Air-to-ground target tracking method based on extended Kalman filtering

The air-to-ground target tracking method addresses inaccuracies in traditional passive direction finding by employing extended Kalman filtering with observation and state equations, improving positioning accuracy and stability through recursive estimation for both fixed and moving targets.

CN119063692APending Publication Date: 2024-12-03BEIJING SHANGSI DIGITAL INTELLIGENCE TECHNOLOGY CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411115441.6
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-14
Publication Date
2024-12-03

AI Technical Summary

Technical Problem

Traditional air-to-ground passive direction finding methods suffer from inaccurate positioning due to reliance on manual judgment, electromagnetic interference, and iterative processes prone to divergence, leading to local minima and positioning deviations.

Method used

An air-to-ground target tracking method using extended Kalman filtering, incorporating observation and state equations, linearization, and recursive estimation to improve accuracy and stability, employing passive and active detection equipment for fixed and moving targets.

Benefits of technology

Enhances positioning accuracy and tracking stability by utilizing a complex mathematical model with recursive estimation, ensuring robustness and consistency, applicable to both passive and active detection systems.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119063692A_ABST
    Figure CN119063692A_ABST
Patent Text Reader

Abstract

The invention belongs to the field of passive and active target positioning and tracking, and provides an air-to-ground target tracking method based on extended Kalman filtering, which comprises the following steps: S1, determining two positioning and tracking targets including an air-to-ground fixed radio radiation source and an air-to-ground moving target; s2, positioning and tracking of a fixed radio radiation source on the ground in the air are realized through passive detection equipment, and positioning and tracking of a moving target on the ground in the air are realized through active detection equipment; s3, based on linearized extended Kalman filtering minimum mean square estimation, providing accurate longitude and latitude positions of the target; according to the method, through iteration from multiple times of prediction to estimation, the result is more accurate, the consistency of state estimation is ensured, an estimated quantity can be estimated in real time by adopting a recursive calculation method and utilizing a computer, and a fixed or moving target can be reliably positioned and tracked in real time.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention belongs to the field of passive and active target positioning and tracking, in particular to an air-to-ground target tracking method based on extended Kalman filtering. Background Art

[0002] Traditional air-to-ground passive direction finding positioning uses the least squares estimation method of direction line crossing. Its disadvantages are that it uses a small amount of information and lacks mathematical modeling of the entire target observation system. Only when the direction finding is more accurate can a more satisfactory estimation performance be obtained. The quality of the direction line and whether it is included in the calculation range heavily rely on manual judgment, which is very arbitrary and easily affected by the operator. In addition, after the direction finder is installed, due to the coupling between the direction finding antenna and the body, the electromagnetic propagation environment does not meet the standard requirements, resulting in a decrease in the direction finding error relative to the factory indicators and a deterioration in the direction finding accuracy. In addition, the convergence of the least squares method is strongly dependent on the nonlinear strength of the cost function and the selection of the initial point of the target position. The iterative process is prone to divergence. The above inherent deficiencies make it easy for the iterative process to fail to converge or fall into a local minimum, resulting in deviations in the positioning results.

[0003] This method, through modeling of the airborne positioning and tracking system, proposes a set of solutions including observation equations, state equations and linearization measures, which can effectively improve the positioning accuracy and tracking stability, robustness and consistency of the source target and the degree of automation.

[0004] To this end, those skilled in the art have proposed an air-to-ground target tracking method based on extended Kalman filtering to solve the problems raised by the background technology. Summary of the invention

[0005] In order to solve the above technical problems, the present invention provides an air-to-ground target tracking method based on extended Kalman filtering to solve the problem in the prior art that the iterative process is prone to failure to converge or fall into a local minimum point, resulting in deviation in positioning results.

[0006] An air-to-ground target tracking method based on extended Kalman filtering comprises the following steps:

[0007] S1: Determine the positioning and tracking target. There are two types of positioning targets, including air-to-ground fixed radio radiation sources and air-to-ground moving targets;

[0008] S2: Use passive detection equipment to realize the positioning and tracking of fixed radio radiation sources on the ground from the air, and use active detection equipment to realize the positioning and tracking of moving targets on the ground from the air;

[0009] S3: Based on the linearized extended Kalman filter minimum mean square estimation, the precise longitude and latitude position of the target is given to ensure the positioning accuracy and tracking stability, robustness and consistency of the target.

[0010] Preferably, the passive detection equipment is used to realize the positioning and tracking of the fixed radio radiation source on the ground by the airborne aircraft (aircraft). First, the linearization factor is extracted. Generally, the nonlinear function f(x) is expressed as follows:

[0011] f(x) is called a correctable nonlinear function. The target direction obtained by the radio direction finder belongs to this type of nonlinear function. After using the earth coordinate system, the linearization factor can be obtained:

[0012]

[0013] in is the estimated value of target longitude and latitude, (λ0,L0) is the current position of the aircraft, and β is the heading.

[0014] Preferably, the observation equation and the state equation are established, and the following model is established for the actual situation of aerial maneuver direction finding and positioning:

[0015] X k+1 =φ k X k +Γ k W k ;

[0016] β k+1 =h k+1 (x k+1 )V k+1 ;

[0017] Where X is the target coordinate, X = [λ, L] -1 , β is the orientation, which is a function of the target coordinates, Γ k ,φ k is the identity matrix h is the linearization factor vector, W and V are noises, and their variance matrices are Q(k) and R(k), which means that Q(k) is non-negative definite, R(k) is positive definite, and X k , W k 、V k are unrelated to each other, where:

[0018] E(W k )=0

[0019] E(V k )=0

[0020] Preferably, the linearized EKF recursion equation can be derived from the above model as follows:

[0021]

[0022] P k+1 / k+1 =[IK k+1 g k+1 ]P k+1 / k ;

[0023]

[0024] Where P is the prediction variance matrix, K is the intermediate variable, and g is the Kalman filter gain.

[0025] Preferably, due to the existence of system errors, when the estimated target position is given, it is usually not a point, but a probability ellipse. Let the rotation angle γ and the major and minor axes of the probability ellipse be And determined by the following formula:

[0026]

[0027] P is the confidence level, They come from the elements on the diagonal of the estimated error variance matrix.

[0028] Preferably, the filter estimation needs to have initial values ​​to start, including the initial position, the initial values ​​of the direction finding error variance matrix R(k) and the state error variance matrix Q(k). The target initial position is obtained by the direction indication obtained by the initial n (n≥2) direction finding and then by the least squares estimation of the angle error.

[0029] Preferably, the initial value of the state variance matrix is ​​a symmetric matrix composed of the error variance empirical values ​​of longitude and latitude on the diagonal, and has the following form:

[0030]

[0031] QLo and QLa are the longitude and latitude error variances, respectively.

[0032] Preferably, the initial value of the observed variance is obtained from the test data obtained during the test of the direction finder after installation, by calculating the χ value of the test data. 2 The results of hypothesis test analysis show that the airborne direction finding error conforms to the normal distribution in each working frequency band. The direction finder is divided into multiple frequency bands such as ultra-short wave and microwave according to the frequency band. Different frequency bands use different antennas and signal receiving equipment, so there are different direction finding errors. The above-mentioned frequency points are tested for normal distribution variance through asymmetric bilateral quantile method to obtain the variance corresponding to each frequency band. In the actual estimation calculation, the corresponding variance matrix is ​​called according to the frequency band where the monitoring frequency is located.

[0033] Preferably, the estimation error threshold is used. If two consecutive estimation values ​​are less than the preset threshold, the iteration is considered to have converged, and the calculation is terminated. The final result is used as the estimated output. If one recursion cannot reach convergence, an iterative method is used:

[0034]

[0035] where φ k is the prediction variance matrix.

[0036] Preferably, active detection equipment is used to realize the positioning and tracking of ground moving targets in the air. The target position calculation method in the state equation can be used to initially locate the ground moving targets using active detection equipment (such as radar, photoelectric detector, etc.). The specific calculation method is as follows:

[0037] The output results of the general three-coordinate radar include the target azimuth α (same as the direction indication given by passive direction finding), straight-line distance c and altitude. The longitude and latitude of the target can be calculated through the longitude and latitude given by the carrier's own position sensor (such as GPS equipment, inertial navigation, Beidou receiver), and this result is used as the previous state input of the state equation;

[0038] In the following spherical triangular coordinate system, let T represent the target position and k represent the position of the carrier: a = (λ T -λ k )cosL T ;

[0039] b=L T -L k =c*cos(α);

[0040] The height of the aircraft is h, the straight-line distance between the aircraft and the target is s, the arc length between the ground projection point of the aircraft and the target sphere is L, and the secant length is l, then we can deduce:

[0041] L T =L k +c*cos(α);

[0042] λ T =sqr(c 2 -b 2 ) / consL T +λ k .

[0043] Compared with the prior art, the present invention has the following beneficial effects:

[0044] 1. Compared with the simple least squares method, the method of the present invention uses a more complex and accurate mathematical model. In addition to the observation equation, the concept of state space is introduced, and the system is described by the state equation. The estimated quantity is taken as the state. Through multiple iterations from prediction to estimation, the result is more accurate and the consistency of state estimation is guaranteed. The basic statistical characteristics (first and second order statistical characteristics) of the estimated quantity and the measured value are included as the initial value of the iterative estimation. This statistical characteristic is based on the statistics obtained by the measurement error distribution hypothesis test and the normal distribution variance test in the real machine environment, and is therefore more accurate.

[0045] 2. The method of the present invention adopts a recursive calculation method, and the estimated quantity can be estimated in real time by a computer; with a slight improvement, it can be applied to the positioning and tracking of ground moving targets by airborne active detection equipment; the present invention is applicable to passive detection systems (such as interferometer direction finding and positioning systems, time difference direction finding and positioning systems, Doppler direction finding and positioning systems, etc.), and their direction finding results can be used as input of the present method, and it is also applicable to active direction finding and positioning systems for ground moving targets (such as various radars, photoelectric detectors), and their outputs can be used as input of the present method. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 The linearization factor graph can be obtained after the earth coordinate system of the present invention;

[0047] Figure 2 A spherical triangular coordinate system diagram for positioning and tracking a ground moving target according to the present invention;

[0048] Figure 3 This is a flow chart of one iteration of the Kalman filter of the present invention. DETAILED DESCRIPTION

[0049] The following embodiments of the present invention are described in further detail in conjunction with the accompanying drawings and examples. The following examples are used to illustrate the present invention, but are not intended to limit the scope of the present invention.

[0050] Reference Figure 1-Figure 3 The present invention provides an air-to-ground target tracking method based on extended Kalman filtering, comprising the following steps:

[0051] S1: Determine the positioning and tracking target. There are two types of positioning targets, including air-to-ground fixed radio radiation sources and air-to-ground moving targets;

[0052] S2: Use passive detection equipment to realize the positioning and tracking of fixed radio radiation sources on the ground from the air, and use active detection equipment to realize the positioning and tracking of moving targets on the ground from the air;

[0053] S3: Based on the linearized extended Kalman filter minimum mean square estimation, the precise longitude and latitude position of the target is given to ensure the positioning accuracy and tracking stability, robustness and consistency of the target.

[0054] As can be seen from the above, this method is used in the passive direction finding and positioning scenario based on the radio radiation source. Passive direction finding and positioning is an important part of radio technical reconnaissance and electronic countermeasures due to its good concealment and long range. It is often used in military and civilian radio monitoring and electronic countermeasures. Airborne air-to-ground passive direction finding and positioning tracking is a common method for radio reconnaissance and direction finding. Since the carrier aircraft can continuously find the direction of the radio radiation source at multiple points during high-altitude flight and perform cross-triangulation to determine the longitude and latitude of the ground radiation source, compared with the ground direction finding station, it can overcome the problem of short detection distance caused by the short line of sight distance during radio straight-line propagation caused by the curvature of the earth, and large direction finding error due to the influence of ground clutter.

