CN115390560A - Ground target track tracking method based on mixed grid multi-model - Google Patents
Ground target track tracking method based on mixed grid multi-model Download PDFInfo
- Publication number
- CN115390560A CN115390560A CN202210992794.9A CN202210992794A CN115390560A CN 115390560 A CN115390560 A CN 115390560A CN 202210992794 A CN202210992794 A CN 202210992794A CN 115390560 A CN115390560 A CN 115390560A
- Authority
- CN
- China
- Prior art keywords
- model
- subset
- fine
- coarse
- probability
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 59
- 230000004927 fusion Effects 0.000 claims abstract description 24
- 238000001914 filtration Methods 0.000 claims description 17
- 238000004422 calculation algorithm Methods 0.000 claims description 16
- 150000001875 compounds Chemical class 0.000 claims description 12
- 238000005259 measurement Methods 0.000 claims description 10
- 230000003993 interaction Effects 0.000 claims description 9
- 238000004364 calculation method Methods 0.000 claims description 5
- 238000004590 computer program Methods 0.000 claims description 4
- 238000000354 decomposition reaction Methods 0.000 claims description 3
- 230000003044 adaptive effect Effects 0.000 claims description 2
- 238000010586 diagram Methods 0.000 description 7
- 230000002452 interceptive effect Effects 0.000 description 5
- 238000004088 simulation Methods 0.000 description 3
- 230000007704 transition Effects 0.000 description 3
- 239000011159 matrix material Substances 0.000 description 2
- 238000007670 refining Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
- G05D1/00—Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots
- G05D1/02—Control of position or course in two dimensions
- G05D1/021—Control of position or course in two dimensions specially adapted to land vehicles
- G05D1/0212—Control of position or course in two dimensions specially adapted to land vehicles with means for defining a desired trajectory
- G05D1/0221—Control of position or course in two dimensions specially adapted to land vehicles with means for defining a desired trajectory involving a learning process
Landscapes
- Engineering & Computer Science (AREA)
- Aviation & Aerospace Engineering (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Automation & Control Theory (AREA)
- Radar Systems Or Details Thereof (AREA)
Abstract
The invention relates to the field of target tracks, in particular to a ground target track tracking method based on a mixed grid multi-model. Step 1: dividing a model set M of the ground target into a coarse model subset and a fine model subset; and 2, step: processing the coarse model subset according to the classification in the step 1 to obtain a coarse model subset with updated probability and carrying out estimation fusion on the coarse model subset; and step 3: step 1, the classified fine model subset is adaptively adjusted according to online data and priori knowledge; and 4, step 4: respectively carrying out probability updating on the rough model subset in the step 2 and the fine model subset in the step 3; and 5: and (4) performing global estimation fusion again on the coarse model subset and the fine model subset which are updated in the step (4). The method is used for solving the problem that the prior technical scheme can not carry out accurate state estimation on the ground target track.
Description
Technical Field
The invention relates to the field of target tracks, in particular to a ground target track tracking method based on a mixed grid multi-model and a readable storage medium.
Background
Since the last 70 s, a variety of fixed structure multi-model algorithms have been applied to state estimation of ground target trajectories. It features that the model set used in the whole state estimation process
Fixed and time invariant, and assuming a set of models
And the real mode space
The same is true. However, when a multi-model algorithm is used to estimate the state of an object with multiple motion patterns, all possible motion patterns cannot be listed, and the model set is
And the real mode space
The same assumption is no longer true.
Disclosure of Invention
The invention provides a ground target track tracking method based on a mixed grid multi-model, which can not accurately estimate the state of a ground target track by utilizing the existing fixed structure multi-model algorithm when the ground target moves in various forms.
The invention is realized by the following technical scheme:
a ground target track tracking method based on a mixed grid multi-model comprises the following steps:
step 1: model set of ground target
Into coarse model subsets M (i =1, \8230;, n) M ) And a fine model subset A (r =1, \8230;, n) A )；
Step 2: sorting according to step 1 on the coarse model subset M (i =1, \ 8230;, n) M ) The processing results in a coarse model subset M (i =1, \ 8230;, n) with probability update M ) And carrying out estimation fusion on the two;
and step 3: step 1 classified fine model subset a (r =1, \8230;, n) A ) Self-adaptive adjustment is carried out according to online data and priori knowledge;
and 4, step 4: for the coarse model subset M of step 2 (i =1, \ 8230;, n) M ) And the fine model subset a of step 3 (r =1, \8230;, n) A ) Respectively updating the probability;
and 5: coarse model subset M (i =1, \ 8230;, n) for step 4 probabilistic update M ) And a fine model subset A (r =1, \8230;, n) A ) And carrying out global state estimation fusion again to realize ground target track tracking.
A ground target track tracking method based on a mixed grid multi-model is disclosed, wherein in the step 2, a coarse model subset M (i =1, \ 8230;, n) is subjected to M ) The treatment specifically comprises the following steps:
step 2.1: sorting according to step 1 on the coarse model subset M (i =1, \ 8230;, n) M ) Performing input interaction;
step 2.2: coarse model subset M (i =1, \ 8230;, n) after input interaction for step 2 M ) And carrying out parallel filtering.
A ground target track tracking method based on a mixed grid multi-model is disclosed, wherein in the step 3, a fine model subset A (r =1, \8230; n) A ) The self-adaptive adjustment according to the online data and the prior knowledge specifically comprises the following steps:
step 3.1: classifying according to step 1 on a fine model subset A (r =1, \8230;, n) A ) Designing;
step 3.2: for the fine model subset A (r =1, \8230;, n) designed in step 3.1 A ) And carrying out parallel filtering.
A ground target track tracking method based on a mixed grid multi-model includes the following steps:
where k denotes the time, x denotes the state quantity, f (-) denotes the state equation, h (-) denotes the measurement equation, w denotes process noise, v denotes measurement noise,
represents an event m k ＝m (j) I.e. model m (j) Acting at time k;
in a hybrid mesh multi-model algorithm, a set of models
The method comprises two parts, namely a coarse model subset M represented by a coarse grid and a fine model subset A represented by a fine grid; model set for time k
Is provided with
Wherein, the coarse model set M is kept unchanged in the whole state estimation process, and the fine model set A is adaptively adjusted according to online data and priori knowledge;
the optimal state estimate based on the minimum mean square error criterion is expressed as
In the formula (I), the compound is shown in the specification,
in order to be a global state estimate,
representing the sequence of measurements from the initial time to time k,
and
respectively a coarse model subset M and a fine model subset A k The posterior probability at time k,
and
based on a coarse model subset M and a fine model subset A, respectively k The state estimate obtained at time k.
3.2, designing a detailed model subset comprises calculating mode moments;
definition of
i =1,2, \ 8230, n is the ith model m (i) First two orders of moment, mu i Is a model m (i) When designing the fine model subset, the desired pattern of the fine model subset at time k
And probability
Not yet acquired, and therefore often utilize the expected pattern at time k-1
And probability
Carrying out replacement; thus, the desired mode
Can be obtained by the following calculation
In the formula (I), the compound is shown in the specification,
a desired pattern that is a subset of the coarse patterns;
from the equation, the desired mode covariance can be obtained as ∑ k Comprises the following steps:
the expected pattern at time k is given by equations (5) and (6)
And covariance ∑ k Next, a set of detailed model subsets is designed by using the method of moment matching
Respectively matching the first two moments of the model subset with the expected patterns
Sum covariance Σ k Equal;
for a ground moving target, designing a model set by considering a constant-speed model and cooperative turning models with different parameters, and adaptively refining the model set
Is designed as follows
In the formula, p is more than or equal to 0 0 < 1 is a predetermined parameter, and subscript n is a vector
Dimension of (2), number of models n A = n +2, number of models n for cooperative turning model A ＝3；
In-process model set
After the design of (2), further utilize
Computing a subset of the fine models
Where B may be represented by ρ Σ k Obtained by Cholseky decomposition and meets rho sigma k ＝BB T (ii) a Thereby completing the design of the fine model set.
A ground target track tracking method based on mixed grid multi-model is used for tracking coarse model subset M (i =1, \8230;, n) M ) Performing input interactions includes model probability prediction
Hybrid weights
Hybrid state estimation and covariance
The coarse model subset M is filtered in parallel into
Wherein KF (·) denotes Kalman filtering;
the probability update of the subset M of the coarse model comprises a likelihood function
Normalized constant
Model probability update
Estimating and fusing the rough model subset M;
where EF (-) denotes the estimated fusion.
A ground target track tracking method based on mixed grid multi-model is obtained by designing a fine model subset A
Wherein FMD (-) represents a fine model subset design;
carrying out parallel filtering on the fine model subset A;
wherein KF (·) represents Kalman filtering;
the probability update of the fine model subset A comprises a likelihood function
Probability of model
Normalized constant
Model probability update
The state estimation fusion of the fine model subset A is specifically
Where EF (-) denotes the estimated fusion.
A ground target track tracking method based on mixed grid multi-model for estimation fusionThe combined coarse model subset M (i =1, \8230;, n) M ) The probability updating is carried out specifically as
For the estimated fused fine model subset A (r =1, \8230;, n) A ) The probability updating is carried out specifically as
A ground target track tracking method based on mixed grid multi-model is used for updating the probability of a coarse model subset M (i =1, \ 8230;, n) M ) And a probability updated fine model subset a (r =1, \8230;, n) A ) The global estimation fusion is performed again specifically in that,
a computer-readable storage medium, in which a computer program is stored which, when being executed by a processor, carries out the above-mentioned method steps.
The invention has the beneficial effects that:
the method adaptively designs the fine model set by using a moment matching method based on the estimation result, and completes the state estimation of the ground target by performing weighted fusion on the estimation result of the fine model set, so that the method has higher estimation precision than a single model algorithm and an interactive multi-model algorithm.
Drawings
Fig. 1 is a schematic diagram of the principle of the present invention.
Fig. 2 is a schematic diagram of the filter structure of the present invention.
FIG. 3 is a graph of the initial moment model distribution and corresponding probability of the present invention.
FIG. 4 is a diagram of the first two moments of the initial time model of the present invention.
Fig. 5 is a diagram of the target motion trajectory of the present invention.
Fig. 6 is a schematic diagram of the target speed profile of the present invention.
FIG. 7 is a schematic diagram of a position estimation error curve of the present invention.
FIG. 8 is a schematic diagram of the velocity estimation error curve of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be described clearly and completely with reference to the accompanying drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
A hybrid mesh multi-model algorithm is applied to track tracking of ground maneuvering targets. The hybrid grid is formed by mixing a coarse grid representing a coarse model and a fine grid representing a fine model (as shown in fig. 1), each grid point corresponds to one target motion model, coordinates of the grid points are relevant parameters in the target motion model, and different grid points represent different models.
A ground target track tracking method based on a mixed grid multi-model comprises the following steps:
step 1: model set of ground target
Into coarse model subsets M (i =1, \8230;, n) M ) And a fine model subset A (r =1, \8230;, n) A )；
The coarse model set M is kept unchanged in the whole state estimation process, and the fine model set A is adaptively adjusted according to online data and priori knowledge;
step 2: classifying the coarse model subset M (i =1, \ 8230;, n) according to step 1 M ) The processing results in a coarse model subset M (i =1, \ 8230;, n) with probability update M ) And carrying out estimation fusion on the two;
and step 3: step 1 classified fine model subset a (r =1, \8230;, n) A ) Self-adaptive adjustment is carried out according to online data and priori knowledge;
and 4, step 4: for the coarse model subset M of step 2 (i =1, \ 8230;, n) M ) And the fine model subset a of step 3 (r =1, \8230;, n) A ) Respectively updating the probability;
and 5: coarse model subset M (i =1, \ 8230;, n) for step 4 probabilistic update M ) And a fine model subset A (r =1, \8230;, n) A ) And carrying out global state estimation fusion again to realize ground target track tracking.
A ground target track tracking method based on mixed grid multi-model is disclosed, in the step 2, a coarse model subset M (i =1, \8230; n) M ) The treatment specifically comprises the following steps:
step 2.1: sorting according to step 1 on the coarse model subset M (i =1, \ 8230;, n) M ) Performing input interaction;
step 2.2: coarse model subset M (i =1, \ 8230;, n) after input interaction for step 2 M ) And carrying out parallel filtering.
A ground target track tracking method based on a mixed grid multi-model is disclosed, wherein in the step 3, a fine model subset A (r =1, \8230; n) A ) The self-adaptive adjustment according to the online data and the prior knowledge specifically comprises the following steps:
step 3.1: classifying according to step 1 on a fine model subset A (r =1, \8230;, n) A ) Designing;
step 3.2: for the fine model subset A (r =1, \8230;, n) designed in step 3.2 A ) And carrying out parallel filtering.
A ground target track tracking method based on a mixed grid multi-model includes the following steps:
where k denotes the time, x denotes the state quantity, f (-) denotes the state equation, h (-) denotes the measurement equation, w denotes process noise, v denotes measurement noise,
represents an event m k ＝m (j) I.e. model m (j) Acting at time k;
in a hybrid mesh multi-model algorithm, a set of models
The method comprises two parts, namely a coarse model subset M represented by a coarse grid and a fine model subset A represented by a fine grid; model set for time k
Is provided with
Wherein, the coarse model set M is kept unchanged in the whole state estimation process, and the fine model set A is adaptively adjusted according to online data and priori knowledge;
the optimal state estimate based on the minimum mean square error criterion is expressed as
In the formula (I), the compound is shown in the specification,
in order to be a global state estimate,
representing the sequence of measurements from the initial time to time k,
and
respectively a coarse model subset M and a fine model subset A k The posterior probability at time k,
and
based on a coarse model subset M and a fine model subset A, respectively k The state estimate obtained at time k.
3.2, designing a detailed model subset comprises calculating mode moments;
in the mixed grid multi-model method, the models in the adaptive fine model subset are not fixed, and are correspondingly adjusted according to the output results of the filters corresponding to other models, so that the accuracy of describing the target motion mode can be improved in the limited number of models. The design of the detailed model subset mainly comprises two parts, namely the calculation of mode moments, namely the calculation of first moment and second moment of grid point coordinates, and the generation of the detailed model subset. Definition of
i =1,2, \ 8230, n is the ith model m (i) First two orders of moment, mu i Is a model m (i) When designing the fine model subset, the desired pattern of the fine model subset at time k
And probability
Not yet acquired, and therefore often utilize the expected pattern at time k-1
And probability
Carrying out replacement; thus, the desired mode
Can be obtained by the following calculation
In the formula (I), the compound is shown in the specification,
a desired pattern that is a subset of the coarse patterns;
from the equation, the covariance of the desired pattern is ∑ k Comprises the following steps:
the expected pattern at time k is given by equations (5) and (6)
And covariance ∑ k Next, a set of detailed model subsets is designed by using the method of moment matching
Respectively connecting the first two moments of the model subset with the expected mode
Sum covariance Σ k Equal;
for a ground moving target, designing a model set by considering a constant-speed model and cooperative turning models with different parameters, and adaptively refining the model set
Is designed as follows
In the formula, p is more than or equal to 0 0 < 1 is a predetermined parameter, and subscript n is a vector
Dimension of (2), number of models n A = n +2, number of models n for cooperative turning model A ＝3；
In-process model set
After the design of (2), further utilize
Computing a subset of the fine models
Where B may be represented by ρ Σ k Obtained by Cholseky decomposition and meets rho sigma k ＝BB T (ii) a Thereby completing the design of the fine model set.
A ground target track tracking method based on mixed grid multi-model is used for tracking coarse model subset M (i =1, \8230;, n) M ) Performing input interactions includes model probability prediction
Mixing weights
Hybrid state estimation and covariance
Parallel filtering the coarse model subset M into
Wherein KF (·) denotes Kalman filtering;
the probability update of the subset M of the coarse model comprises a likelihood function
Normalized constant
Model probability update
Estimating and fusing the rough model subset M;
where EF (-) denotes the estimated fusion.
A ground target track tracking method based on mixed grid multi-model is obtained by designing a fine model subset A
Wherein FMD (-) represents a fine model subset design;
performing parallel filtering on the fine model subset A;
wherein KF (·) represents Kalman filtering;
the probability update of the fine model subset A comprises a likelihood function
Probability of model
Normalized constant
Model probability update
The estimation fusion of the fine model subset A is specifically
Where EF (-) represents the estimated fusion.
A ground target track tracking method based on mixed grid multi-model is used for estimating a fused coarse model subset M (i =1, \8230; n) M ) The probability updating is carried out specifically as
For the estimated fused fine model subset A (r =1, \8230;, n) A ) The probability updating is carried out specifically as
A ground target track tracking method based on mixed grid multi-model is used for updating the probability of a coarse model subset M (i =1, \ 8230;, n) M ) And a probability updated fine model subset a (r =1, \8230;, n) A ) The global estimation fusion is performed again specifically in that,
a computer-readable storage medium, in which a computer program is stored which, when being executed by a processor, carries out the above-mentioned method steps.
To obtain global state estimates
Need to calculate
And
four variables. Since the coarse model subset M remains unchanged throughout the state estimation process, the state estimation can be performed using an interactive multi-model algorithm
And its covariance
Calculating and outputting a coarse model
Probabilities in the coarse model subset M
And the like.
After the state estimation based on the coarse model subset is completed, the state estimation result based on the fine model subset is calculated. First, a fine model subset A at time k is defined k Is composed of
In the formula, m (r) Representing models in a subset of fine models, n A The number of models in the subset of fine models.
State estimation based on a subset of the fine models
And state error covariance
Can be expressed as
In the formula (I), the compound is shown in the specification,
for a fine model m at time k (r) The probability in the fine-model subset a,
and
based on a thin model
State estimation and covariance of
According to the Bayes' theorem,
can be calculated by
In the formula (I), the compound is shown in the specification,
is a thin model
A priori probabilities in the fine model subset A, the probabilities being associated with the fine model
A likelihood function generated according to a certain criterion in the design of the self-adaptive fine model set
And a normalization constant c 2 Are respectively as
In the formula (I), the compound is shown in the specification,
and
which are the measured residual and residual covariance at time k output by the sub-filter r.
Fine model subset A k The posterior probability at time k can be expanded to
In the traditional multi-model algorithm, jump between models is usually represented by a first-order Markov chain, and then the models are refined in the formula
Is predicted with probability of
Can be rewritten as
In the formula (I), the compound is shown in the specification,
π jr representation model
To model
The transition probability of (2).
However, in the mixed grid multi-model algorithm, the model in the fine model set changes in real time according to the output result of each sub-filter, and it is difficult to give a suitable transition probability matrix [ pi ] jr ]。
Therefore, the invention utilizes another way to calculate the model prediction probability
Namely, it is
In the formula (I), the compound is shown in the specification,
is a thin model
In the fine model subset A k A prior probability of (2).
Neglecting the jumps between the subset of fine models and the set of coarse models, have
Substitution of formula (17) into formula (16) gives
In combination of formulae (12), (13) and (18), formula (14) can be further derived as
In the formula (I), the compound is shown in the specification,
is the probability of the coarse model subset M at time k-1.
According to the relation between the probability of the coarse model subset and the probability of the fine model subset, the probability of the coarse model subset M at the time k can be obtained
Is composed of
Therefore, global state estimation of ground target track tracking can be carried out according to the formula.
TABLE 1 hybrid grid Multi-model Algorithm flow
In the above table, KF (·) represents kalman filtering, EF (·) represents estimated fusion, and FMD (·) represents fine model subset design.
And (3) performing track tracking mathematical simulation by respectively using an IMM (inertial measurement model) algorithm and a mixed grid multi-model algorithm by using the ground moving target model as a model set template and angular velocity in the cooperative turning model as a grid point coordinate parameter.
(1) Model set parameters
In addition to the constant velocity model, the coarse model set is defined to include two cooperative turning model components, the turning rate of which is [ 2 ]-15°15°]. Transition probability matrix pi between coarse models HG Is composed of
Initial coarse model probability of
In the fine model design, n A =3 is the number of models in the subset of fine models, and the other preset parameters ρ =0.2 0 ＝0.4。
For the interactive multi-model, a constant-speed model and two cooperative turning models are selected from the model set, and the turning speed of the model is [ -15 degrees and 15 degrees ].
(2) Parameters of object motion
The motion trail of the ground target is shown in fig. 5, and a speed change curve of the target is shown in fig. 6;
the interactive multi-model tracking filter and the hybrid grid multi-model tracking filter are respectively utilized to perform 100 Monte Carlo mathematical simulations, and the obtained simulation results are summarized as shown in FIGS. 7-8:
TABLE 2 root mean square error
The method establishes a coarse model set based on prior information, adaptively designs a fine model set based on an estimation result by using a moment matching method, and completes state estimation of a target by performing weighted fusion on the estimation result of the coarse model set, so that the method has higher estimation precision than an interactive multi-model algorithm.
Claims (10)
1. A ground target track tracking method based on a mixed grid multi-model is characterized by comprising the following steps:
step 1: model set of ground target
Into coarse model subsets M (i =1, \8230;, n) M ) And a fine model subset A (r =1, \8230;, n) A )；
And 2, step: sorting according to step 1 on the coarse model subset M (i =1, \ 8230;, n) M ) The processing results in a coarse model subset M (i =1, \ 8230;, n) with probability update M ) And carrying out estimation fusion on the two;
and step 3: step 1 classified fine model subset a (r =1, \8230;, n) A ) Self-adaptive adjustment is carried out according to online data and priori knowledge;
and 4, step 4: for the coarse model subset M of step 2 (i =1, \ 8230;, n) M ) And the fine model subset a of step 3 (r =1, \8230;, n) A ) Respectively updating the probability;
and 5: coarse model subset M (i =1, \ 8230;, n) for step 4 probabilistic update M ) And a fine model subset A (r =1, \8230;, n) A ) And carrying out global state estimation fusion again to realize ground target track tracking.
2. The method for tracking the ground target trajectory based on the mixed grid multi-model as claimed in claim 1, wherein in the step 2, the coarse model subset M (i =1, \8230;, n) is selected M ) The treatment specifically comprises the following steps:
step 2.1: sorting according to step 1 on the coarse model subset M (i =1, \ 8230;, n) M ) Performing input interaction;
step 2.2: coarse model subset M (i =1, \ 8230;, n) after input interaction for step 2 M ) And performing parallel filtering.
3. The method for tracking the ground target trajectory based on the mixed grid multi-model as claimed in claim 1, wherein the step 3 is performed on a subset A of fine models (r =1, \8230;, n) A ) The self-adaptive adjustment according to the online data and the prior knowledge specifically comprises the following steps:
step 3.1: classifying according to step 1 on a fine model subset A (r =1, \8230;, n) A ) Designing;
step 3.2: for the fine model subset A (r =1, \8230;, n) designed in step 3.1 A ) And carrying out parallel filtering.
4. The method for tracking the ground target trajectory based on the hybrid grid multi-model as claimed in claim 1, wherein the step 1 is specifically a discrete hybrid system as follows:
where k denotes the time, x denotes the state quantity, f (-) denotes the state equation, h (-) denotes the measurement equation, w denotes process noise, v denotes measurement noise,
represents an event m k ＝m (j) I.e. model m (j) Acting at time k;
in the mixed grid multi-model algorithm, a model set
The method comprises two parts, namely a coarse model subset M represented by a coarse grid and a fine model subset A represented by a fine grid; model set for time k
Is provided with
Wherein, the coarse model set M is kept unchanged in the whole state estimation process, and the fine model set A is adaptively adjusted according to online data and priori knowledge;
the optimal state estimate based on the minimum mean square error criterion is expressed as
In the formula (I), the compound is shown in the specification,
in order to be a global state estimate,
representing the sequence of measurements from the initial time to time k,
and
respectively a coarse model subset M and a fine model subset A k The posterior probability at time k is,
and
based on a coarse model subset M and a fine model subset A, respectively k The state estimate obtained at time k.
5. The method for tracking the ground target track based on the hybrid grid multi-model as claimed in claim 3, wherein the step 3.2 of designing the subset of the detailed models comprises calculating mode moments;
definition of
n is the ith model m (i) First two orders of moment, mu i Is a model m (i) When designing the fine model subset, the desired pattern of the fine model subset at time k
And probability
Not yet acquired, and therefore often utilize the expected pattern at time k-1
And probability
Carrying out replacement; thus, the desired mode
Can be obtained by the following calculation
In the formula (I), the compound is shown in the specification,
a desired pattern that is a subset of the coarse patterns;
from the equation, the covariance of the desired pattern is ∑ k Comprises the following steps:
the expected pattern at time k is given by equations (5) and (6)
And covariance ∑ k Below, ofDesigning a group of fine model subsets by using a method of moment matching
Respectively matching the first two moments of the model subset with the expected patterns
Sum covariance Σ k Equal;
for a ground moving target, the design of a model set is carried out by considering a constant-speed model and cooperative turning models with different parameters, and the model set is adaptive to a fine model set
Is designed as follows
In the formula, p is not less than 0 0 < 1 is a predetermined parameter, and subscript n is a vector
Dimension of (2), number of models n A = n +2, number of models n for cooperative turning model A ＝3；
In-process model set
After the design of (2), further utilize
Computing a subset of the fine models
Where B may be represented by ρ Σ k Obtained by Cholseky decomposition and meets rho sigma k ＝BB T (ii) a Thereby completing the design of the fine model set.
6. The method for tracking the ground target trajectory based on the hybrid grid multi-model as claimed in claim 2, wherein the coarse model subset M (i =1, \8230;, n) is selected M ) Performing input interactions includes model probability prediction
Hybrid weights
Hybrid state estimation and covariance
The coarse model subset M is filtered in parallel into
Wherein KF (·) denotes Kalman filtering;
the probability update of the subset M of the coarse model comprises a likelihood function
Normalized constant
Model probability update
Estimating and fusing the rough model subset M;
where EF (-) denotes the estimated fusion.
7. The ground target track tracking method based on the hybrid grid multi-model as claimed in claim 3, characterized in that the fine model subset A is designed to obtain
Wherein FMD (-) represents a fine model subset design;
performing parallel filtering on the fine model subset A;
wherein KF (·) represents kalman filtering;
the probability update of the fine model subset A comprises a likelihood function
Probability of model
Normalized constant
Model probability update
The estimation fusion of the fine model subset A is specifically
Where EF (-) represents the estimated fusion.
8. The method for tracking the ground target trajectory based on the hybrid grid multi-model as claimed in claim 1, wherein the subset M (i =1, \8230;, n) of the coarse models after the estimation fusion is performed M ) The probability updating is carried out specifically as
For the estimation of the fused fine model subset A (r =1, \8230;, n) A ) The probability updating is carried out specifically as
9. The method for tracking the ground target trajectory based on the hybrid grid multi-model as claimed in claim 1, wherein the probability-updated coarse model subset M (i =1, \8230;, n) is selected from M ) And a probability updated fine model subset a (r =1, \8230;, n) A ) The global estimation fusion is performed again specifically in that,
10. a computer-readable storage medium, characterized in that a computer program is stored in the computer-readable storage medium, which computer program, when being executed by a processor, carries out the method steps of any one of the claims 1-9.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210992794.9A CN115390560B (en) | 2022-08-18 | 2022-08-18 | Ground target track tracking method based on mixed grid multi-model |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210992794.9A CN115390560B (en) | 2022-08-18 | 2022-08-18 | Ground target track tracking method based on mixed grid multi-model |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115390560A true CN115390560A (en) | 2022-11-25 |
| CN115390560B CN115390560B (en) | 2023-09-15 |
Family
ID=84120922
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210992794.9A Active CN115390560B (en) | 2022-08-18 | 2022-08-18 | Ground target track tracking method based on mixed grid multi-model |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115390560B (en) |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20130006576A1 (en) * | 2010-03-15 | 2013-01-03 | Bae Systems Plc | Target tracking |
| CN109712172A (en) * | 2018-12-28 | 2019-05-03 | 哈尔滨工业大学 | A kind of pose measuring method of initial pose measurement combining target tracking |
| CN110414130A (en) * | 2019-07-26 | 2019-11-05 | 杭州电子科技大学 | Based on the fuzzy variable structure multi-model maneuvering target tracking method decomposed of error- |
| US20200167934A1 (en) * | 2018-11-27 | 2020-05-28 | GM Global Technology Operations LLC | Systems and methods for applying maps to improve object tracking, lane-assignment and classification |
| CN111562737A (en) * | 2019-12-30 | 2020-08-21 | 中国人民解放军国防科技大学 | Improved AGIMM tracking method controlled by fuzzy logic |
| CN111797478A (en) * | 2020-07-27 | 2020-10-20 | 北京电子工程总体研究所 | Strong maneuvering target tracking method based on variable structure multi-model |
| US20210110411A1 (en) * | 2019-10-11 | 2021-04-15 | Adobe Inc. | Utilizing a tailored machine learning model applied to extracted data to predict a decision-making group |
| CN112784506A (en) * | 2021-01-29 | 2021-05-11 | 中国人民解放军空军工程大学 | Reentry maneuvering trajectory target tracking algorithm based on variable structure multi-model |
| CN114578846A (en) * | 2021-11-29 | 2022-06-03 | 中国人民解放军空军工程大学 | AGIMM tracking method based on maneuver detection sorting |
| CN114611068A (en) * | 2022-03-17 | 2022-06-10 | 重庆大学 | High maneuvering target tracking method |
-
2022
- 2022-08-18 CN CN202210992794.9A patent/CN115390560B/en active Active
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20130006576A1 (en) * | 2010-03-15 | 2013-01-03 | Bae Systems Plc | Target tracking |
| US20200167934A1 (en) * | 2018-11-27 | 2020-05-28 | GM Global Technology Operations LLC | Systems and methods for applying maps to improve object tracking, lane-assignment and classification |
| CN109712172A (en) * | 2018-12-28 | 2019-05-03 | 哈尔滨工业大学 | A kind of pose measuring method of initial pose measurement combining target tracking |
| CN110414130A (en) * | 2019-07-26 | 2019-11-05 | 杭州电子科技大学 | Based on the fuzzy variable structure multi-model maneuvering target tracking method decomposed of error- |
| US20210110411A1 (en) * | 2019-10-11 | 2021-04-15 | Adobe Inc. | Utilizing a tailored machine learning model applied to extracted data to predict a decision-making group |
| CN111562737A (en) * | 2019-12-30 | 2020-08-21 | 中国人民解放军国防科技大学 | Improved AGIMM tracking method controlled by fuzzy logic |
| CN111797478A (en) * | 2020-07-27 | 2020-10-20 | 北京电子工程总体研究所 | Strong maneuvering target tracking method based on variable structure multi-model |
| CN112784506A (en) * | 2021-01-29 | 2021-05-11 | 中国人民解放军空军工程大学 | Reentry maneuvering trajectory target tracking algorithm based on variable structure multi-model |
| CN114578846A (en) * | 2021-11-29 | 2022-06-03 | 中国人民解放军空军工程大学 | AGIMM tracking method based on maneuver detection sorting |
| CN114611068A (en) * | 2022-03-17 | 2022-06-10 | 重庆大学 | High maneuvering target tracking method |
Non-Patent Citations (4)
| Title |
|---|
| HAN SHEN-TU等: "A Novel Variable Structure Multi-Model Tracking Algorithm Based on Error-Ambiguity Decomposition", 2018 21ST INTERNATIONAL CONFERENCE ON INFORMATION FUSION (FUSION) * |
| YU WANG 等: "event-based distributed bias compensation pseudomeasurement information filter for 3D bearing-only target tracking", AEROSPACE SCIENCE AND TECHNOLOGY, no. 117 * |
| 杨云高等: "基于多模型后验概率的机动目标跟踪方法", 雷达科学与技术, no. 03 * |
| 毛征等: "基于高机动目标跟踪的改进变结构IMM算法", 北京工业大学学报, no. 03 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115390560B (en) | 2023-09-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Parmas et al. | PIPPS: Flexible model-based policy search robust to the curse of chaos | |
| Agamennoni et al. | An outlier-robust Kalman filter | |
| WO2020173105A1 (en) | Method and device for tracking maneuvering target | |
| Kalli et al. | Slice sampling mixture models | |
| Stano et al. | Parametric Bayesian filters for nonlinear stochastic dynamical systems: A survey | |
| Nurminen et al. | Skew-$ t $ filter and smoother with improved covariance matrix approximation | |
| Gultekin et al. | Nonlinear Kalman filtering with divergence minimization | |
| WO2017124299A1 (en) | Multi-target tracking method and tracking system based on sequential bayesian filtering | |
| Moldovan et al. | Optimism-driven exploration for nonlinear systems | |
| CN109885883A (en) | A kind of control method of the unmanned vehicle transverse movement based on GK clustering algorithm model prediction | |
| Yu et al. | Nonlinear filtering in unknown measurement noise and target tracking system by variational Bayesian inference | |
| WO2010026417A1 (en) | Estimating a state of at least one target | |
| Qiu et al. | Adaptive genetic particle filter and its application to attitude estimation system | |
| Ulker et al. | Multiple model target tracking with variable rate particle filters | |
| CN109827579A (en) | The method and system of Filtering Model real time correction in a kind of integrated positioning | |
| CN101477623A (en) | Interactive multi-model process based on fuzzy reasoning | |
| Sun et al. | A Gaussian process based method for multiple model tracking | |
| He et al. | A Gaussian mixture unscented Rauch–Tung–Striebel smoothing framework for trajectory reconstruction | |
| CN115390560A (en) | Ground target track tracking method based on mixed grid multi-model | |
| Tang et al. | An improved adaptive unscented FastSLAM with genetic resampling | |
| CN110262235A (en) | A kind of model-free optimal switching method of switching system | |
| Han et al. | Maneuvering target tracking with unknown acceleration using retrospective-cost-based adaptive input and state estimation | |
| Lan et al. | Variational nonlinear Kalman filtering with unknown process noise covariance | |
| Kim et al. | Square root receding horizon information filters for nonlinear dynamic system models | |
| Zhang et al. | Gaussian mixture unscented particle filter with adaptive residual resample for nonlinear model |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |