AGV-based workpiece identification and stacking method and system

Abstract

The invention discloses an AGV-based workpiece identification and stacking method and system, and the method comprises the steps: identifying the model data of a workpiece, obtaining the workpiece, scanning the original three-dimensional point cloud data of the workpiece at a pallet fork through a laser radar, judging whether the original three-dimensional point cloud data is missing or not, and supplementing the virtual point cloud data through a pressure gradient if the original three-dimensional point cloud data is missing, three-dimensional point cloud data are given, if the data are not missing, the three-dimensional point cloud data are directly given, then placing and taking point data are scanned through a laser radar, the moving path of a pallet fork is planned through the placing and taking point data and the three-dimensional point cloud data, and the pallet fork is driven to move according to the pallet fork moving path and then put down a workpiece; supplementing three-dimensional point cloud data of the workpiece on the pallet fork; compared with the prior art, three-dimensional point cloud data missing caused by mirror surface reflection of the mirror surface or high-reflection surface workpiece is reduced, accurate positioning of the workpiece position in the workpiece stacking process is improved, and stacking errors are reduced.

Description

Technical Field

[0001] The present invention relates to an AGV cart, and in particular to a method and system for workpiece recognition and stacking based on an AGV. Background Art

[0002] An Automated Guided Vehicle (AGV) is a mobile transport cart equipped with an automatic guidance and recognition device. With a controller as the core of motion control and a battery or capacitor as the power source, it can perform operation tasks according to a scheduling order under the scheduling of an upper computer scheduling system. With the improvement of industrial automation and the rapid development of computer technology, automated guided vehicles play an increasingly important role in contemporary flexible production lines, assembly lines, and warehousing and logistics automation industry systems. At the same time, in recent years, various mobile robots have emerged. Whether wheeled or tracked, how to make the mobile robot move is the most core work.

[0003] At present, in the process of transporting workpieces by an AGV cart, the AGV cart often uses a lidar to scan it to obtain three-dimensional point cloud data of the workpiece. However, when transporting some workpieces with mirror / highly reflective surfaces, due to specular reflection, some spot noises will be generated, resulting in deviations in the scanned three-dimensional point cloud. Summary of the Invention

[0004] The present invention overcomes the deficiencies of the prior art and provides a method and system for workpiece recognition and stacking based on an AGV.

[0005] To achieve the above object, the technical solution adopted by the present invention is: a method and system for workpiece recognition and stacking based on an AGV, including: S1. Recognize the model data of the workpiece and obtain the workpiece according to the model data;

[0006] S2. Recognize the original three-dimensional point cloud data of the workpiece located on the fork;

[0007] S3. Determine whether the original three-dimensional point cloud data is missing;

[0008] If it is missing, supplement virtual point cloud data using the pressure gradient and give the three-dimensional point cloud data;

[0009] If it is not missing, give the three-dimensional point cloud data;

[0010] S4. Scan the pick-and-place point data, plan the moving path of the fork through the pick-and-place point data and the three-dimensional point cloud data, and the fork moves according to the moving path of the fork to achieve stacking.

[0011] In a preferred embodiment of the present invention,

[0012] In step S1, the model data of the workpiece is the length, width, and height data of the workpiece.

[0013] In a preferred embodiment of the present invention,

[0014] In step S2, the specific process of identifying the three-dimensional point cloud data of the workpiece located on the forklift forks is as follows:

[0015] The forklift fork area is scanned by a lidar at a frequency of 10 - 20 Hz to obtain the lidar point cloud density, and the original three-dimensional point cloud data is generated.

[0016] In a preferred embodiment of the present invention,

[0017] In step S3, the steps for determining whether the original three-dimensional point cloud data is missing are as follows: Determine whether the lidar point cloud density ≤ 500 points / m 2 ,

[0018] If the lidar point cloud density ≤ 500 points / m 2 , it is considered missing. Virtual point cloud data is supplemented using the pressure gradient, and the three-dimensional point cloud data is given;

[0019] If the lidar point cloud density > 500 points / m 2 , it is not considered missing, and the three-dimensional point cloud data is given.

[0020] In a preferred embodiment of the present invention,

[0021] Among them, the specific steps for supplementing virtual point cloud data using the pressure gradient are as follows:

[0022] The virtual point cloud coordinates (x v , y v , z v ) are added to the three-dimensional point cloud data, where the virtual point cloud coordinates (x v , y v , z v ) satisfy:

[0023]

[0024] z v = z 堆叠高度 ,

[0025] where, x contact and y contact are the contact point coordinates detected by the pressure sensor,

[0026] are the gradients of the pressure in the X / Y directions,

[0027] k is the pressure gradient coefficient,

[0028] z 堆叠高度 is the stacking height determined according to the RFID tag.

[0029] In a preferred embodiment of the present invention,

[0030] For the contact point coordinates, they are calculated through the centroid offset formula, and the specific formula is:

[0031]

[0032] where x contact , y contact are the centroid coordinates of the workpiece;

[0033] p i is the measured value of the i-th pressure sensor;

[0034] x i , y i are the installation position coordinates of the i-th pressure sensor on the forklift tine plane,

[0035] n is the total number of pressure sensors.

[0036] In a preferred embodiment of the present invention,

[0037] Gradient in the X direction (along the length of the forklift tine):

[0038] Gradient in the Y direction (along the width of the forklift tine):

[0039] where p(i,j) is the pressure value of the sensor in the i-th row and j-th column, Δx is the spacing between adjacent column pressure sensors, and Δy is the spacing between adjacent row pressure sensors. Similarly calculated according to the above formula

[0040] In a preferred embodiment of the present invention,

[0041] For the pressure gradient coefficient k, it can be obtained through ; specifically:

[0042] Place a known standard test block on the forklift tine and record the pressure gradient

[0043] Manually offset the test block by ΔL 实际位移 = 10 mm, and measure the pressure gradient again to obtain the gradient change

[0044]

[0045] Calculate to obtain k.

[0046] In a preferred embodiment of the present invention,

[0047] In step S4, the specific steps of scanning the pick-and-place point data and planning the forklift movement path through the pick-and-place point data and the three-dimensional point cloud data are as follows:

[0048] The pick-and-place point data includes the three-dimensional coordinates of the target position (x 目标 , y 目标 , z 目标 ), and the target attitude angles (θ x , θ y , θ z );

[0049] The three-dimensional point cloud data of the workpiece includes the pose of the workpiece on the current forklift (x 当前 , y 当前 , z 当前 , θ 当前 );

[0050] Calculate the deviation of the pose between the pick-and-place point and the three-dimensional point cloud:

[0051] Δx = x 目标 - x 当前 , Δy = y 目标 - y 当前 , Δz = z 目标 - z 当前 ;

[0052] Height regulation: Control the planetary roller screw (lead 5mm) to calculate the number of motor rotation turns according to the target height z 目标 Calculate the number of motor rotation turns:

[0053]

[0054] Perform horizontal direction control through PID:

[0055]

[0056] Among them, u y (t) represents the control output of the PID controller in the X-axis direction, e x = x 目标 - x 当前 , and similarly calculate u y (t).

[0057] An AGV-based workpiece recognition and stacking system, based on an AGV-based workpiece recognition and stacking method,

[0058] The RFID recognition module is installed at the front end of the forklift and is used to read the encrypted label information of the workpiece to obtain the model data of the workpiece;

[0059] The lidar is installed at the center of the top of the AGV body and is used to generate the pick-and-place point data and the original three-dimensional point cloud data of the workpiece;

[0060] A pressure feedback compensation module is installed on the working surface of the forklift fork, and calculates the actual contact position of the workpiece through the center of gravity offset formula;

[0061] A point cloud reconstruction and optimization module receives the data from the lidar and the pressure feedback compensation module, determines whether the original three-dimensional point cloud data is missing, and supplements the missing area of the lidar point cloud to improve the three-dimensional point cloud data of the lidar;

[0062] A controller is used to receive the model data of the RFID identification module, generate the moving path of the forklift fork, and send the acquisition information.

[0063] It is used to receive the three-dimensional point cloud data of the point cloud reconstruction and optimization module and the pick-and-place point data of the lidar, calculate the moving path of the forklift fork, and send the pick-and-place information.

[0064] A servo drive unit is connected to the controller, receives the acquisition information and the drive information, and drives the forklift fork to pick up and stack according to the path.

[0065] The present invention solves the defects existing in the background technology, and the present invention has the following beneficial effects:

[0066] (1) By identifying the model data of the workpiece and acquiring the workpiece, scanning the original three-dimensional point cloud data of the workpiece on the forklift fork by the lidar, and determining whether the original three-dimensional point cloud data is missing. If it is missing, virtual point cloud data is supplemented using the pressure gradient to give the three-dimensional point cloud data. If it is not missing, the three-dimensional point cloud data is directly given. Then, the pick-and-place point data is scanned by the lidar, and the moving path of the forklift fork is planned based on the pick-and-place point data and the three-dimensional point cloud data, and the forklift fork is driven to move along the moving path of the forklift fork and then put down the workpiece; supplement the three-dimensional point cloud data of the workpiece on the forklift fork; compared with the prior art, it reduces the missing of the three-dimensional point cloud data caused by specular reflection for workpieces with mirror or highly reflective surfaces, increases the precise positioning of the workpiece position during the stacking process of the workpiece, and reduces stacking errors.

[0067] (2) Detect whether the density of the original three-dimensional point cloud is ≤ 500 points / m 2 , and generate virtual point cloud data through the pressure gradient obtained from the data of the pressure feedback compensation module, and supplement the virtual point cloud data to the original three-dimensional point cloud data; improve the three-dimensional point cloud data of the workpiece on the forklift fork; compared with the prior art, it enhances the positioning of the position of the workpiece with a highly reflective surface in the forklift fork and reduces the influence of data missing caused by specular reflection. BRIEF DESCRIPTION OF THE DRAWINGS

[0068] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings described below are only some embodiments recorded in the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings;

[0069] Figure 1 is a flowchart of a preferred embodiment of the present invention;

[0070] Figure 2 is a schematic flowchart of a preferred embodiment of the present invention;

[0071] Figure 3 is a three-dimensional structure diagram of a preferred embodiment of the present invention; Detailed implementation manners

[0072] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention.

[0073] In the following description, many specific details are set forth in order to fully understand the present invention. However, the present invention can also be implemented in other ways different from those described herein. Therefore, the scope of protection of the present invention is not limited by the specific embodiments disclosed below.

[0074] In the description of the present application, it should be understood that the terms "center", "longitudinal", "transverse", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc. indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings. It is only for the convenience of describing the present application and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation. Therefore, it cannot be understood as a limitation on the scope of protection of the present application. In addition, the terms "first", "second", etc. are only used for descriptive purposes and cannot be understood as indicating or implying relative importance or implicitly indicating the quantity of the indicated technical features. Therefore, the features defined with "first", "second", etc. may explicitly or implicitly include one or more of such features. In the description of the present invention, unless otherwise stated, the meaning of "a plurality" is two or more.

[0075] In the description of this application, it should be noted that unless otherwise clearly specified and defined, the terms "installation", "connection", and "linkage" shall be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be directly connected or indirectly connected through an intermediate medium, and it can be the communication inside two components. For those of ordinary skill in the art, the specific meanings of the above terms in this application can be understood according to specific circumstances.

[0076] As Figure 1 and Figure 2 shown, a method and system for workpiece recognition and stacking based on AGV includes: a method for workpiece recognition and stacking based on AGV:

[0077] S1. Identify the model data of the workpiece and obtain the workpiece according to the model data;

[0078] S2. Identify the original three-dimensional point cloud data of the workpiece located on the forklift forks;

[0079] S3. Determine whether the original three-dimensional point cloud data is missing;

[0080] If it is missing, supplement the virtual point cloud data using the pressure gradient and give the three-dimensional point cloud data;

[0081] If it is not missing, give the three-dimensional point cloud data;

[0082] S4. Scan the pick-and-place point data, plan the moving path of the forklift forks through the pick-and-place point data and the three-dimensional point cloud data, and the forklift forks move according to the moving path of the forklift forks to achieve stacking;

[0083] In step S1, the model data of the workpiece is identified by the RFID identification module to identify the label information on the surface of the workpiece, and the length, width, and height data of the workpiece are obtained;

[0084] Among them, the model data includes the length (L), width (W), and height (H) data of the workpiece;

[0085] The steps for obtaining the workpiece according to the model data are: determining the grasping point according to the model data; adjusting the position of the forklift forks to align with the grasping point and grasping;

[0086] For the confirmation of the grasping point (x 抓取 , y 抓取 , z 抓取 ), it is located on the geometric center symmetry line,

[0087]

[0088] Furthermore, in step S2, the specific method for identifying the three-dimensional point cloud data of the workpiece located on the forklift forks is:

[0089] Scan the forklift area with a lidar at a frequency of 10 - 20 Hz to obtain the lidar point cloud density and generate the original three-dimensional point cloud data.

[0090] Further, in step S3, the step of determining whether the original three-dimensional point cloud data is missing is: determine whether the lidar point cloud density ≤ 500 points / m 2 ,

[0091] If the lidar point cloud density ≤ 500 points / m 2 , then it is missing, supplement the virtual point cloud data using the pressure gradient, and give the three-dimensional point cloud data;

[0092] If the lidar point cloud density > 500 points / m 2 , then it is not missing, and give the three-dimensional point cloud data.

[0093] Further, the specific steps of supplementing the virtual point cloud data using the pressure gradient are:

[0094] Add the virtual point cloud three-dimensional coordinates (x v , y v , z v ) to the three-dimensional point cloud data, where the virtual point cloud coordinates (x v , y v , z v ) satisfy:

[0095]

[0096] z v = z 堆叠高度 ,

[0097] where, x contact and y contact are the contact point coordinates detected by the pressure sensor (calculated by the centroid offset formula),

[0098] are the pressure gradients in the X / Y directions (unit: kg / mm),

[0099] k is the pressure gradient coefficient (unit: mm 2 / kg),

[0100] z 堆叠高度 is the stacking height determined according to the RFID tag.

[0101] More specifically, the formula for calculating the actual contact position of the workpiece by the centroid offset formula is:

[0102]

[0103] where, x contact , y contactis the center-of-gravity coordinate of the workpiece (relative to the geometric center of the fork plane); p i is the measured value of the i-th pressure sensor (unit: kg or N);

[0104] x i , y i are the installation position coordinates of the i-th pressure sensor on the fork plane (unit: mm), and n is the total number of pressure sensors.

[0105] More specifically, the pressure gradient calculation formula is:

[0106] Gradient in the X direction (along the length of the fork):

[0107] Gradient in the Y direction (along the width of the fork):

[0108] where p(i,j) is the pressure value of the sensor in the i-th row and j-th column, Δx is the spacing between adjacent column pressure sensors, and Δy is the spacing between adjacent row pressure sensors. Calculate similarly according to the above formula

[0109] More specifically, for the pressure gradient coefficient k, it can be obtained through ; specifically:

[0110] Place the standard test block (known size) on the fork and record the pressure gradient

[0111] Manually offset the test block by ΔL 实际位移 = 10 mm, measure the pressure gradient again, and obtain the gradient change

[0112]

[0113] Calculate to obtain k.

[0114] Furthermore, in step S4, scan the pick-and-place point data, plan the fork movement path through the pick-and-place point data and the three-dimensional point cloud data, and the specific steps for the fork to move along the fork movement path to achieve stacking are: scan the pick-and-place point area with a lidar at a frequency of 10 - 20 Hz to obtain the pick-and-place point data,

[0115] The pick-and-place point data includes the three-dimensional coordinates of the target position (x 目标 , y 目标 , z 目标 ), and the target attitude angles (θ x , θ y , θ z );

[0116] The three-dimensional point cloud data of the workpiece includes the pose (x 当前 , y 当前 , z 当前 , θ 当前 ) of the workpiece on the current fork;

[0117] Calculate the deviation between the pick-and-place point and the pose of the three-dimensional point cloud:

[0118] Δx = x 目标 - x 当前 , Δy = y 目标 - y 当前 , Δz = z 目标 - z 当前 ;

[0119] Achieve stacked placement through height regulation and horizontal direction control;

[0120] Among them, height regulation is achieved by controlling the planetary roller screw (lead 5mm) according to the target height z 目标 Calculate the number of motor rotation turns:

[0121]

[0122] Perform horizontal direction control through PID:

[0123]

[0124] Among them, u y (t) represents the control output of the PID controller in the X-axis direction, e x = x 目标 - x 当前 , and calculate u y (t) in the same way.

[0125] By confirming height regulation and horizontal direction control, the fork moves along this path and stacking is achieved.

[0126] More specifically, when the fork descends to 10mm away from the surface of the lower workpiece, it stops descending.

[0127] As Figure 3 shown, an AGV-based workpiece recognition and stacking system includes

[0128] an RFID recognition module, installed at the front end of the fork, for reading the encrypted label information of the workpiece and obtaining the model data of the workpiece;

[0129] a lidar, installed at the center of the top of the AGV main body, for generating pick-and-place point data and the original three-dimensional point cloud data of the workpiece;

[0130] The pressure feedback compensation module is installed on the working surface of the forklift fork and calculates the actual contact position of the workpiece through the center of gravity offset formula;

[0131] The point cloud reconstruction and optimization module receives the data from the lidar and the pressure feedback compensation module, determines whether the original three-dimensional point cloud data is missing, and supplements the missing area of the lidar point cloud to improve the three-dimensional point cloud data of the lidar;

[0132] The controller is used to receive the model data of the RFID identification module, generate the forklift fork movement path, and send the acquisition information,

[0133] It is used to receive the three-dimensional point cloud data of the point cloud reconstruction and optimization module and the pick-and-place point data of the lidar, calculate the forklift fork movement path, and send the pick-and-place information,

[0134] The servo drive unit is connected to the controller, receives the acquisition information and the drive information, and drives the forklift fork to pick up and stack according to the path.

[0135] Specifically, the lidar scans through a horizontal viewing angle of 270° at a frequency of 10Hz.

[0136] Based on the ideal embodiments of the present invention as inspiration, through the above description, relevant personnel can completely make various changes and modifications without departing from the technical idea of this invention. The technical scope of this invention is not limited to the content in the specification, and the technical scope must be determined according to the scope of the claims.

Claims

1. An AGV-based workpiece recognition and stacking method, characterized in that, Including: S1. Identify the model data of the workpiece and obtain the workpiece according to the model data; S2. Identify the original three-dimensional point cloud data of the workpiece located on the forklift; S3. Determine whether the original three-dimensional point cloud data is missing; If it is missing, supplement the virtual point cloud data using the pressure gradient and give the three-dimensional point cloud data; If it is not missing, give the three-dimensional point cloud data; S4. Scan the pick-and-place point data, plan the forklift movement path through the pick-and-place point data and the three-dimensional point cloud data, and the forklift moves according to the forklift movement path to achieve stacking.

2. The method and system for workpiece recognition and stacking based on AGV according to claim 1, characterized in that: In step S1, the model data of the workpiece is the length, width, and height data of the workpiece.

3. A method and system for workpiece recognition and stacking based on an AGV according to claim 1, characterized in that: Further, in step S2, the specific method for identifying the three-dimensional point cloud data of the workpiece located on the forklift is: Scan the forklift area with a lidar at a frequency of 10 - 20 Hz to obtain the lidar point cloud density and generate the original three-dimensional point cloud data.

4. A method and system for workpiece recognition and stacking based on an AGV according to claim 1, characterized in that: In step S3, the step of determining whether the original three-dimensional point cloud data is missing is: determining whether the lidar point cloud density is ≤ 500 points / m 2 , If the lidar point cloud density ≤ 500 points / m 2 , it is missing. Virtual point cloud data is supplemented using the pressure gradient to give three-dimensional point cloud data; If the lidar point cloud density > 500 points / m 2 , then there is no loss, and three-dimensional point cloud data is given.

5. The method and system for workpiece recognition and stacking based on AGV according to claim 4, characterized in that: The specific steps for supplementing the virtual point cloud data using the pressure gradient are as follows: Add the virtual point cloud coordinates (x v , y v , z v ) to the 3D point cloud data, where the virtual point cloud coordinates (x v , y v , z v ) satisfy: z v = z 堆叠高度 , where x contact and y contact are the coordinates of the contact point detected by the pressure sensor. is the gradient of the pressure in the X / Y direction, k is the pressure gradient coefficient, z 堆叠高度 is the stacking height determined according to the RFID tag.

6. The method and system for workpiece recognition and stacking based on AGV according to claim 5, characterized in that: For the contact point coordinates, they are calculated through the center-of-gravity offset formula, and the specific formula is: where x contact , y contact are the centroid coordinates of the workpiece; p i is the measured value of the i-th pressure sensor; x i ,y i are the installation position coordinates of the i-th pressure sensor on the forklift tine plane, n is the total number of pressure sensors.

7. A method and system for workpiece recognition and stacking based on an AGV according to claim 6, characterized in that: For is the gradient of the pressure in the X / Y direction, X-direction gradient (along the length of the forklift forks): Y-direction gradient (along the width direction of the fork): where p(i,j) is the pressure value of the sensor at the i-th row and j-th column, Δx is the spacing between adjacent column pressure sensors, and Δy is the spacing between adjacent row pressure sensors. Calculate in the same way according to the above formula 8. A method and system for workpiece recognition and stacking based on an AGV according to claim 7, characterized in that: For the pressure gradient coefficient k, it can be obtained by obtained; Specifically: Place a known standard test block on the fork and record the pressure gradient Manual offset test block ΔL 实际位移 = 10 mm, measure the pressure gradient again, and obtain the gradient change Calculate Obtain k.

9. The method and system for workpiece recognition and stacking based on AGV according to claim 1, wherein: In step S4, the specific steps for scanning the pick-and-place point data and planning the forklift movement path through the pick-and-place point data and the three-dimensional point cloud data are: The pick-up and placement point data includes the three-dimensional coordinates of the target position (x 目标 , y 目标 , z 目标 ), and the target attitude angles (θ x , θ y , θ z ); The three-dimensional point cloud data of the workpiece includes the pose (x 当前 , y 当前 , z 当前 , θ 当前 ) of the workpiece on the current forklift fork; Calculate the deviation of the pose between the pick-and-place point and the three-dimensional point cloud; Δx = x 目标 -x 当前 ,Δy = y 目标 -y 当前 ,Δz = z 目标 -z 当前 ; Height adjustment: By controlling the planetary roller screw (lead 5 mm) according to the target height z 目标 Calculate the number of motor rotation turns: Perform horizontal direction control through PID: where, u y (t) represents the control output of the PID controller in the X-axis direction, and e x = x 目标 - x 当前 . Similarly, calculate u y (t); By confirming the height adjustment and horizontal control, the forklift moves along this path and realizes stacking.

10. An AGV-based workpiece identification and stacking system, based on the AGV-based workpiece identification and stacking method described in any one of claims 1 - 9, characterized in that: An RFID identification module, installed at the front end of the forklift, is used to read the encrypted label information of the workpiece and obtain the model data of the workpiece; A lidar, installed at the center of the top of the AGV main body, is used to generate the pick-and-place point data and the original three-dimensional point cloud data of the workpiece; A pressure feedback compensation module, installed on the working surface of the forklift, calculates the actual contact position of the workpiece through the center-of-gravity offset formula; A point cloud reconstruction and optimization module, receives the data from the lidar and the pressure feedback compensation module, determines whether the original three-dimensional point cloud data is missing, and supplements the missing area of the lidar point cloud to improve the three-dimensional point cloud data of the lidar; A controller, used to receive the model data of the RFID identification module, generate the forklift movement path, and send acquisition information, used to receive the three-dimensional point cloud data of the point cloud reconstruction and optimization module and the pick-and-place point data of the lidar, calculate the forklift movement path, and send pick-and-place information, A servo drive unit, connected to the controller, receives the acquisition information and the drive information, and drives the forklift to pick up and stack according to the path.

Images