[0055] The method can be applied to passive direction finding and positioning and active detection and positioning systems of large and small aircraft, such as electronic reconnaissance aircraft modified from civil passenger aircraft or transport aircraft, electronic warfare aircraft modified from fighter aircraft, or electronic warfare aircraft modified from various strategic / tactical unmanned reconnaissance aircraft.

[0056] Specifically, the passive detection equipment is used to realize the positioning and tracking of the fixed radio radiation source on the ground by the airborne aircraft (aircraft). First, the linearization factor is extracted. Generally, the nonlinear function f(x) is expressed as follows:

[0057] f(x) is called a correctable nonlinear function. The target direction obtained by the radio direction finder belongs to this type of nonlinear function. After using the earth coordinate system, the linearization factor can be obtained:

[0058]

[0059] in is the estimated value of target longitude and latitude, (λ0,L0) is the current position of the aircraft, and β is the heading.

[0060] From the above, we can see that the passive positioning and tracking problem is essentially a nonlinear filtering problem due to the lack of distance measurement information and the inherent nonlinear relationship between angle and position coordinates. It can be solved by extended Kalman filtering. For nonlinear filtering problems, the performance of suboptimal filter estimates constructed using linearization techniques is often related to the linearization accuracy. In the EKF algorithm of passive direction finding and positioning, the linearization error caused by the first-order Taylor expansion directly leads to the inaccuracy of the predicted value, resulting in poor positioning effect. The basic way to improve its filtering performance is to seek a more accurate linearization method. The linearization concept proposed in this method extracts the linearization factor in the nonlinear measurement equation, obtains a more accurate linearization form, and avoids the degradation of filtering performance caused by the use of Taylor expansion.

[0061] Specifically, the observation equation and state equation are established, and the following model is established for the actual situation of aerial maneuver direction finding and positioning:

[0062] X k+1 =φ k X k +Γ k W k ;

[0063] β k+1 =h k+1 (x k+1 )V k+1 ;

[0064] Where X is the target coordinate, X = [λ, L] -1 , β is the orientation, which is a function of the target coordinates, Γ k ,φ k is the identity matrix h is the linearization factor vector, W and V are noises, and their variance matrices are Q(k) and R(k), which means that Q(k) is non-negative definite, R(k) is positive definite, and X k , W k 、V k are unrelated to each other, where:

[0065] E(W k )=0

[0066] E(V k )=0

[0067] Specifically, the linearized EKF recursion equation can be deduced from the above model as follows:

[0068]

[0069] P k+1 / k+1 =[IK k+1 gk+1 ]P k+1 / k ;

[0070]

[0071] Where P is the prediction variance matrix, K is the intermediate variable, and g is the Kalman filter gain.

[0072] Specifically, due to the existence of systematic errors, when the estimated target position is given, it is usually not a point but a probability ellipse. Let the rotation angle γ and the major and minor axes of the probability ellipse be And determined by the following formula:

[0073]

[0074] P is the confidence level, They come from the elements on the diagonal of the estimated error variance matrix.

[0075] Specifically, the filter estimation needs to have initial values ​​to start, including the initial position, the initial values ​​of the direction finding error variance matrix R(k) and the state error variance matrix Q(k). The initial position of the target is obtained by the direction indication obtained by the initial n (n≥2) direction finding and then by the least squares estimation of the angle error.

[0076] Specifically, the initial value of the state variance matrix is ​​a symmetric matrix composed of the error variance empirical values ​​of longitude and latitude on the diagonal, and the form is as follows:

[0077]

[0078] QLo and QLa are the longitude and latitude error variances, respectively.

[0079] Specifically, the initial value of the observed variance is obtained from the test data obtained during the test of the direction finder after installation, and the χ 2 The results of hypothesis test analysis show that the airborne direction finding error conforms to the normal distribution in each working frequency band. The direction finder is divided into multiple frequency bands such as ultra-short wave and microwave according to the frequency band. Different frequency bands use different antennas and signal receiving equipment, so there are different direction finding errors. The above-mentioned frequency points are tested for normal distribution variance through asymmetric bilateral quantile method to obtain the variance corresponding to each frequency band. In the actual estimation calculation, the corresponding variance matrix is ​​called according to the frequency band where the monitoring frequency is located.

[0080] Specifically, the error threshold is estimated. If two consecutive estimated values ​​are less than the preset threshold, the iteration is considered to have converged, the calculation is terminated, and the final result is used as the estimated output. If one recursion cannot reach convergence, an iterative method is used:

[0081]

[0082] where φ k is the prediction variance matrix.

[0083] From the above, we can see that in the iterative calculation process, the variance matrix may be non-positive definite due to rounding errors. In this case, it means that the threshold may be too small, and the filtering iteration can be terminated. By repeatedly calculating the iterative method, the difference between two adjacent estimates can be made less than the threshold value, which can be considered convergence and the iteration is terminated. When the number of iterations in the iterative process is too many, dynamically calculating the error variance matrix and the state noise variance matrix can increase the convergence speed and greatly reduce the number of iterations.

[0084] Specifically, active detection equipment is used to realize the positioning and tracking of ground moving targets in the air. The target position calculation method in the state equation can be used to initially locate the ground moving targets using active detection equipment (such as radar, photoelectric detector, etc.). The specific calculation method is as follows:

[0085] The output results of the general three-coordinate radar include the target azimuth α (same as the direction indication given by passive direction finding), straight-line distance c and altitude. The longitude and latitude of the target can be calculated through the longitude and latitude given by the carrier's own position sensor (such as GPS equipment, inertial navigation, Beidou receiver), and this result is used as the previous state input of the state equation;

[0086] In the following spherical triangular coordinate system, let T represent the target position and k represent the position of the carrier: a = (λ T -λ k )cosL T ;

[0087] b=L T -L k =c*cos(α);

[0088] The height of the aircraft is h, the straight-line distance between the aircraft and the target is s, the arc length between the ground projection point of the aircraft and the target sphere is L, and the secant length is l, then we can deduce:

[0089] L T =L k +c*cos(α);

[0090] λ T =sqr(c 2 -b 2 ) / consL T +λ k .

[0091] From the above, we can see that the initial positioning of ground moving targets is carried out through active detection equipment (such as radar, photoelectric detectors, etc.).

[0092] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example" or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, unless they are contradictory.

[0093] In the drawings of the embodiments disclosed in the present invention, only the structures related to the embodiments disclosed in the present invention are involved, and other structures can refer to the general design. In the absence of conflict, the same embodiment and different embodiments of the present invention can be combined with each other.

[0094] Although the present invention has been described in detail with reference to the aforementioned embodiments, it is still possible for those skilled in the art to modify the technical solutions described in the aforementioned embodiments, or to make equivalent substitutions for some of the technical features therein. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. An air-to-ground target tracking method based on extended Kalman filtering, characterized in that: The steps include: S1: Determine the positioning and tracking target. There are two types of positioning targets, including air-to-ground fixed radio radiation sources and air-to-ground moving targets; S2: Use passive detection equipment to realize the positioning and tracking of fixed radio radiation sources on the ground from the air, and use active detection equipment to realize the positioning and tracking of moving targets on the ground from the air; S3: Based on the linearized extended Kalman filter minimum mean square estimation, the precise longitude and latitude position of the target is given to ensure the positioning accuracy and tracking stability, robustness and consistency of the target.

2. The air-to-ground target tracking method based on extended Kalman filtering according to claim 1, characterized in that: To realize the positioning and tracking of fixed radio radiation sources on the ground by airborne aircraft (aircraft) through passive detection equipment, the linearization factor is first extracted. Generally, the nonlinear function f(x) is expressed as follows: f(x) is called a correctable nonlinear function. The target direction obtained by the radio direction finder belongs to this type of nonlinear function. After using the earth coordinate system, the linearization factor can be obtained: in is the estimated value of target longitude and latitude, (λ0,L0) is the current position of the aircraft, and β is the heading.

3. The air-to-ground target tracking method based on extended Kalman filtering according to claim 2, characterized in that: The observation equation and state equation are established, and the following model is established for the actual situation of aerial maneuver direction finding and positioning: X k+1 =φ k X k +C k W k ; β k+1 =h k+1 (x k+1 )V k+1 ; Where X is the target coordinate, X = [λ, L] -1 , β is the orientation, which is a function of the target coordinates, Γ k ,φ k is the identity matrix h is the linearization factor vector, W and V are noises, and their variance matrices are Q(k) and R(k), which means that Q(k) is non-negative definite, R(k) is positive definite, and X k , W k 、V k Unrelated to each other, where: E(W k )=0 E(V k )=0 4. The air-to-ground target tracking method based on extended Kalman filtering according to claim 3, characterized in that: The linearized EKF recursion equation can be deduced from the above model as follows: Where P is the prediction variance matrix, K is the intermediate variable, and g is the Kalman filter gain.

5. The air-to-ground target tracking method based on extended Kalman filtering according to claim 4, characterized in that: Due to the existence of systematic errors, the estimated target position is usually not a point, but a probability ellipse. Suppose the rotation angle γ and the major and minor axes of the probability ellipse are And determined by the following formula: P is the confidence level, They come from the elements on the diagonal of the estimated error variance matrix.

6. The air-to-ground target tracking method based on extended Kalman filtering according to claim 5, characterized in that: The filter estimation needs initial values ​​to start, including the initial position, the initial values ​​of the direction finding error variance matrix R(k) and the state error variance matrix Q(k). The initial position of the target is obtained by the direction indication obtained by the initial n (n≥2) direction finding and then by the least squares estimation of the angle error.

7. The air-to-ground target tracking method based on extended Kalman filtering according to claim 6, characterized in that: The initial value of the state variance matrix is ​​a symmetric matrix composed of the error variance empirical values ​​of longitude and latitude on the diagonal, and the form is as follows: QLo and QLa are the longitude and latitude error variances, respectively.

8. The air-to-ground target tracking method based on extended Kalman filtering according to claim 7, characterized in that: The initial value of the observed variance is obtained from the test data obtained during the test of the direction finder after installation. 2 The results of hypothesis test analysis show that the airborne direction finding error conforms to the normal distribution in each working frequency band. The direction finder is divided into multiple frequency bands such as ultra-short wave and microwave according to the frequency band. Different frequency bands use different antennas and signal receiving equipment, so there are different direction finding errors. The above-mentioned frequency points are tested for normal distribution variance through asymmetric bilateral quantile method to obtain the variance corresponding to each frequency band. In the actual estimation calculation, the corresponding variance matrix is ​​called according to the frequency band where the monitoring frequency is located.

9. The air-to-ground target tracking method based on extended Kalman filtering according to claim 8, characterized in that: Estimation error threshold. If two consecutive estimation values ​​are less than the preset threshold, the iteration is considered to have converged, the calculation is terminated, and the final result is used as the estimated output. If one recursion cannot reach convergence, an iterative method is used: where φ k is the prediction variance matrix.

10. The air-to-ground target tracking method based on extended Kalman filtering according to claim 1, characterized in that: Active detection equipment is used to realize the positioning and tracking of ground moving targets in the air. The target position calculation method in the state equation can be used to initially locate the ground moving targets using active detection equipment (such as radar, photoelectric detector, etc.). The specific calculation method is as follows: The output results of the general three-coordinate radar include the target azimuth α (same as the direction indication given by passive direction finding), straight-line distance c and altitude. The longitude and latitude of the target can be calculated through the longitude and latitude given by the carrier's own position sensor (such as GPS equipment, inertial navigation, Beidou receiver), and this result is used as the previous state input of the state equation; In the following spherical trigonometric coordinate system, let T represent the target position and k represent the carrier position: a=(λ T -λ k )whatsL T ; b=L T -L k = c*cos(α); The height of the aircraft is h, the straight-line distance between the aircraft and the target is s, the arc length between the ground projection point of the aircraft and the target sphere is L, and the secant length is l, then we can deduce: L T =L k +c*cos(α); l T =sqr(c 2 -b 2 ) / consL T +λ k 。