Robot teaching transfer method and robot

Abstract

The application provides a robot teaching migration method and a robot. In the application, the hand-eye calibration information of a non-reference robot, the height deviation of the mounting height of the mechanical arm of the non-reference robot relative to the mounting height of the mechanical arm of a reference robot, and the TCP installation information deviation of the TCP installation information of the non-reference robot relative to the TCP installation information of the reference robot are obtained by the non-reference robot, so that the teaching action migrated from the reference robot is executed based on the hand-eye calibration information, the height deviation and the TCP installation information deviation, the teaching action is no longer taught to each non-reference robot, the teaching workload is simplified, and the efficiency of the rapid online of the non-reference robot is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of robots, in particular to a robot teaching migration method and a robot. BACKGROUND

[0002] The actions such as grabbing performed by robots are realized based on teaching. The robots here can be composite robots. The composite robots integrate the functions of mobile robots and industrial robots.

[0003] In a field environment, various machines can be different due to differences in process structures, and different machines can require different actions such as grabbing by robots. Under this premise, in order to ensure the accuracy of the actions such as grabbing of robots in various machines, each robot is often taught corresponding actions at each machine. Taking teaching of grabbing actions as an example, if there are 30 robots in the field environment, 40 types of machines, and 10 machines under each type of machine, 12000 (i.e. 30*40*10) grabbing actions will be taught, and the teaching workload is very large. SUMMARY

[0004] The present application provides a robot teaching migration method and a robot to avoid the problem of large teaching workload caused by teaching each robot at each machine.

[0005] The present application provides a teaching migration method applied to multiple robots and multiple machines. The method is applied to a non-reference robot, and the method comprises the following steps:

[0006] Obtaining hand-eye calibration information of the non-reference robot; the hand-eye calibration information comprises a conversion relationship between a robot arm coordinate system corresponding to a robot arm of the non-reference robot and a camera coordinate system corresponding to a camera of the non-reference robot;

[0007] Obtaining a height deviation of a robot arm installation height of the non-reference robot relative to a robot arm installation height of a reference robot, and obtaining a TCP installation information deviation of tool center point (TCP) installation information of the non-reference robot relative to TCP installation information of the reference robot;

[0008] Performing a teaching action migrated from the reference robot based on the hand-eye calibration information, the height deviation, and the TCP installation information deviation; the teaching action is a teaching action collected by the reference robot; and the teaching action is an atomic action split from a preset complete set of actions for reuse.

[0009] The present application provides a teaching migration method applied to multiple robots and multiple machines. The method is applied to a reference robot, and the method comprises the following steps:

[0010] Obtaining teaching actions collected by the reference robot at a reference machine table; the teaching actions at least include atomic actions split from a preset complete set of actions for reuse;

[0011] Obtaining machine table deviations of each non-reference machine table relative to the reference machine table measured by the reference robot;

[0012] Wherein, the teaching actions and the machine table deviations are migrated to each non-reference robot, so that each non-reference robot executes the teaching actions based on hand-eye calibration information obtained by each non-reference robot, a height deviation of a robot arm installation height of each non-reference robot relative to a robot arm installation height of the reference robot, a TCP installation information deviation of TCP installation information of each non-reference robot relative to TCP installation information of the reference robot, and the machine table deviations.

[0013] The application also provides a robot, which at least includes a body, a robot arm, and a camera;

[0014] The body includes a chassis of the robot, and the chassis at least includes a processor, which is used to execute the steps in the first method above when the robot is a non-reference robot, and is used to execute the steps in the second method above when the robot is a reference robot.

[0015] As can be seen from the above technical solutions, in the application, the hand-eye calibration information of the non-reference robot, the height deviation of the robot arm installation height of the non-reference robot relative to the robot arm installation height of the reference robot, and the TCP installation information deviation of the TCP installation information of the non-reference robot relative to the TCP installation information of the reference robot are obtained by the non-reference robot, so that the teaching actions migrated from the reference robot are executed based on the above hand-eye calibration information, height deviation, and TCP installation information deviation, and the teaching work load is simplified, and the efficiency of rapid online of the non-reference robot is improved.

[0016] Further, the embodiment splits the complete set of teaching actions into atomic actions for reuse, which further simplifies the teaching work load and improves the efficiency of project deployment. BRIEF DESCRIPTION OF DRAWINGS

[0017] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present disclosure and serve to explain the principles of the present disclosure, together with the description.

[0018] Figure 1 A robot structure schematic diagram provided by the embodiment of the application;

[0019] Figure 2Method flowchart provided for embodiments of the present application;

[0020] Figure 3a Hand-eye calibration flowchart provided for embodiments of the present application;

[0021] Figure 3b Hand-eye calibration schematic diagram provided for embodiments of the present application;

[0022] Figure 3c Another hand-eye calibration schematic diagram provided for embodiments of the present application;

[0023] Figure 4a Height calibration flowchart provided for embodiments of the present application;

[0024] Figure 4b Mechanical arm installation height deviation calibration schematic diagram provided for embodiments of the present application;

[0025] Figure 4c Another mechanical arm installation height deviation calibration schematic diagram provided for embodiments of the present application;

[0026] Figure 5a TCP installation deviation calibration flowchart provided for embodiments of the present application;

[0027] Figure 5b TCP installation deviation calibration schematic diagram provided for embodiments of the present application;

[0028] Figure 5c Another TCP installation deviation calibration schematic diagram provided for embodiments of the present application;

[0029] Figure 6 Another method flowchart provided for embodiments of the present application;

[0030] Figure 7a Mark plate height deviation calibration flowchart provided for embodiments of the present application;

[0031] Figure 7b Mark plate position deviation calibration flowchart provided for embodiments of the present application;

[0032] Figure 7c Machine table deviation calibration schematic diagram provided for embodiments of the present application;

[0033] Figure 7d Another machine table deviation calibration schematic diagram provided for embodiments of the present application. DETAILED DESCRIPTION

[0034] In order to make the technical solution provided by the embodiments of the present application better understood by those skilled in the art, and make the above-mentioned purposes, features and advantages of the embodiments of the present application more apparent and easy to understand, the robot provided by the embodiments of the present application is described below with reference to the drawings:

[0035] The robot involved in the embodiment at least includes a robot body, a mechanical arm, and a camera. Figure 1 The structure of the robot is exemplified.

[0036] In the embodiment, the robot body at least includes a chassis and the like. The chassis at least includes a processor.

[0037] As an embodiment, the robot body serves as a bridge between the mechanical arm and the camera, and can control the mechanical arm and the camera to cooperate, which avoids direct communication between the camera and the mechanical arm. Specifically, the robot body, such as a robot chassis therein, is configured with an algorithm for controlling the camera and the mechanical arm, so that the robot body serves as a bridge between the camera and the mechanical arm to control the camera and the mechanical arm to cooperate. How to control the camera and the mechanical arm to cooperate is described below, which is not described here.

