Track fusion method based on tracking quality comprehensive evaluation

Abstract

The invention specifically relates to a track fusion method based on tracking quality comprehensive evaluation, comprising: establishing a distributed unmanned aerial vehicle radar target tracking model, the distributed unmanned aerial vehicle radar target tracking model comprising a plurality of radars moving at a constant speed; acquiring a local track of each radar by using a Kalman filtering method; according to a tracking quality evaluation index entropy model, calculating each tracking quality index value of each radar local track; according to each tracking quality index value of each radar local track, performing comprehensive evaluation on the associated track according to a rank sum ratio comprehensive evaluation method; and setting a fusion weight of each radar according to an evaluation result, and carrying out flight path fusion processing based on the fusion weight and the local flight path of each radar to obtain a system flight path of the target. According to the method, the flight path tracking quality is comprehensively evaluated, the fusion weight is adaptively adjusted, the flight path fusion precision is improved, and the tracking quality of the fused flight path is ensured.

Description

Technical Field

[0001] This invention relates to the field of track fusion technology, and specifically to a track fusion method based on comprehensive evaluation of tracking quality. Background Art

[0002] In recent years, with the development of information technologies such as communication networks and artificial intelligence, as well as advanced manufacturing technologies such as integrated circuits and materials science, unmanned aerial vehicle (UAV) systems have gradually developed towards miniaturization, low cost, high flexibility, and clustering. Through multi-radar networking and data fusion processing, multi-level and comprehensive processing of observation data can be achieved. This not only compensates for the inherent shortcomings of low performance of single radars, but also provides observers with a more comprehensive and accurate environmental description.

[0003] Traditional track fusion methods include simple convex combination fusion and Bar-Shalom-Campo fusion. Simple convex combination fusion calculates fusion weights based on the local estimation errors of different radar nodes for the same target. This method is computationally simple and easy to implement, but it struggles to identify the correlations between measurements from different radar nodes, limiting its practical application in engineering. To address the limitations of simple convex combination fusion, the Bar-Shalom-Campo fusion method can better handle track fusion problems under conditions where the mutual covariance of different radar nodes is unknown. However, it involves a large computational load, and its fusion accuracy is significantly affected by the selection of fusion coefficients. Furthermore, both of these track fusion methods calculate fusion weights solely based on tracking accuracy, neglecting the stability of target tracking by different radar nodes.

[0004] It should be noted that the information disclosed in the background section above is only used to enhance the understanding of the background of the present invention, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0005] This invention provides a track fusion method based on comprehensive evaluation of tracking quality, which solves the problem that existing track fusion methods only calculate fusion weights based on tracking accuracy and do not consider the stability of target tracking by different radar nodes. This invention can overcome the defects in the existing technology to a certain extent.

[0006] Other features and advantages of the invention will become apparent from the following detailed description, or may be learned in part by practice of the invention.

[0007] According to a first aspect of the present invention, a track fusion method based on comprehensive tracking quality evaluation is provided, the method comprising: A distributed UAV-borne radar target tracking model is established, which includes multiple radars moving at a constant speed. The local tracks of each radar are obtained using the Kalman filter method; Construct an entropy model for tracking quality evaluation indicators, and calculate the tracking quality index values ​​for each radar local track based on the entropy model; Based on the tracking quality index values ​​of each radar local track, the associated tracks are comprehensively evaluated using the rank-sum ratio comprehensive evaluation method; Based on the evaluation results, a fusion weight is set for each radar. Track fusion processing is performed based on the fusion weight and the local tracks of each radar to obtain the system track of the target.

[0008] In some exemplary embodiments, the method further includes spatiotemporal registration of local tracks of multiple radars prior to track fusion processing, specifically: Spatial registration of local tracks from multiple radars is performed to unify them into the same coordinate system; Time registration is performed on measurement data from multiple radars to unify them to the same time reference.

[0009] In some exemplary embodiments, the tracking quality index values ​​include track update probability, track stability probability, track status stability probability, and track observation stability probability.

