An aircraft flight mode recognition method and system based on an air surveillance system

Abstract

The application discloses an aircraft flight mode recognition method and system based on an air monitoring system, realizes calculation of an aircraft flight route by point flight correlation of fused monitoring data, and realizes estimation of a target state by state estimation of an aircraft track generated by point flight correlation. The heading change rate is obtained by heading calculation of the target state, and the target aircraft flight state is obtained by comprehensively considering the target state and the heading change rate, so that the existing model lacks recognition and processing of aircraft heading and turn rate related modes; meanwhile, the flight mode is recognized by the point flight correlation and state estimation method, so that the method does not need to pre-train sample data or pre-set a flight model, can respond to sudden change of the flight mode of the aircraft in real time, thereby improving real-time performance, and solving the deficiencies of deep learning and neural network algorithms in time complexity and real-time performance.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of air traffic integrated control, in particular to a method and system for identifying flight mode of aircraft based on air surveillance system. BACKGROUND

[0002] The primary radar of air traffic control is the main monitoring means of military and civil aviation control. Its basic principle is to use radio frequency electromagnetic waves to irradiate the aircraft and receive the echo. By analyzing the echo signal, the distance, direction, radial velocity, echo amplitude and other information of the target are obtained. The secondary radar monitoring system of air traffic control is a radar system that locates aircraft equipped with airborne transponders through the inquiry of ground inquiry machines and the response of airborne transponders. According to the response signal of the secondary radar airborne transponder added to the aircraft, the target flight information and its attributes are obtained to provide the flight state monitoring of the aircraft for the ground control personnel. The broadcast automatic dependent surveillance generates accurate aircraft self-positioning information by the airborne satellite-based navigation and positioning system, periodically broadcasts automatic monitoring information through a specific data link and format, and receives and processes the information by a specific ground station device or other airborne equipment, thereby obtaining the accurate position and state information of the aircraft.

[0003] Generally, the primary radar of air traffic control, the secondary radar of air traffic control and the broadcast automatic dependent surveillance together constitute the common air traffic integrated monitoring system. The air traffic integrated monitoring system mainly monitors the military and civil aircraft, and the monitored state information is used as the reference basis for the ground command personnel and the on-board operator to command and control. The flight mode recognition and analysis of the aircraft are important for the control personnel to analyze the flight intention of the aircraft and to verify the flight state of the aircraft.

[0004] The current flight mode recognition technology has certain deficiencies, mainly in the following aspects:

[0005] 1. The current flight mode recognition research classification model is not perfect and comprehensive, and the model lacks recognition and processing of the flight direction and turn rate related mode of the aircraft;

[0006] 2. The current flight model mainly reflects in the flight process of fixed-wing manned aircraft. However, the flight state of unmanned aerial vehicle is quite different from that of manned aircraft. The maneuverability of unmanned aerial vehicle is more variable and the change rate is faster. In this mode, the aircraft needs to actively provide flight data and flight mode, but the aircraft cannot provide relevant data;

[0007] 3. Some studies use deep learning and neural network algorithms to identify aircraft flight patterns. However, this method requires pre-trained sample data and a preset flight model. However, during flight, aircraft may suddenly change their flight mode, and the preset flight model cannot cope with such sudden situations. At the same time, the time complexity of deep learning and neural network algorithms is high and the real-time performance is insufficient. Summary of the Invention

[0008] Based on the problems mentioned above, the purpose of this invention is to provide a method and system for aircraft flight pattern recognition based on an airborne surveillance system. This solves the problems that current models for flight pattern recognition research and classification are incomplete and inadequate, lack recognition and processing of aircraft heading and turn rate related patterns, and that the preset flight models cannot cope with various sudden changes in flight mode during flight.

[0009] This invention is achieved through the following technical solution:

[0010] The first aspect of this invention provides a method for aircraft flight pattern recognition based on an airborne surveillance system, comprising the following steps:

[0011] Acquire surveillance data of the monitored airspace and preprocess the surveillance data;

[0012] The preprocessed monitoring data is fused to generate fused monitoring data.

[0013] The fused surveillance data is correlated with point navigation data, and the target state is estimated based on the results of the point navigation correlation.

[0014] The heading is estimated based on the target state to obtain the heading change rate, and the aircraft mode is generated by combining the target state and the heading change rate.

