Aircraft identification method and system based on multi-radar system
By analyzing point data from multiple radar systems and constructing a matching cost matrix, the problem of aircraft trajectory identification in high-density airspace was solved, enabling accurate identification and continuous tracking of aircraft.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-12-29
- Publication Date
- 2026-04-07
AI Technical Summary
Traditional methods based on kinematic characteristics are insufficient to accurately identify the flight paths of individual aircraft in high-density airspace.
By acquiring point data at each moment using a multi-radar system, and using the DBSCAN clustering algorithm and covariance matrix eigenvalue decomposition, the structural similarity and independence of point clusters are analyzed. A matching cost matrix for track points is constructed, and the Hungarian matching algorithm is applied to obtain the aircraft's track.
Accurately identify and track the flight paths of individual aircraft in high-density airspace, especially when aircraft are close together or their flight paths intersect, by analyzing the continuity of motion trends and structural characteristics, to achieve independent identification and continuous tracking of each aircraft.
Smart Images
Figure CN121806004A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of air traffic management technology, and specifically to an aircraft identification method and system based on a multi-radar system. Background Technology
[0002] A multi-radar system refers to the deployment of radar stations in different geographical locations, enabling these stations to simultaneously conduct multi-angle detection of the same area to accurately identify various aircraft within the detection range. In order to monitor and manage aircraft in the airspace safely, it is necessary to accurately grasp the flight status and movement trend of each aircraft under multi-radar observation conditions, thus requiring the detection and identification of aircraft flight trajectories.
[0003] As air traffic continues to grow, the density of aircraft in airspace such as airport terminal areas is gradually increasing. In airspace with high aircraft density, there may be a large number of parallel or intersecting flight paths, which makes different aircraft similar in instantaneous kinematic characteristics. This makes it difficult to accurately identify the flight paths of individual aircraft in airspace with high aircraft density using traditional kinematic characteristics-based methods. Summary of the Invention
[0004] This invention provides an aircraft identification method and system based on a multi-radar system to solve the existing problem that traditional kinematic features are difficult to accurately identify the flight trajectories of individual aircraft in airspace with high aircraft density.
[0005] The aircraft identification method and system based on a multi-radar system of the present invention adopts the following technical solution: One embodiment of the present invention provides an aircraft identification method based on a multi-radar system, the method comprising the following steps: Acquire the point data of all radars at each time step; Based on the spatial distance between the point data of all radars recorded simultaneously, several point clusters are obtained at each time. Based on the distribution structure of point data within different point clusters at each time, the similarity between different point clusters at each time is obtained. Based on the similarity between point clusters at different times, the independence of point clusters at each time is obtained. Based on the independence of point clusters at each time, several track points at each time are obtained. Based on the spatial position of each track point at each time step and its adjacent time steps, the stability of each predecessor vector and each successor vector of each track point at each time step is obtained, and the stability weights of each predecessor vector and each successor vector of each track point at each time step are obtained. Based on the stability of each predecessor vector and each successor vector of each track point at each time step and its stability weights, the matching cost matrix of all track points at each time step is constructed. Based on the matching cost matrix of all track points at each time point, obtain the track of the aircraft corresponding to the track point.
[0006] Preferably, the specific method for obtaining several point clusters at a given time based on the spatial distance between the point data of all radars simultaneously recorded includes: For any given time, all point data collected by each radar at that time are mapped to the same three-dimensional spatial coordinate system to obtain the point data of all radars at that time. Then, the point data of all radars at that time are clustered using the DBSCAN clustering algorithm to obtain several point clusters at that time.
[0007] Preferably, the method for obtaining the similarity between different trace clusters at each time point based on the distribution structure of trace data within different trace clusters at each time point includes: Preset a local time range For any given time, the time preceding the given time... All moments within a second are considered as local moments of the stated moment; for the th moment of the stated moment... The nth trace cluster class, based on the nth time... All trace data in the trace cluster class are used to construct the first trace at the given time. The covariance matrix of the nth point cluster class, for the nth time... The covariance matrix of the nth trace cluster class is used for eigenvalue decomposition to obtain the nth time step at the given time step. The covariance matrix of the n point clusters is determined by several eigenvalues, and the eigenvalues are ordered in descending order of magnitude to be the nth time step. Each eigenvalue of the covariance matrix of a cluster of points is assigned an index label; For the time mentioned above, the first The cluster of points and the first point at the given time The local time of the first A cluster of points; the first point cluster at the given time... The feature values of each of the nth trace clusters, and the nth time at the given time. The local time of the first The absolute value of the difference between the feature values corresponding to the same index label of the nth trace cluster class is used as the nth time. The cluster of points and the first point at the given time The local time of the first The structural difference factors of each of the n point cluster classes, for the nth time... The cluster of points and the first point at the given time The local time of the first The sum of all structural difference factors of each trace cluster class is negatively correlated and normalized. The result of the negative correlation normalization is taken as the first time step. The cluster of points and the first point at the given time The local time of the first Similarity between clusters of dots.
[0008] Preferably, the method for obtaining the degree of independence of the point clusters at each time step based on the similarity between the point clusters at different times includes: For any time, the th The cluster class of points and its first At a local time, the first time of the stated time will be... Among all the point clusters at a given local time, the one corresponding to the time at that time is... The cluster class with the highest similarity among the cluster classes of dots is denoted as the i-th cluster class at the given time. The corresponding point cluster class at the local time, which will be the first point cluster at the time. The cluster of points and the first point at the given time The similarity between the corresponding point clusters at each local time point is greater than that between the time points mentioned above and their corresponding local time points. The ratio of the temporal distances between local moments is used as the first time step of the stated moment. The stability factor of the corresponding point cluster class at a local time; The mean of the stability factor of the corresponding trace cluster class at all local moments at the stated time is normalized, and the normalized result is used as the first value at the stated time. The degree of independence of each point cluster class.
[0009] Preferably, the specific method for obtaining several waypoints at each time step based on the degree of independence of the waypoint clusters at each time step includes: For any time, the th Each point cluster is defined, and a threshold for independence is preset. If the first time at the stated time The degree of independence of each cluster of points is greater than or equal to , the first time of the stated time The centroid of a point cluster is used as a track point; If the time of the above moment is The degree of independence of each cluster of points is less than According to the time of the statement The differences in each dimension of the data from different point clusters within a given point cluster are used as a distance metric. The DBSCAN algorithm is then used to measure the distance at the given time. Clustering the trace data in the trace clusters of the given time, we obtain the first trace at the given time. Several sub-cluster classes of the point cluster class, at the time of the first point cluster class, will be... The centroid of each sub-cluster of a point cluster class is used as a track point; several track points at each time step are obtained.
[0010] Preferably, the method for obtaining the stability of each predecessor vector and each successor vector of each track point at each time point based on the spatial position of each track point at each time point and its adjacent time points includes the following specific methods: For the The moment of the first The first waypoint will be the first The spatial position of each track point at time n, pointing to the nth time. The moment of the first The vector of the spatial position of the _th waypoint is used as the _th The moment of the first Each preceding vector of the track points; will the _th _ ... The moment of the first The spatial location of the first waypoint points, pointing to the first... The vector of the spatial position of each waypoint at time n is used as the first... The moment of the first Each successor vector of each waypoint; For the The moment of the first Given any predecessor vector and any successor vector of the _ waypoint, the _ ... The moment of the first The cosine similarity between the predecessor vector and the successor vector of the first waypoint is greater than that of the previous waypoint. The moment of the first The absolute value of the difference in magnitude between the preceding vector and the succeeding vector of the first waypoint is used as the first... The moment of the first The stability of the predecessor vector and the successor vector of each waypoint.
[0011] Preferably, the method for obtaining the stability weights of each predecessor vector and each successor vector at each time point includes: In the formula, Indicates the first The moment of the first The stability weights of the predecessor vector and the successor vector for each waypoint; Indicates the first The moment of the first Number of predecessor vectors for each waypoint; Indicates the first The moment of the first The number of successor vectors for each waypoint; Indicates the first The moment of the first The stability of the predecessor vector and the successor vector of each waypoint; Indicates the first The moment of the first The first waypoint The first precursor vector and the second The stability of successor vectors.
[0012] Preferably, the specific method for constructing the matching cost matrix of all waypoints at each time step is as follows: For any predecessor vector and any successor vector of any track point at any time, the product of the stability of the predecessor vector and the successor vector of the track point at the time and its stability weight is negatively correlated and normalized. The result of the negative correlation normalization is used as the matching cost of the predecessor vector and the successor vector of the track point at the time. Obtain the matching cost of each predecessor vector and each successor vector of the track point at the specified time. Based on the matching cost of each predecessor vector and each successor vector of the track point at the specified time, construct the matching cost matrix of each predecessor vector and each successor vector of the track point at the specified time, denoted as the matching cost matrix of the track point at the specified time.
[0013] Preferably, the specific method for obtaining the trajectory of the aircraft corresponding to the trajectory point based on the matching cost matrix of all trajectory points at each time point includes: For any given time, the matching cost matrix of all track points at that time is input into the Hungarian matching algorithm to obtain the track points corresponding to each track point at that time in the adjacent time. The corresponding track points in the adjacent time are connected to obtain the track of the aircraft corresponding to the track point.
[0014] Another embodiment of the present invention provides an aircraft identification system based on a multi-radar system, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps of any one of the above-described aircraft identification methods based on a multi-radar system.
[0015] The beneficial effects of the technical solution of the present invention are: acquiring the point data of all radars at each moment; since multiple aircraft may approach each other when the aircraft density is high, the point data corresponding to multiple aircraft may be grouped into the same point cluster class. However, since the structure of the aircraft is stable, that is, the structure of the point data corresponding to the aircraft will not change with the flight of the aircraft, while the point cluster class corresponding to multiple aircraft will change due to the flight of multiple aircraft, the relative positions between multiple aircraft will change; therefore, based on the structural changes of the point cluster class at consecutive moments, it is possible to distinguish whether the point cluster class corresponds to a single aircraft or multiple aircraft, thereby obtaining the track points corresponding to each aircraft at each moment for subsequent acquisition of the accurate flight trajectory of each aircraft.
[0016] Because aircraft motion is continuous during flight, meaning that their spatial position, velocity direction, and motion trend do not change abruptly between adjacent moments, trackpoints belonging to the same aircraft should exhibit smooth position changes and consistent velocity directions across consecutive moments. This allows us to obtain the stability and stability weights of each preceding and succeeding vector of each trackpoint at each moment based on their spatial positions relative to those at adjacent moments. This enables the construction of a matching cost matrix for all trackpoints at each moment. Based on this matching cost matrix, the aircraft's track for each trackpoint at each moment is obtained. This application achieves accurate identification of multiple aircraft by accurately reconstructing their flight trajectories. Especially in high-density airspace, even when multiple aircraft are spatially close, have intersecting or parallel tracks, the continuity of their motion trends and the stability of their structural characteristics can be analyzed to effectively distinguish different aircraft, thereby achieving independent identification and continuous tracking of each aircraft. Attached Figure Description
[0017] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0018] Figure 1 This is a flowchart illustrating the steps of an aircraft identification method based on a multi-radar system according to the present invention. Detailed Implementation
[0019] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of an aircraft identification method and system based on a multi-radar system proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.
[0020] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.
[0021] The following description, in conjunction with the accompanying drawings, details a specific scheme for an aircraft identification method and system based on a multi-radar system provided by the present invention.
[0022] Please see Figure 1 The diagram illustrates a flowchart of an aircraft identification method based on a multi-radar system according to an embodiment of the present invention. The method includes the following steps: Step S001: Obtain the point data of all radars at each time step.
[0023] It should be noted that this embodiment is an aircraft identification method based on a multi-radar system. Specifically, it obtains the flight trajectory of each aircraft by analyzing the point data collected by each radar in the multi-radar system at each time moment, so as to carry out safety monitoring and precise management of aircraft in the airspace. Therefore, it is necessary to first obtain the point data collected by each radar in the multi-radar system at each time moment.
[0024] Specifically, the point data of each radar at each moment is obtained through a multi-radar system. In this embodiment, a moment is described as 0.1 seconds.
[0025] It should be noted that a multi-radar system refers to the deployment of radar stations in different geographical locations, enabling these stations to simultaneously conduct multi-angle detection of the same area. Each point data contains information in several dimensions, including but not limited to: spatial location information, RCS information (radar cross-section information), etc.
[0026] Step S002: Based on the spatial distance between the point data of all radars recorded simultaneously, obtain several point clusters at each time; based on the distribution structure of point data within different point clusters at each time, obtain the similarity between different point clusters at each time; based on the similarity between point clusters at different times, obtain the degree of independence of point clusters at each time; based on the degree of independence of point clusters at each time, obtain several track points at each time.
[0027] It should be noted that in a multi-radar system, the beam direction, pitch angle, frequency characteristics, and echo sensitivity of different radars may vary. Furthermore, the complex structure of an aircraft (e.g., multiple strong reflection centers on the fuselage, wings, and tail) can cause it to be identified as multiple aircraft in the system. To accurately identify the flight paths of each aircraft, it is first necessary to analyze the point data collected by each radar within the multi-radar system at each moment. Although an aircraft may exhibit multiple point data points in a multi-radar system, these point data points tend to have a tightly clustered distribution in space, which can be used to obtain several point clusters. However, in scenarios with dense aircraft, the point data points in each obtained point cluster may correspond to the point data points of multiple aircraft, thus requiring further analysis of the point clusters.
[0028] It should be further explained that, due to the structural stability of aircraft, that is, the structure of the trace data corresponding to an aircraft does not change with the flight of the aircraft, when the trace data in a trace cluster class corresponds to the trace data of a single aircraft, the structure of the trace cluster class remains stable at consecutive time points. However, when the trace data in a trace cluster class corresponds to the trace data of multiple aircraft, the relative positions between the multiple aircraft will change as they fly, resulting in significant differences in the structure of the trace cluster class at consecutive time points. Therefore, this can be used to distinguish between trace cluster classes corresponding to a single aircraft and trace cluster classes corresponding to multiple aircraft, thereby obtaining the track points corresponding to each aircraft at each time point, which can be used to subsequently obtain the accurate flight trajectories of each aircraft.
[0029] Preferably, in a specific embodiment of the present invention, for any given time, all point data collected by each radar at that time are mapped to the same three-dimensional spatial coordinate system to obtain point data of all radars at that time, and the point data of all radars at that time are clustered by the DBSCAN clustering algorithm to obtain several point clusters at that time. Since the DBSCAN clustering algorithm is a well-known prior art, it will not be described in detail in this embodiment.
[0030] It should be noted that, due to the high density of aircraft, multiple aircraft may approach each other, leading to the grouping of track data corresponding to multiple aircraft into the same track cluster. However, since the structure of aircraft is stable, meaning the structure of the track data corresponding to an aircraft does not change with the flight of the aircraft, the relative positions of the track clusters corresponding to multiple aircraft will change due to the flight of multiple aircraft. Therefore, the structural changes of the track clusters at consecutive time points can be used to distinguish between track clusters corresponding to a single aircraft and track clusters corresponding to multiple aircraft, thereby obtaining the track points corresponding to each aircraft at each time point.
[0031] Preferably, in a specific embodiment of the present invention, a local time range is preset. The The specific value can be set according to the actual situation. This embodiment does not make a hard requirement. In this embodiment, it is used as... To illustrate further, for any given time, the time preceding... All moments within a second are considered local moments of the stated moment (if the time interval preceding the stated moment is less than 1 second). Then all times before the stated time are considered as local times of the stated time); for the first time... The nth trace cluster class, based on the nth time... All trace data in the trace cluster class are used to construct the first trace at the given time. The covariance matrix of the nth point cluster class, for the nth time... The covariance matrix of the nth trace cluster class is used for eigenvalue decomposition to obtain the nth time step at the given time step. The covariance matrix of the n point clusters is determined by several eigenvalues, and the eigenvalues are ordered in descending order of magnitude to be the nth time step. Each eigenvalue of the covariance matrix of the trace cluster class is assigned an index label. Since the process of constructing the covariance matrix and performing eigenvalue decomposition on the covariance matrix is a well-known prior art, it will not be described in detail in this embodiment. Furthermore, for the first time mentioned above... The cluster of points and the first point at the given time The local time of the first A cluster of points; the first point cluster at the given time... The feature values of each of the nth trace clusters, and the nth time at the given time. The local time of the first The absolute value of the difference between the feature values corresponding to the same index label of the n point cluster classes (feature values with the same index label) is used as the nth time. The cluster of points and the first point at the given time The local time of the first The structural difference factors of each of the n point cluster classes, for the nth time... The cluster of points and the first point at the given time The local time of the first The sum of all structural difference factors for each cluster of points is negatively correlated and normalized (using [method name]). The function is normalized for negative correlation. The function represents an exponential function with the natural constant as the base. As input to the model, the implementer can set a negative correlation normalization function according to the actual situation, and use the result obtained by negative correlation normalization as the first time step. The cluster of points and the first point at the given time The local time of the first Similarity between clusters of dots.
[0032] As an example, obtain the first time of the stated moment. The cluster of points and the first point at the given time The local time of the first The specific formula for calculating the similarity between clusters of points is as follows: In the formula, Represents the first time. The cluster of points and the first point at the given time The local time of the first Similarity between clusters of points; This represents the number of eigenvalues of the covariance matrix (the number of eigenvalues of the covariance matrix in three-dimensional space is fixed at 3). Represents the first time. The covariance matrix of the nth point cluster class is... One eigenvalue; Represents the first time. At the local time... The covariance matrix of the nth point cluster class is the first... One eigenvalue; This represents the function that takes the absolute value. This represents an exponential function with the natural constant as its base.
[0033] It should be noted that the covariance matrix of a point cluster class can represent the structural features of the point data within that cluster. Therefore, the structural similarity between different point cluster classes can be assessed by observing the differences in their covariance matrices at corresponding eigenvalues. Furthermore, the lower the similarity between point cluster classes at consecutive time steps, the greater the structural change within those clusters. Therefore, the similarity can be assessed by iterating through the first... The cluster of points and the first point at the given time The similarity between each cluster of points at each local time point is quantified to the first time point. The structural changes of each trace cluster class are used to distinguish the first time point. Each trace cluster class corresponds to a trace cluster class for a single aircraft or a trace cluster class for multiple aircraft.
[0034] Preferably, in a specific embodiment of the present invention, for any given time, the first... The cluster class of points and its first At a local time, the first time of the stated time will be... Among all the point clusters at a given local time, the one corresponding to the time at that time is... The cluster class with the highest similarity among the cluster classes of dots is denoted as the i-th cluster class at the given time. The corresponding point cluster class at the local time, which will be the first point cluster at the time. The cluster of points and the first point at the given time The similarity between the corresponding point clusters at each local time point is greater than that between the time points mentioned above and their corresponding local time points. The ratio of the temporal distances between local moments is used as the first time step of the stated moment. The stability factor of the corresponding point cluster class at a local time; Furthermore, the mean of the stability factor of the corresponding trace cluster class at all local moments of the stated time is normalized (using the sigmoid function), and the normalized result is used as the first value at the stated time. The degree of independence of each point cluster class.
[0035] As an example, obtain the first time of the stated moment. The specific formula for calculating the independence of a cluster of points is as follows: In the formula, Represents the first time. The degree of independence of each point trace cluster class; This indicates the number of local time points at the given moment; Represents the first time. The cluster of points and the first point at the given time The similarity between corresponding point clusters at a local time; Indicates the time and its first The temporal distance between local moments; This represents the sigmoid function, which is used for normalization in this embodiment.
[0036] It should be noted that, since the structure of the point cluster class remains stable at consecutive time steps when the point data in the point cluster class corresponds to the point data of an aircraft, the structure of the point cluster class remains stable at consecutive time steps. The greater the similarity between the nth trace cluster class and the corresponding trace cluster class at each local time point at the given time, the stronger the similarity between the nth trace cluster class at the given time point and the corresponding trace cluster class at each local time point at the given time point. The structure of a cluster of points remains stable at consecutive time points, that is, at the th time point... The more likely the trace data within a trace cluster is to be trace data of a single aircraft, the more likely it is to be trace data of a single aircraft. At the same time, because the relative position of an aircraft to the radar changes continuously during flight, the structure of the trace cluster corresponding to the aircraft observed by the radar will be slightly different. In order to avoid misjudgment due to structural differences caused by observation, the temporal distance between time points is used as a negative correlation weight when calculating the independence of the trace cluster. This is to accurately obtain the independence of the trace cluster to distinguish the trace cluster of a single aircraft from the trace cluster of multiple aircraft, thereby obtaining all track points at each time point.
[0037] Preferably, in a specific embodiment of the present invention, for any given time, the first... Each point cluster is defined, and a threshold for independence is preset. The The specific value can be set according to the actual situation. This embodiment does not make a hard requirement. In this embodiment, it is used as... For example, if the time in question is... The degree of independence of each cluster of points is greater than or equal to Then the first time at the stated time will be... The centroid of a point cluster is used as a track point; If the time of the above moment is The degree of independence of each cluster of points is less than Then according to the time of the stated time The differences in each dimension of the data from different point clusters within a given point cluster are used as a distance metric. The DBSCAN algorithm is then used to measure the distance at the given time. Clustering the trace data in the trace clusters of the given time, we obtain the first trace at the given time. Several sub-cluster classes of the point cluster class, at the time of the first point cluster class, will be... The centroid of each sub-cluster of a point cluster class is used as a track point; several track points at each time step are obtained.
[0038] It should be noted that each track point at each time moment corresponds to each aircraft at that time moment. When the independence of a track cluster at a certain time moment is low, it means that the structure of the cluster changes significantly over consecutive time moments. This indicates that the track data within the cluster no longer stably corresponds to the same aircraft, but may contain track data from multiple aircraft. Therefore, this embodiment measures the differences in multidimensional features of each track data within the track cluster to divide the track data with obvious structural separation trends into several sub-clusters. Since the tracks belonging to the same aircraft have high consistency and compactness in the feature space, the track data in the track cluster with large structural changes over consecutive time moments are further clustered to obtain each track point at all time moments.
[0039] At this point, we have obtained several waypoints for all times.
[0040] Step S003: Based on the spatial position of each track point at each time step and its adjacent time steps, obtain the stability of each predecessor vector and each successor vector of each track point at each time step, and obtain the stability weights of each predecessor vector and each successor vector of each track point at each time step. Based on the stability of each predecessor vector and each successor vector of each track point at each time step and its stability weights, construct the matching cost matrix of all track points at each time step.
[0041] It should be noted that, since the motion of an aircraft during flight is continuous, meaning that the spatial position, velocity direction, and motion trend of an aircraft do not change abruptly between adjacent moments, the track points of the same aircraft should exhibit smooth position changes and consistent velocity directions over consecutive moments. Conversely, if the track points of different aircraft are spatially adjacent but their motion trends are significantly different, then this can be used to solve for the optimal correspondence of each track point in the time series by constructing a matching cost matrix between track points at adjacent moments. This allows each track point at any given moment to find a matching point in subsequent moments, thereby connecting discrete time-based track points into continuous track lines and accurately identifying the flight trajectories of various aircraft in high-density airspace.
[0042] Preferably, in a specific embodiment of the present invention, for the first The moment of the first The first waypoint will be the first The spatial position of each track point at time n, pointing to the nth time. The moment of the first The vector of the spatial position of the _th waypoint is used as the _th The moment of the first Each preceding vector of the track points; will the _th _ ... The moment of the first The spatial location of the first waypoint points, pointing to the first... The vector of the spatial position of each waypoint at time n is used as the first... The moment of the first Each successor vector of each waypoint; Furthermore, regarding the first The moment of the first Given any predecessor vector and any successor vector of the _ waypoint, the _ ... The moment of the first The cosine similarity between the predecessor vector and the successor vector of the first waypoint is greater than that of the previous waypoint. The moment of the first The absolute value of the difference in magnitude between the preceding vector and the succeeding vector of the first waypoint is used as the first... The moment of the first The stability of the predecessor vector and the successor vector of each waypoint.
[0043] It should be noted that the first The moment of the first The predecessor vector and the successor vector of the first waypoint are the hypothetical flight paths of the aircraft corresponding to that waypoint, while the second waypoint... The moment of the first The greater the cosine similarity between the preceding and succeeding vectors of each track point, the more stable the assumed track's direction of motion; The moment of the first The smaller the absolute value of the difference in magnitude between the preceding vector and the succeeding vector at each waypoint, the smaller the change in velocity of the hypothetical waypoint. Therefore, the... The moment of the first The greater the stability value of the predecessor vector and the successor vector of a waypoint, the more likely the hypothetical waypoint is to be the actual waypoint of the aircraft corresponding to that waypoint.
[0044] Furthermore, regarding the first The moment of the first Given any predecessor vector and any successor vector of a given track point, according to the... The moment of the first The stability of the predecessor vector and the successor vector of the first waypoint, combined with the stability of the first waypoint, The moment of the first The stability of other predecessor vectors and other successor vectors of the _th waypoint is obtained. The moment of the first The stability weights of the predecessor vector and the successor vector for each waypoint are calculated using the following formula: In the formula, Indicates the first The moment of the first The stability weights of the predecessor vector and the successor vector for each waypoint; Indicates the first The moment of the first Number of predecessor vectors for each waypoint; Indicates the first The moment of the first The number of successor vectors for each waypoint; Indicates the first The moment of the first The stability of the predecessor vector and the successor vector of each waypoint; Indicates the first The moment of the first The first waypoint The first precursor vector and the second The stability of successor vectors.
[0045] It should be noted that when a certain predecessor and successor vector not only has high stability with the track point itself, but also when other track points are matched with this set of vectors, the stability weight of the predecessor and successor vectors and the track point is amplified, so that it occupies a dominant position in the overall matching judgment. In this way, the stability weight reflects the uniqueness and representativeness of each track point under the global matching constraint, thereby ensuring that the final selected hypothetical track is more in line with the actual motion law of the aircraft.
[0046] Preferably, in a specific embodiment of the present invention, for any predecessor vector and any successor vector of any waypoint at any time, the product of the stability of the predecessor vector and the successor vector of the waypoint at that time with their stability weights is negatively correlated and normalized (using...). The function is normalized for negative correlation. The function represents an exponential function with the natural constant as the base. As input to the model, the implementer can set a negative correlation normalization function according to the actual situation, and use the result of the negative correlation normalization as the matching cost of the predecessor vector and the successor vector of the track point at the time. Similarly, the matching cost of each predecessor vector and each successor vector of the track point at the given time is obtained. Based on the matching cost of each predecessor vector and each successor vector of the track point at the given time, a matching cost matrix of each predecessor vector and each successor vector of the track point at the given time is constructed, denoted as the matching cost matrix of the track point at the given time. in, The cost of matching the first predecessor vector and the first successor vector of the track point at the given time is represented. The matching cost between the first predecessor vector and the second successor vector of the track point at the given time is represented. The matching cost between the second predecessor vector and the first successor vector of the track point at the given time; The matching cost between the second predecessor vector and the second successor vector of the track point at the stated time; The first point of the track at the given time. The matching cost of a predecessor vector with the first successor vector; The first predecessor vector of the track point at the stated time and the second... The matching cost of successor vectors; The first point of the track at the given time. The first precursor vector and the second The matching cost of successor vectors.
[0047] It should be noted that the matching cost matrix of all track points at each time step is used to describe the matching probability between track points at adjacent time steps, and is the core quantization structure for continuous track association; taking the first... The moment and the The matching cost matrix is constructed from the track points at each time point. Each element in the matrix represents the degree of difference between track points at adjacent time points in terms of spatial location, velocity direction, and stability weight. The smaller the cost value, the closer the two points are in terms of physical attributes and the more likely they belong to the same aircraft. By analyzing the matching cost matrix of all track points at each time point, the track of the aircraft corresponding to all track points at each time point can be obtained.
[0048] Thus, the matching cost matrix for all track points at each time step is obtained.
[0049] Step S004: Obtain the trajectory of the aircraft corresponding to the trajectory point based on the matching cost matrix of all trajectory points at each time point.
[0050] It should be noted that after obtaining the matching cost matrix of all track points at each time step S003, the track of the aircraft corresponding to each track point can be obtained based on the matching cost matrix of all track points at each time step, which can better monitor and manage the safety of aircraft in the airspace.
[0051] Preferably, in a specific embodiment of the present invention, for any given time, the matching cost matrix of all track points at that time is input into the Hungarian matching algorithm to obtain the track points corresponding to each track point at that time in the adjacent time. Since the Hungarian matching algorithm is a well-known prior art, it will not be described in detail in this embodiment. The corresponding track points at the adjacent time are connected to obtain the track of the aircraft corresponding to the track point at that time.
[0052] Furthermore, after acquiring the continuous flight paths of each aircraft, the system monitors flight path consistency by comparing flight path data with flight plans in real time. Based on the flight path prediction model, it detects and warns of conflicts between aircraft. At the same time, it statistically analyzes operational indicators such as aircraft density and flight path intersection complexity in the airspace to assess the airspace situation. It also analyzes flight efficiency by combining the deviation between the actual flight path and the optimal path. Finally, it achieves safety monitoring by identifying abnormal patterns in the flight path, thereby constructing a complete airspace safety monitoring and precise management system.
[0053] It should be noted that this embodiment achieves accurate identification of multiple aircraft by accurately reconstructing the flight trajectory of each aircraft; especially in high-density airspace, even if multiple aircraft are close in space, have intersecting or parallel flight paths, the different aircraft can still be effectively distinguished by analyzing the continuity of their motion trends and the stability of their structural characteristics, thereby achieving independent identification and continuous tracking of each aircraft.
[0054] Another embodiment of the present invention provides an aircraft identification system based on a multi-radar system, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements an aircraft identification method based on a multi-radar system in steps S001 to S004.
[0055] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the principles of the present invention should be included within the protection scope of the present invention.
Claims
1. An aircraft identification method based on a multi-radar system, characterized in that, The method includes the following steps: Acquire the point data of all radars at each time step; Based on the spatial distance between the point data of all radars recorded simultaneously, several point clusters are obtained at each time. Based on the distribution structure of point data within different point clusters at each time, the similarity between different point clusters at each time is obtained. Based on the similarity between point clusters at different times, the independence of point clusters at each time is obtained. Based on the independence of point clusters at each time, several track points at each time are obtained. Based on the spatial position of each track point at each time step and its adjacent time steps, the stability of each predecessor vector and each successor vector of each track point at each time step is obtained, and the stability weights of each predecessor vector and each successor vector of each track point at each time step are obtained. Based on the stability of each predecessor vector and each successor vector of each track point at each time step and its stability weights, the matching cost matrix of all track points at each time step is constructed. Based on the matching cost matrix of all track points at each time point, obtain the track of the aircraft corresponding to the track point.
2. The aircraft identification method based on a multi-radar system according to claim 1, characterized in that, The method for obtaining several point clusters at a given time based on the spatial distance between the point data of all radars captured simultaneously includes: For any given time, all point data collected by each radar at that time are mapped to the same three-dimensional spatial coordinate system to obtain the point data of all radars at that time. Then, the point data of all radars at that time are clustered using the DBSCAN clustering algorithm to obtain several point clusters at that time.
3. The aircraft identification method based on a multi-radar system according to claim 1, characterized in that, The method for obtaining the similarity between different trace clusters at each time based on the distribution structure of trace data within different trace clusters at each time step includes the following specific methods: Preset a local time range For any given time, the time preceding the given time... All moments within a second are considered as local moments of the stated moment; for the th moment of the stated moment... The nth trace cluster class, based on the nth time... All trace data in the trace cluster class are used to construct the first trace at the given time. The covariance matrix of the nth point cluster class, for the nth time... The covariance matrix of the nth trace cluster class is used for eigenvalue decomposition to obtain the nth time step at the given time step. The covariance matrix of the n point clusters is determined by several eigenvalues, and the eigenvalues are ordered in descending order of magnitude to be the nth time step. Each eigenvalue of the covariance matrix of a cluster of points is assigned an index label; For the time mentioned above, the first The cluster of points and the first point at the given time The local time of the first A cluster of points; the first point cluster at the stated time... The feature values of each of the nth trace clusters, and the nth time at the given time. The local time of the first The absolute value of the difference between the feature values corresponding to the same index label of the nth trace cluster class is used as the nth time. The cluster of points and the first point at the given time The local time of the first The structural difference factors of each of the n point cluster classes, for the nth time... The cluster of points and the first point at the given time The local time of the first The sum of all structural difference factors of each trace cluster class is negatively correlated and normalized. The result of the negative correlation normalization is taken as the first time step. The cluster of points and the first point at the given time The local time of the first Similarity between clusters of dots.
4. The aircraft identification method based on a multi-radar system according to claim 1, characterized in that, The method for obtaining the degree of independence of point clusters at different times based on the similarity between each point cluster at different times includes the following specific methods: For any time, the th The cluster class of points and its first At a local time, the first time of the said time Among all the point clusters at a given local time, the one corresponding to the time at that time is... The cluster class with the highest similarity among the cluster classes of dots is denoted as the i-th cluster class at the given time. The corresponding point cluster class at the local time, which will be the first point cluster at the time. The cluster of points and the first point at the given time The similarity between the corresponding point clusters at each local time point is greater than that between the time points mentioned above and their corresponding local time points. The ratio of the temporal distances between local moments is used as the first time step of the stated moment. The stability factor of the corresponding point cluster class at a local time; The mean of the stability factor of the corresponding trace cluster class at all local moments at the stated time is normalized, and the normalized result is used as the first value at the stated time. The degree of independence of each point cluster class.
5. The aircraft identification method based on a multi-radar system according to claim 1, characterized in that, The specific method for obtaining several waypoints at each time step based on the degree of independence of the waypoint clusters at each time step includes: For any time, the th Each point cluster is defined, and a threshold for independence is preset. If the first time at the stated time The degree of independence of each cluster of points is greater than or equal to , the first time of the stated time The centroid of a point cluster is used as a track point; If the time of the above moment is The degree of independence of each cluster of points is less than According to the time of the statement The differences in each dimension of the data from different point clusters within a given point cluster are used as a distance metric. The DBSCAN algorithm is then used to measure the distance at the given time. Clustering the trace data in the trace clusters of the given time, we obtain the first trace at the given time. Several sub-cluster classes of the point cluster class, at the time of the first point cluster class, will be... The centroid of each sub-cluster of a point cluster class is used as a track point; several track points at each time step are obtained.
6. The aircraft identification method based on a multi-radar system according to claim 1, characterized in that, The method for obtaining the stability of each predecessor vector and each successor vector of each track point at each time step based on the spatial position of each track point at each time step and its adjacent time steps includes the following specific methods: For the The moment of the first The first waypoint will be the first The spatial position of each track point at time n, pointing to the nth time. The moment of the first The vector of the spatial position of the _th waypoint is used as the _th The moment of the first Each preceding vector of the track points; will the _th _ ... The moment of the first The spatial location of the first waypoint points, pointing to the first... The vector of the spatial position of each waypoint at time n is used as the first... The moment of the first Each successor vector of each waypoint; For the The moment of the first Given any predecessor vector and any successor vector of the _ waypoint, the _ ... The moment of the first The cosine similarity between the predecessor vector and the successor vector of the first waypoint is greater than that of the previous waypoint. The moment of the first The absolute value of the difference in magnitude between the preceding vector and the succeeding vector of the first waypoint is used as the first... The moment of the first The stability of the predecessor vector and the successor vector of each waypoint.
7. The aircraft identification method based on a multi-radar system according to claim 1, characterized in that, The specific method for obtaining the stability weights of each predecessor vector and each successor vector at each time point includes: In the formula, Indicates the first The moment of the first The stability weights of the predecessor vector and the successor vector for each waypoint; Indicates the first The moment of the first Number of predecessor vectors for each waypoint; Indicates the first The moment of the first The number of successor vectors for each waypoint; Indicates the first The moment of the first The stability of the predecessor vector and the successor vector of each waypoint; Indicates the first The moment of the first The first waypoint The first precursor vector and the second The stability of successor vectors.
8. The aircraft identification method based on a multi-radar system according to claim 1, characterized in that, The specific method for constructing the matching cost matrix of all track points at each time step is as follows: For any predecessor vector and any successor vector of any track point at any time, the product of the stability of the predecessor vector and the successor vector of the track point at the time and its stability weight is negatively correlated and normalized. The result of the negative correlation normalization is used as the matching cost of the predecessor vector and the successor vector of the track point at the time. Obtain the matching cost of each predecessor vector and each successor vector of the track point at the specified time. Based on the matching cost of each predecessor vector and each successor vector of the track point at the specified time, construct the matching cost matrix of each predecessor vector and each successor vector of the track point at the specified time, denoted as the matching cost matrix of the track point at the specified time.
9. The aircraft identification method based on a multi-radar system according to claim 1, characterized in that, The specific method for obtaining the flight path of the aircraft corresponding to the flight path point based on the matching cost matrix of all flight path points at each time point includes: For any given time, the matching cost matrix of all track points at that time is input into the Hungarian matching algorithm to obtain the track points corresponding to each track point at that time in the adjacent time. The corresponding track points in the adjacent time are connected to obtain the track of the aircraft corresponding to the track point.
10. An aircraft identification system based on a multi-radar system, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the computer program is executed by the processor, it implements the steps of an aircraft identification method based on a multi-radar system as described in any one of claims 1-9.