CN116047498A - Maneuvering target tracking method based on maximum correlation entropy extended Kalman filtering - Google Patents
Maneuvering target tracking method based on maximum correlation entropy extended Kalman filtering Download PDFInfo
- Publication number
- CN116047498A CN116047498A CN202310063017.0A CN202310063017A CN116047498A CN 116047498 A CN116047498 A CN 116047498A CN 202310063017 A CN202310063017 A CN 202310063017A CN 116047498 A CN116047498 A CN 116047498A
- Authority
- CN
- China
- Prior art keywords
- model
- target
- state
- probability
- matrix
- 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.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 40
- 238000001914 filtration Methods 0.000 title claims abstract description 32
- 239000011159 matrix material Substances 0.000 claims abstract description 61
- 230000007704 transition Effects 0.000 claims abstract description 30
- 238000005259 measurement Methods 0.000 claims abstract description 25
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 18
- 230000002452 interceptive effect Effects 0.000 claims abstract description 13
- 230000003993 interaction Effects 0.000 claims abstract description 7
- 230000008569 process Effects 0.000 claims description 7
- 238000005070 sampling Methods 0.000 claims description 7
- 230000008859 change Effects 0.000 claims description 6
- 238000012804 iterative process Methods 0.000 claims description 6
- 238000000354 decomposition reaction Methods 0.000 claims description 5
- 241000135164 Timea Species 0.000 claims description 3
- 238000010276 construction Methods 0.000 claims description 3
- 230000000694 effects Effects 0.000 description 8
- 238000004364 calculation method Methods 0.000 description 2
- 238000007796 conventional method Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000000342 Monte Carlo simulation Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000007123 defense Effects 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S13/00—Systems using the reflection or reradiation of radio waves, e.g. radar systems; Analogous systems using reflection or reradiation of waves whose nature or wavelength is irrelevant or unspecified
- G01S13/66—Radar-tracking systems; Analogous systems
- G01S13/72—Radar-tracking systems; Analogous systems for two-dimensional tracking, e.g. combination of angle and range tracking, track-while-scan radar
- G01S13/723—Radar-tracking systems; Analogous systems for two-dimensional tracking, e.g. combination of angle and range tracking, track-while-scan radar by using numerical data
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/11—Complex mathematical operations for solving equations, e.g. nonlinear equations, general mathematical optimization problems
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/16—Matrix or vector computation, e.g. matrix-matrix or matrix-vector multiplication, matrix factorization
Abstract
The invention provides a maneuvering target tracking method based on maximum correlation entropy extended Kalman filtering, which comprises the following steps: establishing a model set, a target state equation and a measurement equation of an interactive multi-model algorithm; initializing model probability, target state vector, covariance matrix and Markov state transition matrix; performing input interaction between models, and linearizing a nonlinear model by using an extended Kalman filter; carrying out one-step prediction on each model, and designing a cost function based on a maximum entropy criterion; obtaining posterior update values of the k+1 moment targets in the models through an iteration method; updating the model probability, correcting the transition probability, and finally fusing and outputting the tracking result. By the method, the nonlinear maneuvering target tracking problem under the non-Gaussian noise can be solved, and the target estimation accuracy can be improved.
Description
Technical Field
The invention belongs to the technical field of radar target tracking methods, and particularly relates to a maneuvering target tracking method based on maximum correlation entropy extended Kalman filtering.
Background
Maneuvering target tracking is a process of estimating the state of maneuvering objects based on sensor information, and the tracking technology of maneuvering targets has wide application in the fields of national defense science and technology and national economy. One challenging problem in the area of maneuver target tracking is that maneuver targets have uncertainty, often described by multiple models, as a single model is not sufficient to describe the target motion state, and interactive multimodal algorithms are one of the common multimodal estimation strategies. Under the framework of an interactive multi-model algorithm, when the target estimation state or the sensor measurement vector is nonlinear, a nonlinear filter such as extended Kalman can be used for tracking the maneuvering target. However, the conventional nonlinear filter is derived based on the minimum mean square error criterion and only contains the second-order information of the tracking error, so that a good tracking effect can be obtained under the condition of Gaussian noise, but the tracking effect is poor under the condition of non-Gaussian noise. In order to solve the problem, the invention combines an extended Kalman filtering algorithm based on the maximum correlation entropy criterion with an interactive multi-model algorithm, and can retain high-order moment information of errors because the maximum correlation entropy is adopted instead of the minimum mean square error criterion. Compared with the traditional nonlinear maneuvering target tracking algorithm, the algorithm can obtain better tracking effect under the condition of non-Gaussian noise.
