CN105651284B - The method and device of raising experience navigation interior joint efficiency of selection - Google Patents
The method and device of raising experience navigation interior joint efficiency of selection Download PDFInfo
- Publication number
- CN105651284B CN105651284B CN201511016245.4A CN201511016245A CN105651284B CN 105651284 B CN105651284 B CN 105651284B CN 201511016245 A CN201511016245 A CN 201511016245A CN 105651284 B CN105651284 B CN 105651284B
- Authority
- CN
- China
- Prior art keywords
- experience
- node
- matched
- efficiency
- prior probability
- 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.)
- Active
Links
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
- G01C21/00—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
- G01C21/20—Instruments for performing navigational calculations
Landscapes
- Engineering & Computer Science (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Automation & Control Theory (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
A kind of raising experience is navigated the method and device of interior joint efficiency of selection, and wherein method comprises the following steps, sets up prior probability model, and experience database data substitution prior probability model is obtained the prior probability of node to be matched；Posterior probability model is set up, experience database data substitution posterior probability model is obtained the posterior probability of node to be matched；The matching efficiency score of node to be matched is obtained according to the prior probability and posterior probability, the matching operation object that the node to be matched of highest scoring is preferentially empirically navigated.Above-mentioned technical proposal carries out effective screening criteria by setting up matching efficiency scoring mechanism for node to be matched is given a mark with score, improves the efficiency of selection of experience navigation interior joint, further increasing the computational efficiency of experience navigation.
Description
Technical field
Improved during visual experience is navigated and node selection the present invention relates to unmanned vehicle vision guided navigation field, more particularly to one kind
The method and device of efficiency.
Background technology
Some experiences that have in the prior art of unmanned vehicle navigation are navigated (experience based navigation), this
It is a kind of method of the autonomous navigation that unmanned vehicle is helped using visual experience.So-called experience is not the warp on ordinary meaning
Test, but a series of continuous pictures.Unmanned vehicle is by extracting the current image that perceptron is passed back with the continuous picture for prestoring
Face, and the characteristic (features) of both sides is contrasted, draw the conclusion for whether matching.If the match is successful, unmanned vehicle can be straight
The geographical position for calling the feature for prestoring is connect, is intuitively positioned oneself.
Simple process is seemed, it is necessary to unmanned vehicle is worked in concert with lower component in order to complete this：
1. perceptron
The perceptron of unmanned vehicle is typically 3D camera lenses, or laser radar.Experience may be, but not limited to, picture.
Any computer can recognize that processing, and react the perceptron of Current traffic and background situation can form experience.This
It is a big advantage of empirical algorithms.Perceptron needs the current environment of real-time capture, keeps the stabilization of image and continuous.
2. Character Comparison algorithm
After image is transmitted in real time, the computing unit of unmanned vehicle needs very quickly analysis, extracts the information in image,
Commonly referred to as characteristic.So-called characteristic is a series of points that can portray object image, and these points are typically distributed on the side of object
Edge, produces the sensation of " discontinuous " in the picture.Having many algorithms at present can process less complicated image, there is some algorithms
Extremely complex image can be processed, their degree of accuracy is usual and calculates time correlation.Slow algorithm is restrained, is more had
Possible fine correlation.Herein, it would be desirable to select quick and easy algorithm, in real time positioning.
3. experience matching
Experience matching is this emphasis point.Experience matching can be made very simple, intuitive, be carried from realtime graphic
Feature in the feature and experience storehouse that take all compares to be gone over, which matching, and which mismatches very clear.If calculating single
Unit is powerful enough, and this is nor a good method.Reason is if several scenes seem similar, or further to increase
Plus difficulty, make the illumination condition of several scenes different, possible certain scene and its experience can not be matched, but with other one
Individual unrelated experience matching.This mistake of generation, the navigation of unmanned vehicle can have serious problems.
There is a kind of experience mechanics in some embodiments, allow each scene to possess any many experiences.It is diverse
Scene experience number is more, otherwise experience number is reduced.The thinking of this distribution according to need can to a certain degree alleviate above-mentioned problem.
But, scene matching degree improves and has new problem to occur：Several experiences are all matched, it is difficult to which what if is choiceIf this
Several experiences belong to a scene, and problem is not also serious.If scene is all different, still can not use.
More process necessary fast it is essential that calculating.The speed of vehicle traveling does not allow dividing for one minute a few minutes an of frame
Analysis.It is easy for out originally partially plus prolonged autonomous navigation, it would be desirable to a solution for fast and stable.
The content of the invention
For this reason, it may be necessary to provide the node selecting method in a kind of experience database, the selection of experience navigation interior joint is improved
Efficiency, further improves the computational efficiency of experience navigation.
To achieve the above object, a kind of method of raising experience navigation interior joint efficiency of selection is inventor provided, including
Following steps, set up prior probability model, and the priori that experience database data substitution prior probability model is obtained into node to be matched is general
Rate；Posterior probability model is set up, experience database data substitution posterior probability model is obtained the posterior probability of node to be matched；According to
The prior probability and posterior probability obtain the matching efficiency score of node to be matched, and the node to be matched of highest scoring is preferential
The matching operation object for empirically navigating.
Preferably, described " the matching efficiency score of node to be matched is obtained according to the prior probability and posterior probability "
Specifically, the product of prior probability and posterior probability to be obtained the matching efficiency score of node to be matched multiplied by efficiency factor.
Specifically, the experience database data includes historical experience database data and this experience navigation data.
A kind of device of raising experience navigation interior joint efficiency of selection, including prior probability computing module, posterior probability meter
Calculate module, efficiency score computing module；
The prior probability computing module is used to set up prior probability model, and experience database data is substituted into prior probability model
Obtain the prior probability of node to be matched；
The posterior probability computing module is used to set up posterior probability model, and experience database data is substituted into posterior probability model
Obtain the posterior probability of node to be matched；
The efficiency score computing module be used for according to the prior probability and posterior probability obtain node to be matched
With efficiency score, the matching operation object that the node to be matched of highest scoring is preferentially empirically navigated.
Preferably, the efficiency score computing module is additionally operable to the product of prior probability and posterior probability multiplied by efficiency system
Number obtains the matching efficiency score of node to be matched.
Specifically, the experience database data includes historical experience database data and this experience navigation data.
Prior art is different from, above-mentioned technical proposal is by setting up matching efficiency scoring mechanism for node to be matched is beaten
Point, and effective screening criteria is carried out with score, the experience that improves is navigated the efficiency of selection of interior joint, further increasing through
Test the computational efficiency of navigation.
Brief description of the drawings
Fig. 1 is the execution pipeline schematic diagram of the experience navigation described in the specific embodiment of the invention；
Fig. 2 is three kinds of different experience schematic diagrames described in the specific embodiment of the invention；
Fig. 3 is the method schematic diagram of the raising experience navigation interior joint efficiency of selection described in the specific embodiment of the invention；
Fig. 4 is the posterior probability model schematic diagram described in the specific embodiment of the invention；
Fig. 5 is the apparatus module figure of the raising experience navigation interior joint efficiency of selection described in the specific embodiment of the invention.
Description of reference numerals：
500th, prior probability computing module；
502nd, posterior probability computing module；
504th, efficiency score computing module.
Specific embodiment
To describe technology contents, structural feature, the objects and the effects of technical scheme in detail, below in conjunction with specific reality
Apply example and coordinate accompanying drawing to be explained in detail.
1st, principles and methods
For experience navigation system, preferably operation hardware should possess multi-microprocessor (CPU) or kernel.
What the embodiment shown in Fig. 1 illustrated the execution pipeline (pipeline) of experience navigation is example.Each horizontal stripe in figure (1,
2nd, 3 an independent process (process) 4), is represented, streamline has four processes in parallel running when most.
Streamline is divided into several steps,
1. after obtaining original image from stereo camera, feature extractor is run first, and the feature in original image is carried
Take out.
2. these features are input into the image of visual odometry process, the current characteristics of image of this process contrast and back
Feature, estimates the current position of robot.
3. the information of newest offset estimation and characteristics of image, is next used for updating empirical data by vision navigation system
Storehouse, while attempting the positioning robot in existing correlation experience.
Wherein, the necessary series operation of 1.2. steps, and the 3rd step can be with parallel running.Because every in experience database
Individual experience can carry out computing independently of other experiences, between the data that do not interdepend, it is possible to be separately operable
In different processes.4 such processes (Np=4) have been used in figure.After one process completes a computing for experience, such as
Fruit time restriction is not arrived, and can also continue to carry out the computing of another experience.Vertical dotted line (online- in Fig. 1
Performance) it is the time restriction of streamline execution.Because this is limited, four processes are respective respectively in figure
Have to stop when the computing of second experience of execution, execution pipeline enters next round computing, processes new original graph
Picture.
Because experience navigation needs to provide real-time operation result, so the execution pipeline of whole experience navigation has
Rigid, strict execution time restriction.If it will be apparent that can effectively reduce experience navigation (4 processes parallel portion simultaneously in figure
Point) used by time, then can improve the quality experience positions calculations of more wheels (perform) of experience navigation, or reduce for car
Carry the demand of computer hardware.
The time that experience navigational portions spend mainly attempts positioning one by one in all nodes of current experience
Match somebody with somebody, untill success.Therefore, it is very crucial for for sequencing that these nodes are selected：If select must well, it be likely to
The node of matching is found quickly, and diverted via position success of assaying, computing is completed；, whereas if selecting bad, it is likely that waste a large amount of
Operation time is on unnecessary node.The node selecting method that proposes of a previous piece be based on node and current estimated location it
Between distance, this method is than randomly choosing far better, but the space that is still significantly improved.This piece proposes that a kind of calculating is saved
Put the probability distribution of successfully positioning, and the method for the node that prioritizing selection most probable is successfully positioned.
2 general thoughts
The image that experiential description unmanned vehicle is experienced in the motion in one continuous time and space.The continuity of space-time
In necessarily extending to image.Fine day will not suddenly become the rainy day and become fine day again again, at least in a short period of time, this across not
Too may.
We set up a set of criterion by using this inherent uniformity.For example, unmanned vehicle was in a few minutes in past
All successful match fine days, when next candidate matches experience is found, the experience produced by fine day should be examined first
Consider.It is fine day or rainy day that unmanned vehicle will not go resolution, but unmanned vehicle can rule of thumb in each node relative position,
With passing matching result, can a node be matched and makes anticipation.And this anticipation will determine the elder generation that matching algorithm is performed
Order afterwards.
This sequence does not directly affect the accuracy of matching, can but improve efficiency.
This criterion is a conditional probability model.So-called conditional probability, is some events being known to occur
In the case of, then there is the possibility of something.Set is known from somewhere process in the scene of unmanned vehicle, below may
Reach the possibility in somewhere.In order to calculate this model, it would be desirable to prior model and posterior model.
2.1 prior probability models
In the problem of the navigation of unmanned vehicle, stabilization represents two implications, it is not easy to malfunctions and has gone out mistake and be not to cause
Bigger trouble.
From computing is collected, do not malfunction completely unrealistic.Effective assessment error is probably the first step for reaching stabilization.
By the experience constantly accumulated successfully and fail, it can gradually erect a model for macroscopic view.Need what is made a decision again afterwards
When, this model can help unmanned vehicle to exclude the possibility option of some apparent errors.
Probabilistic model reaches stable state after being run multiple times, it might even be possible to is treated as a part for experience, has supplemented
Kind perceptron cannot directly tell some " Implicit Conditions " of unmanned vehicle.We are the probability for describing this changeless knowledge
Model is called prior model.Specifically, it is possible to understand that into unmanned vehicle traveling possibility somewhither.
2.2 posterior probability models
Posterior probability model is also a conditional probability.That is, its generation depends on the generation of some events.Tool
, to the problem of unmanned vehicle, we are it is to be understood that the present position of known vehicle, vehicle is in the past period by somewhere for body
Possibility.
Relative to prior probability model, posterior probability model is that one kind is searched one's heart, self-examination.That did just now determines
It is fixed, many careful all possible mistakes, posterior probability model can help unmanned vehicle to recognize the possibility of error, so as to avoid more
Further go astray.
Priori and posteriority emphasize particularly on different fields a little, combine them and use, and can just obtain the sort method of our needs.
We are just discussed in detail this method in the following examples.
3. realize in detail
Assuming that unmanned vehicle has run a period of time, have a number of experience.This time is changeable, experience
Number it is The more the better.In embodiment as shown in Figure 2, each experience represents a kind of weather condition, more general feelings
The magnitude of traffic flow and the magnitude of traffic flow of class period of shape, such as school go to school that classes are in front of the door period, or road construction etc. feelings
Condition is also suitable.In order to describe simplicity, we select Changes in weather as unified case.
Since unmanned vehicle has run a period of time, then can define a set W, we are referred to as history collection.Namely
A period of time was once selected as that the node of object may be matched recently, regardless of whether successful match, is completely included into this set.Together
When unmanned vehicle also remember the result that matches each time, we are with Z come referred to as.
Unmanned vehicle, due to the distance in geographical position, can also mark a Candidate Set at each moment from all nodes
Y, comprising be possible to selected node.It is assumed that it is normal solution that must wherein have one and only one.This suppose there is one
Fixed risk, but in the long term without big problem.After all, with the continuous accumulation of experience, the possibility that new scene occurs
It is more and more lower.Our criterion, simple saying is that it is the general of normal solution to calculate each node in the case where assuming known to history collection
Rate.Due to normal solution one and only one, this calculating is meaningful.By Bayesian formula, we can be by this unknown probability
Function decomposition, complete mathematics notation is as follows：
We have seen that unknown probability (equation left side) can be decomposed into two parts.English likelihood signified part is
Our aforementioned posterior probability, prior refers to prior probability.Therefore, please see Figure shown in 3, be of the invention one
The method flow diagram of raising experience navigation interior joint efficiency of selection is planted, is comprised the following steps, S300 sets up prior probability model, will
Experience database data substitutes into the prior probability that prior probability model obtains node to be matched；S302 sets up posterior probability model, will be through
Test database data and substitute into the posterior probability that posterior probability model obtains node to be matched；S304 is general according to the prior probability and posteriority
Rate obtains the matching efficiency score of node to be matched, the matching fortune that the node to be matched of highest scoring is preferentially empirically navigated
Calculate object.Wherein, the node to be matched is select node in Candidate Set, and the experience database data can include history
Experience database data collection, it is also possible to including this experience navigation data, so-called historical experience database data collection can be one or many
The matched data result of experience navigation, traveling navigation is more, and the historical experience database data collection of composition is bigger, this experience
The data of navigation include the node of many matchings of having navigated, and the Data Matching result that these have navigated can be used for experience storehouse number
According to foundation, be more suitable for the node matching in the navigation of current experience.Setting up matching efficiency scoring mechanism by the above method is
Node to be matched is given a mark, and carries out effective screening criteria with score, improves the efficiency of selection of experience navigation interior joint,
Further increasing the computational efficiency of experience navigation.
In some preferred embodiments, can be obtained multiplied by efficiency factor by by the product of prior probability and posterior probability
The matching efficiency score of node to be matched, the β in above formula can be as the inverse of the matching efficiency coefficient in some embodiments.Phase
For can be by the algorithm of weighted scoring respectively than in other embodiment, multiplication obtains matching efficiency score and meets Bayes
Inference, more science, preferably improve the node efficiency of selection in experience navigation.
We introduce specific calculating means respectively by 3.1 and 3.2 two embodiments below.
3.1 posterior probability
Shown in Fig. 2, light color is represented and successfully positioned, and dark color represents positioning failure, and white represents the point for not yet positioning.Solid line and
Dotted line represents two different experiences respectively.Due to n3Successful positioning, the position success rate of experience one is very high, therefore n9Into
Power should be more than n5Success rate.Posterior probability description is exactly this information that counter can be pushed away by accomplished fact.
We define：
θI, j=p (zj| y=i)
Namely known i-th both candidate nodes are selected, once by j-th probability of node.Here which is phase
It is not overall identification for for their own set (Y, W).
Its specific formula is as follows：
Wherein variable ZI, jWhat is represented is the number of times for having in experience storehouse experience to lead to both candidate nodes i from node j.Parameter alpha,
Elect 1 in practical operation as.
Assuming that the positioning of each node is independent, then the posterior probability by whole collection in the past is as follows：
Wherein our defined feature functions：
Whether to distinguish by somewhere.
3.2 prior probabilities
Prior probability is more directly perceived：Known current location, next step reaches the probability in somewhere.