[0015] In the above technical solution, the airspace is monitored jointly by three systems: a primary radar surveillance system, a secondary radar surveillance system, and an ADS-B system. Surveillance data for the monitored airspace is acquired from these three systems, preprocessed, and data from observation points that clearly do not conform to the target's characteristics are removed based on target surveillance information. The preprocessed surveillance data is then fused to generate fused surveillance data. This surveillance data does not rely on data actively provided by the aircraft but is acquired through external surveillance systems. In this way, even if the UAV's maneuverability is highly variable and changes rapidly, its flight status can be captured through surveillance data fusion, thus solving the problem of significant differences between the flight status of UAVs and manned aircraft. Furthermore, acquiring data through three systems and fusing it overcomes the problem of the reliance on single flight data sources.

[0016] The flight path of the aircraft is calculated by point-horizon correlation of the fused monitoring data, and the state of the target aircraft is estimated by point-horizon correlation of the flight path of the aircraft, so that the state information of the target aircraft is obtained, and the state of the target is estimated. By further calculating the heading of the target state, the heading change rate is obtained, and the flight state of the target aircraft is obtained by comprehensively considering the target state and the heading change rate. The flight state can be used to identify the flight mode of the aircraft. This step solves the problem that the existing model lacks identification processing of the aircraft heading and turn rate related mode; at the same time, the method of point-horizon correlation and state estimation is used to identify the flight mode, without the need for pre-training sample data or pre-setting flight model, so that the real-time response to the sudden change of the flight mode of the aircraft can be realized, thereby improving the real-time performance and solving the problems of time complexity and real-time performance of deep learning and neural network algorithm.

[0017] In an optional embodiment, the monitoring data is preprocessed, including:

[0018] determining a reference site center, and converting coordinates of the monitoring data to the reference site center;

[0019] unifying the time of the monitoring data to the time of the reference site center, and compensating the unified time of the monitoring data for time deviation.

[0020] In an optional embodiment, the preprocessed monitoring data is fused, including:

[0021] extracting distance information, angle information and height information from the preprocessed monitoring data;

[0022] The distance information includes data source distance observation results and distance precision variance, and the data source distance observation results and the distance precision variance are weighted and averaged to obtain fused distance;

[0023] The angle information includes data source angle observation results and angle precision variance, and the data source angle observation results and the angle precision variance are calculated by using the direct and inverse tangent function to obtain fused angle;

[0024] The height information includes data source height observation results and height precision variance, and the data source height observation results and the height precision variance are weighted and averaged to obtain fused height information.

[0025] In an optional embodiment, the fused monitoring data is point-horizon correlated, including:

[0026] The observation nodes in the fused monitoring data entering the correlation gate are used as starting track hypothesis to create hypothesis nodes;

[0027] calculate a normalized distance of the hypothesis node by covariance, take the hypothesis node with the normalized distance within a set gate as a candidate hypothesis node;

[0028] traverse the candidate hypothesis node, perform point-horizon association on the candidate hypothesis node by using an improved multi-hypothesis tracking algorithm, and update the candidate hypothesis node according to a point-horizon association result;

[0029] perform node quality evaluation on the updated candidate hypothesis node, and determine a current track target according to a node quality evaluation result.

[0030] In an optional embodiment, state estimation is performed by a result of point-horizon association, including: performing track state estimation on the current track target by using an interacting multiple model algorithm to obtain a target state; wherein the target state includes a horizontal direction state and a vertical direction state;

[0031] the horizontal direction state includes position, speed and acceleration information describing the horizontal direction; wherein the speed describing the horizontal direction is used for heading estimation;

[0032] the vertical direction state includes position, speed and acceleration information describing the vertical direction.

[0033] In an optional embodiment, performing track state estimation on the current track target by using an interacting multiple model algorithm includes:

[0034] constructing a target estimation model; wherein the target estimation model includes a uniform speed model, a uniform acceleration model and a weak maneuvering model;

[0035] initializing model probability and transition probability of the target estimation model;

[0036] performing model interaction on the initialized target estimation model to obtain a model interaction state, a model covariance matrix and a model interaction probability;

[0037] performing filtering processing on the model interaction state and the covariance matrix by using Kalman filtering to obtain filtering new information and covariance corresponding to the filtering new information;

[0038] updating the model interaction probability by using the filtering new information and the covariance corresponding to the filtering new information to obtain an updated probability;

[0039] updating the target estimation model by using the updated probability to obtain an updated target estimation model;

[0040] performing track state estimation on the current track target by using the updated target estimation model to obtain a target state.

[0041] The second aspect of the present application provides an aircraft flight mode recognition system based on an air monitoring system, comprising:

[0042] a preprocessing module configured to acquire monitoring data of a monitoring airspace, and pre-process the monitoring data;

[0043] a data fusion module configured to perform data fusion on the pre-processed monitoring data to generate fused monitoring data;

[0044] a state estimation module configured to perform point-horizon correlation on the fused monitoring data, and perform state estimation based on a result of the point-horizon correlation to obtain a target state;

[0045] a heading estimation module configured to perform heading estimation on the target state to obtain a heading rate, and generate an aircraft mode based on the target state and the heading rate.

[0046] In an optional embodiment, the preprocessing module comprises:

[0047] a coordinate conversion unit configured to determine a reference site center, and convert coordinates of the monitoring data to the reference site center;

[0048] a time unification unit configured to unify time of the monitoring data to time of the reference site center, and perform time deviation compensation on the unified time of the monitoring data.

[0049] In an optional embodiment, the data fusion module comprises:

[0050] a data extraction unit configured to extract distance information, angle information, and height information from the pre-processed monitoring data;

[0051] a distance fusion unit configured to perform weighted average calculation on data source distance observation results and distance precision variances of the distance information to obtain fused distance;

[0052] an angle fusion unit configured to perform calculation on data source angle observation results and angle precision variances of the angle information by using a cotangent function to obtain fused angle;

[0053] a height fusion unit configured to perform weighted average calculation on data source height observation results and height precision variances of the height information to obtain fused height.

[0054] In an optional embodiment, the state estimation module comprises:

[0055] A model construction unit is configured to construct a target estimation model, wherein the target estimation model comprises a uniform speed model, a uniform acceleration model and a weak maneuvering model.

[0056] An initialization unit is configured to initialize model probabilities and transition probabilities of the target estimation model.

[0057] An interaction unit is configured to perform model interaction on the initialized target estimation model to obtain a model interaction state, a model covariance matrix and a model interaction probability.

[0058] A filtering unit is configured to perform filtering processing on the model interaction state and the covariance matrix by using Kalman filtering to obtain filtering new information and a covariance corresponding to the filtering new information.

[0059] A probability updating unit is configured to update the model interaction probability by using the filtering new information and the covariance corresponding to the filtering new information to obtain an updated probability.

[0060] A model updating unit is configured to update the target estimation model by using the updated probability to obtain an updated target estimation model.

[0061] A target state estimation unit is configured to perform track state estimation on the current track target by using the updated target estimation model to obtain a target state.

[0062] Compared with the prior art, the present application has the following advantages and beneficial effects:

[0063] 1. The flight state of a target aircraft is obtained by calculating the heading rate of the target state, and the flight state of the target aircraft is obtained by comprehensively considering the target state and the heading rate, thereby solving the problem that the existing model lacks recognition and processing of the aircraft heading and turning rate related mode.

[0064] 2. The flight mode is recognized by the point-horizon association and state estimation method, without the need for pre-training sample data or pre-setting a flight model, and the real-time response to the sudden change of the flight mode of the aircraft is realized, thereby improving the real-time performance and solving the problems of the deep learning and neural network algorithm in terms of time complexity and real-time performance. BRIEF DESCRIPTION OF DRAWINGS

[0065] In order to more clearly illustrate the technical solutions of the example embodiments of the present application, the following will briefly introduce the drawings needed to be used in the embodiments, and it should be understood that the following drawings only show some embodiments of the present application, and therefore should not be regarded as a limitation on the scope, and for those skilled in the art, other related drawings can also be obtained without creative labor. In the drawings:

[0066] Figure 1A structural schematic diagram of aircraft flight mode recognition provided for the embodiment 1 of the present application is shown in the figure.

[0067] Figure 2 A structural schematic diagram of an electronic device provided for the embodiment 3 of the present application. DETAILED DESCRIPTION

[0068] In order to make the objectives, technical solutions and advantages of the present application clearer, further detailed description will be given to the present application in combination with embodiments and drawings, and the illustrative embodiments of the present application and the description thereof are only used to explain the present application, and do not limit the present application.

[0069] Embodiment 1

[0070] Figure 1 A structural schematic diagram of aircraft flight mode recognition provided for the embodiment 1 of the present application is shown in the figure. Figure 1 As shown in the figure, a method for aircraft flight mode recognition based on an air monitoring system comprises the following steps:

[0071] Obtaining monitoring data of a monitoring airspace, and pre-processing the monitoring data;

[0072] Performing data fusion on the pre-processed monitoring data to generate fused monitoring data;

[0073] Performing point-horizon correlation on the fused monitoring data, and performing state estimation through the result of the point-horizon correlation to obtain a target state;

[0074] Performing heading estimation on the target state to obtain a heading change rate, and generating an aircraft mode by comprehensively considering the target state and the heading change rate.

