CN110261859A - A kind of static alternating state method for tracking target of underwater manoeuvre - Google Patents
A kind of static alternating state method for tracking target of underwater manoeuvre Download PDFInfo
- Publication number
- CN110261859A CN110261859A CN201910557541.7A CN201910557541A CN110261859A CN 110261859 A CN110261859 A CN 110261859A CN 201910557541 A CN201910557541 A CN 201910557541A CN 110261859 A CN110261859 A CN 110261859A
- Authority
- CN
- China
- Prior art keywords
- state
- model
- estimation
- covariance
- 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.)
- Granted
Links
- 230000003068 static effect Effects 0.000 title claims abstract description 18
- 238000000034 method Methods 0.000 title claims description 22
- 238000001914 filtration Methods 0.000 claims abstract description 35
- 230000003993 interaction Effects 0.000 claims abstract description 34
- 238000005259 measurement Methods 0.000 claims abstract description 13
- 239000011159 matrix material Substances 0.000 claims description 26
- 238000004364 calculation method Methods 0.000 claims description 10
- 230000007704 transition Effects 0.000 claims description 6
- 238000004088 simulation Methods 0.000 abstract description 3
- 150000001875 compounds Chemical class 0.000 abstract description 2
- 238000010276 construction Methods 0.000 description 3
- 238000004422 calculation algorithm Methods 0.000 description 2
- 230000007423 decrease Effects 0.000 description 2
- 230000001133 acceleration Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 230000002452 interceptive effect Effects 0.000 description 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 1
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
- G01S15/00—Systems using the reflection or reradiation of acoustic waves, e.g. sonar systems
- G01S15/66—Sonar tracking systems
-
- 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
- G06F17/15—Correlation function computation including computation of convolution operations
-
- 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
-
- 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
- G06F17/18—Complex mathematical operations for evaluating statistical data, e.g. average values, frequency distributions, probability functions, regression analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/20—Analysis of motion
- G06T7/277—Analysis of motion involving stochastic approaches, e.g. using Kalman filters
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20024—Filtering details
Abstract
The invention patent considers that various types of motion states include the static tracking performance of target, and the technical problem to be solved is that providing for the static alternate compound movement target of sub-aqua sport, a kind of applicability is good, the lower tracking of error.The first step, establish at the uniform velocity, speed change, static three kinds of state models;Second step is filtered initialization;Third step carries out input interaction;Input interaction of 4th step to three kinds of obtained models, difference parallel filtering;5th step by error in measurement covariance obtained in the previous step, measures and predicts that error obtains likelihood function, obtains this moment model probability by likelihood function;6th step carries out output interaction.From the point of view of the Simulation results using new interaction models, in the case where this patent proposes new interaction models, filtering error is smaller, and tracking performance is more excellent.
Description
Technical field
Invention belongs to sonar data emulation field, and in particular to a kind of static alternating state target following side of underwater manoeuvre
Method.
Background technique
Underwater manoeuvre and stationary state method for tracking target are to intersect progress between movement and stationary state for assessment
Designed by the tracking performance of submarine target, in the case where being influenced by the unfixed state change factor of submarine target, under water
The motion state of target, which can not be estimated or estimate to shift, leads to the decline of its tracking performance;As part submarine target is leaning on
It is close or far from during, may be moved at certain moment and certain moment it is static, in face of similar situation, need using being transported with it
The adaptable tracking of dynamic state.
Currently, when the unfixed variation of submarine target state, by using it is at the uniform velocity non-maneuver with accelerate maneuver modeling phase
In conjunction with method, it is assumed that target state many places, can be in the situation of practical similar movement state in both motion states
Preferably submarine target is tracked down.In original at the uniform velocity non-maneuver tracking combined with acceleration maneuver modeling,
It is excessively single to the state estimation model of special submarine target, only actual motion is estimated by commonly using target movement model
State, to obtain tracking output corresponding with model as a result, not fully considering the motion states pair such as static, snap maneuver
The influence of tracking performance can not make reasonable tracking and positioning to realistic objective, tracking performance is caused to decline.