There is the number of times that experience passes through both candidate nodes i by node n_k in the experience storehouse that wherein variable N_ (i, k) is represented.Ginseng
1 is elected as in the middle of number γ practices.As can be seen that the node prior probability through multiple experiences is big, this result that exactly we wait.
Finally, in the particular embodiment, we are multiplied priori with posterior probability and obtain：
According to above-mentioned Bayesian formula it is recognised that the size of this numerical value is proportional to matching efficiency score, will determine
The ranking of all nodes in Candidate Set, whenever by a node, experience guider is carried out above method step, finally will
Ranking highest node override to be matched substitutes into experience navigation and carries out computing.By the above method, ranking highest is to be matched
Node most possibly obtains the matching of experience navigation data, saves the calculating time of single match needs, and the experience of improve is led
Boat interior joint efficiency of selection, while also further increasing the efficiency that experience navigation is calculated.
In the embodiment shown in fig. 5, a kind of device of raising experience navigation interior joint efficiency of selection, including elder generation are described
Test probability evaluation entity 500, posterior probability computing module 502, efficiency score computing module 504；
The prior probability computing module 500 is used to set up prior probability model, and experience database data is substituted into prior probability
Model obtains the prior probability of node to be matched；
The posterior probability computing module 502 is used to set up posterior probability model, and experience database data is substituted into posterior probability
Model obtains the posterior probability of node to be matched；
The efficiency score computing module 504 is used to obtain node to be matched according to the prior probability and posterior probability
Matching efficiency score, the matching operation object that the node to be matched of highest scoring is preferentially empirically navigated.Said apparatus lead to
Cross and set up matching efficiency scoring mechanism for node to be matched is given a mark, and effective screening criteria is carried out with score, improve
The efficiency of selection of experience navigation interior joint, further increasing the computational efficiency of experience navigation.
Preferably, the efficiency score computing module 504 is additionally operable to the product by prior probability and posterior probability multiplied by effect
Rate coefficient obtains the matching efficiency score of node to be matched.The effect of raising experience navigation computational efficiency is preferably reached.
Specifically, the experience database data includes historical experience database data and this experience navigation data.
It should be noted that herein, such as first and second or the like relational terms are used merely to a reality
Body or operation make a distinction with another entity or operation, and not necessarily require or imply these entities or deposited between operating
In any this actual relation or order.And, term " including ", "comprising" or its any other variant be intended to
Nonexcludability is included, so that process, method, article or terminal device including a series of key elements not only include those
Key element, but also other key elements including being not expressly set out, or also include being this process, method, article or end
The intrinsic key element of end equipment.In the absence of more restrictions, limited by sentence " including ... " or " including ... "
Key element, it is not excluded that also there is other key element in the process including the key element, method, article or terminal device.This
Outward, herein, " it is more than ", " being less than ", " exceeding " etc. are interpreted as not including this number；" more than ", " below ", " within " etc. understand
It is to include this number.
It should be understood by those skilled in the art that, the various embodiments described above can be provided as method, device or computer program producing
Product.These embodiments can be using the embodiment in terms of complete hardware embodiment, complete software embodiment or combination software and hardware
Form.All or part of step in the method that the various embodiments described above are related to can be instructed by program correlation hardware come
Complete, described program can be stored in the storage medium that computer equipment can read, for performing the various embodiments described above side
All or part of step described in method.The computer equipment, including but not limited to：Personal computer, server, general-purpose computations
Machine, special-purpose computer, the network equipment, embedded device, programmable device, intelligent mobile terminal, intelligent home device, Wearable
Smart machine, vehicle intelligent equipment etc.；Described storage medium, including but not limited to：RAM, ROM, magnetic disc, tape, CD, sudden strain of a muscle
Deposit, USB flash disk, mobile hard disk, storage card, memory stick, webserver storage, network cloud storage etc..
The various embodiments described above are with reference to the method according to embodiment, equipment (system) and computer program product
Flow chart and/or block diagram are described.It should be understood that every during flow chart and/or block diagram can be realized by computer program instructions
The combination of flow and/or square frame in one flow and/or square frame and flow chart and/or block diagram.These computers can be provided
Programmed instruction is to the processor of computer equipment producing a machine so that by the finger of the computing device of computer equipment
Order is produced for realizing what is specified in one flow of flow chart or multiple one square frame of flow and/or block diagram or multiple square frames
The device of function.
These computer program instructions may be alternatively stored in the computer that computer equipment can be guided to work in a specific way and set
In standby readable memory so that instruction of the storage in the computer equipment readable memory is produced and include the manufacture of command device
Product, the command device is realized in one flow of flow chart or multiple one square frame of flow and/or block diagram or multiple square frame middle fingers
Fixed function.
These computer program instructions can be also loaded on computer equipment so that performed on a computing device a series of
Operating procedure is to produce computer implemented treatment, so that the instruction for performing on a computing device is provided for realizing in flow
The step of function of being specified in one flow of figure or multiple one square frame of flow and/or block diagram or multiple square frames.
Although being described to the various embodiments described above, those skilled in the art once know basic wound
The property made concept, then can make other change and modification to these embodiments, so embodiments of the invention are the foregoing is only,
Not thereby scope of patent protection of the invention, the equivalent structure that every utilization description of the invention and accompanying drawing content are made are limited
Or equivalent flow conversion, or other related technical fields are directly or indirectly used in, similarly it is included in patent of the invention
Within protection domain.
Claims (6)
1. a kind of raising experience is navigated the method for interior joint efficiency of selection, it is characterised in that is comprised the following steps, is set up priori general
Rate model, experience database data substitution prior probability model is obtained the prior probability of node to be matched；Set up posterior probability model,
Experience database data substitution posterior probability model is obtained the posterior probability of node to be matched；It is general according to the prior probability and posteriority
Rate obtains the matching efficiency score of node to be matched, the matching fortune that the node to be matched of highest scoring is preferentially empirically navigated
Calculate object；
The posterior probability model is：
Wherein, variable ZI, jWhat is represented is the number of times for having in experience storehouse experience to lead to both candidate nodes i from node j, and α is to make ginseng by oneself
Number, w is the set of node；
The prior probability model is：
Wherein variable NI, kThere is the number of times that experience passes through both candidate nodes i by node Nk in the experience storehouse of representative；γ is to make ginseng by oneself
Number；Y is the set of node.
2. raising experience according to claim 1 is navigated the method for interior joint efficiency of selection, it is characterised in that " the root
The matching efficiency score of node to be matched is obtained according to the prior probability and posterior probability " specifically, by prior probability and posteriority
The product of probability obtains the matching efficiency score of node to be matched multiplied by efficiency factor.
3. raising experience according to claim 1 is navigated the method for interior joint efficiency of selection, it is characterised in that the experience
Database data includes historical experience database data and this experience navigation data.
4. a kind of raising experience is navigated the device of interior joint efficiency of selection, it is characterised in that including prior probability computing module, after
Test probability evaluation entity, efficiency score computing module；
The prior probability computing module is used to set up prior probability model, and experience database data substitution prior probability model is obtained
To the prior probability of node to be matched；
The posterior probability computing module is used to set up posterior probability model, and experience database data substitution posterior probability model is obtained
The posterior probability of node to be matched；
The efficiency score computing module is used to be obtained according to the prior probability and posterior probability the matching effect of node to be matched
Rate score, the matching operation object that the node to be matched of highest scoring is preferentially empirically navigated；
The posterior probability model is：
Wherein, variable ZI, jWhat is represented is the number of times for having in experience storehouse experience to lead to both candidate nodes i from node j, and α is to make ginseng by oneself
Number, w is the set of node；
The prior probability model is：
Wherein variable NI, kThere is experience by node N in the experience storehouse of representativekBy the number of times of both candidate nodes i；γ is to make parameter by oneself；
Y is the set of node.
5. raising experience according to claim 4 is navigated the device of interior joint efficiency of selection, it is characterised in that the efficiency
Points calculating module is additionally operable to obtain the product of prior probability and posterior probability multiplied by efficiency factor the matching of node to be matched
Efficiency score.
