CN115220002B - A method and device for multi-target data association tracking at a fixed single station - Google Patents

Abstract

The embodiment of the present application discloses a multi-target data association tracking method of a fixed single station, which is used to more accurately associate data with the observation values of multiple target objects. The method includes: obtaining a first set of actual observation values of at least two target objects at a target time; determining the association center value of each target object according to the target time period at the target time; based on a joint maximum likelihood estimation model, determining a second set of actual observation values of each target object from the first set of actual observation values according to the association center value; based on a single target tracking model, determining the motion state information of each target object at the target time according to the second set of actual observation values, and the motion state information is used to determine the track of each target object in the target time period.

Description

A method and device for multi-target data association tracking at a fixed single station

Technical Field

The present application relates to the field of radar data processing technology, and in particular to a multi-target data association tracking method and related devices for a fixed single station.

Background technique

In a passive positioning system, due to the presence of environmental noise and random disturbances, multiple observations may be obtained in one detection, and among these observations, it is unknown which ones are from the tracked target and which ones are false observations. When multiple target objects need to be observed, some observations that are far from the predicted values of the target object's motion state are also uncertain to which target object they belong. This requires associating the observations with the target objects, which is called data association.

One method of data association is to predict the motion state of the target object at the next moment based on the observed value at the previous moment, compare the predicted value with the actual value, calculate the data with smaller differences through maximum joint likelihood estimation, and determine the associated pair data corresponding to the actual value and the target object.

However, the positioning error is too large at the beginning of the track, and the filter has not yet reached a stable convergence effect. Therefore, if the track measurement prediction value and the actual measurement are used for correlation at this time, correlation errors are likely to occur, which may lead to incorrect correlation of the entire track in severe cases.

Summary of the invention

The main purpose of the embodiments of the present application is to propose a fixed single-station multi-target data association tracking method and related devices, aiming to improve the accuracy of data association in multi-target tracking.

In a first aspect, an embodiment of the present application provides a multi-target data association tracking method for a fixed single station, the method comprising the following steps: obtaining a first set of actual observation values of at least two target objects at a target moment; determining an association center value of each of the target objects based on a target time period in which the target moment is located; based on a joint maximum likelihood estimation model, determining a second set of actual observation values of each of the target objects from the first set of actual observation values based on the association center value; based on a single target tracking model, determining motion state information of each of the target objects at the target moment based on the second set of actual observation values, the motion state information being used to determine the track of each of the target objects in the target time period.

In a second aspect, an embodiment of the present application provides a passive positioning system, which includes a memory, a processor, a program stored in the memory and executable on the processor, and a data bus for realizing connection and communication between the processor and the memory, wherein the program implements the steps of the aforementioned method when executed by the processor.

In a third aspect, an embodiment of the present application provides a storage medium for computer-readable storage, wherein the storage medium stores one or more programs, and the one or more programs can be executed by one or more processors to implement the steps of the aforementioned method.

The information entropy-based joint maximum likelihood data association method and related devices proposed in the present application determine the association center value in the target time period according to the target time period in which at least two target objects are located at the target moment, determine the motion state information of each target object at the target moment according to the second actual observation value set of each target object, and further track each target object. It is possible to use different association center values at different stages of the track according to the different characteristics of the track at different stages to determine the second actual observation value set of each target object from the first actual observation value set of at least two target objects, thereby more accurately performing data association on the observation values of multiple target objects.

BRIEF DESCRIPTION OF THE DRAWINGS

In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

FIG1 is a schematic diagram of the architecture of a fixed single-station passive positioning multi-target tracking method provided in an embodiment of the present application;

FIG2 is a schematic diagram of a step flow chart of a method for multi-target data association tracking at a fixed single station provided in an embodiment of the present application;

FIG3 is a schematic diagram of the architecture of a multi-target data association method for a fixed single station provided in an embodiment of the present application;

FIG4 is a schematic diagram of another step flow chart of the method for multi-target data association tracking of a fixed single station provided in an embodiment of the present application;

FIG5 is a schematic diagram of another step flow chart of the method for multi-target data association tracking of a fixed single station provided in an embodiment of the present application;

FIG6 is a schematic diagram of another step flow chart of the method for multi-target data association tracking of a fixed single station provided in an embodiment of the present application;

FIG7 is a schematic diagram of another step flow chart of the method for multi-target data association tracking of a fixed single station provided in an embodiment of the present application;

FIG8 is a schematic diagram of another step flow chart of the method for multi-target data association tracking of a fixed single station provided in an embodiment of the present application;

FIG9 is a real trajectory diagram in an application scenario of the multi-target data association tracking method for a fixed single station provided in an embodiment of the present application;

FIG10 is a diagram of real and estimated trajectories in an application scenario of the method for multi-target data association tracking at a fixed single station provided by an embodiment of the present application;

FIG11 is a root mean square error diagram of eight targets in an application scenario of the multi-target data association tracking method of a fixed single station provided in an embodiment of the present application;

FIG12 is a graph showing average position errors of eight targets in an application scenario of the method for multi-target data association tracking at a fixed single station provided by an embodiment of the present application;

FIG. 13 is a schematic diagram of the structure of a passive positioning system provided in an embodiment of the present application.

Detailed ways

Radar data processing and radar signal processing are both important components of modern radar systems. Signal processing is used to detect targets and obtain various useful information about targets, such as distance, speed, and shape, using certain methods. Data processing can further process the target's point traces and tracks, predict the target's position at future times, form a reliable target track, and thus achieve real-time tracking of the target.

Radar data processing includes several main links, such as point track condensation, track initiation, target tracking, and multi-target association. The two basic problems it studies are the association between point tracks and point tracks, and point tracks and tracks in different environments. The former involves track initiation, focusing on the control of the point track correlation range and the selection of related algorithms; the latter involves target tracking, focusing on the application of target motion models and filtering algorithms. The purpose of radar data processing is to use the target information provided by the radar to estimate the target's track and give the target's position at the next moment.

The radar data processing process mainly includes data preprocessing, track initiation, data association, tracking filtering, track disappearance and quality assessment.

The input data of radar data processing is also called observation. Observation is not the data directly scanned by radar, but the data scanned by radar is first processed by radar signal and then obtained by data acquisition. General observation includes radar scanning cycle, radar scanning batch, the number of targets scanned in each batch and the specific information of each target (radial distance, azimuth, pitch angle). In actual engineering, observation is generally mixed with noise pollution, which mainly comes from the following aspects:

1) Random false alarms during the scanning process;

2) Clutter generated by false targets;

3) Interference target;

4) Bait, etc.

Although modern radar signal processing technology has made great progress, there will still be some interference in the observation even after signal processing, and the amount of observation data is generally large, which places high requirements on subsequent computer storage and processing. Data preprocessing is to screen the data before the observation data is started, associated, or other data processing processes, and those data that are not within the threshold are eliminated. Only data that passes all judgment thresholds is retained.

The benefit of observation data preprocessing is that it can significantly reduce the size of data in the subsequent data processing process and greatly reduce the amount of calculation. To a certain extent, it can reduce the burden on the computer, improve the speed of data processing and the accuracy of target tracking, and reduce the possibility of false tracks.

Gate is a very important concept in data processing, and it is used in track initiation, data association, etc. A gate is actually an area, which is generally divided into initial gate and related gate.

The initial gate is generally used in the initial stage of the track. It is an area centered on an arbitrary point, which specifies a spatial range where the target's observation value may appear. Since the target is far away at the beginning of the track, in order to better capture the target, the initial gate is generally established as a large gate.

The so-called correlation gate refers to a spatial area centered on the predicted value of the tracked target. This area specifies the possible range of the observed value of the tracked target. The shape and size of the correlation waveform are determined by, on the one hand, making the probability of the real observation falling into the gate very high, and on the other hand, not allowing too many irrelevant observation points in the correlation gate. Generally, the size of the correlation gate should match the target type. For example, the gate of a fixed target generally only depends on the accuracy of the observation, the gate of a linear target depends on the accuracy of the observed value and the prediction filter, and the gate of a maneuvering target also needs to consider the acceleration factor. The more commonly used correlation gates include rectangular gates, circular gates, elliptical gates, and fan gates in polar coordinate systems.

Track initiation refers to the first point of establishing the target track, that is, the process from the target falling into the radar detection range to the target track establishment. The track initiation process is an important link in the radar data processing process. As the saying goes, "well begun is half done." On the contrary, if the track initiation is not successful, the track establishment will not be smooth, and it is very likely that a reliable track cannot be established, thus failing to achieve correct tracking of the target.

The track initiation process is one of the important links in the radar data processing process. One of the tasks of track initiation is to quickly establish a track for the target entering the radar power area, and the second task is to avoid false tracks as much as possible. However, in order to avoid the establishment of false tracks, it takes a long time to start. It can be seen that there is a certain contradiction between the two tasks, the contradiction between speed and quality, so it is necessary to find an optimal compromise between the two. There are many track initiation algorithms, and the more commonly used ones are intuitive method, logical method, modified logical method, etc.

In the ideal target motion model, the observation environment is always considered to be "clean", and only one observation value is detected each time, and this observation value comes from the target being tracked. However, in the actual system, the environment is not ideal. Due to the existence of factors such as observation noise, false alarms and other phenomena may occur. In addition, the random interference in the observed area will cause clutter in the area where the target may appear. In short, multiple observation values may be obtained in one detection, and among these observation values, it is unknown which ones come from the tracked target and which ones are false observation values. This factor determines that the data association process is an important link in the radar data processing system.

When there is only one target in the radar scanning area and there is no interference, there will be only one point track in the relevant wave gate of the target, and there is no problem of data association. However, when there are multiple targets in the radar scanning area, or there is clutter, the same point track may fall into multiple wave gates or multiple point tracks may appear in the same wave gate, which involves the problem of data association. Data association is to determine the relationship between the radar observation data at a certain moment and the observation data at other moments or the existing tracks, so as to achieve the process of pairing the point track and the track.

Generally speaking, data association can be divided into the following situations according to the different objects that are related to each other:

1) Track start: the connection between track points;

2) Track update: the interconnection between point track and track (track prediction point), which can also be called track keeping;

3) Track fusion: the interconnection of tracks.

There are many methods for data association, which can be roughly divided into two categories: one is the Bayesian data association algorithm, and the other is the maximum likelihood data association algorithm. Bayesian algorithms mainly include the nearest field algorithm, the probabilistic data association algorithm, etc. Bayesian algorithms are all based on the Bayesian criterion. The maximum likelihood algorithms mainly include the aerial integral bifurcation method and the joint maximum likelihood algorithm, etc., and the maximum likelihood algorithms are all based on the likelihood ratio of the observation sequence.

The embodiments of the present application provide a fixed single-station multi-target data association tracking method and related devices, which can use different association center values at different stages of the track according to different characteristics of the track at different stages to determine the second actual observation value set of each target object from the first actual observation value set of at least two target objects, thereby more accurately performing data association on the observation values of multiple target objects.

The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

The flowcharts shown in the accompanying drawings are only examples and do not necessarily include all the contents and operations/steps, nor must they be executed in the order described. For example, some operations/steps may also be decomposed, combined or partially merged, so the actual execution order may change according to actual conditions.

It should be understood that the terms used in this application specification are only for the purpose of describing specific embodiments and are not intended to limit the application. As used in this application specification and the appended claims, unless the context clearly indicates otherwise, the singular forms "a", "an" and "the" are intended to include plural forms.

It should also be understood that the term “and/or” used in the specification and appended claims refers to and includes any and all possible combinations of one or more of the associated listed items.

In conjunction with the accompanying drawings, some embodiments of the present application are described in detail below. In the absence of conflict, the following embodiments and features in the embodiments can be combined with each other.

Please refer to Figure 1, which is a schematic diagram of the architecture of the fixed single-station passive positioning multi-target tracking method provided in an embodiment of the present application.

The passive positioning system obtains a set of actual observation values of multiple target objects at different times. The actual observation values may include azimuth information, azimuth change rate information, Doppler frequency information, Doppler frequency change rate information, etc. Each actual observation value in the set of actual observation values of multiple target objects does not necessarily have a one-to-one correspondence with the target object. Due to the existence of random disturbances and environmental noise, even if the correspondence between the actual observation value at the initial time and each target object is determined, the correspondence between the actual observation value at subsequent times and each target object cannot be determined.

Through data association, the actual observation values of multiple targets can be associated with each target object to determine the actual observation value corresponding to each target object at the target time.

After completing data association and determining the actual observation value corresponding to each target object at the target time, the multi-target tracking problem is transformed into a single target tracking problem. It is necessary to make a maneuver decision for each target object and determine the start and end of the track of each target object by using the tracking gate rules, filtering and prediction.

Once the start and end of the track of each target object are determined, multiple target objects can be comprehensively analyzed to ultimately determine the state of the system composed of the multiple target objects.

Based on the architectural diagram of the fixed single-station passive positioning multi-target tracking method shown in Figure 1, an embodiment of the present application provides a multi-target tracking method with joint maximum likelihood data association.

Please refer to FIG. 2 , which is a schematic flow chart of a step flow of a method for multi-target data association tracking of a fixed single station provided in an embodiment of the present application.

201. Obtain a first actual observation value set of at least two target objects at a target time.

An observation station in a passive positioning system observes a motion system including at least two target objects, and obtains a first actual observation value set of the motion system including at least two target objects at a target time by analyzing electromagnetic wave signals generated by the motion system including at least two target objects and received by the observation station. The first actual observation value set represents the actual observation value set corresponding to the motion system including at least two target objects at the target time, and the correspondence between each target object in the motion system and each actual observation value in the first actual observation value set is not accurately established.

202. Determine the association center of each target object according to the target time period at the target moment.

To realize the positioning and tracking of a motion system including at least two target objects, it is necessary to associate the actual observation values in the first actual observation value set representing the actual observation values of the motion system as a whole with each target object, that is, to determine the actual observation value of each target object. The association center is the basis for establishing the association process between each actual observation value in the first actual observation value set and each target object. The association center is not fixed, but is dynamically adjusted according to the target time period of the target moment when the target object moves at the target moment.

203. Based on the joint maximum likelihood estimation model, determine a second actual observation value set for each target object from the first actual observation value set according to the association center value.

After determining the association center value of the target moment according to the target time period in which the target moment of a motion system including at least two target objects is located, based on the joint maximum likelihood estimation model, the association center value is used as the basis for establishing an association between each actual observation value in the first actual observation value set and each target object, and a second actual observation value set for each target object is determined from the first actual observation value set, and each actual observation value in the second actual observation value set has a corresponding relationship with each target object.

204. Based on the single target tracking model, determine the motion state information of each target object at the target time according to the second actual observation value set, and the motion state information is used to determine the track of each target object in the target time period.

After establishing an association relationship between each actual observation value in the first actual observation value set and each target object, the multi-target tracking problem is transformed into a single target tracking problem, that is, the determination of the motion state of a motion system containing at least two target objects is transformed into the determination of the motion state of each target object in the motion system. Based on the single target tracking model, the motion state information of each target object at the target time is determined according to the second actual observation value set corresponding to each target object, and the track of each target object in the target time period is determined according to the motion state information of each target object at multiple times, so as to realize the positioning and tracking of each target object.

Based on the multi-target data association tracking method of a fixed single station shown in FIG2 , there are multiple ways to determine the association center of each target object according to the target time period at the target moment in step 202. Determining different association centers for data association at different stages according to the motion characteristics and observation value characteristics of a motion system including at least two target objects at different stages can improve accuracy. The following introduces the selection of different association centers at different stages.

Please refer to FIG. 3 , which is a schematic diagram of the architecture of a multi-target data association method for a fixed single station.

Data association is the core of the multi-target tracking solution, which can transform the multi-target tracking problem into single target tracking. Data association in multi-target tracking includes three steps: association preprocessing, selection of association center, calculation and association of information entropy uncertainty.

The first step is to preprocess the measurement information. A tracking gate is set up. All observations received by the observation station will pass through this gate. The correlation gate will filter out observations with excessive errors. Observations that pass the correlation gate are candidate observations.

Then comes the selection of the association center. Since the initial positioning error of a fixed single station is too large, the observed grey prediction value is used as the association center at the beginning of the track. After the target tracking is stable, the observed prediction value is used as the association center.

After determining the association center, we use an uncertainty association method based on information entropy to match the observations and association centers of each target to construct association pairs. Each association pair means that each target has an observation paired with it, which transforms the multi-target tracking problem into a single-target tracking problem.

After data association builds the association pairs, the tracking of each target is a single target tracking problem. The tracking of a single target includes the initial positioning of the target, maneuver judgment, and filtering.

At the beginning of the track, the initial state of the target can be directly calculated using the relationship between the observations of each target and the target state. Taking the initial state of each target as the starting point, the tracking of each target track at subsequent moments is estimated using observation information and filtering algorithms.

Before updating the target state using the observation and filtering algorithm, a judgment of the target maneuver is made. When the target maneuvers are detected, the covariance is adaptively adjusted and then filtered to complete the tracking of the maneuvering target.

Finally, the filtering algorithm and the observations in the associated pairs are used to estimate the state of the targets in the associated pairs.

In combination with the embodiments shown in FIG. 2 and FIG. 3 , different association centers are used in different stages. The track start stage in the embodiment shown in FIG. 3 is introduced below based on the embodiment shown in FIG. 2 .

Please refer to FIG. 4 , which is a schematic diagram of another step flow chart of a method for multi-target data association tracking of a fixed single station at the initial stage of the track. Step 202 in FIG. 4 includes step 2021 and step 2022 .

2021. When the target time period indicates that at least two target objects are at the start stage of the track, the grey prediction value of each target object at the target time is determined based on the grey prediction model.

When the target time period of the current target moment of at least two target objects represents the initial stage of the track, the previous n measurements that have been successfully associated are used as prediction preparation data, and then a set of prediction data is obtained using the gray prediction model. The gray prediction model is a mathematical model that can be established using only a small amount of incomplete information, and then the model can be used to predict the next prediction value. At present, there are many typical gray prediction methods, such as regression analysis.

The first step is cumulative generation. The data of the previous moment is accumulated in sequence to form a new sequence. Let the original sequence be

z ⁽⁰⁾ =(z ⁽⁰⁾ (0),z ⁽⁰⁾ (1),…z ⁽⁰⁾ (k)) (1)

One accumulation generation is expressed as:

GM (1.1) is one of the grey prediction systems and is a first-order differential equation model in the form of:

The prediction formula is as follows:

By the definition of derivative:

When Δt is 1, then:

Arranging equations (4) and (6) to obtain the solution formula for the predicted value is:

The corresponding prediction value at time k+1 is:

2022. Determine the grey prediction value as the association center value.

After the grey prediction value is obtained based on the grey prediction model, the grey prediction value is used as the correlation center value of the target object at the target time in the initial stage of the track.

Based on the embodiment shown in FIG. 4 , the actual observation value corresponding to each object is determined according to the correlation center value and is specifically screened from the first actual observation value set using a correlation gate.

Please refer to FIG. 5 , which is a schematic diagram of another step flow chart of the multi-target data association tracking method for a fixed single station. Step 203 in FIG. 5 includes steps 2031 to 2033 .

2031. Determine the relevant wave gate of each target object according to the grey prediction value.

The observation of a single target should pass through a relevant wave gate, and the gray prediction value of each target object at the target time is used as the wave gate center of each target object. After determining the wave gate shape, the relevant wave gate of each target object at the target time can be determined.

It should be noted that the gate shape can be selected from a variety of shapes, such as an elliptical gate or a square gate. In this embodiment, the fixed single-station multi-target tracking solution selects an elliptical gate as the relevant gate.

2032. Determine a candidate observation value set for each target object from the first actual observation value set according to the relevant wave gate.

The correlation gate can limit the number of observations involved in the correlation judgment and filter out observations with too large errors. Each target object has a correlation gate, and the candidate observation value set of each target object can be screened from the first actual observation value set based on the correlation gate of each target object.

2033. Based on the joint maximum likelihood estimation model, determine the second actual observation value set according to the candidate observation value set and the grey prediction value.

After the candidate observation value set of each target object is determined, a second actual observation value set of each target object is determined according to the candidate observation value set of each target object and the grey prediction value based on the joint maximum likelihood estimation model.

In conjunction with the embodiment shown in FIG. 5 , the following describes how to determine the second actual observation value set for each target object based on the candidate observation value set for each target object in an embodiment of the present application.

Please refer to FIG. 6 , in which step 2033 includes steps 20331 to 20333 .

20331. Calculate the difference between the grey prediction value and each candidate observation value in the candidate observation value set.

Suppose the matrix composed of m system objectives and n measurement indicators is:

Where λ _ij represents the jth measurement indicator of the jth target.

Difference the measured value and the measured predicted value in matrix A, and then take the absolute value, recorded as:

In the formula, _Δij represents the difference between the jth measurement indicator of target i and its corresponding predicted value, that is,

20332. Based on the information entropy calculation model, the difference is taken as input to obtain the expanded uncertainty of each candidate observation value.

Information entropy is a measure of the disorder of a system. Assuming that the system contains multiple measurement indicators j (j = 1, 2, ..., n), the proportion of a single measurement indicator of each target i (i = 1, 2, ..., m) is p _ij , then the entropy of the measurement indicator is defined as:

when When , that is, when the proportions of various measurement indicators are the same, the maximum entropy can be expressed as:

H＝log ₂ m _{j max} (13)

Therefore, if the information entropy of a measurement indicator is smaller, the amount of information provided by the measurement indicator is greater.

The proportion of the jth measurement indicator of target i is:

The expanded uncertainty of the lth measurement indicator is defined as:

The larger s is, the greater the role of the measurement indicator in the target tracking process.

20333. Based on the joint maximum likelihood estimation model, determine the second set of actual observation values according to the expanded uncertainty of each valid observation value.

Assume that the state estimates of n targets are obtained at time k-1 At time k, there are m _k measurements that pass the correlation threshold. The set of all confirmed measurements at time k is recorded as:

The event that the jth measurement in the measurement set at time k is confirmed to belong to target i is recorded as Divide the measurements in each gate into multiple feasible partitions,

The measurements within the feasible partition r shall meet the following requirements:

and

is an empty set, the above formula means that a measurement can only belong to one track.

For each feasible partition r, we can define an event

θ(r) = {partition measurement combination r is true} (20)

The set of all feasible partitions can be defined as:

Γ＝{r} (21)

The uncertainty of information entropy is used as the metric distance between the lth indicator of a certain measurement partition and the lth indicator of the measurement prediction value.

The result of information entropy uncertainty will appear complex, and the result modulo represents the metric distance between the lth indicator of a certain measurement partition and the corresponding indicator of the measurement prediction value, that is,

The improved information entropy uncertainty better reflects the difference between a certain indicator of the measurement partition and the measurement prediction. The smaller the uncertainty, the closer the indicator of the measurement partition is to the corresponding indicator of the measurement prediction value. Therefore, the information entropy uncertainty of all indicators of a certain measurement partition is combined, and the one with the smallest information entropy combined uncertainty value is the maximum possible measurement partition, that is,

Based on the embodiments shown in FIG. 2 to FIG. 6 , the candidate observation value set of each target object is determined from the first actual observation value set according to the relevant wave gate, and can also be screened from the first actual observation value set in a general manner of adding wave gate constraints.

Please refer to FIG. 7 . Step 2032 in FIG. 2 , FIG. 4 to FIG. 6 includes steps 20321 to 20323 .

20321. Determine a set of predicted values of measurement indicators of each target object at a target time.

A measurement indicator prediction value set of each target object at a target time is determined, wherein the measurement indicator prediction value set includes at least one of an azimuth angle prediction value set, a Doppler frequency prediction value set, an azimuth angle change rate prediction value set, and a Doppler frequency change rate prediction value set.

20322. Determine the gate constraints corresponding to the relevant gates according to at least two sets of predicted measurement index values and a set of actual measurement index values of each target object at a moment before the target moment.

At least two measurement indicator prediction value sets are determined. In this embodiment, an azimuth angle prediction value set and a Doppler frequency prediction value set are selected.

Based on the elliptical wave gate as the correlation wave gate, wave gate constraints on two measurement indicators, azimuth angle and Doppler frequency, are added.

The azimuth angle of each target that has been associated and measured at time k-1 is The Doppler frequency is/> The measured azimuth at time k is predicted to be/> The measured Doppler frequency is predicted to be/> Then the constraint range of the azimuth is:

Where β _limit is the azimuth constraint size.

The constraint range of Doppler frequency is:

Where f _limit is the Doppler frequency constraint.

20323. Determine a candidate observation value set from the first actual observation value set according to the gate constraint.

Only observations that fall into the elliptical gate and are within the azimuth constraint and Doppler frequency constraint are valid measurements of the target. If no observation falls into the gate, the point is updated with the predicted value of the observation. If no observation falls on the track for multiple consecutive times, the track is recorded as disappeared. The candidate observation value set for each target object is determined from the first actual observation value set according to the gate constraint.

In combination with the embodiments shown in FIG. 2 and FIG. 3 , different association centers are used in different stages. The track stabilization stage in the embodiment shown in FIG. 3 is introduced below based on the embodiment shown in FIG. 2 .

Please refer to FIG. 8 , which is a schematic flow chart of another step of the multi-target data association tracking method for a fixed single station in the track stabilization phase. Step 202 in FIG. 8 includes step 2023 and step 2024 .

2023. When the target time period indicates that at least two target objects are in a stable track phase, determining an observation prediction value of each target object at the target time based on the track prediction model;

When the target time of at least two target objects is in the target time period representing the track stabilization stage, the observation prediction value of each target object at the target time is determined based on the track prediction model. The observation prediction value is obtained based on the track prediction model using the observation actual value as input.

2024. Determine the observed prediction value as the associated center value.

After the observation prediction value is determined, it is used as the correlation center when the target time period is in the stable stage of the aircraft.

It should be noted that the technical solutions of the embodiments shown in Figures 5 to 7 can all be applied to the technical solutions of the aircraft stabilization stage represented by the embodiment shown in Figure 8, that is, Figure 8 and Figure 5, Figure 8 and Figure 5 and Figure 6, Figure 8 and Figure 5, Figure 6 and Figure 7 can all be combined to form technical solutions similar to the embodiments shown in Figures 5, 6 and 7, and the details will not be repeated here.

An application scenario of the technical solution formed by combining the embodiments shown in FIG. 2 to FIG. 8 is shown in FIG. 9 .

Please refer to FIG. 9 , which is a real trajectory diagram in an application scenario of the multi-target data association tracking method of a fixed single station provided in an embodiment of the present application.

In the two-dimensional scenario, the observation station is located at the origin of the coordinate system, the target is 200-300km away from the observation station, and the target radiation source moves at a uniform speed with acceleration disturbance. The total duration of the target movement is 300s. The frequency of the target radiation source is 10GHz, and the target state is: The observed quantities include azimuth angle α, azimuth angle change rate/> Doppler frequency f, Doppler frequency change rate/> The standard deviations of the observation noise are: σ _α = 2°, /> σ _f = 1 Hz, /> The observation period is T = 1s, and various motion scenarios are simulated to record the impact of various scenarios on performance indicators. Simulation scenario: 8 targets move at a constant speed with turning maneuvers. The initial states of the targets are:

(-20km, 0.3km/s, 285km, 0.1km/s), (100km, -0.3km/s, 285km, 0.1km/s),

(-10km, 0.3km/s, 310km, 0km/s), (-10km, 0.3km/s, 290km, 0km/s),

(60km,-0.3km/s,285km,0.1km/s), (-45km,0.3km/s,270km,0.3km/s),

(90km, 0.3km/s, 305km, -0.2km/s), (-50km, -0.3km/s, 260km, 0.2km/s);

In the two-dimensional plane, the target movement range is in the first and second quadrants. Targets 1-5 perform uniform linear motion, and target 6 performs uniform linear motion with direction adjustment. Maneuvers occur at 100s and 200s respectively. Targets 7 and 8 first perform a section of uniform linear motion, then perform uniform turning motion, and finally perform another section of uniform linear motion. The turning rate is 2.9°/s and the turning radius is 10km.

Figure 9 shows the real track of each target, and Figure 10 shows the tracked track of each target. It can be seen from Figures 9 and 10 that the proposed algorithm can effectively track uniform linear motion targets, uniform turning maneuvering targets, and maneuvering targets with changing heading angles. Figure 11 is the root mean square error diagram of the positions of each target, and Figure 12 is the average position error diagram of each target. Figure 11 clearly reflects the tracking performance of each target. The tracking effect of uniform linear motion targets is ideal, while the tracking error of maneuvering targets is significantly higher than that of uniform linear motion targets. This is because the covariance matrix is adaptively adjusted when the target is detected to be maneuvering during maneuver detection. The fuzzy Gaussian particle filter re-tracks the target in a short time, and the tracking error then gradually decreases. Figure 12 can reflect the performance of the entire multi-target tracking system, and clearly shows the effectiveness of the data association algorithm proposed in this paper.

Please refer to FIG. 13 . An embodiment of the present application provides a passive positioning system, which includes a processor, a memory, and a network interface connected via a system bus, wherein the memory may include a non-volatile storage medium and an internal memory.

The non-volatile storage medium can store an operating system and a computer program. The computer program includes program instructions, and when the program instructions are executed, the processor can execute any video data processing method.

The processor is used to provide computing and control capabilities and support the operation of the entire computer equipment.

The internal memory provides an environment for the operation of the computer program in the non-volatile storage medium. When the computer program is executed by the processor, the processor can execute any fixed single-station multi-target data association tracking method.

The network interface is used for network communication, such as sending assigned tasks, etc. Those skilled in the art will appreciate that the structure shown in FIG13 is only a block diagram of a portion of the structure related to the present application solution, and does not constitute a limitation on the computer device to which the present application solution is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

It should be understood that the processor may be a central processing unit (CPU), and the processor may also be other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. Among them, the general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

A computer-readable storage medium is also provided in an embodiment of the present application, wherein the computer-readable storage medium stores a computer program, wherein the computer program includes program instructions, and the processor executes the program instructions to implement any fixed single-station multi-target data association tracking method provided in an embodiment of the present application.

The computer-readable storage medium may be an internal storage unit of the computer device described in the above embodiment, such as a hard disk or memory of the computer device. The computer-readable storage medium may also be an external storage device of the computer device, such as a plug-in hard disk, a smart memory card (SmartMedia Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), etc. equipped on the computer device.

The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of various equivalent modifications or replacements within the technical scope disclosed in the present application, and these modifications or replacements should be included in the protection scope of the present application. Therefore, the protection scope of the present application shall be based on the protection scope of the claims.

Claims (3)

1. A method for multi-target data association tracking of a fixed single station, characterized by comprising: Acquire a first set of actual observation values of at least two target objects at a target time; Determining the association center value of each target object according to the target time period in which the target moment is located includes: When the target time period indicates that at least two target objects are in the initial stage of the track, determining a grey prediction value of each target object at the target time based on a grey prediction model; determining the grey prediction value as the correlation center value; When the target time period indicates that at least two of the target objects are in a stable track phase, determining an observation prediction value of each of the target objects at the target time based on a track prediction model; determining the observation prediction value as the correlation center value; Based on the joint maximum likelihood estimation model, determining a second actual observation value set for each of the target objects from the first actual observation value set according to the association center value, comprising: Determine the relevant wave gate of each of the target objects according to the grey prediction value; determine the candidate observation value set of each of the target objects from the first actual observation value set according to the relevant wave gate; determine the second actual observation value set according to the candidate observation value set and the grey prediction value based on the joint maximum likelihood estimation model, including: calculating the difference between the grey prediction value and each candidate observation value in the candidate observation value set; based on the information entropy calculation model, take the difference as input to obtain the expanded uncertainty of each candidate observation value; based on the joint maximum likelihood estimation model, determine the second actual observation value set according to the expanded uncertainty of each valid observation value; Determine the relevant wave gate of each of the target objects according to the observation prediction value; determine the candidate observation value set of each of the target objects from the first actual observation value set according to the relevant wave gate; determine the second actual observation value set according to the candidate observation value set and the observation prediction value based on the joint maximum likelihood estimation model, including: calculating the difference between the observation prediction value and each candidate observation value in the candidate observation value set; based on the information entropy calculation model, take the difference as input to obtain the expanded uncertainty of each candidate observation value; based on the joint maximum likelihood estimation model, determine the second actual observation value set according to the expanded uncertainty of each valid observation value; The step of determining a candidate observation value set for each target object from the first actual observation value set according to the correlation gate comprises: Determine a set of measurement index prediction values for each target object at the target time, wherein the set of measurement index prediction values includes at least one of an azimuth angle prediction value set, a Doppler frequency prediction value set, an azimuth angle change rate prediction value set, and a Doppler frequency change rate prediction value set; determine a gate constraint corresponding to the relevant gate according to the set of measurement index prediction values and a set of actual measurement index values of each target object at a moment before the target time; determine the candidate observation value set from the first actual observation value set according to the gate constraint; Based on the single target tracking model, the motion state information of each target object at the target time is determined according to the second actual observation value set, and the motion state information is used to determine the track of each target object in the target time period. 2. A passive positioning system, characterized in that the passive positioning system includes a memory, a processor, a program stored in the memory and executable on the processor, and a data bus for realizing connection and communication between the processor and the memory, wherein when the program is executed by the processor, the steps of the multi-target data association tracking method of a fixed single station as described in claim 1 are realized. 3. A storage medium for computer-readable storage, characterized in that the storage medium stores one or more programs, and the one or more programs can be executed by one or more processors to implement the steps of the fixed single-station multi-target data association tracking method as described in claim 1.