[0010] In some exemplary embodiments, the entropy model for tracking quality evaluation indicators specifically includes: Corresponding to the four evaluation indicators, the following four mutually exclusive events are defined. Each event contains two mutually exclusive sub-events. The information entropy of the event is calculated using the self-information of the sub-events contained in each event, thereby evaluating the quality of the track history over a period of time. event { Track update Track not updated}: Track update probability is The probability that the track has not been updated is ;in, This indicates the number of radar scan cycles and also represents the length of the sliding window. Indicates the number of track update cycles within the sliding window; event { Flight track health The probability of a healthy flight path is: (unhealthy flight path) The probability of an unhealthy flight path is ;in, The number of extrapolation periods of the trajectory. This indicates the maximum number of cycles for extrapolation of a track. Beyond If the target corresponding to the track has stopped moving or is out of the radar's line of sight, then the track can be deleted. event { The flight path is estimated to be stable. The probability of state stability is measured by the return visit time of the track. event { The flight path observation is stable. Track observation oscillation: Among them, the radar observation stability probability is

[0011] in, For the measurement residuals of track observation and correlation, The threshold for maximum correlation between observation and measurement. For the new information covariance matrix; The probability of track observation oscillation is

[0012] The information entropy of each event is:

[0013] In the formula, Indicates the flight path i The Middle j Information entropy of an event Indicates the first j The number of child events in an event. Indicates the flight path i No. j The first event k The probability of each sub-event; For flight path i , No. j Entropy of event information Residual for

[0014] For flight path i The higher the residual value of an event, the better the evaluation of that indicator in track quality.

[0015] In some exemplary embodiments, the method of using the return time of the track to measure the probability of state stability specifically includes: Calculate the trajectory using the VanKeuk formula. i Follow-up visit time

[0016] In the formula, To measure covariance, For the track iPredict covariance, The time constant is the maneuvering time constant. For the track i Predictive control accuracy; Probability of track state stability for

[0017] In the formula, This indicates the preset time window, which is the revisit time corresponding to a stable flight path.

[0018] In some exemplary embodiments, the comprehensive evaluation of associated tracks according to the rank-sum ratio comprehensive evaluation method specifically includes: calculate n Information entropy of four indicators of a radar Arrange the data into a matrix, assign ranks in descending order, and take the average rank for values ​​with the same value, thus forming a rank matrix. for:

[0019] Calculate the rank-sum ratio using the following formula:

[0020] Based on rank-sum ratio Conduct a comprehensive evaluation. The higher the value, the better the evaluation effect.

[0021] In some exemplary embodiments, the step of setting fusion weights for each radar based on the evaluation results, and performing track fusion processing based on the fusion weights and the local tracks of each radar, specifically involves: The fusion weights of each radar are determined based on the rank-sum ratio, using the following formula:

[0022] Determined based on the obtained fusion weights n The radar fusion results are calculated using the following formula:

[0023] in, For the first i The radar track, To integrate flight paths.

[0024] According to a second aspect of the present invention, a storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the track fusion method based on comprehensive evaluation of tracking quality as described in the first aspect.

[0025] According to a third aspect of the present invention, a computer program product is provided, on which a computer program is stored, wherein when the computer program is executed by a processor, the track fusion method based on comprehensive evaluation of tracking quality described in the first aspect is implemented.

[0026] According to a fourth aspect of the present invention, an electronic device is provided, comprising: Processor; and Memory for storing the executable instructions of the processor; The processor is configured to implement the track fusion method based on comprehensive tracking quality evaluation described in the first aspect above by executing the executable instructions.

[0027] The track fusion method based on comprehensive tracking quality evaluation provided by the embodiments of the present invention uses multiple evaluation indicators to comprehensively describe the track tracking quality, constructs a comprehensive track quality evaluation model, and assigns fusion weights by evaluating the quality of tracks from different radars. Tracks with higher quality have higher fusion weights, and vice versa. This achieves the effect of improving track fusion quality and robustness, as well as the system track tracking accuracy. It effectively solves the problem that traditional track fusion methods do not provide a comprehensive evaluation of track quality and do not adequately consider the uncertainty differences between track quality from different radar nodes, resulting in the inability to guarantee the quality of the fused track.

[0028] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit the invention. Attached Figure Description

[0029] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the invention and, together with the description, serve to explain the principles of the invention. It is obvious that the drawings described below are merely some embodiments of the invention, and those skilled in the art can obtain other drawings based on these drawings without any inventive effort.

[0030] Figure 1 This is a schematic diagram illustrating the process of a trajectory fusion method based on comprehensive tracking quality evaluation, which is an exemplary embodiment of the present invention. Figure 2 A schematic diagram of the geometric configuration for distributed unmanned aerial vehicle (UAV) radar target tracking, which is an exemplary embodiment of the present invention; Figure 3 This is a block diagram illustrating dot fusion and filtering processing as an exemplary embodiment of the present invention. DETAILED DESCRIPTION

[0031] Exemplary embodiments will now be described more fully with reference to the accompanying drawings. However, these exemplary embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, they are provided so that the invention will be more comprehensive and complete, and will fully convey the concept of the exemplary embodiments to those skilled in the art. The described features, structures, or characteristics may be combined in any suitable manner in one or more embodiments.

[0032] Furthermore, the accompanying drawings are merely illustrative of the invention and are not necessarily drawn to scale. The same reference numerals in the drawings denote the same or similar parts, and therefore repeated descriptions of them will be omitted. Some block diagrams shown in the drawings are functional entities and do not necessarily correspond to physically or logically independent entities. These functional entities can be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor devices and / or microcontroller devices.

[0033] To address the limitations of traditional track fusion methods, this invention comprehensively considers factors affecting track quality and studies a tracking quality comprehensive evaluation method based on rank-sum ratio. This method provides a more comprehensive evaluation of the target tracking quality of different radar nodes, using the evaluation results as the basis for setting track fusion weights. Higher track quality results in higher fusion weights, and vice versa. Through comprehensive analysis of track quality, the risk of track quality degradation after fusion is effectively reduced, thereby improving the accuracy and robustness of track fusion and demonstrating significant application value.

[0034] This example implementation provides a track fusion method based on comprehensive tracking quality evaluation, referencing... Figure 1 As shown, the specific steps may include: Step S11: Construct a distributed UAV-borne radar target tracking model, wherein the distributed UAV-borne radar target tracking model includes multiple radars moving at a constant speed. Step S12: Use the Kalman filter method to obtain the local tracks of each radar. Step S13: Construct a tracking quality evaluation index entropy model, and calculate the tracking quality index values ​​for each radar local track based on the tracking quality evaluation index entropy model; Step S14: Based on the tracking quality index values ​​of each radar local track, the associated tracks are comprehensively evaluated according to the rank-sum ratio comprehensive evaluation method. Step S15: Set the fusion weight for each radar according to the evaluation results, and perform track fusion processing based on the fusion weight and the local tracks of each radar to obtain the system track of the target.

[0035] This example implementation provides a track fusion method based on comprehensive tracking quality evaluation. It utilizes four indicators—track update rate, track health, track state stability, and track observation stability—to comprehensively evaluate radar tracking quality. A rank-sum ratio comprehensive evaluation method is then used to obtain the comprehensive tracking quality evaluation result of the associated tracks. Fusion weights are assigned based on this comprehensive evaluation result. This method provides a comprehensive evaluation of track tracking quality, adaptively adjusts the fusion weights, improves track fusion accuracy, and ensures the tracking quality of the fused tracks.

[0036] The steps in this exemplary embodiment will now be described in more detail with reference to the accompanying drawings and embodiments.

[0037] In step S11, a distributed UAV-borne radar target tracking model is constructed, wherein the distributed UAV-borne radar target tracking model includes multiple radars moving at a constant speed.

[0038] like Figure 2 The diagram shows the geometric configuration of a radar network system consisting of two UAV-borne radars. The two UAVs are positioned at speeds of... and Flying in the same direction as the target on both sides, the two airborne radars detect and track the target using right forward-looking and left forward-looking radars respectively. Using a geocentric coordinate system, and Radar 1 and Radar 2 respectively Coordinate system. Assume both radars are modern radars, meaning they can provide not only target state estimates but also estimation error covariance. Commonly used fusion architectures include distributed and centralized approaches. This invention considers distributed fusion, where the radar's local trajectory estimates are sent to a fusion center for fusion.

[0039] In step S12, the local tracks of each radar are obtained using the Kalman filter method.

[0040] Specifically, each radar uses the Kalman filter method to obtain its own local track and reports the local track to the fusion center, where the fusion center performs correlation processing on the local tracks reported by each radar node.

[0041] For example, the method further includes: Before track fusion processing at the fusion center, spatiotemporal registration is performed on local tracks from multiple radars, specifically as follows: Spatial registration of local tracks from multiple radars is performed to unify them into the same coordinate system; Time registration is performed on measurement data from multiple radars to unify them to the same time reference.

[0042] In step S13, an entropy model for tracking quality evaluation index is constructed, and the tracking quality index values ​​for each radar local track are calculated based on the entropy model for tracking quality evaluation index.

[0043] For example, the tracking quality indicators include track update probability, track stability probability, track status stability probability, and track observation stability probability.

[0044] Specifically, the tracking accuracy and stability of a track are comprehensively described using track update probability, track stability probability, track state stability probability, and track observation stability probability. Corresponding to these four evaluation metrics, four mutually exclusive events are defined, each containing two mutually exclusive sub-events. The information entropy of each event is calculated using the self-information of its sub-events, thereby evaluating the historical quality of the track over a period of time.

[0045] event { Track update Track not updated}: Track update probability is The probability that the track has not been updated is . This indicates the number of radar scan cycles and also represents the length of the sliding window. Indicates the number of track update cycles within the sliding window. event { Flight track health The probability of a healthy flight path is: (unhealthy flight path) The probability of an unhealthy flight path is . The number of extrapolation periods of the trajectory. This indicates the maximum number of cycles for extrapolation of a track. Beyond If the target corresponding to the track has stopped moving or is out of the radar's line of sight, then the track can be deleted.

[0046] event { The flight path is estimated to be stable. The tracking covariance of a track is constantly changing due to factors such as tracking time, extrapolation time, and target state. Therefore, track state stability cannot be directly converted into event probability. However, the probability of state stability can be measured using the track's revisit time. The VanKeuk formula is used to calculate the track state. i Follow-up visit time (1) In the formula, To measure covariance, For the tracki Predict covariance, The time constant is the maneuvering time constant. For the track i The predictive control accuracy. Among them, , , The value can be preset. The calculation formula is: (2) In the formula, This is the weighting adjustment coefficient. This represents the variance of the northward position state of the flight path. This represents the northbound velocity state variance of the track; the state variances of other tracks can be deduced similarly.

[0047] From equation (1), we can see that the track revisit time is calculated from the track state estimation covariance. The more stable the state, the longer the revisit time, and vice versa. Therefore, we can obtain the track state stability probability. for (3) In the formula, This indicates the preset time window, which is the revisit time corresponding to a stable flight path.

[0048] probability of track instability for (4) event { The flight path observation is stable. Track observation oscillations: Radar observation stability probability is (5) In the formula, For the measurement residuals of track observation and correlation, The threshold for maximum correlation between observation and measurement. For the new information covariance matrix.

[0049] The probability of track observation oscillation is (6) Then the information entropy of each event is (7) In the formula, Indicates the flight path i The Middle j Information entropy of an event Indicates the first j The number of child events in an event. Indicates the flight path i No.j The first event k The probability of each sub-event.

[0050] For flight path i , No. j Entropy of event information Residual for (8) For flight path i The higher the residual value of an event, the better the evaluation of that indicator in track quality. By calculating the comprehensive evaluation value of each event, the fusion weight of the track can be determined.

[0051] In step S14, the associated tracks are comprehensively evaluated according to the rank-sum ratio comprehensive evaluation method based on the tracking quality index values ​​of each radar local track.

[0052] The rank-sum ratio comprehensive evaluation method does not require explicit weighting, but is calculated directly through rank. It is subjective, completely objective, and can eliminate the influence of dimensions. It is suitable for scenarios where data allocation is unknown, the differences in importance between indicators are not significant, or subjective weighting is not required.

[0053] set up The capacity drawn from a univariate population is n The sample has the following ascending order statistics: .like Then it is called k yes The rank in a sample is denoted as . For each ,say It is the i-th rank statistic. These are collectively referred to as rank statistics.

[0054] For local tracks from different radar nodes, the steps of the rank-sum ratio comprehensive evaluation method are as follows: (a) Rank: Calculation n Information entropy of four indicators of a radar Arrange the data into a matrix, assign ranks in descending order, and take the average rank for values ​​with the same value, thus forming a rank matrix. for:

[0055] (b) Calculate the rank-sum ratio according to the following formula:

[0056] (c) Based on rank-sum ratio Conduct a comprehensive evaluation. The higher the value, the better the evaluation effect.

[0057] In step S15, the fusion weight of each radar is set according to the evaluation results, and the trajectory fusion processing is performed based on the fusion weight and the local trajectory of each radar to obtain the system trajectory of the target.

[0058] The fusion weights of each radar are determined based on the rank-sum ratio, using the following formula:

[0059] Based on the obtained fusion weights, n radar fusion results are determined using the following formula:

[0060] in, For the first i The radar track, To integrate flight paths.

[0061] It should be noted that the above figures are merely illustrative of the processes included in the method according to exemplary embodiments of the present invention, and are not intended to be limiting. It is readily understood that the processes shown in the above figures do not indicate or limit the temporal order of these processes. Furthermore, it is readily understood that these processes may, for example, be executed synchronously or asynchronously in multiple modules.

[0062] It should be noted that although several modules or units of the device for performing actions have been mentioned in the detailed description above, this division is not mandatory. In fact, according to embodiments of the present invention, the features and functions of two or more modules or units described above can be embodied in one module or unit. Conversely, the features and functions of one module or unit described above can be further divided and embodied by multiple modules or units.

[0063] The units described in the embodiments of the present invention can be implemented in software or hardware, and the described units can also be located in a processor. The names of these units do not necessarily limit the specific unit itself.

[0064] It should be noted that, as another aspect, this application also provides a storage medium, which may be included in an electronic device or may exist independently without being assembled into the electronic device. The storage medium carries one or more programs, which, when executed by an electronic device, cause the electronic device to perform the methods described in the following embodiments.

[0065] In one embodiment, this application provides a computer program product including a computer program that, when executed by a processor, implements the steps in the above-described method embodiments.

[0066] Furthermore, the above figures are merely illustrative of the processes included in the method according to exemplary embodiments of the present invention, and are not intended to be limiting. It is readily understood that the processes shown in the above figures do not indicate or limit the temporal order of these processes. Additionally, it is readily understood that these processes may be executed synchronously or asynchronously, for example, in multiple modules.

[0067] Other embodiments of the invention will readily occur to those skilled in the art upon consideration of the specification and practice of the invention herein. This application is intended to cover any variations, uses, or adaptations of the invention that follow the general principles of the invention and include common knowledge or customary techniques in the art not disclosed herein. The specification and embodiments are to be considered exemplary only, and the true scope and spirit of the invention are indicated by the claims.

[0068] It should be understood that the present invention is not limited to the precise structure described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of the invention is defined only by the appended claims.

Claims

1. A track fusion method based on comprehensive evaluation of tracking quality, characterized in that, The method includes: A distributed UAV-borne radar target tracking model is established, which includes multiple radars moving at a constant speed. The local tracks of each radar are obtained using the Kalman filter method; Construct an entropy model for tracking quality evaluation indicators, and calculate the tracking quality index values ​​for each radar local track based on the entropy model; Based on the tracking quality index values ​​of each radar local track, the associated tracks are comprehensively evaluated using the rank-sum ratio comprehensive evaluation method; Based on the evaluation results, a fusion weight is set for each radar. Track fusion processing is performed based on the fusion weight and the local tracks of each radar to obtain the system track of the target.

2. The track fusion method based on comprehensive tracking quality evaluation according to claim 1, characterized in that, The method further includes performing spatiotemporal registration of local tracks from multiple radars before track fusion processing, specifically: Spatial registration of local tracks from multiple radars is performed to unify them into the same coordinate system; Time registration is performed on measurement data from multiple radars to unify them to the same time reference.

3. The track fusion method based on comprehensive evaluation of tracking quality according to claim 1, characterized in that, The tracking quality indicators include the probability of track update, the probability of track stability, the probability of track status stability, and the probability of track observation stability.

4. The track fusion method based on comprehensive evaluation of tracking quality according to claim 3, characterized in that, The entropy model for tracking quality evaluation indicators is specifically as follows: Corresponding to the four evaluation indicators, the following four mutually exclusive events are defined. Each event contains two mutually exclusive sub-events. The information entropy of the event is calculated using the self-information of the sub-events contained in each event, thereby evaluating the quality of the track history over a period of time. event { Track update Track not updated}: Track update probability is The probability that the track has not been updated is ;in, This indicates the number of radar scan cycles and also represents the length of the sliding window. Indicates the number of track update cycles within the sliding window; event { Flight track health The probability of a healthy flight path is: (unhealthy flight path) The probability of an unhealthy flight path is ;in, The number of extrapolation periods of the trajectory. This indicates the maximum number of cycles for extrapolation of a track. Exceeding If the target corresponding to the track has stopped moving or is out of the radar's line of sight, then the track can be deleted. event { The flight path is estimated to be stable. The probability of state stability is measured by the return visit time of the track. event { The flight path observation is stable. Track observation oscillation: Among them, the probability of stable radar observation is in, For the measurement residuals of track observation and correlation, The threshold for maximum correlation between observation and measurement. For the new information covariance matrix; The probability of track observation oscillation is The information entropy of each event is: In the formula, Indicates the flight path i The Middle j Information entropy of an event Indicates the first j The number of child events in an event. Indicates the flight path i No. j The first event k The probability of each sub-event; For flight path i , No. j Entropy of event information Residual for For flight path i The higher the residual value of an event, the better the evaluation of that indicator in track quality.

5. The track fusion method based on comprehensive evaluation of tracking quality according to claim 4, wherein the use of track revisit time to measure the probability of state stability specifically comprises: Calculate the trajectory using the VanKeuk formula. i Follow-up visit time In the formula, To measure covariance, For the track i Predict covariance, For the maneuver time constant, For the track i Predictive control accuracy; Probability of track state stability for In the formula, This indicates the preset time window, which is the revisit time corresponding to a stable flight path.

6. The track fusion method based on comprehensive evaluation of tracking quality according to claim 4, characterized in that, The comprehensive evaluation of associated tracks using the rank-sum ratio comprehensive evaluation method is as follows: calculate n Information entropy of four indicators of a radar Arrange the data into a matrix, assign ranks in descending order, and take the average rank for values ​​with the same value, thus forming a rank matrix. for: Calculate the rank-sum ratio using the following formula: Based on rank-sum ratio Conduct a comprehensive evaluation. The higher the value, the better the evaluation effect.

7. The method according to claim 6, characterized in that, The process involves setting fusion weights for each radar based on the evaluation results, and performing track fusion processing based on these fusion weights and the local tracks of each radar. The fusion weights of each radar are determined based on the rank-sum ratio, using the following formula: Determined based on the obtained fusion weights n The radar fusion results are calculated using the following formula: in, For the first i The radar track, To integrate flight paths.

8. A storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the track fusion method based on comprehensive evaluation of tracking quality as described in any one of claims 1 to 7.

9. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the track fusion method based on comprehensive evaluation of tracking quality as described in any one of claims 1 to 7.

10. An electronic device, characterized in that, include: processor; as well as Memory for storing the executable instructions of the processor; The processor is configured to execute the track fusion method based on comprehensive tracking quality evaluation as described in any one of claims 1 to 7 by executing the executable instructions.

Images