Air target interception intention inference method and system based on track prediction

Abstract

The application belongs to the field of aerial target threat estimation, and discloses an aerial target interception intention inference method and system based on track prediction. The application first acquires historical state estimation information of an aircraft and an interceptor, then performs short-time track prediction, then calculates a predicted segment dynamic threat degree to obtain an interception intention matrix, and finally performs long-time prediction on the track of the interceptor with interception intention and the corresponding aircraft to obtain the position and size of an interception encounter area. The application generates pseudo-measurement data through bidirectional smoothing filtering and polynomial fitting to obtain a relatively accurate predicted track, uses an improved nearest encounter point function to more accurately calculate the dynamic threat degree, and then predicts the interception encounter area, so that the inference of the aerial target interception intention and the prediction of the interception encounter area obtain more accurate results.

Description

TECHNICAL FIELD

[0001] The application belongs to the field of aerial target threat estimation, and particularly relates to an aerial target interception intention inference method and system based on track prediction. BACKGROUND

[0002] In the rapidly changing battlefield environment of information-based warfare, it is a very important work to quickly distinguish the enemy target raw data obtained from a large number of complex sensors, and to obtain key information such as target type, position, speed, etc. through data preprocessing, and to judge the threat degree of the incoming target to our side, and to provide data support for the battlefield commander to make corresponding operational deployment decisions. With the rapid equipment and use of new weapons and equipment, and the innovative use of theoretical tactics, the air combat environment is becoming increasingly complex, and online identification of dynamic threats through track prediction to infer interception intention is a prerequisite for target assignment and maneuver decision-making. Especially in the operational scenario of aircraft breakout, an accurate and reliable interception intention inference method can judge the most threatening enemy interceptor in real time, and provide support for subsequent countermeasures or other decisions, thereby ensuring the smooth progress of the aircraft breakout mission.

[0003] The methods currently used to solve the problem of target threat estimation in complex environments at home and abroad mainly include Bayesian network reasoning, Analytic Hierarchy Process (AHP), multi-attribute decision-making theory, fuzzy comprehensive evaluation method, grey correlation method and deep learning method, etc. The coordination form of each terminal platform is still relatively simple, and the self-organization and intelligence degree is not high enough. According to the existing research results, we can find that the main problem of the existing methods is that the algorithm output results are discrete results, and directly using them for multi-target threat sorting will lead to other more difficult problems. For example, the lack of threat degree discretization will make it difficult to compare the threat degree between targets, and if multiple targets belonging to different categories and different states are at the same level of threat degree, they are easy to generate unreasonable sorting results that obviously violate normal cognition in the subsequent threat sorting process due to information loss. In addition, how to further process the threat estimation results for use in related specific intention inference methods is also an important research direction at present.

[0004] The track prediction problem refers to predicting the track of a target in a future period of time as accurately as possible by making full use of known information of the target, including historical track, motion characteristics, and even target attribute category, behavior intention and other information. Its essence belongs to the estimation problem of predicting "unknown" by "known". At present, the methods for solving the target track prediction problem at home and abroad mainly include two categories: model-based track prediction method and deep learning-based track prediction method. Through analysis and comparison of the existing track prediction methods, it can be known that the model-based track prediction method has the advantages of simple structure and strong robustness, but it is difficult to guarantee high prediction accuracy under the conditions of complex model, easy mismatch, or unknown statistical characteristics of noise. Although the deep learning-based track prediction method can improve the track prediction accuracy by combining different neural networks, it has poor interpretability, and the high computational complexity makes it difficult to meet the real-time requirement.

[0005] In the method for predicting the interception intention of an aerial target and the interception encounter area, there are few existing research methods, and more are deep learning methods, which have poor interpretability and lack of real-time performance. How to use track prediction and threat estimation in the inference of interception intention and the prediction of interception encounter area is an important problem to be solved at present. SUMMARY

[0006] The purpose of the present application is to overcome the above-mentioned deficiencies, and to provide an aerial target interception intention inference method and system based on track prediction, which can accurately infer and predict the interception intention of the interceptor and the interception encounter area.

[0007] In order to achieve the above-mentioned purpose, the technical scheme adopted by the present application is as follows:

[0008] In a first aspect, the present application provides an aerial target interception intention inference method based on track prediction, comprising the following steps:

[0009] Obtaining historical state information of the interceptor and the aircraft;

[0010] Respectively performing bidirectional filtering and smoothing processing on the historical state information of the interceptor and the aircraft to obtain bidirectional filtering and smoothing state estimation results and covariance matrices;

[0011] Generating pseudo-measurement data based on polynomial fitting according to the bidirectional filtering and smoothing state estimation results and covariance matrices;

[0012] Performing augmented interacting multiple model-volumetric Kalman filtering on the pseudo-measurement data, updating the historical state information of the interceptor and the aircraft using the filtered pseudo-measurement data, and obtaining the predicted track of the interceptor and the predicted track of the aircraft;

[0013] The predicted flight path of the interceptor and the predicted flight path of the aircraft are processed by using the improved nearest encounter point function, and the dynamic threat degree of the predicted section is calculated;

[0014] According to the dynamic threat degree of the predicted section, the interception intention is judged in combination with the intention inference rule, and an interception intention matrix at each time is obtained.

[0015] According to the interception intention matrix at each time, the nearest distance between the interceptor and the aircraft is obtained in combination with the predicted flight path of the interceptor and the predicted flight path of the aircraft.

[0016] According to the nearest distance between the interceptor and the aircraft, the center point of the interception encounter area and the radius of the interception encounter area are determined.

[0017] The further improvement of the present application lies in that the historical state information of the interceptor comprises:

[0018]

[0019] wherein, the number of the interceptor, an x-axis component of the position estimation value of the interceptor, x a y-axis component of the position estimation value of the interceptor, a z-axis component of the position estimation value of the interceptor, y an x-axis component of the speed estimation value of the interceptor, a y-axis component of the speed estimation value of the interceptor, z a z-axis component of the speed estimation value of the interceptor. x y z

[0020] The further improvement of the present application lies in that the historical state information of the aircraft comprises:

[0021]

[0022] wherein, an x-axis component of the position estimation value of the aircraft, x a y-axis component of the position estimation value of the aircraft, a z-axis component of the position estimation value of the aircraft, y an x-axis component of the speed estimation value of the aircraft, a y-axis component of the speed estimation value of the aircraft, z a z-axis component of the speed estimation value of the aircraft. x y z ​​​​​​​​​​​​

[0023] The further improvement of the present application is that the specific method of bidirectional filtering and smoothing of the historical state information of the interceptor and the aircraft is as follows:

[0024] The prediction duration is set, the prediction interval is divided, the augmented IMM-CKF filtering is adopted for the historical state information of the interceptor or the aircraft, the state estimation is performed from the measurement start time to the measurement end time, and the forward filtering result and the forward filtering covariance matrix are obtained;

[0025] The EKF filtering is adopted for the historical state information of the interceptor or the aircraft, the filtering is performed in reverse from the measurement end time to the measurement start time, and the backward filtering result and the backward filtering covariance matrix are obtained;

[0026] The forward filtering result, the forward filtering covariance matrix, the backward filtering result and the backward filtering covariance matrix are combined, and the bidirectional filtering and smoothing result and the bidirectional filtering and smoothing covariance matrix are obtained

[0027] The further improvement of the present application is that the specific method of generating pseudo measurement data based on polynomial fitting according to the state estimation result and the covariance matrix after bidirectional filtering and smoothing is as follows:

[0028] The state estimation vector of the interceptor from the tracking start time to the end time is extracted, the state estimation vector of the interceptor is split into three groups of position data numbered in time sequence along the X, Y and Z axes of the ECEF coordinate system;

[0029] The polynomial of the order of 4 of the data in the X axis direction of the ECEF coordinate system is obtained; The polynomial of the order of 4 of the data in the Y axis direction of the ECEF coordinate system is obtained;

[0030] Based on the least square fitting, the residual sum of squares between the position prediction data changing with time and the actual data is obtained according to the polynomial coefficients obtained by time According to the necessary condition of finding the extreme value of the multivariate function, the analytical solution expression of the related parameters is obtained by combining the residual sum of squares between the position prediction data changing with time and the actual data;

[0031] The data in the X axis direction of the ECEF coordinate system is brought into the analytical solution expression of the related parameters, and the value of the related parameters is obtained, so that the polynomial of the order of 4 corresponding to the data in the X axis direction of the ECEF coordinate system is obtained;

[0032] The data in the Y axis direction of the ECEF coordinate system is brought into the analytical solution expression of the related parameters, and the value of the related parameters is obtained, so that the polynomial of the order of 4 corresponding to the data in the Y axis direction of the ECEF coordinate system is obtained; The data in the Z axis direction of the ECEF coordinate system is brought into the analytical solution expression of the related parameters, and the value of the related parameters is obtained, so that the polynomial of the order of 4 corresponding to the data in the Z axis direction of the ECEF coordinate system is obtained;

[0033] Similarly, according to the data in the X axis direction of the ECEF coordinate system and the data in the Y axis direction of the ECEF coordinate system, the polynomial of the order of 4 corresponding to the data in the X axis direction of the ECEF coordinate system is obtained; Similarly, according to the data in the X axis direction of the ECEF coordinate system and the data in the Y axis direction of the ECEF coordinate system, the polynomial of the order of 4 corresponding to the data in the X axis direction of the ECEF coordinate system is obtained; ​​​​​Axis and Axis direction data corresponds to Polynomial of order

[0034] According to Axis, Axis and Axis direction data corresponds to Polynomial of order, combined with the required prediction length, the required length of the interceptor in the ECEF coordinate system Coordinate information;

[0035] According to the required length of the interceptor in the ECEF coordinate system Coordinate information, combined with the position information of the aircraft in the prediction time period, the pseudo measurement equation is used to calculate the pseudo measurement data.

