Carrier, distance calculation method, and distance calculation program

Abstract

To provide a transport vehicle capable of detecting the position of a surrounding object without detecting the shape itself.SOLUTION: A transport vehicle includes a point group acquisition unit and an edge identification unit. The point group acquisition unit acquires a point group PG by horizontally irradiating cargo loaded on a cargo loading unit and an object around the transport vehicle with light. The edge identification unit analyzes the acquired point group PG using a frequency distribution with axes representing respectively distances in lateral and longitudinal directions, and identifies sections S1, S2, S3, and S4 with frequencies, adjacent to an area with substantially no frequency, as positions of edges in the lateral direction or the longitudinal direction of the cargo and the object around.SELECTED DRAWING: Figure 6

Description

Technical Field

[0001] The present invention relates to a transport vehicle, an edge position identification method, a distance calculation method, and a position identification program.

Background Art

[0002] Conventionally, as disclosed in Patent Document 1, there is an unmanned transport vehicle that autonomously travels and performs cargo handling. The unmanned transport vehicle disclosed in Patent Document 1 includes a fork, a lifting device for lifting and lowering the fork, and a laser scanner for detecting the position of the vehicle itself. The unmanned transport vehicle is configured to move to a predetermined cargo handling position while detecting its own position, and perform cargo handling work by lifting and lowering the fork.

[0003] By the way, as disclosed in Patent Document 1, the unmanned transport vehicle may perform cargo handling on a moving shelf. The moving shelf moves unlike a fixed shelf, but may deviate from a predetermined moving position during this movement. Then, a deviation will occur between the predetermined cargo handling position and the moving shelf, but the unmanned transport system of Patent Document 1 does not consider this deviation. Also, when performing cargo handling work on a truck parked at a predetermined position, the truck may deviate from the predetermined standby position, and in this case as well, a deviation will occur from the predetermined cargo handling position. Assuming that the cargo handling position is determined assuming that the moving shelf or the truck deviates, there is a problem that it is impossible to stack the loads tightly between the loads.

[0004] Therefore, in order to detect the load etc. loaded on a moving shelf or a truck that has deviated from a predetermined position, for example, the transport vehicle disclosed in Patent Document 2 is equipped with a LiDAR sensor (external sensor), and based on the information obtained by this sensor, it is configured to detect the shape of the pallet and the load to be loaded. However, such a conventional method of detecting the shape of the pallet or the load has high flexibility with respect to the position and orientation (angle) of the object, but depends on the shape and dimensions of the detection target. Therefore, the conventional method has a problem of lacking versatility in that it is necessary to apply different characteristic values and methods according to the detection target.

Prior Art Documents

Patent Document

[0005]

Patent Document 1

Patent Document 2

Summary of the Invention

Problems to be Solved by the Invention

[0006] Therefore, the problem to be solved by the present invention is to provide a carrier that can detect the positions of the edges of surrounding objects without detecting the shape itself.

Means for Solving the Problems

[0007] In order to solve the above problems, the carrier according to the present invention includes: a load loading part, a point cloud acquisition part that irradiates light horizontally on the load loaded on the load loading part and an object around the carrier or any of them to acquire a point cloud, an edge identification part that analyzes the acquired point cloud using a frequency distribution with the distance in the left-right and front-back or any one of the directions as the axis, and identifies an interval with a frequency adjacent to a region without a frequency as the position of the edge of the load and the surrounding object or any of them in the left-right or front-back direction.

[0008] Preferably, the above carrier further includes a distance calculation part, the point cloud acquisition part further irradiates light horizontally on the load and an object near the load to acquire a point cloud, the edge identification part further analyzes the acquired point cloud using a frequency distribution with the distance in the left-right direction as the axis, and respectively identifies the left and right intervals with a frequency adjacent to a region substantially without a frequency as the positions of the edges of the load or the object near it. The distance calculation unit calculates the left-right distance between the load and the nearby object based on the positions of the identified load and the edges of the nearby object. Note that "substantially frequency-free" in the present invention means excluding the case where there is a frequency in an area where there is nothing due to noise or the like. The edge identification unit may analyze by deleting the frequency due to noise or the like or ignoring a small frequency according to a known technique.

[0009] The above-described transport vehicle preferably further includes a distance calculation unit, The point cloud acquisition unit further irradiates light horizontally onto the load loaded on the load-carrying part and an object near the load to acquire a point cloud, The edge identification unit further analyzes the acquired point cloud using a frequency distribution with the distance in the front-rear direction as the axis, and specifies the front and rear sections with frequencies adjacent to the region substantially free of frequencies as the positions of the edges of the load or the nearby object, respectively. The distance calculation unit calculates the front-rear distance between the load and the nearby object based on the positions of the identified load and the edges of the nearby object.

[0010] In order to solve the above problems, an edge position identification method according to the present invention is a method for identifying the position of an edge of an object based on a point cloud acquired by a point cloud acquisition unit disposed on a transport vehicle, a step of irradiating light horizontally onto the object by the point cloud acquisition unit to acquire a point cloud, analyzing the acquired point cloud using a frequency distribution with the distance in the left-right and front-rear directions or either direction as the axis, and specifying the section with a frequency adjacent to the region substantially free of frequencies as the position of the edge of the object in the left-right and front-rear directions or either direction,

[0011] In order to solve the above problems, a gap distance calculation method according to the present invention is a method for calculating the gap distance between a load loaded on a transport vehicle and an object near the load based on a point cloud acquired by a point cloud acquisition unit disposed on the transport vehicle, a step of irradiating light horizontally onto the load and the object by the point cloud acquisition unit to acquire a point cloud, Analyzing the acquired point cloud using a frequency distribution with the distance in the left - right direction as the axis, and specifying an interval with frequency adjacent to a region substantially without frequency as the position of the edge of the load or the object; Based on the positions of the edges of the identified load and object, calculating the distance in the left - right direction of the gap between the load and the object.

[0012] To solve the above problems, the distance calculation method according to the present invention is A method for calculating the distance of the gap between a load loaded on a carrier vehicle and an object near the load based on the point cloud acquired by a point cloud acquisition unit arranged on the carrier vehicle, The step of horizontally irradiating light on the load and the object by the point cloud acquisition unit to acquire a point cloud; Analyzing the acquired point cloud using a frequency distribution with the distance in the front - rear direction as the axis, and specifying an interval with frequency adjacent to a region substantially without frequency as the position of the edge of the load or the object; Based on the positions of the edges of the identified load and object, calculating the distance in the front - rear direction of the gap between the load and the object.

[0013] To solve the above problems, the distance calculation method according to the present invention is A method for calculating the distance of the gap between a load loaded on a carrier vehicle and an object near the load based on the point cloud acquired by a point cloud acquisition unit arranged on the carrier vehicle, The step of horizontally irradiating light on the load and the object by the point cloud acquisition unit to acquire a point cloud; Analyzing the acquired point cloud using a frequency distribution with the distance in the front - rear direction as the axis, specifying the interval of the peak value in the upper region as the edge of the object, and specifying the interval of the peak value in the lower region as the position of the edge of the load; Based on the positions of the edges of the identified load and object, calculating the distance in the front - rear direction between the load and the object.

[0014] To solve the above problems, the position specification program according to the present invention is A point cloud acquisition unit that horizontally irradiates an object with light and acquires a point cloud, a computer, and causes a computer of a carrier vehicle equipped with the computer to analyze the acquired point cloud using a frequency distribution with distances in the left-right and front-back directions or any one of the directions as axes, and execute a step of specifying an interval with a frequency adjacent to a region substantially without a frequency as the position of an edge in the left-right or front-back direction of the object.

Effect of the Invention

[0015] The carrier vehicle according to the present invention can detect the position of the edge of a surrounding object without detecting the shape itself.

Brief Description of the Drawings

[0016]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

Figure 12

Mode for Carrying Out the Invention

[0017] Hereinafter, with reference to the accompanying drawings, an embodiment of the carrier vehicle, edge position specifying method, distance calculation method, and position specifying program of the present invention will be described. The double arrow X in the figure indicates the left-right direction, the double arrow Y indicates the front-back direction, and the double arrow Z indicates the up-down direction.

[0018] FIG. 1 is a side view of the carrier vehicle 1 according to this embodiment, and FIG. 2 is a functional block diagram of the control unit 30. The carrier vehicle 1 according to this embodiment is an unmanned carrier vehicle that autonomously travels and performs loading and unloading, but this is merely an example, and the carrier vehicle 1 according to the present invention is not limited thereto. For example, the carrier vehicle 1 may be a carrier vehicle that can be used for both manned and unmanned operations.

[0019] As shown in FIGS. 1 and 2, the carrier vehicle 1 includes a plurality of wheels 10, a vehicle body 11, a drive unit 12, a laser scanner 13, left and right masts 14, a lift bracket 15, left and right forks 16, a lifting unit 17, a backrest 18, a side shift unit 19, left and right carriages 20, left and right reach legs 21, left and right two-dimensional LiDAR sensors 22, left and right connecting parts 23, and a control unit 30. The carrier vehicle 1 is a reach-type forklift, but this is merely an example, and the carrier vehicle 1 according to the present invention may be a counterbalanced forklift.

[0020] The vehicle body 11 is disposed on the wheels 10, and the drive unit 12 is disposed inside the vehicle body 11. The drive unit 12 is configured to rotate or stop the wheels 10.

[0021] The laser scanner 13 is disposed above the vehicle body 11 and rotates horizontally to irradiate a laser. Then, the laser scanner 13 identifies the current position of the carrier vehicle 1 by identifying the position of a reflector disposed in the facility by scanning the reflected light of the laser.

[0022] The left and right masts 14 extend vertically and are disposed in front of the vehicle body 11. The lift bracket 15 has a finger bar for fixing the left and right forks 16 and is configured to be lifted and lowered along the left and right masts 14 by the lifting unit 17. The left and right forks 16 correspond to the "load loading part" of the present invention. In this embodiment, the number of forks 16 is four, but it may be two or six and is not particularly limited. By including four forks 16, the carrier vehicle 1 can simultaneously lift two pallets (loads).

[0023] The backrest 18 is formed in a frame shape, extends vertically and horizontally, and is configured to receive the loaded load W1. Note that only the outer frame of the backrest 18 shown in FIGS. 3 and 5 is illustrated, and this outer frame is disposed outside the forks 16 in the left and right directions.

[0024] The side shift unit 19 has an actuator and is configured to move the backrest 18 together with the fork 16 in the left - right direction by the actuator. Thereby, the side shift unit 19 can adjust the left - right position of the fork 16 with respect to the fork insertion hole of the pallet or adjust the position where the load W1 is loaded. The actuator may be a hydraulic actuator or an electric actuator and is not particularly limited.

[0025] The left - and - right carriages 20 are respectively provided outside the left - and - right masts 14, and the left - and - right reach legs 21 extend forward from the vehicle body 11. Guides for guiding the carriage 20 are provided inside the left - and - right reach legs 21, and the mast 14 is moved to the forward position or the backward position together with the carriage 20 by a reach cylinder (not shown).

[0026] The left - and - right 2D LiDAR sensors 22 are composed of laser scanners and are configured to irradiate a laser while rotating in the horizontal direction and scan the reflected light of the laser to obtain the distance to objects around the 2D LiDAR sensor 22 as a point cloud PG. The 2D LiDAR sensor 22 corresponds to the "point cloud acquisition unit" of the present invention. The point cloud acquisition unit may be, for example, a 3D LiDAR sensor or a 3D ToF (Time of Flight) camera instead of the 2D LiDAR sensor 22 and is not limited to the 2D LiDAR sensor.

[0027] As shown in FIGS. 1 and 3, the left - and - right connecting parts 23 have a first end part 23a, an intermediate part 23b, and a second end part 23c.

[0028] The first end part 23a is fixed to the left - and - right ends of the backrest 18, and the intermediate part 23b extends obliquely backward from the first end part 23a in a plan view of the backrest 18. The second end part 23c has a horizontal plane continuous with the intermediate part 23b and supports the 2D LiDAR sensor 22 by the horizontal plane.

[0029] The length of the middle portion 23b is configured such that the two-dimensional LiDAR sensor 22 supported by the second end portion 23c is positioned outside the side surface of the load loaded on the fork 16. That is, if the width of the backrest 18 is narrow and the load protrudes left and right beyond the backrest 18, the length of the middle portion 23b will be configured to be long accordingly.

[0030] FIG. 4 is a plan view showing the laser irradiation range LE of the two-dimensional LiDAR sensor 22, and FIG. 5 is a perspective view showing the laser irradiation range LE of the two-dimensional LiDAR sensor 22. Further, FIGS. 4 and 5 show the load W1 loaded on the fork 16 and the load W2 loaded adjacent to the load loading position P in front thereof. The load loading position P is, for example, a predetermined loading position of a moving shelf included in the cargo handling plan, a predetermined loading position of the cargo bed of the truck T, or the like.

[0031] As shown in FIGS. 4 and 5, the two-dimensional LiDAR sensor 22 is arranged at a position where it can horizontally irradiate the laser to the load W1 loaded on the fork 16 and the load loading position P. Then, the two-dimensional LiDAR sensor 22 irradiates the laser while rotating horizontally and receives the reflected light, thereby obtaining the distance to the object for each irradiation angle. This distance data is obtained as a point cloud PG.

[0032] FIG. 6A is a diagram showing the point cloud PG obtained by the two-dimensional LiDAR sensor 22 on the left side. The X-axis in FIGS. 6A and 6B indicates the distance in the left-right direction, the Y-axis in FIGS. 6A and 6C indicates the distance in the front-rear direction, and the intersection (origin) of the X-axis and the Y-axis indicates the position of the two-dimensional LiDAR sensor 22. Also, the point cloud PG in the attached drawings is an image diagram for showing an example of the point cloud PG to be obtained, and is not the actually obtained point cloud PG. As shown in FIG. 6A, the point cloud PG is obtained along the end faces of the load W1 loaded on the fork 16 and the load W2 loaded adjacent to the load loading position P.

[0033] As shown in FIG. 1, the control unit 30 is disposed inside the vehicle body 11. The control unit 30 is constituted by a computer having a storage device, an arithmetic unit, and a memory. The storage device stores a position identification program for causing the computer to execute an edge position identification method and a distance calculation method.

[0034] As shown in FIG. 2, the control unit 30 includes a storage unit 32, a travel control unit 34, an edge identification unit 35, a distance calculation unit 36, a load position identification unit 37, a lifting control unit 38, a side shift control unit 40, and a side shift stop unit 41.

[0035] The storage unit 32 stores a handling schedule, and the handling schedule includes a load loading position P. The storage unit 32 also includes the positions of the left and right two-dimensional LiDAR sensors 22 and the distance from the retracted position to the advanced position of the mast 14.

[0036] The travel control unit 34 is configured to control the drive unit 12, and causes the transport vehicle 1 to travel to the load loading position P with reference to the load loading position P stored in the storage unit 32 and the current position acquired by the laser scanner 13.

[0037] As will be described later, the edge identification unit 35 analyzes the acquired point group PG using a frequency distribution with the distances in the front-rear and left-right directions as the X-axis and Y-axis, respectively, and identifies the intervals with frequencies adjacent to the regions substantially without frequencies as the positions of the edges of the load W1 or the surrounding objects (for example, the load W2) in the left-right and front-rear directions. As described above, in the present invention, "substantially without frequencies" means excluding the case where there is a frequency in a region where there is nothing due to noise or the like. The edge identification unit 35 may analyze by deleting the frequencies due to noise or the like or ignoring the frequencies with a small number by a known technique. Hereinafter, the description of "substantially without frequencies" will be omitted as the description of "without frequencies".

[0038] As will be described later, the distance calculation unit 36 calculates the distances in the left-right and front-back directions between the load W1 and the object (e.g., load W2) based on the positions of the edges of the load W1 and the object (e.g., load W2) in the left-right, front-back directions specified by the edge specification unit 35.

[0039] Based on the distance D2 in the front-back direction between the load W1 and the load W2 calculated by the distance calculation unit 36 and the distance from the retracted position to the advanced position of the mast 14, the travel control unit 34 may calculate the forward distance required for unloading, and advance the transport vehicle 1 based on the calculated distance.

[0040] The load position specification unit 37 specifies the positions of the edges of the load W1 (W3) in the left-right and front-back directions specified by the edge specification unit 35 as the side position and the front position on either the left or right side of the load W1 (W3). Further, the load position specification unit 37 calculates the position at the center in the left-right direction of the load W1 (W3) based on the specified side positions on the left and right of the load W1 (W3). In addition, the load position specification unit 37 detects a change in the positional relationship between the two-dimensional LiDAR sensor 22 and the load W1 (W3) when the fork 16 is moving by the side shift unit 19 and when the fork 16 is pulled out from the load.

[0041] The lifting control unit 38 is configured to control the lifting unit 17, and raises and lowers the fork 16 by the lifting unit 17 based on the load loading position P stored in the storage unit 32.

[0042] The side shift control unit 40 is configured to control the side shift unit 19, and based on the distance D1 in the left-right direction between the load W1 specified by the distance calculation unit 36 and the object adjacent to the load loading position P, the side shift unit 19 moves the load W1 closer to or away from the object adjacent to the load loading position P. Thereby, it is possible to squeeze between the load W1 and the load W2 to load the load W1, or to avoid a state where the load W1 overlaps.

[0043] When the side shift stop unit 41 detects a change in the positional relationship between the 2D LiDAR sensor 22 and the load W1 while the fork 16 is being moved by the side shift unit 19, it stops the operation of the side shift unit 19. This prevents the fork 16 from moving left and right by the side shift unit 19 when the load W1 contacts an object (e.g., load W2) adjacent to the load loading position P, thereby preventing damage to, for example, the front panel of the truck T.

[0044] <Edge Position Identification Method and Distance Calculation Method> Next, with reference to FIG. 6, a method in which the edge identification unit 35 identifies the positions of the edges of the load W1 and the object, and the distance calculation unit 36 calculates the left - right distance D1 between the load W1 and the load W2 will be described again. FIGS. 6B and 6C show the point cloud PG of FIG. 6A as histograms in the left - right direction and the up - down direction. In this description, the explanation is based on the point cloud PG acquired by the left 2D LiDAR sensor 22. Therefore, when implementing the edge position identification method and the distance calculation method according to the present invention based on the point cloud PG acquired by the right 2D LiDAR sensor 22, the left and right are reversed.

[0045] As shown in FIG. 6B, according to the frequency distribution on the X - axis, there is a region with no frequency in the center. This region indicates an area where the reflection of the laser by the 2D LiDAR sensor 22 is extremely low or non - existent compared to other regions.

[0046] The edge identification unit 35 identifies the right - hand frequency - containing section S1 adjacent to the region with no frequency as the position (left - right coordinate) of the left edge of the load W1. Also, the edge identification unit 35 identifies the left - hand frequency - containing section S2 adjacent to the region with no frequency as the position (left - right coordinate) of the right edge of the load W2.

[0047] Further, as shown in FIG. 6C, according to the frequency distribution on the Y-axis, there are a region without frequency closest to the origin and a region without frequency second closest to the origin. These regions also indicate regions where the laser reflection by the two-dimensional LiDAR sensor 22 is extremely low or absent compared to other regions.

[0048] The edge specifying unit 35 specifies, as the position (vertical coordinate) of the front edge of the load W1 viewed from the carrier vehicle 1, the interval S3 with frequency adjacent to the upper side of the region without frequency closest to the origin. Further, the edge specifying unit 35 specifies, as the position (vertical coordinate) of the front edge of the load W2, the upper interval S4 with frequency adjacent to the region without frequency second closest to the origin.

[0049] Note that since each of the intervals S1, S2, S3, and S4 has a numerical width, the average value of the numerical values of each interval may be used as the position of each edge, or the minimum value or the maximum value in the interval S1 may be used as the position of each edge.

[0050] The distance calculation unit 36 calculates the distance D1 between the positions of the edges of the load W1 and the load W2 specified in the left-right direction, that is, the distance D1 between the coordinates of each edge in the left-right direction. Next, the distance calculation unit 36 calculates the distance D2 between the positions of the edges of the load W1 and the load W2 specified in the up-down direction, that is, the distance D2 between the coordinates of each edge in the up-down direction.

[0051] Incidentally, as shown in FIG. 6C, according to the frequency distribution on the Y-axis, peak values exist in the upper interval group and the lower interval group sandwiching the second region without frequency. Therefore, the edge specifying unit 35 may specify, as the position (vertical coordinate) of the front edge of the load W1 viewed from the carrier vehicle 1, the lower interval among these two intervals with peak values. Further, the edge specifying unit 35 may specify, as the position (vertical coordinate) of the front edge of the load W2, the upper interval among these two intervals with peak values.

[0052] In this way, the carrier vehicle 1 can identify the position of the load W1 and the positions of the edges of each object adjacent to the load placement position P by analyzing the point cloud PG acquired by the two-dimensional LiDAR sensor 22 using the frequency distribution. Furthermore, the carrier vehicle 1 can calculate the left-right distance D1 and the front-rear distance D2 between the load W1 and the objects adjacent to the load placement position P.

[0053] As a result, even if the mobile shelf, the truck T, etc. are displaced from the predetermined position, the carrier vehicle 1 can correct the load placement position P later, so that the load W1 can be loaded at an appropriate position. Note that the histograms in FIGS. 6B and 6C are for the purpose of explaining the frequency distribution in this specification, and it is not particularly necessary for the edge identification unit 35 to create a histogram.

[0054] FIGS. 7 to 10 show examples of information that can be obtained by frequency distribution analysis by the carrier vehicle 1.

[0055] FIG. 7A shows the point cloud PG acquired by the two-dimensional LiDAR sensor 22 when the load placement position P is a frame-shaped rack. The left side of FIG. 7A shows two point clouds PG obtained by irradiating the two frames with laser light. FIGS. 7B and 7C show the acquired point cloud PG as histograms in the left-right direction and the up-down direction.

[0056] The edge identification unit 35 identifies the right-side frequency interval S1 adjacent to the region without frequency as the position (left-right coordinate) of the left edge of the load W1 by the same method as described above. Also, the edge identification unit 35 identifies the left-side frequency interval S2 adjacent to the region without frequency as the position (left-right coordinate) of the right edge of the frame.

[0057] Furthermore, the edge identification unit 35 identifies the position (vertical coordinate) of the front edge of the load W1 as viewed from the carrier 1 as the upper frequency interval S3 adjacent to the area with no frequency closest to the origin. Also, the edge identification unit 35 identifies the position (vertical coordinate) of the front edge of the frame as the upper frequency interval S4 adjacent to the area with no frequency second closest to the origin.

[0058] Next, the distance calculation unit 36 calculates the gap distance D1 between the positions of the edges of the identified load W1 and the edges of the load W2 in the left-right direction by the same method as described above. Also, the distance calculation unit 36 calculates the gap distance D2 between the position of the edge of the identified frame in the vertical direction and the position of the edge of the load W2.

[0059] Also, Fig. 8A shows the point cloud PG acquired by the 2D LiDAR sensor 22 when the position of the 2D LiDAR sensor 22 is arranged at the center of the height of the backrest 18. On the right side of Fig. 8A, the point cloud PG acquired by the reflection of the laser on the end of the backrest 18 is shown.

[0060] In this case, the edge identification unit 35 identifies the lower interval S5 as the position of the front edge of the load W1 as viewed from the carrier 1 among the peak value intervals of the upper interval group and the lower interval group, and identifies the upper interval S6 as the position (vertical coordinate) of the front edge of the load W2. Note that the edge identification unit 35 identifies the positions of the edges of the loads W1 and W2 in the left-right direction by the same method.

[0061] Next, the distance calculation unit 36 calculates the gap distances D1 and D2 by the same method.

[0062] Also, Fig. 9A shows the point cloud PG acquired by the 2D LiDAR sensor 22 when there is an abnormality in the loading destination space such as load collapse. On the upper side of Fig. 9A, the point cloud PG acquired by the reflection of the laser on the abnormality occurrence location is shown.

[0063] In this case, as shown in FIG. 9B, there is no region without frequency in the center in the left-right direction. Therefore, the edge identification unit 35 cannot identify the edges of the load W1 and the load W2. In other words, the edge identification unit 35 can identify that there is no gap between the load W1 and the load W2. Thereby, it can be recognized that it is impossible to perform the cargo handling without interference between the cargos, such as when there is an abnormality in the loading destination space. At this time, the control unit 30 may stop the cargo handling operation of the transport vehicle 1.

[0064] FIG. 10A shows a point group PG obtained by irradiating only the load W1 with a laser by the two-dimensional LiDAR sensor 22.

[0065] As shown in FIGS. 10B and 10C, the edge identification unit 35 identifies the frequency sections S7 and S8 closer to the origin as the positions of the edges of the load W1 in the left-right direction and the up-down direction.

[0066] Next, the distance calculation unit 36 calculates the distances D3 and D4 from the two-dimensional LiDAR sensor 22 (origin) to the positions of the left side and the front side edges of the load W1 identified. Thereby, the positional relationship between the load W1 and the two-dimensional LiDAR sensor 22 can be obtained.

[0067] As briefly described in the background art paragraph, conventionally, in the analysis using a LiDAR sensor, the distance between the surrounding objects and the LiDAR sensor is specified by comparing and matching the shape and features of the objects specified in advance with the obtained point group PG. In this method, when the unloading destination is a thin frame-shaped structure, when the surrounding structure including the backrest 18 is detected by the LiDAR sensor, or when there is an abnormality in the loading destination space, it is difficult to stably obtain the distance to the surrounding objects.

[0068] In addition, in the analysis using a conventional LiDAR sensor, since it is a method of recognizing the shape and features of a previously specified object, the position of the LiDAR sensor is adjusted so that the laser is irradiated onto the load W1 and the laser is not blocked. Therefore, with the conventional method, the mutual positional relationship among the carrier vehicle 1, the load W1, and the object adjacent to the load loading position P cannot be obtained only by the LiDAR sensor. Thus, with the conventional method, it is necessary to separately perform other distance measurements, interference confirmation, etc., and for that purpose, it was necessary to separately arrange other sensors, etc.

[0069] On the other hand, according to the method of the present invention, by using only the two-dimensional LiDAR sensors 22 on the left and right, the mutual positional relationship among the carrier vehicle 1, the loaded load W1, the object or load W2 adjacent to the load loading position P can be obtained. Moreover, according to the method of the present invention, even when the unloading destination is a thin frame-like structure, or when the surrounding structure including the backrest 18 is detected by the two-dimensional LiDAR sensor 22, the distance between the load W1 and the load W2 can always be stably obtained. Further, according to the method of the present invention, when there is an abnormality in the loading space, the handling operation can be promptly stopped.

[0070] Next, with reference to FIGS. 11 and 12, an example of a series of operations of the carrier vehicle 1 according to the present invention will be described. In this description, the carrier vehicle 1 in FIGS. 11 and 12 is described as a counterbalanced forklift. Therefore, it is described that the position of the mast 14 in the front-rear direction does not move.

[0071] (1)(1-1) As shown in FIG. 11A, before the carrier vehicle 1 scoops up the load W3, the laser is irradiated onto the load W3 by the two-dimensional LiDAR sensors 22 on the left and right. (1-2) Next, the carrier vehicle 1 analyzes the acquired point cloud PG using a frequency distribution by the edge specifying unit 35 to specify the edge positions of the left and right ends of this load W3. (1-3) Next, the carrier vehicle 1 calculates the position at the center in the left-right direction of the load W3 by the load position specifying unit 37, and calculates the distance D5 between the position at the center in the left-right direction of the load W3 and the position at the center in the left-right direction of the backrest 18 by the distance calculating unit 36. (1-4) Further, the carrier vehicle 1 moves the fork 16 in the left-right direction by the side shift unit 19 based on the calculated distance D5, thereby correcting the center deviation between the fork 16 and the load W3.

[0072] Incidentally, for example, when the carrier vehicle 1 is a side fork vehicle, the carrier vehicle 1 can correct the center deviation between the fork 16 and the load W3 by moving the vehicle body 11 by the travel control unit 34 based on the specified distance D5.

[0073] (2)(2-1) Next, the carrier vehicle 1 scoops up the load W3 (W1) and conveys it to the truck T, and as shown in FIG. 11B, while irradiating the laser toward the loading platform of the truck T by the two-dimensional LiDAR sensor 22 on the truck T side, it travels in parallel with the loading platform of the truck T. (2-2) Next, the carrier vehicle 1 analyzes the acquired point group PG by the edge specifying unit 35 using the frequency distribution, thereby specifying the position of the edge of the object (load W2) adjacent to the load loading position P on the loading platform, and calculating the mutual distances (positional relationships) between the specified edge position, the position of the two-dimensional LiDAR sensor 22, and the position of the load W1 by the distance calculating unit 36. (2-3) Next, when the carrier vehicle 1 specifies the position of the edge of this object, it changes its direction toward the truck T side by the travel control unit 34 based on the positional relationship between the specified edge position, the position of the two-dimensional LiDAR sensor 22, and the load W1.

[0074] (3)(3-1) Next, as shown in FIG. 11C, before the carrier vehicle 1 advances toward the load loading position P, it irradiates the laser in the horizontal direction by the two-dimensional LiDAR sensor 22. (3-2) Next, the carrier vehicle 1 identifies the positions of the edges of the load W1 and the load W2 by analyzing the acquired point cloud PG using a frequency distribution by the edge identification unit 35. (3-3) Next, the carrier vehicle 1 determines whether or not the load W1 interferes with the load W2 by calculating the distance D1 between the load W1 and the load W2 by the distance calculation unit 36. At this time, as described above, when the carrier vehicle 1 identifies that an abnormality has occurred, it may stop the handling operation.

[0075] (4)(4-1) Next, the carrier vehicle 1 advances to the load loading position P by the travel control unit 34 based on the distance D2 in the front-rear direction between the load W1 and the load W2 calculated by the distance calculation unit 36. (4-2) Next, as shown in FIG. 12A, before lowering the load W1, the carrier vehicle 1 irradiates a laser in the horizontal direction by the two-dimensional LiDAR sensor 22. (4-3) Next, the carrier vehicle 1 identifies the edges of the load W1 and the load W2 by analyzing the acquired point cloud PG using a frequency distribution by the edge identification unit 35. (4-4) Next, the carrier vehicle 1 calculates the distance D1 between the load W1 and the load W2 by the distance calculation unit 36. (4-5) Next, the carrier vehicle 1 calculates an appropriate control amount of the side shift unit 19 by the side shift control unit 40 based on the calculated distance D1.

[0076] Thereby, the carrier vehicle 1 can appropriately move the load W1 closer to the load W2 by the side shift unit 19.

[0077] (5)(5-1) Next, as shown in FIG. 12B, when the side shift unit 19 moves the fork 16 to the left, the carrier vehicle 1 irradiates a laser by the two-dimensional LiDAR sensor 22. (5-2) Next, while the carrier vehicle 1 identifies the edge of the load W1 by the edge identification unit 35 and identifies the positional relationship between the two-dimensional LiDAR sensor 22 and the load W1 by the load position identification unit 37, the side shift unit 19 moves the fork 16. (5-3) At this time, when the transport vehicle 1 detects a change in the positional relationship between the 2D LiDAR sensor 22 and the load W1 by the load position specifying unit 37, the side shift stop unit 41 stops the operation of the side shift unit 19.

[0078] Thereby, the transport vehicle 1 can detect that the load W1 starts to slide on the fork 16. Therefore, for example, the transport vehicle 1 can detect that the load W1 has been pressed against an object such as the front panel or the rear panel of the truck T, and after this detection, by stopping the movement of the fork 16, it is possible to prevent damage to the front panel or the rear panel.

[0079] On the other hand, if the transport vehicle 1 wants to press the load W1 against the load W2, it may be configured such that after detecting that the load W1 starts to slide on the fork 16, the side shift stop unit 41 stops the operation of the side shift unit 19.

[0080] (6)(6-1) Next, as shown in FIG. 12C, when the transport vehicle 1 pulls out the fork 16 from the load W3, the 2D LiDAR sensor 22 irradiates a laser. (6-2) Next, the transport vehicle 1 analyzes the acquired point group PG using the frequency distribution by the edge specifying unit 35 to specify the edge of the load W3, and while specifying the positional relationship between the 2D LiDAR sensor 22 and the load W3 by the load position specifying unit 37, pulls out the fork 16. (6-3) At this time, when the transport vehicle 1 detects that there is no change in the positional relationship between the 2D LiDAR sensor 22 and the load W3 by the load position specifying unit 37, the travel control unit 34 stops the movement of the transport vehicle 1.

[0081] Thereby, the transport vehicle 1 prevents pulling the load W3 by the fork 16.

[0082] As described above, since the two-dimensional LiDAR sensor 22 is arranged at a position where it can irradiate the load W1 and the load placement position P with a laser, it irradiates the load W1 and an object (for example, the load W2) adjacent to the load placement position P with a laser, and can detect the reflected light from the load W1 and the object adjacent to the load placement position P to obtain the point group PG. As a result, the transport vehicle 1 can specify the positions of the edges of the load W1 and the load W2 by the edge specifying unit 35, and can calculate the distances D1 and D2 by the distance calculating unit 36. Therefore, even if a mobile rack, a truck T, etc. are displaced from a predetermined position, the load placement position P can be corrected later, and the loading and unloading work can be appropriately performed.

[0083] Moreover, the transport vehicle 1 can specify the three relative positional relationships among the load W1, the load W2, and the two-dimensional LiDAR sensor 22 (transport vehicle 1) by analyzing the point group PG using the frequency distribution, so that the series of operations (1) to (6) described above can be performed.

[0084] As described above, one embodiment of the transport vehicle, the edge position specifying method, the distance calculating method, and the position specifying program of the present invention has been described. However, the present invention is not limited to the above embodiment. For example, the transport vehicle according to the present invention may be implemented by the following modification examples.

[0085] <Modification Example> · The second end portion 23c of the connecting portion 23 may not be located above the backrest 18. In this case, the point group PG obtained by the two-dimensional LiDAR sensor 22 becomes the point group PG shown in FIG. 8A. However, as already described, the edge specifying unit 35 can specify the positions of the edges of the loads W1 and W2, and the distance calculating unit 36 can calculate the distance between the loads W1 and W2. Also, the first end portion 23a of the connecting portion 23 may be provided at the upper end of the backrest 18.

[0086] ·If the two-dimensional LiDAR sensor 22 is arranged at a position where it can irradiate a laser beam, for example, on the load W1 loaded on the load-carrying part 16 and an object adjacent to the load-carrying position P, it may be fixed to, for example, the vehicle body 11 or the finger bar, or the first end portion 23a of the connecting part 23 may be fixed to the side surface (see FIG. 1) of the vertically extending portion 16a of the fork 16 in the vertical direction. Alternatively, the two-dimensional LiDAR sensor 22 may be fixed to the vehicle body 11, the vertically extending portion 16a of the fork 16, or the finger bar via the connecting part.

[0087] ·The load-carrying part may be constituted by a platen, for example, instead of the fork 16.

Explanation of Reference Numerals

[0088] W1 Load loaded on the load-carrying part W2 Load loaded adjacent to the load-carrying position W3 Load loaded at the load-carrying position D1 Distance in the left-right direction D2 Distance in the front-rear direction P Load-carrying position LE Laser irradiation range PG Point cloud T Truck 1 Carrier vehicle 10 Wheels 11 Vehicle body 12 Driving part 13 Laser scanner 14 Mast 15 Lift bracket 16 Fork (load-carrying part) 16a Vertically extending portion 17 Lifting part 18 Backrest 19 Side shift part 20 Carriage 21 Reach leg 22 Two-dimensional LiDAR sensor (point cloud acquisition part) 23 Connecting part 23a First end portion 23b Intermediate portion 23c Second end portion 30 Control Unit 32 Memory Unit 34 Travel Control Unit 35 Edge Identification Unit 36 Distance Calculation Unit 37 Load Position Identification Unit 38 Lifting Control Unit 40 Side Shift Control Unit 41 Side Shift Stop Unit

Claims

1. A load carrying section, a point cloud acquisition section that irradiates light horizontally onto the load placed on the load carrying section, an object around the transport vehicle, or any of them to acquire a point cloud, an edge identification section that analyzes the acquired point cloud using a frequency distribution with distances in the left-right and front-back directions or any one of the directions as axes, and identifies intervals with frequencies adjacent to a region substantially without frequencies as positions of edges of the load, the surrounding object, or any of them in the left-right or front-back direction, a distance calculation section, and the point cloud acquisition section further irradiates light horizontally onto the load and an object near the load to acquire the point cloud, the edge identification section further analyzes the acquired point cloud using a frequency distribution with the distance in the left-right direction as an axis, and identifies left and right intervals with frequencies adjacent to a region substantially without frequencies as positions of edges of the load or the object near it, respectively, the distance calculation section calculates the distance in the left-right direction between the load and the object near it based on the identified positions of the edges of the load and the object near it. A transport vehicle.

2. A load carrying section, a point cloud acquisition section that irradiates light horizontally onto the load placed on the load carrying section, an object around the transport vehicle, or any of them to acquire a point cloud, an edge identification section that analyzes the acquired point cloud using a frequency distribution with distances in the left-right and front-back directions or any one of the directions as axes, and identifies intervals with frequencies adjacent to a region substantially without frequencies as positions of edges of the load, the surrounding object, or any of them in the left-right or front-back direction, a distance calculation section, and the point cloud acquisition section further irradiates light horizontally onto the load and an object near the load to acquire the point cloud, the edge identification section further analyzes the acquired point cloud using a frequency distribution with the distance in the front-back direction as an axis, and identifies front and back intervals with frequencies adjacent to a region substantially without frequencies as positions of edges of the load or the object near it, respectively, the distance calculation section calculates the distance in the front-back direction between the load and the object near it based on the identified positions of the edges of the load and the object near it. A transport vehicle.

3. A method for calculating the distance of the gap between the load loaded on the transport vehicle and an object near the load based on the point cloud acquired by the point cloud acquisition section arranged on the transport vehicle, a step of irradiating light horizontally onto the load and the object by the point cloud acquisition section to acquire the point cloud, Analyzing the obtained point cloud using a frequency distribution with the distance in the left - right direction as the axis, and specifying an interval with a frequency adjacent to a region substantially without a frequency as the position of the edge of the load or the object; Calculating the distance in the left - right direction of the gap between the load and the object based on the positions of the edges of the load and the object thus specified, the distance calculation method comprising:

4. A method for calculating the distance of a gap between a load loaded on a transport vehicle and an object near the load based on a point cloud obtained by a point cloud acquisition unit arranged on the transport vehicle, The step of horizontally irradiating the load and the object with light by the point cloud acquisition unit to obtain the point cloud; Analyzing the obtained point cloud using a frequency distribution with the distance in the front - rear direction as the axis, and specifying an interval with a frequency adjacent to a region substantially without a frequency as the position of the edge of the load or the object; Calculating the distance in the front - rear direction of the gap between the load and the object based on the positions of the edges of the load and the object thus specified, the distance calculation method comprising:

5. A method for calculating the distance of a gap between a load loaded on a transport vehicle and an object near the load based on a point cloud obtained by a point cloud acquisition unit arranged on the transport vehicle, The step of horizontally irradiating the load and the object with light by the point cloud acquisition unit to obtain the point cloud; Analyzing the obtained point cloud using a frequency distribution with the distance in the front - rear direction as the axis, specifying the interval of the peak value in the upper region as the edge of the object, and specifying the interval of the peak value in the lower region as the position of the edge of the load; Calculating the distance in the front - rear direction between the load and the object based on the positions of the edges of the load and the object thus specified, the distance calculation method comprising:

6. A distance calculation program for calculating the distance of a gap between a load loaded on a transport vehicle and an object near the load, The transport vehicle Comprises a point cloud acquisition unit that irradiates light and acquires a point cloud, And a computer, The program causes the computer to Perform the step of horizontally irradiating the load and the object with light by the point cloud acquisition unit to obtain the point cloud; Analyze the obtained point cloud using a frequency distribution with the distance in the left - right direction as the axis, and specify an interval with a frequency adjacent to a region substantially without a frequency as the position of the edge of the load or the object; A distance calculation program that causes a computer to execute a step of calculating a lateral distance of a gap between the load and the object based on the specified load and the position of the edge of the object. **Claim 7**: A distance calculation program for calculating the distance of a gap between a load loaded on a transport vehicle and an object near the load, wherein the transport vehicle comprises a point cloud acquisition unit that irradiates light and acquires a point cloud, and a computer, and the distance calculation program causes the computer to execute a step of horizontally irradiating the load and the object with light by the point cloud acquisition unit to acquire the point cloud, analyze the acquired point cloud using a frequency distribution with the distance in the front-back direction as the axis, and specify an interval with a frequency adjacent to a region substantially without a frequency as the position of the edge of the load or the object, and execute a step of calculating the front-back distance of the gap between the load and the object based on the specified positions of the edges of the load and the object. **Claim 8**: A distance calculation program for calculating the distance of a gap between a load loaded on a transport vehicle and an object near the load, wherein the transport vehicle comprises a point cloud acquisition unit that irradiates light and acquires a point cloud, and a computer, and the distance calculation program causes the computer to execute a step of horizontally irradiating the load and the object with light by the point cloud acquisition unit to acquire the point cloud, analyze the acquired point cloud using a frequency distribution with the distance in the front-back direction as the axis, specify an interval of the peak value in the upper region as the edge of the object, and specify an interval of the peak value in the lower region as the position of the edge of the load, and execute a step of calculating the front-back distance between the load and the object based on the specified positions of the edges of the load and the object.

Images