CN109000649A - A kind of all directionally movable robot pose calibration method based on right angle bend feature - Google Patents
A kind of all directionally movable robot pose calibration method based on right angle bend feature Download PDFInfo
- Publication number
- CN109000649A CN109000649A CN201810529469.2A CN201810529469A CN109000649A CN 109000649 A CN109000649 A CN 109000649A CN 201810529469 A CN201810529469 A CN 201810529469A CN 109000649 A CN109000649 A CN 109000649A
- Authority
- CN
- China
- Prior art keywords
- line segment
- quarter bend
- coordinate system
- feature
- abstract
- 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
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
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
- G01C25/00—Manufacturing, calibrating, cleaning, or repairing instruments or devices referred to in the other groups of this subclass
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S17/00—Systems using the reflection or reradiation of electromagnetic waves other than radio waves, e.g. lidar systems
- G01S17/88—Lidar systems specially adapted for specific applications
Abstract
The invention discloses all directionally movable robot pose calibration methods based on right angle bend feature, method includes the following steps: S1. establishes quarter bend model；S2. quarter bend model is standardized, then establishes the right angle crook overall situation with reference to point set；S3. trolley original state coordinate system is established, and calculates preliminary pose estimation of the AGV trolley in global coordinate system；S4. quarter bend feature is identified；S5. Global localization is carried out to AGV trolley based on quarter bend feature and pose updates.The present invention is in the case where abundant analysis typical warehouse aisles environment, it is described using environmental characteristic of the laser radar data to right angle corner, and carry out Global localization and pose update under channel environment to comprehensive trolley with this feature, the accumulated error in relative positioning is eliminated, the whole positioning accuracy of robot is improved.
Description
Technical field
The invention belongs to Mobile Robotics Navigation field of locating technology, specially a kind of full side based on right angle bend feature
Variable mobile robot pose calibration method.
Background technique
All directionally movable robot namely AGV, with its high the degree of automation, that flexibility is good, space utilization rate is high etc. is many
Advantage is widely used in logistic industry and other industry rapidly, and becomes manufacturing industry system and automated warehouse storage system
Important logistics delivery equipment.
The pose of intelligent vehicle is estimated, referred to as positions, and is indispensable important link during ACV is automatically controlled and navigated.
In general environment, the pose of mobile robot is indicated usually using three-dimensional vector, i.e. the cross relative to world coordinates position
To, longitudinal translation component and the rotation angle component for representing its direction.Accurate pose estimation is for the automatic of mobile robot
Map generates, path planning and control, target detection and tracking etc. are of great significance.
Common location technology mainly has relative positioning based on telemetry and inertial navigation and GPS positioning, road sign fixed
The absolute fixs technology such as position and map match.There is the contradiction between technical costs and precision more in existing localization method, with position of navigating
Relative positioning technical method based on reckoning is simple, technical costs is low, but there are problems that accumulated error；Absolute fix technology does not have
Accumulated error, but such technology there are technologies it is complicated, method transplantability is poor, technical costs is high the problems such as.
Since individual relative positioning and absolute fix technology all have certain shortcoming, wanted in some positioning accuracies
It asks in high complex environment, positioning result is often not ideal enough.In practical application, the mode mostly combined in two ways, with
Absolute posture information updates the calibration that relative pose information carries out, and eliminates positioning accumulated error, estimates to improve whole pose
The precision of calculation.
In existing pose calibration method, technology is more mature mainly to be had based on GPS data fusion, navigation beacon and mark
Label, Landmarks etc., but GPS data is chiefly used in outdoor environment, indoor signal is vulnerable to blocking interference, and positioning accuracy is mostly with rice
For magnitude, indoor accurate position precision is inadequate；Localization method positioning accuracy based on beacon and label is high, and real-time is good, but should
Class localization method needs place the apparent artificial landmark of big measure feature or label within the scope of sensors sense environmental, workload compared with
Greatly, road sign is easy to damage in the scene of the environment complexity such as storage room, and cost of equipment maintenance is high；Secondly, the position based on map match
It is more that appearance calibration method was studied under environment indoors also compare, and has laser radar SLAM, vision SLAM technology etc. than more typical,
But such algorithm complexity is higher, computationally intensive, and under the higher environment of this similarity of warehouse aisles, Map recognition
Error is also easy to produce with matched process；In addition, in some pose calibration methods based on environmental characteristic, it is more representational
There is the technology based on beam like features and indoor right angle feature, such method flexibility is high, but the former is in spacious indoor environment
Effect is preferable, undesirable in the occasion effect of some environment complexity；And in the method based on indoor right angle feature, although in structure
Changing has good applicability under obvious indoor environment, but such method during extracting angle point information vulnerable to sensing
The influence of device noise or environmental disturbances factor, precision susceptible.
Therefore, it is badly in need of one kind and is suitable for omni-directional mobile robots, under automated warehousing channel environment, applicability is good, cuts
Real feasible pose calibration method.
Summary of the invention
In view of this, to solve the above-mentioned problems, the present invention provides a kind of all-around mobile based on right angle bend feature
Robot pose calibration method, this method utilize laser radar data in the case where abundant analysis typical warehouse aisles environment
The environmental characteristic of right angle corner is described, and Global localization is carried out to comprehensive trolley under channel environment with this feature
It is updated with pose, eliminates the accumulated error in relative positioning, improve the whole positioning accuracy of robot.
The purpose of the present invention is achieved through the following technical solutions: provided by the invention a kind of special based on right angle bend
The all directionally movable robot pose calibration method of sign, method includes the following steps:
S1. quarter bend model, including L-type quarter bend model, I type quarter bend model, T-type quarter bend model are established；
Wherein, the L-type quarter bend model includes by two lines of the abstract perpendicular intersection fitted of two sides metope
Section and one is by abstract the first line segment being fitted of package stacking, by two lines of the abstract perpendicular intersection fitted of two sides metope
A wherein line segment for section and the first line segment horizontal line by the abstract fitting of package stacking；
The T-type quarter bend model includes being abstracted by the abstract second line segment fitted of a face metope and by two package stackings
Two line segments in parastate of fitting are abstracted two line segments being fitted in parastate by two package stackings and by a face metope
The abstract second line segment fitted is vertical but non-intersecting；
The I type quarter bend model includes the third line segment taken out by a face metope as fitting and is abstracted by two package stackings
Two line segments in parastate of fitting take out the third line segment as fitting by a face metope and are abstracted fitting with by two package stackings
Two line segments in parastate it is vertical and intersect；Or the I type quarter bend model includes being taken out by a package stacking as fitting
The 4th line segment and two line segments in parastate by the abstract fitting of two metopes, the 4th that picture fits is taken out by a package stacking
Line segment is vertical with two line segments in parastate by the abstract fitting of two metopes and intersects；
S2. quarter bend model is standardized, then establishes the right angle crook overall situation with reference to point set；
S3. trolley original state coordinate system is established, and calculates preliminary pose estimation of the AGV trolley in global coordinate system；
S4. quarter bend feature is identified；
S5. Global localization is carried out to AGV trolley based on quarter bend feature and pose updates.
Preferably, the L-type quarter bend model are as follows:
The T-type quarter bend model are as follows:
The I type quarter bend model are as follows:
In two-dimensional Cartesian coordinate system lower line segment Li=(pi,θi, start, end, len), wherein start_q and end_q points
Do not indicate small front side wall matching line segment rise, beginning extreme coordinates, start_ (q+1) trolley left channel inner wall matching line segment
Starting point coordinate, end_ (q-1) indicate trolley right channel inner wall matching line segment initial point coordinate, piIt indicates from origin to straight line
LiDistance, θiIndicated origin and LiVertical line and positive direction of the x-axis angle, start and end respectively indicate line segment LiRise
Beginning extreme coordinates, len indicate line segment LiLength, δdminAnd δdmaxIndicate the minimum threshold and maximum of Euclidean distance between two o'clock
Threshold value, δdmaxFor judging whether adjacent two lines section head and the tail intersect, δdminFor judging the distance between adjacent two lines section is the first
Whether certain threshold value is greater than.
Preferably, in the step S2, to the standardized method of the L-type quarter bend aspect of model are as follows:
It will be fitted by two line segments of the abstract perpendicular intersection fitted of two sides metope and by two sides metope is abstract
Perpendicular intersection two line segments in a line segment as fixed character；
To T-type quarter bend aspect of model standardized method are as follows:
It is found out respectively first by the two line segment extended lines and second line segment in parastate of the abstract fitting of two package stackings
Intersecting point coordinate；
Then the arithmetic mean of instantaneous value for seeking two intersecting point coordinates, using the arithmetic mean of instantaneous value as key point Key_P；
And using key point Key_P and second line segment as the key feature of positioning；
To the standardized method of the I type quarter bend aspect of model are as follows:
Find out the friendship prolonged by two line segments in parastate of the abstract fitting of two package stackings with third line segment respectively first
Point coordinate；
Then the arithmetic mean of instantaneous value for seeking two intersecting point coordinates, using the arithmetic mean of instantaneous value as key point Key_P'；
And using key point Key_P' and third line segment as the key feature of positioning；
Or to the standardized method of the I type quarter bend aspect of model are as follows:
It finds out to be prolonged by two line segments in parastate of the abstract fitting of two metopes respectively first and be sat with the intersection point of the 4th line segment
Mark；
Then the arithmetic mean of instantaneous value for seeking two intersecting point coordinates, using the arithmetic mean of instantaneous value as key point Key_P "；
And using key point Key_P " and the 4th line segment as the key feature of positioning.
Preferably, in the step S2, when establishing global coordinate system；Point with the abstract fitting of warehouse entrance is original
Point, using the direction of the line segment perpendicular to the abstract fitting of package stacking as y-axis, to be parallel to the abstract line segment being fitted of package stacking
Direction is that x-axis establishes global coordinate system xoy；And the global reference point of right angle crook positioning key point is established under global coordinate system
Collect P={ P1,P2,...,Pn, wherein Pi=(xi,yi,θri), i ∈ [1, n], (xi,yi) indicate quarter bend two after standardization
Coordinate of the intersecting straight lines intersection point at global coordinate system xoy, θriIt indicates with the straight line and global coordinate system xoy of the abstract fitting of metope
X-axis angle.
Preferably, the step S3 specifically includes following sub-step:
S31. the vehicle body coordinate system x ' o ' y ' that the original state of trolley is established using the geometric center of AGV trolley as origin, builds
Vertical all-around mobile moving of car model:
Wherein, (w1,w2,w3,w4) indicate AGV trolley four wheels revolving speed, r is radius of wheel, and m, n are respectively indicated
The half of body width, length, (vx,vy, w) and indicate mass center under vehicle body initial coordinate system respectively along x-axis, y-axis, headstock direction
Movement speed；
S32. it is calculated according to dead reckoning principle at t=kT moment, at vehicle body coordinate system x ' o ' y ', counting of carriers mass center
Coordinate (x 'k,y′k,θ′k), θk' indicate trolley headstock direction and trolley geocentric coordinate system y positive direction angle, T indicates boat position
Calculate the sampling period, k indicates positive integer；
S33. according to the vehicle body coordinate system x ' o ' y ' conversion relational expression counting of carriers of global coordinate system xoy and trolley in the overall situation
Center-of-mass coordinate (x under coordinate system xoyk,yk,θk), θkIndicate the angle of the y positive direction of trolley headstock direction and global coordinate system.
Preferably, the step S4 specifically includes following sub-step:
S41. firstly, manual remote control AGV trolley carries out the study in preliminary path in channel environment, it is preliminary to generate quarter bend
The global path key point map of identification；During AGV trolley carries out subsequent independent navigation, according to priori map, use
Whether posture tracking method real time discriminating AGV trolley enters the neighborhood model for the global path key point map that quarter bend tentatively identifies
In enclosing, contiguous range is defined as centered on key point, radius r1Border circular areas；
S42. the line segment feature based on laser radar extracts, using split_merge algorithm to laser radar scanning data
Line segment feature is carried out to extract to obtain line segment aggregate L'；The rejecting for carry out invalid data to the line segment extracted in environment and line segment are again
Merging obtains new line segment aggregate L "；
S43. Model checking is carried out to the line segment aggregate L " near quarter bend in a frame laser radar using quarter bend model
And classify to it.
Preferably, the step S5 specifically includes following sub-step:
S51. reference coordinate is associated with；After AGV trolley completes the identification of quarter bend feature, need the small parking stall AGV at this time
The place's of setting quarter bend feature is associated with preset global reference coordinate point and matches, and obtains correct associated reference point；
S52. the AGV pose based on line segment feature resolves, and utilizes the standardization right angle extracted in laser radar scanning data
Curved model line segment and key feature carry out pose to laser radar and resolve to obtain the pose of laser radar under global coordinate system xoy
o″(xo,yo,θo)；
S53.AGV pose updates, and finds out AGV trolley in world coordinates by laser radar and the conversion of AGV trolley coordinate system
It is the position coordinates and attitude angle (x under xoyr,yr,θr), using the Global localization result of quarter bend to the relatively fixed of AGV trolley
Position result is updated.
By adopting the above-described technical solution, the present invention has the advantage that:
The present invention is in the case where abundant analysis typical warehouse aisles environment, using laser radar data to right angle corner
Environmental characteristic be described, and Global localization is carried out with this feature under channel environment to comprehensive trolley and pose updates,
The accumulated error in relative positioning is eliminated, the whole positioning accuracy of robot is improved.
Other advantages, target and feature of the invention will be illustrated in the following description to a certain extent, and
And to a certain extent, based on will be apparent to those skilled in the art to investigating hereafter, Huo Zheke
To be instructed from the practice of the present invention.Target of the invention and other advantages can be realized by following specification and
It obtains.
Detailed description of the invention
To make the objectives, technical solutions, and advantages of the present invention clearer, below in conjunction with attached drawing to the present invention make into
The detailed description of one step:
Fig. 1 is typical automated warehouse applications environment；
Fig. 2 is the curved schematic diagram of L-type；
Fig. 3 is the curved schematic diagram of T-type；
Fig. 4 is the curved schematic diagram of I type；
Fig. 5 is the curved standardization of L-type；
Fig. 6 is the curved standardization of T-type；
Fig. 7 is the curved standardization of I type；
Fig. 8 is reference coordinate point set schematic diagram；
Fig. 9 is all-around mobile trolley vehicle body coordinate system；
Figure 10 is the quarter bend Primary Location schematic diagram differentiated based on key point；
Figure 11 is that IEPF algorithm divides principle；
Figure 12 is the channel line segment schematic diagram that split_merge algorithm extracts；
Figure 13 is the AGV positioning schematic based on line segment feature；
Figure 14 is the flow chart of the method for the present invention.
Specific embodiment
Illustrate embodiments of the present invention below by way of specific specific example, those skilled in the art can be by this specification
Other advantages and efficacy of the present invention can be easily understood for disclosed content.The present invention can also pass through in addition different specific realities
The mode of applying is embodied or practiced, the various details in this specification can also based on different viewpoints and application, without departing from
Various modifications or alterations are carried out under spirit of the invention.
Fig. 1 is please referred to Figure 14.It should be noted that diagram provided in the present embodiment only illustrates this in a schematic way
The basic conception of invention, only shown in schema then with related component in the present invention rather than package count when according to actual implementation
Mesh, shape and size are drawn, when actual implementation kenel, quantity and the ratio of each component can arbitrarily change for one kind, and its
Assembly layout kenel may also be increasingly complex.
The all directionally movable robot pose calibration method based on right angle bend feature of the present embodiment, including walk as follows
It is rapid:
Step 1: quarter bend model is established.
Warehouse aisles environment is mostly made of stereo storage rack and metope, due to the generally relatively more neat rule of the disposing way of package
Then, it is formed in channel interior than more typical structuring geometrical characteristic, at channel turning, is formed obvious by line segment group
At quarter bend feature, as shown in Figure 1.
By environmental analysis it is found that exemplary right angle present in warehouse aisles environment is curved mainly following a few classes:
Step 11: L-type quarter bend model is established
Fig. 2 show L-type quarter bend schematic diagram, and the channel under the scene is by two sides metope and one side package stacking
Composition, typical L-type quarter bend is mainly by three line segment (L1,L2,L3) be composed, L-type quarter bend model includes by two face walls
Two line segments of the abstract perpendicular intersection fitted in face and one are by abstract the first line segment being fitted of package stacking, by two face walls
A wherein line segment for two line segments of the abstract perpendicular intersection fitted in face and the First Line by the abstract fitting of package stacking
Section horizontal line.In two-dimensional Cartesian coordinate system lower line segment Li=(pi,θi, start, end, len), wherein piIt indicates from origin to straight line Li
Distance, θiIndicated origin and LiVertical line and positive direction of the x-axis angle, start and end respectively indicate line segment LiStarting
Extreme coordinates, len indicate line segment LiLength, based on combination of channels characteristic in the above line segment parameter and actual environment, the scene
Under L-type quarter bend model description are as follows:
In above formula, formula (1) is the rough description constituted to L-type quarter bend model silhouette, and formula (2) is to meeting rough profile
Model make further accurate description, be mainly used for judge in L-type quarter bend model two sides metope be abstracted fit it is perpendicular
The right angle feature that two line segments of intersection are constituted；Num_L indicates the line extracted from one frame scan data of laser radar
The total number of section, the shortest length of len_min expression group l-shaped curved line section, ε are indicated for angle between judging two lines section most
Small neighbourhood, δdminAnd δdmaxIndicate the minimum value and max-thresholds of Euclidean distance between two o'clock, δdmaxFor judging adjacent two lines
Whether section head and the tail intersect, δdminFor judging whether the distance between adjacent two lines section is the first is greater than certain threshold value, in formula (2) " |
| " indicate mathematical operator "or".
Step 12: T-type quarter bend model is established
Fig. 3 show T-type quarter bend model schematic, and the channel under the scene is by a face metope and two sides package stacking group
At.Bend abstract model under the scene is also made of three line segments, and T-type quarter bend model includes by the abstract fitting of a face metope
Second line segment out and two line segments in parastate by the abstract fitting of two package stackings, by the abstract fitting of two package stackings
It is vertical but non-intersecting with by the abstract second line segment fitted of a face metope in two line segments of parastate.In conjunction with line segment parameter and
Combination of channels characteristic in actual environment, the model description of T-type right angle bend under the scene are as follows:
Wherein, formula 4 mainly identifies the line segment that metope fits, in meaning of parameters therein and T-type quarter bend
Parameter description it is consistent.
Step 13:I type quarter bend model
Fig. 4 show the quarter bend that the I type channel end belongs to a kind of specific type, since it has and the curved phase of exemplary right angle
As line segment combined characteristic, the honest example of this reality belonged to the curved one kind of exemplary right angle for being used for Global localization.According to constitute such
The difference of quarter bend ambient enviroment, the quarter bend of the type are broadly divided into two types, and I type quarter bend model includes by a face wall
The third line segment as fitting and two line segments in parastate by the abstract fitting of two package stackings are taken out in face, are taken out by a face metope
As the third line segment fitted is vertical with two line segments in parastate by the abstract fitting of two package stackings and intersects；Or it is described
I type quarter bend model includes being taken out by a package stacking as the 4th line segment that fits and by the abstract fitting of two metopes in parallel shape
Two line segments of state, by a package stacking take out as the 4th line segment that fits with by the abstract fitting of two metopes in the two of parastate
Line segment is vertical and intersects.The quarter bend model of the type describes are as follows:
Step 2: quarter bend feature normalization and foundation are global with reference to point set
In order to make the localization method based on line segment feature that there is versatility in different models, will be identified first by model
The quarter bend of classification is standardized, that is, extracts the same characteristic features for the positioning of AGV trolley；Then the right angle crook overall situation is established
With reference to point set.
Step 21: quarter bend feature normalization
The standardization of Step 211:L type quarter bend
As shown in figure 5, (a) (b) is respectively the quarter bend model after practical quarter bend model and table convert, the type it is straight
The straight line L of the curved metope fitting with two sides intersection in angle2And L3Intersection point Key_P and a fitting a straight line therein as positioning institute
Fixed character, in practical application, generally using with the immediate straight line of vehicle body axis of abscissas direction slope as positioning
Crucial linear feature.
The standardization of Step 212:T type quarter bend
As shown in fig. 6, finding out be abstracted by two package stackings respectively first in the scene during quarter bend feature normalization
The two line segment extended lines and second line segment L in parastate of fitting2Intersecting point coordinate P23And P21, then by P23And P21Arithmetic
Average value finds out key point Key_P, and the key point Key_P and metope that find out are abstracted to the second line segment L fitted2As fixed
The key feature of position, in which:
The standardization of Step 213:I type quarter bend
As shown in fig. 7, finding out be abstracted by two package stackings respectively first in the scene during quarter bend feature normalization
The two line segment extended lines and second line segment L in parastate of fitting2Intersecting point coordinate P23And P21, then by P23And P21Arithmetic
Average value finds out key point Key_P, and the key point Key_P and metope that find out are abstracted to the second line segment L fitted2As fixed
The key feature of position, in which:
Step 22: it establishes global with reference to point set
In the case where standardizing warehouse environment, the spatial position of shelf is put generally perpendicular with the wherein face metope in warehouse
Or parastate, based on this situation, for the present embodiment when establishing global coordinate system, the point with the abstract fitting of warehouse entrance is original
Point, using the direction of the line segment perpendicular to the abstract fitting of package stacking as y-axis, to be parallel to the abstract line segment being fitted of package stacking
Direction is that x-axis establishes global coordinate system xoy；And the global reference point of right angle crook positioning key point is established under global coordinate system
Collect P={ P1,P2,...,Pn, wherein Pi=(xi,yi,θri), i ∈ [1, n], (xi,yi) indicate quarter bend two after standardization
Coordinate of the intersecting straight lines intersection point at global coordinate system xoy, θriIt indicates with the straight line and global coordinate system xoy of the abstract fitting of metope
X-axis angle.As shown in figure 8, will be stored in database profession in order with reference to point set.
Step 3: establishing trolley original state coordinate system, obtains moving of car parameter, knot using the encoder of interior of body
Close preliminary pose estimation of the omni-directional mobile robots kinematics model counting of carriers in global coordinate system.
As shown in figure 9, establishing the original state vehicle body coordinate system x ' o ' of trolley using the geometric center of AGV trolley as origin
Y ', wherein y-axis direction is parallel with headstock direction, and laser radar coordinate system x " o " y " is vehicle body coordinate system x ' o ' y ' along headstock in figure
Positive direction horizontal-shift distance n.
Step 31: all-around mobile moving of car model is established
In above formula, (w1,w2,w3,w4) indicate that the revolving speed that four wheels are calculated according to wheel encoder data, r are wheel half
Diameter, m, n respectively indicate the half of body width, length, (vx,vy, w) and indicate mass center under vehicle body initial coordinate system respectively along x
Axis, y-axis, the movement speed in headstock direction.
Step 32: calculating according to dead reckoning principle at the t=kT moment, under vehicle body coordinate system x ' o ' y ' coordinate system,
Coordinate (the x ' of AGV mass centerk,y′k,θ′k) are as follows:
(Δxi, Δ yi, Δ θi) indicate the offset of AGV trolley mass center within i-th of period.Because T is sufficiently small,
It is assumed that in cycle T, (the Δ x under original state coordinatei,Δyi,Δθi) it is steady state value, then have:
Step 33: existed according to global coordinate system xoy and vehicle body coordinate system x ' o ' y ' coordinate system conversion relational expression counting of carriers
Center-of-mass coordinate (x under global coordinate system xoyk,yk,θk):
Wherein, (x0,y0,θ0) be the mass center at global coordinate system xoy initial coordinate, θ is trolley headstock direction and global
The angle of the y positive direction of coordinate system xoy, (x 'k,y′k,θ′k) indicate that trolley mass center is sat under vehicle body coordinate system x ' o ' y ' coordinate system
Mark.
Step 4: quarter bend feature identification
During carrying out Global localization to AGV trolley based on quarter bend feature, accurately identify that quarter bend is characterized in
The first step of Global localization is carried out, the process that the present embodiment identifies quarter bend is broadly divided into Primary Location and accurate judgement two is big
Step.
Step 41: quarter bend Primary Location
Firstly, manual remote control AGV trolley carries out the study in preliminary path in channel environment.When AGV trolley marches to directly
When near angle is curved, the small truck position (x obtained at the position based on dead reckoning under computer record is controlled in operationi,yi), and should
Principium identification key point when positioning result is as near AGV moving of car to quarter bend, is stored in database for it in order,
Generate the global path key point map that quarter bend tentatively identifies.
AGV trolley carry out subsequent independent navigation during, using posture tracking method real time discriminating AGV trolley whether
Into in the contiguous range of global path key point map, contiguous range is defined as centered on key point, radius r1's
Border circular areas, specific schematic diagram are as shown in Figure 10.
Step 42: quarter bend accurately determines
After carrying out Primary Location judgement to quarter bend, using the quarter bend decision model of front to Primary Location contiguous range
Interior environment does further identification, determines whether there is the quarter bend feature for meeting model in the contiguous range, and which belongs to
Class quarter bend feature.Should in the process mainly include two links: the line segment feature based on laser radar extracts and quarter bend model
Determine.
Step 421: the line segment feature based on laser radar extracts
The present embodiment carries out line segment feature extraction, the algorithm to laser radar scanning data using split_merge algorithm
It is broadly divided into segmentation-step of fitting-merging three.
The segmentation stage: being split the point set for being not belonging to same straight line, using IEPF clustering method, by right angle
The point cloud data of curved one frame scan of laser radar nearby is split, and obtains multiple point set set D=for belonging to different straight lines
(D1,D2,...,Dn)。
As shown in figure 11, to the laser data point D in a certain regioni, choose the head and the tail two o'clock in the region first as end
Point fitting a straight line li, concurrently set the distance threshold d for a little arriving straight linethd, all the points of head and the tail point-to-point transmission are traversed, are calculated
Point arrives the distance d of straight line outi, and find out a little to straight line liMaximum distance dmaxIf dmax< dthd, then determine own in the region
Point is on same straight line, otherwise, to put fitting a straight line apart from maximum point P to first place in the regionmaxFor separation, by this
Point set in region is divided into two point set { Pj(xj,yj) | j=1,2 ..., i } and { Pj(xj,yj) | j=i+1, i+2 ...,
N }, it is then concentrated respectively in each point and repeats above-mentioned segmentation step, and so on, until all point sets meet dividing strip
Until part, multiple point set set D=(D for belonging to different straight lines are finally obtained1,D2,...,Dn)。
The fitting stage: to the point set set D=(D after segmentation1,D2,...,Dn) carry out least square respectively line segment it is quasi-
It closes, the line segment for obtaining each set indicates L=(L1,L2,...,Ln), i ∈ [1, n], Li=(pi,θi, start, end, len),
Wherein piIt indicates from origin to straight line LiDistance, θiIndicated origin and LiVertical line and positive direction of the x-axis angle, start and
End respectively indicates line segment LiStarting endpoint coordinate, len indicate line segment LiLength.
Merging phase: to line segment aggregate L=(L1,L2,...,Ln) in meet the line segment of certain threshold range and merge,
It prevents same line segment to be divided into multistage, causes line segment over-fitting, adjacent two straight line Li,LjMerging criterion are as follows: | pi-pj
| < δp, and | θi-θj| < δθ, wherein δpAnd δθThe respectively merging threshold of line segment parameter merges place to line segment aggregate
Reason obtains new line segment aggregate L '=(L1,L2,...,Lm)。
It can be completed by segmentation-step of fitting-merging three and the accurate of scan data swept to a frame laser radar environment
It extracts, but in the physical channel environment being made of package stacking, between existing centainly due to putting between package and package
Gap can extract a plurality of intermittent conllinear line segment from one frame environmental scanning data of laser radar, as shown in figure 12.
In order to make to meet foretype quarter bend model to the description of quarter bend based on line segment feature in actual scene,
It needs to carry out the rejecting of invalid data to the line segment extracted in environment herein and line segment merges again.
Invalid data is rejected: mainly being rejected between the line segment being fitted in crack stacking package, is sat in laser radar
Under mark system, to any line segment Li=(pi,θi, start, end, len), there is following two rejecting principle are as follows: 1) to satisfaction | | θi|-
90 ° | < thdθLine segment in, if pi< thdp, which is rejected；If pi> thdpAnd len_i < dmin, reject；2) to satisfaction
||θi| -180 ° | < thdθOr | | θi| -0 ° | < thdθLine segment, if len_i < dmin, it rejects,；Wherein, parameter thdθFor line
Discrimination threshold between section polar diameter angle and 90 °, thdpThe threshold value met, d are needed for line segment polar diameter lengthminFor minimum length along path
Spend threshold value.
Merge again: the line segment after rejecting invalid data is carried out using the line segment merging method of split_merge algorithm
Merge again and finally obtains new line segment aggregate L "=(L1,L2,...,Lk)。
Step 422: the model based on line segment feature determines
Using the curved model of exemplary right angle to line segment aggregate L "=(L near quarter bend in a frame laser radar1,L2,...,
Lk) carry out Model checking and classify to it.
Step 5: Global localization and pose based on quarter bend feature update
Step 51: reference coordinate association
After AGV trolley completes the identification of quarter bend feature, need quarter bend feature at the position and the preset overall situation
Reference coordinate point is associated matching.
The present embodiment is associated matching to reference coordinate using " arest neighbors " method, in matching process, in conjunction with AGV trolley boat
Position calculates that resulting estimate goes out in quarter bend standardized model with reference to key point coordinate Pi′(Xi′,Yi'), " arest neighbors " method of utilization exists
Reference coordinate point concentrates Searching point Pi', when searching some reference point PjWith point Pi' between Euclidean distance minimum when, then it is assumed that
PjIt is the correct associated reference point of the right angle crook.
Step 52: the AGV pose based on line segment feature resolves
As shown in figure 13, there are two cartesian coordinate systems, global coordinate system xoy and laser radar coordinate system x ' o ' y ',
Two intersecting straight lines L under known global coordinate system1=(k1,b1,p1,θ1), L2=(k2,b2,p2,θ2), the intersecting point coordinate A of straight line
(xA,yA), wherein ki,biRespectively represent the slope and intercept of straight line, pi,θiThe respectively corresponding polar diameter of straight line and polar angle；Swashing
It is respectively L that measurement, which obtains straight line parameter, under optical radar coordinate system1'=(k1′,b1′,p1′,θ1'), L2'=(k2′,b2′,p2′,
θ2'), straight-line intersection coordinate A ' (x 'A,y′A), if the pose of laser radar coordinate origin is o " (x under global coordinate systemo″,
yo″,θo"), wherein θoIndicate that posture of the laser radar in global coordinate system towards angle, is expressed as radar fix system x-axis in figure
Positive direction and global coordinate system positive direction of the x-axis angle, can obtain in conjunction with Figure 13 and above-mentioned condition:
Pose the o " (x of laser radar under global coordinate system can be found out by bringing formula 14 into formula 13o,yo,θo)。
Step 53:AGV pose updates
Herein, laser radar and AGV car body belong to rigid body connection status, as shown in figure 9, similarly, passing through laser radar
Position coordinates and attitude angle of the AGV under global coordinate system can be found out with the conversion of AGV trolley coordinate system.
In Fig. 9, pose of the laser radar obtained using quarter bend feature under global coordinate system is o " (xo,yo,θo),
Position coordinates and attitude angle (x of the AGV trolley at global coordinate system xoy are converted to by coordinater,yr,θr):
Finally, the Global localization result using quarter bend is updated the relative positioning result of AGV trolley, eliminate opposite
Accumulated error in positioning result completes the global pose calibration of AGV trolley.
The all directionally movable robot pose calibration method based on right angle bend feature of the present embodiment make full use of it is sharp
In the case where the environmental information of optical radar acquisition, feature is linearized in conjunction with the structuring of composition exemplary right angle bend, proposes base
In the omni-directional mobile robots pose calibration method of quarter bend feature, the method is that omni-directional mobile robots channel is realized in engineering
Interior accurate positioning provides a solution, and the program is easily achieved, is adaptable, is able to solve mobile robot in channel
Position error problem.
Finally, it is stated that the above examples are only used to illustrate the technical scheme of the present invention and are not limiting, although referring to compared with
Good embodiment describes the invention in detail, those skilled in the art should understand that, it can be to skill of the invention
Art scheme is modified or replaced equivalently, and without departing from the objective and range of the technical program, should all be covered in the present invention
Protection scope in.
Claims (7)
1. a kind of all directionally movable robot pose calibration method based on right angle bend feature, which is characterized in that this method packet
Include following steps:
S1. quarter bend model, including L-type quarter bend model, I type quarter bend model, T-type quarter bend model are established；
Wherein, the L-type quarter bend model include by the abstract perpendicular intersection fitted of two sides metope two line segments and
One, by abstract the first line segment being fitted of package stacking, is abstracted two line segments of the perpendicular intersection fitted by two sides metope
Wherein a line segment and the first line segment horizontal line by the abstract fitting of package stacking；
The T-type quarter bend model includes by the abstract second line segment fitted of a face metope and by the abstract fitting of two package stackings
Two line segments in parastate, be abstracted by two line segments in parastate of the abstract fitting of two package stackings with by a face metope
The second line segment fitted is vertical but non-intersecting；
The I type quarter bend model includes the third line segment taken out by a face metope as fitting and is fitted by two package stackings are abstract
Two line segments in parastate, by a face metope take out as the third line segment that fits with by two package stackings it is abstract be fitted be in
Two line segments of parastate are vertical and intersect；Or the I type quarter bend model include taken out by a package stacking as fit the
Four line segments and two line segments in parastate that fitting is abstracted by two metopes take out the 4th line segment as fitting by a package stacking
It is vertical with two line segments in parastate by the abstract fitting of two metopes and intersect；
S2. quarter bend model is standardized, then establishes the right angle crook overall situation with reference to point set；
S3. trolley original state coordinate system is established, and calculates preliminary pose estimation of the AGV trolley in global coordinate system；
S4. quarter bend feature is identified；
S5. Global localization is carried out to AGV trolley based on quarter bend feature and pose updates.
2. a kind of all directionally movable robot pose calibration method based on right angle bend feature according to claim 1,
It is characterized in that, the L-type quarter bend model are as follows:
The T-type quarter bend model are as follows:
The I type quarter bend model are as follows:
In two-dimensional Cartesian coordinate system lower line segment Li=(pi,θi, start, end, len), wherein start_q and end_q distinguishes table
Show small front side wall matching line segment rise, beginning extreme coordinates, start_ (q+1) trolley left channel inner wall matching line segment rise
Point coordinate, end_ (q-1) indicate the initial point coordinate of trolley right channel inner wall matching line segment, piIt indicates from origin to straight line Li's
Distance, θiIndicated origin and LiVertical line and positive direction of the x-axis angle, start and end respectively indicate line segment LiStarting point
Point coordinate, len indicate line segment LiLength, δdminAnd δdmaxIndicate the minimum threshold and max-thresholds of Euclidean distance between two o'clock,
δdmaxFor judging whether adjacent two lines section head and the tail intersect, δdminFor whether judging the distance between adjacent two lines section is the first
Greater than certain threshold value.
3. a kind of all directionally movable robot pose calibration method based on quarter bend feature according to claim 1,
It is characterized in that, in the step S2, to the standardized method of the L-type quarter bend aspect of model are as follows:
By by two line segments of the abstract perpendicular intersection fitted of two sides metope and by two sides metope it is abstract fit be in
A line segment in two line segments of the state that intersects vertically is as fixed character；
To T-type quarter bend aspect of model standardized method are as follows:
Find out the friendship of the two line segment extended lines and second line segment in parastate by the abstract fitting of two package stackings respectively first
Point coordinate；Then the arithmetic mean of instantaneous value for seeking two intersecting point coordinates, using the arithmetic mean of instantaneous value as key point Key_P；
And using key point Key_P and second line segment as the key feature of positioning；
To the standardized method of the I type quarter bend aspect of model are as follows:
It finds out to be prolonged by two line segments in parastate of the abstract fitting of two package stackings respectively first and be sat with the intersection point of third line segment
Mark；
Then the arithmetic mean of instantaneous value for seeking two intersecting point coordinates, using the arithmetic mean of instantaneous value as key point Key_P'；
And using key point Key_P' and third line segment as the key feature of positioning；
Or to the standardized method of the I type quarter bend aspect of model are as follows:
Find out the intersecting point coordinate prolonged by two line segments in parastate of the abstract fitting of two metopes with the 4th line segment respectively first；
Then the arithmetic mean of instantaneous value for seeking two intersecting point coordinates, using the arithmetic mean of instantaneous value as key point Key_P "；
And using key point Key_P " and the 4th line segment as the key feature of positioning.
4. a kind of all directionally movable robot pose calibration method based on quarter bend feature according to claim 3,
It is characterized in that, in the step S2, when establishing global coordinate system；Using the point of the abstract fitting of warehouse entrance as origin, to hang down
It is directly y-axis in the direction of the line segment of the abstract fitting of package stacking, is abstracted the direction for the line segment being fitted to be parallel to package stacking as x
Axis establishes global coordinate system xoy；And the overall situation of right angle crook positioning key point is established under global coordinate system with reference to point set P=
{P1,P2,...,Pn, wherein Pi=(xi,yi,θri), i ∈ [1, n], (xi,yi) indicate quarter bend two intersections after standardization
Coordinate of the straight-line intersection at global coordinate system xoy, θriIndicate the x with the straight line of the abstract fitting of metope and global coordinate system xoy
The angle of axis.
5. a kind of all directionally movable robot pose calibration method based on quarter bend feature according to claim 4,
It is characterized in that, the step S3 specifically includes following sub-step:
S31. the vehicle body coordinate system x ' o ' y ' of the original state of trolley is established using the geometric center of AGV trolley as origin, is established complete
Orientation moving trolley kinematics model:
Wherein, (w1,w2,w3,w4) indicate AGV trolley four wheels revolving speed, r is radius of wheel, and it is wide that m, n respectively indicate vehicle body
The half of degree, length, (vx,vy, w) and indicate that mass center moves respectively along x-axis, y-axis, headstock direction under the vehicle body initial coordinate system
Speed；
S32. it is calculated according to dead reckoning principle at t=kT moment, at vehicle body coordinate system x ' o ' y ', counting of carriers center-of-mass coordinate
(x′k,y′k,θ′k), θ 'kIndicate that the angle of the y positive direction of trolley headstock direction and trolley geocentric coordinate system, T indicate dead reckoning
Sampling period, k indicate positive integer；
S33. according to the vehicle body coordinate system x ' o ' y ' conversion relational expression counting of carriers of global coordinate system xoy and trolley in world coordinates
It is center-of-mass coordinate (x under xoyk,yk,θk), θkIndicate the angle of the y positive direction of trolley headstock direction and global coordinate system.
6. a kind of all directionally movable robot pose calibration method based on quarter bend feature according to claim 5,
It is characterized in that, the step S4 specifically includes following sub-step:
S41. it firstly, manual remote control AGV trolley carries out the study in preliminary path in channel environment, generates quarter bend and tentatively identifies
Global path key point map；During AGV trolley carries out subsequent independent navigation, according to priori map, using pose
Whether tracking real time discriminating AGV trolley enters in the contiguous range of global path key point map, contiguous range definition
For centered on key point, radius r1Border circular areas；
S42. the line segment feature based on laser radar extracts, and is carried out using split_merge algorithm to laser radar scanning data
Line segment feature extracts to obtain line segment aggregate L'；The rejecting of invalid data is carried out to the line segment extracted in environment and line segment merges again
Obtain new line segment aggregate L "；
S43. Model checking and right is carried out to the line segment aggregate L " near quarter bend in a frame laser radar using quarter bend model
It is classified.
7. a kind of all directionally movable robot pose calibration method based on quarter bend feature according to claim 6,
It is characterized in that, the step S5 specifically includes following sub-step:
S51. reference coordinate is associated with；After AGV trolley completes the identification of quarter bend feature, needing will be at the small truck position AGV at this time
Quarter bend feature is associated with preset global reference coordinate point and matches, and obtains correct associated reference point；
S52. the AGV pose based on line segment feature resolves, and utilizes the standardization quarter bend mould extracted in laser radar scanning data
Molded line section and key feature carry out pose to laser radar and resolve to obtain the pose o " of laser radar under global coordinate system xoy
(xo,yo,θo)；
S53.AGV pose updates, and finds out AGV trolley in global coordinate system xoy by laser radar and the conversion of AGV trolley coordinate system
Under position coordinates and attitude angle (xr,yr,θr), using the Global localization result of quarter bend to the relative positioning result of AGV trolley
It is updated.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810529469.2A CN109000649B (en) | 2018-05-29 | 2018-05-29 | Omni-directional mobile robot pose calibration method based on right-angle bend characteristics |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810529469.2A CN109000649B (en) | 2018-05-29 | 2018-05-29 | Omni-directional mobile robot pose calibration method based on right-angle bend characteristics |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109000649A true CN109000649A (en) | 2018-12-14 |
| CN109000649B CN109000649B (en) | 2022-02-15 |
Family
ID=64573689
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810529469.2A Active CN109000649B (en) | 2018-05-29 | 2018-05-29 | Omni-directional mobile robot pose calibration method based on right-angle bend characteristics |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109000649B (en) |
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110146110A (en) * | 2019-05-20 | 2019-08-20 | 哈尔滨工程大学 | A kind of error hiding determination method of indoor environment robot line feature ICNN data correlation |
| CN110260875A (en) * | 2019-06-20 | 2019-09-20 | 广州蓝胖子机器人有限公司 | A kind of method in Global motion planning path, Global motion planning device and storage medium |
| CN110442136A (en) * | 2019-08-12 | 2019-11-12 | 北京特种机械研究所 | A kind of any attitude air navigation aid of Mecanum wheel AGV |
| CN110823225A (en) * | 2019-10-29 | 2020-02-21 | 北京影谱科技股份有限公司 | Positioning method and device under indoor dynamic situation |
| CN111060093A (en) * | 2019-12-31 | 2020-04-24 | 芜湖哈特机器人产业技术研究院有限公司 | Mobile robot positioning method and system based on single corner folding plate |
| CN111142117A (en) * | 2019-12-31 | 2020-05-12 | 芜湖哈特机器人产业技术研究院有限公司 | Hybrid navigation map construction method and system fusing corner plates and occupying grids |
| CN112068128A (en) * | 2020-09-19 | 2020-12-11 | 重庆大学 | Method for processing data and acquiring pose of linear radar in straight road scene |
| CN112230256A (en) * | 2019-07-15 | 2021-01-15 | 苏州宝时得电动工具有限公司 | Autonomous robot, positioning calibration method and device thereof, and storage medium |
| CN112327325A (en) * | 2020-09-16 | 2021-02-05 | 安徽意欧斯物流机器人有限公司 | Method for improving 2D-SLAM precision and stability based on characteristic road sign |
| CN112577494A (en) * | 2020-11-13 | 2021-03-30 | 上海航天控制技术研究所 | SLAM method, electronic device and storage medium suitable for lunar vehicle |
| CN112731354A (en) * | 2020-12-24 | 2021-04-30 | 芜湖哈特机器人产业技术研究院有限公司 | Self-calibration method for laser radar pose on AGV |
| CN113391318A (en) * | 2021-06-10 | 2021-09-14 | 上海大学 | Mobile robot positioning method and system |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103941264A (en) * | 2014-03-26 | 2014-07-23 | 南京航空航天大学 | Positioning method using laser radar in indoor unknown environment |
| CN104729500A (en) * | 2015-02-15 | 2015-06-24 | 南京航空航天大学 | Global positioning method of laser-navigated AGV (automatic guided vehicle) |
| CN105987697A (en) * | 2016-04-26 | 2016-10-05 | 重庆大学 | Right-angle bend Mecanum wheel type AGV (Automatic Guided Vehicle) navigation and positioning method and system |
| CN106643805A (en) * | 2016-12-30 | 2017-05-10 | 上海交通大学 | Position calibration method of laser positioning sensor in AGV (automated guided vehicle) |
| CN107065887A (en) * | 2017-05-26 | 2017-08-18 | 重庆大学 | Backward air navigation aid in omni-directional mobile robots passage |
| CN107933557A (en) * | 2017-11-17 | 2018-04-20 | 吉林大学 | A kind of vehicle right-angled bend control system and its control method based on multi-cam |
| CN108036792A (en) * | 2017-12-11 | 2018-05-15 | 苏州中德睿博智能科技有限公司 | A kind of data fusion method of odometer for mobile robot and measurement pose |
-
2018
- 2018-05-29 CN CN201810529469.2A patent/CN109000649B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103941264A (en) * | 2014-03-26 | 2014-07-23 | 南京航空航天大学 | Positioning method using laser radar in indoor unknown environment |
| CN104729500A (en) * | 2015-02-15 | 2015-06-24 | 南京航空航天大学 | Global positioning method of laser-navigated AGV (automatic guided vehicle) |
| CN105987697A (en) * | 2016-04-26 | 2016-10-05 | 重庆大学 | Right-angle bend Mecanum wheel type AGV (Automatic Guided Vehicle) navigation and positioning method and system |
| CN106643805A (en) * | 2016-12-30 | 2017-05-10 | 上海交通大学 | Position calibration method of laser positioning sensor in AGV (automated guided vehicle) |
| CN107065887A (en) * | 2017-05-26 | 2017-08-18 | 重庆大学 | Backward air navigation aid in omni-directional mobile robots passage |
| CN107933557A (en) * | 2017-11-17 | 2018-04-20 | 吉林大学 | A kind of vehicle right-angled bend control system and its control method based on multi-cam |
| CN108036792A (en) * | 2017-12-11 | 2018-05-15 | 苏州中德睿博智能科技有限公司 | A kind of data fusion method of odometer for mobile robot and measurement pose |
Non-Patent Citations (1)
| Title |
|---|
| 孙棣华等: "基于滤波反演法的参数不确定自动引导车的运动控制", 《控制与决策》 * |
Cited By (19)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110146110A (en) * | 2019-05-20 | 2019-08-20 | 哈尔滨工程大学 | A kind of error hiding determination method of indoor environment robot line feature ICNN data correlation |
| CN110146110B (en) * | 2019-05-20 | 2022-11-15 | 哈尔滨工程大学 | Mismatching judgment method for ICNN data association of robot line characteristics in indoor environment |
| CN110260875A (en) * | 2019-06-20 | 2019-09-20 | 广州蓝胖子机器人有限公司 | A kind of method in Global motion planning path, Global motion planning device and storage medium |
| CN112230256B (en) * | 2019-07-15 | 2024-04-09 | 苏州宝时得电动工具有限公司 | Autonomous robot, positioning calibration method and device thereof, and storage medium |
| CN112230256A (en) * | 2019-07-15 | 2021-01-15 | 苏州宝时得电动工具有限公司 | Autonomous robot, positioning calibration method and device thereof, and storage medium |
| CN110442136A (en) * | 2019-08-12 | 2019-11-12 | 北京特种机械研究所 | A kind of any attitude air navigation aid of Mecanum wheel AGV |
| CN110823225A (en) * | 2019-10-29 | 2020-02-21 | 北京影谱科技股份有限公司 | Positioning method and device under indoor dynamic situation |
| CN111060093B (en) * | 2019-12-31 | 2021-07-30 | 芜湖哈特机器人产业技术研究院有限公司 | Mobile robot positioning method and system based on single corner folding plate |
| CN111060093A (en) * | 2019-12-31 | 2020-04-24 | 芜湖哈特机器人产业技术研究院有限公司 | Mobile robot positioning method and system based on single corner folding plate |
| CN111142117A (en) * | 2019-12-31 | 2020-05-12 | 芜湖哈特机器人产业技术研究院有限公司 | Hybrid navigation map construction method and system fusing corner plates and occupying grids |
| CN111142117B (en) * | 2019-12-31 | 2021-11-05 | 芜湖哈特机器人产业技术研究院有限公司 | Hybrid navigation map construction method and system fusing corner plates and occupied grids |
| CN112327325A (en) * | 2020-09-16 | 2021-02-05 | 安徽意欧斯物流机器人有限公司 | Method for improving 2D-SLAM precision and stability based on characteristic road sign |
| CN112068128A (en) * | 2020-09-19 | 2020-12-11 | 重庆大学 | Method for processing data and acquiring pose of linear radar in straight road scene |
| CN112577494A (en) * | 2020-11-13 | 2021-03-30 | 上海航天控制技术研究所 | SLAM method, electronic device and storage medium suitable for lunar vehicle |
| CN112577494B (en) * | 2020-11-13 | 2022-11-18 | 上海航天控制技术研究所 | SLAM method, electronic device and storage medium suitable for lunar vehicle |
| CN112731354A (en) * | 2020-12-24 | 2021-04-30 | 芜湖哈特机器人产业技术研究院有限公司 | Self-calibration method for laser radar pose on AGV |
| CN112731354B (en) * | 2020-12-24 | 2024-04-05 | 芜湖哈特机器人产业技术研究院有限公司 | Self-calibration method for laser radar pose on AGV |
| CN113391318A (en) * | 2021-06-10 | 2021-09-14 | 上海大学 | Mobile robot positioning method and system |
| CN113391318B (en) * | 2021-06-10 | 2022-05-17 | 上海大学 | Mobile robot positioning method and system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109000649B (en) | 2022-02-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109000649A (en) | A kind of all directionally movable robot pose calibration method based on right angle bend feature | |
| CN109916393B (en) | Multi-grid-value navigation method based on robot pose and application thereof | |
| CN113168717B (en) | Point cloud matching method and device, navigation method and equipment, positioning method and laser radar | |
| CN109186606B (en) | Robot composition and navigation method based on SLAM and image information | |
| KR101003168B1 (en) | Multidimensional Evidence Grids and System and Methods for Applying Same | |
| CN107741234A (en) | The offline map structuring and localization method of a kind of view-based access control model | |
| CN111693050B (en) | Indoor medium and large robot navigation method based on building information model | |
| Eppenberger et al. | Leveraging stereo-camera data for real-time dynamic obstacle detection and tracking | |
| US11543530B1 (en) | Indoor SLAM method based on 3D lidar and UWB | |
| Bista et al. | Appearance-based indoor navigation by IBVS using line segments | |
| Chen et al. | Real-time identification and avoidance of simultaneous static and dynamic obstacles on point cloud for UAVs navigation | |
| Simond et al. | Homography from a vanishing point in urban scenes | |
| CN114047750A (en) | Express delivery warehousing method based on mobile robot | |
| Huang et al. | Mobile robot localization using ceiling landmarks and images captured from an rgb-d camera | |
| CN114998276A (en) | Robot dynamic obstacle real-time detection method based on three-dimensional point cloud | |
| Shi et al. | Online topological map building and qualitative localization in large-scale environment | |
| Zhu et al. | Indoor multi-robot cooperative mapping based on geometric features | |
| Yang et al. | Multi-agv tracking system based on global vision and apriltag in smart warehouse | |
| CN110320912A (en) | The AGV positioning navigation device and method that laser is merged with vision SLAM | |
| Li et al. | Pallet detection and localization with RGB image and depth data using deep learning techniques | |
| CN115454055B (en) | Multi-layer fusion map representation method for indoor autonomous navigation and operation | |
| CN115755888A (en) | AGV obstacle detection system with multi-sensor data fusion and obstacle avoidance method | |
| Yi et al. | Map representation for robots | |
| Tazaki et al. | Outdoor autonomous navigation utilizing proximity points of 3D Pointcloud | |
| Noaman et al. | Landmarks exploration algorithm for mobile robot indoor localization using VISION sensor |
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 |