[0036] The further improvement of the present application is that the improved nearest encounter point function is used to process the predicted track of the interceptor and the predicted track of the aircraft, and the specific method for calculating the dynamic threat degree of the prediction section is as follows:

[0037] According to the predicted track of the interceptor and the predicted track of the aircraft, the relative position and relative speed of the interceptor to the aircraft are calculated;

[0038] According to the relative position and relative speed of the interceptor to the aircraft, the nearest encounter point distance and the nearest encounter point time are calculated;

[0039] The nearest encounter point time is normalized, and the interceptor time threat degree value is obtained according to the normalized nearest encounter point time;

[0040] The nearest encounter point distance is normalized, and the interceptor distance threat degree value is obtained according to the normalized nearest encounter point distance;

[0041] The interceptor speed and the included angle between the interceptor and the connecting line between the aircraft are used to correct the interceptor time threat degree value and the interceptor distance threat degree value;

[0042] According to the corrected interceptor time threat degree value and the interceptor distance threat degree value, the dynamic threat degree of the prediction section is obtained.

[0043] The further improvement of the present application is that according to the dynamic threat degree of the prediction section, the intention inference rule is combined to judge the interception intention, and the specific method for obtaining the interception intention matrix of each moment is as follows:

[0044] The preset intention inference rule is used to establish a judgment model of the prediction section dynamic threat degree and the intention inference rule;

[0045] The prediction section dynamic threat degree is input into the judgment model to obtain the interception intention of each moment.

[0046] According to the interception intention of each moment, an interception intention matrix of each moment is obtained.

[0047] The interception encounter area center point is determined according to the interception intention matrix of each moment, and the interception encounter area radius is calculated, and the specific method for obtaining the closest distance between the interceptor and the aircraft is as follows:

[0048] According to the interception intention matrix of each moment, the corresponding interceptor and aircraft are extracted, and the interception encounter area is predicted according to the interception intention matrix of the interceptor and the aircraft, the interception encounter area including a point set that can be hit by the interceptor on the aircraft;

[0049] The closest distance between the interceptor and the aircraft is obtained according to the predicted track of the interceptor and the predicted track of the aircraft.

[0050] The specific method for determining the interception encounter area center point and the interception encounter area radius according to the closest distance between the interceptor and the aircraft is as follows:

[0051] At the moment of the closest distance between the interceptor and the aircraft, the midpoint of the line connecting the predicted track of the interceptor and the predicted track of the aircraft is the interception encounter area center point.

[0052] The interception encounter area radius is obtained according to the closest distance between the interceptor and the aircraft and the predicted time length of the closest distance between the interceptor and the aircraft.

[0053] In a second aspect, the present application provides an air target interception intention inference system based on track prediction, including the following steps:

[0054] An information acquisition module is configured to acquire historical state information of the interceptor and the aircraft.

[0055] A filtering module is configured to perform bidirectional filtering and smoothing processing on the historical state information of the interceptor and the aircraft respectively, to obtain state estimation results and covariance matrices after bidirectional filtering and smoothing.

[0056] A pseudo two-side data acquisition module is configured to generate pseudo measurement data based on polynomial fitting according to the state estimation results and the covariance matrices after bidirectional filtering and smoothing.

[0057] A predicted track acquisition module is configured to perform augmented interactive multiple model-volumetric Kalman filtering on the pseudo measurement data, to update the historical state information of the interceptor and the aircraft by using the filtered pseudo measurement data, to obtain the predicted track of the interceptor and the predicted track of the aircraft.

[0058] A threat degree acquisition module is configured to process the predicted track of the interceptor and the predicted track of the aircraft by using an improved closest point of encounter function, to calculate a predicted segment dynamic threat degree.

[0059] An intercept intention matrix acquisition module is configured to determine the intercept intention according to the predicted segment dynamic threat degree and in combination with the intention inference rule, and obtain the intercept intention matrix at each time point;

[0060] A nearest distance acquisition module is configured to obtain the nearest distance between the interceptor and the aircraft according to the intercept intention matrix at each time point and in combination with the predicted flight path of the interceptor and the predicted flight path of the aircraft;

[0061] An intercept encounter area acquisition module is configured to determine the intercept encounter area center point and the intercept encounter area radius according to the nearest distance between the interceptor and the aircraft.

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

[0063] The present application first acquires the historical state estimation information of the aircraft and the interceptor, then performs short-time flight path prediction, then calculates the predicted segment dynamic threat degree to obtain the intercept intention matrix, and finally performs long-time prediction on the flight path of the interceptor with the intercept intention and the corresponding aircraft to obtain the position and size of the intercept encounter area. The present application generates pseudo-measurement data through bidirectional smoothing filtering and polynomial fitting to obtain a relatively accurate predicted flight path, uses an improved nearest encounter point function to more accurately calculate the dynamic threat degree, then predicts the intercept encounter area, so that the inference of the aerial target intercept intention and the prediction of the intercept encounter area obtain more accurate results. BRIEF DESCRIPTION OF DRAWINGS

[0064] Figure 1 The flowchart of the present application;

[0065] Figure 2 The flight path prediction result graph;

[0066] Figure 3 The simulation scene and flight path prediction result graph at the 4th s;

[0067] Figure 4 The threat prediction and intercept intention result graph of the aircraft 1 at the 4th s;

[0068] Figure 5 The threat prediction and intercept intention result graph of the aircraft 2 at the 4th s;

[0069] Figure 6 The threat prediction and intercept intention result graph of the aircraft 3 at the 4th s;

[0070] Figure 7 The intercept encounter area prediction result graph of the aircraft 1 and the interceptor 2 at the 3rd s;

[0071] Figure 8 The system diagram of the present application. DETAILED DESCRIPTION

[0072] For further understanding of the present application, the following will make a detailed description of the present application in combination with the accompanying drawings and specific embodiments. It should be understood that the embodiments are only to explain the present application and not to limit it.

[0073] Embodiment 1:

[0074] Referring to Figure 1 The air target interception intention inference method based on track prediction includes the following steps:

[0075] S1, obtaining the historical state information of the interceptor and the aircraft.

[0076] S2, performing bidirectional filtering smoothing processing on the historical state information of the interceptor and the aircraft respectively to obtain the bidirectional filtering smoothed state estimation result and the covariance matrix.

[0077] S3, generating pseudo-measurement data based on polynomial fitting according to the bidirectional filtering smoothed state estimation result and the covariance matrix.

[0078] S4, performing augmented interacting multiple model-volumetric Kalman filtering on the pseudo-measurement data, updating the historical state information of the interceptor and the aircraft using the filtered pseudo-measurement data to obtain the predicted track of the interceptor and the predicted track of the aircraft.

[0079] S5, processing the predicted track of the interceptor and the predicted track of the aircraft using the improved nearest encounter point function to calculate the predicted segment dynamic threat degree.

[0080] S5, judging the interception intention according to the predicted segment dynamic threat degree combined with the intention inference rule to obtain the interception intention matrix at each time.

[0081] S7, obtaining the closest distance between the interceptor and the aircraft according to the interception intention matrix at each time combined with the predicted track of the interceptor and the predicted track of the aircraft.

[0082] S8, determining the interception encounter area center point and the interception encounter area radius according to the closest distance between the interceptor and the aircraft.

[0083] Embodiment 2:

[0084] The present embodiment includes the following steps:

[0085] Step one, obtaining the historical state information of the interceptor and the aircraft;

[0086] Step 2: Two-Filter Smoothing (TFS) is performed on the interceptor's historical state estimation information. Then, pseudo-measurement data is generated based on polynomial fitting and combined with augmented interactive multi-model-volumetric Kalman (IMM-CKF) filtering to obtain the predicted track.

[0087] Step 3: Use the improved closest encounter point function (ICPA function) to calculate the dynamic threat level of the predicted segment, combine it with the intention inference rule to determine the interception intention, and obtain the interception intention matrix at each moment;

[0088] Step 4: Make a long-term prediction of the trajectory of the interceptor with interception intention and the corresponding aircraft, determine the center point of the interception encounter area, calculate the radius of the interception encounter area, and obtain the position and size (spherical) of the interception encounter area.

[0089] Example 3:

[0090] In step 1, the number is The interceptor to k The state estimation information at a given moment is represented as follows, consisting of 10-12 dimensions:

[0091] Formula 1

[0092] The first six dimensions are motion state information, which are expressed as follows in the ECEF coordinate system:

[0093] Formula 2

[0094] in, The interceptor position estimate x Axis component, The interceptor position estimate y Axis component, The interceptor position estimate z Axis component, is the interceptor velocity estimate x Axis component, is the interceptor velocity estimate y Axis component, is the interceptor velocity estimate z Axis component.

[0095] Number The aircraft status information is shown as follows:

[0096] Formula 3

[0097] in, The estimated value of the aircraft position x Axis component, The estimated value of the aircraft positiony axis component, for an aircraft position estimate z axis component, for an aircraft velocity estimate x axis component, for an aircraft velocity estimate y axis component, for an aircraft velocity estimate z axis component.

[0098] Embodiment 4:

[0099] In the step of predicting the trajectory of the interceptor, based on the bidirectional filtering and smoothing of the historical state estimation information of the interceptor, the generation of pseudo-measurement data based on polynomial fitting, and the combination of the augmented IMM-CKF filter, the specific process is as follows:

[0100] Step 1: Set the prediction time length, segment the prediction interval, and use the TFS algorithm to perform reverse filtering on the measurement values of the last segment of the interceptor's trajectory to smooth the estimation.

[0101] Forward filtering uses the augmented IMM-CKF filter to perform state estimation from the measurement start time to the measurement end time, obtaining the forward filtering result and the covariance matrix .

[0102] Backward filtering uses the EKF (Extended Kalman Filter) filter to perform filtering in reverse from the measurement end time to the measurement start time, obtaining the backward filtering result and the covariance matrix .

[0103] Step 2: Data fusion is performed on the results to obtain the bidirectional filtering and smoothing results and the covariance matrix .

[0104] Formula 4

[0105] Formula 5

[0106] Step 3: According to the smoothed historical state information of the interceptor, the optimal polynomial function is found.

[0107] For the selected interceptor , the state estimation vector k from the start time to the end time of tracking m is extracted, and along the three-axis direction of the ECEF coordinate system, the The state estimation vector is split into three sets of position data numbered in time sequence, as shown below:

[0108] Formula 6

[0109] In the ECEF coordinate system As an example, assume that there is data in the axis direction Order polynomial:

[0110] Formula 7

[0111] Approximate satisfaction The overall distribution of the data set that changes over time along the axis direction.

[0112] According to the Least Squares (LS) fitting requirements, it is necessary to calculate the time Obtain The polynomial coefficients can make the residual sum of squares between the time-varying position prediction data calculated by the polynomial and the actual data The minimum mathematical form is as follows:

[0113] Formula 8

[0114] According to the necessary conditions for finding the extreme value of a multivariate function, Find the partial derivatives and set each equation equal to 0, then solve to get the parameters The analytical solution expression is as follows:

[0115] Formula 9

[0116] ECEF coordinate system Data in the axis direction Substitute into formula 9 to solve the parameters At this point, the interceptor's old track in the ECEF coordinate system is obtained. Axis direction data corresponding to Order polynomial .

[0117] Similarly, the ECEF coordinate system Axis direction data and Axis direction data Substitute them into formula 9 respectively, and we can get Axis and Axis direction data corresponding to polynomial of order.

[0118] In finding the three coordinate axes Based on the order polynomial, according to the required prediction time , brought into In the above example, we can directly calculate the ~ The interceptor in the ECEF coordinate system within the time interval Coordinate information. Combined with the aircraft in the prediction interval ~ The position information is used to calculate the pseudo measurement using the measurement equation:

[0119] Formula 10

[0120] in, is the pseudo relative distance between the sensor and the target, is the pseudo target sight azimuth, is the elevation angle of the pseudo target line of sight, The variance is Gaussian white noise.

[0121] In the fourth step, after obtaining the pseudo-measurement, the interceptor state in the prediction period is updated based on the augmented IMM-CKF filtering technology to obtain the predicted track.

[0122] Example 5:

[0123] The specific process for calculating the dynamic threat level of the predicted segment using the ICPA function, determining the interception intention based on the intention inference rules, and obtaining the interception intention matrix at each moment is as follows:

[0124] In the first step, the ICPA function is used to calculate the dynamic threat degree of the predicted segment based on the trajectory status information of the interceptor and the aircraft.

[0125] Consider the distance between the interceptor's current trajectory and the closest point to our aircraft object and the time of closest arrival , which are the two indicators of closest encounter time and closest encounter distance. The following formulas are all based on a certain moment, so the time parameter is omitted.

[0126] (1) Calculation number is The enemy interceptor numbered The relative position of our aircraft and relative speed for:

[0127] Formula 11

[0128] Formula 12

[0129] Calculate the distance to the nearest encounter point for:

[0130] Equation 13

[0131] Time of the closest approach point is:

[0132] Equation 14

[0133] where, is the minimum distance between the interceptor and the aircraft at the point on the straight line trajectory of the interceptor according to the current speed and direction of movement of the interceptor, is the time of straight line movement of the interceptor from the current position to the closest point to the aircraft according to the current speed of the interceptor.

[0134] (2) After normalizing , the threat degree value of the interceptor defined according to the time of the closest approach point is:

[0135] Equation 15

[0136] where, is the time required for the enemy interceptor and the aircraft to reach the closest approach when the threat degree is 1, is the time required for the two to reach the closest approach when the threat degree is 0.7, is the time required for the two to reach the closest approach when the threat degree is 0.3, is the time required for the two to reach the closest approach when the threat degree is 0.

[0137] Similarly, the threat degree value of the interceptor defined according to the distance is:

[0138] Equation 16

[0139] where, is the closest approach distance between the enemy interceptor and the aircraft, is the closest approach distance between the enemy interceptor and the aircraft when the threat degree is 1, is the closest approach distance between the two when the threat degree is 0.7, is the closest approach distance between the two when the threat degree is 0.3, is the closest approach distance between the two when the threat degree is 0.