[0038] In the embodiment, the camera described above can be a 2D camera. As an embodiment, the camera described above can be carried on a specified position of the mechanical arm, such as the end of the mechanical arm. In the embodiment, the camera and the mechanical arm have no binding relationship, and the camera or the mechanical arm can be replaced individually according to requirements. For example, when the camera is replaced, the robot body only needs to re-adapt the driving of the camera; similarly, when the mechanical arm is replaced, the robot body only needs to re-adapt the control driving of the mechanical arm.

[0039] Based on the above description, the method provided by the embodiment of the application is described below:

[0040] Referring to Figure 2 , Figure 2 The method flowchart is provided by the embodiment of the application. The method is applied to a non-reference robot. In the embodiment, a robot is often selected as a reference robot from a plurality of robots applied to a target working environment as shown in Figure 1 The remaining robots are non-reference robots. Figure 2 The flowchart shown can be applied to any non-reference robot.

[0041] To realize teaching migration of multiple robots and multiple machines, Figure 2 The core idea of the flowchart shown is to determine the difference of the non-reference robot relative to the reference robot, so that the non-reference robot can execute the teaching action migrated from the reference robot (such as the teaching action collected by the reference robot on the reference machine) based on the determined difference. Compared with the existing method of teaching each robot, the method of migrating the teaching action effectively reduces the number of teaching, saves labor cost and project deployment time.

[0042] The flowchart shown in Figure 2 is described below:

[0043] AsFigure 2 As shown in the figure, the flow can include the following steps:

[0044] Step 201, obtaining hand-eye calibration information of the non-reference robot.

[0045] In this embodiment, the hand-eye calibration information of the non-reference robot is obtained through hand-eye calibration. Alternatively, in this embodiment, the hand-eye calibration is completed by the body of the non-reference robot, specifically by the chassis controlling the cooperation of the robot arm and the camera. For example Figure 3a The specific implementation of hand-eye calibration is exemplified, which is not described here.

[0046] In this embodiment, the hand-eye calibration is completed by the body of the non-reference robot, such as the chassis controlling the cooperation of the robot arm and the camera, without only completing the hand-eye calibration at the camera end, and the camera end does not need to store any parameters required by the hand-eye calibration, nor does it need to communicate directly with the robot arm, which makes it only need to re-execute the hand-eye calibration when replacing the camera, simplifies the logic of the camera end, and enhances the compatibility of different types of cameras.

[0047] In this embodiment, the hand-eye calibration information refers to the conversion relationship between the robot arm coordinate system corresponding to the robot arm of the non-reference robot and the camera coordinate system corresponding to the camera of the non-reference robot. Alternatively, the robot arm coordinate system is, for example, the coordinate system in which the base of the robot arm is located, which can be the world coordinate system. In addition, in this embodiment, the camera coordinate system can be the coordinate system corresponding to the installation position of the camera, such as the coordinate system established with the focusing center of the camera as the origin and the optical axis as the Z axis, which is not specifically limited in this embodiment. After obtaining the hand-eye calibration information of the non-reference robot, the camera and the robot arm can be unified for auxiliary positioning.

[0048] Step 202, obtaining a height deviation of the installation height of the robot arm of the non-reference robot relative to the installation height of the robot arm of the reference robot, and obtaining a TCP installation information deviation of the tool center point (TCP: Tool Center Point) installation information of the non-reference robot relative to the TCP installation information of the reference robot.

[0049] In this embodiment, the height deviation is realized by the body of the non-reference robot, such as the chassis controlling the cooperation of the robot arm and the camera of the non-reference robot, under the premise of the reference machine, and specifically as Figure 4a The flowchart is not described here.

[0050] In this embodiment, the TCP installation information deviation is determined by the body of the non-reference robot, such as the chassis controlling the robot arm of the non-reference robot to execute the action input by the external, and specifically as Figure 5aThe flowchart is not described here.

[0051] It should be noted that in the present embodiment, the above-mentioned hand-eye calibration information, the above-mentioned height deviation, and the above-mentioned TCP installation information deviation are inherent differences between the non-reference robot and the reference robot, and are unique to the non-reference robot. They can be stored as configuration information of the non-reference robot in the body of the non-reference robot (such as the chassis of the non-reference robot).

[0052] In step 203, the non-reference robot performs the teaching action migrated from the reference robot based on the above-mentioned hand-eye calibration information, the above-mentioned height deviation, and the above-mentioned TCP installation information deviation.

[0053] In the present embodiment, the teaching action is an atomic action for reuse which is split from a preset complete set of actions collected by the reference robot under the reference machine. Here, the reference machine includes a reference machine corresponding to each machine type, and the reference machine corresponding to any machine type is selected from all machines of this type. When the reference robot collects the above-mentioned teaching action, the above-mentioned teaching action can be imported from the reference robot to each non-reference robot in batch import mode. Finally, Figure 2 The non-reference robot applying the flowchart will obtain the above-mentioned teaching action.

[0054] It should be noted that in the present embodiment, after the non-reference robot is imported with the above-mentioned teaching action collected by the reference robot, the non-reference robot will store the teaching action in the robot arm of the non-reference robot. When the non-reference robot executes the teaching action, it will take into account the above-mentioned hand-eye calibration information, the above-mentioned height deviation, and the above-mentioned TCP installation information deviation to ensure smooth execution of the teaching action.

[0055] The above-mentioned teaching action is described below by way of example:

[0056] Taking the example of teaching the picking and placing actions of the reference robot (such as an AGV specially used for teaching) on the reference machine of any machine type:

[0057] In order to simplify the teaching workload, the present embodiment will split the complete set of picking actions and the complete set of placing actions to be taught according to the needs, mainly into atomic actions that can be reused and atomic actions that cannot be reused. The purpose of this splitting is to atomize the teaching action as much as possible, to realize the reuse of the reusable teaching action among multiple robots, and to reduce the teaching workload.

[0058] Specifically, the complete set of picking actions and the complete set of placing actions can be split into:

[0059] 1. A photographing action. As an embodiment, the photographing action here can be summarized as: in the scenario of picking up or putting down goods from / to a machine table (such as a reference machine table), the robot arm of the reference robot moves from a designated hang-up position to a designated photographing position on the machine table (such as the reference machine table). Here, the hang-up of the robot arm means that the robot arm is retracted to a designated posture, which avoids the robot arm from colliding with external obstacles when the entire robot moves (at this time, the chassis of the robot also moves), so as to prevent damage to the robot arm. As described above, in this embodiment, the reference robot performs the photographing action described above when picking up or putting down goods from / to any machine table (such as the reference machine table or a non-reference machine table). That is, the photographing action described above is an atomic action that can be reused, which can be reused in multiple different robots.

[0060] 2. A grabbing action. As an embodiment, the grabbing action here can be summarized as: in the scenario of picking up goods from a machine table (such as a reference machine table), the robot arm of the reference robot moves from a designated photographing position (such as a designated photographing position on the reference machine table) to a designated grabbing position on the machine table. In a specific implementation, when the reference robot picks up goods from the machine table, no matter which cache bin the goods are grabbed from, the robot arm of the reference robot moves from the designated photographing position to the grabbing position on the machine table. That is, the grabbing action described above in the scenario of picking up goods is an atomic action that can be reused, which can be reused in multiple different robots.