Disclosure of Invention
The invention aims to provide a maneuvering target tracking method based on maximum correlation entropy expansion Kalman filtering, which solves the problem that the tracking effect of the traditional nonlinear filter is poor under the condition of non-Gaussian noise; the tracking effect is optimized, so that the effect is more accurate.
The technical scheme adopted by the invention is as follows: the maneuvering target tracking method based on maximum correlation entropy extended Kalman filtering comprises the following specific operation steps:
and 6, updating the model probability: calculating likelihood functions of the models, obtaining probability updating values of the models according to a Bayesian probability formula, and adopting the probability change rate of the models to construct a correction function to correct transition probabilities among the models;
And 8, repeating the steps 3-7 until the target tracking process is finished.
The invention is also characterized in that:
the step 1 is specifically as follows:
step 1.1, the target state at time k is expressed asThe linear CV model of the uniform linear motion of the target is simulated, and a target state equation is constructed as follows:
x k+1 =F k x k +G k w k (1)
wherein the state vector x k Wherein x, y respectively represent the position of the target along the x direction and the position along the y direction;is the speed of the target in the x-direction, the speed in the y-direction; t represents the sampling interval, w k Mean 0, covariance Q Gaussian white noise, F k Representing a state transition matrix, G k Representing a noise driving matrix;
step 1.2 toThe method comprises the steps of representing a target state at the moment k, simulating a nonlinear CT model of coordinated turning motion of a target with unknown angular velocity, and constructing a target state equation:
x k+1 =f(x k )+G k w k (3)
wherein the angular velocity omega k Not a constant;
step 1.3, taking the x-direction position and the y-direction position as observables, and establishing a measurement equation with non-Gaussian noise:
z k =H k x k +v k (5)
wherein ,zk Observation vector representing k time, measurement noise v k Is non-Gaussian noise obeying a mixed Gaussian distribution, and is specifically expressed as follows: v k ~(1-α)N(0,R 1k )+αN(0,R 2k ) Wherein alpha represents weight, the value range is 0-1, N (DEG,) represents normal distribution, R 1k To measure the noise covariance, R 2k To measure the covariance of noise when it is abnormally disturbed.
initializing parameter values, including determining a probability that the target is in model i at time k=1State estimation->Covariance matrix->And a Markov state transition matrix pi k ={π ij,k } M×M Wherein each element pi ij,k The transition probability of the system from the model i to the model j at the moment k is represented, and M represents the total number of the models.
The step 3 is specifically as follows:
step 3.1, according to the model probabilityAnd a Markov state transition matrix pi k Calculating the mixing probability between the k moment models by the formulas (7), (8)>
wherein ,representing the probability that the target is in the model j after the input interaction;
step 3.2, according to the target state at the time of kCovariance matrix->And the mixing probability between the models obtained in step 3.1 +.>Obtaining the mixed state estimation +.A mixed state estimation of the k moment target in each model is obtained by the formulas (9) (10) respectively>And hybrid covariance matrix->
wherein ,is the hybrid state estimate for the object in each model at time k,>is the time kThe target is at the hybrid covariance of each model.
step 4.1, calculating a Jacobian matrix of a nonlinear system state equation according to formulas (11) and (12), and linearizing the nonlinear state equation (3):
wherein the partial derivatives of the state quantities over angular rate are:
wherein ,ωk Is angular velocity, T is sampling interval,The speed in the x direction and the speed in the y direction are respectively;
step 4.2, estimating the mixed state of each model by using the k moment targets obtained in the step 3.2And hybrid covariance matrix->According to whether the state equation is linear, the prior prediction value is filtered by using the formula (13) or (14) respectively>Solving for a priori covariance estimate +.>
The step 5 is specifically as follows:
step 5.1, filtering priori predicted values obtained in step 4.2 according to the target state equations (1) (3), the measurement equation (5) and />The system equation is extended as follows:
wherein ,
wherein ,for a priori prediction covariance matrix->Cholesky score was performedThe resulting lower triangular matrix is solved for,for measuring covariance matrix R k+1 Performing Cholesky decomposition to obtain a lower triangular matrix; />
The two sides of the formula (16) are respectively multiplied byThe following expansion equation is obtained:
wherein
Wherein I represents an identity matrix;
step 5.2, designing a cost function based on maximum entropy:
wherein ,Gσ Representing a Gaussian kernel function, L isAnd l=n+m, n representing the state dimension, m representing the measurement dimension, +.>Representation->The r element of (2)>Representation->R line of (2);
the optimal estimate of the target state is:
deriving the optimal target state for equation (21) is:
wherein ,diagonal matrix representing kernel function structure related to target motion state prediction +.>A diagonal matrix representing the construction of a kernel function related to target measurements,/->Is-> Lines 1 to n, +.>Is->N+1th to n+mth rows;
step 5.3, let iteration number t=1, use the priori predicted value obtained in step 4.2 as the initial value of the iterative process:and determining a kernel bandwidth value sigma and a threshold epsilon;
step 5.4, calculating the state estimation value obtained by the t-th iteration process by using the following formula
wherein ,
step 5.5, comparing whether the estimated values of the t iteration and the t-1 iteration obtained in the step 5.4 meet the formula (31), and if not, repeating the step 5.4; if yes, ending the iterative process, and setting the iteration result of the t timeAs a filtered posterior update value +.>And updating covariance matrix +.>
step 6.1, filtering a priori prediction value obtained according to step 4.2And covariance matrix->Calculating likelihood functions of the models:
step 6.2, updating the probability of each model according to a Bayesian probability formula by the following formula:
step 6.3, constructing a transition probability correction function of the model j:
step 6.4, using the correction function of the model j to correct the transition probability of other models to the model:
normalizing the transition probability:
the step 7 is specifically as follows:
step 7.1, updating the value according to the filtering posterior obtained in step 5.5And the updated probability of each model obtained in the step 6.2 is fused with the state estimation of each model:
step 7.2, updating the value of the covariance matrix obtained in step 5.5Covariance of each model was fused:
the beneficial effects of the invention are as follows:
1. the invention provides a maneuvering target tracking algorithm based on maximum correlation entropy expansion Kalman filtering, which can solve the tracking problem of maneuvering targets under the nonlinear and non-Gaussian noise conditions.
2. Under the framework of the interactive multi-model algorithm, an extended Kalman filter based on the maximum entropy is adopted, and because the maximum correlation entropy is used instead of the minimum mean square error criterion, the high-order moment information of errors can be reserved, and under the non-Gaussian noise condition, the tracking result is superior to that of the traditional interactive multi-model algorithm based on the extended Kalman filter.
Drawings
FIG. 1 is a global flow chart of the maneuver target tracking method of the present invention based on maximum correlation entropy extended Kalman filtering;
FIG. 2 is a graph of a target tracking trajectory in embodiment 1 of the maneuver target tracking method of the present invention based on maximum correlation entropy extended Kalman filtering;
FIG. 3 (a) is a graph of position mean square error versus conventional method for example 1 of the maneuver target tracking method of the present invention based on maximum correlation entropy extended Kalman filtering;
fig. 3 (b) is a velocity mean square error comparison graph of the motorized target tracking method of the present invention based on the maximum correlation entropy-extended kalman filter with the conventional method.
Detailed Description
The present invention is described in detail below with reference to the drawings and specific embodiments so that the advantages and features of the present invention will be more readily understood by those skilled in the art.
Example 1
In order to verify the effectiveness of the invention, a maneuvering target is tracked, the radar sampling period is 0.1s, the Monte Carlo simulation times are 100 times, and the specific movement process of the target is as follows: starting from an origin point, performing uniform linear motion along the x direction at a speed of 5m/s, starting to perform left turning at 30 degrees after 3 seconds, stopping turning at the target after 5 seconds, continuing to perform uniform linear motion along the x direction at a speed of 5m/s, starting to perform right turning at 30 degrees after 4 seconds, stopping turning at the target after 6 seconds, continuing to perform uniform linear motion along the x direction at a speed of 5m/s, and stopping moving at the target after 2 seconds; the method is implemented according to the following steps:
and (3) executing the step (1) to establish a model set, a target state equation and a measurement equation of the interactive multi-model algorithm.
