CN106707272A - Multi-target tracking method based on theory of random sets - Google Patents
Multi-target tracking method based on theory of random sets Download PDFInfo
- Publication number
- CN106707272A CN106707272A CN201610522044.XA CN201610522044A CN106707272A CN 106707272 A CN106707272 A CN 106707272A CN 201610522044 A CN201610522044 A CN 201610522044A CN 106707272 A CN106707272 A CN 106707272A
- Authority
- CN
- China
- Prior art keywords
- target
- probability
- subset
- label
- density function
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 14
- 238000005259 measurement Methods 0.000 claims abstract description 11
- 230000004083 survival effect Effects 0.000 claims description 12
- 238000010606 normalization Methods 0.000 claims description 3
- 238000000151 deposition Methods 0.000 claims 1
- 238000001514 detection method Methods 0.000 description 2
- 238000004587 chromatography analysis Methods 0.000 description 1
- 238000005094 computer simulation Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- UAJUXJSXCLUTNU-UHFFFAOYSA-N pranlukast Chemical compound C=1C=C(OCCCCC=2C=CC=CC=2)C=CC=1C(=O)NC(C=1)=CC=C(C(C=2)=O)C=1OC=2C=1N=NNN=1 UAJUXJSXCLUTNU-UHFFFAOYSA-N 0.000 description 1
- 229960004583 pranlukast Drugs 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
- 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
- G01S13/726—Multiple target tracking
Landscapes
- Engineering & Computer Science (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Computer Networks & Wireless Communication (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Radar Systems Or Details Thereof (AREA)
Abstract
The invention provides a multi-target tracking method based on the theory of random sets, and aims at realizing multi-target state effective tracking in a complex scene under the restricted condition of limited resources. According to the method, firstly each continue live track is further predicted according to the Bayes rule in the track prediction stage, and then a target track is self-adaptively generated according to target prior information; and in the track updating stage, firstly the corresponding existence weight probability and a combined multi-target probability density function are computed under each assumption, and then the posterior probability parameter of each target mark number is computed, including the existence probability and the corresponding probability density function. The method has the advantages of being low in approximation cost and excellent in performance under the limited resources and has the robustness suitable for any measurement models. Besides, the algorithm occupies less system computing resources so as to have great engineering application prospect.
Description
Technical field
The invention belongs to multiple target tracking field, its more particularly to theoretical lower Multitarget Tracking field of random set.
Background technology
In recent years, the multi-target detection tracking technique based on random set statistical theory has obtained extensive concern, this kind of side
Method avoids the data correlation of traditional multiple target tracking, and can process that target number is unknown and situation of time-varying.At present, mostly
Several random set track algorithms, such as probability hypothesis density wave filter, the probability hypothesis density wave filter of radix, label multiple target
Bernoulli Jacob's wave filter etc. both for standard measurement model (referring to document R.Mahler, Statistical
Multisource-Multitarget Information Fusion,Norwood,MA:Artech House, 2007.) and set
Meter.However, in actual multiple target tracking scene, the measurement model of standard has many limitation, many sensor models
Can not be described with standard measurement model, such as before detections of radar in trace model, wireless sensor network multi-user positioning, it is many
Multi-target position in input multi output radar, sonar amplitude sensor, radio frequency chromatography imaging tracing system, video frequency following system
Deng.
Asked for the multiple target tracking under the measurement model (the especially non-standard measurement model of nonlinearity) of broad sense
Topic, document (F.Papi, B.N.Vo, B.T.Vo, C.Fantacci and M.Beard, " Generalized labeled
multi-Bernoulli Approximation of multi-object densities,”IEEE Trans.on Signal
Process.Vol.63, No.20, pp.5487-5497,2015.) a kind of broad sense label multiple target Bernoulli Jacob's wave filter is proposed,
The wave filter realizes effective tracking of multiple target under off-gauge measurement model, and can effectively recognize target identities, shape
Into targetpath.But, the multiple target posteriority component number of the wave filter is presented super exponential increase with the growth of target number, makes
Greatly, poor real is limited by very large computing resource needed for it in actual applications.
The content of the invention
The technical problems to be solved by the invention are, the existing random set multiple target tracking side for non-standard measurement model
The system resources in computation that method takes increases with target number in super index, so as to be unfavorable for its practical engineering application.
The present invention solves the technical scheme that is used of above-mentioned technical problem, a kind of multiple target theoretical based on random set with
Track method, it is comprised the following steps:
Step 1, initialization system parameter
Initialization system parameter includes:Radar surveillance scope:That is datum plane, radar resolution ratio △ r, radar scanning
Cycle T, observation totalframes K;Target survival probabilitySingle goal probability density function fk|k-1(xk|xk-1);Birth mesh
Mark model is distributed for multiple target Bernoulli JacobWhereinRepresent theThe presence probability of individual birth targetpath,Represent theThe individual corresponding distribution probability density function of targetpath of being born,Represent birth target label ensemble space;
Multiple target likelihood function g (Z | X), wherein Z represents measurement set, and X represents multiple target state set;
Step 2:Initialization label multiple target-Bernoulli parameterAnd makeWhereinRepresent current
The target label ensemble space of frame;Initialization iteration time k=1,Represent theThe individual presence for continuing survival targetpath is general
Rate,Represent theThe individual probability density function for continuing survival target;
Step 3, the prediction flight path for calculating kth frame:
3.1st, according to the prior information of target birth position, and based on target birth model birth fresh target flight path, birth
Object module is distributed for multiple target Bernoulli Jacob
3.2 predictions for continuing survival flight path, according to Bayesian forecasting equation, theBar continues Trajectory Prediction of surviving
Presence probabilityWith corresponding probability density functionIt is as follows respectively,
Wherein, mathematic sign<f1(x),f2(x)>Representative function f1(x) and function f2Inner product between (x);Represent and continue
Survival targetpath label ensemble space;
Step 3, the finite subset space for building the target label set predicted;
3.1 determine target label setTarget label collection is combined into target birth label set and continues mark of surviving with target
Number union of sets collection;
3.2 build target label setAll finite subsets space, its mathematical symbolism isFinite subset number beHere I is defined+It is finite subset spaceAny one have
Subset is limited, and its mathematical symbolism is
Step 4, according to bayesian criterion, calculate each finite subset of kth framePosteriority weight probability and its
Corresponding multiple target probability density function;
4.1. each subset is calculatedPriori weight probability w+(I+):
Wherein mathematic sign ∏ represents that company multiplies symbol,Represent theThe presence probability of individual prediction targetpath,Table
Show indicator function, it is defined as follows:Indicator function
4.2. each subset is calculatedMultiple target posterior probability density function normalization factor ηZ(I+):
WhereinRepresent theThe probability density function of individual prediction targetpath;
4.3. each subset is calculatedPosteriority weight probability w (I+):
w(I+)=w+(I+)ηZ(I+)
4.4. each subset is calculatedMultiple target posterior probability density function:
Step 5, to each subsetPosteriority weight probability w (I+) be normalized:
Step 6, each target label of calculatingPresence probabilityAnd its corresponding probability density function
6.1st, each target label is calculatedPresence probabilityBy each subset of step 5In forgive target labelPosteriority weight probability w (I+) sued for peace, obtain target labelPresence probability
6.2nd, each target label is calculatedPosterior probability density functionBy each subset of step 5In forgive target labelMarginal probability density functionSummation is weighted, its weighting weight is
The posteriority weight probability w (I of corresponding subset+), and the presence probability obtained with step 6.1The posteriority obtained to weighted sum is general
Rate density functionIt is normalized:
Wherein
WhereinRepresent label subset I+IncludingMultiple target posterior probability density function P
(X|Z);Step 7, dbjective state are extracted;Radix corresponding to cardinal of the set distribution maximum is the estimation of k moment targets number
Nk, and the N is estimated respectivelykThe state of individual target;
If step 8, k<K, makes k be equal to k+1, and return to step 3, wherein k are frame number.
Innovative point of the invention is, 1) in forecast period, independent prediction to be carried out between multiple Bernoulli Jacob's components, greatly section
Computing resource needed for having saved multiple target tracking;2) in the more new stage, by building the finite subset space of target label, joint
Multiple target probability density function under multiple hypotheses is considered, therefore the tracking is adapted to more broadly multiple target and measures mould
Type.
An advantage of the invention is that it provides one kind take computing resource it is few, and the excellent effect of tracking performance multiple target tracking side
Method, and target tracking algorism proposed by the present invention can be applicable any measurement model, with compared with strong algorithms robustness.
Brief description of the drawings
Fig. 1 is flow chart of the invention;
Fig. 2 is the schematic diagram of inventive algorithm;
Fig. 3 is simulating scenes figure under inventive algorithm typical scene;
Fig. 4 is the tracking effect figure of inventive algorithm.
Specific embodiment
The content of present invention description is described for convenience, and following term definition is done first:
Random collection:It refer to given state spaceByAll finite subsets constitute a superspaceThen
It is defined onOn stochastic variable be referred to as random collection.
Cardinal of the set:It refer to the number of element in set.
The main method using Computer Simulation of the invention is verified that all steps, conclusion are all in MATLAB-R2014b
Upper checking is correct.Specific implementation step is as follows:
Step 1, initialization system parameter.
Initialization system parameter includes:Radar surveillance scope [0,60m] × [0,60m], radar resolution ratio △ r=1m, thunder
Up to scan period T=1s, observation totalframes K=28;Target survival probabilityBirth object module is multiple target Bernoulli Jacob
DistributionWherein The measuring value vector of present frameRepresent, wherein mathematic signRepresent real number
Domain, zjJ-th measuring value of radar resolution cell of (j=1 ..., M) expression, the number M=3600 of resolution cell, multiple target is seemingly
The mathematical expression of right function g (Z | X) be for:
Wherein mathematic signRepresent the general of under the conditions of multiple target state X j-th measuring value of resolution cell
Rate density function, here using Gaussian Profile, mathematic(al) representation is:
WhereinExpression average is μ, and covariance matrix is the Gaussian Profile of Γ,Represent dbjective state x
To j-th power contribution amount of resolution cell, σNX () represents noise power.
Step 2:Initialization label multiple target-Bernoulli parameterAnd makeWhereinRepresent current
The target label ensemble space of frame;Initialization iteration time k=1.
Step 3, the prediction flight path for calculating kth frame:
3.1st, according to the prior information of target birth position, and based on target birth model birth fresh target flight path, birth
Object module is distributed for multiple target Bernoulli Jacob
3.2 predictions for continuing survival flight path.According to Bayesian forecasting equation, theBar continues Trajectory Prediction of surviving
Presence probabilityWith corresponding probability density functionIt is as follows respectively,
Wherein, mathematic sign<f1(x),f2(x)>Representative function f1(x) and function f2X the inner product between (), its mathematics is determined
Justice isRepresent and continue targetpath label ensemble space of surviving.
Step 3, the finite subset space for building the target label set predicted.
3.1 determine target label setTarget label collection is combined into target birth label set and continues mark of surviving with target
Number union of sets collection, i.e.,Wherein.Element number is then in target label setWherein symbol | f | represents f
Length.
3.2 build target label setAll finite subsets space, its mathematical symbolism isFinite subset number beHere I is defined+It is finite subset spaceAny one have
Subset is limited, and its mathematical symbolism is
Step 4, according to bayesian criterion, calculate each finite subset of kth framePosterior probability and its correspondence
Multiple target probability density function.
4.1. each subset is calculatedPriori weight probability w+(I+), its mathematical computations expression formula is
Wherein mathematic sign ∏ is represented and even multiply symbol.
