CN106934324A - Based on the radar data correlating methods for simplifying many hypothesis algorithms - Google Patents
Based on the radar data correlating methods for simplifying many hypothesis algorithms Download PDFInfo
- Publication number
- CN106934324A CN106934324A CN201511018622.8A CN201511018622A CN106934324A CN 106934324 A CN106934324 A CN 106934324A CN 201511018622 A CN201511018622 A CN 201511018622A CN 106934324 A CN106934324 A CN 106934324A
- Authority
- CN
- China
- Prior art keywords
- association
- algorithms
- flight path
- optimal
- hypothesis
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/12—Classification; Matching
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2415—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on parametric or probabilistic models, e.g. based on likelihood ratio or false acceptance rate versus a false rejection rate
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Data Mining & Analysis (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Probability & Statistics with Applications (AREA)
- Signal Processing (AREA)
- Life Sciences & Earth Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Radar Systems Or Details Thereof (AREA)
Abstract
The present invention proposes a kind of based on the radar data correlating methods for simplifying many hypothesis algorithms.The problem that " multiple shot array " and amount of calculation exponential type caused for many hypothesis algorithmic delay decision-making mechanisms of tradition in multiple-target system rise, introduce likelihood ratio scoring function and linear distribution (LAP) pruning method, compared using track association log-likelihood and assume that association probability calculating is simplified, " optimal " hypothesis of M at current time are obtained by the level beta pruning of LAP algorithms flight path, beta pruning is assumed by global level again, quick for obtaining is optimal to assume matching sequence as efficient association measuring point, the final purpose for realizing reducing amount of calculation.Specifically include:Flight path score based on likelihood ratio function is calculated, the flight path level M- based on LAP is optimal assumes beta pruning and three major parts of optimal hypothesis generation of the global level based on LAP.Using method proposed by the present invention, can largely reduce the operands for assuming algorithm data association process on the basis of association accuracy is ensured more.
Description
Technical field
It is more particularly to a kind of based on letter the invention belongs to Radar Multi Target tracking data correlating method technical field
Change the radar data correlating methods for assuming algorithm more.
Background technology
Under radar echo,high level environment, interfered between multiple target, easily occur measuring and closed with real goal mistake
The phenomenon such as connection or targetpath interruption, target tracking accuracy and efficiency are severely impacted.Data association algorithm
Effect be exactly to be associated metric data with real goal flight path to match, so as to realize the accurate of multiple target
Tracking and the correct renewal of targetpath are maintained.In Radar Multi Target tracking, data association algorithm is mainly solved
Certainly two major class problem:First, when a measurement for target simultaneously fall into two target association Bo Mennei when how
Correct target trajectory is selected to be associated;Second, when simultaneously multiple amounts occurs in a target association Bo Mennei
Correct measurement how is selected during survey carries out flight path renewal maintenance.Under echo,high level environment, tracked target
Kinetic characteristic is complex.Around under the frequent interference of noise and clutter, make using only single frames echo information
The problems such as association according to real goal can be caused to be associated with the false erroneous association for measuring or loss.
Relative to single frames correlating method (such as arest neighbors, probabilistic data association, JPDA),
Multiframe correlating method with multiple hypotheis tracking (MHT) algorithm as representative is accumulated using follow-up many frame informations
The tired uncertainty to reduce association.Its biggest advantage is that complicated difficult related question can be postponed
Decision guidance obtains more information, and the past interrelated decision that has an opportunity to change is to lift interrelating effect,
The division and merging of target can also be processed.Multiple hypotheis tracking algorithm maintains and flight path track initiation, flight path
Termination is unified on a framework and processes, and is a kind of most powerful optimal multiple target tracking of function generally acknowledged at present
Method, is widely adopted in current many advanced Multi-Radar Tracking systems.
