Nothing Special   »   [go: up one dir, main page]

CN109687844A - A kind of intelligent maneuver method for tracking target - Google Patents

A kind of intelligent maneuver method for tracking target Download PDF

Info

Publication number
CN109687844A
CN109687844A CN201810938387.3A CN201810938387A CN109687844A CN 109687844 A CN109687844 A CN 109687844A CN 201810938387 A CN201810938387 A CN 201810938387A CN 109687844 A CN109687844 A CN 109687844A
Authority
CN
China
Prior art keywords
model
probability
covariance
transition probability
steps
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
Application number
CN201810938387.3A
Other languages
Chinese (zh)
Inventor
谢国
孙澜澜
刘涵
王文卿
梁莉莉
金永泽
张永艳
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Xian University of Technology
Original Assignee
Xian University of Technology
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Xian University of Technology filed Critical Xian University of Technology
Priority to CN201810938387.3A priority Critical patent/CN109687844A/en
Publication of CN109687844A publication Critical patent/CN109687844A/en
Pending legal-status Critical Current

Links

Classifications

    • HELECTRICITY
    • H03ELECTRONIC CIRCUITRY
    • H03HIMPEDANCE NETWORKS, e.g. RESONANT CIRCUITS; RESONATORS
    • H03H21/00Adaptive networks
    • H03H21/0012Digital adaptive filters
    • H03H21/0025Particular filtering methods
    • H03H21/0029Particular filtering methods based on statistics
    • H03H21/003KALMAN filters

Landscapes

  • Physics & Mathematics (AREA)
  • Probability & Statistics with Applications (AREA)
  • Feedback Control In General (AREA)

Abstract

The invention discloses a kind of intelligent maneuver method for tracking target, transition probability matrix and model probability are modeled and initialized to the maneuvering target tracked first, input interaction is carried out to system mode and covariance according to the transition probability of system and model probability, obtain the filter input of admixture estimated value and its covariance estimated value as M concurrent working based on model, calculate state filtering value and its covariance based on model, then model probability is updated, computation model probability change rate, and transition probability is modified using model probability change rate, then it is normalized, finally carry out output interaction, obtain final state estimation and its variance, the present invention solves existing in the prior art since the fixed caused interacting multiple model algorithm of transition probability matrix is smooth The weak problem of noise immune, realizes the adaptive of transition probability, effectively improves the ability of Tracking Maneuvering Targets.

Description

Intelligent maneuvering target tracking method
Technical Field
The invention belongs to the technical field of maneuvering target tracking, and particularly relates to an intelligent maneuvering target tracking method.
Background
Maneuvering target tracking is always one of the hot points of research, and an interactive multi-model algorithm is the most common maneuvering target tracking algorithm at present, and has the advantage of realizing concurrent filtering of multiple models, and the system state estimation value is the weighting of the state estimation values based on the models. The interactive multi-model algorithm is characterized in that jumping between models is guided by a finite Markov chain. The interactive multi-model algorithm is essentially a filtering algorithm with special parameters, which are the transition probability matrix and have profound influence on the performance of the algorithm. The transition probability matrix of the standard interactive multi-model algorithm is set by prior information and is fixed, which obviously does not accord with the actual situation and influences the improvement of the algorithm performance. The interactive multi-model algorithm can generate a plurality of posterior information after iteration, the adaptation of the algorithm can be realized by correcting the transition probability by adopting the posterior information, and the improvement is realized by amplifying the inertia of the system essentially, so that the noise smoothing capability of the algorithm can be effectively improved, and the accuracy of tracking the maneuvering target is improved.
Disclosure of Invention
The invention aims to provide an intelligent maneuvering target tracking method, which solves the problem of weak smooth noise capability of an interactive multi-model algorithm caused by fixed transition probability matrix in the prior art, realizes the self-adaption of transition probability and effectively improves the capability of tracking maneuvering targets.
The technical scheme adopted by the invention is that the intelligent maneuvering target tracking method is implemented according to the following steps:
step 1, modeling a maneuvering target to be tracked to obtain M models, and initializing a transition probability matrix and a model probability;
step 2, performing input interaction on the system state and the covariance according to the transition probability and the model probability of the system to obtain a model-based mixed state estimation valueAnd its covariance estimate P0j(k-1);
Step 3, the model-based mixed state estimation value obtained by the calculation in the step 2And its covariance estimate P0j(k-1) works as M pieces in parallelCalculating model-based state filter valuesAnd its covariance Pj(k);
Step 4, updating the model probability according to the posterior information obtained in the step 3;
step 5, calculating a model probability change rate according to the model probability obtained in the step 4, correcting the transition probability by adopting the model probability change rate, and then carrying out normalization processing on the transition probability;
step 6, performing output interaction on the state estimation value based on the model and the variance thereof obtained in the step 3 and the model probability obtained in the step 4 to obtain a final state estimation value and the variance thereof;
and 7, repeating the steps 2 to 6 to iteratively track the maneuvering target until the iteration is finished.
The present invention is also characterized in that,
the step 1 is implemented according to the following steps:
step 1.1, modeling a system of a target needing maneuvering tracking to obtain M models:
x(k+1)=F(k,mk)x(k,mk)+G(k,mk)w(k,mk)
z(k)=H(k,mk)x(mk)+v(k,mk)
wherein,andrespectively representing a state vector and a measurement vector of a system at the moment k, n and m are positive integers, mkRepresents the current model of the system at time k, and the value is {1, 2.., M }, F (k, M) }k)、G(k,mk)、H(k,mk) Are all model mkSystem matrix at time k, w (k, m)k)、v(k,mk) Respectively represent the model mkThe process noise and the measurement noise of (1), which are uncorrelated white Gaussian noise with a mean value of zero and a variance matrix of Q (k, m), respectivelyk),R(k,mk);
Step 1.2, initializing a transition probability matrix pi:
Π=[πij]M×M
wherein, piijRepresenting the probability of model i jumping to model j.
Step 1.3, initializing a transition probability matrix mu:
μ=[μi]1×M
wherein, muiRepresenting the probability of model i.
The step 2 is implemented according to the following steps:
step 2.1, according to the transition probability matrix pi and the model probability muiCalculating the mixing probability mui|j
Step 2.2, obtaining the mixing probability mu according to the step 2.1i|jComputing model-based hybrid state estimatesAnd its covariance estimate P0j(k-1)。
Step 3 is specifically implemented according to the following steps:
the estimated value of the mixed state based on the model calculated in the step 2And its covariance estimate P0j(k-1) calculating model-based state filter values as input to M parallel-operating Kalman filtersAnd its covariance Pj(k) Error mean vj(k) And error covariance Sj(k).。
Step 4 is specifically implemented according to the following steps:
step 4.1, obtaining the error mean value v based on the model according to the step 3j(k) And error covariance Sj(k) Computing a model-based likelihood function Λj(k);
Step 4.2, the likelihood function Lambda based on the model obtained by calculation according to the step 4.1j(k) Updating the model probability mui(k)。
Step 5 is specifically implemented according to the following steps:
step 5.1, calculating the probability change rate of the model according to the model probability obtained in the step 4, and defining a transition probability correction function f as follows:
step 5.2, the transfer probability is corrected by adopting the correction function obtained in the step 5.1, and normalization processing is carried out to obtain the transfer probability pi of intelligent processingij(k):
πij(k)'=fj(k)*πij(k-1),i=1,2,…M
Normalization:
step 6 is implemented according to the following steps:
filtering the model-based state obtained in step 3And its covariance Pj(k) And the model probability mu obtained in the step 4i(k) Weighting to obtain the final state estimation value
The invention has the beneficial effects that the transition probability matrix in the standard interactive multi-model algorithm is determined according to the prior information, which does not accord with the actual situation and seriously influences the performance of the algorithm. According to research, the model probability change rate is adopted to generate a correction function, the transition probability is corrected, the transition probability self-adaption can be achieved, noise is effectively smoothed, model misjudgment is reduced, algorithm performance is improved, and therefore the maneuvering target tracking capacity is improved.
Drawings
FIG. 1 is a flow chart of an intelligent maneuvering target tracking method of the invention;
FIG. 2 is a schematic diagram of a standard interactive multi-model algorithm in the intelligent maneuvering target tracking method of the invention.
Detailed Description
The present invention will be described in detail below with reference to the accompanying drawings and specific embodiments.
The invention discloses an intelligent maneuvering target tracking method, which is implemented by the following steps of:
step 1, modeling a maneuvering target to be tracked to obtain M models, initializing a transition probability matrix and a model probability, and specifically implementing the following steps:
step 1.1, modeling a system of a target needing maneuvering tracking to obtain M models:
x(k+1)=F(k,mk)x(k,mk)+G(k,mk)w(k,mk)
z(k)=H(k,mk)x(mk)+v(k,mk)
wherein,andrespectively representing a state vector and a measurement vector of a system at the moment k, n and m are positive integers, mkRepresents the current model of the system at time k, and the value is {1, 2.., M }, F (k, M) }k)、G(k,mk)、H(k,mk) Are all model mkSystem matrix at time k, w (k, m)k)、v(k,mk) Respectively represent the model mkThe process noise and the measurement noise of (1), which are uncorrelated white Gaussian noise with a mean value of zero and a variance matrix of Q (k, m), respectivelyk),R(k,mk);
Step 1.2, initializing a transition probability matrix pi:
Π=[πij]M×M
wherein, piijRepresenting the probability of model i jumping to model j.
Step 1.3, initializing a transition probability matrix mu:
μ=[μi]1×M
wherein, muiRepresenting the probability of model i;
step 2, performing input interaction on the system state and the covariance according to the transition probability and the model probability of the system to obtain a model-based mixed state estimation valueAnd its covariance estimate P0j(k-1), specifically comprising the following steps:
step 2.1, according to the transition probability matrix pi and the model probability muiCalculating the mixing probability mui|j
Step 2.2, obtaining the mixing probability mu according to the step 2.1i|jComputing model-based hybrid state estimatesAnd its covariance estimate P0j(k-1);
Step 3, the model-based mixed state estimation value obtained by the calculation in the step 2And its covariance estimate P0j(k-1) computing model-based state filter values as M filter inputs operating in parallelAnd its covariance Pj(k) The method is implemented according to the following steps:
the estimated value of the mixed state based on the model calculated in the step 2And its covariance estimate P0j(k-1) calculating model-based state filter values as input to M parallel-operating Kalman filtersAnd its covariance Pj(k) Error mean vj(k) And error covariance Sj(k).;
And 4, updating the model probability according to the posterior information obtained in the step 3, and specifically implementing the following steps:
step 4.1, obtaining the error mean value v based on the model according to the step 3j(k) And error covariance Sj(k) Computing a model-based likelihood function Λj(k);
Step 4.2, the likelihood function Lambda based on the model obtained by calculation according to the step 4.1j(k) Updating the model probability mui(k);
Step 5, calculating a model probability change rate according to the model probability obtained in the step 4, correcting the transition probability by adopting the model probability change rate, and then carrying out normalization processing on the transition probability, wherein the method is implemented according to the following steps:
step 5.1, calculating the probability change rate of the model according to the model probability obtained in the step 4, and defining a transition probability correction function f as follows:
step 5.2, the transfer probability is corrected by adopting the correction function obtained in the step 5.1, and normalization processing is carried out to obtain the transfer probability pi of intelligent processingij(k):
πij(k)'=fj(k)*πij(k-1),i=1,2,…M
Normalization:
and 6, performing output interaction on the state estimation value and the variance thereof based on the model obtained in the step 3 and the model probability obtained in the step 4 to obtain a final state estimation value and a variance thereof, and specifically performing the following steps:
filtering the model-based state obtained in step 3And its covariance Pj(k) And the model probability mu obtained in the step 4i(k) Weighting to obtain the final state estimation value
And 7, repeating the steps 2 to 6 to iteratively track the maneuvering target until the iteration is finished.
The interactive multi-model algorithm is the most widely applied maneuvering target tracking algorithm, the transition probability matrix is an important parameter of the interactive multi-model algorithm, the parameter has great influence on the performance of the algorithm, but the parameter in the standard interactive multi-model algorithm is set by prior information and is fixed, which is not in accordance with the actual situation, in order to improve the performance of the interactive multi-model algorithm, the scheme provides that the posterior information is utilized to realize transition probability self-adaptation, the inertia of the system is increased, and therefore the capability of the algorithm for smoothing noise is effectively improved.

Claims (7)

1. An intelligent maneuvering target tracking method is characterized by being implemented according to the following steps:
step 1, modeling a maneuvering target to be tracked to obtain M models, and initializing a transition probability matrix and a model probability;
step 2, performing input interaction on the system state and the covariance according to the transition probability and the model probability of the system to obtain a model-based mixed state estimation valueAnd its covariance estimate P0j(k-1);
Step 3, the model-based mixed state estimation value obtained by the calculation in the step 2And its covariance estimate P0j(k-1) computing model-based state filter values as M filter inputs operating in parallelAnd its covariance Pj(k);
Step 4, updating the model probability according to the posterior information obtained in the step 3;
step 5, calculating a model probability change rate according to the model probability obtained in the step 4, correcting the transition probability by adopting the model probability change rate, and then carrying out normalization processing on the transition probability;
step 6, performing output interaction on the state estimation value based on the model and the variance thereof obtained in the step 3 and the model probability obtained in the step 4 to obtain a final state estimation value and the variance thereof;
and 7, repeating the steps 2 to 6 to iteratively track the maneuvering target until the iteration is finished.
2. The intelligent maneuvering target tracking method according to claim 1, characterized in that the step 1 is implemented specifically according to the following steps:
step 1.1, modeling a system of a target needing maneuvering tracking to obtain M models:
x(k+1)=F(k,mk)x(k,mk)+G(k,mk)w(k,mk)
z(k)=H(k,mk)x(mk)+v(k,mk)
wherein,andrespectively representing a state vector and a measurement vector of a system at the moment k, n and m are positive integers, mkRepresents the current model of the system at time k, and the value is {1, 2.., M }, F (k, M) }k)、G(k,mk)、H(k,mk) Are all model mkSystem matrix at time k, w (k, m)k)、v(k,mk) Respectively represent the model mkThe process noise and the measurement noise of (1), which are uncorrelated white Gaussian noise with a mean value of zero and a variance matrix of Q (k, m), respectivelyk),R(k,mk);
Step 1.2, initializing a transition probability matrix pi:
Π=[πij]M×M
wherein, piijRepresenting the probability of model i jumping to model j.
Step 1.3, initializing a transition probability matrix mu:
μ=[μi]1×M
wherein, muiRepresenting the probability of model i.
3. The intelligent maneuvering target tracking method according to claim 2, characterized in that the step 2 is specifically implemented according to the following steps:
step 2.1, according to the transition probability matrix pi and the model probability muiCalculating the mixing probability mui|j
Step 2.2, obtaining the mixing probability mu according to the step 2.1i|jComputing model-based hybrid state estimatesAnd its covariance estimate P0j(k-1)。
4. The intelligent maneuvering target tracking method according to claim 3, characterized in that the step 3 is implemented specifically according to the following steps:
calculating the step 2Resulting model-based hybrid state estimatesAnd its covariance estimate P0j(k-1) calculating model-based state filter values as input to M parallel-operating Kalman filtersAnd its covariance Pj(k) Error mean vj(k) And error covariance Sj(k).。
5. The intelligent maneuvering target tracking method according to claim 4, characterized in that the step 4 is specifically implemented according to the following steps:
step 4.1, obtaining the error mean value v based on the model according to the step 3j(k) And error covariance Sj(k) Computing a model-based likelihood function Λj(k);
Step 4.2, the likelihood function Lambda based on the model obtained by calculation according to the step 4.1j(k) Updating the model probability mui(k)。
6. The intelligent maneuvering target tracking method according to claim 5, characterized in that the step 5 is implemented specifically according to the following steps:
step 5.1, calculating the probability change rate of the model according to the model probability obtained in the step 4, and defining a transition probability correction function f as follows:
step 5.2, the transfer probability is corrected by adopting the correction function obtained in the step 5.1, and normalization processing is carried out to obtain the transfer probability pi of intelligent processingij(k):
πij(k)'=fj(k)*πij(k-1),i=1,2,…M
Normalization:
7. the intelligent maneuvering target tracking method according to claim 6, characterized in that the step 6 is implemented specifically according to the following steps:
filtering the model-based state obtained in step 3And its covariance Pj(k) And the model probability mu obtained in the step 4i(k) Weighting to obtain the final state estimation value
CN201810938387.3A 2018-08-17 2018-08-17 A kind of intelligent maneuver method for tracking target Pending CN109687844A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201810938387.3A CN109687844A (en) 2018-08-17 2018-08-17 A kind of intelligent maneuver method for tracking target

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201810938387.3A CN109687844A (en) 2018-08-17 2018-08-17 A kind of intelligent maneuver method for tracking target

Publications (1)

Publication Number Publication Date
CN109687844A true CN109687844A (en) 2019-04-26

Family

ID=66185564

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201810938387.3A Pending CN109687844A (en) 2018-08-17 2018-08-17 A kind of intelligent maneuver method for tracking target

Country Status (1)

Country Link
CN (1) CN109687844A (en)

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110261859A (en) * 2019-06-25 2019-09-20 北京中科海讯数字科技股份有限公司 A kind of static alternating state method for tracking target of underwater manoeuvre
CN110378411A (en) * 2019-07-16 2019-10-25 浙江大学 Maneuvering target tracking method under a kind of support vector machines auxiliary water based on interactive multi-model
CN110426671A (en) * 2019-07-04 2019-11-08 重庆邮电大学 Model probability modified IMM method for tracking target and device in real time are based in WSN
CN112615604A (en) * 2020-12-08 2021-04-06 苏州挚途科技有限公司 Filtering method and device of intelligent driving perception system and electronic equipment

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20030200065A1 (en) * 2001-04-20 2003-10-23 Li Luo Wen Maneuvering target tracking method via modifying the interacting multiple model (IMM) and the interacting acceleration compensation (IAC) algorithms
CN102568004A (en) * 2011-12-22 2012-07-11 南昌航空大学 Tracking algorithm for high maneuvering targets
CN103853908A (en) * 2012-12-04 2014-06-11 中国科学院沈阳自动化研究所 Self-adapting interactive multiple model mobile target tracking method
CN103955600A (en) * 2014-04-03 2014-07-30 深圳大学 Target tracking method and truncated integral Kalman filtering method and device
CN106933106A (en) * 2016-05-26 2017-07-07 哈尔滨工程大学 A kind of method for tracking target based on fuzzy control Multiple Models Algorithm
CN107193009A (en) * 2017-05-23 2017-09-22 西北工业大学 A kind of many UUV cooperative systems underwater target tracking algorithms of many interaction models of fuzzy self-adaption
CN107704432A (en) * 2017-07-28 2018-02-16 西安理工大学 A kind of adaptive Interactive Multiple-Model method for tracking target of transition probability

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20030200065A1 (en) * 2001-04-20 2003-10-23 Li Luo Wen Maneuvering target tracking method via modifying the interacting multiple model (IMM) and the interacting acceleration compensation (IAC) algorithms
CN102568004A (en) * 2011-12-22 2012-07-11 南昌航空大学 Tracking algorithm for high maneuvering targets
CN103853908A (en) * 2012-12-04 2014-06-11 中国科学院沈阳自动化研究所 Self-adapting interactive multiple model mobile target tracking method
CN103955600A (en) * 2014-04-03 2014-07-30 深圳大学 Target tracking method and truncated integral Kalman filtering method and device
CN106933106A (en) * 2016-05-26 2017-07-07 哈尔滨工程大学 A kind of method for tracking target based on fuzzy control Multiple Models Algorithm
CN107193009A (en) * 2017-05-23 2017-09-22 西北工业大学 A kind of many UUV cooperative systems underwater target tracking algorithms of many interaction models of fuzzy self-adaption
CN107704432A (en) * 2017-07-28 2018-02-16 西安理工大学 A kind of adaptive Interactive Multiple-Model method for tracking target of transition probability

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
许登荣;程水英;包守亮;: "自适应转移概率交互式多模型跟踪算法", 电子学报, no. 09, pages 60 - 67 *

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110261859A (en) * 2019-06-25 2019-09-20 北京中科海讯数字科技股份有限公司 A kind of static alternating state method for tracking target of underwater manoeuvre
CN110261859B (en) * 2019-06-25 2023-10-31 北京中科海讯数字科技股份有限公司 Underwater maneuvering static alternating state target tracking method
CN110426671A (en) * 2019-07-04 2019-11-08 重庆邮电大学 Model probability modified IMM method for tracking target and device in real time are based in WSN
CN110378411A (en) * 2019-07-16 2019-10-25 浙江大学 Maneuvering target tracking method under a kind of support vector machines auxiliary water based on interactive multi-model
CN110378411B (en) * 2019-07-16 2021-03-23 浙江大学 Method for assisting underwater maneuvering target tracking by support vector machine based on interactive multi-model
CN112615604A (en) * 2020-12-08 2021-04-06 苏州挚途科技有限公司 Filtering method and device of intelligent driving perception system and electronic equipment

Similar Documents

Publication Publication Date Title
CN109687844A (en) A kind of intelligent maneuver method for tracking target
CN108599737B (en) Design method of nonlinear Kalman filter of variational Bayes
CN100587719C (en) Method for tracking dimension self-adaptation video target with low complex degree
Aghasadeghi et al. Maximum entropy inverse reinforcement learning in continuous state spaces with path integrals
CN106933106B (en) Target tracking method based on fuzzy control multi-model algorithm
CN107704432A (en) A kind of adaptive Interactive Multiple-Model method for tracking target of transition probability
CN106487358B (en) A kind of maneuvering target turning tracking
CN108682023A (en) Close coupling Unscented kalman tracking filter algorithm based on Elman neural networks
CN109410247A (en) A kind of video tracking algorithm of multi-template and adaptive features select
CN106021194B (en) A kind of multi-sensor multi-target tracking bias estimation method
CN112986977B (en) Method for overcoming radar extended Kalman track filtering divergence
CN112051569B (en) Radar target tracking speed correction method and device
CN112564557A (en) Control method, device and equipment of permanent magnet synchronous motor and storage medium
CN112965487A (en) Mobile robot trajectory tracking control method based on strategy iteration
CN102252672A (en) Nonlinear filtering method for underwater navigation
CN114063131A (en) GNSS/INS/wheel speed combined positioning real-time smoothing method
CN106934124B (en) Adaptive variable window method based on measurement change detection
CN108154231B (en) System error-based parameter self-tuning method for MISO full-format model-free controller
CN106845016B (en) One kind being based on event driven measurement dispatching method
CN108470016B (en) System state prediction method of industrial dryer
CN116527515A (en) Remote state estimation method based on polling protocol
CN115828533A (en) Interactive multi-model robust filtering method based on Student's t distribution
CN108507593A (en) A kind of dimensionality reduction RTS ellipsoid collection person's smoothing methods of INS errors model
CN110034746B (en) Kalman filtering method based on maximum collaborative entropy
CN106709569A (en) Parameter estimation method for FitzHugh-Nagumo neuron system

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
RJ01 Rejection of invention patent application after publication
RJ01 Rejection of invention patent application after publication

Application publication date: 20190426