CN109813319A - A kind of open loop optimization method and system for building figure based on SLAM - Google Patents
A kind of open loop optimization method and system for building figure based on SLAM Download PDFInfo
- Publication number
- CN109813319A CN109813319A CN201910172096.2A CN201910172096A CN109813319A CN 109813319 A CN109813319 A CN 109813319A CN 201910172096 A CN201910172096 A CN 201910172096A CN 109813319 A CN109813319 A CN 109813319A
- Authority
- CN
- China
- Prior art keywords
- information
- map
- building
- slam
- open loop
- 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
Landscapes
- Control Of Position, Course, Altitude, Or Attitude Of Moving Bodies (AREA)
Abstract
The present disclosure proposes a kind of open loop optimization method and system for building figure based on SLAM, the locating local environment cartographic information for needing to build figure equipment is obtained, and map is carried out based on the local environment cartographic information；It under open loop situations, needs to build figure equipment when receiving external absolute fix information, establishes the constraint between absolute fix information and current kinetic location information, the building of global map is carried out according to constraint information.Disclosed technique scheme optimizes the addition for receiving absolute location information by open loop, ensure that the unification of coordinate system, helps to solve the problems, such as map fusion when multirobot cooperates with map structuring.
Description
Technical field
The present invention relates to field of navigation technology, the open loop optimization method of figure is built based on SLAM and it is more particularly to a kind of
System.
Background technique
Immediately positioning and map structuring (SimultaneousLocalization and Mapping, SLAM) technology, i.e.,
In a static circumstances not known, by the movement and measurement of robot, carry out academic environment map, and at the same time determining machine
Position of the people in map.
For popular, SLAM answer two problems: " I where? " " what is around me? ", just as people to one
The same in a foreign environment, what SLAM attempted to solve is exactly the space pass for recovering observer itself and ambient enviroment
System, corresponding " I where " is exactly orientation problem, and it is exactly to build figure problem that " what is around me " is corresponding, provides ring around
One description in border.
SLAM problem is proposed by the article that Cheeseman and Smith in 1985 writes earliest, has founded description geometry in text
The Principle of Statistics of correlation between uncertain and feature and feature, these principles constitute the mathematics for solving SLAM problem
Basis.Smith, Self and Cheeseman are in SLAM algorithm of the proposition based on Kalman filtering in 1987, the Theoretical Research Framework
The SLAM algorithm based on Kalman filtering is set to become most widely used method.Since the proposition of the SLAM concept eighties in last century
Till now, SLAM technology has been passed by more than 30 years history.The sensor that SLAM system uses constantly is being expanded, from early stage
Sonar, 2D/3D laser radar finally, then arrive the various cameras such as monocular, binocular, RGBD, ToF, and with inertia measurement list
The fusion of the sensors such as first IMU；The algorithm of SLAM is also from the method (EKF, PF etc.) based on filter of beginning to based on optimization
Method transformation, technological frame is also from the single thread of beginning to multithreading evolution.
Unmanned vehicle is a kind of wheeled robot.In recent years, with the development of artificial intelligence technology, unmanned vehicle technology at
For currently very popular one of field.Wherein satellite-based navigator fix technology development relative maturity, but due to
Satellite-signal is vulnerable to external interference, and since blocking for high building, bridge, tunnel etc. causes GPS signal in urban road
It loses, GPS navigation is not available under indoor environment, therefore development does not depend on the location technology of satellite and has become necessarily to become
Gesture, SLAM technology become the important method for solving this technical need.Unmanned vehicle is as a kind of special machine people, to ambient enviroment
The building of map and for itself accurate positioning in the environment be realize unmanned vehicle safety traffic premise.Unmanned vehicle SLAM
Technology is the process for carrying out perception cognition to ambient enviroment by onboard sensor (camera, laser radar, ultrasonic wave etc.),
It is a main investigative technique in the following artificial intelligence direction.
It includes vision SLAM and laser SLAM that SLAM, which builds diagram technology, wherein laser SLAM technology is due to developing more early, technology
Comparative maturity is constantly in technology leading position.
Vision SLAM is cheaper compared to laser SLAM, light, and image information is more abundant, but to the reality of data calculation process
When property there are certain requirements, and the influence vulnerable to environmental change.Laser radar is high with precision, speed is fast, it is dynamic to be enable to respond quickly
The features such as environmental change under quiet state, therefore the reliability of laser SLAM technology and safety are superior to image SLAM.
No matter which kind of SLAM technology will rely on the position where last moment itself during founding a capital figure come under determining
The position at one moment itself.Since any sensor has certain error, figure process is built in the robot based on laser SLAM
In, error can gradually add up, and be modified when that can be detected by local closed loop in the lesser environment of scene to deviation.But
Under giant single and gallery environment, accumulated error is excessive, will lead to robot can not closed loop, build figure failure.
Summary of the invention
The purpose of this specification embodiment is to provide a kind of open loop optimization method that figure is built based on SLAM, can be according to ring
Absolute location information in border carries out the real-time map building that slam builds figure, suitable for robot stabilized under large scene build figure and
Multirobot cooperates with map structuring.
This specification embodiment provides a kind of open loop optimization method that figure is built based on SLAM, real by the following technical programs
It is existing:
The locating local environment cartographic information for needing to build figure equipment is obtained, and is carried out based on the local environment cartographic information
Build figure；
Under open loop situations, need to build figure equipment when receiving external absolute fix information, establish absolute fix information with
Constraint between current kinetic location information carries out the building of global map according to constraint information.
This specification embodiment provides a kind of open loop optimization system that figure is built based on SLAM, real by the following technical programs
It is existing:
The construction unit of local environment map, is configured as: obtaining locating local environment cartographic information, and is based on the office
Portion's Environmental Map Information carries out building figure；
The construction unit of global map, is configured as: when receiving external absolute fix information, establishing absolute fix letter
Constraint between breath and current kinetic location information carries out the building of global map according to constraint information.
This specification embodiment provides a kind of computer equipment, including memory, processor and storage are on a memory
And the computer program that can be run on a processor, which is characterized in that the processor realizes a kind of base when executing described program
The open loop optimization method step of figure is built in SLAM.
This specification embodiment provides a kind of computer readable storage medium, is stored thereon with computer program and (refers to
Enable), which is characterized in that the program (instruction) realizes a kind of open loop optimization method step that figure is built based on SLAM when being executed by processor
Suddenly.
This specification embodiment provides a kind of robot, and the robot includes robot body, the robot sheet
Sensor on body equipped with laser radar sensor and the external absolute fix information of acquisition；
The data of laser radar sensor detection and external absolute fix information are transmitted to operation processing unit, the fortune
Processing unit is calculated to be configured as:
The locating local environment cartographic information for needing to build figure equipment is obtained, and is carried out based on the local environment cartographic information
Map；
It needs to build figure equipment when receiving external absolute fix information, establishes absolute fix information and current kinetic position
Constraint between information carries out the building of global map according to constraint information.
Compared with prior art, the beneficial effect of the disclosure is:
Disclosed technique scheme builds nomography in mobile the improving on the basis of building chart-pattern of traditional single machine, using open loop
The method of optimization increases the robustness for building figure by the absolute fix information in reception environment.
Disclosed technique scheme can eliminate tiring out in mobile modeling mode by receiving the absolute fix information in environment
Product error, obtains more accurate navigation map under large scene environment.
Disclosed technique scheme by it is original build figure on the basis of increase discontinuous absolute location information, figure matter is built in guarantee
Amount does not influence to build the renewal frequency of figure while raising.
Disclosed technique scheme optimizes the addition for receiving absolute location information by open loop, ensure that the unification of coordinate system,
Help to solve the problems, such as more map references fusion when multirobot cooperates with map structuring.
Detailed description of the invention
The Figure of description for constituting a part of this disclosure is used to provide further understanding of the disclosure, and the disclosure is shown
Meaning property embodiment and its explanation do not constitute the improper restriction to the disclosure for explaining the disclosure.
Fig. 1 is that the sub- slam of the embodiment of the present disclosure builds drawing system hardware architecture diagram；
Fig. 2 is that the sub- slam of the embodiment of the present disclosure builds drawing system architecture diagram；
Fig. 3 is the map structuring work flow diagram that embodiment of the present disclosure subbase optimizes in open loop；
Fig. 4 is the map structuring algorithm flow chart that embodiment of the present disclosure subbase optimizes in open loop.
Specific embodiment
It is noted that following detailed description is all illustrative, it is intended to provide further instruction to the disclosure.Unless another
It indicates, all technical and scientific terms used herein has usual with disclosure person of an ordinary skill in the technical field
The identical meanings of understanding.
It should be noted that term used herein above is merely to describe specific embodiment, and be not intended to restricted root
According to the illustrative embodiments of the disclosure.As used herein, unless the context clearly indicates otherwise, otherwise singular
Also it is intended to include plural form, additionally, it should be understood that, when in the present specification using term "comprising" and/or " packet
Include " when, indicate existing characteristics, step, operation, device, component and/or their combination.
The patent No. " CN201810804313.0 " discloses " a kind of real-time SLAM scene map structuring system, navigation system
And method ", which builds figure by multirobot collaboration and uploads to cloud respectively, carries out whole map optimization.Disclosed technique
Scheme is to build figure using preset absolute location information to robot and carry out open loop optimization.The method can be used for optimizing list
Figure error is built under a robot large scene, the map after can also optimizing multiple robots directly splices, and realizes multimachine
Figure is built in device people collaboration.
Under large scene, if carrying out collaboration using multiple robots builds figure, addition that can be random under scene disperses
The absolute fix information two dimensional code of itself world coordinates (such as have), after two dimensional code is seen by each robot, according to machine
The Global localization information for including in the pose and two dimensional code of people's relative two dimensional code may know that each robot in world coordinates
Position, and robot builds during figure the absolute fix information for having two dimensional code to include and is modified.According to absolute fix information
The map optimized that each robot is built is stitched together, and is multirobot collaboration composition.This method need to only obtain scene
In sparse absolute fix information.
In common closed-loop optimization, just progress closed loop is full when the scene of process only before robot detects
Office's optimization, and there are certain errors.When the two dimensional code such as comprising absolute fix information is added during building figure, robot inspection
Global optimization can be carried out when measuring two dimensional code, need not wait for the scene re-optimization passed through before returning to.I.e. by external exhausted
Open loop optimization is carried out to information.
Examples of implementation one
This specification embodiment provides a kind of open loop optimization method that figure is built based on SLAM, real by the following technical programs
It is existing:
The locating local environment cartographic information for needing to build figure equipment is obtained, and is carried out based on the local environment cartographic information
Build figure；
It needs to build figure equipment when receiving external absolute fix information, establishes absolute fix information and current kinetic position
Constraint between information carries out the building of global map according to constraint information.
Referring to shown in attached drawing 3, for environmental information, the absolute posture information of robot is obtained using sensor, using swashing
Optical radar obtains laser data and carries out judging whether it is the absolute posture information of first frame first for absolute posture information, if
It is that global coordinate system is then determined based on the information, later using absolute pose as observation pose, otherwise, directly by absolute pose
Robot motion's pose is found out using Scanmatcher scan matching algorithm for laser data as observation pose, is utilized
The detection of BBS closed loop finds out observation pose.
Then judge whether absolute pose data occur, if so, the Weighted Constraint between addition scan and absolute pose,
If it is not, then adding the closed loop constraint between scan and submap, building residual equation carries out SPA optimization.
Overall step can be found in shown in attached drawing 4, each module electrifying startup of robot, runs ros operating system, acquires each section
Point sensor data carry out global map building, and open loop optimization is carried out when there is absolute pose.
In the specific implementation, the cartographic information of current location is got using 360 ° of laser radars, specific progress map
Building is referring to following scheme:
The laser data in the mobile platform acquisition environment of laser radar is carried, front end builds figure use and builds figure side end to end
Each frame laser data is mapped in sub- map by formula by way of scan-to-map (scanning element maps directly to map),
I.e. many adjacent scan scanning elements estimate pose by scan matching, are inserted into the sub- map of submap and form sub- map.It is logical
Overscanning matching, scan scanning element can be matched to a best estimation pose in sub- map.Due to the whole process time
Error also very little that is very short, therefore accumulating, and the motion profile of robot has been determined by scan matching.
After sub- map quantity increases, determined between the motion profile of robot and actual motion track by scan matching
Cumulative errors slowly become larger, it is therefore desirable to carry out rear end optimization.Each frame scan scanning frame not only with previous frame submap
Sub- map is matched, and also the sub- map of the submap of former frames is compared therewith, is established local winding, is carried out to local map
Optimization.
When conventional closed loop detection process is the position passed by before motion platform is returned to, by the scanning of present laser
Frame is compared with the map stored in the position before, finds the corresponding position in storage map by branch-bound algorithm
It sets.If it is ideal without deviation in the case where, current track pose and storage by scan-matcher scan matching
Corresponding pose is consistent in map.But due to the presence of kinematic error, the two is made to generate deviation.
Therefore, in the specific embodiment of the disclosure, by current laser data frame storage map in corresponding position
Appearance constructs residual equation as constraining as observation pose, and the deviation that motion pose generates, and it is excellent to carry out SPA to global map
Change, corrects error, the map after generating optimization.
Under open loop Optimal State, due to absolute fix information be it is discontinuous, when not detecting external absolute position
When confidence ceases, figure process is built according to above-mentioned, carries out the creation of front-end map and motion profile, it is excellent that rear end carries out global map closed loop
Change.
When receiving external absolute location information, judged to receive the category information whether for the first time first, such as
Fruit is then transformed into robot local Coordinate System in the coordinate system of external absolute position；If not then omitting the step.It
Robot is passed through to the absolute pose received afterwards as observation pose, the motion pose calculated when generating map with front end carries out
Weighted Constraint constructs residual equation, carries out the SPA optimization of global map, corrects error.As long as receiving absolute pose,
Construct the constraint between current kinetic pose and absolute observations pose；When not receiving absolute observations pose, carry out just
The creation of normal closed loop constraint.Open loop optimization process occurs during figure is built in front end, and when progress open loop optimization, no longer
Carry out the closed-loop optimization of rear end.
In a particular embodiment, if the posture information of scan scanning element is by ξ=(ξx,ξy,ξθ) indicate, including ξx,ξyTranslation
Component and ξθRotational component.In partial sweep, each frame scan scanning element is only matched with a part of map.Match party
Formula calculates the alignment relation of scan scanning element and the sub- map of submap by Nonlinear Least Squares Method.
Firstly, scan number of scan points evidence passes through Formula of Coordinate System Transformation TξIt indicates.Each scan number of scan points is according in scan
It is expressed as in coordinate systemK is the number of scanning element in a frame laser data, hkIt indicates
Scanpoint coordinate points under scan coordinate system.Each pose is expressed as ξ=(ξ in the sub- map coordinates system of submapx,ξy,
ξθ).Conversion regime is as follows:
P indicates each coordinate of scan scanning element in scan coordinate system, TξP indicates to convert scan coordinate system to
Coordinate in submap coordinate system.All scanning elements in each frame scan are converted by same set of ξ parameter, multiple adjacent
The continuous iteration of scan, formed submap.
Before scan is inserted into submap, being found by Ceres-scanmatch scan matching mode sweeps scan
The process of optimal solution is found in one position of described point maximum probability in the sub- map of submap.
Ceres solution is classified as solving a non-linear least square problem:
Wherein, hkIndicate scan point, T under scan coordinate systemξhkIndicate the point being transformed into submap coordinate system
Position.MsmoothIndicate the smoothed version of the grid probability value M after bicubic interpolation.K is to scan in a frame laser data
The number of point.
Formula essence are as follows: each of scan scanning element point is transformed into submap by the same conversion parameter,
Calculate each scan point corresponding M after bicubic interpolationsmoothOn probability, that is, find a pose pose make
∑MsmoothIt is maximum.
The form of Submap is substantially that the form of probabilistic grid stores.A series of similar scan form submap
Probabilistic grid
I.e. the probability of each mesh point is made of the probability of point closest around.When a needle scan scanning element quilt
When being inserted into probabilistic grid, hit indicates that the point in scan is transformed into immediate one in the pose pose in submap
A grid is occupied primary.Miss indicates that the point in scan is transformed into the pose pose in submap to scan scan data origin
Between grid on line do not have it is occupied.
If the unassigned mistake of Grid (hit or miss), p=0.5.If Grid had been observed, update
Formula is as follows:
Mnew(x)=clamp (odds-1(odds(Mold(x))·odds(p))) (4)。
In an examples of implementation, if not detecting absolute location information, conventional closed loop constraint is carried out:
In another examples of implementation, when the position that robot motion passes by before, by present laser scan data frame
The position to match is quickly found out by branch-bound algorithm with the map of storage, that is, finding an optimum position ξ makes at this
A frame scan scan data in all scan point scanning elements where the sum of the probability value of probabilistic grid maximum.
By the way that a best match position, i.e. observation position can be found after the algorithm.If it is ideal, do not have error
Under conditions of, which should be consistent with current motion pose.Due to the presence of deviation, so that observation pose and movement
Pose is inconsistent, constructs residual equation according to constraint.
Residual error E's is calculated as
Finally it is attributed to a non-linear least square problem:
Wherein,For sub- map posture information submap_pose,For scanning frame
Pose, i.e. scan_pose.And Σij,ξijIndicate constraint information, ΣijIndicate that covariance matrix, i.e. jth frame scan are inserted into i-th
Confidence level in frame submap, ξijIndicate that partial sweep matching and whole scan are matched as a result, i.e. present laser scanning frame
The constraint relationship and present laser scanning frame between scan and former frame submap by branch-bound algorithm find out on sub- ground
Matching position in figure.
By solving the non-linear least square problem, global error can be made to reach minimum under this constraint, optimization is complete
Local figure.
When sensor receives external absolute location information ξ for the first timeAbsolute=(ξx,ξy,ξθ) when, first absolutely according to this
Location information is coordinately transformed, i.e., is transformed to the map coordinates system of storage in absolute location coordinates system.Become by coordinate
Change matrix TξAbsolute,
Coordinate transform can be realized.P indicates each coordinate of scan scanning element in scan coordinate system, TξP is indicated
Coordinate scan coordinate system converted in world coordinate system
The absolute position for after receiving the second frame absolute posture information, being just no longer coordinately transformed, but this being observed
Appearance is compared as observation pose with motion pose.Ideally, the two should be identical, but due to the presence of error,
So that motion pose generates accumulated error.Therefore according to the constraint of motion pose and observation pose, the two is weighted, is constructed
Residual equation.Wherein since the confidence level of absolute location information is very high, observation bit need to be increased when constructing residual equation
The weight of appearance.
Equally it is attributed to a least square problem:
Wherein, due to absolute pose data be not it is continuous, exist when do not observe absolute posture information, therefore i,
J only indicates to observe a certain section of local map parameter of absolute posture information, i.e. i, j are to observe from a certain section of path absolutely
When to pose, terminate when being interrupted to absolute pose.For the i-th frame map pose submap_ of observation
Pose,For the pose of jth frame scan frame, i.e. scan_pose.And ΣAb_ij,ξAb_ijIndicate constraint information,
ΣAb_jIndicate covariance matrix, i.e. confidence level of the jth frame scan in the absolute pose coordinate system that jth frame observes, ξAb_jTable
Show the absolute posture information that jth frame scan is observed.
By solving the non-linear least square problem, external absolutely posture information can be fused to the local map of building
In, i.e., the constraint between laser scanning data scan and external absolutely posture information is established by open loop optimization.The method can
The error generated during repairing the front end slam creation local map.
It should be noted that the explanation about external absolutely posture information:
The sensor for obtaining external absolute fix information has very much, and if there is GPS in outdoor, there is UWB in interior or with absolute position
The two dimensional code etc. of confidence breath, embodiment of the present disclosure obtain the corresponding absolute position of two dimensional code by taking two dimensional code as an example, through camera
It sets.
In room conditions, absolute posture information can be provided by modes such as UWB or two dimensional codes.By taking two dimensional code as an example,
It needs to determine global absolute coordinate system indoors in advance when running open loop optimization, the arrangement of two dimensional code is carried out according to this coordinate system, often
A two dimensional code includes the posture information of its position.And two dimensional code need to be by certain Density Distribution, under the less environment of feature
Density is high, and density is low in the environment of feature rich.When observing two dimensional code using camera, the position that includes according to two dimensional code
It is corresponding absolute under two dimensional code coordinate system can to find out robot for relative pose between appearance information and robot and two dimensional code
Pose.
Under outdoor conditions, the sensor that such as GPS has absolute posture information can be used.Since GPS signal is being built
It builds the more place of object to have when do not receive, the absolute pose received can be made full use of to believe using open loop optimization
Breath.
This method is applicable not only to build figure under single machine device National People's Congress scene, is also applied for multirobot collaboration and builds figure.It is based on
Drawing method is built in robot open loop optimization described above, and in the case where the scene of figure is built in multirobot collaboration, each robot is from the overall situation
The sparse position of difference under scene starts to carry out building figure at random, carries out open loop optimization by absolute fix information, figure mistake is built in optimization
Difference, and global map coordinate is converted by the sub- map in part.It will be uploaded to same host by the local map of open loop optimization, directly
It connects splice by the local map that each robot is built and global map information can be obtained.
Examples of implementation two
This specification embodiment provides a kind of open loop optimization system that figure is built based on SLAM, real by the following technical programs
It is existing:
The construction unit of local environment map, is configured as: obtaining locating local environment cartographic information, and is based on the office
Portion's Environmental Map Information carries out building figure；
The construction unit of global map, is configured as: when receiving external absolute fix information, establishing absolute fix letter
Constraint between breath and current kinetic location information carries out the building of global map according to constraint information.
Above system can based on server realize, be configured in the server include local environment map construction unit and
The construction unit of global map, technology of the said units in specific embodiment are realized reference can be made to specific in examples of implementation one
Algorithm implementation procedure.
It should be noted that although being referred to several modules or submodule of equipment in the detailed description above, it is this
Division is only exemplary rather than enforceable.In fact, in accordance with an embodiment of the present disclosure, two or more above-described moulds
The feature and function of block can embody in a module.Conversely, the feature and function of an above-described module can be with
Further division is to be embodied by multiple modules.
Examples of implementation three
This specification embodiment provides a kind of computer equipment, including memory, processor and storage are on a memory
And the computer program that can be run on a processor, which is characterized in that the processor realizes a kind of base when executing described program
The open loop optimization method step of figure is built in SLAM.
This specification embodiment also provides a kind of computer readable storage medium, is stored thereon with computer program and (refers to
Enable), which is characterized in that the program (instruction) realizes a kind of open loop optimization method step that figure is built based on SLAM when being executed by processor
Suddenly.
In the present embodiment, computer program product may include computer readable storage medium, containing for holding
The computer-readable program instructions of row various aspects of the disclosure.Computer readable storage medium, which can be, can keep and store
By the tangible device for the instruction that instruction execution equipment uses.Computer readable storage medium for example can be-- but it is unlimited
In-- storage device electric, magnetic storage apparatus, light storage device, electric magnetic storage apparatus, semiconductor memory apparatus or above-mentioned
Any appropriate combination.
Example IV
This specification embodiment provides a kind of robot, and the robot includes robot body, the robot sheet
Sensor on body equipped with laser radar sensor and the external absolute fix information of acquisition；Robot body can be common machine
The mechanical equipment of device people.
The data of laser radar sensor detection and external absolute fix information are transmittedOperation processing unit, Neng Goujie
Receive building and amendment that sensor information carries out map.
The local environment cartographic information and absolute location information that laser radar detects, the received absolute position of observation sensor
Confidence breath is sent to after timestamp information is added to operation processing unit.
Since absolute fix information is not necessarily continuously, when not receiving absolute fix information, operation processing unit
The local environment cartographic information for only receiving laser carries out front-end map building, the constraint of rear end closed loop.
When receiving absolute location information, the pact between absolute location information and current kinetic location information is established immediately
Beam carries out the building of global map according to constraint information.
Specific hardware block diagram can be found in shown in attached drawing 1, and software architecture is referring to shown in attached drawing 2, in fig. 1, at operation
Reason unit is arm processor, which is connected by network with user terminal, and user terminal can pass through user interface realization pair
The operation of robot controls, which is also connected with sensor, for example, IMU, laser radar, GPS, UWB and camera etc., biography
Sensor carries out the acquisition of data as basic data acquisition unit and is transmitted to operation processing unit, and operation processing unit is also logical
It crosses interface layer to communicate with embedded processing unit, the driving to motor is realized by motor driven using embedded processing unit,
The embedded processing unit is also connected with safety sensor and Boolean value output sensor, receives corresponding data.
In the realization of software, referring to shown in attached drawing 2, the main program of mobile platform bottom controls upper layer decision-making process,
Handle node, laser data, absolute pose data provide data supporting for upper layer decision-making process, and encoder data is mobile platform
The main program of bottom provides data supporting, and upper layer decision-making process exports corresponding cartographic information according to the input data.
In the examples of implementation, above-mentioned robot is intended only as the form of expression of mobile platform, in specific implementation, can be with
For the form of other mobile platform, for example, unmanned vehicle etc..
It is understood that in the description of this specification, reference term " embodiment ", " another embodiment ", " other
The description of embodiment " or " first embodiment~N embodiment " etc. means specific spy described in conjunction with this embodiment or example
Sign, structure, material or feature are contained at least one embodiment or example of the disclosure.In the present specification, to above-mentioned
The schematic representation of term may not refer to the same embodiment or example.Moreover, the specific features of description, structure, material
Person's feature can be combined in any suitable manner in any one or more of the embodiments or examples.
The foregoing is merely preferred embodiment of the present disclosure, are not limited to the disclosure, for the skill of this field
For art personnel, the disclosure can have various modifications and variations.It is all within the spirit and principle of the disclosure, it is made any to repair
Change, equivalent replacement, improvement etc., should be included within the protection scope of the disclosure.
Claims (10)
1. a kind of open loop optimization method for building figure based on SLAM, characterized in that include:
The locating local environment cartographic information for needing to build figure equipment is obtained, and carries out ground based on the local environment cartographic information
Figure；
It under open loop situations, needs to build figure equipment when receiving external absolute fix information, establishes absolute fix information and current
Constraint between movement position information carries out the building of global map according to constraint information.
2. a kind of open loop optimization method for building figure based on SLAM as described in claim 1, characterized in that need to build figure equipment and exist
It when receiving external absolute fix information, carries out judging whether it is the absolute posture information of first frame first, if so, being based on the letter
It ceases and determines global coordinate system, later using absolute pose as observation pose, otherwise, directly using absolute pose as observation pose.
3. a kind of open loop optimization method for building figure based on SLAM as described in claim 1, characterized in that acquisition needs to build figure and sets
When standby locating local environment cartographic information, the laser data in environment is obtained using laser radar, by each frame laser number
It is mapped in sub- map according to by scan-to-map.
4. a kind of open loop optimization method for building figure based on SLAM as claimed in claim 3, characterized in that increase in sub- map quantity
After mostly, the cumulative errors between the motion profile and actual motion track of the robot determined by scan matching become larger, and carry out
Rear end optimization, each frame scan are not matched with previous frame submap only, and also the submap of former frames is compared therewith,
Local winding is established, local map is optimized.
5. a kind of open loop optimization method for building figure based on SLAM as claimed in claim 3, characterized in that outer when not detecting
When the absolute location information in portion, using current laser data frame, corresponding pose is as observation pose in storage map, with fortune
The deviation that dynamic pose generates constructs residual equation as constraint, carries out SPA optimization to global map, corrects error, generates optimization
Map afterwards.
6. a kind of open loop optimization method for building figure based on SLAM as claimed in claim 3, characterized in that according to motion pose and
The constraint for observing pose, is weighted the two, constructs residual equation, is solved using nonlinear least square method, and optimization is global
Map.
7. a kind of open loop optimization system for building figure based on SLAM, characterized in that include:
The construction unit of local environment map, is configured as: obtaining locating local environment cartographic information, and is based on the local ring
Border cartographic information carries out building figure；
The construction unit of global map, is configured as: when receiving external absolute fix information, establish absolute fix information with
Constraint between current kinetic location information carries out the building of global map according to constraint information.
8. a kind of computer equipment, characterized in that on a memory and can be on a processor including memory, processor and storage
The computer program of operation, which is characterized in that the processor realizes that claim 1-6 is any described when executing described program
A kind of open loop optimization method step for building figure based on SLAM.
9. a kind of computer readable storage medium, is stored thereon with computer program, which is characterized in that the program is held by processor
A kind of claim 1-6 any open loop optimization method step that figure is built based on SLAM is realized when row.
10. a kind of robot, the robot includes robot body, is sensed on the robot body equipped with laser radar
Device and the sensor for obtaining external absolute fix information；
The data of laser radar sensor detection and external absolute fix information are transmittedOperation processing unit, feature exist
In the operation processing unit is configured as:
The locating local environment cartographic information for needing to build figure equipment is obtained, and carries out ground based on the local environment cartographic information
Figure；
It needs to build figure equipment when receiving external absolute fix information, establishes absolute fix information and current kinetic location information
Between constraint, according to constraint information carry out global map building.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910172096.2A CN109813319B (en) | 2019-03-07 | 2019-03-07 | Open loop optimization method and system based on SLAM (Simultaneous localization and mapping) mapping |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910172096.2A CN109813319B (en) | 2019-03-07 | 2019-03-07 | Open loop optimization method and system based on SLAM (Simultaneous localization and mapping) mapping |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109813319A true CN109813319A (en) | 2019-05-28 |
| CN109813319B CN109813319B (en) | 2021-09-28 |
Family
ID=66608359
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910172096.2A Active CN109813319B (en) | 2019-03-07 | 2019-03-07 | Open loop optimization method and system based on SLAM (Simultaneous localization and mapping) mapping |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109813319B (en) |
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110726413A (en) * | 2019-10-25 | 2020-01-24 | 中国人民解放军国防科技大学 | Multi-sensor fusion and data management mechanism facing large-scale SLAM |
| CN110749901A (en) * | 2019-10-12 | 2020-02-04 | 劢微机器人科技(深圳)有限公司 | Autonomous mobile robot, map splicing method and device thereof, and readable storage medium |
| CN111551953A (en) * | 2020-05-06 | 2020-08-18 | 天津博诺智创机器人技术有限公司 | Indoor map construction optimization method based on SLAM algorithm |
| CN112097774A (en) * | 2020-09-16 | 2020-12-18 | 东北大学秦皇岛分校 | Distributed map fusion method based on adaptive Kalman filtering and average tracking |
| CN112161635A (en) * | 2020-09-22 | 2021-01-01 | 中山大学 | Cooperative graph building method based on minimum loop detection |
| CN112595320A (en) * | 2020-11-23 | 2021-04-02 | 北京联合大学 | ROS-based high-precision positioning autonomous navigation method and system for indoor intelligent wheelchair |
| CN112923919A (en) * | 2021-01-21 | 2021-06-08 | 湖南格纳微信息科技有限公司 | Pedestrian positioning method and system based on graph optimization |
| CN112985416A (en) * | 2021-04-19 | 2021-06-18 | 湖南大学 | Robust positioning and mapping method and system based on laser and visual information fusion |
| CN113192153A (en) * | 2021-05-26 | 2021-07-30 | 清华大学 | Multi-agent cooperative map construction method and device based on sub-map |
| CN113252051A (en) * | 2020-02-11 | 2021-08-13 | 北京图森智途科技有限公司 | Map construction method and device |
| CN113297259A (en) * | 2021-05-31 | 2021-08-24 | 深圳市优必选科技股份有限公司 | Robot and environment map construction method and device thereof |
| CN115423965A (en) * | 2022-11-04 | 2022-12-02 | 安徽蔚来智驾科技有限公司 | Map construction method, map construction apparatus, vehicle, and storage medium |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106643720A (en) * | 2016-09-28 | 2017-05-10 | 深圳市普渡科技有限公司 | Method for map construction based on UWB indoor locating technology and laser radar |
| CN106840148A (en) * | 2017-01-24 | 2017-06-13 | 东南大学 | Wearable positioning and path guide method based on binocular camera under outdoor work environment |
| JP2017162457A (en) * | 2016-03-11 | 2017-09-14 | 株式会社東芝 | Image analysis system and method |
| CN107515891A (en) * | 2017-07-06 | 2017-12-26 | 杭州南江机器人股份有限公司 | A kind of robot cartography method, apparatus and storage medium |
| CN107727104A (en) * | 2017-08-16 | 2018-02-23 | 北京极智嘉科技有限公司 | Positioning and map building air navigation aid, apparatus and system while with reference to mark |
| US20180120116A1 (en) * | 2016-11-01 | 2018-05-03 | Brain Corporation | Systems and methods for robotic mapping |
| CN108303710A (en) * | 2018-06-12 | 2018-07-20 | 江苏中科院智能科学技术应用研究院 | Drawing method is built in the more scene positioning of unmanned plane based on three-dimensional laser radar |
| CN108520543A (en) * | 2018-04-09 | 2018-09-11 | 网易（杭州）网络有限公司 | A kind of method that relative accuracy map is optimized, equipment and storage medium |
| CN109357676A (en) * | 2018-10-19 | 2019-02-19 | 北京三快在线科技有限公司 | The localization method and device and mobile device of a kind of mobile device |
| CN109407677A (en) * | 2018-12-24 | 2019-03-01 | 清华大学 | The trace tracking method of automatic driving vehicle |
-
2019
- 2019-03-07 CN CN201910172096.2A patent/CN109813319B/en active Active
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2017162457A (en) * | 2016-03-11 | 2017-09-14 | 株式会社東芝 | Image analysis system and method |
| CN106643720A (en) * | 2016-09-28 | 2017-05-10 | 深圳市普渡科技有限公司 | Method for map construction based on UWB indoor locating technology and laser radar |
| US20180120116A1 (en) * | 2016-11-01 | 2018-05-03 | Brain Corporation | Systems and methods for robotic mapping |
| CN106840148A (en) * | 2017-01-24 | 2017-06-13 | 东南大学 | Wearable positioning and path guide method based on binocular camera under outdoor work environment |
| CN107515891A (en) * | 2017-07-06 | 2017-12-26 | 杭州南江机器人股份有限公司 | A kind of robot cartography method, apparatus and storage medium |
| CN107727104A (en) * | 2017-08-16 | 2018-02-23 | 北京极智嘉科技有限公司 | Positioning and map building air navigation aid, apparatus and system while with reference to mark |
| CN108520543A (en) * | 2018-04-09 | 2018-09-11 | 网易（杭州）网络有限公司 | A kind of method that relative accuracy map is optimized, equipment and storage medium |
| CN108303710A (en) * | 2018-06-12 | 2018-07-20 | 江苏中科院智能科学技术应用研究院 | Drawing method is built in the more scene positioning of unmanned plane based on three-dimensional laser radar |
| CN109357676A (en) * | 2018-10-19 | 2019-02-19 | 北京三快在线科技有限公司 | The localization method and device and mobile device of a kind of mobile device |
| CN109407677A (en) * | 2018-12-24 | 2019-03-01 | 清华大学 | The trace tracking method of automatic driving vehicle |
Non-Patent Citations (1)
| Title |
|---|
| WOLFGANG HESS等: "Real-Time Loop Closure in 2D LIDAR SLAM", 《2016 IEEE INTERNATIONAL CONFERENCE ON ROBOTICS AND AUTOMATION》 * |
Cited By (17)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110749901A (en) * | 2019-10-12 | 2020-02-04 | 劢微机器人科技(深圳)有限公司 | Autonomous mobile robot, map splicing method and device thereof, and readable storage medium |
| CN110749901B (en) * | 2019-10-12 | 2022-03-18 | 劢微机器人科技(深圳)有限公司 | Autonomous mobile robot, map splicing method and device thereof, and readable storage medium |
| CN110726413A (en) * | 2019-10-25 | 2020-01-24 | 中国人民解放军国防科技大学 | Multi-sensor fusion and data management mechanism facing large-scale SLAM |
| CN113252051A (en) * | 2020-02-11 | 2021-08-13 | 北京图森智途科技有限公司 | Map construction method and device |
| CN111551953A (en) * | 2020-05-06 | 2020-08-18 | 天津博诺智创机器人技术有限公司 | Indoor map construction optimization method based on SLAM algorithm |
| CN112097774A (en) * | 2020-09-16 | 2020-12-18 | 东北大学秦皇岛分校 | Distributed map fusion method based on adaptive Kalman filtering and average tracking |
| CN112097774B (en) * | 2020-09-16 | 2022-08-23 | 东北大学秦皇岛分校 | Distributed map fusion method based on adaptive Kalman filtering and average tracking |
| CN112161635A (en) * | 2020-09-22 | 2021-01-01 | 中山大学 | Cooperative graph building method based on minimum loop detection |
| CN112161635B (en) * | 2020-09-22 | 2022-07-05 | 中山大学 | Cooperative graph building method based on minimum loop detection |
| CN112595320A (en) * | 2020-11-23 | 2021-04-02 | 北京联合大学 | ROS-based high-precision positioning autonomous navigation method and system for indoor intelligent wheelchair |
| CN112923919A (en) * | 2021-01-21 | 2021-06-08 | 湖南格纳微信息科技有限公司 | Pedestrian positioning method and system based on graph optimization |
| CN112985416B (en) * | 2021-04-19 | 2021-07-30 | 湖南大学 | Robust positioning and mapping method and system based on laser and visual information fusion |
| CN112985416A (en) * | 2021-04-19 | 2021-06-18 | 湖南大学 | Robust positioning and mapping method and system based on laser and visual information fusion |
| CN113192153A (en) * | 2021-05-26 | 2021-07-30 | 清华大学 | Multi-agent cooperative map construction method and device based on sub-map |
| CN113297259A (en) * | 2021-05-31 | 2021-08-24 | 深圳市优必选科技股份有限公司 | Robot and environment map construction method and device thereof |
| CN115423965A (en) * | 2022-11-04 | 2022-12-02 | 安徽蔚来智驾科技有限公司 | Map construction method, map construction apparatus, vehicle, and storage medium |
| CN115423965B (en) * | 2022-11-04 | 2023-02-28 | 安徽蔚来智驾科技有限公司 | Map construction method, map construction apparatus, vehicle, and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109813319B (en) | 2021-09-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109813319A (en) | A kind of open loop optimization method and system for building figure based on SLAM | |
| US11328158B2 (en) | Visual-inertial positional awareness for autonomous and non-autonomous tracking | |
| CN108303099B (en) | Autonomous navigation method in unmanned plane room based on 3D vision SLAM | |
| CN104236548B (en) | Autonomous navigation method in a kind of MAV room | |
| CN108827306A (en) | A kind of unmanned plane SLAM navigation methods and systems based on Multi-sensor Fusion | |
| CN106840148B (en) | Wearable positioning and path guiding method based on binocular camera under outdoor working environment | |
| CN109509230A (en) | A kind of SLAM method applied to more camera lens combined type panorama cameras | |
| WO2019119289A1 (en) | Positioning method and device, electronic apparatus, and computer program product | |
| CN107544501A (en) | A kind of intelligent robot wisdom traveling control system and its method | |
| CN110118556A (en) | A kind of robot localization method and device based on covariance mixing together SLAM | |
| CN106168805A (en) | The method of robot autonomous walking based on cloud computing | |
| CN105411490A (en) | Real-time positioning method of mobile robot and mobile robot | |
| CN110211228A (en) | For building the data processing method and device of figure | |
| CN103926933A (en) | Indoor simultaneous locating and environment modeling method for unmanned aerial vehicle | |
| CN111077907A (en) | Autonomous positioning method of outdoor unmanned aerial vehicle | |
| CN112518739A (en) | Intelligent self-navigation method for reconnaissance of tracked chassis robot | |
| CN113325837A (en) | Control system and method for multi-information fusion acquisition robot | |
| CN107063242A (en) | Have the positioning navigation device and robot of virtual wall function | |
| An et al. | Development of mobile robot SLAM based on ROS | |
| CN110260866A (en) | A kind of robot localization and barrier-avoiding method of view-based access control model sensor | |
| CN113238554A (en) | Indoor navigation method and system based on SLAM technology integrating laser and vision | |
| CN112652001B (en) | Underwater robot multi-sensor fusion positioning system based on extended Kalman filtering | |
| CN113763548A (en) | Poor texture tunnel modeling method and system based on vision-laser radar coupling | |
| CN112068152A (en) | Method and system for simultaneous 2D localization and 2D map creation using a 3D scanner | |
| Karam et al. | Integrating a low-cost mems imu into a laser-based slam for indoor mobile mapping |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |