Tracking processing device and tracking processing method
The tracking processing device accurately selects observation points by calculating signal similarity, addressing the issue of incorrect tracking due to different frequencies, ensuring precise target tracking.
Patent Information
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- KK TOSHIBA
- Filing Date
- 2024-11-26
- Publication Date
- 2026-06-05
AI Technical Summary
Conventional tracking processing methods risk incorrect tracking when selecting observation values that are targets with different frequencies, leading to inaccurate tracking.
A tracking processing device that calculates the direction and angle of arrival of radio waves, predicts the next observation point, collects characteristic information within a correlation gate, and selects the observation point most similar to the target based on calculated similarity using signal features like frequency, pulse width, and bandwidth.
Enables highly accurate target tracking by considering the specifications of observation values, preventing mistaken tracking of targets with different frequencies.
Smart Images
Figure 2026092461000001_ABST
Abstract
Description
Technical Field
[0006] , , , ,
[0005] , , , , , ,
[0001] This embodiment relates to a tracking processing device and a tracking processing method.
Background Art
[0002] As a method for tracking a moving target such as a flying object, the arrival direction of radio waves reflected or radiated from the moving target is estimated, the target angle is calculated from the estimated direction, and the position at the next observation time is predicted (see Patent Document 1). When performing tracking processing, there is a method of selecting the observation value (target angle) closest to the predicted value within a correlation gate centered on the predicted position (predicted value) (see Patent Document 2).
[0003] However, in the conventional tracking processing method as described above, since the observation value at the closest distance from the predicted value is automatically selected, if the observation value is a target with specifications of a completely different frequency, there is a risk of incorrect tracking.
Prior Art Documents
Patent Documents
[0004]
Patent Document 1
Patent Document 2
Summary of the Invention
Problems to be Solved by the Invention
[0005] As described above, in the conventional tracking processing device, there is a problem that, as a tracking target, even if the observation value at the closest distance from the predicted value is a target with specifications of a completely different frequency, it may be selected and incorrect tracking may be performed.
[0006] An object of this embodiment is to provide a tracking processing device and a tracking processing method that realize highly accurate target tracking in consideration of the specifications of observation values. [[ID=4
[0007] To solve the above problems, the tracking processing device according to the embodiment receives radio waves from a target, calculates the direction and angle of arrival of the radio waves, predicts the position of the next observation point, collects characteristic information of radio waves observed within a correlation gate centered on the predicted position, calculates similarity by comparing it with the characteristic information of the target, and selects the observation point of the radio waves that is most similar as the tracking target based on the calculated similarity. [Brief explanation of the drawing]
[0008] [Figure 1] Figure 1 is a block diagram showing the configuration of a tracking device according to the first embodiment. [Figure 2] Figure 2 is a flowchart showing the tracking process flow of the tracking processing device shown in Figure 1. [Figure 3] Figure 3 is a conceptual diagram showing the specific tracking process of the tracking processing device shown in Figure 1. [Figure 4] Figure 4 is a conceptual diagram showing the simulation results of target tracking processing using the tracking processing device shown in Figure 1. [Figure 5] Figure 5 is a vector diagram showing the vector similarity between data (b, c, d) and data a, when the feature quantities are frequency and pulse width, in the tracking processing device shown in Figure 1, with the feature quantities left unchanged. [Figure 6] Figure 6 is a vector diagram showing the vector similarity between data (b, c, d) and data (a) when the tracking processing device shown in Figure 1 scales the features to a range of 0 to 1 and then subtracts the median from the population, in the case where the features are frequency and pulse width. [Figure 7] Figure 7 is a block diagram showing the configuration of the tracking processing device according to the second embodiment. [Modes for carrying out the invention]
[0009] The embodiments will be described below with reference to the drawings.
[0010] (First Embodiment) The first embodiment will be described with reference to Figures 1 to 6.
[0011] Figure 1 is a block diagram showing the configuration of a target tracking device according to the first embodiment. In Figure 1, antennas 111 to 11N each capture radio waves from a target (a device that transmits radio waves). Receiving units 121 to 12N each receive radio waves captured by antennas 111 to 11N and output a received signal corresponding to the radio wave strength. The direction of arrival estimation unit 13 estimates the direction of arrival of the radio waves from the received signals output from the receiving units 121 to 12N and calculates the target angle. Meanwhile, the signal conversion unit 14 receives the received signal output from any of the receiving units 121, performs a predetermined conversion process, and outputs information (feature quantities) that indicate the characteristics of the signal. The tracking processing unit 15 performs tracking processing on the target angle originating from the arriving radio waves corresponding to the signal feature quantities obtained by the signal conversion unit 14, based on the target angle estimated by the direction of arrival estimation unit 13. The display unit 16 displays the target position based on the target angle after the tracking processing.
[0012] Figure 2 is a flowchart showing the processing flow of the tracking processing unit 15 shown in Figure 1. In the tracking processing unit 15, once the target to be tracked is determined (step S11), the destination of the target to be tracked is predicted and the predicted data is updated (step S12), and a correlation gate is opened centered on the position of the updated predicted data (step S13). Here, it is determined whether or not there is observed data within the correlation gate (step S14), and if it is determined that there is no observed data (NO), the target is considered lost (step S15), and the series of tracking processes is terminated.
[0013] In step S14, if it is determined that there is observed data within the correlation gate (YES), the similarity of the observed data is calculated (step S16), and it is determined whether or not there is observed data that satisfies the similarity criteria (step S17). If it is determined that there is no observed data (NO), the target is considered lost (step S15), and the series of tracking processes is terminated.
[0014] In step S17, if it is determined that there is observation data in the correlation gate (YES), it is further determined whether there is a plurality of observation data that satisfy the determination condition (step S18). If it is determined that there is not a plurality (NO), the observation data in the correlation gate is selected as the target (step S19). If it is determined that there is a plurality (YES), the observation data is selected based on the similarity calculation result (step S20), and smoothed data is calculated from the selected observation data and the prediction data and set as the target (step S22). When the target is specified in step S19 or step S21, it is determined whether to continue the tracking (step S22). If continuing, the time is advanced by t seconds (step S23), and the process returns to step S12 to continue the tracking process. If not continuing (NO), the tracking process is terminated.
[0015] In the above tracking process, in this embodiment, as a method for selecting observation data, within the correlation gate, the similarity of the signal is calculated by the following formula (1) using information indicating the characteristics of the following signals, and the observation data is selected based on the calculated similarity. By doing so, it becomes possible to select a more accurate target as the observation data.
[0016] The information indicating the characteristics of the signal (hereinafter referred to as the feature quantity) is shown below. · Frequency (Freq) · Pulse Repetition Interval (PRI) · Pulse Width (PW) · Bandwidth (BW) In formula (1), the smaller the value of the signal similarity Ex, the more similar the signal information is indicated.
[0017]
Equation
[0018] Figure 3 is a conceptual diagram showing the specific tracking process of the tracking processing device shown in Figure 1. (a) shows the smoothed data S(N) when the observed data O1 to O5 enter a correlation gate centered on the Nth predicted data F(N), and the result of calculating the similarity to the smoothed data S(N) using equation (1) from the signal parameters (frequency [MHz], pulse width [μs]) of the observed data O1 to O5. Furthermore, (b) shows how, within a correlation gate using the predicted data F(N-1), the observed data with the highest similarity (minimum value) to the predicted data F(N-1) is selected to calculate smoothed data S(N-1), the predicted data F(N) is determined from this smoothed data S(N-1), and within a correlation gate using the predicted data F(N), the observed data O1 with the highest similarity (minimum value) to the predicted data F(N) among the observed data O1~O5 in the correlation gate is selected to calculate smoothed data S(N), and the next predicted data F(N+1) is determined from this smoothed data S(N).
[0019] In this case, the conventional tracking method shown in Patent Document 2 selects the observation data O4 that is closest to the predicted data F(N), calculates the smoothed value between the predicted data F(N) and the observed data O4, and obtains the N+1th predicted data F(N+1). In contrast, in this embodiment, the signal similarity Ex of the observed data O1 to O5 is calculated using equation (1), and the observed data with the smallest similarity value (the one with the most similar signal information) is selected. In the example in Figure 3(a), the observed data O1 has the smallest similarity value, and as shown in Figure 3(b), the smoothed data S(N) is calculated from the observed data O1 and the predicted data F(N) to obtain the N+1th predicted data F(N+1).
[0020] To confirm the effectiveness of the tracking method in this embodiment, simulation data was created under certain conditions and evaluated.
[0021] Figure 4 is a conceptual diagram showing the simulation results of target tracking processing by the tracking processing device shown in Figure 1, where (a) is the conditions and (b) is the data of the simulation results. This simulation simulates a case where the actual target being tracked and multiple dummy targets are at similar distances and orientations from the perspective of the tracking processing device, which receives radio waves from the target, calculates the direction and angle of arrival, performs tracking processing, and displays the result.
[0022] Even under these conditions, we confirmed that by using the tracking method for calculating signal similarity according to this embodiment, the target being tracked can be correctly tracked without being mistaken for a dummy target, even as time passes.
[0023] Therefore, the tracking processing device of this embodiment receives radio waves from a target, calculates the direction and angle of arrival of the radio waves, predicts the position at the next observation time, collects characteristic information of the radio waves observed within a correlation gate centered on the predicted position, and tracks the observation point of the radio waves that is most similar to the characteristic information of the target. Thus, it is possible to achieve highly accurate target tracking that takes into account the parameters of the observed values.
[0024] (Variation 1) A modification 1 of the similarity calculation method of the above embodiment will be described. In this example, if there are N pieces of information indicating the characteristics of the signal for the observed data and the predicted data, these are treated as vectors with N dimensions, and the dot product of the vectors of the observed data and the predicted data is calculated as the similarity, and the most similar observed data is selected within the correlation gate.
[0025] The procedure for calculating similarity using the dot product is: (1) Scaling of each feature, (2) Calculation of vector centering, (3) Feature weighting, (4) Calculation of similarity using the dot product of vectors The calculations are performed in this order, and the most similar observed data is selected. The advantage of this modified method 1 is that the observed data can be normalized with respect to the population of data observed for each feature, so even if there is data with an extreme feature, it is possible to prevent the data from being affected by its value.
[0026] (1) Scaling of each feature Specify the range of possible values for each feature and transform them so that their values fall within the range of [0 to 1].
[0027]
number
[0028] (2) Calculation of vector centering The median of each feature in the observed data (the center of the vector for each feature) is calculated, and the vector is centered by subtracting the median from each feature.
[0029]
number
[0030] When performing vector centering, if the population distribution of each feature lies on a normal distribution, the value subtracted from each feature can be the population mean. However, if the population is skewed from a normal distribution, the mean may not work, so subtracting the median is more advantageous.
[0031] (3) Feature weighting Multiply each feature by the weighting coefficient using the following formula.
[0032]
number
[0033] (4) Calculation of similarity using the dot product of vectors The similarity between two vectors is calculated using the dot product of the two vectors as shown in the formula below.
[0034]
number
[0035] cos(x,y) = 0: The angle between the two vectors is 90 degrees, and they are independent and orthogonal vectors. This shows that the two features are independent of whether they are similar or dissimilar.
[0036] cos(x,y) = -1: The angle between the two vectors is 180 degrees, and they are vectors pointing in opposite directions. This indicates that the two features are not completely similar.
[0037] As an example of similarity calculation, we will explain the case with two features (frequency and pulse width) by referring to Figures 5 and 6.
[0038] Figure 5 is a vector diagram showing the vector similarity between data (b, c, d) and data a, when feature 1 is frequency and feature 2 is pulse width, as shown in Table 1, with feature 1 and 2 left unchanged.
[0039] [Table 1]
[0040] In other words, Table 1 shows the features as they are, and when we calculate the similarity to data a, as shown in Figure 5, the vectors of each data (b, c, d) point in the same direction, so the similarity values are almost the same.
[0041] Figure 6 is a vector diagram showing the vector similarity between data (b, c, d) and data (a) when, as shown in Table 2, feature 1 is frequency and feature 2 is pulse width, and the features are scaled to a range of 0 to 1 and then the median from the population is subtracted.
[0042] [Table 2]
[0043] In other words, as shown in Figure 6, the vector of data b points in the same direction as data a, while the vectors of data c and d point in opposite directions. Therefore, when calculating the similarity, data b will be approximately 1, and c and d will be -1, indicating that data b is similar to data a.
[0044] (Modification 2) In the above modified example 1, the similarity was calculated using frequency and pulse width, but it may also be calculated by taking into account other feature quantities (pulse repetition interval, bandwidth) as shown in the embodiment.
[0045] Furthermore, in this embodiment, a correlation gate is opened to select the observed values, but the terms for distance and arrival angle may also be included in the similarity formula in equation (1).
[0046] Furthermore, selecting observed values within the correlation gate is equivalent to filtering by distance and angle of arrival beforehand, so it is also possible to include terms for distance and angle of arrival in equation (1) to select observed values.
[0047] Similarly to the above, in Modification 1, distance and arrival angle terms may also be included in the features.
[0048] (Variation 3) In Modification 1, similarity was calculated to compare the predicted and actual values of the tracking process, but it may also be used to compare the features registered in the database with the target features.
[0049] Figure 7 is a block diagram showing the configuration of a tracking processing device using a database, as a third modified example. In Figure 7, the same parts as in Figure 1 are denoted by the same reference numerals, and redundant explanations are omitted.
[0050] The configuration shown in this modified example 3 adds a classification unit 17 and a database 18 to the tracking processing device shown in Figure 1. The signal features obtained by the signal conversion unit 14 are sent to the classification unit 17. The database 18 has various features of the true target registered in advance. The classification unit 17 compares the signal features from the signal conversion unit 14 with the signal features of the true target registered in the database 18 to calculate the similarity, and then identifies the most similar signal from the calculation result and sends it to the tracking processing unit 15.
[0051] In the tracking processing device with the above configuration, the classification unit 17 identifies the signal with the most similar signal features by comparing it with the signal features of the true target registered in the database 18, and then performs tracking processing. This enables highly accurate target tracking that takes into account the characteristics of the observed values.
[0052] It should be noted that the present invention is not limited to the above embodiments, and the components can be modified and implemented in practice without departing from the spirit of the invention. Furthermore, various inventions can be formed by appropriately combining the multiple components disclosed in the above embodiments. For example, some components may be deleted from all the components shown in the embodiments. Moreover, components from different embodiments may be appropriately combined. [Explanation of Symbols]
[0053] 111-11N... Antenna, 121-12N... Receiving unit, 13... Direction of arrival estimation unit, 14... Signal conversion unit, 15... Tracking processing unit, 16... Display unit, 17... Classification unit, 18... Database.
Claims
1. A receiving unit that receives radio waves from the target, A processing unit calculates the direction and angle of arrival of the radio waves received by the receiving unit to predict the position of the next observation point, collects characteristic information of the radio waves observed within a correlation gate centered on the predicted point at the predicted position, calculates the similarity by comparing it with the characteristic information of the target, and selects the observation point of the radio waves that is most similar to the calculated similarity as the target for tracking. A display unit that displays the tracking results of the aforementioned processing unit and A tracking processing device equipped with the following.
2. The tracking processing device according to claim 1, wherein the selected observation point is within the correlation gate.
3. The tracking processing device according to claim 1, wherein the processing unit calculates the dot product of the vectors of feature quantities representing the characteristics of the radio waves as the similarity of the feature information.
4. The tracking apparatus according to claim 3, wherein the processing unit performs scaling of each feature quantity used in calculating the similarity.
5. The tracking apparatus according to claim 3, wherein the processing unit calculates the center of the vector of each feature, which is the median of each feature in the population used to calculate the similarity, and performs vector centering by subtracting the median from each feature.
6. The tracking apparatus according to claim 3, wherein the processing unit weights each feature quantity when calculating the similarity.
7. The tracking apparatus according to claim 1, wherein the processing unit compares characteristic information of radio waves observed within the correlation gate with characteristic information of tracking targets registered in a database in advance to calculate similarity, and determines whether the observed radio waves are tracking targets registered in the database based on the calculated similarity.
8. The tracking processing device according to claim 7, wherein the processing unit calculates the dot product of the vectors of each feature quantity representing the characteristics of the radio waves in order to calculate the similarity of the feature information.
9. The tracking apparatus according to claim 7, wherein the processing unit performs scaling of each feature quantity indicating the characteristics of the radio waves in order to calculate the similarity of the feature information.
10. The tracking processing device according to claim 7, wherein the processing unit calculates the median value (center of the vector of each feature) within the population of each feature quantity that represents the characteristics of the radio waves, and performs vector centering by subtracting the median value from each feature quantity, in order to calculate the similarity of the feature information.
11. The tracking apparatus according to claim 7, wherein the processing unit performs a calculation in which weights each feature quantity indicating the characteristics of the radio waves in calculating the similarity of the feature information.
12. Receiving radio waves from the target, The direction and angle of arrival of the received radio waves are calculated to predict the position of the next observation point. We collect characteristic information of radio waves observed within a correlation gate centered on the predicted point of the predicted location. The similarity is calculated by comparing it with the aforementioned target feature information. Based on the calculated similarity, the observation point of the most similar radio waves is selected as the target for tracking. Display tracking results for the selected tracking target. Tracking method.