MHT relevant treatments mainly include point mark and faciation close, point mark and flight path is related, flight path hypothesis branch,
The links such as flight path score is calculated, assumes generation, assuming beta pruning, the merging of cluster and separate, Target track displaying,
Handling process is as shown in Figure 3.MHT algorithms improve the association degree of accuracy, Chang Beiyong using Delayed Decision mechanism
Multiple-target system under complicated ghost environments.But in tracking destination number and the increased feelings of clutter quantity
Under condition, MHT algorithms inevitably face hypothesis branch's number and computation complexity exponential type to be increased this and two asks greatly
Topic.Traditional MHT algorithms calculate newest amount on the basis of known measurement distribution probability using Bayesian formula
The association probability with reliable flight path is surveyed, and the hypothesis branch association probability after association is screened to be assumed to multiframe,
Finally give optimal global hypothesis.The tracking effect of the method relies on priori, to the initial of relevant parameter
Changing has high requirement, and computation complexity is high.Although using the common optimal hypothesis same energy of prune approach of m-
Access it is optimal global it is assumed that but need to be associated probability calculation to all hypothesis branches during its beta pruning,
Therefore operand is big, and real-time is restricted.
The content of the invention
It is a kind of based on the radar data correlating methods for simplifying many hypothesis algorithms it is an object of the invention to propose.This
Invention uses linear distribution (LAP) to assume the cost function of association as flight path by the use of likelihood ratio function
Algorithm is assumed multiframe to carry out many hard cuttings from flight path level and global level, simplifies the data correlation of multiple target tracking
Process.
To realize above-mentioned technical purpose, the present invention is adopted the following technical scheme that and is achieved.
A kind of radar data correlating methods based on simplified many hypothesis algorithms, comprise the following steps:
1) the oval association ripple door centered on Kalman prediction value is formed, newest measurement is closed
Connection judges.If measure being located at association Bo Mennei, step 2 is carried out);Otherwise, subsequent time is received to measure,
And to its repeat step 1);
2) relevance assumption is generated to each measurement for entering confirmation target association Bo Mennei, using likelihood ratio
Function, calculates the score matrix of relevance assumption flight path;
3) the relevance assumption flight path score formed according to current time and each target, is obtained using LAP algorithms
The flight path level optimal hypothesis of M-;
4) the processed measurement frame number of N-Scan methods is judged, if now reception processing N frame amount is surveyed, profit
Associating all hypothesis retained after judging to N frames with LAP algorithms carries out global level beta pruning, carries out step 5);
If current time does not receive nth frame measurement yet, receive next frame and measure, repeat step 1) -4);
5) measured using the efficient association of each target, carry out Kalman filtering renewal.Screen out credible target
Uncorrelated measurement information in relevance assumption cluster, the processed frame numbers of N-Scan subtract 1, and update reliable flight path shelves
Case information.Return to step 1), next frame is measured and is associated judgement treatment.
The present invention puts forth effort on hypothesis generation, probability for data correlation efficiency in Radar Multi Target tracking
Three links of beta pruning are calculated and assumed, assumes that algorithm simplifies more to.The present invention is equally applicable for complicated returning
Under ripple environment, the accurate tracking of multiple targets is realized, with computation complexity is low, speed is fast, tracking error
Small advantage.
It should be appreciated that all combinations of aforementioned concepts and the extra design for describing in greater detail below are only
One of the subject matter of the disclosure is can be viewed as in the case where such design is not conflicting
Point.In addition, all combinations of theme required for protection are considered as a part for the subject matter of the disclosure.
Can be more fully appreciated with from the following description with reference to accompanying drawing present invention teach that foregoing and other side
Face, embodiment and feature.The feature of other additional aspects such as illustrative embodiments of the invention and/or have
Beneficial effect will be obvious in the following description, or by according to present invention teach that specific embodiment practice
In learn.
Brief description of the drawings
Accompanying drawing is not intended to drawn to scale.In the accompanying drawings, each the identical or approximate phase for showing in each figure
Same part can be indicated by the same numeral.For clarity, in each figure, not each
Part is labeled.Now, example will be passed through and various aspects of the invention is described in reference to the drawings
Embodiment, wherein:
Fig. 1 is based on the flow chart for simplifying the radar data correlating methods for assuming algorithm more.
Fig. 2 is flight path level and global level flight path beta pruning design sketch (3 " optimal ", 3Scan) based on LAP.
Fig. 3 is cross track target following design sketch.
Fig. 4 is circular trace tracking error figure.
Fig. 5 is that circular trace target data association process assumes number change figure.
Specific embodiment
In order to know more about technology contents of the invention, especially exemplified by specific embodiment and institute's accompanying drawings are coordinated to be described as follows:
Each side with reference to the accompanying drawings to describe the present invention in the disclosure, shown in the drawings of the reality of many explanations
Apply example.Embodiment of the disclosure must not be intended to include all aspects of the invention.It should be appreciated that being situated between above
The various designs for continuing and embodiment, and those designs for describing in more detail below and implementation method can be with
Any one is implemented in many ways, because design disclosed in this invention and embodiment are not limited
In any implementation method.In addition, it is disclosed by the invention some aspect can be used alone, or with the present invention
It is disclosed otherwise any appropriately combined to use.
Radar data correlating methods based on simplified many hypothesis algorithms proposed by the invention, it is main to include being based on
The flight path score of likelihood ratio function is calculated, the flight path level M- based on LAP is optimal assumes beta pruning and based on LAP
Global level it is optimal assume generation three major parts.
Below in conjunction with the accompanying drawings, some one exemplary embodiments of the invention are illustrated.
Embodiments in accordance with the present invention, propose a kind of radar data correlating methods based on simplified many hypothesis algorithms,
Easily there is " multiple shot array " and association process amount of calculation with existing many hypothesis algorithms under overcoming complicated ghost environments
Excessive problem.
With reference to shown in Fig. 1 flow charts, the realization of the method generally comprises following 5 steps:
1) the oval association ripple door centered on Kalman prediction value is formed using its velocity interval, to most
New measurement is associated judgement.If measure being located at association Bo Mennei, step 2 is carried out);Otherwise, under reception
One moment measured, and to its repeat step 1);
2) relevance assumption is generated to each measurement for entering confirmation target association Bo Mennei, using likelihood ratio letter
Number, calculates the score matrix of relevance assumption flight path;
3) the relevance assumption flight path score formed according to current time and each target, is obtained using LAP algorithms
The flight path level optimal hypothesis of M-;
4) the processed measurement frame number of N-Scan methods is judged, if now reception processing N frame amount is surveyed, profit
Associating all hypothesis retained after judging to N frames with LAP algorithms carries out global level beta pruning, carries out step 5);
If current time does not receive nth frame measurement yet, receive next frame and measure, repeat step 1) -4);
5) measured using the efficient association of each target, carry out Kalman filtering renewal.Screen out credible target pass
Connection assumes uncorrelated measurement information in cluster, and the processed frame numbers of N-Scan subtract 1, and update reliable flight path archives
Information.Return to step 1), next frame is measured and is associated judgement treatment.
In the above method, the step 1) it is specially:
11) k moment radar sampling data are received, using " present statistical model is to target k moment states
Estimated;
12) centered on k moment state estimations, oval association ripple door is formed according to formula (1);
d2=v (k) S-1(k)vT(k)≤γ (1)
Wherein d2To obey the m frees degreeThe new breath covariance of distribution, parameter γ can be by acquisition of tabling look-up.
13) judge whether to measure in ellipse association Bo Mennei, if being located at association Bo Mennei, carry out step 2);
Otherwise, subsequent time is received to measure, and to its repeat step 1).
In the above method, the step 2) it is specially:
21) the LLR ratio LLR that every flight path assumes branch is calculated according to formula (2-3)k, as every
The score of individual hypothesis:
LLRk=LLRK-1+ln(ΔLRk) (2)
Δ LRk is the likelihood ratio increment at k moment, and above formula is to measure the hypothesis flight path score being associated with flight path, under
Formula is not associated flight path score.PDIt is detection probability, PFAIt is false-alarm probability, PFAIt is the probability density of clutter
Function (generally assume that for be uniformly distributed or Poisson distribution), vkIt is new breath, SkIt is new breath covariance, N is represented
Gaussian Profile.
22) score of all hypothesis branches is collected, is converted into track association score matrix, be represented by:It is public
Formula (4).
As shown in Fig. 2 in the above method, the step 3) it is specially:
31) by step 2) in obtain hypothesis flight path likelihood ratio track association score matrix substitute LAP methods in
The association probability for being used, as the property value assumed;
32) utilize step 31) in track association score matrix P_value, to it take the conversion of negative,
And then obtain LAP decision matrixs;
33) weight matrix is determined according to priori weight vectors;
34) M- optimal sequencing matrixes P is obtained according to LAP linear programming methods;
35) according to M- optimal sequencing matrix P, the optimal hypothesis of M- of all targets of present frame are obtained.
In the above method, the step 32) it is specially:
321) to track association score matrix P_value take the conversion of negative:
322) track association cost matrix is obtained:
323) a decision matrix D is definedm×n(m is represented and is measured number, and the flight path that n represents previous moment is assumed
Numbers of branches), the element in decision matrix D represents the k moment and measures and assume branch with the flight path at k-1 moment
Associate the property value of matching.Order:Dm×n=P_cost.
In the above method, the step 33) specific sub-step include:
331) according to the actual scene of multiple target tracking, a weight vectors ω=(ω is defined1,ω2), wherein element
Represent the association probability measured with target:ω=(p_Track1, p_Track2);
332) current time reliability flight path is defined to assume to be scored at Track1_value and Track2_value, according to
Below equation calculates now weight vectors;
333) according to weight vectors, final weight matrix π is determined according to below equation.
Π=D × ωT (8)
In the above method, the step 34) specific sub-step include:
341) whole weight matrix is traveled through, an initiation sequence a is found0Make global costMinimum, for example
Minimum cost sequence is X1With Track1Association, X3With Track2Association:
342) with initiation sequence a0Based on, one of them or several matching sequence numbers are replaced successively, search power square
Battle array, is replaced with secondary dominating sequence, forms a new sequence;
343) the hypothesis cost representated by the sequence for obtaining increases successively, is assumed to be until forming M " optimal "
Only, M values are 3 in the present embodiment.
In the above method, the step 35) specific sub-step include:
351) according to step 34) M of all targets that obtains is " optimal " assumes (M=3), retain " most
It is excellent " assume in qualified measurement, and confirm to delete the relatively low unqualified measurement of score in target cluster from changing;
352) delete step 351 in temporary transient flight path archives) the middle unqualified relevance assumption rejected;
353) score of each credible target flight path is recalculated.
As shown in Fig. 2 in the above method, the step 4) specific sub-step include:
41) judge the processed measurement frame number of N-Scan methods, used in the present embodiment at the N-Scan of N=3
Reason method;
If 42) now the frame amount of reception processing 3 is surveyed, retain after being judged the association of N frames using LAP algorithms
All hypothesis carry out global level beta pruning, for all confirmation targets, obtain it is optimal association matching, walked
It is rapid 5);
If 43) current time do not receive yet the 3rd frame amount survey, receive next frame measure, repeat step 1)
-4)。
In the above method, the step 42) specific sub-step include:
421) flight path obtained after being associated to multiframe is it is assumed that according to its track association score, repeat step 31)
- 33) global level LAP weight matrixs, are obtained;
422) weight matrix is traveled through, the optimal sequence for making global Least-cost is found;
423) according to optimal sequence, determine that the efficient association of credible target is measured.
Assume with reference to the circular trace tracking error figure of Fig. 4 and the circular trace target data association process of Fig. 5
Number change figure, it is proposed by the present invention based on the radar data correlating methods for assuming algorithm are simplified more, for many
" multiple shot array " and amount of calculation that many hypothesis algorithmic delay decision-making mechanisms of tradition are caused in Target Tracking System refer to
The problem that number type rises, invention introduces likelihood ratio scoring function and LAP pruning methods, is closed using flight path
Connection log-likelihood is compared assumes that association probability calculating is simplified, and is assumed by the level beta pruning pair of LAP algorithms flight path
The M " optimal " at current time filter out it is assumed that again to all measurement shapes of N-Scan (N=3) sliding window
Into hypothesis branch carry out LAP global levels and assume beta pruning, quickly obtain one optimal to assume that matching sequence is made
It is efficient association measuring point, the final purpose for realizing reducing amount of calculation.Specifically include:Based on likelihood ratio function
Flight path score calculate, flight path based on LAP level M- is optimal assumes beta pruning and the global level based on LAP most
It is excellent to assume three major parts of generation.Using method proposed by the present invention, association accuracy can ensured
On the basis of, largely reduce the operands for assuming algorithm data association process more.
Although the present invention is disclosed above with preferred embodiment, so it is not limited to the present invention.The present invention
Those of ordinary skill in the art, it is without departing from the spirit and scope of the present invention, each when that can make
The change planted and retouching.Therefore, protection scope of the present invention is worked as and is defined depending on those as defined in claim.
Claims (9)
1. it is a kind of based on the radar data correlating methods for simplifying many hypothesis algorithms, it is characterised in that including following
Step:
1) the oval association ripple door centered on Kalman prediction value is formed, newest measurement is closed
Connection judges:If measure being located at association Bo Mennei, step 2 is carried out);Otherwise, the measurement of subsequent time is received,
And repeat step 1) deterministic process;
2) relevance assumption is generated to each measurement for entering confirmation target association Bo Mennei, using likelihood ratio
Function, calculates the score matrix of relevance assumption flight path;
3) the relevance assumption flight path score formed according to current time and each target, is obtained using LAP algorithms
The flight path level optimal hypothesis of M-;
4) the processed measurement frame number of N-Scan methods is judged, if now reception processing N frame amount is surveyed, profit
Associating all hypothesis retained after judging to N frames with LAP algorithms carries out global level beta pruning, carries out step 5);
If current time does not receive nth frame measurement yet, receive next frame and measure, repeat step 1) -4);
5) measured using the efficient association of each target, carry out Kalman filtering renewal:Screen out credible target
Uncorrelated measurement information in relevance assumption cluster, the processed frame numbers of N-Scan subtract 1, and update reliable flight path shelves
Case information.Return to step 1), next frame is measured and is associated judgement treatment.
2. described in claim 1 based on the radar data correlating methods for assuming algorithms are simplified more, its feature exists
In the step 2) likelihood ratio score matrix calculation procedure include:
21) LLR ratio that every flight path assumes branch, the score assumed as each are calculated;
22) score of all hypothesis branches is collected, is converted into track association score matrix.
3. described in claim 1 based on the radar data correlating methods for assuming algorithms are simplified more, its feature exists
In the step 3) using LAP algorithms obtain flight path level the optimal hypothesis of M- specific sub-step be:
31) by step 2) in obtain hypothesis flight path likelihood ratio track association score matrix substitute LAP methods in
The association probability for being used, as the property value assumed;
32) utilize step 31) in track association score matrix P_value, to it take the conversion of negative,
And then obtain LAP decision matrixs;
33) weight matrix is determined according to priori weight vectors;
34) M- optimal sequencing matrixes P is obtained according to LAP linear programming methods;
35) according to M- optimal sequencing matrix P, the optimal hypothesis of M- of present frame are obtained.
4. described in claim 3 based on the radar data correlating methods for assuming algorithms are simplified more, its feature exists
In the step 32) in the calculation procedure of LAP decision matrixs include:
321) to track association score matrix P_value take the conversion of negative:
322) track association cost matrix P_cost is obtainedk;
323) decision matrix D is obtained according to association cost matrixm×n, m is represented and is measured number, when n represents previous
The flight path at quarter assumes numbers of branches, decision matrix Dm×nIn element represent the k moment and measure and the boat at k-1 moment
Mark assumes the property value that customs station lump is matched somebody with somebody.
5. described in claim 3 based on the radar data correlating methods for assuming algorithms are simplified more, its feature exists
In the step 33) specific sub-step include:
331) according to the actual scene of multiple target tracking, a weight vectors ω=(ω is defined1,ω2), wherein element
Represent the association probability measured with target:ω=(p_Track1, p_Track2);
332) define current time reliability flight path to assume to be scored at Track1_value and Track2_value, calculate
Now weight vectors;
333) according to weight vectors, final weight matrix π is determined according to below equation.
Π=D × ωT
6. described in claim 3 based on the radar data correlating methods for assuming algorithms are simplified more, its feature exists
In the step 34) sub-step include:
341) whole weight matrix is traveled through, finding an initiation sequence makes global costIt is minimum;
342) based on initiation sequence, one of them or several matching sequence numbers are replaced successively, search weight matrix,
Replaced with secondary dominating sequence, form a new sequence;
343) the hypothesis cost representated by the sequence for obtaining increases successively, is assumed to be until forming M " optimal "
Only.
7. described in claim 3 based on the radar data correlating methods for assuming algorithms are simplified more, its feature exists
In the step 35) specific sub-step include:
351) according to step 34) M " optimal " of all targets that obtains " optimal " assumes it is assumed that retaining
In qualified measurement, and confirm to delete the relatively low unqualified measurement of score in target cluster from changing;
352) delete step 351 in temporary transient flight path archives) the middle unqualified relevance assumption rejected;
353) score of each credible target flight path is recalculated.
8. described in claim 3 based on the radar data correlating methods for assuming algorithms are simplified more, its feature exists
In the step 4) specific sub-step include:
41) the processed measurement frame number of N-Scan methods is judged;
If 42) now reception processing N frame amount is surveyed, retain after being judged the association of N frames using LAP algorithms
All hypothesis carry out global level beta pruning, for all confirmation targets, obtain it is optimal association matching, walked
It is rapid 5);
If 43) current time do not receive yet nth frame measurement, receive next frame measure, repeat step 1)
-4)。
9. described in claim 3 based on the radar data correlating methods for assuming algorithms are simplified more, its feature exists
In the step 42) specific sub-step include:
421) flight path obtained after being associated to multiframe is it is assumed that according to its track association score, repeat step 31)
- 33) global level LAP weight matrixs, are obtained;
422) weight matrix is traveled through, the optimal sequence for making global Least-cost is found;
423) according to optimal sequence, determine that the efficient association of credible target is measured.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201511018622.8A CN106934324A (en) | 2015-12-30 | 2015-12-30 | Based on the radar data correlating methods for simplifying many hypothesis algorithms |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201511018622.8A CN106934324A (en) | 2015-12-30 | 2015-12-30 | Based on the radar data correlating methods for simplifying many hypothesis algorithms |
Publications (1)
Publication Number | Publication Date |
---|---|
CN106934324A true CN106934324A (en) | 2017-07-07 |
Family
ID=59442300
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201511018622.8A Pending CN106934324A (en) | 2015-12-30 | 2015-12-30 | Based on the radar data correlating methods for simplifying many hypothesis algorithms |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106934324A (en) |
Cited By (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108303684A (en) * | 2018-01-31 | 2018-07-20 | 长沙深之瞳信息科技有限公司 | Ground surveillance radar multi-object tracking method based on radial velocity information |
CN109656271A (en) * | 2018-12-27 | 2019-04-19 | 杭州电子科技大学 | A kind of soft correlating method of track based on data correlation thought |
CN110145966A (en) * | 2019-02-27 | 2019-08-20 | 中国人民解放军空军工程大学 | A kind of finite time convergence control Global robust sliding formwork method of guidance |
CN110414396A (en) * | 2019-07-19 | 2019-11-05 | 中国人民解放军海军工程大学 | A kind of unmanned boat perception blending algorithm based on deep learning |
CN110646798A (en) * | 2019-09-23 | 2020-01-03 | 森思泰克河北科技有限公司 | Target track association method, radar and terminal equipment |
CN111812638A (en) * | 2020-07-17 | 2020-10-23 | 北京理工大学 | IMM-MHT multi-target tracking method |
CN113109761A (en) * | 2021-04-13 | 2021-07-13 | 北京工业大学 | Trajectory-oriented calculation time reduction method based on multi-hypothesis tracking algorithm |
CN113589252A (en) * | 2021-08-03 | 2021-11-02 | 东风汽车集团股份有限公司 | Multi-radar sensor multi-target tracking method based on MHT algorithm |
CN113687347A (en) * | 2021-06-22 | 2021-11-23 | 中国电子科技集团公司电子科学研究院 | Flight path cluster data processing method and system and readable storage medium |
CN114994633A (en) * | 2022-08-04 | 2022-09-02 | 四川九洲空管科技有限责任公司 | Method for secondary radar point track correlation processing |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103759732A (en) * | 2014-01-14 | 2014-04-30 | 北京航空航天大学 | Angle information assisted centralized multi-sensor multi-hypothesis tracking method |
CN103886196A (en) * | 2014-03-14 | 2014-06-25 | 中国电子科技集团公司第二十八研究所 | Track associating method resistant to big system errors |
CN104794345A (en) * | 2015-04-21 | 2015-07-22 | 四川大学 | Flat type multi-hypothesis associative processing method during trajectory tracking |
-
2015
- 2015-12-30 CN CN201511018622.8A patent/CN106934324A/en active Pending
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103759732A (en) * | 2014-01-14 | 2014-04-30 | 北京航空航天大学 | Angle information assisted centralized multi-sensor multi-hypothesis tracking method |
CN103886196A (en) * | 2014-03-14 | 2014-06-25 | 中国电子科技集团公司第二十八研究所 | Track associating method resistant to big system errors |
CN104794345A (en) * | 2015-04-21 | 2015-07-22 | 四川大学 | Flat type multi-hypothesis associative processing method during trajectory tracking |
Non-Patent Citations (3)
Title |
---|
KUSHA PANTA等: "Novel Data Association Schemes for the Probility Hypothesis Density Filter", 《IEEE》 * |
仵浩等: "基于LAP的MHT算法在多目标跟踪中的应用", 《现代防御技术》 * |
翟海涛: "多假设跟踪算法研究及其应用", 《信息化研究》 * |
Cited By (15)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108303684A (en) * | 2018-01-31 | 2018-07-20 | 长沙深之瞳信息科技有限公司 | Ground surveillance radar multi-object tracking method based on radial velocity information |
CN109656271A (en) * | 2018-12-27 | 2019-04-19 | 杭州电子科技大学 | A kind of soft correlating method of track based on data correlation thought |
CN110145966A (en) * | 2019-02-27 | 2019-08-20 | 中国人民解放军空军工程大学 | A kind of finite time convergence control Global robust sliding formwork method of guidance |
CN110414396A (en) * | 2019-07-19 | 2019-11-05 | 中国人民解放军海军工程大学 | A kind of unmanned boat perception blending algorithm based on deep learning |
CN110414396B (en) * | 2019-07-19 | 2021-07-16 | 中国人民解放军海军工程大学 | Unmanned ship perception fusion algorithm based on deep learning |
CN110646798A (en) * | 2019-09-23 | 2020-01-03 | 森思泰克河北科技有限公司 | Target track association method, radar and terminal equipment |
CN110646798B (en) * | 2019-09-23 | 2021-09-28 | 森思泰克河北科技有限公司 | Target track association method, radar and terminal equipment |
CN111812638B (en) * | 2020-07-17 | 2022-07-29 | 北京理工大学 | IMM-MHT multi-target tracking method |
CN111812638A (en) * | 2020-07-17 | 2020-10-23 | 北京理工大学 | IMM-MHT multi-target tracking method |
CN113109761A (en) * | 2021-04-13 | 2021-07-13 | 北京工业大学 | Trajectory-oriented calculation time reduction method based on multi-hypothesis tracking algorithm |
CN113687347A (en) * | 2021-06-22 | 2021-11-23 | 中国电子科技集团公司电子科学研究院 | Flight path cluster data processing method and system and readable storage medium |
CN113687347B (en) * | 2021-06-22 | 2023-11-17 | 中国电子科技集团公司电子科学研究院 | Track cluster data processing method, system and readable storage medium |
CN113589252A (en) * | 2021-08-03 | 2021-11-02 | 东风汽车集团股份有限公司 | Multi-radar sensor multi-target tracking method based on MHT algorithm |
CN114994633A (en) * | 2022-08-04 | 2022-09-02 | 四川九洲空管科技有限责任公司 | Method for secondary radar point track correlation processing |
CN114994633B (en) * | 2022-08-04 | 2022-11-08 | 四川九洲空管科技有限责任公司 | Method for secondary radar point track correlation processing |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN106934324A (en) | Based on the radar data correlating methods for simplifying many hypothesis algorithms | |
CN107063259B (en) | Track association method and electronic equipment | |
CN104237879B (en) | A kind of multi-object tracking method in radar system | |
CN104931934B (en) | A kind of radar plot condensing method based on PAM cluster analyses | |
CN109508000A (en) | Isomery multi-sensor multi-target tracking method | |
CN106680806A (en) | Multi-radar plot fusion method | |
CN106054169B (en) | Multistation Radar Signal Fusion detection method based on tracking information | |
CN109002835A (en) | A kind of particle filter data correlation method based on maximum entropy fuzzy clustering | |
CN103759732B (en) | A kind of centralized multisensor multiple hypotheis tracking method of angle information auxiliary | |
CN103941233B (en) | The radar interval alternately radiation control method of tracking worked in coordination with by a kind of multi-platform main passive sensor | |
CN105761276B (en) | Based on the iteration RANSAC GM-PHD multi-object tracking methods that adaptively newborn target strength is estimated | |
CN107192998A (en) | A kind of adapter distribution track data fusion method based on covariance target function | |
CN104793200B (en) | Dynamic planning track-before-detect method based on iterative processing | |
CN110376581B (en) | Explicit multi-target tracking method based on Gaussian mixture probability hypothesis density filter | |
CN110780269A (en) | Explicit multi-target tracking method based on GM-PHD filter under self-adaptive new growth intensity | |
Smith et al. | Systematic analysis of the PMBM, PHD, JPDA and GNN multi-target tracking filters | |
CN103247057A (en) | Road target multi-hypothesis tracking algorithm under target-echo-road network data association | |
Kwon et al. | A robust mobile target localization method for cooperative unmanned aerial vehicles using sensor fusion quality | |
CN104008403B (en) | A kind of SVM(Vector machine)The multi-targets recognition decision method of pattern | |
CN108152789A (en) | Utilize the passive track-corelation data correlation and localization method of RSS information | |
CN104501812A (en) | Filtering algorithm based on self-adaptive new target strength | |
CN107346020A (en) | A kind of distribution for asynchronous multi-static radar system batch estimation fusion method | |
CN109190647B (en) | Active and passive data fusion method | |
Solonar et al. | General construction principles and performance features of trajectory processing by data from one radar data source | |
CN106291530A (en) | A kind of probabilistic data association optimization method based on nearest neighbor method |
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 | ||
WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20170707 |
|
WD01 | Invention patent application deemed withdrawn after publication |