A pre-observation overall fusion tracking method for Doppler measurement

By constructing a motion model under the Cartesian coordinate system and using the linear fusion of weighted least squares normal, the error problem caused by nonlinear transformation in Doppler radar target tracking is solved, and the tracking accuracy and performance are improved.

CN117784064BActive Publication Date: 2024-11-08KUNMING UNIV OF SCI & TECH
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202311825135.7
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2023-12-28
Publication Date
2024-11-08
Estimated Expiration
2043-12-28

AI Technical Summary

Technical Problem

In target tracking, Doppler radar has increased the complexity of the tracking system due to the nonlinear transformation between the polar coordinate system and the Cartesian coordinate system, and the algorithm's estimation performance has decreased. The existing nonlinear filtering method has a large error when dealing with strong nonlinearity. , and it is difficult to select parameters.

Method used

Using the pre-observation overall fusion tracking method, a uniform linear motion and uniform acceleration linear motion model is constructed under the Cartesian coordinate system. By linearly fusion of Doppler measurement and position measurement by weighted least squares method, nonlinear filtering can avoid the estimation accuracy Influence.

Benefits of technology

It effectively improves the target tracking accuracy, reduces the calculation amount, overcomes the strong nonlinearity between Doppler measurement and target state, and improves the tracking performance.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN117784064B_ABST
    Figure CN117784064B_ABST
Patent Text Reader

Abstract

The present invention discloses a pre-observation overall fusion tracking method for Doppler measurement, which relates to the technical field of Doppler radar target tracking. Specifically, it includes: obtaining the target state and covariance of the uniform linear motion and uniform accelerated linear motion models in the Cartesian coordinate system and performing one-step prediction; one-step prediction of the target state and covariance of the obtained uniform linear motion and uniform accelerated linear motion models to calculate the measurement prediction and covariance; obtaining the Doppler radar measurement value and converting the Doppler radar measurement value to the Cartesian coordinate, and calculating the measurement conversion error covariance; completing the update of the target tracking state and covariance at the current moment in the uniform linear motion and uniform accelerated linear motion models through weighted least squares linear fusion; and looping the above steps until the target tracking is completed. The present invention reasonably utilizes the Doppler measurement information and avoids the influence of nonlinear filtering on the estimation accuracy, which can effectively improve the target tracking performance.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention belongs to the technical field of Doppler radar target tracking, and in particular relates to a pre-observation overall fusion tracking method for Doppler measurement. Background Art

[0002] Since Doppler radar can denoise in the frequency domain, its anti-interference and anti-clutter capabilities are stronger than ordinary radar. Therefore, Doppler radar is widely used in target tracking, battlefield reconnaissance, traffic control and other fields.

[0003] When tracking a target, in addition to obtaining the distance and azimuth of the target, Doppler radar can also obtain the target's Doppler measurement, i.e., radial velocity. Doppler measurement is the only measurement value among all Doppler radar measurements that contains target velocity information. Studies have shown that making full use of Doppler measurement can effectively improve target tracking accuracy. However, when tracking moving targets with the help of radar measurement, physical quantities in the Cartesian coordinate system are usually used to characterize the target's motion state, while the measurements provided by Doppler radar, i.e., distance, azimuth, and Doppler measurements, are all in the polar coordinate system. The coordinate system transformation introduces nonlinearity that is difficult to eliminate between the target's motion state and the radar measurement. The nonlinearity of the measurement will undoubtedly increase the complexity of the tracking system and lead to a decrease in the algorithm estimation performance. Therefore, solving the target state and measurement nonlinearity is the main problem of target tracking.

[0004] To solve the problem of radar target tracking with Doppler measurement, the following nonlinear filtering is usually used:

[0005] Sequential Extended Kalman Filter (SEKF) first uses the measurement conversion Kalman filter to linearly filter the position measurement, and then uses EKF to process the Doppler measurement. However, since the EKF discards the higher-order terms above the second order during the linearization process, the error caused by strong nonlinearity is larger.

[0006] Sequential Unscented Kalman Filter (SUKF) first decorrelates the distance and Doppler measurements with measurement error correlation, and then uses UKF to sequentially process the position measurement and pseudo-Doppler measurement. The disadvantage is that the problem of selecting the parameters of the unscented Kalman filter has not been fully solved.

[0007] Converted Measurement Kalman Filter with Doppler Measurement (CMKFRR) This method converts the range, angle and Doppler measurements to position and velocity in a Cartesian coordinate system. This conversion is unbiased and consistent, but the measurement conversion requires prior knowledge of the distribution of the tangential velocity.

[0008] Although these methods have improved the target tracking effect of Doppler radar to a certain extent, there is still room for improvement in tracking accuracy and computational complexity.

[0009] Therefore, it is an urgent problem for those skilled in the art to propose a pre-observation overall fusion tracking method for Doppler measurement to solve the difficulties existing in the prior art. Summary of the invention

[0010] The present invention proposes a pre-observation overall fusion tracking method for Doppler measurement, which reasonably utilizes Doppler measurement information and avoids the influence of nonlinear filtering on estimation accuracy, and can effectively improve target tracking performance.

[0011] In order to achieve the above object, the present invention adopts the following technical solution:

[0012] A pre-observation overall fusion tracking method for Doppler measurement obtains the posterior estimated mean and covariance of the target state at time k. The specific steps of filtering from time k to time k+1 are as follows:

[0013] S1: Obtain target motion data, construct a uniform linear motion model and a uniform accelerated linear motion model, and perform one-step prediction of the target state and covariance of the uniform linear motion model and the uniform accelerated linear motion model in a Cartesian coordinate system;

[0014] S2: Calculate the measurement prediction and its covariance by one-step prediction of the target state and covariance of the uniform linear motion model and the uniform accelerated linear motion model obtained in S1;

[0015] S3: Obtain target Doppler radar measurement values ​​and convert the Doppler radar measurement values ​​to Cartesian coordinates, and calculate the measurement conversion error covariance;

[0016] S4: In the uniform linear motion model and the uniform accelerated linear motion model, the target tracking state and covariance are updated at the current moment by linearly fusing the calculation results in S2 and S3 using the weighted least squares method;

[0017] S5: Execute steps S1-S4 repeatedly until the target tracking is completed.

[0018] In the above method, optionally, the specific content of S1 is:

[0019] The specific process of one-step prediction of the target state and covariance of the uniform linear motion model in Cartesian coordinates is:

[0020] X cv (k+1,k)=Φ cv (k)X cv (k,k);

[0021] P cv (k+1,k)=Φ cv (k) cv (k,k)Φ cv(k) T +Γ cv (k)Q cv (k)Γ cv (k) T ;

[0022]

[0023]

[0024] In the formula, represents the state estimate of the target at time k, x, y and is the position and velocity of the target; X cv (k+1, k) represents the state prediction of the target at time k+1; cv (k, k) represents the estimated covariance of the target at time k; cv (k+1,k) represents the one-step prediction covariance of the target at time k+1; Φ cv (k) is the state transfer matrix, Γ cv (k) is the process noise driving matrix, Q cv (k) is process noise;

[0025] The specific process of one-step prediction of the target state and covariance of the uniformly accelerated linear motion model under Cartesian coordinates is:

[0026] X ca (k+1,k)=Φ ca (k)X ca (k,k);

[0027] P ca (k+1,k)=Φ ca (k) ca (k,k)Φ ca (k) T +Γ ca (k)Q ca (k)Γ ca (k) T ;

[0028]

[0029]

[0030] In the formula, represents the state estimate of the target at time k, is the acceleration of the target; X ca (k+1, k) represents the state prediction of the target at time k+1; ca (k, k) represents the estimated covariance of the target at time k; ca(k+1,k) represents the one-step prediction covariance of the target at time k+1; Φ ca (k) is the state transfer matrix, Γ ca (k) is the process noise driving matrix, Q ca (k) is the process noise.

[0031] In the above method, optionally, the specific formula for calculating the measurement prediction and its covariance in S2 is as follows:

[0032]

[0033]

[0034]

[0035]

[0036]

[0037] In the formula, r c ,θ c , and They are distance prediction, angle prediction, radial velocity prediction and tangential velocity prediction respectively; and is the corresponding covariance; is the state one-step prediction value, P(k+1,k) is the covariance one-step prediction value, P 2×2 (k+1,k) is the first 2 rows and 2 columns of P(k+1,k).

[0038] The above method is optional. The specific content in S3 is:

[0039] Convert the target Doppler measurements to Cartesian coordinates:

[0040]

[0041] Calculate the expectation for the above formula:

[0042]

[0043]

[0044]

[0045]

[0046] The unbiased measurement is converted to:

[0047]

[0048] In the formula, xm ,y m and is the measurement conversion value in the Cartesian coordinate system. The above formula is the measurement conversion of the uniform linear motion model and the uniformly accelerated linear motion model.

[0049] In the above method, optionally, the measurement conversion error covariance is specifically expressed as:

[0050]

[0051] In the formula,

[0052]

[0053]

[0054]

[0055]

[0056]

[0057]

[0058]

[0059]

[0060]

[0061]

[0062] In the above method, the specific content of S4 is optional: constructing a linear regression model of the uniform linear motion model as follows:

[0063]

[0064] Let the above matrix equation be recorded as Ax=b+e, and the covariance Var(e) of e be recorded as V, then we have:

[0065]

[0066]

[0067] Status Update: X cv (k+1,k+1)=(A T V -1 A) -1 A T V -1 b;

[0068] Covariance update: Ρcv (k+1,k+1)=(A T V -1 A) -1 ;

[0069] Among them, x, y and is the a posteriori estimate of position and velocity in Cartesian coordinates; is the one-step prediction value of the state; x m ,y m and is the measurement conversion value in the Cartesian coordinate system; is the one-step prediction covariance; is the measurement transformation covariance;

[0070] In the above method, optionally, the specific content of S4 is:

[0071] The linear regression model for constructing the uniformly accelerated linear motion model is:

[0072]

[0073]

[0074] Let the above matrix equation be recorded as Ax=b+e, and the covariance Var(e) of e be recorded as V, then we have:

[0075]

[0076]

[0077] Status Update: X ca (k+1,k+1)=(A T V -1 A) -1 A T V -1 b;

[0078] Covariance update: Ρ ca (k+1,k+1)=(A T V -1 A) -1 ;

[0079] Among them, x, y, and is the a posteriori estimate of position, velocity, and acceleration in the Cartesian coordinate system; and is the one-step prediction value of the state; x m ,y m , and is the measurement conversion value in the Cartesian coordinate system; is the one-step prediction covariance, obtained from S1; is the measurement transformation covariance.

[0080] It can be seen from the above technical solution that, compared with the prior art, the pre-observation overall fusion tracking method for Doppler measurement of the present invention has the following beneficial effects:

[0081] The present invention adopts a completely different technical idea from the previous algorithms, converting the position measurement and Doppler measurement to the Cartesian coordinate system, making the state and observation linearly related, and finally obtaining the posterior estimate of the target state at each moment through weighted least squares linear fusion. It overcomes the strong nonlinearity between Doppler measurement and target state, can track the target with a smaller amount of calculation, and can greatly improve the target tracking accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0082] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.

[0083] Figure 1 A flow chart of a pre-observation overall fusion tracking method for Doppler measurement disclosed in the present invention;

[0084] Figure 2(a) is a graph showing the error curves of different position tracking algorithms under the uniform linear motion model;

[0085] Figure 2(b) is a graph showing the error curves of different algorithms for tracking velocity under the uniform linear motion model;

[0086] Figure 3(a) is a graph showing the error curves of different position tracking algorithms under the uniformly accelerated linear motion model;

[0087] Figure 3(b) is the error curve of different tracking speed algorithms under the uniformly accelerated linear motion model. DETAILED DESCRIPTION

[0088] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0089] In this application, relational terms such as first and second, etc. are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. The terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of more restrictions, the elements defined by the sentence "comprise one..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0090] Reference Figure 1 As shown in FIG. 1 , a pre-observation overall fusion tracking method for Doppler measurement is used to obtain the posterior estimated mean and covariance of the target state at time k. The specific steps of filtering from time k to time k+1 are as follows:

[0091] S1: Obtain target motion data, construct a uniform linear motion model and a uniform accelerated linear motion model, and perform one-step prediction of the target state and covariance of the uniform linear motion model and the uniform accelerated linear motion model in a Cartesian coordinate system;

[0092] S2: Calculate the measurement prediction and its covariance by one-step prediction of the target state and covariance of the uniform linear motion model and the uniform accelerated linear motion model obtained in S1;

[0093] S3: Obtain target Doppler radar measurement values ​​and convert the Doppler radar measurement values ​​to Cartesian coordinates, and calculate the measurement conversion error covariance;

[0094] S4: In the uniform linear motion model and the uniform accelerated linear motion model, the target tracking state and covariance are updated at the current moment by linearly fusing the calculation results in S2 and S3 using the weighted least squares method;

[0095] S5: Execute steps S1-S4 repeatedly until the target tracking is completed.

[0096] Furthermore, the specific content of S1 is:

[0097] The specific process of one-step prediction of the target state and covariance of the uniform linear motion model in Cartesian coordinates is:

[0098] X cv (k+1,k)=Φ cv (k)X cv (k,k);

[0099] P cv(k+1,k)=Φ cv (k) cv (k,k)Φ cv (k) T +Γ cv (k)Q cv (k)Γ cv (k) T ;

[0100]

[0101]

[0102] In the formula, represents the state estimate of the target at time k, x, y and is the position and velocity of the target; X cv (k+1, k) represents the state prediction of the target at time k+1; cv (k, k) represents the estimated covariance of the target at time k; cv (k+1,k) represents the one-step prediction covariance of the target at time k+1; Φ cv (k) is the state transfer matrix, Γ cv (k) is the process noise driving matrix, Q cv (k) is process noise;

[0103] The specific process of one-step prediction of the target state and covariance of the uniformly accelerated linear motion model under Cartesian coordinates is:

[0104] X ca (k+1,k)=Φ ca (k)X ca (k,k);

[0105] P ca (k+1,k)=Φ ca (k) ca (k,k)Φ ca (k) T +Γ ca (k)Q ca (k)Γ ca (k) T ;

[0106]

[0107]

[0108] In the formula, represents the state estimate of the target at time k, is the acceleration of the target; X ca (k+1, k) represents the state prediction of the target at time k+1;ca (k, k) represents the estimated covariance of the target at time k; ca (k+1,k) represents the one-step prediction covariance of the target at time k+1; Φ ca (k) is the state transfer matrix, Γ ca (k) is the process noise driving matrix, Q ca (k) is the process noise.

[0109] Furthermore, the specific formula for calculating the measurement prediction and its covariance in S2 is as follows:

[0110]

[0111]

[0112]

[0113]

[0114]

[0115] In the formula, r c ,θ c , and They are distance prediction, angle prediction, radial velocity prediction and tangential velocity prediction respectively; and is the corresponding covariance; is the state one-step prediction value, P(k+1,k) is the covariance one-step prediction value, P 2×2 (k+1,k) is the first 2 rows and 2 columns of P(k+1,k).

[0116] Furthermore, the specific content of S3 is:

[0117] Convert the target Doppler measurements to Cartesian coordinates:

[0118]

[0119] Calculate the expectation for the above formula:

[0120]

[0121]

[0122]

[0123]

[0124] The unbiased measurement is converted to:

[0125]

[0126] In the formula, x m ,y m and is the measurement conversion value in the Cartesian coordinate system. The above formula is the measurement conversion of the uniform linear motion model and the uniformly accelerated linear motion model.

[0127] Furthermore, the measurement conversion error covariance is specifically expressed as:

[0128]

[0129] In the formula,

[0130]

[0131]

[0132]

[0133]

[0134]

[0135]

[0136]

[0137]

[0138]

[0139]

[0140] Furthermore, the specific content of S4 is:

[0141] The linear regression model for constructing the uniform linear motion model is:

[0142]

[0143] Let the above matrix equation be recorded as Ax=b+e, and the covariance Var(e) of e be recorded as V, then we have:

[0144]

[0145]

[0146] Status Update: X cv (k+1,k+1)=(A T V -1 A) -1 A T V-1 b;

[0147] Covariance update: Ρ cv (k+1,k+1)=(A T V -1 A) -1 ;

[0148] Among them, x, y and is the a posteriori estimate of position and velocity in Cartesian coordinates; is the one-step prediction value of the state; x m ,y m and is the measurement conversion value in the Cartesian coordinate system; is the one-step prediction covariance; is the measurement transformation covariance;

[0149] Furthermore, the specific content of S4 is:

[0150] The linear regression model for constructing the uniformly accelerated linear motion model is:

[0151]

[0152]

[0153] Let the above matrix equation be recorded as Ax=b+e, and the covariance Var(e) of e be recorded as V, then we have:

[0154]

[0155]

[0156] Status Update: X ca (k+1,k+1)=(A T V -1 A) -1 A T V -1 b;

[0157] Covariance update: Ρ ca (k+1,k+1)=(A T V -1 A) -1 ;

[0158] Among them, x, y, and is the a posteriori estimate of position, velocity, and acceleration in the Cartesian coordinate system; and is the one-step prediction value of the state; x m ,y m , and is the measurement conversion value in the Cartesian coordinate system; is the one-step prediction covariance, obtained from S1; is the measurement transformation covariance.

[0159] Example 1: Using the radar target tracking method of the present invention, the target is tracked to move in a uniform linear motion in a two-dimensional space. The initial position of the target is (10km, 10km), the initial velocity is (8m / s, 10m / s), and the Doppler radar located at the origin of the coordinates provides the distance, azimuth and Doppler measurement of the target with a sampling period of 1s. The method proposed by the present invention is compared with the SEKF, SUKF and CMKFRR algorithms with Doppler measurement. The standard deviations of their measurement noise are σ r =100m,σ θ =1°, The correlation coefficient ρ = 0.5, and the simulation results of the root mean square error of position and velocity of different algorithms are shown in Figure 2(a) and Figure 2(b).

[0160] Example 2: Using the radar target tracking method of the present invention, the target is tracked to perform uniformly accelerated linear motion in two-dimensional space. The initial position of the target is (10km, 10km), the initial velocity is (8m / s, 10m / s), and the initial acceleration is (2m / s2, 2m / s2). The Doppler radar located at the origin of the coordinate system provides the distance, azimuth and Doppler measurement of the target with a sampling period of 1s. The method proposed by the present invention is compared with the SEKF, SUKF and CMKFRR algorithms with Doppler measurement. The standard deviations of their measurement noise are σ and σ, respectively. r =100m,σ θ =1°,σ r =0.1(m / s), correlation coefficient ρ=0.5, the simulation results of the position and velocity root mean square errors of different algorithms are shown in Figure 3(a) and Figure 3(b).

[0161] It can be seen from Figure 2(a), Figure 2(b), Figure 3(a), and Figure 3(b) that with the progress of time, the position and velocity estimation accuracy of each method is continuously improved. Under different motion models, the method proposed in the present invention can reduce the influence of strong nonlinearity of Doppler measurement, and improve the tracking accuracy and convergence speed.

[0162] Each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the system or system embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0163] The above description of the disclosed embodiments enables one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A pre-observation overall fusion tracking method for Doppler measurement, characterized in that: Obtain the posterior estimated mean and covariance of the target state at time k, and the specific steps of filtering from time k to time k+1 are as follows: S1: Obtain target motion data, construct a uniform linear motion model and a uniform accelerated linear motion model, and perform one-step prediction of the target state and covariance of the uniform linear motion model and the uniform accelerated linear motion model in a Cartesian coordinate system; S2: Calculate the measurement prediction and its covariance by one-step prediction of the target state and covariance of the uniform linear motion model and the uniform accelerated linear motion model obtained in S1; S3: Obtain target Doppler radar measurement values ​​and convert the Doppler radar measurement values ​​to Cartesian coordinates, and calculate the measurement conversion error covariance; The specific content in S3 is: Convert the target Doppler measurements to Cartesian coordinates: Calculate the expectation for the above formula: The unbiased measurement is converted to: In the formula, x m ,y m and is the measurement conversion value in the Cartesian coordinate system. The above formula is the measurement conversion of the uniform linear motion model and the uniformly accelerated linear motion model; S4: In the uniform linear motion model and the uniform accelerated linear motion model, the target tracking state and covariance are updated at the current moment by linearly fusing the calculation results in S2 and S3 using the weighted least squares method; The specific contents of S4 are: The linear regression model for constructing the uniform linear motion model is: Let the above matrix equation be recorded as Ax=b+e, and the covariance Var(e) of e be recorded as V, then we have: Status Update: X cv (k+1,k+1)=(A T V -1 A) -1 A T V -1 b; Covariance update: Ρ cv (k+1,k+1)=(A T V -1 A) -1 ; Among them, x, y and is the a posteriori estimate of position and velocity in Cartesian coordinates; is the one-step prediction value of the state; x m ,y m and is the measurement conversion value in the Cartesian coordinate system; is the one-step prediction covariance; is the measurement transformation covariance; S5: Execute steps S1-S4 repeatedly until the target tracking is completed.

2. A pre-observation overall fusion tracking method for Doppler measurement according to claim 1, characterized in that: The specific contents of S1 are: The specific process of one-step prediction of the target state and covariance of the uniform linear motion model in Cartesian coordinates is: X cv (k+1,k)=Φ cv (k)X cv (k,k); R cv (k+1,k)=Φ cv (k)P cv (k,k)Φ cv (k) T +C cv (k)Q cv (k)C cv (k) T ; In the formula, represents the state estimate of the target at time k, x, y and is the position and velocity of the target; X cv (k+1, k) represents the state prediction of the target at time k+1; cv (k, k) represents the estimated covariance of the target at time k; cv (k+1,k) represents the one-step prediction covariance of the target at time k+1; Φ cv (k) is the state transfer matrix, Γ cv (k) is the process noise driving matrix, Q cv (k) is process noise; The specific process of one-step prediction of the target state and covariance of the uniformly accelerated linear motion model under Cartesian coordinates is: X ca (k+1,k)=Φ ca (k)X ca (k,k); R ca (k+1,k)=Φ ca (k)P ca (k,k)Φ ca (k) T +C ca (k)Q ca (k)C ca (k) T ; In the formula, represents the state estimate of the target at time k, is the acceleration of the target; X ca (k+1, k) represents the state prediction of the target at time k+1; ca (k, k) represents the estimated covariance of the target at time k; ca (k+1,k) represents the one-step prediction covariance of the target at time k+1; Φ ca (k) is the state transfer matrix, Γ ca (k) is the process noise driving matrix, Q ca (k) is the process noise.

3. The pre-observation overall fusion tracking method for Doppler measurement according to claim 1, characterized in that: The specific formula for calculating measurement prediction and its covariance in S2 is as follows: In the formula, r c ,θ c , and They are distance prediction, angle prediction, radial velocity prediction and tangential velocity prediction respectively; and is the corresponding covariance; is the state one-step prediction value, P(k+1,k) is the covariance one-step prediction value, P 2×2 (k+1,k) is the first 2 rows and 2 columns of P(k+1,k).

4. The pre-observation overall fusion tracking method for Doppler measurement according to claim 1, characterized in that: The measurement conversion error covariance is specifically expressed as: In the formula, 5. The pre-observation overall fusion tracking method for Doppler measurement according to claim 1, characterized in that: The specific contents of S4 are: The linear regression model for constructing the uniformly accelerated linear motion model is: Let the above matrix equation be recorded as Ax=b+e, and the covariance Var(e) of e be recorded as V, then we have: Status Update: X ca (k+1,k+1)=(A T V -1 A) -1 A T V -1 b; Covariance update: Ρ ca (k+1,k+1)=(A T V -1 A) -1 ; Among them, x, y, and is the a posteriori estimate of position, velocity, and acceleration in the Cartesian coordinate system; and is the one-step prediction value of the state; x m ,y m , and is the measurement conversion value in the Cartesian coordinate system; is the one-step prediction covariance, obtained from S1; is the measurement transformation covariance.