[0140] The angle between the speed of the interceptor and the line connecting the interceptor and the aircraft is introduced to represent the influence of the two cases of the interceptor approaching and moving away on the threat degree, and the processing method is as follows:

[0141] Equation 17

[0142] wherein, is the original CPA threat level, is the improved ICPA threat level, is a set coefficient, which will affect the sensitivity of the threat level to the change of the target speed direction.

[0143] (3) The ICPA threat level is calculated as:

[0144] Equation 18

[0145] It should be noted that the above method and formula only give a dynamic threat estimation model at a certain moment, in actual situations, , , , and etc. change with time, and the threat estimation result needs to be calculated in real time.

[0146] Second step, according to the intention inference rule, if there is an interception intention, set the corresponding item of the intention matrix to 1, and evaluate the intention level according to the average threat level of the predicted segment as follows: if the average threat level < 0.6, the interception intention level = 0; if the average threat level > 0.9, the interception intention level = 3. If neither is satisfied, calculate the difference between the threat level at the next moment and the current threat level, if the number of moments when the difference is greater than 0 accounts for 70% of the number of predicted moments, at this time if 0.6 if the average threat level < 0.7, the interception intention level = 1; if 0.7 if the average threat level < 0.8, the interception intention level = 2; if neither is satisfied, the interception intention level = 3. If there is no interception intention, set the corresponding item of the intention matrix to 0, and finally obtain the interception intention matrix at each moment.

[0147] Example 6:

[0148] In the step of long-time prediction of the trajectories of the interceptor with interception intention and the corresponding aircraft, determination of the center point of the interception encounter area, calculation of the radius of the interception encounter area, and obtaining of the position and size of the interception encounter area, the specific process is as follows:

[0149] First step, extract the interceptor with interception intention and the corresponding aircraft from the intention matrix, and perform interception encounter area prediction on this interception pair. The interception area includes the points around the aircraft that may be hit by the interceptor trajectory. Through trajectory prediction, the moment when the predicted trajectories of the two are closest is found, and the midpoint of the line connecting the two at this moment is the center position of the interception encounter area.

[0150] Second step, calculate the radius of the interception encounter area R :

[0151] Equation 19

[0152] wherein, is the closest distance between the enemy interceptor and our aircraft in the predicted time period, is the time length from the beginning of the prediction to the time when the enemy interceptor and our aircraft are closest, , and are the standard deviations of the measurement noise of the confrontation scene respectively (the sensor measurement noise is zero-mean Gaussian white noise).

[0153] Example 7:

[0154] Referring to Figure 8 The air target interception intention inference system based on track prediction comprises the following steps:

[0155] An information acquisition module is configured to acquire historical state information of the interceptor and the aircraft.

[0156] A filtering module is configured to perform bidirectional filtering and smoothing processing on the historical state information of the interceptor and the aircraft respectively, to obtain state estimation results and covariance matrices after bidirectional filtering and smoothing.

[0157] A pseudo two-sided data acquisition module is configured to generate pseudo measurement data based on polynomial fitting according to the state estimation results and the covariance matrices after bidirectional filtering and smoothing.

[0158] A predicted track acquisition module is configured to perform augmented interacting multiple model-volumetric Kalman filtering on the pseudo measurement data, to update the historical state information of the interceptor and the aircraft using the filtered pseudo measurement data, to obtain a predicted track of the interceptor and a predicted track of the aircraft.

[0159] A threat degree acquisition module is configured to process the predicted track of the interceptor and the predicted track of the aircraft using an improved closest point of encounter function, to calculate a predicted segment dynamic threat degree.

[0160] An interception intention matrix acquisition module is configured to determine the interception intention according to the predicted segment dynamic threat degree in combination with intention inference rules, to obtain an interception intention matrix at each time.

[0161] A closest distance acquisition module is configured to obtain the closest distance between the interceptor and the aircraft according to the interception intention matrix at each time in combination with the predicted track of the interceptor and the predicted track of the aircraft.

[0162] An interception encounter area acquisition module is configured to determine a center point and a radius of an interception encounter area according to the closest distance between the interceptor and the aircraft.

[0163] Example 8:

[0164] Referring to Figure 2 , if the prediction time length is set to 0.5s, i.e. 10 sampling periods, the double-sided filter smoothing result is represented by the thick solid line, the former half of the thin solid line represents the historical track, and the latter half of the thin solid line represents the track prediction result. It can be seen that after the bidirectional filter smoothing, the originally fluctuating initial track of the interceptor is well smoothed, so that the accuracy of the polynomial fitting is improved, the generation accuracy of the pseudo-measurement is improved, and the rationality of the track prediction is improved.

[0165] Referring to Figure 3 , the time of the current state estimation is the 4th second, three aircrafts and four interceptors are set in the simulation scene, aircraft 1 is guided by interceptor 1 and 2, and aircraft 3 is guided by interceptor 3 and 4. Referring to Figure 4 , Figure 5 , Figure 6 , for aircraft 1, the average threat degree of interceptor 1 and interceptor 2 is greater than that of interceptor 3 and interceptor 4, so interceptor 1 and interceptor 2 have high interception intention for the aircraft, for aircraft 2, since the average threat degree is less than the threshold 0.6 of the interception intention, at this time it is determined that the four interceptors have no interception intention, for aircraft 3, although the average of the threat degrees of the four interceptors in the threat prediction part is high, since interceptor 1 and interceptor 2 pass by aircraft 3, the threat degree has many falling parts, so it can be directly determined that there is no logical intention, and the predicted segment threat degree of interceptor 3 and interceptor 4 continues to rise, which is determined as high interception intention. In summary, the simulation results are consistent with the simulation scene settings.

[0166] Referring to Figure 7 , the simulation parameters remain unchanged, considering the simulation results at a single time, if the time of the current state estimation is the 3rd second, which is a middle period of the interception intention, aircraft 1 and interceptor 2 are selected for interception encounter area prediction. From the above simulation results, it can be seen that the proposed air target interception intention inference method based on track prediction can better predict and label the possible encounter area in the interception confrontation scene.

[0167] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present application but not to limit it, although the present application has been described in detail with reference to the above embodiments, those skilled in the art should understand that the specific embodiments of the present application can be modified or replaced by equivalents without departing from the spirit and scope of the present application, any modification or equivalent replacement without departing from the spirit and scope of the present application should be covered in the protection scope of the claims of the present application.

Claims

1. A method for inferring air target interception intention based on trajectory prediction, characterized in that: The following steps are involved: Obtain historical status information of interceptors and aircraft; The historical state information of the interceptor and the aircraft is subjected to two-way filtering and smoothing respectively to obtain the state estimation results and covariance matrix after two-way filtering and smoothing; According to the state estimation results and covariance matrix after bidirectional filtering and smoothing, pseudo measurement data is generated based on polynomial fitting. The specific method is as follows: Extract the interceptor's state estimation vector from the start to the end of tracking, along the ECEF coordinate system In the three-axis direction, the interceptor's state estimation vector is split into three groups of position data numbered in time sequence; Get the ECEF coordinate system Axis direction data polynomial of order; Based on the least squares fitting, according to the time The obtained polynomial coefficients are used to obtain the residual sum of squares between the time-varying position prediction data and the actual data; According to the necessary conditions for finding the extreme value of multivariate functions, combined with the residual sum of squares between the time-varying position prediction data and the actual data, the analytical solution expression of the relevant parameters is obtained; ECEF coordinate system The data in the axial direction are brought into the analytical solution expressions of the relevant parameters to obtain the values ​​of the relevant parameters, thereby obtaining the interceptor in the ECEF coordinate system. Axis direction data corresponding to polynomial of order; Similarly, according to the ECEF coordinate system Axis direction data and Axis direction data, get Axis and Axis direction data corresponding to polynomial of order; according to axis, Axis and Axis direction data corresponding to The order polynomial, combined with the required prediction time, is used to obtain the interceptor's position in the ECEF coordinate system within the required time. Coordinate information; According to the required time, the interceptor is located in the ECEF coordinate system. The coordinate information is combined with the position information of the aircraft in the predicted time period, and the pseudo-measurement calculation is performed using the measurement equation to obtain the pseudo-measurement data; Perform augmented interactive multi-model-volumetric Kalman filtering on the pseudo-measurement data, and use the filtered pseudo-measurement data to update the historical state information of the interceptor and the aircraft to obtain the predicted trajectory of the interceptor and the aircraft; The predicted trajectory of the interceptor and the aircraft are processed by using the improved closest approach point function to calculate the dynamic threat degree of the predicted section. According to the dynamic threat level of the predicted segment, the interception intention is judged in combination with the intention inference rules to obtain the interception intention matrix at each moment; According to the interception intention matrix at each moment, the predicted trajectory of the interceptor and the predicted trajectory of the aircraft are combined to obtain the closest distance between the interceptor and the aircraft; According to the minimum distance between the interceptor and the aircraft, the center point and radius of the interception encounter area are determined.

2. The method for inferring the aerial target interception intention based on trajectory prediction according to claim 1 is characterized in that: The interceptor's historical status information includes: in, is the interceptor number, The interceptor position estimate x Axis component, The interceptor position estimate y Axis component, The interceptor position estimate z Axis component, is the interceptor velocity estimate x Axis component, is the interceptor velocity estimate y Axis component, is the interceptor velocity estimate z Axis component.

3. The method for inferring the aerial target interception intention based on trajectory prediction according to claim 1, characterized in that: The historical status information of the aircraft includes: in, The estimated value of the aircraft position x Axis component, The estimated value of the aircraft position y Axis component, The estimated value of the aircraft position z Axis component, is the estimated value of the aircraft speed x Axis component, is the estimated value of the aircraft speed y Axis component, is the estimated value of the aircraft speed z Axis component.

4. The method for inferring the aerial target interception intention based on trajectory prediction according to claim 1, characterized in that: The specific method of performing bidirectional filtering and smoothing on the historical status information of the interceptor and the aircraft is as follows: Set the prediction time, divide the prediction interval, use augmented IMM-CKF filtering on the interceptor or aircraft historical state information, perform state estimation from the start time to the end time of measurement, and obtain the forward filtering result and forward filtering covariance matrix; The interceptor or aircraft historical state information is filtered using EKF, and the filtering is performed in reverse from the end of the measurement to the start of the vector measurement to obtain the backward filtering result and the backward filtering covariance matrix; The forward filtering result, the forward filtering covariance matrix, the backward filtering result and the backward filtering covariance matrix are combined to obtain the bidirectional filtering smoothing result and the bidirectional filtering covariance matrix.

5. The method for inferring the aerial target interception intention based on trajectory prediction according to claim 1, characterized in that: The improved closest encounter point function is used to process the interceptor's predicted trajectory and the aircraft's predicted trajectory. The specific method for calculating the dynamic threat degree of the predicted section is as follows: Calculate the relative position and relative speed of the interceptor to the aircraft based on the predicted trajectory of the interceptor and the predicted trajectory of the aircraft; Calculate the closest approach distance and closest approach time based on the interceptor's relative position and relative speed to the aircraft; Normalize the closest encounter point time, and obtain the interceptor time threat level value based on the normalized closest encounter point time; Normalize the closest encounter point distance, and obtain the interceptor distance threat level value based on the normalized closest encounter point distance; The interceptor time threat level value and the interceptor distance threat level value are corrected using the interceptor speed and the angle between the interceptor and the aircraft; According to the revised interceptor time threat level value and interceptor distance threat level value, the predicted segment dynamic threat level is obtained.

6. The method for inferring the aerial target interception intention based on trajectory prediction according to claim 1, characterized in that: The specific method of determining the interception intention matrix at each moment based on the dynamic threat level of the predicted segment and the intention inference rules is as follows: Preset intention inference rules and establish a judgment model based on the dynamic threat level of the prediction segment and intention inference rules; Input the predicted segment dynamic threat level into the judgment model to obtain the interception intention at each moment; According to the interception intention at each moment, the interception intention matrix at each moment is obtained.

7. The method for inferring the aerial target interception intention based on trajectory prediction according to claim 1, characterized in that: According to the interception intention matrix at each moment, the specific method to determine the center point of the interception encounter area, calculate the radius of the interception encounter area, and obtain the closest distance between the interceptor and the aircraft is as follows: According to the interception intention matrix at each moment, the corresponding interceptor and aircraft are extracted, and the interception encounter area is predicted based on the interception intention matrix of the interceptor and aircraft. The interception encounter area includes the set of points where the interceptor can hit the aircraft; According to the predicted trajectory of the interceptor and the predicted trajectory of the aircraft, the minimum distance between the interceptor and the aircraft is obtained.

8. The method for inferring aerial target interception intention based on trajectory prediction according to claim 1, characterized in that: The specific method for determining the center point and radius of the interception encounter area based on the closest distance between the interceptor and the aircraft is as follows: At the moment when the interceptor and the aircraft are closest, the midpoint of the line connecting the predicted track of the interceptor and the predicted track of the aircraft is the center point of the interception encounter area; The radius of the interception encounter zone is obtained based on the minimum distance between the interceptor and the aircraft and the predicted time for the interceptor and the aircraft to reach the minimum distance.

9. The air target interception intention inference system based on trajectory prediction is characterized by: The method for inferring the aerial target interception intention based on trajectory prediction according to claim 1 comprises the following steps: Information acquisition module, used to obtain historical status information of interceptors and aircraft; The filtering module is used to perform bidirectional filtering and smoothing on the historical state information of the interceptor and the aircraft, respectively, to obtain the state estimation result and covariance matrix after bidirectional filtering and smoothing; The pseudo two-sided data acquisition module is used to generate pseudo measurement data based on polynomial fitting according to the state estimation results and covariance matrix after bilateral filtering and smoothing; The predicted track acquisition module is used to perform augmented interactive multi-model-volumetric Kalman filtering on the pseudo-measurement data, and use the filtered pseudo-measurement data to update the historical state information of the interceptor and the aircraft to obtain the predicted track of the interceptor and the predicted track of the aircraft; The threat degree acquisition module is used to process the interceptor's predicted trajectory and the aircraft's predicted trajectory using an improved closest approach point function to calculate the dynamic threat degree of the predicted segment; The interception intention matrix acquisition module is used to determine the interception intention based on the dynamic threat level of the predicted segment and the intention inference rules, and obtain the interception intention matrix at each moment; The closest distance acquisition module is used to obtain the closest distance between the interceptor and the aircraft based on the interception intention matrix at each moment and the predicted trajectory of the interceptor and the aircraft; The interception encounter area acquisition module is used to determine the center point and radius of the interception encounter area according to the minimum distance between the interceptor and the aircraft.

Images