[0061] For the scenario of the reference robot putting down goods to the machine table, the grabbing action performed by the reference robot is different, which is caused by the different cache bins where the goods are grabbed by the robot arm of the reference robot. For example, when the goods are placed in the head cache bin, the grabbing action performed by the reference robot can be that the robot arm of the reference robot first moves from the designated hang-up position to the cache bin at the head, and when the goods are placed in the tail cache bin, the grabbing action performed by the reference robot can be that the robot arm of the reference robot first moves from the designated hang-up position to the cache bin at the tail. It can be seen that the grabbing action performed by the reference robot when the goods are placed in the head cache bin is different from the grabbing action performed by the reference robot when the goods are placed in the tail cache bin. That is, the grabbing action described above in the scenario of putting down goods is an atomic action that cannot be reused, and needs to be taught to each non-reference robot one by one.

[0062] 3. Placement action. The placement action here can be summarized as: the action of moving above the machine table after the robot arm of the reference robot grasps the goods, and then placing the goods on the machine table. In the scenario of the reference robot placing goods on the machine table, the robot arm of the reference robot grasps the goods from the head buffer or the tail buffer, and after grasping the goods, the robot arm will move above the goods placement point of the machine table and then place the goods on the machine table. That is, the placement action in the scenario of placing goods on the machine table is an atomic action that can be reused by multiple different robots.

[0063] However, in the scenario of taking goods from the machine table, after the reference robot grasps the goods from the machine table, the placement action performed by the robot arm of the reference robot on the goods is different. This difference is caused by the different buffer of the goods. For example, when placing the grasped goods in the head buffer, the placement action performed by the reference robot can be: the robot arm of the reference robot places the goods in the buffer at the head, and when the goods are placed in the tail buffer, the placement action performed by the reference robot can be: the robot arm of the reference robot places the goods in the buffer at the tail. It can be seen that the placement action performed by the reference robot when placing the goods in the head buffer is different from the placement action performed by the reference robot when placing the goods in the tail buffer. That is, the placement action described above in the scenario of taking goods is an atomic action that cannot be reused, and each non-reference robot needs to be taught one by one.

[0064] The embodiment splits the whole set of actions taught to the reference robot to extract reusable atomic actions, and then migrates each reusable atomic action to each non-reference robot, simplifies the teaching work, and improves the efficiency of project deployment.

[0065] The teaching action is described above.

[0066] When the non-reference robot obtains the teaching action migrated from the reference robot, if the non-reference robot is currently working on the reference machine table, the non-reference robot can execute the teaching action migrated from the reference robot based on the hand-eye calibration information, the height deviation, and the TCP installation information deviation. That is, the hand-eye calibration information, the height deviation, and the TCP installation information deviation need to be considered when executing the teaching action.

[0067] If the reference robot is not currently working on the non-reference machine, the teaching motion migrated from the reference robot can be executed on the non-reference machine based on the obtained machine deviation of the non-reference machine relative to the reference machine, the hand-eye calibration information, the height deviation, and the TCP installation information deviation.

[0068] As an embodiment, the machine deviation is obtained by the reference robot working on the non-reference machine and the reference machine respectively. When the reference robot obtains the machine deviation, the machine deviation can be imported from the reference robot to each non-reference robot in batch. The following Figure 7a An example is given to show how to determine the machine deviation, which is not described here.

[0069] So far, the completion Figure 2 The flow is shown.

[0070] Through the Figure 2 The flow is shown, it can be seen that in the embodiment, the hand-eye calibration information of the non-reference robot, the height deviation of the installation height of the mechanical arm of the non-reference robot relative to the installation height of the mechanical arm of the reference robot, and the TCP installation information deviation of the TCP installation information of the non-reference robot relative to the TCP installation information of the reference robot are obtained by the non-reference robot, so that the teaching motion migrated from the reference robot can be executed based on the hand-eye calibration information, the height deviation, and the TCP installation information deviation, and each non-reference robot does not need to be taught the teaching motion, which simplifies the teaching workload and improves the efficiency of rapid online of the non-reference robot.

[0071] It can be seen from the Figure 2 The flow is shown, it can be seen that in the embodiment, the hand-eye calibration information of the non-reference robot, the height deviation of the installation height of the mechanical arm of the non-reference robot relative to the installation height of the mechanical arm of the reference robot, and the TCP installation information deviation of the TCP installation information of the non-reference robot relative to the TCP installation information of the reference robot are obtained by the non-reference robot, so that the teaching motion migrated from the reference robot can be executed based on the hand-eye calibration information, the height deviation, and the TCP installation information deviation, and each non-reference robot does not need to be taught the teaching motion, which simplifies the teaching workload and improves the efficiency of rapid online of the non-reference robot.

[0072] Further, the embodiment splits the whole teaching motion into atomic motions for reuse, which further simplifies the teaching workload and improves the efficiency of project deployment.

[0073] The following describes the Figure 3a The flow is shown.

[0074] Referring to Figure 3a , Figure 3aA hand-eye calibration flowchart is provided for the embodiments of the present application. It should be noted that, in the present embodiment, before hand-eye calibration, the installation and setting of TCP on the robot arm are first completed, and the robot pose information involved below is the TCP pose information in the robot coordinate system. Here, the TCP of the robot arm is a reference point on the end effector of the robot arm, which is used to determine the position and direction of the tool. The TCP is usually installed on the end effector of the robot arm, such as the center point of the gripper or the tip of the welding tool.

[0075] Based on the above description, as shown in Figure 3a , the flowchart can include the following steps:

[0076] Step 301: Send a collection instruction to the non-reference robot arm through the chassis of the non-reference robot, so that the robot arm moves based on the collection instruction and sends a notification to the chassis of the non-reference robot when it moves to the collection point.

[0077] Optionally, in the present embodiment, a host computer such as a PC is first opened, which is connected to the chassis of the non-reference robot through a network. When the host computer is successfully logged in, the host computer sends a hand-eye calibration debugging command to the chassis of the non-reference robot based on the current hand-eye calibration requirement, so that the non-reference robot enters a hand-eye calibration debugging mode based on the hand-eye calibration debugging command. Then, the host computer displays an interface as shown in Figure 3b .

[0078] When the start calibration on the above interface is selected, the host computer sends a start calibration instruction to the chassis of the non-reference robot, so that the chassis of the non-reference robot initiates a hand-eye calibration flowchart as shown in Figure 3a .

[0079] Specifically, as shown in the hand-eye calibration flowchart Figure 3a , in step 301, the chassis of the non-reference robot first sends the above collection instruction to the robot arm of the non-reference robot. When the robot arm of the non-reference robot receives the collection instruction, it starts to move and sends a notification to the chassis of the non-reference robot when it moves to the designated collection point. Finally, step 301 is realized.