The method is implemented according to the following steps:
step 1.1, the target state at time k is expressed asThe linear CV model of the uniform linear motion of the target is simulated, and a target state equation is constructed as follows:
x k+1 =F k x k +G k w k (1)
wherein the state vector x k Wherein x, y respectively represent the position of the target along the x direction and the position along the y direction;is the speed of the target in the x-direction, the speed in the y-direction; t represents the sampling interval, w k Mean 0, covariance Q Gaussian white noise, F k Representing a state transition matrix, G k Representing a noise driving matrix;
step 1.2 toRepresenting the state of the target at the moment k, and coordinating unknown angular velocity by the simulation targetThe nonlinear CT model of turning motion is constructed as follows:
x k+1 =f(x k )+G k w k (3)
wherein the angular velocity omega k Not a constant;
step 1.3, taking the x-direction position and the y-direction position as observables, and establishing a measurement equation with non-Gaussian noise:
z k =H k x k +v k (5)
wherein ,zk Observation vector representing k time, measurement noise v k Is non-Gaussian noise obeying a mixed Gaussian distribution, and is specifically expressed as follows: v k ~(1-α)N(0,R 1k )+αN(0,R 2k ) Wherein alpha represents weight, the value range is 0-1, N (DEG,) represents normal distribution, R 1k To measure the noise covariance, R 2k To measure the covariance of noise when it is abnormally disturbed.
The method is implemented according to the following steps:
step 3.1, according to the model probabilityAnd a Markov state transition matrix pi k Calculating the mixing probability between the k moment models by the formulas (7), (8)>
wherein ,representing the probability that the target is in the model j after the input interaction;
step 3.2, according to the target state at the time of kCovariance matrix->And the mixing probability between the models obtained in step 3.1 +.>Obtaining the mixed state estimation +.A mixed state estimation of the k moment target in each model is obtained by the formulas (9) (10) respectively>And hybrid covariance matrix->
wherein ,is the hybrid state estimate for the object in each model at time k,>is the mixed covariance of the model with the target at time k.
The method is implemented according to the following steps:
step 4.1, calculating a Jacobian matrix of a nonlinear system state equation according to formulas (11) and (12), and linearizing the nonlinear state equation (3):
wherein the partial derivatives of the state quantities over angular rate are:
wherein ,ωk Is angular velocity, T is sampling interval,Velocity sum in x-direction respectivelyVelocity in the y direction;
step 4.2, estimating the mixed state of each model by using the k moment targets obtained in the step 3.2And hybrid covariance matrix->According to whether the state equation is linear, the prior prediction value is filtered by using the formula (13) or (14) respectively>Solving for a priori covariance estimate +.>
The method is implemented according to the following steps:
step 5.1, filtering priori predicted values obtained in step 4.2 according to the target state equations (1) (3), the measurement equation (5) and />The system equation is extended as follows:
wherein ,
wherein ,for a priori prediction covariance matrix->A lower triangular matrix obtained by Cholesky decomposition,for measuring covariance matrix R k+1 Performing Cholesky decomposition to obtain a lower triangular matrix;
the two sides of the formula (16) are respectively multiplied byThe following expansion equation is obtained:
wherein
Wherein I represents an identity matrix;
step 5.2, designing a cost function based on maximum entropy:
wherein ,Gσ Representing a Gaussian kernel function, L isAnd l=n+m, n representing the state dimension, m representing the measurement dimension, +.>Representation->The r element of (2)>Representation->R line of (2);
the optimal estimate of the target state is:
deriving the optimal target state for equation (21) is:
wherein ,diagonal matrix representing kernel function structure related to target motion state prediction +.>A diagonal matrix representing the construction of a kernel function related to target measurements,/->Is-> Lines 1 to n, +.>Is->N+1th to n+mth rows;
step 5.3, let iteration number t=1, use the priori predicted value obtained in step 4.2 as the initial value of the iterative process:and determining a kernel bandwidth value sigma and a threshold epsilon;
step 5.4, calculating the state estimation value obtained by the t-th iteration process by using the following formula
wherein ,
step 5.5, comparing whether the estimated values of the t iteration and the t-1 iteration obtained in the step 5.4 meet the formula (31), and if not, repeating the step 5.4; if yes, ending the iterative process, and setting the iteration result of the t timeAs a filtered posterior update value +.>And updating covariance matrix +.>
And 6, updating the model probability: calculating likelihood functions of the models, obtaining probability updating values of the models according to a Bayesian probability formula, and adopting the probability change rate of the models to construct a correction function to correct transition probabilities among the models; the method is implemented according to the following steps:
step 6.1, filtering a priori prediction value obtained according to step 4.2And covariance matrix->Calculating likelihood functions of the models:
step 6.2, updating the probability of each model according to a Bayesian probability formula by the following formula:
step 6.3, constructing a transition probability correction function of the model j:
step 6.4, using the correction function of the model j to correct the transition probability of other models to the model:
normalizing the transition probability:
The method is implemented according to the following steps:
step 7.1, updating the value according to the filtering posterior obtained in step 5.5And the updated probability of each model obtained in the step 6.2 is fused with the state estimation of each model:
step 7.2, updating the value of the covariance matrix obtained in step 5.5Covariance of each model was fused:
and 8, repeating the steps 3-7 until the target tracking is finished, and obtaining a final target tracking effect as shown in fig. 2 and 3 (a) -3 (b).
As can be seen from fig. 2, under the non-gaussian noise condition, the maximum entropy extended kalman filtering algorithm based on the interactive multi-model is close to the real state of the target, which indicates that the algorithm can track the maneuvering target well, and as can be seen from fig. 3 (a) and fig. 3 (b), the position root mean square error and the speed root mean square error of the maximum entropy extended kalman filtering algorithm based on the interactive multi-model are smaller than the corresponding error value of the extended kalman filtering algorithm based on the interactive multi-model, which indicates that the accuracy of the maximum entropy extended kalman filtering algorithm based on the interactive multi-model is higher and the tracking effect is better.
Claims (8)
1. The maneuvering target tracking method based on maximum correlation entropy extended Kalman filtering is characterized by comprising the following specific steps of:
step 1, establishing a model set, a target state equation and a measurement equation of an interactive multi-model algorithm;
step 2, initializing parameters: initializing model probability, a target state vector, a covariance matrix and a Markov state transition matrix of each model at the moment k=1;
step 3, input interaction: calculating the mixing probability among the k moment models according to the model probability, the target state vector, the covariance matrix and the Markov state transition matrix of each model at the k moment to obtain a mixing state estimated value and a mixing covariance matrix of each model of the target at the k moment;
step 4, filter prior estimation: linearizing nonlinear models in the model set by using an extended Kalman filter, and predicting each model in one step to obtain a priori predicted value of each model of which the target is at the time of k+1;
step 5, updating a filter posterior: according to the target state equation obtained in the step 1, the measurement equation and the priori predicted value obtained in the step 4, designing a cost function based on the maximum entropy criterion, and obtaining posterior updated values of the k+1 moment targets in the models through an iteration method;
and 6, updating the model probability: calculating likelihood functions of the models, obtaining probability updating values of the models according to a Bayesian probability formula, and adopting the probability change rate of the models to construct a correction function to correct transition probabilities among the models;
step 7, fusing output results: fusing posterior update values and covariance of each model, and outputting state vector of target at k+1 timeAnd covariance matrix P k+1 ;
And 8, repeating the steps 3-7 until the target tracking process is finished.
2. The maneuvering target tracking method based on maximum correlation entropy expansion kalman filtering according to claim 1, wherein the step 1 is specifically as follows:
step 1.1, the target state at time k is expressed asThe linear CV model of the uniform linear motion of the target is simulated, and a target state equation is constructed as follows:
x k+1 =F k x k +G k w k (1)
wherein the state vector x k Wherein x, y respectively represent the position of the target along the x direction and the position along the y direction;is the speed of the target in the x-direction, the speed in the y-direction; t represents the sampling interval, w k Mean 0, covariance Q Gaussian white noise, F k Representing a state transition matrix, G k Representing a noise driving matrix;
step 1.2 toThe method comprises the steps of representing a target state at the moment k, simulating a nonlinear CT model of coordinated turning motion of a target with unknown angular velocity, and constructing a target state equation:
x k+1 =f(x k )+G k w k (3)
wherein the angular velocity omega k Not a constant;
step 1.3, taking the x-direction position and the y-direction position as observables, and establishing a measurement equation with non-Gaussian noise:
z k =H k x k +v k (5)
wherein ,zk Observation vector representing k time, measurement noise v k Is non-Gaussian noise obeying a mixed Gaussian distribution, and is specifically expressed as follows: v k ~(1-α)N(0,R 1k )+αN(0,R 2k ) Wherein alpha represents weight, the value range is 0-1, N (DEG,) represents normal distribution, R 1k To measure the noise covariance, R 2k To measure the covariance of noise when it is abnormally disturbed.
