KR101632167B1 - Method for clustering an radar signal - Google Patents

Abstract

The present invention relates to a method of clustering radar signals using difference in arrival direction and arrival time of collected radar signals, and more particularly, to a method of clustering radar signals using a difference in arrival direction of a radar signal, Receiving a PDW list sorted according to a TOA; inputting PDW data constituting the input PDW list to a grouping signal having similarity by calling a distance calculation function, The difference between the value of the signal feature parameters including the pulse arrival time of the PDW data to be searched and the value of the signal feature parameters representative of the grouped group is obtained and compared with a predetermined threshold value for each signal feature variable The PDW data to be grouped may be added to a group existing at the closest distance, And clustering the group signals into group signals having similarity.

Description

{METHOD FOR CLUSTERING AN RADAR SIGNAL}

The present invention relates to a method for clustering radar signals, and more particularly, to a method for clustering collected radar signals using difference in arrival direction and arrival time of collected radar signals.

In general, electronic warfare support (ES) systems help identify and locate enemy power structures and deployments by identifying and locating threat emitters after receiving enemy signals. The main functions of the ES system are detection of threat signal source, type of threat and operation, location of threat signal source, and indication of threat information to support situation recognition.

The ES system measures the pulse characteristics of the received signal, identifies the rules, correlation, and continuity of pulse trains from the collected data, analyzes the characteristics of the data, and obtains Emitter Identification Data (EID ) To identify the source of the signal.

Since the recent signal environment has a very high pulse density and various types of signal sources, an ES system that confirms each radar signal in real time requires faster and more accurate analysis capability. For this purpose, ES system has developed a clustering method of radar pulses as a preprocessing technique in order to reduce the burden of signal analysis and to support reliable analysis.

Clustering in the ES system is a special application of data clustering that classifies unknown radar sources from received radar pulses. Compared to normal data clustering, classification of radar sources has a special situation.

First, radar pulses have many dimensions. Second, the number of received pulses is very variable depending on the signal environment. In a good environment, the number of received pulses may reach several million per second. However, under the harsh environment, the number of received pulses is very small. The number is about a dozen for the radar signal.

Third, the radar signal has various modulation types, and the characteristics of the pulse are determined according to the type thereof. Therefore, these factors must be considered for the clustering method in the ES system.

Radar pulse clustering, on the other hand, is performed as a preprocessing for signal analysis between the radar signal acquisition and analysis process and should provide reliable cluster information for the signal analysis process. For this purpose, the clustering method should not 1) distribute the pulses of one radar source to the other clusters, 2) do not form clusters that are too large, and 3) minimize processing time.

However, the known radar pulse clustering method (commonly known as a three-dimensional sequential clustering method or a three-dimensional simultaneous clustering method) uses a frequency or arrival direction (AOA, If the azimuth is used, there is a disadvantage in that one signal is divided into different groups when the azimuth is out of the predetermined azimuth, and the azimuth, frequency , And if the pulse widths are the same, it is possible to form a cluster which is too large, and the time for analyzing the signal may be prolonged.

Korean Patent Publication No. 10-1007662 Korean Patent Publication No. 10-1026144

It is an object of the present invention to solve the above-mentioned problems, and it is an object of the present invention to solve the above-mentioned problems, and it is an object of the present invention to provide a radar- Dimensional clustering method of a collecting radar signal that can track signals of a radiator and prevent over-grouping.

According to another aspect of the present invention, there is provided a method for clustering a radar signal,

A first step of receiving a radar signal and a PDW list in which a plurality of PDW (Pulse Description Word) data generated by a signal collecting unit are sorted according to a pulse arrival time (TOA);

The PDW data of the input PDW list is called a distance calculation function and is clustered into group signals having similarity, and the value of the signal characteristic variables including the pulse arrival time of PDW data to be grouped by using the distance calculation function And a signal feature parameter value representative of each of the grouped groups, and compares the signal feature parameters with a predetermined threshold value for each signal feature parameter to find a group that exists in the closest distance and to group the PDW data to be grouped And clustering the group signals into groups having similarity, the method comprising the steps of:

In addition, in the second step, the first PDW data to be grouped is generated as a first group when the generated group is not present, and the value of signal characteristic variables representing the first group to be generated is the same as the value of the first PDW data And the pulse arrival time is set to the value of the signal characteristic variables.