[0075] It should be noted that the method performs monitoring on the airspace through a primary radar monitoring system, a secondary radar monitoring system and an ADS-B system, obtains monitoring data of the monitoring airspace from the above three systems, pre-processes the monitoring data, removes the monitoring data of observation points that do not obviously conform to the target characteristics according to target monitoring information, and performs data fusion on the pre-processed monitoring data to generate fused monitoring data. The above monitoring data does not depend on data actively provided by an aircraft, but is obtained through an external monitoring system. In this way, even if the maneuverability of the unmanned aerial vehicle is variable and the change rate is fast, the flight state of the unmanned aerial vehicle can also be captured through monitoring data fusion, thereby solving the problem of large difference in flight state between the unmanned aerial vehicle and manned aircraft. Meanwhile, the data obtained through the three systems is fused, thereby overcoming the problem of single flight data.

[0076] The flight path of the aircraft is calculated by point flight correlation of the fusion monitoring data, and the state estimation of the aircraft track generated by the point flight correlation is performed to obtain the state information of the target aircraft, so as to realize the estimation of the target state. By further calculating the heading of the target state, the heading change rate is obtained, and the target state and the heading change rate are integrated to obtain the flight state of the target aircraft. The flight state can be used to identify the flight mode of the aircraft. This step solves the problem that the existing model lacks identification processing of the aircraft heading and turn rate related mode; at the same time, the method of point flight correlation and state estimation is used to identify the flight mode, without pre-training sample data or pre-setting flight model, so that the real-time response to the sudden change of the flight mode of the aircraft can be realized, thereby improving the real-time performance and solving the problems of time complexity and real-time performance of deep learning and neural network algorithm.

[0077] In an optional embodiment, the monitoring data is preprocessed, including:

[0078] The coordinates of the monitoring data are converted to the reference station center;

[0079] The time of the monitoring data is unified to the time of the reference station center, and the time of the unified monitoring data is compensated for time deviation.

[0080] It should be noted that the purpose of preprocessing in this embodiment is twofold. One is to calibrate the time and space of the target data received from multiple monitoring data sources, and the other is to eliminate observation points that do not obviously conform to the target characteristics according to the target monitoring information.

[0081] In an optional embodiment, the preprocessed monitoring data is fused, including:

[0082] Distance information, angle information and height information are extracted from the preprocessed monitoring data;

[0083] The distance information includes data source distance observation results and distance precision variance, and the data source distance observation results and the distance precision variance are weighted and averaged to obtain the fused distance;

[0084] The angle information includes data source angle observation results and angle precision variance, and the data source angle observation results and the angle precision variance are calculated by using the direct and inverse tangent function to obtain the fused angle;

[0085] The height information includes data source height observation results and height precision variance, and the data source height observation results and the height precision variance are weighted and averaged to obtain the fused height information.

[0086] It should be noted that data fusion is a process of combining information from multiple data sources into a consistent, accurate, and useful information representation, and in the embodiments of the present application, the unique target information of each monitoring data source is retained, and (2) the common target information of each monitoring data source is fused.

[0087] In the embodiments, the data fusion involves fusion of distance, angle and height information. The fusion of distance, angle and height information is crucial for aircraft state estimation. Among them, the distance information can determine the distance between the aircraft and the surrounding objects or target points or the distance between the aircraft and specific landmarks or navigation points, and is also used for subsequent aircraft speed estimation and heading estimation; the angle information can help to identify the attitude of the aircraft, perceive the direction of the aircraft in space, and be used for subsequent identification of aircraft turning, rolling and other action recognition; the height information can be used to monitor the vertical movement of the aircraft, which is very important for the control of flight modes such as take-off, landing and hovering.

[0088] Specifically, the calculation formula for distance information fusion with k data sources is as follows:

[0089]

[0090] In the above formula, is the fused distance, r i is the distance observation result of the i th data source, denotes the distance accuracy variance observed by the i th data source.

[0091] Specifically, the calculation formula for angle information fusion with k data sources is as follows:

[0092]

[0093] In the above formula, is the fused angle, θ i is the angle observation result of the i th data source, denotes the angle accuracy variance observed by the i th data source.

[0094] Specifically, the calculation formula for height information fusion with k data sources is as follows:

[0095]

[0096] In the above formula, is the fused height, h i is the height observation result of the i th data source, denotes the height accuracy variance observed by the i th data source.

[0097] In an alternative embodiment, the fusion monitoring data is point-horizon correlated, comprising:

[0098] The observed nodes entering the correlation gate in the fusion monitoring data are created as starting track hypotheses to create hypothesis nodes;

[0099] The normalized distance of the hypothesis nodes is calculated by covariance, and the hypothesis nodes within the set gate are selected as candidate hypothesis nodes;

[0100] The candidate hypothesis nodes are traversed, the improved multi-hypothesis tracking algorithm is used for point-horizon correlation of the candidate hypothesis nodes, and the candidate hypothesis nodes are updated according to the point-horizon correlation result;

[0101] The updated candidate hypothesis nodes are evaluated in terms of node quality, and the current track target is determined according to the node quality evaluation result.

[0102] It should be noted that the present step corresponds to the track management in Figure 1 The purpose of the track management is to realize point-horizon correlation by using the improved multi-hypothesis tracking algorithm. The gate is a threshold for determining whether the observation value matches multiple tracks, thereby avoiding false updates. It plays a key role in multi-target track correlation and helps to determine whether the observation value is associated with a specific track. The correlation gate in the present embodiment refers to the gate threshold of track matching. All observation points entering the correlation gate, i.e. within the gate threshold, are created as hypotheses of the started track to create hypothesis nodes for updating, wherein the improved nearest neighbor algorithm is used for target correlation when correlation is performed.

[0103] The innovation of the target and the covariance are calculated by target prediction, and the influence of target measurement noise is considered, and the target within the set gate becomes a candidate hypothesis. The innovation of the target is a key concept, which represents the difference between the actual observation value and the predicted observation value. The innovation is an intermediate variable in the filtering process, which is used to update the state estimation and covariance matrix of the filter.

[0104] All hypothesis nodes of all existing targets are traversed for correlation update, and the nodes are pruned according to the filtering state during target hypothesis node update. The node quality is evaluated through the continuous filtering results of each observation node of the target, and the current most hypothesis is selected as the update result of the current track target according to the evaluation score.

[0105] Further, the present embodiment uses m-n logic criteria for track initiation and track termination. Specifically, in an observation window of m times, there are n times of observations that meet the conditions, i.e. the track can be started; in an observation window of m times, there are n times of observations that do not meet the conditions, i.e. the track can be terminated.

[0106] In an alternative embodiment, the state estimation is performed by point association results, comprising: performing track state estimation on the current track target by using an interacting multiple model algorithm to obtain a target state; wherein the target state comprises a horizontal direction state and a vertical direction state;

[0107] The horizontal direction state comprises position, velocity and acceleration information in the horizontal direction; wherein the velocity in the horizontal direction is used for heading estimation;

[0108] The vertical direction state comprises position, velocity and acceleration information in the vertical direction.

[0109] It should be noted that the state estimation is performed in the horizontal plane and the vertical direction, the horizontal direction is represented by <x, Vx, Ax, y, Vy, Ay> to describe the position, velocity and acceleration information in the horizontal direction, and the vertical direction is represented by <z, Vz, Az> to describe the position, velocity and acceleration information in the vertical direction.

[0110] In the embodiment, the target heading and turning rate states are added, specifically, the heading estimation is performed by using the Vx and Vy information of the target, and the heading change rate is obtained by differentiating the continuous heading estimation.

[0111] In an alternative embodiment, the track state estimation on the current track target is performed by using an interacting multiple model algorithm, comprising:

[0112] constructing a target estimation model; wherein the target estimation model comprises a constant speed model, a constant acceleration model and a weak maneuvering model;

[0113] initializing the model probability and transition probability of the target estimation model;

[0114] performing model interaction on the initialized target estimation model to obtain a model interaction state, a model covariance matrix and a model interaction probability;

[0115] performing Kalman filtering on the model interaction state and the covariance matrix to obtain a filter innovation and a covariance corresponding to the filter innovation;

[0116] updating the model interaction probability by using the filter innovation and the covariance corresponding to the filter innovation to obtain an updated probability;

[0117] updating the target estimation model by using the updated probability to obtain an updated target estimation model;

[0118] performing track state estimation on the current track target by using the updated target estimation model to obtain a target state.

[0119] Specifically, the calculation process of the model interaction state is as follows:

[0120]

[0121] In the above formula, is the state estimation of the jth model at time k, u k-∥k-1 (i|j) is the transition probability from the ith model to the jth model, and N is the total number of models.

[0122] Specifically, the model covariance matrix calculation process is as follows:

[0123]

[0124] In the above formula, is the covariance matrix of the jth model at time k-1.

[0125] Specifically, the model interaction probability calculation process is as follows:

[0126]

[0127] In the above formula, is the transition probability of the ith sub-model to the jth sub-model at the next time; u k-1 (i) is the probability of the ith sub-model at time k-1; is a normalization factor.

[0128] Wherein, the model interaction state and covariance matrix are filtered by using Kalman filtering, including:

[0129] State transition: X k = F k X k-1 + B k u k +w k

[0130] Covariance prediction: P k = F k P k-1 F k ′ + Q k

[0131] Filtering gain: K k = P k H ′ k (H k P k H ′ k + R k ) -1

[0132] State update:

[0133] Covariance update: P k = (I - K k H k )P k|k-1

[0134] The model interaction probability is updated using the filter innovation and the covariance corresponding to the filter innovation, including calculating the likelihood function of the model, the filter innovation and the covariance corresponding to the innovation, wherein the calculation is as follows:

[0135] Likelihood function of the model:

[0136] Filter innovation of the model:

[0137] Covariance corresponding to the filter information:

[0138] Model probability update: wherein

[0139] In the above formula, is the likelihood function of the jth model at time k, indicating the matching degree of the observation data under the model; m is the dimension of the observation vector; is the determinant of the information covariance of the jth model at time k; is the innovation of the jth sub-model at time k, indicating the difference between the observation value and the predicted value; is the information covariance of the jth model at time k; H j (k) is the observation matrix of the jth model at time k, used to map the state space to the observation space; R(k) is the observation noise covariance at time k.

[0140] The target estimation model is updated using the updated probability, including integrating the state of each model and calculating the covariance matrix of each model, and the calculation process is as follows:

[0141] Integrate the model state:

[0142] Integrate the model covariance matrix:

[0143]

[0144] Further, the configuration management of the communication interface and communication protocol of the target data acquisition channel; the configuration management of the track management data start / stop criterion and filter parameters, multi-model initialization parameters. Statistics of the processing state of the whole system, realization of the management of the processing load, automatic exclusion of the remote target when exceeding the system processing capability, adaptive system load management;

[0145] The display end displays the target original information of each data source received; and displays the target and its mode information after state estimation and mode analysis.

[0146] Embodiment 2

[0147] The embodiment 2 of the present application provides an aircraft flight mode recognition system based on an air monitoring system, comprising:

[0148] A preprocessing module is configured to acquire monitoring data of a monitoring airspace, and pre-process the monitoring data;

[0149] A data fusion module is configured to perform data fusion on the pre-processed monitoring data, and generate fused monitoring data;

[0150] A state estimation module is configured to perform point-track association on the fused monitoring data, and perform state estimation through the result of the point-track association, to obtain target state;

[0151] A heading estimation module is configured to perform heading estimation on the target state, to obtain heading rate, and generate aircraft mode by integrating the target state and the heading rate.

[0152] In an optional embodiment, the preprocessing module comprises:

[0153] A coordinate conversion unit is configured to determine a reference site center, and convert the coordinates of the monitoring data to the reference site center;

[0154] A time unification unit is configured to unify the time of the monitoring data to the time of the reference site center, and perform time deviation compensation on the unified time of the monitoring data.

[0155] In an optional embodiment, the data fusion module comprises:

[0156] A data extraction unit is configured to extract distance information, angle information and height information from the pre-processed monitoring data;

[0157] A distance fusion unit is configured to perform weighted average calculation on the distance source observation result and the distance precision variance of the distance information, to obtain fused distance;

[0158] An angle fusion unit is configured to calculate the data source angle observation and the angle precision variance by using an arc tangent function to obtain a fused angle.

[0159] A height fusion unit is configured to perform weighted average calculation on the data source height observation and the height precision variance to obtain a fused height.

[0160] In an optional embodiment, the state estimation module comprises:

[0161] A model construction unit is configured to construct a target estimation model, wherein the target estimation model comprises a uniform speed model, a uniform acceleration model and a weak maneuvering model.

[0162] An initialization unit is configured to initialize model probability and transition probability of the target estimation model.

[0163] An interaction unit is configured to perform model interaction on the initialized target estimation model to obtain a model interaction state, a model covariance matrix and a model interaction probability.

[0164] A filter unit is configured to perform filter processing on the model interaction state and the covariance matrix by using Kalman filtering to obtain filter innovation and covariance corresponding to the filter innovation.

[0165] A probability updating unit is configured to update the model interaction probability by using the filter innovation and the covariance corresponding to the filter innovation to obtain an updated probability.

[0166] A model updating unit is configured to update the target estimation model by using the updated probability to obtain an updated target estimation model.

[0167] A target state estimation unit is configured to perform track state estimation on the current track target by using the updated target estimation model to obtain a target state.

[0168] Embodiment 3

[0169] Figure 2 A structural schematic diagram of an electronic device provided for the embodiment 3 of the present application is shown in FIG. 1, which comprises a processor 21, a memory 22, an input device 23 and an output device 24. Figure 2 The number of the processors 21 in the computer device can be one or more, and one processor 21 is taken as an example in the embodiment. Figure 2 The processor 21, the memory 22, the input device 23 and the output device 24 in the electronic device can be connected through a bus or other means, and the connection through the bus is taken as an example in the embodiment. Figure 2 ​

[0170] The memory 22 can be used to store software programs, computer executable programs and modules as a computer readable storage medium. The processor 21 executes various functions and data processing of the electronic device by running the software programs, instructions and modules stored in the memory 22, that is, implements the method for identifying flight mode of aircraft based on air monitoring system in embodiment 1.

[0171] The memory 22 can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system and at least one application program required by a function; the data storage area can store data created according to the use of the terminal, etc. In addition, the memory 22 can include a high-speed random access memory, and can also include a non-volatile memory, such as at least one magnetic disk storage device, a flash memory device, or other non-volatile solid-state memory device. In some examples, the memory 22 can further include a memory remotely arranged with respect to the processor 21, which can be connected to the electronic device through a network. Examples of the above network include but are not limited to the Internet, an intranet, a local area network, a mobile communication network, and a combination thereof.

[0172] The input device 23 can be used to receive the id and password input by the user, etc. The output device 24 is used to output the network configuration page.

[0173] Embodiment 4

[0174] The embodiment 4 of the present application also provides a computer readable storage medium, and the computer executable instructions are used to implement the method for identifying flight mode of aircraft based on air monitoring system as provided in embodiment 1 when executed by a computer processor.

[0175] The storage medium provided by the embodiment of the present application contains computer executable instructions, which are not limited to the method operations provided in embodiment 1, but can also perform related operations in the method for identifying flight mode of aircraft based on air monitoring system provided by any embodiment of the present application.

[0176] The above specific embodiments further explain the purpose, technical solutions and beneficial effects of the present application. It should be understood that the above description is only a specific embodiment of the present application and is not used to limit the protection scope of the present application. Any modification, equivalent replacement, improvement, etc. within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. An aerial vehicle flight mode identification method based on an over-the-air surveillance system, the method comprising: The method comprises the following steps: acquiring monitoring data of a monitoring airspace, and preprocessing the monitoring data; performing data fusion on the preprocessed monitoring data to generate fused monitoring data; The specific process comprises: extracting distance information, angle information and height information from the preprocessed monitoring data; The distance information comprises data source distance observation results and distance precision variances, and the data source distance observation results and the distance precision variances are subjected to weighted average calculation to obtain fused distance; The angle information comprises data source angle observation results and angle precision variances, and the data source angle observation results and the angle precision variances are calculated by using a cotangent function to obtain fused angle; The height information comprises data source height observation results and height precision variances, and the data source height observation results and the height precision variances are subjected to weighted average calculation to obtain fused height information; performing point-track association on the fused monitoring data, and performing state estimation through a result of the point-track association to obtain target state; The point-track association on the fused monitoring data comprises: creating a hypothesis node by taking an observation node in the fused monitoring data that enters an association gate as a starting track hypothesis; calculating a normalized distance of the hypothesis node by using a covariance, and taking a hypothesis node with the normalized distance within a set gate as a candidate hypothesis node; iterating through the candidate hypothesis node, performing point-track association on the candidate hypothesis node by using an improved multi-hypothesis tracking algorithm, and updating the candidate hypothesis node according to a result of the point-track association; performing node quality evaluation on the updated candidate hypothesis node, and determining a current track target according to a result of the node quality evaluation; performing heading estimation on the target state to obtain a heading change rate, and generating an aircraft mode by comprehensively considering the target state and the heading change rate.

2. The aircraft flight mode identification method based on the air surveillance system according to claim 1, wherein, The preprocessing of the monitoring data comprises: determining a reference site center, and converting coordinates of the monitoring data to the reference site center; unifying time of the monitoring data to time of the reference site center, and performing time bias compensation on the unified time of the monitoring data.