[0080] It should be noted that in the present step 301, the robot arm is controlled by a built-in motion program, which controls how the robot arm moves. As an embodiment, in the present step 301, during the movement of the robot arm of the non-reference robot, the chassis of the non-reference robot is fixed, and the robot arm of the non-reference robot moves according to the control of the above motion program, such as stretching, rotating or lifting, etc.

[0081] Step 302: When the chassis of the non-reference robot receives a notification sent by the robotic arm of the non-reference robot, it sends a photo-taking command to the camera of the non-reference robot, so that the camera takes a picture of the calibration object based on the photo-taking command, obtains the current image, and outputs the template matching result of the current image to the chassis of the non-reference robot.

[0082] As one embodiment, the calibration object here is, for example, a Mark board. As another embodiment, the calibration object here can also be a feature mark of an object, such as a circle or dot on a Mark board; this embodiment is not specifically limited.

[0083] Optionally, in this embodiment, the template matching result is represented by (u, v, θ); (u, v) represents the deviation between the pixel position of the calibrator in the current image and the pixel position of the calibrator in the template image. θ represents the rotation angle deviation between the calibrator in the current image and the calibrator in the template image. This template matching result is used to determine the hand-eye calibration parameters, as detailed in step 303.

[0084] Step 303: Using the current pose information of the robotic arm of the non-reference robot collected from the chassis of the non-reference robot and the template matching results mentioned above, determine the current hand-eye calibration parameters, and determine the hand-eye calibration information of the non-reference robot based on the current hand-eye calibration parameters.

[0085] As an example, determining the current hand-eye calibration parameters using the current pose information of the robotic arm acquired from the chassis of a non-reference robot, along with the template matching results, may include: inputting the current pose information of the robotic arm and the template matching results into a pre-set hand-eye calibration algorithm to obtain the current hand-eye calibration parameters. The current hand-eye calibration parameters refer to the transformation relationship between the robotic arm coordinate system corresponding to the robotic arm of the non-reference robot and the camera coordinate system corresponding to the camera of the non-reference robot.

[0086] Once the current hand-eye calibration parameters are determined, the interface of the host computer will be updated, as shown below. Figure 3c As shown. Figure 3c As shown, the updated interface will display the robotic arm pose information used to determine the current hand-eye calibration parameters, as well as the template matching results mentioned above.

[0087] It should be noted that after the current pose information of the robot arm and the template matching result are input into the preset hand-eye calibration algorithm, the current hand-eye calibration parameter is obtained, and the confidence of the current hand-eye calibration parameter is further obtained. Based on this, as an embodiment, the hand-eye calibration information of the non-reference robot in step 303 is determined based on the current hand-eye calibration parameter, which can be implemented based on the confidence. For example, if the confidence of the current hand-eye calibration parameter meets the set confidence requirement, such as more than 90%, the current hand-eye calibration parameter is determined as the hand-eye calibration information; if the confidence of the current hand-eye calibration parameter does not meet the set confidence requirement, the step of sending the collection instruction to the robot arm by the chassis of the non-reference robot is returned until it is found that the robot arm has not traversed all the collection points. Of course, if it is found that the robot arm has traversed all the collection points, a failure message can be directly output at this time to indicate that the robot arm of the non-reference robot is adjusted or the collection points are adjusted, and the embodiment is not specifically limited.

[0088] After the hand-eye calibration information is determined, when the end calibration is selected as shown in Figure 3c , the host computer sends an end calibration instruction to the chassis of the non-reference robot to make the chassis of the non-reference robot end the hand-eye calibration based on the end calibration instruction.

[0089] At this point, the process shown in Figure 3a is completed.

[0090] The hand-eye calibration of the non-reference robot is implemented through the process shown in Figure 3a .

[0091] The process shown in Figure 4a will be described below.

[0092] Referring to Figure 4a , Figure 4a , a height calibration flowchart provided by the embodiment of the present application is provided. The height calibration flowchart is mainly used to calibrate the height deviation of the installation height of the robot arm of the non-reference robot relative to the installation height of the robot arm of the reference robot. In specific applications, the installation height deviation of the robot arm between the non-reference robot and the reference robot affects the shooting accuracy of the camera on the non-reference robot and the grabbing height of the robot arm on the non-reference robot. For scenes with high accuracy requirements, the installation height deviation of the robot arm needs to be calibrated before the teaching motion is migrated.

[0093] As shown in Figure 4a , the process can include the following steps:

[0094] At step 401, the height deviation calibration instruction is sent to the robot arm by the chassis of the non-reference robot, so that the robot arm moves in a fixed environment based on the height deviation calibration instruction and sends a notification to the chassis of the non-reference robot when it moves to the designated shooting position.

[0095] Optionally, in this embodiment, a host computer such as a PC is first opened, which is connected to the chassis of the non-reference robot through a network. When the host computer is successfully logged in, the host computer sends a height deviation debugging command to the chassis of the non-reference robot based on the current height deviation calibration requirement, so that the non-reference robot enters a height deviation debugging mode based on the height deviation debugging command. Then, the host computer displays an interface as shown in Figure 4b .

[0096] When the start calibration on the above interface is selected, the host computer sends a start calibration instruction to the chassis of the non-reference robot, so that the chassis of the non-reference robot initiates a height deviation calibration process as shown in Figure 4a based on the start calibration instruction.

[0097] Specifically, as shown in the hand-eye calibration process in Figure 4a , at step 401, the chassis of the non-reference robot first sends the height deviation calibration instruction to the robot arm of the non-reference robot. When the robot arm of the non-reference robot receives the height deviation calibration instruction, it starts to move in a fixed environment and sends a notification to the chassis of the non-reference robot when it moves to the designated shooting position. Finally, step 401 is realized.

[0098] It should be noted that in this step 401, the robot arm is controlled by a built-in motion program, which controls how the robot arm moves. As an example, in this step 401, the chassis of the non-reference robot is fixed during the movement of the robot arm of the non-reference robot, and the robot arm of the non-reference robot moves according to the control of the motion program, such as extension, rotation or lifting, etc.

[0099] As an example, the fixed environment described above can be the environment before the non-reference robot is shipped, or a reference machine, and this embodiment does not specifically limit it.

[0100] At step 402, when the chassis of the non-reference robot receives the notification sent by the robot arm of the non-reference robot, a shooting instruction is sent to the camera of the non-reference robot, so that the camera shoots the calibration object based on the shooting instruction to obtain an image and outputs it to the chassis of the non-reference robot.

[0101] Here, the calibration object is as described above.

[0102] Step 403, comparing the calibration object in the image with the calibration object in the obtained reference image to determine the height deviation of the mounting height of the robot arm of the non-reference robot relative to the mounting height of the robot arm of the reference robot. The reference image refers to an image obtained by photographing the calibration object by the camera of the reference robot at the photographing position in the fixed environment.

[0103] Optionally, in the embodiment, there are differences between the calibration object in the image and the calibration object in the obtained reference image, such as differences in area size, differences in center position, etc. Based on the differences, the height deviation algorithm set in advance is input, and the height deviation is obtained.

[0104] After the height deviation is determined, the interface of the host computer at least displays the robot arm pose information and the height deviation, as shown in Figure 4c When the end calibration on the interface is selected, the host computer sends an end calibration instruction to the chassis of the non-reference robot to make the chassis of the non-reference robot end the mounting height deviation calibration of the robot arm based on the end calibration instruction.

[0105] At this point, the process shown in Figure 4a is completed.

[0106] Through the process shown in Figure 4a , the height deviation of the mounting height of the robot arm of the non-reference robot relative to the mounting height of the robot arm of the reference robot is calibrated.

[0107] The process shown in Figure 5a will be described below.

[0108] Referring to Figure 5a , Figure 5a , a TCP installation deviation calibration flowchart provided by the embodiment of the present application. The flowchart is mainly used to calibrate the TCP installation information deviation of the TCP installation information of the non-reference robot relative to the TCP installation information of the reference robot. The TCP installation information deviation of the TCP installation information of the non-reference robot relative to the TCP installation information of the reference robot will affect the migration of the teaching motion in the high-precision grabbing scene. If the TCP installation information deviation cannot be controlled within millimeter precision, the grabbing motion of the non-reference robot will be inaccurate. Based on this, the embodiment calibrates the TCP installation information deviation of the TCP installation information of the non-reference robot relative to the TCP installation information of the reference robot to avoid the above defects.

[0109] As shown in Figure 5a , the flowchart can include the following steps:

[0110] Step 501, receiving an externally input motion parameter.

[0111] Optionally, in this embodiment, a host computer such as a PC is first opened, which is connected with the chassis of the non-reference robot through a network. When the host computer is successfully logged in, the host computer sends a TCP installation bias debugging command to the chassis of the non-reference robot based on the TCP installation bias requirement, so that the non-reference robot enters the TCP installation bias debugging mode based on the TCP installation bias debugging command. Then, the host computer displays an interface as shown in Figure 5b .

[0112] When the action parameter is input on the interface as shown in Figure 5b , the host computer will transmit the action parameter to the chassis of the non-reference robot, that is, the non-reference robot receives the externally input action parameter as described in step 501. As shown in Figure 5b , the action parameter may include an action identification (ID) such as taking a photo 3000, grabbing 3000, placing 3000, etc.

[0113] In step 502, the TCP calibration instruction is sent to the mechanical arm of the non-reference robot through the chassis of the non-reference robot, so that the mechanical arm of the non-reference robot moves based on the TCP calibration instruction and sends a notification to the chassis of the non-reference robot when it moves to the designated calibration point.

[0114] In this embodiment, when the start calibration on the interface as shown in Figure 5b displayed by the host computer is selected, the host computer sends a start calibration instruction to the non-reference robot, so that the non-reference robot performs step 502 based on the start calibration instruction.

[0115] In step 503, after receiving the notification sent by the mechanical arm of the non-reference robot, the chassis of the non-reference robot collects the current pose information of the mechanical arm (denoted as first pose information).

[0116] In step 504, the pose of the mechanical arm is adjusted in response to the above-mentioned pose adjustment to collect the current second pose information of the mechanical arm.

[0117] In this embodiment, when the mechanical arm performs the action corresponding to the action parameter such as the grabbing action at the calibration point, there will be a certain bias. Taking the grabbing action as an example, the bias is the grabbing bias. For this bias, the embodiment can adaptively adjust the pose of the mechanical arm to avoid the above-mentioned bias. That is, here, the pose of the mechanical arm is adjusted due to the bias when the mechanical arm performs the action corresponding to the action parameter at the calibration point.

[0118] After completing the adjustment of the pose of the mechanical arm, the current pose information of the mechanical arm of the non-reference robot (denoted as second pose information) is also collected through the chassis of the non-reference robot.

[0119] Step 505: Based on the first pose information and the second pose information mentioned above, obtain the TCP installation information deviation of the non-reference robot relative to the TCP installation information of the reference robot.

[0120] As an example, step 505 can be implemented using the chassis of a non-reference robot. Specifically, step 505 can obtain the TCP installation information deviation of the non-reference robot relative to the TCP installation information of the reference robot based on the aforementioned first pose information and the aforementioned second pose information, combined with a TCP calibration algorithm. For example, the aforementioned first pose information and the aforementioned second pose information can be input into the TCP calibration algorithm to obtain the aforementioned TCP installation information deviation.

[0121] It should be noted that the above TCP installation information deviation includes position information (x, y, z), which also includes attitude information (rx, ry, rz).

[0122] After determining the above-mentioned deviations in TCP installation information, such as Figure 5b The interface shown will then update to look like this. Figure 5c The interface shown. Compared to, as... Figure 5b The interface shown is as follows: Figure 5c The interface shown now includes the aforementioned TCP installation information discrepancies, as detailed below. Figure 5c As shown. When as Figure 5c When the end calibration on the interface shown is triggered, the host computer will send an end calibration command to the non-baseline robot, so that the non-baseline robot can end the TCP installation information deviation calibration based on the end calibration command.

[0123] This concludes the process. Figure 5a The process is shown below.

[0124] pass Figure 5a The process shown obtained the deviation of the TCP installation information of the non-baseline robot from the TCP installation information of the baseline robot.

[0125] The above description of the method provided in the embodiments of this application is from the perspective of a non-reference robot. The following description is from the perspective of a reference robot:

[0126] See Figure 6 , Figure 6 Another method flowchart provided for an embodiment of this application. This method is applied to a benchmark robot. For example... Figure 6 As shown, the process may include the following steps:

[0127] In step 601, a teaching motion of a reference robot collected on a reference machine table is obtained; the teaching motion at least includes an atomic motion split from a preset complete set of motions for reuse.

[0128] The teaching motion is described in step 203 above, and will not be repeated here.

[0129] In step 602, a table deviation of each non-reference machine table relative to the reference machine table measured by the reference robot is obtained.

[0130] Optionally, in the embodiment, the reference robot measures the table deviation of each non-reference machine table relative to the reference machine table on each non-reference machine table. In a specific implementation, the table deviation of any non-reference machine table relative to the reference machine table can include: a Mark plate height deviation from the ground; here, the Mark plate height deviation from the ground refers to the deviation of the height of the Mark plate on the non-reference machine table from the ground relative to the height of the Mark plate on the reference machine table; or,

[0131] The above-mentioned Mark plate height deviation from the ground, and a Mark plate position deviation; here, the Mark plate position deviation refers to the deviation between the position of the Mark plate on the non-reference machine table in a two-dimensional (2D) plane and the position of the Mark plate on the reference machine table in the 2D plane.

[0132] As an embodiment, the above-mentioned Mark plate height deviation from the ground aims to combine the height deviation of the mounting height of the mechanical arm of the non-reference robot relative to the mounting height of the mechanical arm of the reference robot to compensate for the photographing height of the camera of the non-reference robot, so that the height of the camera of each non-reference robot to the Mark plate on each machine table (including the reference machine table and the non-reference machine table) is a fixed value in the multi-robot and multi-machine table operation.

[0133] As an embodiment, to obtain the above-mentioned Mark plate height deviation from the ground, the embodiment will first use the reference robot to collect an image of the Mark plate on each non-reference machine table of each non-reference machine table of the same type as the machine table, and then the chassis of the reference robot combines the image with the image of the Mark plate collected on the reference machine table of the same type as the machine table and calculates the above-mentioned Mark plate height deviation from the ground by using an algorithm, as shown in Figure 7a .

[0134] In the embodiment, after the deviation of the mark plate height from the ground is calculated, if it is found that the position of the mark plate on the non-reference machine in the two-dimensional plane still deviates from the position of the mark plate on the reference machine in the two-dimensional plane (i.e., the mark plate position deviation), the deviation between the position of the mark plate on the non-reference machine in the two-dimensional plane and the position of the mark plate on the reference machine in the two-dimensional plane (i.e., the mark plate position deviation) needs to be further calibrated. For details, see Figure 7b

[0135] In the embodiment, the mark plate is a means for positioning the mechanical arm on the visual auxiliary robot, and is generally attached to the machine. For example, a mark plate is attached beside the box (OC: Open Casette) on the machine for placing the silicon wafer.

[0136] In the embodiment, the teaching motion obtained by the reference robot and the machine deviation will be migrated to each non-reference robot, so that each non-reference robot performs the teaching motion based on the obtained hand-eye calibration information, the height deviation of the mechanical arm mounting height of each non-reference robot from the mechanical arm mounting height of the reference robot, the TCP mounting information deviation of the TCP mounting information of each non-reference robot from the TCP mounting information of the reference robot, and the machine deviation.

[0137] It should be noted that in the embodiment, the machine deviation can be stored as shared configuration information of the reference robot to the chassis of the reference robot. The shared configuration information can be allowed to be migrated in batches to each non-reference robot, for example, as shared configuration information to be migrated and stored to the chassis of each non-reference robot.

[0138] It should be further noted that the shared configuration information stored by the reference robot can also be updated according to actual needs, such as updating due to the addition of a non-reference platform. When the shared configuration information stored by the reference robot is updated, the updated shared configuration information will be migrated to each non-reference robot again and overwrite the original shared configuration information stored by each non-reference robot.

[0139] At this point, the process shown in Figure 6

[0140] By Figure 6 ​​As shown in the flow, in the embodiment, the reference robot collects each atomic operation that can be reused, measures the machine table deviation of each non-reference machine table relative to the reference machine table, and then imports each atomic operation that can be reused and the machine table deviation into each non-reference robot, so that each non-reference robot combines the machine table deviation and performs the teaching action migrated from the reference robot based on the difference between each non-reference robot and the reference robot (such as the height deviation of the mounting height of the mechanical arm of the non-reference robot relative to the mounting height of the mechanical arm of the reference robot, and the TCP installation information deviation of the TCP installation information of the non-reference robot relative to the TCP installation information of the reference robot), which realizes the need to no longer teach each non-reference robot the teaching action, simplifies the teaching workload, and improves the efficiency of the rapid online of the non-reference robot.

[0141] The Mark plate height deviation from the ground is described as follows:

[0142] Referring to Figure 7a , Figure 7a The Mark plate height deviation calibration flowchart is provided for the embodiment of the application. The flow is applied to a reference robot, as shown in Figure 7a , the flow can include the following steps:

[0143] Step 701a, obtaining a first Mark plate image; the first Mark plate image is an image of a Mark plate on a non-reference machine table collected by a camera of the reference robot when the mechanical arm of the reference robot moves to a photographing position of the non-reference machine table.

[0144] Optionally, in the embodiment, a host computer such as a PC is first opened, and the host computer is connected to the chassis of the reference robot through a network. When the host computer is successfully logged in, the host computer sends a Mark plate height deviation from the ground debugging command to the chassis of the reference robot based on the Mark plate height deviation from the ground calibration requirement, so that the reference robot enters a Mark plate height deviation from the ground debugging mode based on the Mark plate height deviation from the ground debugging command. Then, the host computer displays an interface as shown in Figure 7c .

[0145] When the interface as shown in Figure 7cWhen the start calibration on the interface of the upper host is selected, the upper host sends a start calibration instruction to the reference robot, so that the reference robot sends a Mark plate off-ground height deviation calibration instruction to the mechanical arm of the reference robot based on the start calibration instruction through the chassis of the reference robot. Then, the mechanical arm moves on the non-reference machine based on the Mark plate off-ground height deviation calibration instruction, and sends a notification to the chassis of the reference robot when it moves to the shooting position on the non-reference machine. When the chassis of the reference robot receives the notification sent by the mechanical arm, it sends a shooting instruction to the camera of the reference robot, so that the camera acquires an image based on the shooting instruction, i.e., the first Mark plate image described above, and outputs it to the chassis of the reference robot. That is, the first Mark plate image is finally obtained as described in step 701a.

[0146] It should be noted that the process of obtaining the first Mark plate image involves the hand-eye calibration information of the reference robot, so as to obtain the first Mark plate image based on the hand-eye calibration information of the reference robot. For the hand-eye calibration information of the reference robot, please refer to the process of obtaining the hand-eye calibration information of the non-reference robot shown in Figure 3a The process of obtaining the hand-eye calibration information of the non-reference robot will not be described here.

[0147] In step 702a, the height difference of the Mark plate in the first Mark plate image and the obtained second Mark plate image is determined to be the Mark plate off-ground height deviation.

[0148] In this embodiment, the second Mark plate image is an image of a Mark plate on the reference machine table acquired by the camera of the reference robot when the mechanical arm of the reference robot moves to the shooting position of the reference machine table. It is similar to the process of acquiring the image of the Mark plate on the non-reference machine table by the camera of the reference robot. It should be noted that if multiple Mark plates are deployed on the non-reference machine table and the reference machine table, the Mark plates on the non-reference machine table and the reference machine table can be assigned with the same identifier in the same way. Correspondingly, the Mark plates in the first Mark plate image and the second Mark plate image described above can be the Mark plates with the same identifier.

[0149] As an embodiment, the height difference of the Mark plate in the first Mark plate image and the obtained second Mark plate image can be directly determined as the Mark plate off-ground height deviation.

[0150] At this point, the process shown in Figure 7a is completed.

[0151] Through the process shown in Figure 7a , the calibration of the Mark plate off-ground height deviation can be completed.

[0152] The Mark plate position deviation calibration described above is described as follows:

[0153] Referring to Figure 7b , Figure 7b A Mark plate position deviation calibration flowchart is provided in the embodiments of the present application. The flowchart is applied to a reference robot. The reference robot is in a non-reference machine.

[0154] As shown in Figure 7b , the flowchart can include the following steps:

[0155] In step 701b, the base plate of the reference robot sends the Mark plate height-from-ground deviation to the mechanical arm of the reference robot, so that the mechanical arm adjusts the shooting height of the camera of the reference robot based on the Mark plate height-from-ground deviation.

[0156] That is, in the present embodiment, after the Mark plate height-from-ground deviation is obtained through the flowchart shown in Figure 7a , the Mark plate height-from-ground deviation is sent to the mechanical arm, so that the mechanical arm re-adjusts the shooting height of the camera of the reference robot based on the Mark plate height-from-ground deviation, to ensure that the camera can accurately collect the image of the Mark plate on the non-reference machine.

[0157] In step 702b, the base plate of the reference robot collects the current pose information (denoted as third pose information) of the mechanical arm of the reference robot.

[0158] In step 703b, the pose of the mechanical arm of the reference robot is adjusted in response to the current pose information (denoted as fourth pose information) of the mechanical arm.

[0159] In the present embodiment, because the position deviation calibration described above has not been completed, when the mechanical arm of the reference robot performs a specified action such as a grabbing action, it will have a certain deviation. Based on this, the present embodiment can adjust the pose of the mechanical arm to avoid the deviation. That is, the pose adjustment of the mechanical arm is caused by the deviation of the mechanical arm when performing the specified action after adjusting the shooting height of the camera. After adjusting the pose of the mechanical arm, the current pose information of the mechanical arm, i.e., the fourth pose information described above, can be collected.

[0160] It should be noted that, as described above, in the present step 703b, the pose adjustment of the mechanical arm involves the hand-eye calibration information of the reference robot, which unifies the mechanical arm and the camera of the reference robot based on the hand-eye calibration information of the reference robot, so that the camera of the reference robot can take pictures, and the mechanical arm of the reference robot can perform actions such as grabbing, placing, etc. based on the image output by the camera.

[0161] In step 704b, the Mark plate position deviation is determined based on the third pose information and the fourth pose information.

[0162] Optionally, in the embodiment, the third pose information and the fourth pose information are input into the Mark plate position deviation calibration algorithm to obtain the Mark plate position deviation.

[0163] At this point, the calibration of the Mark plate position deviation is completed. Figure 7b

[0164] Figure 7b The calibration of the Mark plate position deviation is achieved through the above process.

[0165] After the calibration of the Mark plate height deviation and the Mark plate position deviation is completed, the host computer displays an interface as shown in Figure 7d The interface contains the Mark plate height deviation and the Mark plate position deviation.

[0166] When the end calibration on the interface as shown in Figure 7d is selected, the host computer sends an end calibration instruction to the reference robot, so that the non-reference robot ends the calibration of the Mark plate height deviation and the Mark plate position deviation based on the end calibration instruction.

[0167] It can be seen that in the embodiment, when the reusable atomic operations collected by the reference robot and the machine table deviations of each non-reference machine table relative to the reference machine table measured by the reference robot are migrated to each non-reference robot, each non-reference robot combines the machine table deviation and executes the teaching motion migrated from the reference robot based on the differences between each non-reference robot and the reference robot (such as the height deviation of the arm mounting height of the non-reference robot relative to the arm mounting height of the reference robot, and the TCP mounting information deviation of the TCP mounting information of the non-reference robot relative to the TCP mounting information of the reference robot). This achieves the teaching of the teaching motion without the need to teach each non-reference robot, simplifies the teaching workload, and improves the efficiency of the rapid online of the non-reference robot.

[0168] ​​For ease of understanding, the hand-eye calibration information process of the reference robot is briefly described as follows: the chassis of the reference robot sends a collection instruction to the robot arm of the reference robot, so that the robot arm of the reference robot moves based on the collection instruction and sends a notification to the chassis of the reference robot when it moves to the collection point; when the chassis of the reference robot receives the notification sent by the robot arm of the reference robot, the chassis sends a photographing instruction to the camera of the reference robot, so that the camera photographs the calibration object based on the photographing instruction to obtain a current image and outputs a template matching result of the current image to the chassis of the reference robot; the current hand-eye calibration parameter is determined based on the current pose information of the robot arm of the reference robot collected by the chassis of the reference robot and the template matching result; and the hand-eye calibration information of the reference robot is determined according to the current hand-eye calibration parameter.

[0169] In the embodiment, the template matching result is as described above and can be represented by (u, v, θ); (u, v) represents the deviation of the pixel position of the calibration object in the current image from the pixel position of the calibration object in the template image, and θ represents the rotational angle deviation of the calibration object in the current image from the calibration object in the template image.

[0170] In the embodiment, the hand-eye calibration information of the reference robot is determined according to the current hand-eye calibration parameter, including: if the confidence of the current hand-eye calibration parameter meets the set confidence requirement, the current hand-eye calibration parameter is determined as the hand-eye calibration information of the reference robot; if the confidence of the current hand-eye calibration parameter does not meet the set confidence requirement, the step of sending a collection instruction to the robot arm of the reference robot by the chassis of the reference robot is returned when the robot arm of the reference robot has not traversed all the collection points by the time the current hand-eye calibration parameter is found. The hand-eye calibration information process of the reference robot can refer to the hand-eye calibration information process of the non-reference robot.

[0171] The method provided by the embodiment of the application is described above, and the robot provided by the embodiment of the application is described below:

[0172] The robot provided by the embodiment of the application is as described above, and can include a robot body, a robot arm, and a camera. The body includes a processor, which is used to execute the steps in the flow shown in Figure 2 or the flow shown in Figure 6 .

[0173] Based on the same application concept as the above method, the embodiment of the application further provides a computer program product, which stores a computer program. The computer program is executed by a processor to implement the method disclosed in the above examples.

[0174] The above merely provides an example of the present application, and is not intended to limit the present application. Any modification, equivalent replacement, improvement, etc. within the spirit and principle of the present application should be included in the scope of claims of the present application.

Claims

1. A teaching migration method applied to a multi-robot multi-machine, characterized by, The method is applied to a non-reference robot, and the method comprises: obtaining hand-eye calibration information of the non-reference robot; the hand-eye calibration information comprises a conversion relationship between a robot arm coordinate system corresponding to a robot arm of the non-reference robot and a camera coordinate system corresponding to a camera of the non-reference robot; obtaining a height deviation of a robot arm installation height of the non-reference robot relative to a robot arm installation height of a reference robot, and obtaining a tool center point (TCP) installation information deviation of TCP installation information of the non-reference robot relative to TCP installation information of the reference robot; performing teaching actions migrated from the reference robot based on the hand-eye calibration information, the height deviation and the TCP installation information deviation; the teaching actions are teaching actions collected by the reference robot; the teaching actions are atomic actions split from a preset complete set of actions and used for reuse.

2. The method of claim 1, wherein, The hand-eye calibration information of the non-reference robot is obtained by: sending a collection instruction to the robot arm of the non-reference robot through the chassis of the non-reference robot, so that the robot arm moves based on the collection instruction and sends a notification to the chassis of the non-reference robot when moving to a collection point; when the chassis of the non-reference robot receives the notification sent by the robot arm of the non-reference robot, sending a photographing instruction to the camera of the non-reference robot, so that the camera photographs a calibration object based on the photographing instruction to obtain a current image and outputs a template matching result of the current image to the chassis of the non-reference robot; the template matching result is represented by (u, v, θ); (u, v) represents a deviation between a pixel position of the calibration object in the current image and a pixel position of the calibration object in a template image, and θ represents a rotation angle deviation of the calibration object in the current image relative to the calibration object in the template image; determining a current hand-eye calibration parameter based on current pose information of the robot arm collected by the chassis of the non-reference robot and the template matching result; determining the hand-eye calibration information of the non-reference robot according to the current hand-eye calibration parameter.

