阅读说明：本技术 一种基于多观测者测速信息的分布式目标跟踪方法 (Distributed target tracking method based on multi-observer speed measurement information ) 是由 梁源 徐兵 于 2019-10-22 设计创作，主要内容包括：本发明提供了一种基于多观测者测速信息的分布式目标跟踪方法,包含四个阶段：本地目标定位阶段：首先每个观测者利用自身所观测到相对角度,相对距离信息对目标进行定位计算,得到目标的本地定位结果；信息通信阶段：每个观测者将自身本地定位结果发送至融合中心,融合中心在综合所有观测者信息后,得到全局定位结果,并将该结果反馈至每一个观测者；本地目标跟踪阶段：利用阶段2中反馈得到的全局定位结果与自身所观测到的相对轴向速度信息对目标再次进行目标跟踪计算,得到目标的本地跟踪结果；信息通信阶段：每个观测者将自身本地跟踪结果发送至融合中心,融合中心在综合所有观测者信息后,得到全局跟踪结果,并将该结果反馈至每一个观测者。(The invention provides a distributed target tracking method based on multi-observer speed measurement information, which comprises four stages: local target positioning stage: firstly, each observer carries out positioning calculation on a target by utilizing the relative angle and relative distance information observed by the observer to obtain a local positioning result of the target; and (3) information communication stage: each observer sends the local positioning result to the fusion center, and the fusion center obtains a global positioning result after integrating the information of all the observers and feeds the result back to each observer; a local target tracking stage: performing target tracking calculation again on the target by using the global positioning result obtained by feedback in the stage 2 and the relative axial speed information observed by the target, so as to obtain a local tracking result of the target; and (3) information communication stage: and each observer sends the local tracking result to the fusion center, and the fusion center obtains a global tracking result after integrating the information of all the observers and feeds the result back to each observer.)

1. A distributed target tracking method based on multi-observer speed measurement information is characterized by comprising the following steps:

step 1: local target positioning stage: each observer estimates the position information of the target by using the target relative angle and the target relative distance information received by the observer to obtain a local positioning result of the observer;

step 2: and (3) information communication stage: each observer sends the local positioning result of the observer per se in the step 1 to a fusion center, the fusion center obtains a global positioning result after integrating the information of all the observers, and the result is fed back to each observer;

and step 3: a local target tracking stage: performing target tracking calculation again on the target by using the global positioning result obtained by feedback in the step 2 and the relative axial speed information observed by the target, so as to obtain a local tracking result of the target;

and 4, step 4: and (3) information communication stage: each observer sends the local tracking result of the observer in the step 3 to a fusion center, the fusion center obtains a global tracking result after integrating the information of all the observers, and the result is fed back to each observer;

and 5: and finishing the target tracking calculation of the current step and outputting a target tracking result.

2. The distributed target tracking method based on multi-observer speed measurement information according to claim 1, wherein the specific method for obtaining the local positioning result in step 1 is as follows:

take observer a as an example, and the rest observers are analogized by analogy; using the object localization model, the r sensed by observer a at time k_a(k),θ_a(k),

3. The distributed target tracking method based on multi-observer speed measurement information according to claim 2, wherein the target positioning model is:

x₁(k+1)＝Φ₁x₁(k)+w₁(k)

z₁(k)＝H₁(x₁(k))+r₁(k)

the corresponding state vector and state transition matrix are:

sequentially setting the target three-axis position and the target three-axis speed under X, Y and Z axes;

t is the system sampling interval, and the state error w (k) is zero-mean Gaussian white noise of the corresponding dimension;

the corresponding observation vectors are:

r₁(k) to correspond toThe measurement noise of the dimension is zero mean Gaussian white noise;

4. The distributed target tracking method based on multi-observer speed measurement information according to claim 1, wherein in the information communication stage in step 2, all observers send their own positioning results and corresponding positioning error covariance matrices to the fusion center, and in the calculation at time k, the fusion center can receive the following information:P_a(k)，

based on the above information, the following fusion calculation is made:

P(k)＝((P_a(k))^-1+(P_b(k))^-1+P_c(k)^-1)^-1

wherein the content of the first and second substances,p (k) is a covariance matrix of the positioning error of the fused positioning result; on the basis of the above, the fusion center will

5. The distributed target tracking method based on multi-observer speed measurement information according to claim 1, wherein the target tracking calculation method in step 3 is as follows:

firstly, estimating a relative angle of a target by using a global positioning result obtained by feedback in the step 2;

taking observer a as an example, the rest observers can be analogized in the same way;

then, dv sensed by the observer a at time k is analyzed by the target tracking model_a(k) As filtering input, EKF calculation is carried out by using the target tracking model to obtain the local target tracking result of the observer a in the target tracking stage at the moment k

6. The distributed target tracking method based on multi-observer speed measurement information according to claim 5, wherein the specific method for estimating the relative angle of the target by using the global positioning result obtained by feedback in step 2 in step 3 is as follows:

θ₁and

wherein

7. The distributed target tracking method based on multi-observer speed measurement information according to claim 5, wherein the target tracking model in step 3 is:

x₂(k+1)＝Φ₂x₂(k)+w₂(k)

z₂(k)＝H₂(x₂(k))+r₂(k)

the corresponding state vector and state transition matrix are:

sequentially setting target three-axis speed and target three-axis acceleration under X, Y and Z axes;

the corresponding observation vectors are: z₂(k)＝[dv]^TI.e. target relative axial velocity; the corresponding observation matrix is:

8. the distributed target tracking method based on multi-observer speed measurement information according to claim 1, wherein in the information communication stage in step 4, all observers send their tracking results and corresponding tracking error covariance matrices to the fusion center, and in the calculation at time k, the fusion center can receive the following information:

based on the above information, the following fusion calculation is made:

wherein the content of the first and second substances,for fused trace results, P²(k) And the covariance matrix of the positioning error of the fused tracking result is fed back to each observer by the fusion center.

Technical Field

The invention relates to the field of target tracking and positioning, in particular to a distributed target tracking method based on multi-observer speed measurement information.

Background

The target tracking problem is caused by the occurrence of radar stations, the occurrence of SCR-28 of the first tracking radar station in British before the second war in 1937 marks the formal occurrence of the target tracking problem, and various tracking technical means including target tracking systems such as radar, laser, infrared and sonar are developed successively later, and the technical means are continuously improved in the future development. Earlier, the target tracking technology was not well known, and until the 70 s, after the kalman filtering theory was successfully applied in the field of target tracking, people generally paid attention to the maneuvering target tracking technology and generated great interest. Later, the maneuvering target tracking technology is rapidly developed, and is one of the most popular fields of domestic and foreign research. In recent decades, new technologies such as extended kalman filtering, multi-model algorithms, and particle filtering have been developed and developed in order to meet the needs of science and technology and applications. With the development of the times, the related sensor technology is also continuously improved. The current target tracking sensor can acquire information such as a relative distance and a relative angle of a target, and can measure a relative axial velocity (a velocity component in a target-observer connecting line direction) of the target in real time to acquire the relative axial velocity of the target.

Disclosure of Invention

In order to further enhance the target tracking performance, the method introduces the relative axial speed information into the target tracking system, and can greatly improve the tracking capability and the tracking precision of the target; meanwhile, the same target is tracked by adopting a multi-observer cooperative observation mode, and the accuracy and stability of target tracking can be more effectively improved by utilizing the information of multiple observers, so that distributed target tracking and positioning under the condition that the speed measurement information of multiple observers exists are realized.

The purpose of the invention is realized by the following technical scheme:

a distributed target tracking method based on multi-observer speed measurement information comprises the following steps:

step 1: local target positioning stage: each observer estimates the position information of the target by using the target relative angle and the target relative distance information received by the observer to obtain a local positioning result of the observer;

step 2: and (3) information communication stage: each observer sends the local positioning result of the observer per se in the step 1 to a fusion center, the fusion center obtains a global positioning result after integrating the information of all the observers, and the result is fed back to each observer;

and step 3: a local target tracking stage: performing target tracking calculation again on the target by using the global positioning result obtained by feedback in the step 2 and the relative axial speed information observed by the target, so as to obtain a local tracking result of the target;

and 4, step 4: and (3) information communication stage: each observer sends the local tracking result of the observer in the step 3 to a fusion center, the fusion center obtains a global tracking result after integrating the information of all the observers, and the result is fed back to each observer;

and 5: and finishing the target tracking calculation of the current step and outputting a target tracking result.

Further, the specific method for obtaining the local positioning result in step 1 is as follows:

using the object localization model, the r sensed by observer a at time k_a(k),θ_a(k),

As the filtering input, the extended kalman filtering algorithm (EKF) is used for filtering estimation to obtain the local target positioning result of the observer a in the target positioning stage at the moment k

The positioning error covariance matrix P corresponding thereto_a(k) In that respect Take observer a as an example, and so on for the rest of the observers.

Further, the target location model is:

x₁(k+1)＝Φ₁x₁(k)+w₁(k)

z₁(k)＝H₁(x₁(k))+r₁(k)

the corresponding state vector and state transition matrix are:

sequentially setting the target three-axis position and the target three-axis speed under X, Y and Z axes;

t is the system sampling interval, and the state error w (k) is zero-mean Gaussian white noise of the corresponding dimension;

the corresponding observation vectors are:

namely a target relative distance, a target relative azimuth angle and a target relative pitch angle; the corresponding observation matrix is:

r₁(k) the measurement noise of the corresponding dimension is zero mean Gaussian white noise;

at time k, tester a is in its own triaxial position.

Further, in the information communication stage in step 2, all observers send their own positioning results and corresponding positioning error covariance matrices to the fusion center, and in the calculation at time k, the fusion center may receive the following information:

P_a(k)，

P_b(k)，

P_c(k)；

based on the above information, the following fusion calculation is made:

P(k)＝((P_a(k))^-1+(P_b(k))^-1+P_c(k)^-1)^-1

wherein the content of the first and second substances,

p (k) is a covariance matrix of the positioning error of the fused positioning result; on the basis of the above, the fusion center will

And P (k) is fed back to the observers a, b, c.

Further, the target tracking calculation method in step 3 is as follows:

firstly, estimating a relative angle of a target by using a global positioning result obtained by feedback in the step 2;

then, dv sensed by the observer a at time k is analyzed by the target tracking model_a(k) As filtering input, EKF calculation is carried out by using the target tracking model to obtain the local target tracking result of the observer a in the target tracking stage at the moment kThe covariance matrix of tracking error corresponding thereto

Further, the specific method for estimating the target relative angle by using the global positioning result obtained by feedback in step 2 in step 3 is as follows: take observer a as an example, and the rest observers are analogized by analogy;

θ₁and

the calculated relative azimuth angle and the calculated relative pitch angle are both

And calculating to obtain the following results:

wherein

And

are respectively as

The 1 st 2 nd and 3 rd elements in (b).

Further, the target tracking model in step 3 is:

x₂(k+1)＝Φ₂x₂(k)+w₂(k)

z₂(k)＝H₂(x₂(k))+r₂(k)

the corresponding state vector and state transition matrix are:

sequentially setting target three-axis speed and target three-axis acceleration under X, Y and Z axes;

the corresponding observation vectors are: z₂(k)＝[dv]^TI.e. target relative axial velocity; the corresponding observation matrix is:

further, in the information communication stage in step 4, all observers send their tracking results and corresponding tracking error covariance matrices to the fusion center, and in the calculation at time k, the fusion center may receive the following information:

based on the above information, the following fusion calculation is made:

wherein the content of the first and second substances,for fused trace results, P²(k) And the covariance matrix of the positioning error of the fused tracking result is fed back to each observer by the fusion center.

The input of the target tracking algorithm of the invention additionally introduces the relative axial velocity of the target besides the information of the relative angle and the relative distance of the target; the target tracking algorithm is a distributed algorithm under the condition that multiple observers exist, and the calculation of target tracking is not only executed in a fusion center, but also executed locally by the observers; the target tracking algorithm comprises two tracking calculations and two communication between an observer and a fusion center, wherein the first tracking calculation is mainly used for estimating the position of a target, and the communication content of the first communication is the result of the first tracking calculation; the second tracking calculation is mainly used for estimating the speed and the acceleration of the target, and the communication content of the second communication is the result of the second tracking calculation; and the final output of the algorithm is the target tracking result after the second communication.

Compared with the prior art, the invention has the beneficial effects that:

1. compared with the traditional target tracking algorithm, the method has the advantages that the tracking accuracy and the tracking stability of the target tracking algorithm are remarkably improved by introducing the target axial speed observed quantity and introducing multiple observers;

2. the distributed algorithm architecture is adopted, so that on one hand, the complexity of calculation is reduced, and the operation efficiency is improved; on the other hand, the robustness of the target tracking system is enhanced, and the stability of the algorithm is enhanced;

3. the method adopts a target tracking calculation mode of two-stage calculation, effectively reduces the calculation complexity and the coupling between the parameters to be estimated, and is more convenient to apply in engineering practice.

Drawings

FIG. 1 is a block diagram of a distributed target tracking algorithm designed in the present invention;

FIG. 2 is a timing diagram of a distributed target tracking algorithm according to the present invention;

FIG. 3 is a schematic diagram of simulation of an operation result of the distributed target tracking algorithm of the present invention 1;

FIG. 4 is a schematic diagram of simulation of an operation result of the distributed target tracking algorithm of the present invention 2;

FIG. 5 is a schematic diagram of simulation of the operation result of the distributed target tracking algorithm of the present invention 3;

fig. 6 is a schematic diagram 4 illustrating simulation of an operation result of the distributed target tracking algorithm according to the present invention.

FIG. 7 is a schematic diagram 5 illustrating simulation of the operation result of the distributed target tracking algorithm according to the present invention;

fig. 8 is a schematic diagram 6 illustrating simulation of an operation result of the distributed target tracking algorithm according to the present invention.

Detailed Description

The invention is described in detail below by way of example with reference to the accompanying drawings.

The embodiment provides a distributed target tracking method based on multi-observer speed measurement information, and the overall flow block diagram of the method is shown in fig. 1, and the system timing diagram is shown in fig. 2.

For convenience of subsequent description, the following definitions are made in this embodiment:

defining that 3 observers in the system observe the target, which are named as a, b and c respectively (it should be noted that the method of the present invention is not limited to 3 observers, and the method of the present invention can also be extended to the case where more observers exist); at time k, the obtained corresponding observation information is sequentially:

wherein r is_a(k),θ_a(k),dv_a(k) Sequentially and respectively obtaining a target relative distance, a target relative azimuth angle, a target relative pitch angle and a target relative axial speed of the observer a at the moment k;

wherein r is_b(k),θ_b(k),

dv_b(k) Sequentially and respectively obtaining a target relative distance, a target relative azimuth angle, a target relative pitch angle and a target relative axial speed of the observer b at the moment k;

wherein r is_a(k),θ_a(k),dv_a(k) The relative distance of the target, the relative azimuth angle of the target, the relative pitch angle of the target and the relative axial velocity of the target obtained by the observer c at the time k are sequentially and respectively.

At time k, the motion information of three observers is:

observer a: three-axis positions of the self-body are

Self triaxial velocity of

Observer b: three-axis positions of the self-body are

Self triaxial velocity of

An observer c: three-axis positions of the self-body are

Self triaxial velocity of

The movement information of the observer is provided by a navigation system equipped with the observer.

Step 1: local target positioning stage: and each observer estimates the position information of the target by using the target relative angle and the target relative distance information received by the observer to obtain a local positioning result of the observer.

The following description will be made by taking the calculation flow of the observer a as an example:

firstly, the following target positioning model is constructed:

x₁(k+1)＝Φ₁x₁(k)+w₁(k)

z₁(k)＝H₁(x₁(k))+r₁(k)

the corresponding state vector and state transition matrix are:

sequentially setting the target three-axis position and the target three-axis speed under X, Y and Z axes;

t is the system sampling interval, and the state error w (k) is zero-mean white Gaussian noise of the corresponding dimension.

The corresponding observation vectors are:

namely a target relative distance, a target relative azimuth angle and a target relative pitch angle; the corresponding observation matrix is:

r₁(k) the measurement noise corresponding to the dimension is zero mean white gaussian noise.

Using the object localization model described above, the r sensed by observer a at time k is measured_a(k),θ_a(k),

As the filtering input, the extended kalman filtering algorithm (EKF) is used for filtering estimation to obtain the local target positioning result of the observer a in the target positioning stage at the moment k

The positioning error covariance matrix P corresponding thereto_a(k) In that respect The EKF algorithm is a commonly used nonlinear filtering algorithm, and is specifically described as follows:

for a nonlinear continuous discrete system as follows:

X_k+1＝f(X_k)+w_k

Z_k＝h(X_k)+v_k

wherein, X_kIs a system state, Z_kFor systematic observation of information, randomVariable w_kAnd v_kThe noise is the process noise and the measurement noise of the system respectively, and is the zero mean value Gaussian white noise which is independent from each other. Wherein, the variance of the process noise is Q, and the variance of the measurement noise is R. f is the sum system state variable X in the state equation_kA time k dependent nonlinear function; h is the sum state variable X in the observation equation_kA non-linear function related to time k.

Applying an EKF algorithm to the model:

P_k,k-1＝AP_k-1A^T+Q

K_k＝P_k,k-1H^T[HP_k,k-1H^T+R]^-1

P_k＝[I-K_kH]P_k,k-1

wherein the content of the first and second substances,

represents the result of the filtering of the previous step,

one-step prediction of the representative state, P_k,k-1Representing the one-step prediction error variance, K_kRepresenting the filter gain, P_kRepresenting the current step filter error variance, P_k-1Representing the filtering error variance of the k-1 step; i represents an identity matrix of corresponding dimension, wherein A and H represent f and H respectively

The jacobian matrix is a matrix in which the first partial derivatives of a function commonly used in mathematics are arranged in a certain way.

The same operations are performed for observers b and c, and the result isPositioning result of observer b in this step

The positioning error covariance matrix P corresponding thereto_b(k) And the positioning result of the observer c in the current step

The positioning error covariance matrix P corresponding thereto_c(k)。

Step 2: and (3) information communication stage: and (3) each observer sends the local positioning result of the observer to the fusion center in the step (1), the fusion center obtains a global positioning result after integrating the information of all the observers, and the result is fed back to each observer.

After the calculation in step 1 is completed, all observers send their local positioning results and the corresponding positioning error covariance matrix to the fusion center, and in the calculation at time k, the fusion center can receive the following information:

P_a(k)，

P_b(k)，

P_c(k) in that respect Here, 3 observers are taken as an example for explanation, and it should be noted that the method can be extended to a case where more observers exist.

Based on the above information, the following fusion calculation is made:

P(k)＝((P_a(k))^-1+(P_b(k))^-1+P_c(k)^-1)^-1

wherein the content of the first and second substances,

p (k) is a covariance matrix of the positioning error of the fused positioning result. On the basis of the above, the fusion center will

And P (k) is fed back to the observers a, b, c.

Meanwhile, in order to further improve the target precision and stability for a long time, each observer replaces the local result of the observer with a global result, namely:

observer a:

P_a(k)＝P(k)

observer b:

P_b(k)＝P(k)

an observer c:

P_c(k)＝P(k)。

and step 3: a local target tracking stage: and (3) performing target tracking calculation again on the target by using the global positioning result obtained by feedback in the step (2) and the relative axial speed information observed by the target, so as to obtain a local tracking result of the target.

The following description will be made by taking the calculation flow of the observer a as an example:

firstly, estimating a target relative angle by using a positioning result obtained by feedback in the step 2, and as follows:

θ₁and

the calculated relative azimuth angle and the calculated relative pitch angle are both

And calculating to obtain the following results:

whereinAnd

are respectively as

The 1 st 2 nd and 3 rd elements in (b).

Based on the above calculation results, the target tracking model is introduced as follows:

x₂(k+1)＝Φ₂x₂(k)+w₂(k)

z₂(k)＝H₂(x₂(k))+r₂(k)

the corresponding state vector and state transition matrix are:

sequentially setting target three-axis speed and target three-axis acceleration under X, Y and Z axes;

the corresponding observation vectors are: z₂(k)＝[dv]^TI.e. target relative axial velocity; the corresponding observation matrix is:

dv sensed by observer a at time k is measured using the above model_a(k) As filtering input, EKF calculation is carried out by using the target tracking model to obtain the local target tracking result of the observer a in the target tracking stage at the moment kThe covariance matrix of tracking error corresponding thereto

The same operation is carried out on observers b and c, and the tracking result of the observer b in the step can be obtained

The covariance matrix of tracking error corresponding thereto

And the tracking result of the observer c in the current step

The covariance matrix of tracking error corresponding thereto

And 4, step 4: and (3) information communication stage: and (3) each observer sends the local tracking result of the observer to the fusion center, and the fusion center obtains a global tracking result after integrating the information of all the observers and feeds the result back to each observer.

After the calculation in step 3 is completed, all observers send their tracking results and the corresponding tracking error covariance matrix to the fusion center, and in the calculation at time k, the fusion center may receive the following information:

here, 3 observers are taken as an exampleIt should be noted that the method can be extended to situations where more observers are present.

Based on the above information, the following fusion calculation is made:

wherein the content of the first and second substances,

for fused trace results, P²(k) And the covariance matrix of the positioning error of the fused tracking result is fed back to each observer by the fusion center.

Meanwhile, in order to further improve the target precision and stability for a long time, each observer replaces the local result of the observer with a global result, namely:

observer a:

observer b:

an observer c:

the target tracking result at the time k is finally obtained as follows:

target three-axis position:

target triaxial speed:

target triaxial acceleration:

wherein the content of the first and second substances,

in turn represent

The 1 st, 2 nd and 3 rd elements of (a);

in turn represent

1 st, 2 nd, 3 rd, 4 th, 5 th and 6 th elements.

And (3) outputting the filtering result obtained in the step (5) as a final target tracking result, wherein the final output result is shown in fig. 3 to fig. 7, fig. 3 is a schematic diagram of a three-axis position real value and a tracking result, fig. 4 is a schematic diagram of a three-axis tracking position error, fig. 5 is a schematic diagram of a three-axis velocity real value and a tracking result, fig. 6 is a schematic diagram of a three-axis tracking velocity error, fig. 7 is a schematic diagram of a three-axis acceleration real value and a tracking result, and fig. 8 is a schematic diagram of a three-. It can be seen from the simulation graphs that, by using the target tracking algorithm designed by the invention, not only can the position tracking result and the velocity tracking result in the three axial directions converge to the vicinity of the real value quickly and accurately, but also the acceleration tracking result in the three axial directions can realize accurate long-term tracking of the real acceleration, that is, the accurate positioning and tracking of the target can be effectively realized by using the algorithm.

In summary, the above description is only a preferred embodiment of the present invention, and is not intended to limit the scope of the present invention. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.