WO2012160630A1 - 軌跡補正方法、軌跡補正装置および移動体装置 - Google Patents
軌跡補正方法、軌跡補正装置および移動体装置 Download PDFInfo
- Publication number
- WO2012160630A1 WO2012160630A1 PCT/JP2011/061685 JP2011061685W WO2012160630A1 WO 2012160630 A1 WO2012160630 A1 WO 2012160630A1 JP 2011061685 W JP2011061685 W JP 2011061685W WO 2012160630 A1 WO2012160630 A1 WO 2012160630A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- node
- trajectory
- measurement
- unit
- probability
- Prior art date
Links
Images
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
- G01C21/005—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 with correlation of navigation data from several sources, e.g. map or contour matching
-
- 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
- G01C21/26—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 specially adapted for navigation in a road network
- G01C21/28—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00 specially adapted for navigation in a road network with correlation of data from several navigational instruments
- G01C21/30—Map- or contour-matching
Definitions
- the present invention relates to a trajectory correction method for correcting a trajectory of a moving body, a trajectory correction apparatus, and a technology of the mobile body apparatus.
- Patent Document 1 discloses an environment identification device, an environment identification method, a program, a recording medium, and a robot that use a camera as an external sensor, recognize a specific shape as a landmark, and perform autonomous movement while generating a map.
- An apparatus is disclosed.
- Patent Document 2 using a laser scanner as an external sensor, the shape data of a peripheral object acquired at the current time and the shape data of the peripheral object acquired at a position different from the current time at the previous time are sequentially recorded.
- An environment map generation method and a mobile robot are disclosed in which a map is generated by enlarging a region where shape data of peripheral objects is measured by matching (superimposing) to a target object.
- Measured error occurs in the acquired trajectory when the moving body travels. Therefore, the sequential map generation method as in the prior art has a problem that the cumulative error increases as the area of the map to be created increases. In particular, when passing a previously visited point again with a different route or when generating a map with multiple vehicles, the same point is stored as a different point on the map due to measurement error, and consistency can be maintained. May cause wrinkles. As a result, there is a problem that the accuracy of the map is deteriorated, the moving body loses its own position or the target route, and it is difficult to continue the movement.
- the present invention has been made in view of such a background, and an object of the present invention is to calculate a highly accurate trajectory.
- the present invention provides a trajectory correction method by a trajectory correction apparatus that corrects a trajectory traveled by a mobile object, the trajectory correction apparatus including trajectory data of the mobile object acquired by a single measurement unit.
- the trajectory correction apparatus including trajectory data of the mobile object acquired by a single measurement unit.
- a plurality of nodes are set, and the position data of the moving body acquired by the first measuring unit is associated with the node, and the moving body acquired by another measuring unit different from the first measuring unit
- the position data is associated with the node, the position where the node is likely to occur is represented by probability, the position where the position data associated with the node is likely to occur is represented by probability, and each probability
- An evaluation function including the node and the position data as variables is calculated based on the above, and a trajectory having the highest probability of occurrence of each node is calculated based on the evaluation function.
- a highly accurate trajectory can be calculated.
- FIG. 1 is a diagram illustrating a configuration example of an autonomous mobile system according to the present embodiment.
- the autonomous mobile system 1 includes an in-vehicle unit 10 that collects data necessary for map creation, and a management unit 20 that performs trajectory correction based on the collected data.
- the in-vehicle unit 10 is mounted on a moving body V that is a moving body device such as an autonomous mobile robot or a vehicle.
- the management unit 20 is installed in a management facility such as an office building.
- the in-vehicle unit 10 and the management unit 20 can communicate via a wireless network or the like.
- the in-vehicle unit 10 includes a peripheral object shape measuring unit 11, a peripheral object shape matching unit 12, a geographic information system data acquiring unit 13, a geographic information system matching unit 14, a GNSS (Global Navigation Satellite System) positioning unit 15, and a wheel rotation amount measuring unit. 16, a control unit 17 and the like.
- the peripheral object shape matching unit 12, the geographic information system matching unit 14, the GNSS positioning unit 15, and the wheel rotation amount measuring unit 16 may be collectively referred to as measurement means.
- the peripheral object shape measuring unit 11 measures the shape of an object (a building, a road tree, a road pole such as a power pole, a person, or another moving object V) existing around the moving object V.
- a laser scanner, a stereo camera, a TOF (Time Of Flight) distance image camera, or the like can be used as the peripheral object shape measuring unit 11.
- the peripheral object shape matching unit 12 receives, at the current position, the peripheral object shape data (structure shape data) measured by the peripheral object shape measurement unit 11 of the moving object V and the peripheral object shape data measured at different positions. By matching (superimposing), the difference from the previous time is calculated for the relative position between the two points of the moving object V.
- matching for example, the paper “Distance Data Processing-Shape Model Generation Technology from Multiple Distance Images” (authors: Ken Masuda, Kyoko Okaya, Riksho Sagawa, Conference Meeting: Proc. Of the 146th CVIM, year held : 2004) can be used.
- the geographic information system data acquisition unit 13 is a part that acquires map shape data of road structures in the city, road structures such as buildings, roadside trees, and utility poles from a geographic information system having map data.
- map shape data As a data format of the map shape data, a city three-dimensional model used in a recent car navigation system, City GML (Geography Markup Language) standardized by OGC (Open Geospatial Consortium), or the like can be used.
- the geographic information system matching unit 14 maps the shape data of the surrounding object measured by the surrounding object shape measuring unit 11 of the moving object V at the current position and the surrounding object map near the current position acquired by the geographic information system data acquiring unit 13. By matching (superimposing) the shape data, the difference from the previous time is calculated for the relative position of the current position of the moving object V to the map of the geographic information system. For the matching, the same method as that of the peripheral object shape matching unit 12 can be used.
- the GNSS positioning unit 15 uses a positioning system such as GPS (Global Positioning System) to calculate the current position of the moving object V with respect to a reference coordinate system such as a planar rectangular coordinate system.
- GPS Global Positioning System
- the wheel rotation amount measuring unit 16 calculates the relative difference of the current position with respect to the position of the moving object V at the previous time by accumulating the wheel rotation amount.
- an inertial sensor or gyro sensor called IMU (Inertial Measurement Unit) is used. The method described in the conference: Proc. OfRA ICRA '96, held year: 1996) can be used.
- the in-vehicle unit 10 does not have to include all of the units 11 to 16 and only needs to have at least two of these units.
- the control unit 17 performs overall control of the units 11 to 16.
- Each unit 11 to 17 is realized by a program stored in a ROM (Read Only Memory) or the like being executed by a CPU (Central Processing Unit).
- ROM Read Only Memory
- CPU Central Processing Unit
- the management unit 20 includes an evaluation function generation unit 21, a trajectory optimization calculation unit 22, and a shape map data generation unit 23.
- the evaluation function generation unit 21 uses the difference in relative position (details will be described later) of the moving body V with respect to the previous time calculated by each of the measuring units 11 to 16 to move the moving body V when traveling in the traveling environment. This is a part for generating an evaluation function expressed by a probability function using a difference of each measured relative position as a variable.
- the trajectory error is corrected by obtaining a variable (that is, a relative position difference) when the evaluation function is optimized. Details will be described later.
- the trajectory optimization calculation unit 22 optimizes the evaluation function generated by the evaluation function generation unit 21 using a relative position difference as a variable. As a result, the trajectory optimization calculation unit 22 corrects the trajectory of the moving object V in which the accumulated error has occurred to the trajectory of the moving object V that has no accumulated error and maintains consistency throughout. Details will be described later.
- the shape map data generation unit 23 affixes the shape data of the peripheral object measured by the peripheral object shape measurement unit 11 to the trajectory of the moving object V corrected by the trajectory optimization calculation unit 22 so that there is no accumulated error. Generate a map.
- the management unit 20 is realized by a PC (Personal Computer) or the like, and each of the units 21 to 23 is a program stored in a ROM or HDD (Hard Disk Drive) is expanded in a RAM (Random Access Memory) and executed by the CPU. By embodying it.
- FIG. 2 is a flowchart showing a processing procedure of the autonomous mobile system according to the present embodiment.
- the autonomous mobile system 1 calculates the exact trajectory by executing the processing of the flowchart shown in FIG. 2, and further generates a map based on this trajectory, so that the moving body V does not lose sight of its own position or target route. You can reach your destination.
- the surrounding object shape measuring unit 11 is an object (a road structure such as a building, a roadside tree, or a power pole, a person or another mobile object) that exists around the mobile object V.
- the shape (object shape) of the body V etc. is measured (S101).
- the traveling here may be autonomous traveling by the moving body V or may be steered by a pilot.
- the surrounding object shape measuring unit 11 measures the shape of an object such as a road structure 411 that falls within its own measurement range 401 with the moving body V traveling in the travelable region (road) 412 in FIG.
- the peripheral object shape measuring unit 11 emits a laser to the surroundings, and measures the shape of the surrounding objects by reflection thereof.
- the peripheral object shape measuring unit 11 is configured so that the periphery of the region 501 having a predetermined height as shown in FIG.
- Object shape data may be extracted.
- the peripheral object shape measurement unit 11 may extract characteristic peripheral object shape data such as a plane or a cylinder.
- each of the measuring means 12, 14 to 16 acquires measurement data (S102). At this time, each measurement means 12, 14 to 16 transmits measurement data to the management unit 20.
- the control unit 17 determines whether or not a predetermined amount of measurement data has been accumulated (S103). As a result of step S103, when a predetermined amount of measurement data is not accumulated (S103 ⁇ No), the control unit 17 returns the process to step S101.
- the evaluation function generation unit 21 sets a node for the trajectory traveled by the moving object V created based on the transmitted data.
- a graph structure is generated by generating an arc (S104).
- FIG. 5 the same elements as those in FIG. A travel locus t is a diagram illustrating an example of a travel locus when the moving body V actually travels. As shown in FIG. 5, it is assumed that the moving object V is traveling along an actual travel locus t. The travel locus t is calculated based on the distance measured from the number of wheel rotations by the wheel rotation amount measuring unit 16. The travel locus t is actually closed.
- FIG. 6 is a diagram for explaining the relationship between nodes and arcs.
- the evaluation function generation unit 21 divides the travel locus t (FIG. 5) for each predetermined length, sets the divided point as a node p, and connects the nodes p with a straight arc g. Then, a graph structure is generated and a trajectory X is generated. At this time, the evaluation function generation unit 21 generates the node p and the arc g with the position of the moving body V calculated by the wheel rotation amount measurement unit 16 accumulating the wheel rotation amount as an initial position. Since a cumulative error occurs in the measurement of the wheel rotation amount measurement unit 16, the graph structure (trajectory X) generated by the evaluation function generation unit 21 shown in FIG.
- FIG. 6 shows the travel locus t in the actual environment shown in FIG. Measurement error (deviation) has occurred.
- the autonomous mobile system 1 executes the processing procedure (S105 to S109) of the subsequent flowchart shown in FIG.
- the i-th node p is expressed as a node p i and its position is expressed as a vector x i .
- the vector x i that is the position of the node p i is expressed as in Expression (1).
- u i and v i are, for example, world coordinates
- ⁇ i is the posture (orientation) of the moving object V.
- the trajectory X represented by the node p and the arc g is expressed as a set of vectors x i representing the positions of the n nodes p i as shown in Expression (2).
- n is the number of nodes.
- the peripheral function shape matching unit 12, the geographic information system matching unit 14, the GNSS positioning unit 15, the wheel rotation are applied to the graph structure composed of the node p and the arc g generated by the evaluation function generation unit 21 in step S 104.
- Each of the quantity measuring units 16 calculates the correspondence between the measurement z and the node p, and calculates the probability distribution of the difference measurement error between the acquired measurement z and the node p (S105).
- FIG. 7 is a diagram for explaining definitions of terms according to the present embodiment.
- FIG. 7 (a) shown as the position of the wheel rotation quantity measuring unit 16 has measured, the example using the node p 1 and the node p 2.
- the measurement error distribution (elliptical distribution y 1 ) at the node p 2 is represented by a normal distribution accuracy matrix ⁇ .
- the ellipse y 1 is a covariance ellipse represented by the accuracy matrix ⁇ .
- the accuracy matrix ⁇ is also called an information matrix and corresponds to an inverse matrix of a normal distribution covariance matrix.
- elliptic y1 in FIG. 7 (a) the normal distribution centered on node p 2
- the small arrow q indicates the orientation of the moving body V and corresponds to ⁇ i in the equation (1).
- FIG. 7B a method of associating a position measured by another measuring unit with a node measured by the wheel rotation amount measuring unit 16 will be described.
- a position where there is a measuring means (for example, the GNSS positioning unit 15) other than the wheel rotation amount measuring unit 16 is measured.
- This position is described as measurement z (m 1 ) (position data).
- “m 1 ” is an identification number indicating the measurement means (for example, the GNSS positioning unit 15) that measured the measurement z. That is, the measurement z (m 1 ) means the measurement z measured by the measurement unit m 1 .
- the evaluation function generation unit 21 determines which node p is associated with the measurement z (m 1 ). For example, if the evaluation function generation unit 21 determines that the measurement z (m 1 ) corresponds to the node p 2 based on the time and the like, the measurement z is expressed as measurement z 2 (m 1 ).
- the node p 1 and the measurement z 2 (m 1) and the measurement of the relative position of the difference of the line connecting the (hereinafter, referred to as differential measurement) is and be expressed as Z 12 (m 1).
- the component of the difference measurement Z 12 (m 1 ) is represented by the difference between the measurement z 2 and the node p 1 .
- the correspondence relationship is that the covariance ellipse of the measurement z 2 (m 1 ) is associated with the node p 2 (in other words, the difference measurement Z 12 (m 1 ) and the arc g 12 are associated).
- the node p measured by the wheel rotation amount measuring unit 16 becomes an initial value of the correspondence relationship with the measurement z measured by other measuring means, that is, a reference value.
- the measurement z 2 (m 1 ) Since the measurement z 2 (m 1 ) is also considered to contain an error, the measurement z 2 (m 1 ) has a covariance ellipse y2 with an accuracy matrix ⁇ as in FIG.
- the difference source for the difference measurement is arbitrarily set by the user. For example, as shown in FIG. 7D, a predetermined number of nodes p before the node p associated with the measurement z may be set as the difference source.
- the previous node p is the difference source. That is, a difference source of differential measurement generated for measurement z 3 (m 1) (Z 13) is in two previous node p measurement z 3 (m 1) and are associated with the node p 3 and it has a certain node p 1. Further, a difference source of differential measurements are generated for measurement z 4 (m 1) (Z 24) is a measurement z 4 2 preceding node (m 1) and are associated with the node p 4 p and it has a certain node p 2.
- the method of determining the difference source for the difference measurement is not limited to the method of FIG.
- a node p previously associated with the measurement z may be set as the difference source.
- the measurement z 1 (m 1 ) and the node p 1 are associated
- the measurement z 3 (m 1 ) and the node p 3 are associated
- the measurement z 4 Assume that (m 1 ) and node p 4 are associated with each other.
- the difference source of the difference measurement (Z 13 ) generated for the measurement z 3 (m 1 ) is the node p 1 associated with the measurement z before that.
- the difference source of the difference measurement (Z 34 ) generated for the measurement z 4 (m 1 ) is the node p 3 associated with the measurement z before that.
- the measurement z 4 (m 1) is a difference source of other differential measurement.
- the node p is also measured z.
- the correspondence relationship c 1,2 (m 0 ) can be defined (not shown).
- each of the measuring means 12 and 14 to 16 calculates the measurement z j (m k ), the difference measurement Z ij (m k ), and the correspondence relationship c i, j (m k ) measured by itself.
- i, j, and k are integers, and i ⁇ j.
- FIG. 8 is a diagram for explaining locus determination.
- the trajectory in FIG. 8 is the same as the trajectory X in FIG. In FIG. 8, nodes p 0 , p 1 , p 2 ,... Are calculated, and the corresponding measurements z 0 (m 1 ), z 1 (m 1 ), z 2 (m 1 ), z 5 are calculated. (M 1 ),... Are calculated ((m 1 ) is omitted in FIG.
- the difference source node is a node three nodes before the node for which the correspondence is defined.
- the nodes p 0 , p 1 , p 2 ,... Can also be measured as z 0 (m 0 ), z 1 (m 0 ), and z 2 (m 0 ), respectively.
- the illustration is omitted here (“m 0 ” is an identification number indicating the wheel rotation amount measuring unit 16).
- the probability of each position x ⁇ X of all the nodes p under the condition where the correspondence c i, j (m k ) (not shown in FIG.
- ⁇ i, j> means the combination of all i, j for which the correspondence c i, j (m k ) is defined, and C is a possible correspondence c i, j (m k ).
- M is the set of all measurement means used.
- equation (3) is the probability that the locus X will occur when the correspondence c i, j (m k ) occurs. For example, taking the measurement z j (m 1 ) related to the identification number m 1 of the measuring means in FIG.
- the correspondence between the nodes is ⁇ z 0 (m 1 ), z 1 ( m 1 ), z 2 (m 1 ), z 5 (m 1 ), z 6 (m 1 ), z 9 (m 1 ), z 10 (m 1 ), z 13 (m 1 ), z 14 (m 1 ) ⁇
- the probability density function multiplied by equation (3) is ⁇ p (x
- the description regarding m 1 is omitted.
- the evaluation function generation unit 21 maximizes the probability density function p (x) expressed by Equation (3) to obtain the trajectory Xc of the moving object V that has no accumulated error and maintains consistency throughout.
- the trajectory X indicates a trajectory before correction
- the trajectory Xc indicates a trajectory after correction. That is, Equation (3) probability density function p (x) represented by the fact that the greatest is that each position x j is occurring at the highest probability.
- the evaluation function generation unit 21 obtains each position x as described above, and obtains a trajectory Xc of the moving object V that has no accumulated error and maintains consistency throughout by connecting the obtained positions x as nodes.
- N (•) is a probability density function representing a normal distribution.
- the formula (4) is generated by the method described in each reference described later. This equation (4) is a probability distribution of difference measurement errors.
- the evaluation function generation unit 21 Based on the probability distribution of the difference measurement error, the evaluation function generation unit 21 generates an evaluation function (S106).
- the evaluation function generation unit 21 derives the evaluation function F (x) by the following procedure. First, when Expression (4) is substituted into Expression (3) and expanded according to the normal distribution formula, the following Expression (5) is derived.
- ⁇ ij is a normalization variable
- d ij (x) is a relative position difference (difference measurement Z ij ) of the measurement z j corresponding to the j-th node p j viewed from the i-th node p i.
- This is a function to obtain.
- This function is a function of x.
- ⁇ ij is an accuracy matrix in the measurement z j (m k ) having the correspondence c i, j .
- ⁇ ij is calculated by means such as a technique described in each reference described later. If the natural logarithm of both sides is calculated
- Equation (7) e ij is represented by the following equation (8).
- maximizing the probability density function p (x) shown in Expression (3) is equivalent to minimizing the evaluation function F (x) shown in Expression (7). Therefore, minimizing the evaluation function F (x) corresponds to optimizing the evaluation function F (x). That is, in step S106 of FIG. 2, the evaluation function generation unit 21 measures the measurement z j (m acquired by the surrounding object shape matching unit 12, the geographic information system matching unit 14, the GNSS positioning unit 15, and the wheel rotation amount measuring unit 16. k )), an evaluation function F (x) shown in Expression (7) is calculated.
- control unit 17 determines whether or not the traveling in the traveling environment targeted by the moving object V has been completed, that is, whether or not the traveling of the moving object V has been completed (S107), so that the peripheral object It is determined whether or not the shape data has been collected. If the result of step S107 is that the travel has not ended (S107 ⁇ No), the control unit 17 returns the process to step S101.
- step S107 when the traveling is finished (S107 ⁇ Yes), the trajectory optimization calculating unit 22 optimizes the evaluation function F (x) shown in the equation (7), thereby generating a cumulative error.
- the trajectory X of the moving object V is corrected to the trajectory Xc of the moving object V that has no accumulated error and maintains consistency throughout (S108).
- the trajectory optimization calculation unit 22 optimizes the evaluation function F (x) by obtaining the position x of the node p that minimizes the evaluation function F (x) by the following equation (9). Accordingly, the trajectory optimization calculation unit 22 uses the trajectory X (FIG. 6) acquired by the wheel rotation amount measurement unit 16 as a moving object V having no accumulated error and maintaining consistency as shown in FIG. 9. To the locus Xc. The trajectory optimization calculation unit 22 minimizes the evaluation function F (x) by a simultaneous equation solving method or a nonlinear optimization method.
- the shape map data generation unit 23 generates a map by attaching the shape data of the peripheral object measured by the peripheral object shape measurement unit 11 to the corrected trajectory Xc of the moving body V (S109).
- the shape map data generation unit 23 instead of generating a map while estimating its own position by sequential matching, all the shape data of surrounding objects in the driving environment for which the map is to be generated are collected and accumulated over the entire measurement data
- An accurate map can be generated by estimating the trajectory Xc of the moving object V having no error and maintaining consistency, and pasting the measured shape data of the surrounding objects to the trajectory Xc.
- the node p is generated from the data measured by the wheel rotation amount measuring unit 16, but the node p may be generated from the data measured by other measuring means. That is, data measured by other measuring means may be used as the reference value.
- FIG. 10 is a flowchart showing the procedure of the process of step S105 by the peripheral object shape matching unit.
- the peripheral object shape matching unit 12 has the shape data of the peripheral object measured by the peripheral object shape measuring unit 11 at the point of the i-th node p i for the i-th node p i and the j-th node p j ,
- the correspondence between the measurement z acquired by the surrounding object shape matching unit 12 and the nodes p i and p j is expressed as follows. calculate.
- the peripheral object shape measuring unit 11 searches for a node p that is within a predetermined threshold distance from the i-th node p i that is currently targeted (S201), and determines a corresponding j-th node p j .
- the peripheral object shape measuring unit 11 measures the peripheral object shape data measured by the peripheral object shape measuring unit 11 itself at the point of the i-th node p i and the peripheral object measured at the point of the j-th node p j. Are matched with each other (S202). The result of this matching is measurement z.
- matching for example, "Distance data processing-Shape model generation technology from multiple distance images" (author: Takeshi Masuda, Kyoko Okaya (Shimizu), Ritsumasa Sagawa, conference: Proc. Of the 146th CVIM, 2004) can be used.
- the peripheral object shape measurement unit 11 calculates a probability distribution (probability density function: equation (4)) of the difference measurement error in the measurement z j by using the matching result in step S202 (S203).
- the probability distribution of the difference measurement error in the measurement z j is, for example, the paper “Calculation of likelihood distribution of robot posture in scan matching” (author: Masahiro Tomono, conference: Proc. Of RSJ'10, held year: 2010) Can be used.
- the peripheral object shape measuring unit 11 determines whether or not the processing of steps S201 to S203 has been completed for all nodes in the travel environment that are targets of map generation (S204). If the result of step S204 is not complete (S204 ⁇ No), the peripheral object shape measuring unit 11 returns the process to step S201. If the result of step S204 is complete (S204 ⁇ Yes), the peripheral object shape measuring unit 11 ends the process.
- FIG. 11 is a flowchart showing the procedure of the process of step S105 by the geographic information system matching unit.
- the geographic information system matching unit 14 matches (superimposes) the shape data of the surrounding objects measured by the surrounding object shape measuring unit 11 at the target node p with reference to the map shape data obtained by the geographic information system data obtaining unit 13. ) To calculate the correspondence relationship between the map data of the geographic information system and the probability distribution of the difference measurement error (probability density function: equation (4)).
- geographical information system matching unit 14 searches for data of the geographic information system from the node p i which is currently the subject via the geographic information system data acquiring unit 13 within a distance of a predetermined threshold value (S301). Then, the geographic information system matching unit 14 uses the searched object data with the map shape data acquired by the geographic information system data acquisition unit 13 as a reference, and the shape of the peripheral object measured by the peripheral object shape measurement unit 11 at the target node p. Data is matched (S302). The result of this matching is measurement z. For the matching, a method similar to the processing in step S202 of FIG. 10 can be used.
- the geographic information system matching unit 14 calculates the correspondence relationship between the measurement z and the node p using the matching result in step S302, and calculates the probability distribution of the difference measurement error (probability density function: equation (4). ) Is calculated (S303).
- the calculation of the probability distribution of the difference measurement error can use the same method as in step S203 of FIG.
- FIG. 12 is a flowchart showing the procedure of the process of step S105 by the GNSS positioning unit.
- the GNSS positioning unit 15 calculates a correspondence c i, j with respect to a reference coordinate system such as a planar rectangular coordinate system, for example, using a positioning system.
- the GNSS positioning unit 15 measures the measurement z with respect to the reference coordinate system of the current target node p using GNSS (S401). Then, the GNSS positioning unit 15 calculates the correspondence relationship with the node p and the probability distribution of the error of the measurement z using the positioning result (measurement z) in step S401 (S402).
- the probability distribution of the error of the measurement z for example, information on the GST sentence in the NMEA-0183 format, which is a communication protocol used in GNSS, can be used.
- FIG. 13 is a flowchart showing the procedure of the process of step S105 by the wheel rotation amount measuring unit.
- the wheel rotation amount measuring unit 16 accumulates the wheel rotation amount, thereby obtaining a correspondence relationship c i, j when the node p at the current position with respect to the node p at which the moving object V was at the previous time is measured z. calculate.
- the wheel rotation amount measurement unit 16 sets the node p j at the current position as the measurement z with reference to the node p i at which the moving object V was at the previous time, and the difference measurement Z ij as the wheel rotation amount. Measurement is performed by accumulating (S501). At this time, using an inertial sensor or gyro sensor called IMU, the paper “Gyrodometry: A New Method for Combining Data from Gyros and Odometry in Mobile Robots” (Author: Johann Borenstein, and Liqiang Feng, Conference Meeting: Proc. Of The method described in ICRA '96, held year: 1996) can be used.
- the wheel rotation quantity measuring unit 16 uses the measurement result in the step S501, the node p j corresponding relation c i of the measurement z for, j and the probability distribution of the error in the difference measurement Z ij (probability density function: Equation (4)) is calculated (S502).
- the method described in the book “Vehicle” (author: Kimio Kanai et al., Publisher: Corona, publication year: 2003) can be used to calculate the probability density distribution of measurement z errors.
- FIG. 14 is a diagram illustrating another embodiment of the autonomous traveling system according to the present embodiment.
- one mobile unit V and the management unit 20 communicate with each other.
- the in-vehicle unit 10 mounted on a plurality of mobile units V1, V2 (V) Communication with the management unit 20 may be performed.
- the management unit 20 generates an evaluation function based on the measurement data collected from each in-vehicle unit 10 and obtains the trajectory Xc.
- each in-vehicle unit 10 may not have the respective units 11 to 15 (however, all the in-vehicle units 10 need to have the wheel rotation amount measuring unit 16).
- the management unit 20 may integrate the measurement data collected from each in-vehicle unit 10 to generate an evaluation function and obtain the trajectory Xc.
- FIG. 15 is a diagram showing another embodiment of the autonomous traveling system according to the present embodiment.
- the functions of the in-vehicle unit 10 and the management unit 20 in FIG. 1 may be provided in the in-vehicle unit 10 a of the moving body V. Thereby, it is possible to correct the trajectory and create a map only by the moving body Va.
- an evaluation function is generated based on data obtained from a plurality of measurement means, and the evaluation function is optimized, so that the trajectory X of the moving object V including a measurement error can be obtained throughout. It is possible to estimate the trajectory Xc of the moving object V that has no accumulated error and maintains consistency. Then, by pasting the shape data of the surrounding objects on the locus Xc, it is possible to generate an accurate map without accumulating errors. Thereby, the moving body V can reach the destination without losing sight of its own position or the target route by moving the moving body V autonomously based on such a map.
Landscapes
- Engineering & Computer Science (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Automation & Control Theory (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Control Of Position, Course, Altitude, Or Attitude Of Moving Bodies (AREA)
- Navigation (AREA)
- Instructional Devices (AREA)
Abstract
Description
特許文献2には、レーザスキャナを外界センサとして用いて、現在時刻に取得した周辺物体の形状データと、1つ前の時刻に現在とは異なる位置で取得した周辺物体の形状データとを逐次的にマッチング(重ね合わせ)することにより、周辺物体の形状データを測定した領域を拡大していくことで地図を生成する環境地図生成方法および移動ロボットが開示されている。
図1は、本実施形態に係る自律移動システムの構成例を示す図である。
自律移動システム1は、地図作成に必要な各データを収集する車載部10と、収集された各データを基に、軌跡補正を行う管理部20とを有する。
車載部10は、例えば、自律移動ロボットや乗物などの移動体装置である移動体Vに搭載されている。また、管理部20は、例えば、オフィスビルなどの管理設備に設置されている。
車載部10と、管理部20とは、無線ネットワークなどを介して通信可能である。
車載部10は、周辺物体形状測定部11、周辺物体形状マッチング部12、地理情報システムデータ取得部13、地理情報システムマッチング部14、GNSS(Global Navigation Satellite System)測位部15、車輪回転量測定部16、制御部17などを有している。なお、周辺物体形状マッチング部12、地理情報システムマッチング部14、GNSS測位部15、車輪回転量測定部16をまとめて、測定手段と称することがある。
管理部20は、評価関数生成部21、軌跡最適化計算部22、形状地図データ生成部23を有している。
次に、図1を参照しつつ、図2から図9に沿って本実施形態に係る自律移動システム1の具体的な処理内容について説明する。
図2は、本実施形態に係る自律移動システムの処理手順を示すフローチャートである。
自律移動システム1が図2に示すフローチャートの処理を実行することで正確な軌跡を算出し、さらにこの軌跡を基に地図を生成することで、移動体Vが自己位置や目標経路を見失うことなく目的地に到達することができる。
図3における走行可能領域(道路)412を走行している移動体Vを中心とし、周辺物体形状測定部11が、自身の測定範囲401に入る道路構造物411などの物体形状を測定する。
これは、例えば、周辺物体形状測定部11がレーザを周囲に放射し、その反射によって周囲の物体の形状を測定する。
この際、人や他の移動体Vなどの移動体Vを地図に記憶してしまうことを避けるため、周辺物体形状測定部11は、図4に示すように所定の高さの領域501の周辺物体の形状データを抽出してもよい。あるいは、周辺物体形状測定部11は平面や円柱などの、特徴的な周辺物体の形状データを抽出するようにしてもよい。
次に、制御部17が、測定のデータが所定量蓄積されたか否かを判定する(S103)。
ステップS103の結果、測定のデータが所定量蓄積されていない場合(S103→No)、制御部17はステップS101へ処理を戻す。
ステップS103の結果、測定のデータが所定量蓄積されている場合(S103→Yes)、送信されたデータを基に作製される移動体Vが走行した軌跡に対して、評価関数生成部21がノードとアークを生成する(S104)ことによって、グラフ構造を生成する。
走行軌跡tは移動体Vが実際に走行した際の走行軌跡の例を示す図である。
図5に示すように、移動体Vが実際の走行軌跡tに沿って走行しているものとする。走行軌跡tは、車輪回転量測定部16による車輪の回転数から測定される距離を基に算出されるものである。なお、走行軌跡tは実際には閉じている。
図6に示すように、評価関数生成部21は、走行軌跡t(図5)を所定の長さ毎に区切り、区切った地点をノードpとし、ノードp間を直線のアークgで繋ぐことで、グラフ構造を生成し、軌跡Xを生成する。
この際、評価関数生成部21は、車輪回転量測定部16が車輪回転量を累積することで算出した移動体Vの位置を初期位置として、ノードpとアークgを生成する。
車輪回転量測定部16の測定には累積誤差が発生するため、図6に示す評価関数生成部21が生成したグラフ構造(軌跡X)には、図5に示す実際の環境での走行軌跡tとは測定誤差(ずれ)が発生してしまっている。自律移動システム1は、この測定誤差(ずれ)を修正して正確な地図を生成するために、図2に示す以降のフローチャートの処理手順(S105~S109)を実行する。
このとき、ノードpとアークgで表される軌跡Xは、式(2)に示すようにn個のノードpiの位置を表すベクトルxiの集合として表現される。
図7は、本実施形態に係る各用語の定義を説明するための図である。
まず、図7(a)に示すように、車輪回転量測定部16が測定した位置として、ノードp1とノードp2を用いた例を示す。
ノードp1と、ノードp2とを結ぶ線をアークg12と表現する。ノードp2は、測定誤差を含んでいるので、その測定誤差は正規分布に従うと仮定すると、真のノードp2は楕円y1の範囲内に存在すると考えられる。ここで、ノードp2の測定誤差の分布(楕円分布y1)を正規分布の精度行列Ωによって表現される。ここで、楕円y1は、精度行列Ωが表す共分散楕円である。精度行列Ωは情報行列とも呼ばれ、正規分布の共分散行列の逆行列に相当する。具体的には、図7(a)における楕円y1は、正規分布において、ノードp2を中心とし、その中心から標準偏差σの範囲を示している。
なお、小さな矢印qは、移動体Vの姿勢の向きを示すものであり、式(1)のθiに相当する。
図7(b)に示すように、車輪回転量測定部16以外の測定手段(例えば、GNSS測位部15)がある位置を測定したとする。この位置を測定z(m1)(位置データ)と記載する。ここで、「m1」とは、測定zを測定した測定手段(例えば、GNSS測位部15)を示す識別番号である。つまり、測定z(m1)とは、測定手段m1によって測定された測定zを意味する。
そして、測定z(m1)が、どのノードpに対応付けられるのかを評価関数生成部21が判定する。例えば、時刻などを基に、測定z(m1)がノードp2に対応すると評価関数生成部21によって判定されたとすると、この測定zを測定z2(m1)と表現することとする。
ここで、ノードp1と測定z2(m1)とを結ぶ線を相対的な位置の差分の測定(以下、差分測定と称する)をZ12(m1)と表現することとする。差分測定Z12(m1)の成分は、測定z2と、ノードp1との差で表現される。
このように、車輪回転量測定部16によって測定されたノードpは、その他の測定手段によって測定された測定zとの対応関係の初期値、つまり基準値となる。
図7(c)のように、測定z4(m1)がノードp4と対応付けられ、ノードp2,p3と対応付けられる測定z(m1)が得られなかった場合、この測定z(m1)は、測定z4(m1)となり、差分測定は、ノードp1に対する差分測定はZ14(m1)となる。
例えば、図7(d)に示すように、測定zと対応付けられているノードpの所定個前のノードpが差分元とされてもよい。図7(d)の例では、2つ前のノードpが差分元となっている。つまり、測定z3(m1)に対して生成される差分測定(Z13)の差分元は、測定z3(m1)と対応付けられているノードp3の2つ前のノードpであるノードp1となっている。また、測定z4(m1)に対して生成される差分測定(Z24)の差分元は、測定z4(m1)と対応付けられているノードp4の2つ前のノードpであるノードp2となっている。
例えば、図7(e)に示すように、その前に測定zとが対応付けられたノードpが差分元とされてもよい。図7(e)に示されているように、測定z1(m1)とノードp1とが対応付けられ、測定z3(m1)とノードp3とが対応付けられ、測定z4(m1)とノードp4とが対応付けられているとする。ここで、測定z3(m1)に対して生成される差分測定(Z13)の差分元は、その前に測定zと対応付けられているノードp1となる。同様に、測定z4(m1)に対して生成される差分測定(Z34)の差分元は、その前に測定zと対応付けられているノードp3となる。そして、測定z4(m1)は、他の差分測定の差分元となる。
図7で説明した定義を基に、評価関数生成部21は軌跡Xを確定するための評価関数を算出し、この評価関数が最大となる各ノードpの各位置xを算出することにより、最も尤もらしい軌跡Xcを確定する。
図8は、軌跡確定を説明するための図である。なお、図8における軌跡は、図6における軌跡Xと同じものである。
図8では、ノードp0,p1,p2,・・・が算出されており、それに対応する測定z0(m1),z1(m1),z2(m1),z5(m1),・・・が算出されている(図8では(m1)は省略)。ここで、差分元のノードは、対応関係が定義されたノードから3つ前のノードとしている。
なお、前記したようにノードp0,p1,p2,・・・も、それぞれ測定z0(m0),z1(m0),z2(m0)とすることができるが、煩雑になるのを避けるため、ここでは図示省略してある(「m0」は車輪回転量測定部16を示す識別番号)。
ここで、対応関係ci,j(mk)(図8において図示せず)が定義されている条件下におけるすべてのノードpの各位置x∈Xの確率を確率密度関数p(x|ci,j(mk))で表すと、対応関係ci,j(mk)は独立に生起すると見なされるため、位置xが生起する確率は、対応関係ci,j(mk)が生起する条件におけるxj(ノードpjの位置)が生起する確率の混合分布で表現され、その式は式(3)で示される確率密度関数p(x)で示される。ここで、xは式(2)で定義されるxである。
式(3)が意味するところは、対応関係ci,j(mk)が生じたときに軌跡Xが生起する確率である。
例えば、図8における測定手段の識別番号m1に関する測定zj(m1)を例にすると、ノードとの間で対応関係がとれているのは、{z0(m1),z1(m1),z2(m1),z5(m1),z6(m1),z9(m1),z10(m1),z13(m1),z14(m1)}であるので、式(3)にて乗算される確率密度関数は、{p(x|c16,0(m1)),p(x|c14,1(m1)),p(x|c15,2(m1)),p(x|c2,5(m1)),p(x|c3,6(m1)),p(x|c6,9(m1)),p(x|c7,10(m1)),p(x|c10,13(m1)),p(x|c11,14(m1))}となる。なお、図8ではm1に関する記載を省略している。
図7(b)で説明したように、軌跡Xにおける各ノードpの位置xが存在する確率的位置を、精度行列Ωの正規分布で表現すると、式(3)における確率密度関数p(x|ci,j(mk))は、以下の式(4)で示される。
評価関数生成部21は、評価関数F(x)を以下のような手順で導出する。
まず、式(3)に式(4)を代入して正規分布の公式に従って展開すると、以下の式(5)が導かれる。
式(5)に関して両辺の自然対数を求めると、式(6)が導かれる。
すなわち、図2のステップS106において、評価関数生成部21は、周辺物体形状マッチング部12、地理情報システムマッチング部14、GNSS測位部15、車輪回転量測定部16によって取得された測定zj(mk)を基に、式(7)に示す評価関数F(x)を算出する。
ステップS107の結果、走行終了していない場合(S107→No)、制御部17は、ステップS101へ処理を戻す。
次に、図10~図15を参照して、図2のステップS105における各測定手段の処理を説明する。
図10は、周辺物体形状マッチング部によるステップS105の処理の手順を示すフローチャートである。
周辺物体形状マッチング部12は、i番目のノードpiとj番目のノードpjに対して、i番目のノードpiの地点で周辺物体形状測定部11が測定した周辺物体の形状データと、j番目のノードpjの地点で測定した周辺物体の形状データをマッチング(重ね合わせ)することにより、周辺物体形状マッチング部12が取得した測定zと、ノードpi,pjとの対応関係を算出する。
ステップS204の結果、完了していない場合(S204→No)、周辺物体形状測定部11はステップS201へ処理を戻す。
ステップS204の結果、完了している場合(S204→Yes)、周辺物体形状測定部11は処理を終了する。
図11は、地理情報システムマッチング部によるステップS105の処理の手順を示すフローチャートである。
地理情報システムマッチング部14は、地理情報システムデータ取得部13が取得した地図形状データを基準として、対象とするノードpで周辺物体形状測定部11が測定した周辺物体の形状データをマッチング(重ね合わせ)することにより、地理情報システムが有する地図データに対するノードとの対応関係および差分測定の誤差の確率分布(確率密度関数:式(4))を算出する。
そして、地理情報システムマッチング部14は、探索したデータを地理情報システムデータ取得部13が取得した地図形状データを基準として、対象とするノードpで周辺物体形状測定部11が測定した周辺物体の形状データをマッチングする(S302)。このマッチングの結果が測定zとなる。マッチングには、図10のステップS202の処理と同様の方法を用いることができる。
図12は、GNSS測位部によるステップS105の処理の手順を示すフローチャートである。
GNSS測位部15は、測位システムを用いて、例えば、平面直角座標系などの基準座標系に対する対応関係ci,jを算出する。
そして、GNSS測位部15は、ステップS401での測位結果(測定z)を用いて、ノードpに対する対応関係、および、測定zの誤差の確率分布を算出する(S402)。測定zの誤差の確率分布は、例えば、GNSSで用いられる通信プロトコルであるNMEA‐0183フォーマットのGSTセンテンスの情報などを用いることができる。
図13は、車輪回転量測定部によるステップS105の処理の手順を示すフローチャートである。
車輪回転量測定部16は、車輪回転量を累積することで、移動体Vが1つ前の時刻にいたノードpに対する現在位置のノードpを測定zとしたときの対応関係ci,jを算出する。
図1では、1台の移動体Vと、管理部20とが通信を行っているが、図14に示すように複数台の移動体V1,V2(V)に搭載された車載部10が、管理部20と通信を行うようにしてもよい。この場合、管理部20は、各々の車載部10から収集した測定のデータを基に、評価関数を生成し、軌跡Xcを求める。
あるいは、各々の車載部10が、各部11~15を有していなくてもよい(ただし、車輪回転量測定部16は、すべての車載部10が有する必要がある)。この場合、管理部20は、各々の車載部10から収集した測定のデータを統合して、評価関数を生成し、軌跡Xcを求めるようにしてもよい。
図15に示すように、移動体Vの車載部10aに図1の車載部10および管理部20の機能が設けられてもよい。
これにより、移動体Vaのみで軌跡の補正および地図の作成が可能となる。
本実施形態によれば、複数の測定手段から得られたデータを基に、評価関数を生成し、その評価関数を最適化することにより、測定誤差を含む移動体Vの軌跡Xを、全体にわたって累積誤差がなく整合性を保った移動体Vの軌跡Xcを推定することができる。そして、その軌跡Xcに、周辺物体の形状データを貼付することで、誤差が累積しない正確な地図の生成が可能となる。
これにより、このような地図を基に、移動体Vが自律移動することで、移動体Vは自己位置や目標経路を見失うことなく目的地に到達することができる。
10,10a 車載部
11 周辺物体形状測定部
12 周辺物体形状マッチング部
13 地理情報システムデータ取得部
14 地理情報システムマッチング部
15 GNSS測位部
16 車輪回転量測定部
17 制御部
20 管理部(軌跡補正装置)
21 評価関数生成部
22 軌跡最適化計算部
23 形状地図データ生成部
V,V1,V2,Va 移動体(移動体装置)
Claims (5)
- 移動体が走行した軌跡を補正する軌跡補正装置による軌跡補正方法であって、
前記軌跡補正装置は、
1の測定手段が取得した前記移動体の軌跡データに、複数のノードを設定し、
前記1の測定手段が取得した前記移動体の位置データを、前記ノードに対応付けるとともに、前記1の測定手段とは別の他の測定手段が取得した前記移動体の位置データを、前記ノードに対応付け、
前記ノードが生起する可能性がある位置を確率で示し、
前記ノードに対応付けられた位置データが生起する可能性がある位置を確率で示し、
各確率を基に、前記ノードおよび前記位置データを変数として含む評価関数を算出し、
前記評価関数を基に、各ノードが生起する確率が最も大きい軌跡を算出する
ことを特徴とする軌跡補正方法。 - 前記確率は、確率密度関数で示されており、
前記評価関数は、尤度関数であり、
前記軌跡補正装置は、
最尤推定法によって、前記各ノードが生起する確率が最も大きい軌跡を算出する
ことを特徴とする請求の範囲第1項に記載の軌跡補正方法。 - 前記軌跡補正装置は、
前記算出された軌跡の周囲に、構造物の形状データを貼付することにより、地図データを作成する
ことを特徴とする請求の範囲第1項または請求の範囲第2項に記載の軌跡補正方法。 - 移動体が走行した軌跡を補正する軌跡補正装置であって、
1の測定手段が取得した前記移動体の軌跡データに、複数のノードを設定し、
前記1の測定手段が取得した前記移動体の位置データを、前記ノードに対応付けるとともに、前記1の測定手段とは別の他の測定手段が取得した前記移動体の位置データを、前記ノードに対応付け、
前記ノードが生起する可能性がある位置を確率で示し、
前記ノードに対応付けられた位置データが生起する可能性がある位置を確率で示し、
各確率を基に、前記ノードおよび前記位置データを変数として含む評価関数を算出する評価関数生成部と、
前記評価関数を基に、各ノードが生起する確率が最も大きい軌跡を算出する軌跡最適化計算部と、
を有することを特徴とする軌跡補正装置。 - 請求の範囲第4項に記載の軌跡補正装置を搭載している
ことを特徴とする移動体装置。
Priority Applications (4)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
DE112011105210.3T DE112011105210T5 (de) | 2011-05-20 | 2011-05-20 | Ortskurvenkorrekturverfahren, Ortskurvenkorrekturvorrichtung und mobiles Objektgerät |
JP2013516092A JP5852645B2 (ja) | 2011-05-20 | 2011-05-20 | 軌跡補正方法、軌跡補正装置および移動体装置 |
US14/118,899 US9182235B2 (en) | 2011-05-20 | 2011-05-20 | Locus correcting method, locus correcting apparatus, and mobile object equipment |
PCT/JP2011/061685 WO2012160630A1 (ja) | 2011-05-20 | 2011-05-20 | 軌跡補正方法、軌跡補正装置および移動体装置 |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
PCT/JP2011/061685 WO2012160630A1 (ja) | 2011-05-20 | 2011-05-20 | 軌跡補正方法、軌跡補正装置および移動体装置 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2012160630A1 true WO2012160630A1 (ja) | 2012-11-29 |
Family
ID=47216735
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/JP2011/061685 WO2012160630A1 (ja) | 2011-05-20 | 2011-05-20 | 軌跡補正方法、軌跡補正装置および移動体装置 |
Country Status (4)
Country | Link |
---|---|
US (1) | US9182235B2 (ja) |
JP (1) | JP5852645B2 (ja) |
DE (1) | DE112011105210T5 (ja) |
WO (1) | WO2012160630A1 (ja) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2014122894A (ja) * | 2012-12-19 | 2014-07-03 | Toyota Motor Engineering & Manufacturing North America Inc | 車両のための装置及び方法、及び、その方法を実行するための命令を含む記憶媒体 |
JP2018504650A (ja) * | 2014-12-26 | 2018-02-15 | ヘーレ グローバル ベスローテン フェンノートシャップ | 装置の位置特定のための幾何学的指紋法 |
CN108303075A (zh) * | 2017-12-29 | 2018-07-20 | 广州斯马特信息科技有限公司 | 轨迹生成方法和系统 |
Families Citing this family (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP6878045B2 (ja) * | 2017-02-28 | 2021-05-26 | 国立研究開発法人理化学研究所 | 点群データの抽出方法、及び点群データの抽出装置 |
DE102017212603A1 (de) * | 2017-07-21 | 2019-01-24 | Robert Bosch Gmbh | Verfahren zum Bereitstellen und zum Verbessern einer Positionswahrscheinlichkeitsverteilung für GNSS-Empfangsdaten |
KR102063534B1 (ko) * | 2017-11-30 | 2020-01-09 | 주식회사 모빌테크 | 라이다를 이용한 지도 생성 방법 |
JP6974189B2 (ja) * | 2018-01-16 | 2021-12-01 | 株式会社豊田中央研究所 | 地図作成装置 |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2002357437A (ja) * | 1984-06-07 | 2002-12-13 | Tele Atlas North America Inc | 移動体追跡情報提供装置 |
JP2008002906A (ja) * | 2006-06-21 | 2008-01-10 | Toyota Motor Corp | 測位装置 |
JP2008026282A (ja) * | 2006-07-25 | 2008-02-07 | Toyota Motor Corp | 測位装置、ナビゲーションシステム |
JP2008175786A (ja) * | 2007-01-22 | 2008-07-31 | Zhencheng Hu | 移動体位置検出方法および移動体位置検出装置 |
Family Cites Families (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP0779998B1 (de) * | 1994-09-06 | 1998-03-25 | Siemens Aktiengesellschaft | Verfahren zur bestimmung der position einer landmarke in der umgebungskarte einer selbstbeweglichen einheit, deren abstand zur einheit dynamisch von dieser ermittelt wird |
JP2004110802A (ja) | 2002-08-26 | 2004-04-08 | Sony Corp | 環境同定装置、環境同定方法、プログラム及び記録媒体、並びにロボット装置 |
JP2008076348A (ja) | 2006-09-25 | 2008-04-03 | Dainippon Toryo Co Ltd | 塗料中の六価クロムの分析方法 |
JP4788722B2 (ja) * | 2008-02-26 | 2011-10-05 | トヨタ自動車株式会社 | 自律移動ロボット、自己位置推定方法、環境地図の生成方法、環境地図の生成装置、及び環境地図のデータ構造 |
-
2011
- 2011-05-20 JP JP2013516092A patent/JP5852645B2/ja active Active
- 2011-05-20 DE DE112011105210.3T patent/DE112011105210T5/de active Pending
- 2011-05-20 US US14/118,899 patent/US9182235B2/en active Active
- 2011-05-20 WO PCT/JP2011/061685 patent/WO2012160630A1/ja active Application Filing
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2002357437A (ja) * | 1984-06-07 | 2002-12-13 | Tele Atlas North America Inc | 移動体追跡情報提供装置 |
JP2008002906A (ja) * | 2006-06-21 | 2008-01-10 | Toyota Motor Corp | 測位装置 |
JP2008026282A (ja) * | 2006-07-25 | 2008-02-07 | Toyota Motor Corp | 測位装置、ナビゲーションシステム |
JP2008175786A (ja) * | 2007-01-22 | 2008-07-31 | Zhencheng Hu | 移動体位置検出方法および移動体位置検出装置 |
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2014122894A (ja) * | 2012-12-19 | 2014-07-03 | Toyota Motor Engineering & Manufacturing North America Inc | 車両のための装置及び方法、及び、その方法を実行するための命令を含む記憶媒体 |
JP2018504650A (ja) * | 2014-12-26 | 2018-02-15 | ヘーレ グローバル ベスローテン フェンノートシャップ | 装置の位置特定のための幾何学的指紋法 |
US10145956B2 (en) | 2014-12-26 | 2018-12-04 | Here Global B.V. | Geometric fingerprinting for localization of a device |
CN108303075A (zh) * | 2017-12-29 | 2018-07-20 | 广州斯马特信息科技有限公司 | 轨迹生成方法和系统 |
Also Published As
Publication number | Publication date |
---|---|
DE112011105210T5 (de) | 2014-05-28 |
JP5852645B2 (ja) | 2016-02-03 |
JPWO2012160630A1 (ja) | 2014-07-31 |
US9182235B2 (en) | 2015-11-10 |
US20140088863A1 (en) | 2014-03-27 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
JP5852645B2 (ja) | 軌跡補正方法、軌跡補正装置および移動体装置 | |
KR102425272B1 (ko) | 디지털 지도에 대한 위치를 판별하기 위한 방법 및 시스템 | |
US11726208B2 (en) | Autonomous vehicle localization using a Lidar intensity map | |
US20190073794A1 (en) | Survey data processing device, survey data processing method, and survey data processing program | |
JP5927735B2 (ja) | 地図データ作成装置、自律移動システムおよび自律移動制御装置 | |
US9098088B2 (en) | Method for building outdoor map for moving object and apparatus thereof | |
WO2012086029A1 (ja) | 自律移動システム | |
US20230083965A1 (en) | Map construction method, apparatus and storage medium | |
CN105044668A (zh) | 一种基于多传感器装置的wifi指纹数据库构建方法 | |
CN110887493B (zh) | 基于局部地图匹配的轨迹推算方法、介质、终端和装置 | |
US20200364883A1 (en) | Localization of a mobile unit by means of a multi-hypothesis kalman filter method | |
US11798225B2 (en) | 3D building generation using topology | |
JP2017090239A (ja) | 情報処理装置、制御方法、プログラム及び記憶媒体 | |
WO2024027350A1 (zh) | 车辆定位方法、装置、计算机设备、存储介质 | |
US11578991B2 (en) | Method and system for generating and updating digital maps | |
US11670169B2 (en) | Systems and methods for distributed cooperative localization in a connected vehicular platform | |
WO2017175433A1 (ja) | 地図データ生成システムおよび地図データ生成方法 | |
CN112487234B (zh) | 基于分裂型层次聚类的轨迹分析方法、设备及存储介质 | |
JP2019174191A (ja) | データ構造、情報送信装置、制御方法、プログラム及び記憶媒体 | |
EP3663973A1 (en) | Automatic detection and positioning of structure faces | |
KR102506411B1 (ko) | 차량의 위치 및 자세 추정 방법, 장치 및 이를 위한 기록매체 | |
CN112747752A (zh) | 基于激光里程计的车辆定位方法、装置、设备和存储介质 | |
WO2019188874A1 (ja) | データ構造、情報処理装置、及び地図データ生成装置 | |
Tang et al. | Enhanced indoor positioning through human-robot collaboration | |
CN118244785B (zh) | 空地双模的两栖无人机及其定位方法、装置、存储介质 |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 11865972 Country of ref document: EP Kind code of ref document: A1 |
|
ENP | Entry into the national phase |
Ref document number: 2013516092 Country of ref document: JP Kind code of ref document: A |
|
WWE | Wipo information: entry into national phase |
Ref document number: 14118899 Country of ref document: US Ref document number: 1120111052103 Country of ref document: DE Ref document number: 112011105210 Country of ref document: DE |
|
122 | Ep: pct application non-entry in european phase |
Ref document number: 11865972 Country of ref document: EP Kind code of ref document: A1 |