Summary of the invention
1, goal of the invention
The invention patent considers that various types of motion states include the static tracking performance of target, skill to be solved
Art problem is to provide that a kind of applicability is good, the lower track side of error for the static alternate compound movement target of sub-aqua sport
Method.
2, technical solution
Motor-driven and stationary state target following is realized by the following method in the present invention:
More complex underwater movement objective motor pattern mainly includes non-maneuver uniform motion, motor-driven turning variable motion, quiet
Only equal prevailing operating states, therefore when carrying out target following to it, mainly consider associated state model.Consider one section
Target trajectory, when beginning target be in straight line at the uniform velocity with static alternating state, it is subsequent in turning accelerate with it is static, at the uniform velocity
Alternating state compares real trace and its filter tracking as a result, observation tracking performance by various forms of tracking.
The first step, establish at the uniform velocity, speed change, static three kinds of state models;
Second step is filtered initialization, initializes the Markov Transition Probabilities matrix and every kind of model of three kinds of models
Probability, initialize the state estimation and covariance estimation of every kind of model;
Third step carries out input interaction, and the model prediction probability respectively obtained with three kinds of models acquires input interaction probability;
Initialization or the state estimation that filters of previous moment are mixed with inputting to interact probability calculation and obtain the state of every kind of model
It closes;On this basis, state covariance is calculated using the state covariance that initialization or previous moment obtain to mix.
Input interaction of 4th step to three kinds of obtained models, biography is respectively adopted in parallel filtering, three kinds of state models respectively
System Kalman filtering obtains its error in measurement covariance, measures prediction error, state estimation, state variance estimated value;
5th step by error in measurement covariance obtained in the previous step, measures and predicts that error obtains likelihood function, by likelihood
Function obtains this moment model probability;
6th step carries out output interaction, state estimation, state variance estimation and the update obtained by each model parallel filtering
Model probability afterwards obtains state output estimation and state covariance estimation.
Specifically, uniform motion model is so established, and in Kalman Filter Estimation, is related to relevant parameter and matrix of variables such as
Under:
Wherein Fm1 is state-transition matrix, G1For input matrix, Q1For state error covariance matrix, H1To measure square
Battle array, R are error in measurement covariance matrix, and T is the scan period.
Initial state estimation:
Original state covariance:
Wherein, X (3), X (2) difference 3,2 state estimation of X-coordinate moment, Y (3), Y (2) difference 3,2 state of Y-coordinate moment
Estimation, D are diagonal entry data value in error in measurement covariance matrix.
Using classical kalman filter method to carrying out state estimation under at the uniform velocity model.Target trajectory is constructed, speed is set as-
15, scan period 2, error in measurement is set as 50, and filtering passes through 50 Monte Carlo calculations, and emulation obtains result such as attached drawing
1。
Respectively obtain two dimension comparison, filtering error mean value and the error criterion value of real trace and filtering estimation.
It can be seen that, at the uniform velocity model is acted normally at the uniform velocity, under stationary state alternating situation from attached drawing 1, is become in turning
Speed, at the uniform velocity, it is static alternately filter not normal, X under situation, the equal filtering error of Y direction is larger.
Specifically, it is at the uniform velocity built such that with variable motion model, in Kalman Filter Estimation, is related to relevant parameter and variable
Matrix is as follows:
Parameter meaning is identical as at the uniform velocity model in matrix.
Switching between two states is controlled using interacting multiple model algorithm.
X (k+1)=FmjX(k)+GjWj(k) j=1,2
Wherein WjIt (k) is k moment state error.
The conversion between these models, the transition probability matrix of Markov chain can be controlled with a Markov chain
Are as follows:
By multi-model Kalman filtering, state estimation and state covariance are obtained are as follows:
Wherein μj(k) it is in j-th of shape probability of state for k moment target,Pj(k/k) to filter respectively the when
The state estimation and state covariance matrix of j state.
Using multi-model kalman filter method at the uniform velocity, speed change interaction models carry out state estimation.To same construction
Track, filtering pass through 50 Monte Carlo calculations, and emulation obtains result such as attached drawing 2.
Respectively obtain two dimension comparison, filtering error mean value and the error criterion value of real trace and filtering estimation.
From attached drawing 2 it can be seen that, at the uniform velocity, speed change interaction models at the uniform velocity, stationary state alternating situation under X-axis filtering miss
Poor very little, Y-axis filtering have certain error, in turning speed change, at the uniform velocity, it is static alternately under situation, the equal filtering error of X, Y direction has
Centainly become larger.
By the above experiment simulation, it has been found that X-axis or Y-axis error become larger occur motor-driven or non-maneuver movement with
When interaction occurs for stationary state, the tracking performance to solve the problems, such as stationary state transfer is poor, is added in multimode interaction quiet
Only model is allowed to adapt to the variation of complicated submarine target stationary state.
In interdiction model Kalman Filter Estimation, it is related to relevant parameter and matrix of variables is as follows:
Specifically, the switching Markov Transition Probabilities matrix in the step 2, between three models are as follows:
At the uniform velocity, speed change, the Initial state estimation of interdiction model are as follows:
At the uniform velocity, speed change, the original state covariance of interdiction model are as follows:
Specifically, the step 3 carries out input interaction by state mixing, on the basis of three kinds of possible models respectively
According to subsequent time update motion state probability make respectively this model state mixing, wherein the update probability of models switching by
Parallel filtering is determined, obtains model prediction probability by the obtained likelihood function of parallel filtering and model probability, and then obtain
Input interaction probability.
State mixing is as follows:
Wherein μj/i(k) be to input interaction probability, i represent at the uniform velocity, speed change, interdiction model mark, j/i represents under i state
Subsequent time j state probability, for example it is the probability of stationary state or frogman by static shape that speed change state transfer is inscribed when frogman k
State comes back to speed change or at the uniform velocity shape probability of state.
Covariance mixing is as follows:
Wherein PjIt (k/k) is state variance.
In step 4, parallel filtering is carried out using traditional Kalman filtering, three kinds of model state estimations:
Three kinds of model state variances:
Pi(k+1/k+1)=[I-Ki(k+1)Hi(k+1)]Pi(k+1/k)
Model likelihood function:
Model probability:
Export interaction mode estimation and state covariance are as follows:
Meaning of parameters is identical with more than.
State estimation is carried out to new interaction models using multi-model kalman filter method.To same construction track, filter
Wave process passes through 50 Monte Carlo calculations, and emulation obtains result such as attached drawing 3.
Respectively obtain two dimension comparison, filtering error mean value and the error criterion value of real trace and filtering estimation.
From attached drawing 3 it can be seen that, new interaction models at the uniform velocity, stationary state alternating situation under or turning speed change,
At the uniform velocity, under static alternately situation, filtering track is clear, X, and the equal filtering error of Y direction is equal compared with both the above model method
It is smaller.
3, technical effect
From using new interaction models Simulation results from the point of view of, at the uniform velocity, accelerate interaction models algorithm compared with, with mesh
The progress of track is marked, X, Y-axis filtering error value are averagely reduced close to one times respectively, can sum up and propose new interaction in this patent
Under model, filtering error is smaller, and tracking performance is more excellent.
Detailed description of the invention
Fig. 1 constructs target trajectory, and speed is set as -15, and scan period 2, error in measurement is set as 50, and filtering passes through 50
Secondary Monte Carlo calculations, emulation obtain result.
Fig. 2 using multi-model kalman filter method at the uniform velocity, speed change interaction models carry out state estimation.To same structure
Track is made, filtering passes through 50 Monte Carlo calculations, and emulation obtains result.
Fig. 3 carries out state estimation to new interaction models with multi-model kalman filter method.To same construction track, filter
Wave process passes through 50 Monte Carlo calculations, and emulation obtains result.
The flow chart of Fig. 4 embodiment of the present invention.
Specific embodiment
The invention will be described in further detail with reference to the accompanying drawing.Flow chart of the present invention is as shown in Figure 4.
It is filtered initialization first, initializes the Markov Transition Probabilities matrix P of three kinds of modelsijAnd every kind of model
Probability μj(k), the state estimation of every kind of model is initializedAnd covariance estimates Pj(k/k)。
Next carries out input interaction, the model prediction probability μ respectively obtained with three kinds of modelsi(k+1/k) input is acquired to hand over
Mutual probability μj/i(k);The state estimation that initialization or previous moment filter is interacted into probability calculation with input and obtains every kind
The state of model mixesOn this basis, it is calculated using the state covariance that initialization or previous moment obtain
Obtain state covariance mixing
The input interaction of three kinds of models is obtained, respectively parallel filtering.Traditional Kalman's filter is respectively adopted in three kinds of state models
Wave obtains its error in measurement covariance Si(k+1), prediction error v is measuredi(k+1), state estimationState
Estimate of variance Pi(k+1/k+1).Wherein state estimation, state variance estimated value interacted the input for being used for subsequent time
Cheng Zhong.
By error in measurement covariance obtained in the previous step, measures and predict that error obtains likelihood function Λi(k+1), by seemingly
Right function arrives this moment model probability μi(k+1).Wherein state model probability will be used for the input interactive process of subsequent time
In.
Output interaction is finally carried out, after the state estimation, state variance estimation and the update that are obtained by each model parallel filtering
Model probability obtain state output estimationAnd state covariance estimation P (k+1/k+1).
Claims (5)
1. a kind of motor-driven and stationary state method for tracking target, specifically includes the following steps:
The first step, establish at the uniform velocity, speed change, static three kinds of state models;
Second step is filtered initialization, initialize three kinds of models Markov Transition Probabilities matrix and every kind of model it is general
Rate initializes the state estimation and covariance estimation of every kind of model;
Third step carries out input interaction, and the model prediction probability respectively obtained with three kinds of models acquires input interaction probability;It will be first
The state estimation that beginningization or previous moment filter interacts probability calculation with input and obtains the state mixing of every kind of model;?
On the basis of this, state covariance is calculated using the state covariance that initialization or previous moment obtain and mixes;
Input interaction of 4th step to three kinds of obtained models, legacy card is respectively adopted in parallel filtering, three kinds of state models respectively
Kalman Filtering obtains its error in measurement covariance, measures prediction error, state estimation, state variance estimated value;
5th step by error in measurement covariance obtained in the previous step, measures and predicts that error obtains likelihood function, by likelihood function
Obtain this moment model probability;
6th step carries out output interaction, the state estimation that is obtained by each model parallel filtering, state variance estimation and updated
Model probability obtains state output estimation and state covariance estimation.
2. a kind of motor-driven and stationary state method for tracking target as described in claim 1, which is characterized in that the at the uniform velocity state mould
Relevant parameter and matrix of variables involved in type are as follows:
Wherein Fm1 is state-transition matrix, G1For input matrix, Q1For state error covariance matrix, H1For measurement matrix, R is
Error in measurement covariance matrix, T are the scan period;
Relevant parameter involved in the speed change state model and matrix of variables are as follows:
Parameter meaning is identical as at the uniform velocity model in matrix;
Relevant parameter involved in the stationary state model and matrix of variables are as follows:
Parameter meaning is identical as at the uniform velocity model in matrix.
3. a kind of motor-driven and stationary state method for tracking target as claimed in claim 2, which is characterized in that in the step 2,
Switching Markov Transition Probabilities matrix between three models are as follows:
At the uniform velocity, speed change, the Initial state estimation of interdiction model are as follows:
At the uniform velocity, speed change, the original state covariance of interdiction model are as follows:
4. a kind of motor-driven and stationary state method for tracking target as claimed in claim 3, which is characterized in that in the step 3,
Input interaction is carried out, the model prediction probability μ respectively obtained with three kinds of modelsi(k+1/k) input interaction probability μ is acquiredj/i(k);
Initialization or the state estimation that filters of previous moment are mixed with inputting to interact probability calculation and obtain the state of every kind of model
It closes:
On this basis, state covariance is calculated using the state covariance that initialization or previous moment obtain to mix
5. a kind of motor-driven and stationary state method for tracking target as claimed in claim 4, which is characterized in that in the step 4,
Parallel filtering is carried out using traditional Kalman filtering, three kinds of model state estimations:
Three kinds of model state variances:
Pi(k+1/k+1)=[I-Ki(k+1)Hi(k+1)]Pi(k+1/k)
Model likelihood function:
Model probability:
Export interaction mode estimation and state covariance are as follows:
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910557541.7A CN110261859B (en) | 2019-06-25 | 2019-06-25 | Underwater maneuvering static alternating state target tracking method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910557541.7A CN110261859B (en) | 2019-06-25 | 2019-06-25 | Underwater maneuvering static alternating state target tracking method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN110261859A true CN110261859A (en) | 2019-09-20 |
CN110261859B CN110261859B (en) | 2023-10-31 |
Family
ID=67921478
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910557541.7A Active CN110261859B (en) | 2019-06-25 | 2019-06-25 | Underwater maneuvering static alternating state target tracking method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110261859B (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111208506A (en) * | 2020-01-08 | 2020-05-29 | 中国船舶重工集团公司第七二四研究所 | Simplified interactive multi-model tracking method |
CN111289967A (en) * | 2020-03-31 | 2020-06-16 | 四川长虹电器股份有限公司 | Personnel detection tracking and counting algorithm based on millimeter wave radar |
CN111625766A (en) * | 2020-04-27 | 2020-09-04 | 中国人民解放军63921部队 | Generalized continuation approximation filtering method, storage medium and processor |
Citations (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102568004A (en) * | 2011-12-22 | 2012-07-11 | 南昌航空大学 | Tracking algorithm for high maneuvering targets |
JP2012251904A (en) * | 2011-06-03 | 2012-12-20 | Mitsubishi Electric Corp | Tracking device |
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 |
CN104021285A (en) * | 2014-05-30 | 2014-09-03 | 哈尔滨工程大学 | Interactive multi-model target racking method with optimal motion pattern switching parameters |
CN104020480A (en) * | 2014-06-17 | 2014-09-03 | 北京理工大学 | Satellite navigation method for interactive multi-model UKF with self-adapting factors |
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 |
CN107292265A (en) * | 2017-06-20 | 2017-10-24 | 中国电子科技集团公司第二十八研究所 | A kind of target trajectory rapid extracting method based on motor-driven detection |
CN109508000A (en) * | 2018-12-16 | 2019-03-22 | 西南电子技术研究所(中国电子科技集团公司第十研究所) | Isomery multi-sensor multi-target tracking method |
CN109687844A (en) * | 2018-08-17 | 2019-04-26 | 西安理工大学 | A kind of intelligent maneuver method for tracking target |
-
2019
- 2019-06-25 CN CN201910557541.7A patent/CN110261859B/en active Active
Patent Citations (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2012251904A (en) * | 2011-06-03 | 2012-12-20 | Mitsubishi Electric Corp | Tracking device |
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 |
CN104021285A (en) * | 2014-05-30 | 2014-09-03 | 哈尔滨工程大学 | Interactive multi-model target racking method with optimal motion pattern switching parameters |
CN104020480A (en) * | 2014-06-17 | 2014-09-03 | 北京理工大学 | Satellite navigation method for interactive multi-model UKF with self-adapting factors |
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 |
CN107292265A (en) * | 2017-06-20 | 2017-10-24 | 中国电子科技集团公司第二十八研究所 | A kind of target trajectory rapid extracting method based on motor-driven detection |
CN109687844A (en) * | 2018-08-17 | 2019-04-26 | 西安理工大学 | A kind of intelligent maneuver method for tracking target |
CN109508000A (en) * | 2018-12-16 | 2019-03-22 | 西南电子技术研究所(中国电子科技集团公司第十研究所) | Isomery multi-sensor multi-target tracking method |
Non-Patent Citations (1)
Title |
---|
孙福明: ""机动目标跟踪状态估计与数据关联技术的研究"", 《中国博士学位论文全文数据库 信息科技辑》, no. 08, pages 136 - 7 * |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111208506A (en) * | 2020-01-08 | 2020-05-29 | 中国船舶重工集团公司第七二四研究所 | Simplified interactive multi-model tracking method |
CN111289967A (en) * | 2020-03-31 | 2020-06-16 | 四川长虹电器股份有限公司 | Personnel detection tracking and counting algorithm based on millimeter wave radar |
CN111625766A (en) * | 2020-04-27 | 2020-09-04 | 中国人民解放军63921部队 | Generalized continuation approximation filtering method, storage medium and processor |
Also Published As
Publication number | Publication date |
---|---|
CN110261859B (en) | 2023-10-31 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN103853908B (en) | A kind of maneuvering target tracking method of adaptive interaction formula multi-model | |
CN110261859A (en) | A kind of static alternating state method for tracking target of underwater manoeuvre | |
CN106054170B (en) | A kind of maneuvering target tracking method under constraints | |
Lermusiaux et al. | Science of autonomy: Time-optimal path planning and adaptive sampling for swarms of ocean vehicles | |
CN106933106A (en) | A kind of method for tracking target based on fuzzy control Multiple Models Algorithm | |
US7030809B2 (en) | Multiple model radar tracking filter and systems and methods employing same | |
CN104252178A (en) | Strong maneuver-based target tracking method | |
CN101661104A (en) | Target tracking method based on radar/infrared measurement data coordinate conversion | |
CN102568004A (en) | Tracking algorithm for high maneuvering targets | |
CN102622520A (en) | Distributed multi-model estimation fusion method of maneuvering target tracking | |
CN101819782A (en) | Variable-step self-adaptive blind source separation method and blind source separation system | |
CN109300144A (en) | A kind of pedestrian track prediction technique of mosaic society's power model and Kalman filtering | |
CN107462882A (en) | A kind of multiple maneuver target tracking methods and system suitable for flicker noise | |
Wang et al. | Multi-agent sensitivity enhanced iterative best response: A real-time game theoretic planner for drone racing in 3D environments | |
CN109212519B (en) | Narrow-band radar target tracking method based on BF-DLSTM | |
Jin et al. | Hierarchical and stable multiagent reinforcement learning for cooperative navigation control | |
Fang et al. | Autonomous underwater vehicle formation control and obstacle avoidance using multi-agent generative adversarial imitation learning | |
CN105424043B (en) | It is a kind of based on judging motor-driven estimation method of motion state | |
CN109444841A (en) | Smooth structure changes filtering method and system based on amendment switching function | |
CN104021285B (en) | A kind of interactive multi-model method for tracking target with optimal motion pattern switching parameter | |
CN102663771A (en) | Interactive multi-model estimation method based on covariance intersection | |
CN114219066A (en) | Unsupervised reinforcement learning method and unsupervised reinforcement learning device based on Watherstein distance | |
CN109766569A (en) | Submarine movement Model Simplification Method and device | |
CN107391446A (en) | Irregular shape based on random matrix extends target shape and method for estimating state more | |
Liu | RETRACTED: Research on decision-making strategy of soccer robot based on multi-agent reinforcement learning |
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 |