In addition, in the second step of the clustering method of the collecting radar signal according to the embodiment of the present invention described above,

Wherein the PDW data to be grouped is determined as similarity between the PDW data to be grouped and each of the grouped groups, and if there is no similarity, the PDW data to be grouped is created as a new group,

The difference between the value of the signal feature parameters including the pulse arrival time of the PDW data to be grouped and the value of the signal feature parameters representative of the grouped one group is obtained and compared with the predetermined threshold value for each signal feature parameter The value of the signal characteristic variables representing the newly generated group is set to the value of the signal characteristic variables including the pulse arrival time of the PDW data to be grouped, .

In the above-described embodiments, each of the signal characteristic parameters of the PDW data and the signal characteristic parameters representative of the group includes at least the arrival direction of the pulse, the frequency, the pulse width, and the pulse arrival time information,

The signal characteristic parameters representative of the group include the arrival direction, frequency, pulse width, and pulse arrival time information of the pulses, and the values of the respective characteristic variables are average values of the respective signal characteristic parameters of the PDW data belonging to the same group .

According to the embodiment of the present invention as described above, the method according to the embodiment of the present invention is capable of tracking a signal of a moving radiator by using a difference in arrival direction of a radar signal, Therefore, it is possible to minimize an excessive grouping of signals, and it is possible to improve the accuracy of signal analysis.

According to a further embodiment of the present invention, when the clustering groups are excessive, there is an effect that the number of groups can be reduced by merging groups located at a high similarity again, i.e., at a close distance.

1 is a diagram for explaining a signal processing process of an electronic warfare support system.
FIG. 2 is a diagram for comparing grouping in the case where the difference between the arrival direction (a) of the signal and the arrival direction (b) is used.
FIG. 3 is a diagram for comparing a case where a pulse arrival time difference is applied and a case where a pulse arrival time difference is not applied.
4 is a typical PDW list example in which a plurality of pieces of PDW data are listed.
5 illustrates a program routine for implementing a method of clustering radar signals in accordance with an embodiment of the present invention.
FIG. 6 to FIG. 8 illustrate simulation results for further explaining the effect of the clustering method according to the embodiment of the present invention. FIG.

Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings. In the following description of the present invention, a detailed description of known functions and configurations incorporated herein will be omitted when it may make the subject matter of the present invention rather unclear.

1 is a diagram for further illustrating a signal processing process of the electronic warfare support system.

First, in the electronic warfare environment, the signal collector converts the received signal into a PDW (Pulse Description Word) format. (PA), a pulse width (PW), a pulse arrival time (TOA: Time), and a pulse arrival time (TOA) indicating characteristics of a signal with respect to a received radar signal by down- of Arrival, Frequency, Angle of Arrival (AOA), and the like. For convenience of description, the parameters (PA, PW, TOA, AOA, etc.) representing the characteristics of the received signal with respect to the radar signal are defined as signal characteristic variables, and these signal characteristic parameters are gathered to construct PDW data . The signal collecting unit receives the radar signal and generates a plurality of PDW (Pulse Description Word) data by arranging the PDW list according to the pulse arrival time (TOA). An example of such a PDW list is shown in FIG.

For reference, the mixed signal shown in FIG. 1 (a) is a set of PDW data obtained by collecting signals radiated by several radars in chronological order. The clustering method divides the mixed signal into signals that are determined to be radiated by one radar, and generates a grouped signal as shown in FIG. 1 (b). After that, in the signal determination step, the characteristic of the signal is analyzed (Fig. 1 (c)). The present invention relates to a method for clustering mixed radar signals collected in such signal processing.

The clustering method according to an embodiment of the present invention will be described in more detail with reference to a method of clustering a collection radar signal according to an embodiment of the present invention. have.

The 4-dimensional simultaneous clustering method is a method utilizing additionally the arrival direction, frequency, pulse width and arrival time of the pulse. The present invention is different from the three-dimensional clustering method in which it is known to use the arrival direction difference of the pulse and the arrival time difference of the pulse.

1. Difference in arrival direction

Since the moving radar signal is characterized in that the receiving direction of the signal is not abruptly changed, the possibility of forming a more accurate group is increased by using the difference in arrival direction of the signal (or pulse). The difference (DAOA) between the arrival directions of the signals expressed by the following equation (1) is the difference in arrival direction of continuously received PDW data.

은 현재 PDW 데이터의 도착방향이고,

은 직전의 PDW 데이터의 도착방향을 나타낸다.

Is the arrival direction of the current PDW data,

Indicates the arrival direction of the immediately preceding PDW data.

Fig. 2 is a view for explaining a comparison of grouping in the case where the difference (b) between the arrival direction (a) and the arrival direction of the signal is used, assuming that the signal moves in a certain direction.

Referring to FIG. 2, the horizontal axis represents a pulse arrival time (TOA) and the vertical axis represents a pulse arrival direction (AOA). In the three-dimensional clustering method using the arrival direction as it is, the difference in the arrival direction of the pulses is compared with the threshold value (AOA Thresholds), and the two groups are separated as shown in FIG.

However, in the 4 - dimensional simultaneous clustering method, the groups are separated based on the difference in arrival direction (DAOA). Therefore, as shown in FIG. 2 (b), when the difference in the arrival direction of the pulses does not deviate within a certain range (ΔdATH), the groups are not separated and form one group.

Four-dimensional simultaneous clustering uses the difference in arrival direction of signals for grouping, and signals generated from moving emitters can be classified into one group. Four-dimensional simultaneous clustering improves the accuracy of the signal analysis by using the difference of the arrival direction in the grouping, as compared with the grouping using only the arrival direction, frequency and pulse width of the existing three-dimensional method, have.

2. Difference of arrival time

The interval between pulses emitted by one radar is called the pulse repetition interval (PRI). The PRI of a signal emitted from one radar has characteristics that are included in a certain time interval, and the PRI differs depending on the radar type. The PRI varies from several microseconds to several tens of milliseconds. Depending on the distribution of the PRI, the radar can be classified into fixed, periodic repetition, jittering, and so on. Therefore, even if the distribution of arrival direction, frequency, and pulse width are the same, if the arrival time of pulses is different, it is more likely that the signals are emitted by other radars. Therefore, the accuracy of clustering can be improved by adding difference of arrival time in clustering collected radar signals.

For reference, FIG. 3 shows a diagram for comparing the case where the arrival time difference of the pulse is applied and the case where the difference is not applied.

은 현재 PDW의 도착시간이며,

은 직전 PDW의 도착시간이다.The difference (DTOA: Difference Time Of Arrival) expressed by Equation (2) is a difference in arrival time of consecutively received PDWs.

Is the arrival time of the current PDW,

Is the arrival time of the immediately preceding PDW.

)를 기준으로 일정한 범위(

) 안에 속한다.Expressing this as a formula, PRI for a signal radiated by one radar is a set of DTOA (n), and DTOA (n) is an average value

) To a certain range (

).

The four-dimensional clustering method uses the arrival time characteristic of the signal to separate the group when k times the DTOAmean of the signal (k is a threshold integer). By setting the threshold value in this way, even if k pulses are missed due to the processing delay of the receiver when receiving a signal, it can be classified into the same group, and it is possible to prevent the group from being formed small.

In the three-dimensional clustering method, when the arrival time intervals of the received signals are different from each other, they are classified into one group as shown in FIG. 3 (a) .

A method of clustering four-dimensional clusters of a radar signal using difference in arrival direction and arrival time will be described in more detail with reference to FIG.

5 illustrates a program routine for implementing the method of clustering radar signals according to an embodiment of the present invention.

5, a method for clustering a radar signal according to an embodiment of the present invention includes:

A first step of receiving a radar signal and a PDW list in which a plurality of PDW (Pulse Description Word) data generated by a signal collecting unit are sorted according to a pulse arrival time (TOA);

The PDW data including the pulse arrival time of the PDW data to be grouped by using the distance calculation function (calc_distance ()) by clustering each PDW data constituting the input PDW list into group signals having similarity by calling a distance calculation function A group that exists at the closest distance is found by finding the difference between the value of the signal characteristic variables and the value of the signal characteristic variables representing each group group and comparing the value with the predetermined threshold value for each signal characteristic variable, And a second step of adding PDW data or generating a new group and clustering the PDW data into group signals having similarity.

In this clustering method, each of the signal characteristic parameters of the PDW data and the signal characteristic parameters representative of the group includes at least the arrival direction, frequency, pulse width, and pulse arrival time information of the pulse, And the values of the variables have an average value of the respective signal characteristic parameters of the PDW data belonging to the same group.

In the second step, the first PDW data to be grouped is generated as the first group when the generated group is not present, and the value of the signal characteristic variables representing the first group to be generated is the pulse of the first PDW data And the arrival time is set to the value of the signal characteristic variables.

In the second step,

Wherein the PDW data to be grouped is determined as similarity between the PDW data to be grouped and each of the grouped groups, and if there is no similarity, the PDW data to be grouped is created as a new group,

The difference between the value of the signal feature parameters including the pulse arrival time of the PDW data to be grouped and the value of the signal feature parameters representative of the grouped one group is obtained and compared with the predetermined threshold value for each signal feature parameter And judges that there is no similarity when the threshold value is exceeded.

[0033] The configuration of the present invention will be described in further detail with reference to FIG.

First, it is assumed that a PDW list is generated by receiving a radar signal and arranging a plurality of PDW (Pulse Description Word) data according to a pulse arrival time (TOA). The PDW list is classified into a clustering When the PDW data is applied to the processing unit, the clustering processing unit terminates when all the PDW data in the PDW list are grouped, and repeats the clustering routine when PDW data that is not grouped remains in the PDW list.

If there is PDW data to be grouped, first, as described in lines 5 to 7 of FIG. 5, a variable for storing the closest group group (minGroup) is set to -1, and a variable for storing the minimum distance is set to max Initializes a group list (GROUP list), which is a variable that stores a group to be created.

Since the generated PDW data in the first PDW list does not exist yet, the minGroup is -1, and a new group is created (row 19 in FIG. 5) And the values of the signal characteristic variables (DAOA, DFreq, DPW, DTOA) representative of the group are initialized to the values of the signal characteristic parameters of the first PDW data )do. At this time, DTOA is initialized to -1.

From the second PDW data in the PDW list, the distance to the group in the group list is calculated and added to the closest group (that is, having similarity) or a new group is created.

Calculation of the distance between each grouped group and the PDW data to be grouped is performed by calling the distance calculation function (cal_distance ()) described in line 10 of FIG. The distance calculation function cal_distance () shown in line 30 of FIG. 5 includes four signal characteristic variables of the PDW data to be grouped, namely, the arrival direction (PDW.AOA), the frequency (PDW.Freq) (DAOA, DFreq, DPW, DTOA) between the arrival time (TOA) and the four signal characteristic variables (group.AvgAoA, group.Avgfreq, group.AvgPW, group.AvgTOA) (Distance = DAOA + DFreq + DPW + DTOA) through a method of obtaining the absolute value of the signal characteristic variables and comparing the signal characteristic parameters with predetermined threshold values (AOA_TH, Freq_TH, PW_TH, DTOA_TH) PDW data to be grouped is added or a new group is created and grouped into group signals having similarity.

When the difference between the four signal characteristic variables is compared with each threshold value, -1 is input to the variable dist when the difference is out of the threshold value range. When the difference is included in the threshold value range, the difference is added to the distance (distance = DAOA + DFreq + DPW + DTOA) (line 41 in FIG. 5). In this case, since the units of the four signal characteristic variables are different, it is preferable to normalize the difference of each signal characteristic variable to calculate the distance when calculating the distance.

Meanwhile, in the process of performing line 16 through line 8 in FIG. 5, which is a distance calculation interval between PDW data to be grouped with each group, minDist, which is a variable for storing the shortest distance among the groups, If the value is found, store the distance (dist) in minDist and simultaneously store the current group in the minGroup variable. If the nearest group is not found after performing all the 16 rows in line 8 of FIG. 5, the minGroup variable is maintained at -1 initially set.

If the closest group (minGroup) is -1 as shown in line 17 of FIG. 5, since there is no group to include the PDW data to be grouped, the PDW data to be grouped is created as a new group (row 19 in FIG. 5) And initializes the values of the four signal characteristic variables (AOA, Freq, PW, TOA) representative of the generated group to the values of the four signal characteristic parameters of the corresponding PDW data. If the closest group exists, the PDW data is added to the nearest group, and the values of the signal characteristic variables representing the group to which the PDW data is newly added include the values of the signal characteristic parameters of the newly added PDW data, The maximum, minimum, and average values are updated for each of Freq, PW, and TOA. If grouping is performed on the last PDW data in the PDW list in this order, the PDW data having similar characteristics will generate several groups.

As described above, according to the present invention, by using the difference of the arrival direction instead of the arrival direction of the signal, it becomes possible to track a moving emitter that could not be provided by the existing three-dimensional clustering method, This method can minimize over-grouping compared to existing 3D clustering method.

If the number of groups generated in the clustering process exceeds a predetermined threshold value, a step of finding and merging proximity groups existing within a predetermined range may be added to reduce the number of groups.

The 4-dimensional simultaneous clustering method of the collecting radar signal according to the embodiment of the present invention can further reduce the grouping compared to the existing three-dimensional clustering method by adding the arrival time difference between consecutive PDWs, It is possible to group the group with a minimum fragmentation due to the tracking effect of the signal radiation source using the difference in the arrival direction between consecutive PDWs rather than the comparison between the simple arrival directions.

In order to confirm the degree of segmentation of the group according to the clustering method of the present invention, three methods of three-dimensional sequential clustering method, three-dimensional simultaneous clustering method, and clustering method according to the embodiment of the present invention The simulator was implemented.

The input of the simulator was tested with a total of 120 signal data with a precise number of groups, and three input scenarios with two elements were used to confirm the difference between 3D and 3D clustering. .

Input scenario 1 is a signal whose arrival direction changes in the range of 0.3 to 0.8. The three-dimensional sequential clustering method and the three-dimensional simultaneous clustering method are separated into five and four groups, respectively, as shown in FIG. 6 (a), based on the threshold range of the arrival direction. On the other hand, according to the present invention, the reference for separating groups is a method for comparing a difference in the arrival direction of a continuous input signal with a threshold value. Therefore, the difference in arrival direction is all within the threshold value range, Group G1. Grouping considering differences in arrival directions may result in excessive grouping. Considering the performance of the signal analysis system, we set the time to limit the overload group and output the signal processing result within the response time limit.

Input scenario 2 is a signal whose arrival direction changes in the range of 0.1 to 0.5 and the arrival time also changes. The three-dimensional sequential clustering method and the three-dimensional simultaneous clustering method output three or two clusters according to the arrival direction threshold as shown in FIG. 7 (a). The present invention (four-dimensional simultaneous clustering method) is such that even if the difference in arrival direction is within the threshold value range as shown in FIG. 7B, the difference in arrival time between two consecutive signals in FIG. 7A exceeds the threshold value range It was confirmed that the cotton groups were separated and divided into different groups. When the difference between the two consecutive signals is larger than the threshold value by utilizing the difference in arrival time, the groups are divided into groups. Group is small and signal analysis time is reduced but signal analysis may not be possible in case of too small a signal group is merged after grouping is completed.