3. The method of claim 2, wherein, The determination of the current hand-eye calibration parameter further comprises determining a confidence degree of the current hand-eye calibration parameter; The determination of the hand-eye calibration information of the non-reference robot according to the current hand-eye calibration parameter comprises: if the confidence degree of the current hand-eye calibration parameter meets a set confidence degree requirement, the current hand-eye calibration parameter is determined as the hand-eye calibration information; if the confidence degree of the current hand-eye calibration parameter does not meet the set confidence degree requirement, the step of sending the collection instruction to the robot arm through the chassis of the non-reference robot is returned when it is found that the robot arm has not traversed all the collection points by the time being.

4. The method of claim 1, wherein, The height deviation of the robot arm installation height of the non-reference robot relative to the robot arm installation height of the reference robot is obtained by: sending a height deviation calibration instruction to the robot arm through the chassis of the non-reference robot, so that the robot arm moves in a fixed environment based on the height deviation calibration instruction and sends a notification to the chassis of the non-reference robot when moving to a designated photographing position; When the non-reference robot chassis receives the notification sent by the robot arm, the non-reference robot chassis sends a photographing instruction to the camera of the non-reference robot, so that the camera photographs the calibration object to obtain an image based on the photographing instruction and outputs the image to the non-reference robot chassis; The image of the calibration object is compared with the calibration object in the obtained reference image to determine the height deviation of the installation height of the robot arm of the non-reference robot relative to the installation height of the robot arm of the reference robot. The reference image refers to an image obtained by photographing the calibration object by the camera of the reference robot at the photographing position in the fixed environment.

5. The method of claim 1, wherein, The TCP installation information deviation of the TCP installation information of the non-reference robot relative to the TCP installation information of the reference robot includes: Receiving an externally input action parameter; The non-reference robot chassis sends a TCP calibration instruction to the robot arm of the non-reference robot, so that the robot arm of the non-reference robot moves based on the TCP calibration instruction and sends a notification to the non-reference robot chassis when it moves to the designated calibration point; After the non-reference robot chassis receives the notification sent by the robot arm of the non-reference robot, the non-reference robot chassis collects the first pose information of the robot arm at present; In response to the pose adjustment of the robot arm, the second pose information of the robot arm at present is collected by the non-reference robot chassis; the pose adjustment of the robot arm is caused by the deviation of the robot arm when performing the action corresponding to the action parameter at the calibration point; Based on the first pose information and the second pose information, the TCP installation information deviation of the TCP installation information of the non-reference robot relative to the TCP installation information of the reference robot is obtained.

6. The method of claim 1, wherein, The execution of the teaching action migrated from the reference robot based on the hand-eye calibration information, the height deviation and the TCP installation information deviation includes: If the non-reference robot is currently working on the reference machine table, the teaching action migrated from the reference robot is executed on the reference machine table based on the hand-eye calibration information, the height deviation and the TCP installation information deviation; If the non-reference robot is currently working on the non-reference machine table, the teaching action migrated from the reference robot is executed on the non-reference machine table based on the obtained machine table deviation of the non-reference machine table relative to the reference machine table and the hand-eye calibration information, the height deviation and the TCP installation information deviation.

7. A teaching migration method applied to a multi-robot multi-machine platform, characterized in that, The method is applied to a reference robot, and the method includes: Obtaining a teaching action collected by the reference robot on a reference machine table; the teaching action at least includes an atomic action used for reuse which is split from a preset complete set of actions; Obtaining machine table deviations of each non-reference machine table relative to the reference machine table measured by the reference robot; The teaching action and the machine table deviation are migrated to each non-reference robot, so that each non-reference robot performs the teaching action based on hand-eye calibration information obtained by each non-reference robot, a height deviation of a mounting height of a mechanical arm of each non-reference robot relative to a mounting height of a mechanical arm of the reference robot, a TCP mounting information deviation of TCP mounting information of each non-reference robot relative to TCP mounting information of the reference robot, and the machine table deviation.

8. The method of claim 7, wherein, The machine table deviation at least includes: a Mark plate height deviation, which refers to a deviation of a height of a Mark plate on a non-reference machine table relative to a height of a Mark plate on a reference machine table; or the Mark plate height deviation and a Mark plate position deviation, which refers to a deviation between a position of a Mark plate on a non-reference machine table in a two-dimensional plane and a position of a Mark plate on a reference machine table in the two-dimensional plane.

9. The method of claim 8, wherein, The Mark plate height deviation is determined by: obtaining a first Mark plate image, which is an image of a Mark plate on a non-reference machine table captured by a camera of the reference robot when a mechanical arm of the reference robot moves to a photographing position of the non-reference machine table; determining the Mark plate height deviation based on a height difference of the Mark plate in the first Mark plate image and a second Mark plate image obtained, which is an image of a Mark plate on a reference machine table captured by the camera of the reference robot when the mechanical arm of the reference robot moves to a photographing position of the reference machine table.

10. The method of claim 8, wherein, The Mark plate position deviation is determined by: sending, by a chassis of the reference robot, the Mark plate height deviation to a mechanical arm of the reference robot, so that the mechanical arm adjusts a photographing height of a camera of the reference robot based on the Mark plate height deviation; obtaining, by the chassis of the reference robot, third pose information of the mechanical arm at present; in response to a pose adjustment of the mechanical arm, obtaining, by the chassis of the reference robot, fourth pose information of the mechanical arm at present, the pose adjustment of the mechanical arm being caused by a deviation of the mechanical arm after adjusting the photographing height of the camera and performing a specified action; determining the Mark plate position deviation based on the third pose information and the fourth pose information.

11. The method of claim 1 or 7, wherein, The teaching action at least includes: a photographing action, which refers to an action of the mechanical arm of the reference robot moving from a specified suspended position to a specified photographing position in a scenario of taking goods from or placing goods on a machine table; a grabbing action, which refers to an action of the mechanical arm of the reference robot moving from the specified photographing position to a specified grabbing position on the machine table in a scenario of taking goods from the machine table; a placing action, which refers to an action of the mechanical arm of the reference robot moving above a goods placing point on the machine table after grabbing goods and then placing the goods on the machine table in a scenario of placing goods on the machine table.

12. A robot, characterized in that The robot comprises at least a body, a robotic arm, a camera; The body comprises a chassis of the robot, the chassis comprising at least a processor, the processor being used to perform steps in the method of any of claims 1 to 6 when the robot is a non-reference robot, and being used to perform steps in the method of any of claims 7 to 11 when the robot is a reference robot.

Images