6. raising experience according to claim 4 is navigated the method for interior joint efficiency of selection, it is characterised in that the experience
Database data includes historical experience database data and this experience navigation data.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201511016245.4A CN105651284B (en) | 2015-12-29 | 2015-12-29 | The method and device of raising experience navigation interior joint efficiency of selection |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201511016245.4A CN105651284B (en) | 2015-12-29 | 2015-12-29 | The method and device of raising experience navigation interior joint efficiency of selection |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105651284A CN105651284A (en) | 2016-06-08 |
| CN105651284B true CN105651284B (en) | 2017-06-20 |
Family
ID=56478520
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201511016245.4A Active CN105651284B (en) | 2015-12-29 | 2015-12-29 | The method and device of raising experience navigation interior joint efficiency of selection |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105651284B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111474932B (en) * | 2020-04-23 | 2021-05-11 | 大连理工大学 | Mobile robot mapping and navigation method integrating scene experience |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104807465A (en) * | 2015-04-27 | 2015-07-29 | 安徽工程大学 | Method and device for realizing simultaneous localization and mapping of robots |
| CN104833964A (en) * | 2015-04-28 | 2015-08-12 | 南京邮电大学 | Object detection tracking integrated method and device aiming at radar/sonar system |
| CN105157704A (en) * | 2015-06-03 | 2015-12-16 | 北京理工大学 | Bayesian estimation-based particle filter gravity-assisted inertial navigation matching method |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8478642B2 (en) * | 2008-10-20 | 2013-07-02 | Carnegie Mellon University | System, method and device for predicting navigational decision-making behavior |
-
2015
- 2015-12-29 CN CN201511016245.4A patent/CN105651284B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104807465A (en) * | 2015-04-27 | 2015-07-29 | 安徽工程大学 | Method and device for realizing simultaneous localization and mapping of robots |
| CN104833964A (en) * | 2015-04-28 | 2015-08-12 | 南京邮电大学 | Object detection tracking integrated method and device aiming at radar/sonar system |
| CN105157704A (en) * | 2015-06-03 | 2015-12-16 | 北京理工大学 | Bayesian estimation-based particle filter gravity-assisted inertial navigation matching method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105651284A (en) | 2016-06-08 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US10970854B2 (en) | Visual target tracking method and apparatus based on deep adversarial training | |
| US20230186486A1 (en) | Vehicle tracking method and apparatus, and electronic device | |
| Huang et al. | Robust visual tracking via constrained multi-kernel correlation filters | |
| Adhikari et al. | Iterative bounding box annotation for object detection | |
| CN112560656A (en) | Pedestrian multi-target tracking method combining attention machine system and end-to-end training | |
| CN113657560B (en) | Weak supervision image semantic segmentation method and system based on node classification | |
| KR102305230B1 (en) | Method and device for improving accuracy of boundary information from image | |
| CN111523463B (en) | Target tracking method and training method based on matching-regression network | |
| CN111126249A (en) | Pedestrian re-identification method and device combining big data and Bayes | |
| CN115512251A (en) | Unmanned aerial vehicle low-illumination target tracking method based on double-branch progressive feature enhancement | |
| Yao et al. | Dynamicbev: Leveraging dynamic queries and temporal context for 3d object detection | |
| CN112258557B (en) | Visual tracking method based on space attention feature aggregation | |
| KR101700030B1 (en) | Method for visual object localization using privileged information and apparatus for performing the same | |
| CN105651284B (en) | The method and device of raising experience navigation interior joint efficiency of selection | |
| CN113592015A (en) | Method and device for positioning and training feature matching network | |
| CN115272705B (en) | Training method, device and equipment for saliency object detection model | |
| CN116543417A (en) | Human body posture estimation method, device, equipment and storage medium | |
| CN112541515A (en) | Model training method, driving data processing method, device, medium and equipment | |
| CN115239889A (en) | Training method of 3D reconstruction network, 3D reconstruction method, device, equipment and medium | |
| CN115424012A (en) | Lightweight image semantic segmentation method based on context information | |
| CN114664410A (en) | Video-based focus classification method and device, electronic equipment and medium | |
| US11106344B2 (en) | Methods and devices for capturing heuristic information via a weighting tool | |
| CN113887501A (en) | Behavior recognition method and device, storage medium and electronic equipment | |
| CN113326885A (en) | Method and device for training classification model and data classification | |
| CN114237394B (en) | Motion recognition method, device, equipment and medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |