Method for estimating loading platform posture and forklift

Abstract

To easily estimate the posture of a loading platform of a transport vehicle parked in a stopping area.SOLUTION: A side wing 21 of a transport vehicle 1 has a rear surface 210a which is exposed outwardly at a position higher than an overall height H of the transport vehicle 1 when in the open position. A scanner 34 is provided on a forklift 3 so as to perform scanning at least in the horizontal direction. An estimation method includes steps of: irradiating, with a laser beam from the scanner 34, the rear surface 210a of the side wing 21 of the transport vehicle 1 in the open position parked in a stopping area to obtain point cloud data; processing the point cloud data to obtain a regression line of the point cloud representing an object in the stopping area at the height position of the rear surface 210a of the side wing 21 in the open position; and estimating information on the posture of the loading platform 2 of the transport vehicle parked in the stopping area based on the regression line.SELECTED DRAWING: Figure 6

Description

[Technical field]

[0001] The present application relates to a method for estimating the attitude of the platform of a transport vehicle, and to a forklift configured to implement said method. [Background technology]

[0002] A system has been developed in which a forklift automatically loads and unloads cargo from the bed of a transport vehicle (such as a truck or trailer) or a container. The forklift drives up to the bed of a transport vehicle parked in a designated parking area within the work area and loads and unloads cargo from the bed.

[0003] Since the transport vehicle is driven by a driver, the posture (direction) and position of the transport vehicle parked in the parking area are different each time. Therefore, in order for the forklift to load the cargo bed with high accuracy, the loading position where the forklift should stop during loading must be changed according to the posture of the cargo bed, etc.

[0004] For this purpose, the loading platform of a parked transport vehicle is detected, data is acquired, and the posture of the loading platform is estimated by processing the data, and the loading position is determined or corrected based on the estimated posture.

[0005] The method of Patent Document 2 irradiates a laser beam from a forklift to a loading platform, obtains three-dimensional point cloud data, processes the point cloud data to determine a straight line representing the edge of the loading platform (loading platform floor), and determines the loading position (loading position) based on this straight line. This method eliminates the need to attach markers to the loading target or pillars in the work area as in Patent Documents 3 and 4, and the need to install sensors such as cameras on structures in the work area.

[0006] The method of Patent Document 1 involves irradiating the inside of a side-opening container with a laser beam from a forklift to obtain point cloud data, processing the point cloud data to determine a straight line representing the side wall of the container, and calculating the relative angle of the container with respect to the forklift from the inclination of the straight line. In Patent Document 1, when a container is loaded with cargo, the laser beam must be irradiated onto the container while taking into consideration the arrangement of the cargo so as not to be interfered with by the cargo, and point cloud data must be obtained. [Prior art documents] [Patent documents]

[0007] [Patent Document 1] JP 2020-175997 A [Patent Document 2] JP 2023-30983 A [Patent Document 3] JP 2021-160931 A [Patent Document 4] JP 2022-144133 A Summary of the Invention [Problem to be solved by the invention]

[0008] An object of the present application is to easily estimate the posture of a loading platform of a transport vehicle parked in a parking area. [Means for solving the problem]

[0009] The present application provides an estimation method for use in a loading / unloading system in which a forklift automatically handles loading and unloading onto the loading platform of a transport vehicle, the method estimating the attitude of the loading platform of the transport vehicle parked in a parking area. Here, the side wings of the loading platform are configured to rotate up and down between a closed position in which the loading platform is closed on its side and an open position in which the loading platform is opened on its side. When the side wing is in the open position, the side wing has a back surface exposed outwardly at a position higher than the overall height of the transporter when in the closed position. A scanner is also mounted on the forklift so as to be capable of scanning at least horizontally. And, the estimation method is: irradiating the laser light from the scanner toward the underside of the side wing in the open position of the transport vehicle parked in the parking area to acquire point cloud data; Processing the point cloud data to obtain a regression line of the point cloud representing objects in the parking area at a height position of the underside of the side wing in the open position; and and estimating information regarding the attitude of the loading platform of the transport vehicle parked in the parking area based on the regression line.

[0010] As information regarding the posture of the loading platform, the estimation method may estimate the direction of the regression line as the fore-and-aft direction of the loading platform.

[0011] As the information regarding the attitude of the loading platform, the estimation method may estimate a relative angle of the loading platform with respect to the forklift from a slope of the regression line.

[0012] The estimation method further comprises: a first line that is perpendicular to the regression line and passes through one end point of the group of points; A second straight line that is perpendicular to the regression line and passes through the other end point of the group of points is obtained; The positions of the front and rear ends of the platform may be estimated from the first straight line and the second straight line.

[0013] The estimation method further comprises: determining a distance between the first straight line and the second straight line; The longitudinal length of the loading platform may be estimated from the distance.

[0014] The side wing is a top surface portion that constitutes a top surface of the loading platform when the loading platform is in the closed position; a side portion that constitutes a side surface of the loading platform when the loading platform is in the closed position; The back surface is the back surface of the top surface portion. The estimation method according to claim 1.

[0015] The scanner may be a 2D-LiDAR scanner or a 3D-LiDAR scanner.

[0016] The scanner may be used as an external sensor to estimate the position of the forklift in a work area.

[0017] The present application also provides an automatically operable fourlift that handles cargo on the loading platform of a transport vehicle parked in a parking area. The forklift is a scanner that is provided so as to be capable of scanning at least in a horizontal direction and that irradiates a laser beam toward the underside of the side wing of the transport vehicle that is in the open position and that is parked in the parking area, thereby acquiring point cloud data; a platform information estimation unit that processes the point cloud data to obtain a regression line of the point cloud representing objects in the parking area at the height position of the underside of the side wing in the open position, and calculates information regarding the posture of the platform of the transport vehicle parked in the parking area based on the regression line, It is configured to implement the estimation method described above. [Brief description of the drawings]

[0018] [Figure 1] 1A and 1B are schematic views partially showing a working area, where in FIG. 1A the loading platform of the transport vehicle is closed, and in FIG. 1B the loading platform of the transport vehicle is open. [Diagram 2] FIG. 2A is a schematic side view of an exemplary transporter with the cargo bed closed, and FIG. 2B is a schematic side view of the transporter of FIG. 2A with the cargo bed open. [Diagram 3] FIG. 3 is a schematic perspective view of an exemplary transport vehicle with an open loading platform. [Figure 4] FIG. 4 is a schematic side view of an exemplary forklift. [Diagram 5]FIG. 5 is a block diagram of a forklift showing a configuration for implementing the estimation method. [Figure 6] FIG. 6 illustrates an example of the positional relationship between a forklift (scanner) and a transport vehicle (side wing) for implementing the estimation method. [Figure 7] FIG. 7 is a flow diagram illustrating steps of an exemplary estimation method. [Figure 8] FIG. 8 illustrates an example of point cloud data. [Figure 9] FIG. 9 is a diagram for explaining setting of a region of interest and obtaining a regression line. [Figure 10] FIG. 10 is a diagram for explaining estimation of loading platform information based on a regression line. [Figure 11] 11A and 11B are photographs showing the inside of the side wings in the open position. [Figure 12] 12A-12C show test data in which a regression line was actually obtained using 2D-LiDAR. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0019] Hereinafter, exemplary embodiments of the present application will be described with reference to the drawings. Note that the components shown in the drawings are not necessarily in exact dimensions or proportions, but merely represent their functions or operations.

[0020] FIG 1A partially illustrates a working area 5 of an exemplary cargo handling system. The cargo handling system defines a parking area 50 within the working area 5. A transport vehicle 1 is driven by a driver and parked within the parking area 50. Next, as shown in FIG 1B, the side wings 21 of the cargo platform 2 are opened to open the cargo platform 2. A forklift 3 performs cargo handling (loading and unloading) on ​​the cargo platform 2 by automatic driving.

[0021] [Transport vehicle] 2A shows an example transport vehicle 1. The transport vehicle 1 is a truck or trailer driven by a driver, and includes a cab 10 with a driver's seat, a plurality of wheels 11 for driving, and a wing body type cargo bed 2 provided behind the cab 10.

[0022] The loading platform 2 includes a loading platform floor 20 (FIG. 2B), side wings 21, tailgates 22, a front panel 23, a rear frame 24, a rear door 25 (FIG. 3), a center frame 26 (FIGS. 2B and 3), etc. These components enclose a loading space 27 shown in FIG. 2B, thereby closing the loading platform 2. The cargo W is placed on the loading platform floor 20 within the loading space 27. Here, the cargo W refers to the pallet and all the items placed on it.

[0023] The side wings 21 are attached to the center frame 26 via hinges or the like, and can rotate up and down around an axis parallel to the center line of the vehicle body (the center line of the loading platform 2 in the front-rear direction) of the transport vehicle 1. The side wings 21 are rotated up and down by an actuator (not shown), such as a cylinder, between a closed position in which the loading platform 2 (loading space 27) is closed at the side as shown in Fig. 2A, and an open position in which the loading platform 2 (loading space 27) is opened at the side as shown in Figs. 2B and 3B.

[0024] 2B, the open position of the side wing 21 is set higher than the overall height H of the transport vehicle 1 when the side wing 21 is in the closed position. In other words, when the bed 2 is open, the side wing 21 is located higher than the loading space 27.

[0025] The tailgate 22 is also configured to be rotatable up and down. The side wings 21 are rotated upward from the closed position to the open position to open the platform 2, and then the tailgate 22 is rotated downward to hang down, thereby opening the platform 2 widely on the sides. With the platform 2 opened in this manner, the forklift 3 can load or unload cargo W onto or from the platform floor 20.

[0026] 3, the side wings 21 are provided on both the left and right sides of the bed 2. The tailgates 22 are also provided on both the left and right sides of the bed 2. Therefore, the forklift 3 can handle cargo from both the left and right sides of the bed 2.

[0027] [forklift] 4 shows an example of a forklift 3. The forklift 3 is an unmanned forklift and is configured to be automatically operable. The forklift 3 includes a vehicle body 30 having drive wheels 31, forks 32 for supporting cargo W, and a cargo handling device 33 for raising and lowering the forks 32 and, optionally, shifting them left and right and / or moving them forward and backward.

[0028] The method for estimating the posture of the loading platform 2 is carried out by the forklift 3. For this purpose, the forklift 3 is provided with an optical scanner 34 that is arranged to scan at least in the horizontal direction. The scanner 34 is a LiDAR (Light Detection And Ranging) scanner that projects a pulsed laser beam around the forklift 3 and receives the laser beam reflected by surrounding objects to obtain point cloud data. The point cloud data includes a point cloud that represents the surface of the surrounding objects. Each point in the point cloud indicates a position from the scanner 34 (forklift 3), that is, coordinates (polar coordinates or Cartesian coordinates) based on the scanner 34.

[0029] The scanner 34 may be a 2D (two-dimensional) LiDAR having a scanning range only in the horizontal direction, or a 3D (three-dimensional) LiDAR having a scanning range in both the horizontal and vertical directions. The scanner 34 is supported at a height to be described later by a bracket 35 attached to the vehicle body 30.

[0030] 5 is a block diagram of a forklift 3 showing a configuration for implementing a method for estimating the posture of the loading platform 2. The forklift 3 includes a storage medium 36 and a control unit 37. The storage medium 36 stores environmental information (e.g., an environmental map) of the work area 5, work instructions related to loading and unloading received from a management server of the loading and unloading system, and a program for executing automatic operation of the forklift 3. The storage medium 36 also stores point cloud data acquired by the scanner 34, information calculated based on the point cloud data (described later), and the like.

[0031] The control unit 37 controls the automatic operation of the forklift 3. The control unit 37 is a calculation device, and includes a calculation circuit such as a CPU (Central Processing Unit).

[0032] An automatically operable forklift 3 is usually equipped with an external sensor (not shown), and the control unit 37 estimates the position and orientation of the forklift 3 itself (the forklift 3) in the work area 5 based on the output from the external sensor. This method of estimating the self-position may be based on the principle of triangulation using a reflector provided in the work area 5, or on SLAM (Simultaneous Localization and Mapping) technology. These are well-known technologies, and therefore a description thereof will be omitted. The scanner 34 for estimating the attitude of the loading platform 2 may also function as this external sensor. This contributes to a reduction in the number of measuring means required for the automatic operation of the forklift 3.

[0033] The control unit 37 drives and steers the drive wheels 31 and activates the loading / unloading device 33 to operate the forks 32 based on its estimated position and work instructions, so that the forklift 3 automatically travels within the work area 5 and unloads cargo onto the loading platform 2.

[0034] The control unit 37 includes a loading platform information estimation unit 370 (hereinafter simply referred to as the "estimation unit 370") for processing the point cloud data acquired by the scanner 34 and determining information relating to the attitude, position, length, etc. of the loading platform 2. The control unit 37 (estimation unit 370) functions by executing a program stored in the storage medium 36.

[0035] [Scanner mounting height] 6, the side wing 21 has a back surface 210a that is exposed outward at a position higher than the overall height H when in the open position. Specifically, the side wing 21 is made up of a top surface portion 210 that extends from its rotation axis and forms the top surface of the loading platform 2 when in the closed position, and a side surface portion 211 that extends perpendicularly to the top surface portion 210 and forms the side surface of the loading platform 2 when in the closed position. The above-mentioned back surface 210a exposed outward is the back surface of the top surface portion 210.

[0036] The estimation method of the present application utilizes this back surface 210a. The scanner 34 is attached so as to be able to scan the inside of the side wing 21 in the horizontal direction at least at the height position (horizontal plane) where this back surface 210a exists.

[0037] When the scanner 34 is a 2D-LiDAR scanner, it has a scanning range (scanning surface) 34a that is horizontal to the ground. The scanner 34 is supported within the height range of the back surface 210a of the side wing 21 in the open position, as shown in Fig. 6. Note that, as shown in the example of Fig. 6, depending on the type of the loading platform 2, the side surface portion 211 may be present within the height range of the back surface 210a. In this case, the height of the scanner 34 is naturally adjusted in advance so that the laser light does not interfere with the side surface portion 211.

[0038] When the scanner 34 is a 3D-LiDAR scanner, it has a scanning range in the vertical direction in addition to the horizontal direction, so it may be mounted at the same height as the 2D-LiDAR scanner, and even if it is mounted at a position close to the overall height H or a slightly lower position, the laser light can be irradiated onto the back surface 201a over the entire length of the loading platform in the front-rear direction. Although not shown, if the scanner 34 is too low compared to the overall height H, depending on the positional relationship between the forklift 3 and the transport vehicle 1 when acquiring the point cloud data, the laser light may interfere with the main structure of the transport vehicle 1 (the cab 10, the front panel 23, the rear frame 24, the rear door 25, etc.). Taking this into consideration, the height of the scanner 34 is adjusted in advance.

[0039] [Method of estimating platform posture] 7, the estimation method 4 includes steps S1 to S3. Steps S1 to S3 are performed by the forklift 3 (controller 37).

[0040] First, step S1 is performed. Step S1 is a step for acquiring point cloud data required for estimating the posture of the loading platform 2 by the scanner 34. Specifically, in step S1, the forklift 3 uses the scanner 34 attached to itself to scan the underside 210a of the side wing 21 in the open position of the transport vehicle 1 parked in the parking area 50 at least in the horizontal direction.

[0041] 1B, the transport vehicle 1 is parked in the parking area 50, the side wings 21 are in the open position, and the loading platform 2 is open. The forklift 3 causes the scanner 34 to project laser light at a position near the parking area 50, preferably at a position facing the rear surface 210a of the side wing 21. The laser light is projected horizontally toward at least the parking area 50, and thus irradiated onto the rear surface 210a of the side wing 21 over the entire length of the loading platform in the front-rear direction, and is also irradiated onto other surrounding objects and reflected, and the reflected light is received by the scanner 34 to obtain point cloud data.

[0042] Next, step S2 is performed. Step S2 is a step of processing the point cloud data acquired in step S1 to obtain a regression line of the point cloud representing the object in the parking area 50 at the height position of the back surface 210a. This step S2 is performed by the estimation unit 370 through calculation.

[0043] An example of point cloud data acquired by a 2D-LiDAR scanner in the positional relationship of FIG. 1B is shown in FIG. 8. This is horizontal plane two-dimensional coordinate (XY coordinate) data at the height position where the back surface 210a exists (specifically, the height of the scanning range 34a in FIG. 6). The two-dimensional coordinate data has the forklift 3 (scanner 34) as its origin.

[0044] Although it depends on the type, the 2D-LiDAR has a horizontal scanning range of about 270° to 360°. Therefore, the point cloud data includes not only a point cloud PC1 representing the surface (mainly the back surface 210a) of an object in the parking area 50, but also another point cloud PC2 representing structures in the work area 5 (for example, walls and pillars of a building).

[0045] The purpose of step S2 is to obtain a regression line of the point cloud PC1, and the point cloud PC2 is not necessary. Therefore, the estimation unit 370 sets an appropriate ROI (Region of Interest) for the point cloud data. The forklift 3 recognizes its own position and orientation in the work area 5 by an external sensor. The parking area 50 is a predefined area within the work area 5 and is known. Therefore, the positional relationship between the forklift 3 and the parking area 50 is self-evident. Therefore, it is very easy to set the ROI.

[0046] 9, the estimation unit 370 sets an ROI for the point cloud data, and obtains a regression line RL of the points in the ROI, that is, the regression line RL of the point cloud PC1, for example, by the least squares method, etc. In this manner, the regression line RL of the point cloud PC1 is acquired.

[0047] In the case of a 3D-LiDAR scanner, three-dimensional point cloud data is acquired, but by setting the ROI as described above at the height position (known) where the back surface 210a exists, the regression line RL of the point cloud PC1 can be similarly acquired.

[0048] Next, step S3 is performed. Step S3 is a step of determining (estimating) information about the posture of the platform 2 based on the regression line RL. This step S3 is performed by the estimation unit 370 through calculation.

[0049] The point cloud PC1 is a point cloud representing the surface of an object present in the parking area 50 at the height position of the back surface 210a. When the side wing 21 is in the open position, it is located higher than the overall height H. Therefore, the point cloud PC1 captures the inside of the side wing 21 located higher than the overall height H, and does not include points representing other main structures of the transport vehicle 1 such as the cab 10, the front panel 23, the rear frame 24, and the rear door 25. The point cloud PC1 can be said to be a point cloud that substantially represents the back surface 210a. Therefore, the regression line RL can be regarded as a line substantially parallel to the center line of the loading platform 2 in the fore-aft direction.

[0050] Although the point cloud PC1 may include points representing the rain covers and cylinders on both sides of the side wing 21, these are very small compared to the length of the side wing 21 in the fore-aft direction of the loading platform, and if the points of the point cloud PC1 are dense, their influence is extremely small and can be ignored. Therefore, even in such a case, the regression line RL can be regarded as being substantially parallel to the center line of the loading platform 2 in the fore-aft direction.

[0051] Therefore, the estimation unit 370 may determine that the direction of the regression line RL is the front-rear direction of the cargo bed 2.

[0052] The estimation unit 370 estimates the relative angle of the platform 2 of the transport vehicle 1 in the stopping area 50 with respect to the forklift 3 at the point where the point cloud data is acquired from the gradient of the regression line RL. Specifically, the estimation unit 370 may obtain the angle between the regression line RL and the X-axis, and determine this as the relative angle.

[0053] Furthermore, in step S3, the estimation unit 370 may determine information regarding the position and length of the loading platform 2 based on the regression line RL. Specifically, the estimation unit 370 determines the positions of the front and rear ends of the loading platform 2 and the loading platform length.

[0054] 10, the estimation unit 370 obtains a first straight line PL1 that is perpendicular to the regression line RL and passes through one end point of the point group PC1, and a second straight line PL2 that is perpendicular to the regression line RL and passes through the other end point of the point group PC1. As shown in Figs. 2A and 2B, the side wing 21 exists over the entire fore-aft direction of the loading platform 2, so that the points at both ends of the point group PC1 correspond to the positions of the front and rear ends of the loading platform 2. Therefore, the estimation unit 370 may determine that the front and rear ends of the loading platform 2 are on these straight lines PL1 and PL2.

[0055] Furthermore, the estimation unit 370 obtains the distance d between the first straight line PL1 and the second straight line PL2. As described above, it can be determined that the front and rear ends of the loading platform 2 are located on the first and second straight lines PL1, PL2. Therefore, this distance d corresponds to the front-to-rear length of the loading platform 2. In this manner, the estimation unit 370 may calculate the front-to-rear length of the loading platform 2 from the distance d.

[0056] From the above, the estimation unit 370 can determine that the loading platform 2 is present in a region between the straight lines PL1 and PL2 and parallel to the regression line RL. Note that the regression line RL essentially represents the rear surface 210a of the side wing 21 on one side, and is essentially parallel to the center line of the loading platform 2 in the front-rear direction, but does not necessarily coincide with it.

[0057] In this way, the information about the loading platform 2 acquired in step S3 is used to determine or correct the loading position where the forklift 3 will stop when loading a load onto the loading platform 2. The above method does not provide information such as the position of the left or right end of the loading platform 2. The information obtained by estimation method 4 alone is insufficient to finally determine (correct) the loading position, but the forklift 3 is usually equipped with an external sensor for self-location estimation and other measuring means, and by combining this information with information obtained from these, the loading position can be determined.

[0058] This application focuses on the back surface 210a of the side wing 21 in the open position, which is exposed outwardly and is higher than the overall height H. No major structure of the transport vehicle 1, except for the side wing 21, is present at the height of the side wing 21 in the open position. Therefore, the regression line RL of the point cloud PC1 accurately represents the longitudinal direction of the back surface 210a of the side wing 21, which is parallel to the center line of the loading platform 2 in the front-rear direction.

[0059] Unlike the embodiment of the present application, for example, if the scanner 34 scans the platform 2 at a position lower than the overall height H, the point cloud data will include points representing various structures of the transport vehicle 1 and possibly even the luggage W in the loading space 27. From this, complex processing will be required to remove points that are not necessary for estimating the posture of the platform 2.

[0060] In the above embodiment, such complicated processing is not required, and there is an advantage that information regarding the posture of platform 2 can be easily and accurately estimated by only the simple processing steps of calculating the regression line (and setting an ROI if the scanning range of scanner 34 is wide).

[0061] The back surface 210a of the side wing 21 is not limited to the flat end as shown in the above example or FIG. 11A, but may have a rib structure as shown in FIG. 11B and have unevenness over the entire surface. In the case of FIG. 11B, the points of the point cloud are not arranged in a straight line as in FIG. 9. However, even in this case, the unevenness is averaged, and the obtained regression line itself is a straight line that is substantially parallel to the center line of the cargo bed 2 in the front-rear direction. Therefore, the estimation method 4 does not necessarily require the back surface 210a to be flat, and can be applied even when the back surface is not flat.

[0062] [Test data] Figure 12A shows the whole surrounding point cloud data actually obtained using a 2D-LiDAR scanner with a horizontal scanning range of 360°. Here, the point clouds existing linearly in the left and right direction capture the inside of the side wing of the transport vehicle. The remaining point clouds are unnecessary for estimating the posture of the loading platform, representing the structures of the work area.

[0063] Figure 11B is an enlarged view of the portion of the point cloud data in Figure 11A that is related to the side wing. After setting an ROI and removing unnecessary points, the regression line RL (Figure 11C) of the point cloud PC1 was obtained by the least squares method. The regression line RL was y = -0.0044x + 4.0135. It was confirmed that this line was substantially parallel to the center line of the loading platform in the longitudinal direction.

[0064] As is clear from FIG. 11B and FIG. 11C, points representing the surface of an object other than the back surface of the side wing were confirmed at both ends of the point cloud PC1. These are cylinders for opening and closing the side wing, located at both ends of the side wing. A regression line RL was obtained including these points, and it was confirmed that this regression line RL is substantially parallel to the center line of the loading platform in the front-rear direction. In other words, it was found that these points are small compared to the length of the side wing in the front-rear direction of the loading platform, and can be ignored if the points of the point cloud are dense. It was confirmed that even if such points representing objects other than the back surface are included, the posture of the loading platform 2 can be accurately estimated without processing to remove the points from the point cloud PC1. [Explanation of symbols]

[0065] 1 Transport vehicle 2 Cargo platform 21 Side Wing 210 Top surface 210a Back 211 Side part 3. Forklift 34 Scanner 37 Control Unit 370 Cargo Information Estimation Department 4 Estimation method 5. Working Area 50 Parking area H Overall height of transport vehicle PC1 point cloud (representing the inside of the side wing) PL1 1st straight line PL2 2nd straight line RL Regression Line W Luggage d Distance

Claims

1. 1. A method for estimating a posture of a platform of a transport vehicle parked in a parking area, the method being used in a loading and unloading system in which a forklift automatically handles loading and unloading on a platform of the transport vehicle, the method comprising: the side wings of the loading platform are configured to rotate up and down between a closed position in which the loading platform is closed at its side and an open position in which the loading platform is opened at its side, When the side wing is in the open position, the side wing has a back surface exposed outwardly at a position higher than the overall height of the transport vehicle when in the closed position, a scanner is mounted on the forklift so as to be capable of scanning at least in a horizontal direction; The estimation method includes: irradiating the laser light from the scanner toward the underside of the side wing in the open position of the transport vehicle parked in the parking area to acquire point cloud data; Processing the point cloud data to obtain a regression line of the point cloud representing objects in the parking area at a height position of the underside of the side wing in the open position; and and estimating information regarding an attitude of the platform of the transport vehicle parked in the parking area based on the regression line. Estimation method.

2. The estimation method estimates the direction of the regression line to be the front-rear direction of the loading platform. The estimation method according to claim 1 .

3. The estimation method includes estimating a relative angle of the platform with respect to the forklift from a slope of the regression line. The estimation method according to claim 1 .

4. The estimation method further comprises: a first straight line that is perpendicular to the regression line and passes through one end point of the group of points; a second straight line that is perpendicular to the regression line and passes through the other end point of the group of points; estimating positions of the front end and the rear end of the loading platform from the first straight line and the second straight line; The estimation method according to claim 1 .

5. The estimation method further comprises: determining a distance between the first straight line and the second straight line; Estimating the front-to-rear length of the loading platform from the distance; The estimation method according to claim 4.

6. The side wing is a top surface portion that constitutes a top surface of the loading platform when the loading platform is in the closed position; a side portion that constitutes a side surface of the loading platform when the loading platform is in the closed position; The back surface is the back surface of the top surface portion. The estimation method according to claim 1 .

7. The scanner is a 2D-LiDAR scanner. The estimation method according to claim 1 .

8. The scanner is a 3D-LiDAR scanner. The estimation method according to claim 1 .

9. using the scanner as an external sensor for estimating the position of the forklift in a work area; The estimation method according to claim 1 .

10. An automatically operable forklift that handles cargo on the loading platform of a transport vehicle parked in a parking area, the side wings of the loading platform are configured to rotate up and down between a closed position in which the loading platform is closed at its side and an open position in which the loading platform is opened at its side, When the side wing is in the open position, the side wing has a back surface exposed outwardly at a position higher than the overall height of the transport vehicle when in the closed position, The forklift is a scanner that is provided so as to be capable of scanning at least in a horizontal direction and that irradiates a laser beam toward the underside of the side wing of the transport vehicle that is in the open position and that is parked in the parking area, thereby acquiring point cloud data; a platform information estimation unit that processes the point cloud data to obtain a regression line of the point cloud representing objects in the parking area at the height position of the underside of the side wing in the open position, and calculates information regarding the posture of the platform of the transport vehicle parked in the parking area based on the regression line, A forklift configured to implement the estimation method of claim 1.

Images