3. The maneuvering target tracking method based on maximum correlation entropy expansion kalman filtering according to claim 1, wherein the step 2 is specifically implemented in the following manner:
initializing parameter values, including determining a probability that the target is in model i at time k=1State estimation->Covariance matrix->And a Markov state transition matrix pi k ={π ij,k } M×M Wherein each element pi ij,k The transition probability of the system from the model i to the model j at the moment k is represented, and M represents the total number of the models.
4. A maneuvering target tracking method based on maximum correlation entropy expansion kalman filtering according to claim 3, wherein step 3 is specifically as follows:
step 3.1, according to the model probabilityAnd a Markov state transition matrix pi k Calculating the mixing probability between the k moment models by the formulas (7), (8)>
wherein ,representing the probability that the target is in the model j after the input interaction;
step 3.2, according to the target state at the time of kCovariance matrix->And the mixing probability between the models obtained in step 3.1 +.>Obtaining the mixed state estimation +.A mixed state estimation of the k moment target in each model is obtained by the formulas (9) (10) respectively>And hybrid covariance matrix->
5. The maneuvering target tracking method based on maximum correlation entropy expansion kalman filtering according to claim 1, wherein the step 4 is specifically implemented in the following manner:
step 4.1, calculating a Jacobian matrix of a nonlinear system state equation according to formulas (11) and (12), and linearizing the nonlinear state equation (3):
wherein the partial derivatives of the state quantities over angular rate are:
wherein ,ωk Is angular velocity, T is sampling interval,The speed in the x direction and the speed in the y direction are respectively;
step 4.2, estimating the mixed state of each model by using the k moment targets obtained in the step 3.2And hybrid covariance matrix->According to whether the state equation is linear, the prior prediction value is filtered by using the formula (13) or (14) respectively>Solving for a priori covariance estimate +.>
6. The maneuvering target tracking method based on maximum correlation entropy expansion kalman filtering according to claim 1, wherein the step 5 is specifically as follows:
step 5.1, filtering priori predicted values obtained in step 4.2 according to the target state equations (1) (3), the measurement equation (5) and />The system equation is extended as follows:
wherein ,
wherein ,for a priori prediction covariance matrix->Lower triangular matrix obtained by Cholesky decomposition>For measuring covariance matrix R k+1 Performing Cholesky decomposition to obtain a lower triangular matrix;
the two sides of the formula (16) are respectively multiplied byThe following expansion equation is obtained:
wherein
Wherein I represents an identity matrix;
step 5.2, designing a cost function based on maximum entropy:
wherein ,Gσ Representing a Gaussian kernel function, L isAnd l=n+m, n representing the state dimension, m representing the measurement dimension,representation->The r element of (2)>Representation->R line of (2);
the optimal estimate of the target state is:
Deriving the optimal target state for equation (21) is:
wherein ,diagonal matrix representing kernel function structure related to target motion state prediction +.>A diagonal matrix representing the construction of a kernel function related to target measurements,/->Is->Lines 1 to n, +.>Is->N+1th to n+mth rows;
step 5.3, let iteration number t=1, use the priori predicted value obtained in step 4.2 as the initial value of the iterative process:and determining a kernel bandwidth value sigma and a threshold epsilon;
step 5.4, calculating the state estimation value obtained by the t-th iteration process by using the following formula
wherein ,
step 5.5, comparing whether the estimated values of the t iteration and the t-1 iteration obtained in the step 5.4 meet the formula (31), and if not, repeating the step 5.4; if yes, ending the iterative process, and setting the iteration result of the t timeAs a filtered posterior update value +.>And updating covariance matrix +.>
7. The maneuvering target tracking method based on maximum correlation entropy expansion kalman filtering according to claim 1, wherein the step 6 is specifically implemented in the following manner:
step 6.1, filtering a priori prediction value obtained according to step 4.2And covariance matrix->Calculating likelihood functions of the models: />
Step 6.2, updating the probability of each model according to a Bayesian probability formula by the following formula:
step 6.3, constructing a transition probability correction function of the model j:
step 6.4, using the correction function of the model j to correct the transition probability of other models to the model:
normalizing the transition probability:
8. the maneuvering target tracking method based on maximum correlation entropy expansion kalman filtering according to claim 1, wherein the step 7 is specifically as follows:
step 7.1, updating the value according to the filtering posterior obtained in step 5.5And the updated probability of each model obtained in the step 6.2 is fused with the state estimation of each model:
step 7.2, updating the value of the covariance matrix obtained in step 5.5Covariance of each model was fused:
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202310063017.0A CN116047498A (en) | 2023-01-16 | 2023-01-16 | Maneuvering target tracking method based on maximum correlation entropy extended Kalman filtering |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202310063017.0A CN116047498A (en) | 2023-01-16 | 2023-01-16 | Maneuvering target tracking method based on maximum correlation entropy extended Kalman filtering |
Publications (1)
Publication Number | Publication Date |
---|---|
CN116047498A true CN116047498A (en) | 2023-05-02 |
Family
ID=86113154
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202310063017.0A Pending CN116047498A (en) | 2023-01-16 | 2023-01-16 | Maneuvering target tracking method based on maximum correlation entropy extended Kalman filtering |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN116047498A (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN117353705A (en) * | 2023-10-08 | 2024-01-05 | 哈尔滨工业大学 | One-step delay tracking filtering method and system with unknown delay probability |
CN117565877A (en) * | 2024-01-19 | 2024-02-20 | 华东交通大学 | Layered fusion estimation method for running state of distributed driving electric automobile |
-
2023
- 2023-01-16 CN CN202310063017.0A patent/CN116047498A/en active Pending
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN117353705A (en) * | 2023-10-08 | 2024-01-05 | 哈尔滨工业大学 | One-step delay tracking filtering method and system with unknown delay probability |
CN117565877A (en) * | 2024-01-19 | 2024-02-20 | 华东交通大学 | Layered fusion estimation method for running state of distributed driving electric automobile |
CN117565877B (en) * | 2024-01-19 | 2024-04-09 | 华东交通大学 | Layered fusion estimation method for running state of distributed driving electric automobile |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN116047498A (en) | Maneuvering target tracking method based on maximum correlation entropy extended Kalman filtering | |
CN111178385B (en) | Target tracking method for robust online multi-sensor fusion | |
CN108896986B (en) | Measurement conversion sequential filtering maneuvering target tracking method based on predicted value | |
CN107402381B (en) | Iterative self-adaptive multi-maneuvering target tracking method | |
CN109633590B (en) | Extended target tracking method based on GP-VSMM-JPDA | |
CN107688179B (en) | Comprehensive probability data interconnection method based on Doppler information assistance | |
CN111291471B (en) | Constraint multi-model filtering method based on L1 regular unscented transformation | |
WO2018098926A1 (en) | Multi-target tracking method and system applicable to flicker noise | |
CN112114308A (en) | Space-time joint target tracking method for sector-scan radar | |
CN111259332B (en) | Fuzzy data association method and multi-target tracking method in clutter environment | |
CN111693984A (en) | Improved EKF-UKF moving target tracking method | |
CN115204212A (en) | Multi-target tracking method based on STM-PMBM filtering algorithm | |
CN112034445B (en) | Vehicle motion trail tracking method and system based on millimeter wave radar | |
CN111798494B (en) | Maneuvering target robust tracking method under generalized correlation entropy criterion | |
CN110989341B (en) | Constraint auxiliary particle filtering method and target tracking method | |
Konatowski et al. | A comparison of estimation accuracy by the use of KF, EKF & UKF filters | |
Gurajala et al. | Derivation of the Kalman filter in a Bayesian filtering perspective | |
CN107886058B (en) | Noise-related two-stage volume Kalman filtering estimation method and system | |
CN115828533A (en) | Interactive multi-model robust filtering method based on Student's t distribution | |
Hu et al. | Hybrid sampling-based particle filtering with temporal constraints | |
CN113189578B (en) | Extended target tracking method | |
CN113191427B (en) | Multi-target vehicle tracking method and related device | |
CN111262556B (en) | Multi-target tracking method for simultaneously estimating unknown Gaussian measurement noise statistics | |
CN113030945A (en) | Phased array radar target tracking method based on linear sequential filtering | |
CN108572362B (en) | TWS radar space-time joint association tracking method and device |
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 |