4.2. each subset is calculatedMultiple target posterior probability density function normalization factor ηZ(I+),
Its mathematical computations expression formula is
4.3. each subset is calculatedPosteriority weight probability w (I+), its mathematical computations expression formula is
w(I+)=w+(I+)ηZ(I+)
4.4. each subset is calculatedMultiple target posterior probability density function, its mathematical computations expression formula is
Step 5, to each subsetPosteriority weight probability w (I+) be normalized, i.e.,
Step 6, each target label of calculatingPresence probabilityAnd its corresponding probability density function
6.1st, each target label is calculatedPresence probabilityBy each subset of step 5In forgive target labelPosteriority weight probability w (I+) sued for peace, obtain target labelPresence probabilityI.e.
6.2nd, each target label is calculatedPosterior probability density functionBy each subset of step 5In forgive target labelMarginal probability density functionSummation is weighted, its weighting weight is
The posteriority weight probability w (I of corresponding subset+), and the presence probability obtained with step 6.1The posteriority obtained to weighted sum is general
Rate density functionIt is normalized:
Wherein
WhereinRepresent that label set includesMultiple target posterior probability density function P (X
|Z);Step 7, dbjective state are extracted.Radix corresponding to cardinal of the set distribution maximum is k moment targets number and estimates Nk,
And the N is estimated respectivelykThe state of individual target.
If step 8, k<K, makes k be equal to k+1, and return to step 3, wherein k are frame number.
By above step, it is possible to multiple target tracking under effectively realizing complex scene under system limited resources.
Claims (1)
1. a kind of multi-object tracking method theoretical based on random set, it is comprised the following steps:
Step 1, initialization system parameter
Initialization system parameter includes:Radar surveillance scope:That is datum plane, radar resolution ratio △ r, radar scanning cycle
T, observation totalframes K;Target survival probability pS(l), single goal probability density function fk|k-1(xk|xk-1);Birth target mould
Type is distributed for multiple target Bernoulli JacobWhereinL-th presence probability of birth targetpath is represented,L-th corresponding distribution probability density function of targetpath of being born is represented,Represent birth target label ensemble space;
Multiple target likelihood function g (Z | X), wherein Z represents measurement set, and X represents multiple target state set;
Step 2:Initialization label multiple target-Bernoulli parameterAnd makeWhereinRepresent present frame
Target label ensemble space;Initialization iteration time k=1, r(l)Represent l-th presence probability of continuation survival targetpath, p(l)Represent l-th probability density function of continuation survival target;
Step 3, the prediction flight path for calculating kth frame:
3.1st, according to the prior information of target birth position, and based on target birth model birth fresh target flight path, target of being born
Model is distributed for multiple target Bernoulli Jacob
3.2 predictions for continuing survival flight path, according to Bayesian forecasting equation, theBar continues depositing for survival Trajectory Prediction
In probabilityWith corresponding probability density functionIt is as follows respectively,
Wherein, mathematic sign < f1(x),f2(x)>Representative function f1(x) and function f2Inner product between (x);Expression continues to deposit
Targetpath label ensemble space living;
Step 3, the finite subset space for building the target label set predicted;
3.1 determine target label setTarget label collection is combined into target birth label set and continues label set of surviving with target
The union of conjunction;
3.2 build target label setAll finite subsets space, its mathematical symbolism is 's
Finite subset number isHere I is defined+It is finite subset spaceAny one finite subset, and its mathematics accord with
Number it is expressed as
Step 4, according to bayesian criterion, calculate each finite subset of kth framePosteriority weight probability and its correspondence
Multiple target probability density function;
4.1. each subset is calculatedPriori weight probability w+(I+):
Wherein mathematic sign ∏ represents that company multiplies symbol,L-th presence probability of prediction targetpath is represented,Expression refers to
Show function, it is defined as follows:Indicator function
4.2. each subset is calculatedMultiple target posterior probability density function normalization factor ηZ(I+):
WhereinRepresent theThe probability density function of individual prediction targetpath;
4.3. each subset is calculatedPosteriority weight probability w (I+):
w(I+)=w+(I+)ηZ(I+)
4.4. each subset is calculatedMultiple target posterior probability density function:
Step 5, to each subsetPosteriority weight probability w (I+) be normalized:
Step 6, each target label of calculatingPresence probability r(l)And its corresponding probability density function p(l)(x);
6.1st, the presence probability r of each target label l is calculated(l);By each subset of step 5In forgive target label l's
Posteriority weight probability w (I+) sued for peace, obtain the presence probability r of target label l(l):
6.2nd, each target label is calculatedPosterior probability density function p(l)(x);By each subset of step 5
In forgive the marginal probability density function of target label lSummation is weighted, its weighting weight is corresponding subset
Posteriority weight probability w (I+), and the presence probability r obtained with step 6.1(l)The posterior probability density letter obtained to weighted sum
Number p(l)X () is normalized:
WhereinWherein
P({(x,l)}∪{(x1,l1),…,(xn,ln) | Z) represent label subset I+Multiple target posterior probability density function P including l
(X|Z);
Step 7, dbjective state are extracted;Radix corresponding to cardinal of the set distribution maximum is k moment targets number and estimates Nk,
And the N is estimated respectivelykThe state of individual target;
If step 8, k < K, k is made to be equal to k+1, return to step 3, wherein k are frame number.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610522044.XA CN106707272B (en) | 2016-07-01 | 2016-07-01 | A kind of multi-object tracking method based on random set theory |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610522044.XA CN106707272B (en) | 2016-07-01 | 2016-07-01 | A kind of multi-object tracking method based on random set theory |
Publications (2)
Publication Number | Publication Date |
---|---|
CN106707272A true CN106707272A (en) | 2017-05-24 |
CN106707272B CN106707272B (en) | 2019-11-05 |
Family
ID=58939708
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201610522044.XA Expired - Fee Related CN106707272B (en) | 2016-07-01 | 2016-07-01 | A kind of multi-object tracking method based on random set theory |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106707272B (en) |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109508444A (en) * | 2018-12-18 | 2019-03-22 | 桂林电子科技大学 | Section measures the fast tracking method of the more Bernoulli Jacob of lower interactive multimode broad sense label |
CN111414843A (en) * | 2020-03-17 | 2020-07-14 | 森思泰克河北科技有限公司 | Gesture recognition method and terminal device |
CN111457916A (en) * | 2020-03-30 | 2020-07-28 | 中国人民解放军国防科技大学 | Space debris target tracking method and device based on expansion mark random finite set |
CN112215146A (en) * | 2020-10-12 | 2021-01-12 | 西安交通大学 | Weak and small target joint detection and tracking system and method based on random finite set |
CN114491413A (en) * | 2022-01-25 | 2022-05-13 | 中国人民解放军海军航空大学航空作战勤务学院 | Probability density hypothesis track generation method and system based on minimum cross entropy |
Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104199006A (en) * | 2014-07-16 | 2014-12-10 | 电子科技大学 | Random set tracking method based on multi-hypothesis combined distributed filter |
-
2016
- 2016-07-01 CN CN201610522044.XA patent/CN106707272B/en not_active Expired - Fee Related
Patent Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104199006A (en) * | 2014-07-16 | 2014-12-10 | 电子科技大学 | Random set tracking method based on multi-hypothesis combined distributed filter |
Non-Patent Citations (5)
Title |
---|
BAILU WANG ET AL.: ""Distributed Multi-Target Tracking Via Generalized Multi-Bernoulli Random Finite Sets"", 《18TH INTERNATIONAL CONFERENCE ON INFORMATION FUSION》 * |
BA-TUONG VO ET AL.: ""Labeled Random Finite Sets and Multi-Object Conjugate Priors"", 《IEEE TRANSACTIONS ON SIGNAL PROCESSING》 * |
冯新喜 等: ""基于随机有限集理论的多扩展目标跟踪技术综述"", 《空军工程大学学报(自然科学版)》 * |
吴卫华 等: ""基于随机有限集的多目标跟踪算法综述"", 《电光与控制》 * |
张英杰 等: ""基于PHD滤波的多传感器多目标跟踪融合算法"", 《测控技术》 * |
Cited By (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109508444A (en) * | 2018-12-18 | 2019-03-22 | 桂林电子科技大学 | Section measures the fast tracking method of the more Bernoulli Jacob of lower interactive multimode broad sense label |
CN109508444B (en) * | 2018-12-18 | 2022-11-04 | 桂林电子科技大学 | Quick tracking method for interactive multimode generalized label multi-Bernoulli under interval measurement |
CN111414843A (en) * | 2020-03-17 | 2020-07-14 | 森思泰克河北科技有限公司 | Gesture recognition method and terminal device |
CN111414843B (en) * | 2020-03-17 | 2022-12-06 | 森思泰克河北科技有限公司 | Gesture recognition method and terminal device |
CN111457916A (en) * | 2020-03-30 | 2020-07-28 | 中国人民解放军国防科技大学 | Space debris target tracking method and device based on expansion mark random finite set |
CN111457916B (en) * | 2020-03-30 | 2021-05-07 | 中国人民解放军国防科技大学 | Space debris target tracking method and device based on expansion mark random finite set |
CN112215146A (en) * | 2020-10-12 | 2021-01-12 | 西安交通大学 | Weak and small target joint detection and tracking system and method based on random finite set |
CN114491413A (en) * | 2022-01-25 | 2022-05-13 | 中国人民解放军海军航空大学航空作战勤务学院 | Probability density hypothesis track generation method and system based on minimum cross entropy |
Also Published As
Publication number | Publication date |
---|---|
CN106707272B (en) | 2019-11-05 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Yang et al. | Semisupervised multilabel deep learning based nonintrusive load monitoring in smart grids | |
CN106707272A (en) | Multi-target tracking method based on theory of random sets | |
Vo et al. | A random finite set conjugate prior and application to multi-target tracking | |
CN109752710B (en) | Rapid target angle estimation method based on sparse Bayesian learning | |
Li et al. | Multi-EAP: Extended EAP for multi-estimate extraction for SMC-PHD filter | |
Liang-Qun et al. | Intuitionistic fuzzy joint probabilistic data association filter and its application to multitarget tracking | |
Yang et al. | Linear-time joint probabilistic data association for multiple extended object tracking | |
WO2018098926A1 (en) | Multi-target tracking method and system applicable to flicker noise | |
Addabbo et al. | HRR profile estimation using SLIM | |
Mahler | A brief survey of advances in random-set fusion | |
CN105354860A (en) | Box particle filtering based extension target CBMeMBer tracking method | |
Chen et al. | Extended target probability hypothesis density filter based on cubature Kalman filter | |
Tang et al. | Seamless tracking of apparent point and extended targets using Gaussian process PMHT | |
CN114636975A (en) | LPI radar signal identification method based on spectrogram fusion and attention mechanism | |
Wu et al. | Cost-sensitive latent space learning for imbalanced PolSAR image classification | |
Chang et al. | Applied state estimation and association | |
Panwar et al. | Maximum likelihood algorithm for time-delay based multistatic target localization | |
Pegoraro et al. | Deep learning for accurate indoor human tracking with a mm-wave radar | |
Qu et al. | Improving the reliability for confidence estimation | |
Li et al. | Robust generalized labeled multi-Bernoulli filter and smoother for multiple target tracking using variational Bayesian | |
CN105353353B (en) | Multiple search particle probabilities assume the multi-object tracking method of density filtering | |
Zhu et al. | Multi-resolution spatio-temporal prediction with application to wind power generation | |
Wu et al. | An MCMC based EM algorithm for mixtures of Gaussian processes | |
Turkmen et al. | Genetic tracker with neural network for single and multiple target tracking | |
Wang et al. | Sliding residual network for high-speed target detection in additive white Gaussian noise environments |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant | ||
CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20191105 |