiation source pulse sequence based on the radiation source grid position, and then re-sorting the signal is adopted to solve the iterative signal sorting of the remaining unknown target-related pulses. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 This is a flow chart of a rotating interferometer antenna signal sorting method based on iterative peak extraction according to an embodiment of the present invention.

[0020] Figure 2 Schematic diagram of the principle of iterative peak selection by voting positioning according to an embodiment of the present invention. DETAILED DESCRIPTION

[0021] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0022] The technical solutions between the various embodiments of the present invention can be combined with each other, but they must be based on the fact that ordinary technicians in this field can implement them. When the combination of technical solutions is mutually contradictory or cannot be implemented, it should be deemed that such a combination of technical solutions does not exist and is not within the scope of protection required by the present invention.

[0023] The following will further introduce the specific implementation methods, as well as the technical difficulties and inventive points of this invention in combination with this design example.

[0024] In response to the challenges faced by this technical scenario, the present invention adopts a signal sorting method based on voting positioning and iterative peak extraction, using the process of sorting known signals first and then unknown signals, combined with step-by-step sorting and iterative peak extraction and gradual dilution logic to achieve precise sorting under the rotating interferometer positioning system.

[0025] Combine Figure 1 The present invention provides a method for sorting antenna signals of a rotating interferometer based on iterative peak extraction, comprising the following steps:

[0026] Step 1: Classify the known target knowledge base information. Based on the characteristics of the radiation source, the frequency characteristics of the radiation source include two types: fixed frequency range and non-agile type, and agile frequency type. The pulse width characteristics of the radiation source include two types: fixed pulse width with optional pulse width, and pulse width interval with optional pulse width. The repetition characteristics of the radiation source include two types: fixed repetition with optional pulse width, and repetition interval with variable repetition.

[0027] According to the frequency, pulse width and repetition constraint of the radiation source signal, the target knowledge base information is divided into 8 categories, namely:

[0028] Category 1: Radiating sources with fixed frequency range and non-agile type, fixed and optional pulse width, and fixed and optional repetition frequency;

[0029] Category 2: Radiators with fixed frequency range and non-agile type, fixed and optional pulse width, and variable frequency interval;

[0030] Category 3: The frequency range is fixed and the type is non-agile, the pulse width interval varies, and the radiation source is fixed and optional;

[0031] Category 4: Radiators with fixed frequency range and non-agile type, varying pulse width and repetition frequency ranges;

[0032] Category 5: Fixed frequency range and agile type, fixed and optional pulse width, fixed and optional radiation source;

[0033] Category 6: A radiation source with a fixed frequency range and agile type, fixed and optional pulse width, and variable frequency range;

[0034] Category 7: Fixed frequency range and agile type, variable pulse width interval, fixed frequency and optional radiation source;

[0035] Category 8: Radiators with fixed frequency range and agile type, varying pulse width and repetition frequency.

[0036] Each type of known target knowledge base information target has multiple radiation source targets, so the i-th known target of type a in the target knowledge base is represented as target a,i , a=1,2,…,8, the radiation source parameter information in the known target library is expressed as rf represents the frequency information of the known target, pri represents the repetition information of the known target, and pw represents the pulse width information of the known target. {pdw} represents the initial pulse sequence in the current processing cycle, and {pdw a,i} represents the pulse sequence for the radiation source pulse screening for the i-th known target knowledge base information of category a, where the m-th pulse is pdw a,i,m , its parameter information is expressed as

[0037] The two types of known target frequency parameter information, fixed frequency range and non-agile type and frequency type agile, are expressed as is the minimum value of the frequency interval, is the maximum value of the frequency interval.

[0038] The known target pulse width information with fixed optional pulse width is expressed as pw a,i,j Indicates the radiation source target a,i The jth pulse width value, 1≤j≤N a,i , N a,i target a,i The number of pulse width values.

[0039] The known target pulse width information with optional pulse width interval is expressed as is the minimum value of the pulse width interval, is the maximum value of the pulse width interval.

[0040] The weight information of the known target with fixed and optional weight is expressed as pria, i, k, represents the radiation source target a,i The kth multi-cycle value, 1≤k≤M a,i , M a,i target a,i The number of repeated values.

[0041] The known target pulse width information that can be selected in the repetition interval is expressed as is the minimum value of the repeated period interval, is the maximum value of the repeated period interval.

[0042] Step 2: Classify and filter known target-associated pulses based on the target knowledge base information. This step, based on the constraint capability of the known target knowledge base information, first filters pulse sequences of strongly constrained radiation sources, then filters pulse sequences of weakly constrained radiation sources. By continuously diluting the pulse sequences, the initial pulse sequence {pdw} is filtered and sorted according to steps 2.1 to 2.8. The pulse sequences corresponding to the successfully sorted radiation sources are gradually extracted from the initial pulse sequence {pdw}, improving the reliability of known target pulse screening.

[0043] Step 2.1, perform correlation pulse screening on the first type of known targets, and perform correlation pulse screening on the radiation sources with fixed frequency range, non-agile type, fixed and optional pulse width, and fixed and optional repetition period (i.e., when a=1).

[0044] Step 2.1.1: Filter by frequency, pulse width, and multiplication. The filtering criteria are as follows:

[0045]

[0046] in, represents the frequency screening gate threshold, Represents the threshold value for the heavy cycle screening gate, Represents the pulse width screening gate threshold.

[0047] Step 2.1.2, the pulse sequence obtained after filtering by the above judgment conditions is {pdw_filter a,i}, perform frequency-based histogram statistical clustering on the pulse sequence, sort it from most to least according to the corresponding number of pulses, and extract the corresponding Q-class pulse sequence after frequency clustering {pdw_filter a,i,q}, where {pdw_filter a,i,q} represents the pulse sequence corresponding to the qth carrier frequency cluster, 0≤q≤Q, in descending order, the number of pulses greater than a certain threshold threshold key (Because it is a frequency non-agile signal) the pulse sequence {pdw_filter a,i,q}, merged into a non-agile frequency clustered pulse sequence {pdw_filter a,i} non-agile The signal sorting based on voting positioning and iterative peak selection is carried out, and the pulse sequence corresponding to the radiation source obtained by sorting is expressed as edw a,i (pdw), the signal sorting method based on voting positioning and iterative peak selection is specifically described in step 3.

[0048] Step 2.1.3, set the pulse sequence {pdw a,i}Remove edw a,i (pdw) Generate the subsequent pulse sequence for the first category, i+1th known target, and repeat step 2.1 until all the first category known targets are traversed.

[0049] Step 2.2, perform correlation pulse screening on the second type of known targets, and perform correlation pulse screening on radiation sources with a fixed frequency range, non-agile type, fixed and optional pulse width, and variable repetition interval (i.e., when a=2).

[0050] The pulse filtering judgment conditions are as follows:

[0051]

[0052] For the pulse sequence that meets the judgment condition {pdw_filter a,i}, execute step 2.1.2. Repeat step 2.2 until all known targets of category 2 are traversed.

[0053] Step 2.3, perform correlation pulse screening on the third type of known targets, with a fixed frequency range and non-agile type, variable pulse width interval, and fixed frequency optional radiation source screening (i.e., when a=3).

[0054] The pulse filtering judgment conditions are as follows:

[0055]

[0056] For the pulse sequence that meets the judgment condition {pdw_filter a,i}, execute step 2.1.2. Loop through step 2.3 until all known targets of the third category are traversed.

[0057] Step 2.4, perform correlation pulse screening on the fourth type of known targets, screening of radiation sources with fixed frequency range and non-agile type, varying pulse width interval, and varying repetition interval (i.e., when a=4).

[0058] The pulse filtering judgment conditions are as follows:

[0059]

[0060] For the pulse sequence that meets the judgment condition {pdw_filter a,i}, execute step 2.1.2. Loop through step 2.4 until all known targets of the fourth category are traversed.

[0061] Step 2.5, perform correlation pulse screening on the fifth type of known targets, with a fixed frequency range and agile type, fixed and optional pulse width, and fixed and optional radiation source screening (i.e., when a=5).

[0062] Step 2.5.1, pulse filtering decision conditions are as follows:

[0063]

[0064] Step 2.5.2, the pulse sequence obtained after filtering by the above judgment conditions is {pdw_filter a,i}, perform frequency-based histogram statistical clustering on the pulse sequence, sort it from most to least according to the corresponding number of pulses, and extract the corresponding Q-class pulse sequence after frequency clustering {pdw_filter a,i,q}, where {pdw_filter a,i,q} represents the pulse sequence corresponding to the qth carrier frequency cluster, 0≤q≤Q, in descending order, the number of pulses is less than a certain threshold threshold key (Because it is a frequency agile signal) the pulse sequence {pdw_filter a,i,q}, merged into a frequency-agile clustered pulse sequence {pdw_filtera,i} agile Introducing the signal sorting based on voting positioning and iterative peak selection, the pulse sequence corresponding to the selected radiation source is expressed as edw a,i (pdw), the signal sorting based on voting localization iterative peak selection is described in detail in step 3.

[0065] Step 2.5.3, set the pulse sequence {pdw a,i}Remove edw a,i (pdw) Generate the subsequent pulse sequence for the fifth category, i+1th known target, and repeat step 2.5 until all the fifth category known targets are traversed.

[0066] Step 2.6, perform correlation pulse screening on the sixth type of known targets, with a fixed frequency range and agile type, a fixed and optional pulse width, and screening of radiation sources with varying repetition intervals (i.e., when a=6).

[0067] The pulse filtering judgment conditions are as follows:

[0068]

[0069] For the pulse sequence that meets the judgment condition {pdw_filter a,i}, then execute step 2.5.2. Repeat step 2.6 until all known targets of the sixth category are traversed.

[0070] Step 2.7, perform correlation pulse screening on the 7th type of known targets, with a fixed frequency range and agile type, variable pulse width interval, and fixed repetition frequency for screening of optional radiation sources (i.e., when a=7).

[0071] The pulse filtering judgment conditions are as follows:

[0072]

[0073] For the pulse sequence that meets the judgment condition {pdw_filter a,i}, then execute step 2.5.2. Repeat step 2.7 until all known targets of the seventh category are traversed.

[0074] Step 2.8, perform correlation pulse screening on the 8th type of known targets, and screen the radiation sources with fixed frequency range and agile type, variable pulse width interval, and variable repetition interval (i.e., when a=8).

[0075] The pulse filtering judgment conditions are as follows:

[0076]

[0077] For the pulse sequence that meets the judgment condition {pdw_filter a,i}, then execute step 2.5.2. Repeat step 2.8 until all the known targets of the eighth category are traversed.

[0078] Step 3: Use the pulse sequence {pdw_filter pdw_id}, pdw_id represents the number of the pulse sequence, and the signal sorting of the iterative peak is carried out by voting and positioning.

[0079] Step 3.1: Construct the voting location geographic grid. With the projection point of the observation station at the current processing cycle as the center, draw a grid with equal latitude and longitude intervals (grid_lat, grid_lon, grid_X, grid_Y, grid_Z) covering the coverage of the observation station's receiving signal. grid_id , where grid_lat represents the center latitude of the grid, grid_lon represents the center longitude of the grid, grid_X, grid_Y, grid_Z represent the WGS84 coordinates of the grid center, and grid_id represents the grid ID.

[0080] Step 3.2: For the pulse sequence {pdw_filter pdw_id} vote_id Perform phase difference voting positioning and generate voting grid histogram vote_id , initialized to 0, where vote_id represents the number of times the voting location iteration peak is taken, the grid with the largest number of pulses in the voting location geographic grid is selected, and all corresponding pulse sequences in the grid are extracted {pdw_filter pdw_id} vote_id，peak , where peak represents the voting peak grid, and the criteria for pulse train voting are as follows:

[0081] Using the pulse sequence {pdw_filter pdw_id} vote_id The arrival time of each pulse in the interpolation is used to obtain the attitude information of the observation station (pitch sat ,roll sat ,yaw sat ) and the orbital WGS84 coordinate system coordinates (x sat ,y sat ,z sat ), then the incident vector of each geographic grid relative to the observation station antenna in the WGS84 coordinate system is It can be expressed as:

[0082]

[0083] The incident vector Convert from the WGS84 coordinate system to the observation station antenna coordinate system to obtain

[0084]

[0085] in, The rotation matrix from the WGS84 coordinate system to the observation station coordinate system can be expressed by (pitch sat ,roll sat ,yaw sat ) is calculated, Represents the rotation matrix from the observation station body to the antenna coordinate system.

[0086] Using the pulse sequence {pdw_filter pdw_id} vote_id The arrival time of each pulse in the interpolation calculation is the three-dimensional baseline vector of the longest baseline of the rotating interferometer in the antenna coordinate system: The theoretical phase difference Tphase of each geographic grid corresponding to a single pulse is grid_id,pdw_id It can be expressed as:

[0087]

[0088] Among them, λ pdw_id Represents the pulse wavelength, by calculating the theoretical phase difference Tphase of each pulse corresponding to each geographic grid grid_id,pdw_id , and then the measured phase difference Cphase of the pulse pdw_id The phase difference residual Δphase can be obtained by doing the difference grid_id ,pdw_id, that is:

[0089] Δphase grid_id,pdw_id =|Tphase grid_id,pdw_id -Cphase pdw_id |

[0090] Based on the phase difference residual number Δphase grid_id,pdw_id Voting is carried out, and the voting criteria are: if Δphase grid_id,pdw_id <π / 2, voting grid histogram vote_id =histogram vote_id +1; otherwise, histogram vote_id =histogram vote_id +0.

[0091] After the voting, the histogram vote_id Sort and determine the grid with the largest number of pulses in the voting location geographic grid, and extract all the corresponding pulse sequences in the peak grid

[0092] This step mainly completes the pre-sorting of pulse sequences that depend on position constraints, extracts the pulse sequence of the radiation source using position constraints, and reduces the pressure of dense pulse sequences on the main sorting.

[0093] Step 3.3: Extract the pulse sequence {pdw_filter pdw_id} vote_id，peak Perform main sorting and extract the pulse sequence corresponding to the radiation source that meets the sorting conditions {pdw_filter pdw_id} vote_id，peak,sort , where sort represents the identification after traditional signal main sorting, and obtains the grid position edw_loc (grid_X, grid_Y, grid_Z) of the radiation source to be sorted grid_id,edw_id , where edw_id represents the ID of the selected radiation source. The main sorting method can use traditional sorting methods, such as the classic CDIF method, which can sort out the true correlated pulse sequence of the radiation source from the pulse sequence within the peak position of the peak grid.

[0094] Step 3.4: Set the pulse train {pdw_filter pdw_id} vote_id , remove the pulse sequence {pdw_filter pdw_id} vote_id，peak,sort , get the pulse sequence {pdw_filter pdw_id} vote_id＝vote_id+1 , then repeat step 3.2, and a new voting grid histogram will be formed vote_id＝vote_id+1 , and a new highest peak grid can be obtained, and then step 3.3 is repeated, as Figure 2 The advantage of this step is that it iteratively removes the successfully sorted radiation source pulse sequence and updates the voting grid histogram, which can reduce the influence of the voting positioning interference pulse caused by the antenna phase difference measurement error and phase difference ambiguity until the extracted pulse sequence {pdw_filter pdw_id} vote_id，peak After step 3.3, it is no longer possible to use the main sorting method to separate out effective radiation sources.

[0095] The advantages of the method of iterative peak extraction by voting and positioning in step 3 are as follows: the pulse sequence is pre-sorted under the geographical location constraint by using geographical grid constraints and pulse voting, which solves the disadvantage that a single pulse cannot be pre-sorted for direction finding and positioning under a rotating system, cleans the pulse sequence entering the main sorting, and reduces the difficulty of the main sorting; the iterative peak extraction by voting and positioning adopts the idea of ​​iteratively extracting the peak pulse sequence, and updates the voting grid histogram after each sorting, which reduces the influence of interference pulses, further reduces the difficulty of the main sorting, and can solve the problem of geographical distribution of known target radiation sources of the same model in multiple locations.

[0096] Step 4: Carry out signal sorting for the remaining unknown target-related pulses. Based on steps 2 and 3, complete the signal sorting based on the known target knowledge base information target, and remove the associated pulse sequence of the radiation source from the original pulse sequence {pdw} of the processing cycle, and use the remaining unknown target pulse sequence {pdw} non-key , carry out signal sorting of the remaining unknown target associated pulses.

[0097] Step 4.1: Perform iterative peak signal sorting on the frequency non-agile unknown target signal.

[0098] Using the remaining unknown target pulse sequence {pdw} non-key Conduct frequency histogram-based clustering, {pdw} non-key,s represents the pulse sequence corresponding to the sth carrier frequency cluster, and {pdw} non-key,s Sort by the number of corresponding pulses from most to least, and sort by the number of pulses greater than a certain threshold in order from most to least non-key Pulse sequence {pdw} non-key,s Bring them into step 3 respectively, carry out signal sorting based on voting positioning iterative peak, and sort out the frequency non-agile unknown target radiation source edw non-key,non-fixed,s , and each type of {pdw} non-key,s The correlated pulses of the selected radiation sources are taken from {pdw} non-key Eliminate them one by one until the number of all pulses is greater than a certain threshold non-key Pulse sequence {pdw} non-key,s , forming the remaining pulse sequence {pdw} non-key,non-fixed The difference between this step and step 2.1.2 and step 2.5.2 is that this step substitutes the pulse sequences after frequency clustering into step 3 separately, instead of merging them and then substituting them into step 3. The reason is that without the constraint of known target library knowledge information, the remaining unknown target pulse sequence {pdw} non-key The pulse density will be high, and it is necessary to perform the signal sorting method of voting, locating, iteratively extracting peaks in step 3 in batches according to the frequency clustering characteristics.

[0099] Step 4.2: Use the radiation source positioning constraints to carry out radiation source batching and parameter re-statistical update within the cycle.

[0100] Based on the frequency non-agile and unknown target signal sorting in step 4.1, the fixed and unknown target signals can be sorted out, but the radiation sources with frequency hopping or agile types will also be sorted into multiple fixed frequency radiation sources, resulting in an increase in the number of radiation sources. Therefore, according to the radiation source grid position edw_loc (grid_X, grid_Y, grid_Z) obtained in step 3 grid_id,edw_id , based on the radiation source edw selected in step 4.1 non-key,non-fixed,s , the k-means algorithm is used for position clustering, and the distance threshold can be set to 2 times the grid distance. The pulse sequences corresponding to the radiation sources clustered into the same category are merged and re-sorted by the classic CDIF method to complete the radiation source batching and parameter re-statistical update within the cycle.

[0101] Step 4.3: Perform iterative peak signal sorting on the frequency-agile unknown target signal.

[0102] Using the remaining unknown target pulse sequence {pdw} non-key,non-fixed All of them are brought into step 3 to carry out signal sorting based on voting positioning and iterative peak selection, thus completing the signal sorting of all pulse sequences.

[0103] In specific implementation, the above process can be automatically run through computer software, and the hardware system of the running method should also be within the scope of protection.

Claims

1. A method for sorting antenna signals of a rotating interferometer based on iterative peak extraction, characterized in that: Here are the steps: Step 1: Classify the known target knowledge base information target according to the characteristics of the radiation source: The characteristics of the radiation source include the frequency characteristics of the radiation source, the pulse width characteristics of the radiation source and the repetition characteristics of the radiation source; The frequency characteristics of the radiation source include two types: fixed frequency range and non-agile type, and agile frequency type; The pulse width characteristics of the radiation source include two types: fixed pulse width and pulse width interval. The radiation source's repetition characteristics include two types: fixed and optional repetition, and interval-varying repetition. Step 2: According to the known target knowledge base information target, classify and filter the known target-related pulses; according to the constraint ability of the known target knowledge base information target, the idea of ​​first filtering the pulse sequence of the strong constraint type radiation source signal and then filtering the pulse sequence of the weak constraint type radiation source signal is adopted. By continuously diluting the pulse sequence, the initial pulse sequence {pdw} is screened and sorted in sequence, and the pulse sequence corresponding to the successfully sorted radiation source is gradually extracted from the initial pulse sequence {pdw}, thereby improving the reliability of the known target pulse screening and obtaining the initial screening pulse sequence {pdw_filter pdw_id }, where pdw_id represents the number of the pulse sequence; Step 3: Use the initial filtering pulse sequence {pdw_filter pdw_id } Carry out voting positioning iterative peak signal sorting to obtain the pulse sequence {pdw_filter pdw_id } vote_id，peak ; Step 4: Perform signal sorting on the remaining unknown target-related pulses to complete the signal sorting of all pulse sequences: Based on the known target knowledge base information target, the signal is sorted and the associated pulse sequence of the radiation source is removed from the original pulse sequence {pdw} of the processing cycle, and the remaining unknown target pulse sequence {pdw} is used. non-key , carry out signal sorting of the remaining unknown target associated pulses.

2. The method for selecting antenna signals of a rotating interferometer based on iterative peak selection according to claim 1, characterized in that: In step 1, the known target knowledge base information target is divided into 8 categories according to the frequency, pulse width, and repetition constraint of the radiation source signal, namely: Category 1: Radiating sources with fixed frequency range and non-agile type, fixed and optional pulse width, and fixed and optional repetition frequency; Category 2: Radiators with fixed frequency range and non-agile type, fixed and optional pulse width, and variable frequency interval; Category 3: The frequency range is fixed and the type is non-agile, the pulse width interval varies, and the radiation source is fixed and optional; Category 4: Radiators with fixed frequency range and non-agile type, varying pulse width and repetition frequency ranges; Category 5: Fixed frequency range and agile type, fixed and optional pulse width, fixed and optional radiation source; Category 6: A radiation source with a fixed frequency range and agile type, fixed and optional pulse width, and variable frequency range; Category 7: Fixed frequency range and agile type, variable pulse width interval, fixed frequency and optional radiation source; Category 8: Radiators with fixed frequency range and agile type, variable pulse width range and variable repetition range; There are multiple radiation source targets in each type of known target knowledge base information target, so the i-th known target of type a in the known target knowledge base information target is represented as target a,i , a=1,2,…,8, the radiation source parameter information in the known target knowledge base information target is expressed as rf represents the frequency information of the known target, pri represents the repetition information of the known target, and pw represents the pulse width information of the known target; {pdw} represents the initial pulse sequence in the current processing cycle, and {pdw a,i } represents the pulse sequence when screening the radiation source pulse for the a-th category and the i-th known target knowledge base information, where the m-th pulse is pdw a,i,m , its parameter information is expressed as The two types of known target frequency parameter information, fixed frequency range and non-agile type and frequency type agile, are expressed as is the minimum value of the frequency interval, is the maximum value of the frequency interval; The known target pulse width information with fixed optional pulse width is expressed as pw a,i,j Indicates the radiation source target a,i The jth pulse width value, 1≤j≤N a,i , N a,i target a,i The number of pulse width values; The known target pulse width information with optional pulse width interval is expressed as is the minimum value of the pulse width interval, is the maximum value of the pulse width interval; The weight information of the known target with fixed and optional weight is expressed as pri a,i,k Indicates the radiation source target a,i The kth multi-cycle value, 1≤k≤M a,i , M a,i target a,i The number of repeated values ​​of ; The known target pulse width information that can be selected in the repetition interval is expressed as is the minimum value of the repeated period interval, is the maximum value of the repeated period interval.

3. The method for selecting antenna signals of a rotating interferometer based on iterative peak selection according to claim 2, characterized in that: Step 2 is as follows: In step 2.1, when a=1, perform correlation pulse screening on the first type of known targets, that is, perform correlation pulse screening on radiation sources with a fixed frequency range, non-agile type, fixed and optional pulse width, and fixed and optional repetition period, as follows: Step 2.1.1: Filter by frequency, pulse width, and multiplication. The filtering criteria are as follows: in, represents the frequency screening gate threshold, Represents the threshold value for the heavy cycle screening gate, Represents the pulse width screening gate threshold; Step 2.1.2, the pulse sequence obtained after filtering by the above judgment conditions is {pdw_filter a,i }, for this pulse sequence {pdw_filter a,i Perform frequency-based histogram statistical clustering, sort the corresponding pulses from most to least, and extract the corresponding Q-class pulse sequence after frequency clustering {pdw_filter a,i,q }, where {pdw_filter a,i,q } represents the pulse sequence corresponding to the qth carrier frequency cluster, 0≤q≤Q, in descending order, the number of pulses greater than the threshold threshold key The pulse sequence {pdw_filter a,i,q }, merged into a non-agile frequency clustered pulse sequence {pdw_filter a,i } non-agile The signal sorting based on voting positioning and iterative peak selection is carried out, and the pulse sequence corresponding to the radiation source obtained by sorting is expressed as edw a,i (pdw); Step 2.1.3, set the pulse sequence {pdw a,i }Remove edw a,i (pdw) Generate the subsequent pulse sequence for the first category, i+1 known target, and repeat step 2.1 until all the first category known targets are traversed; In step 2.2, when a=2, perform correlation pulse screening on the second type of known targets, and perform correlation pulse screening on the radiation sources with a fixed frequency range, non-agile type, fixed and optional pulse width, and variable frequency interval, as follows: The pulse filtering judgment conditions are as follows: For the pulse sequence that meets the judgment condition {pdw_filter a,i }, execute step 2.1.2, and obtain the pulse sequence corresponding to the radiation source, which is expressed as edw a,i (pdw), repeat step 2.2 until all known targets of the second category are traversed; In step 2.3, when a=3, the third type of known target is screened for associated pulses. The frequency range is fixed and the type is non-agile. The pulse width interval varies. The screening of optional radiation sources with fixed repetition period is as follows: The pulse filtering judgment conditions are as follows: For the pulse sequence that meets the judgment condition {pdw_filter a,i }, execute step 2.1.2, and obtain the pulse sequence corresponding to the radiation source, which is expressed as edw a,i (pdw), repeat step 2.3 until all the known targets of the third category are traversed; Step 2.4, a=4, perform correlation pulse screening on the fourth type of known targets, with a fixed frequency range and non-agile type, pulse width interval variation, and repeated period interval variation radiation source screening, as follows: The pulse filtering judgment conditions are as follows: For the pulse sequence that meets the judgment condition {pdw_filter a,i }, execute step 2.1.2, and obtain the pulse sequence corresponding to the radiation source, which is expressed as edw a,i (pdw), and then repeat step 2.4 until all the known targets of the fourth category are traversed.

4. The method for selecting antenna signals of a rotating interferometer based on iterative peak selection according to claim 3, characterized in that: Step 2 is as follows: In step 2.5, when a=5, perform correlation pulse screening on the fifth type of known targets. The frequency range is fixed and the type is agile, the pulse width is fixed and optional, and the repetition period is fixed and optional. The screening of radiation sources is as follows: Step 2.5.1, pulse filtering decision conditions are as follows: Step 2.5.2, the pulse sequence obtained after filtering by the above judgment conditions is {pdw_filter a,i }, perform frequency-based histogram statistical clustering on the pulse sequence, sort it from most to least according to the corresponding number of pulses, and extract the corresponding Q-class pulse sequence after frequency clustering {pdw_filter a,i,q }, where {pdw_filter a,i,q } represents the pulse sequence corresponding to the qth carrier frequency cluster, 0≤q≤Q, in descending order, the number of pulses is less than a certain threshold threshold key The pulse sequence {pdw_filter a,i,q }, merged into a frequency-agile clustered pulse sequence {pdw_filter a,i } agile Introducing the signal sorting based on voting positioning and iterative peak selection, the pulse sequence corresponding to the selected radiation source is expressed as edw a,i (pdw); Step 2.5.3, set the pulse sequence {pdw a,i }Remove edw a,i (pdw) Generate the subsequent pulse sequence for the fifth category, i+1th known target, and repeat step 2.5 until all the fifth category known targets are traversed; In step 2.6, when a=6, perform correlation pulse screening on the sixth type of known targets. The frequency range is fixed and the type is agile. The pulse width is fixed and optional. The screening of radiation sources with varying frequency intervals is as follows: The pulse filtering judgment conditions are as follows: For the pulse sequence that meets the judgment condition {pdw_filter a,i } Then execute step 2.5.2, and then loop through step 2.6 until all known targets of the sixth category are traversed; Step 2.7, when a=7, perform correlation pulse screening on the seventh type of known target, with a fixed frequency range and agile type, variable pulse width interval, and fixed frequency. The screening of optional radiation sources is as follows: The pulse filtering judgment conditions are as follows: For the pulse sequence that meets the judgment condition {pdw_filter a,i } Then execute step 2.5.2, and then loop through step 2.7 until all known targets of the seventh category are traversed; Step 2.8, when a=8, perform correlation pulse screening on the eighth type of known targets, with a fixed frequency range and agile type, pulse width interval variation, and repeated period interval variation. The screening of radiation sources is as follows: The pulse filtering judgment conditions are as follows: For the pulse sequence that meets the judgment condition {pdw_filter a,i }, then execute step 2.5.2, and loop through step 2.8 until all the eighth category of known targets are traversed.

5. The method for selecting antenna signals of a rotating interferometer based on iterative peak selection according to claim 4, characterized in that: In step 3, the initial filtering pulse sequence {pdw_filter pdw_id } Carry out voting positioning iterative peak signal sorting to obtain the pulse sequence {pdw_filter pdw_id } vote_id，peak , as follows: Step 3.1: Construct a voting location geographic grid; with the projection point of the observation station at the current processing cycle as the center, draw a grid of equal latitude and longitude intervals (grid_lat, grid_lon, grid_X, grid_Y, grid_Z) covering the coverage area of ​​the observation station's receiving signal. grid_id , where grid_lat represents the center latitude of the grid, grid_lon represents the center longitude of the grid, grid_X, grid_Y, grid_Z represent the WGS84 coordinates of the grid center, and grid_id represents the grid ID; Step 3.2: For the pulse sequence {pdw_filter pdw_id } vote_id Perform phase difference voting positioning and generate voting grid histogram vote_id , initialized to 0, where vote_id represents the number of times the voting location iteration peak is taken, the grid with the largest number of pulses in the voting location geographic grid is selected, and all corresponding pulse sequences in the grid are extracted {pdw_filter pdw_id } vote_id，peak , where peak represents the voting peak grid; after the voting is completed, the histogram vote_id Sort and determine the grid with the largest number of pulses in the voting location geographic grid, and extract all the corresponding pulse sequences in the peak grid {pdw_filter pdw_id } vote_id，peak ; Step 3.3: Extract the pulse sequence {pdw_filter pdw_id } vote_id，peak Perform main sorting and extract the pulse sequence corresponding to the radiation source that meets the sorting conditions {pdw_filter pdw_id } vote_id，peak,sort , where sort represents the identification after traditional signal main sorting, and obtains the grid position edw_loc (grid_X, grid_Y, grid_Z) of the radiation source to be sorted grid_id,edw_id , where edw_id represents the number of the selected radiation source; Step 3.4: Set the pulse train {pdw_filter pdw_id } vote_id , remove the pulse sequence {pdw_filter pdw_id } vote_id，peak,sort , get the pulse sequence {pdw_filter pdw_id } vote_id＝vote_id+1 , then repeat step 3.2, and a new voting grid histogram will be formed vote_id＝vote_id+1 , and a new highest peak grid can be obtained, and then step 3.3 is repeated until the pulse sequence {pdw_filter pdw_id } vote_id，peak .

6. The method for selecting antenna signals of a rotating interferometer based on iterative peak selection according to claim 5, characterized in that: In step 3.3, the main sorting method uses the classic CDIF method to sort out the true correlated pulse sequences of the radiation source from the pulse sequences within the peak position of the peak grid.

7. The method for selecting antenna signals of a rotating interferometer based on iterative peak selection according to claim 4, characterized in that: In step 4, signal sorting is performed on the remaining unknown target-related pulses to complete the signal sorting of all pulse sequences, as follows: Step 4.1: Iterative peak signal sorting for frequency non-agile unknown target signals: Using the remaining unknown target pulse sequence {pdw} non-key Conduct frequency histogram-based clustering, {pdw} non-key,s represents the pulse sequence corresponding to the sth carrier frequency cluster, and {pdw} non-key,s Sort by the number of corresponding pulses from most to least, and sort by the number of pulses greater than the threshold in order from most to least non-key Pulse sequence {pdw} non-key,s Bring them into step 3 respectively, carry out signal sorting based on voting positioning iterative peak, and sort out the frequency non-agile unknown target radiation source edw non-key,non-fixed,s , and each type of {pdw} non-key,s The correlated pulses of the selected radiation sources are taken from {pdw} non-key Eliminate them one by one until the number of all pulses is greater than the threshold non-key Pulse sequence {pdw} non-key,s , forming the remaining pulse sequence {pdw} non-key,non-fixed ; Step 4.2: Use the radiation source positioning constraints to carry out radiation source batching and parameter re-statistical update within the cycle: According to the radiation source grid position edw_loc (grid_X, grid_Y, grid_Z) obtained in step 3 grid_id,edw_id , based on the radiation source edw selected in step 4.1 non-key,non-fixed,s , use the k-means algorithm to perform position clustering, the distance threshold can be set to 2 times the grid distance, merge the pulse sequences corresponding to the radiation sources clustered into the same category and re-perform the classic CDIF method main sorting, complete the radiation source batching and parameter re-statistical update within the cycle; Step 4.3: Perform iterative peak signal sorting on the frequency-agile unknown target signal: Using the remaining unknown target pulse sequence {pdw} non-key,non-fixed All of them are brought into step 3 to carry out signal sorting based on voting positioning and iterative peak selection, thus completing the signal sorting of all pulse sequences.

Images