Input scenario 3 is a signal whose arrival direction changes in the range of 0.1 to 0.4 and the arrival time also changes. The three-dimensional sequential clustering method and the three-dimensional simultaneous clustering method are shown in FIG. 8 (a), and the results are shown as five clusters according to the threshold value of the arrival direction. In the present invention, Were separated by 10 clusters. The cluster G5 of FIG. 8 (b) forms a group differently according to the three-dimensional clustering method and the four-dimensional clustering method. In the three - dimensional clustering method, the group is classified by simply comparing the arrival direction of signals to a fixed threshold value range. The present invention is capable of generating a more detailed group by comparing and classifying threshold values for differences in arrival direction and arrival time.

In conclusion, the method of the present invention has the effect of tracking the signal of the moving radiator in comparison with the three-dimensional clustering method, and the difference of the arrival direction of the signal has the effect of tracking the moving radiator signal, . However, it is also confirmed that the group is divided into small pieces, and then a process of merging the groups having similar characteristics is performed after the group is completed as shown in FIG. 8 (c). Based on the characteristics of the G5, G6, and G9 groups in FIG. 8 (b), similar groups were merged to create one G5 group as shown in FIG. 8 (c). G5 were transmitted from the same radiation source, and the results were also classified into one group.

As a result of analysis of 73 input signals as shown in the following table, it is confirmed that the 4-dimensional clustering method can be grouped by 44% more than the 3-dimensional clustering method in order to obtain the difference between the three-dimensional cluster and the four- I could.

Clustering
Key influencers Input Count Number of 3D clusters Number of 4-dimensional clusters Arrival direction 14 5.7 4.3 arrival time 52 1.4 4.3 Arrival direction, arrival time 7 One 3.1 Total 73 Average 2.7 Average 3.9

While the present invention has been particularly shown and described with reference to exemplary embodiments thereof, it is to be understood that the invention is not limited to the disclosed exemplary embodiments, but, on the contrary, is intended to cover various modifications and equivalent arrangements included within the spirit and scope of the invention. Accordingly, the true scope of the present invention should be determined only by the appended claims.

Claims (9)

A first step of receiving a radar signal and a PDW list in which a plurality of PDW data generated by a signal collecting unit are sorted according to a pulse arrival time (TOA);
The distance to each group is calculated using the distance calculation function for each PDW data constituting the input PDW list, and the PDW data to be grouped is searched for a group existing at the closest distance or added to a new group, And a second step of clustering the group signals into group signals,
(DAOA, DFreq, DPW, and DPW) between the four signal characteristic variables of the PDW data to be grouped, ie, arrival direction, frequency, pulse width, arrival time, (DAOA, DFreq, DPW, DTOA) when the absolute values of the obtained four signal characteristic variables are within the respective threshold values defined for the respective signal characteristic parameters, And the distance is calculated by summing A method for clustering radar signals. The method according to claim 1, wherein, in the second step, the first PDW data to be grouped is generated as a first group in the absence of the generated group, and the value of the signal characteristic variables representing the first group to be generated is The arrival direction, the frequency, and the pulse width of the PDW data. The method according to claim 1, wherein in the second step,
The PDW data to be grouped is judged whether or not there is similarity between the PDW data and the grouped group, and if there is no similarity, the PDW data to be grouped is generated as a new group,
And determines that there is no similarity when the absolute values of the four signal characteristic parameters obtained from the distance calculation function deviate from the respective threshold values defined for the four signal characteristic parameters. Clustering method. [Claim 4] The method according to claim 3, wherein the value of the signal characteristic variables representative of the newly generated group is set as a pulse arrival time, an arrival direction, a frequency, and a pulse width value of PDW data to be grouped according to a result of the similarity determination result. Clustering method of radar signals. delete The method of any one of claims 1 to 4, wherein the value of the signal characteristic parameters representative of the group is an average value of each of the four signal characteristic parameters of the PDW data belonging to the same group. delete 2. The method according to claim 1, wherein, in calculating the distance between the PDW data and the group to be grouped by using the distance calculation function, the units of the four signal characteristic variables are different, A method for clustering a collection radar signal. 3. The method of claim 1, further comprising: a third step of finding and merging a group of proximate distances existing within a predetermined range when the number of groups generated after completing the second step exceeds a predetermined threshold value Way.

Images