3. The aircraft flight mode identification method based on the air surveillance system according to claim 1, wherein, The state estimation through the result of the point-track association comprises: performing track state estimation on the current track target by using an interacting multiple model algorithm to obtain target state; wherein the target state comprises horizontal direction state and vertical direction state; The horizontal direction state comprises position, speed and acceleration information in the horizontal direction; wherein the speed in the horizontal direction is used for heading estimation; The vertical direction state comprises position, speed and acceleration information in the vertical direction.

4. The aircraft flight mode identification method based on the air surveillance system according to claim 3, characterized in that, The track state estimation on the current track target by using the interacting multiple model algorithm comprises: constructing a target estimation model; wherein the target estimation model comprises a uniform speed model, a uniform acceleration model and a weak maneuvering model; initializing model probability and transition probability of the target estimation model; performing model interaction on the initialized target estimation model to obtain model interaction state, model covariance matrix and model interaction probability; Filtering the model interaction state and the covariance matrix by using Kalman filtering to obtain filter innovation and covariance corresponding to the filter innovation; Updating the model interaction probability by using the filter innovation and the covariance corresponding to the filter innovation to obtain an updated probability; Updating the target estimation model by using the updated probability to obtain an updated target estimation model; Estimating the track state of the current track target by using the updated target estimation model to obtain a target state.

5. An aerial vehicle flight mode identification system based on an over-the-air surveillance system, characterized in that, The method comprises the following steps: A preprocessing module is configured to acquire monitoring data of a monitoring airspace and pre-process the monitoring data; A data fusion module is configured to fuse the pre-processed monitoring data to generate fused monitoring data; The data fusion module comprises: A data extraction unit is configured to extract distance information, angle information and height information from the pre-processed monitoring data; A distance fusion unit is configured to perform weighted average calculation on data source distance observation results and distance precision variances of the distance information to obtain fused distance; An angle fusion unit is configured to perform calculation on data source angle observation results and angle precision variances of the angle information by using a cotangent function to obtain fused angle; A height fusion unit is configured to perform weighted average calculation on data source height observation results and height precision variances of the height information to obtain fused height information; A state estimation module is configured to perform point-track association on the fused monitoring data and perform state estimation based on a result of the point-track association to obtain a target state; The specific process of the state estimation module performing point-track association on the fused monitoring data comprises the following steps: An observation node in the fused monitoring data that enters a correlation gate is taken as a starting track hypothesis to create a hypothesis node; A normalized distance of the hypothesis node is calculated by using covariance, and a hypothesis node with the normalized distance within a set gate is taken as a candidate hypothesis node; The candidate hypothesis nodes are traversed, the improved multi-hypothesis tracking algorithm is used to perform point-track association on the candidate hypothesis nodes, and the candidate hypothesis nodes are updated according to a result of the point-track association; The updated candidate hypothesis nodes are subjected to node quality evaluation, and a current track target is determined according to a result of the node quality evaluation; A heading estimation module is configured to perform heading estimation on the target state to obtain a heading change rate, and generate an aircraft mode by comprehensively considering the target state and the heading change rate.

6. An aerial vehicle flight mode identification system based on an over-the-air surveillance system as defined in claim 5, wherein, The preprocessing module comprises: A coordinate conversion unit is configured to determine a reference site center and convert coordinates of the monitoring data to the reference site center; A time unification unit is configured to unify time of the monitoring data to time of the reference site center and compensate for time deviation of the unified monitoring data.

7. An aerial vehicle flight mode identification system based on an aerial surveillance system as defined in claim 5, wherein, The state estimation module further comprises: A model construction unit is configured to construct a target estimation model; wherein the target estimation model comprises a uniform speed model, a uniform acceleration model and a weak maneuvering model. An initialization unit is configured to initialize model probabilities and transition probabilities of the target estimation model; An interaction unit is configured to perform model interaction on the initialized target estimation model to obtain a model interaction state, a model covariance matrix, and model interaction probabilities; A filtering unit is configured to perform Kalman filtering on the model interaction state and the covariance matrix to obtain filtering new information and covariance corresponding to the filtering new information; A probability updating unit is configured to update the model interaction probabilities by using the filtering new information and the covariance corresponding to the filtering new information to obtain updated probabilities; A model updating unit is configured to update the target estimation model by using the updated probabilities to obtain an updated target estimation model; A target state estimation unit is configured to perform track state estimation on the current track target by using the updated target estimation model to obtain a target state.

Images