Robotic object movement systems and methods for movement of objects

Abstract

A robot object moving system includes an image acquisition module for acquiring an image of an object, an image processing module for processing the image to determine a pick-up position of the object, a pick-up module for picking up the object at the pick-up position, a sensing module for sensing the pick-up position of the object, and a control module for controlling the sensing module to sense the pick-up position of the object. The sensing module senses sensing data of at least two dimensions of the pickup module, the control module determines whether the pickup position needs to be adjusted by using a first pre-training model with a first standard according to the sensing data, and the control module adjusts the pickup position by using a second pre-training model with a second standard according to the sensing data. The pick-up module is used for adjusting the pick-up position to obtain the adjusted pick-up position and picking up the object according to the adjusted pick-up position.

Description

Technical Field

[0001] This case is about a robotic object moving system and method for moving objects, especially a robotic object moving system and method that can be used to move objects with unknown weight distribution. Background Art

[0002] In the prior art, robotic systems are often used in industrial applications because they can handle tasks that are relatively complex for humans (such as manufacturing and logistics, etc.), and the robotic system can be a robot or a robotic arm. And due to the rapid development of related technologies, the movement of robotic systems can be controlled automatically without continuous manual control. For example, a robotic arm system can be used to pick up boxes in a factory area or a logistics site, so as to move smaller objects or goods in the boxes from one location to another. When the robotic arm system detects a box that needs to be picked up, the robotic arm executes a picking task according to a preset program to pick up the box.

[0003] In a traditional robotic arm system, a camera is used to take a photo of the box, and then the picking position of the box is determined based on the information obtained from the photo. For example, the robotic arm system obtains the center position of the box through the photo of the box, and then picks up the box at the position of the center of the box surface. However, in practical applications, the robotic arm system needs to handle boxes of various different sizes and weights, and it cannot judge the weight distribution of the box from the photo taken by the camera. Therefore, it may cause instability during the picking and moving process of the box. The smaller objects or goods in the box may move positions and collide with each other. Even more seriously, the box may shake or fall off the robotic arm system, resulting in serious damage or loss. Since the sizes and weight distributions of different boxes are different, the picking method of using the center position of the box as the picking position is not applicable to all boxes. Specifically, when picking up the box at the center position of the box, since the weight distribution of the box is not evenly distributed, the box may tilt or shake during the picking process, resulting in the box falling or being damaged. In addition, even if the robotic arm system can pick up all boxes at the center position, a larger clamping force needs to be used to avoid instability during the picking process.

[0004] Therefore, how to develop a robotic object moving system and method that can improve the defects of the above-mentioned prior art is an urgent need at present. Summary of the Invention

[0005] The object of the present invention is to provide a robotic object moving system and method for moving objects with unknown weight distribution. In the system and method of the present invention, the picking position of the object is adjusted based on the sensing data obtained during the initial picking attempt, thereby improving the stability of picking objects with unfixed sizes and uneven weights. In addition, compared with the traditional method of using the center position of the object surface as the picking position, when picking the object at the adjusted picking position, the present invention can pick the object with a smaller picking force, and the adjusted picking position is closer to the center of gravity of the object.

[0006] According to an embodiment of the present invention, there is provided a robotic object moving method for moving objects with unknown weight distribution, comprising the following steps: obtaining at least one image of an object placed at a starting position; performing image processing on the image to determine a pickable area on a first surface of the object and a first position as the picking position, wherein the first position is the center of the pickable area; using a picking module to pick the object at the picking position on the first surface of the object with a preset first-level picking force, wherein the picking distance and time for picking the object are a preset distance and a first time period respectively; obtaining sensing data of at least two dimensions of the picking module during the first time period; determining whether to adjust the picking position according to the sensing data and using a first pre-trained model with a first criterion; adjusting the picking position according to the sensing data and the pickable area on the first surface of the object and using a second pre-trained model with a second criterion to obtain an adjusted picking position; and picking the object at the adjusted picking position and moving the object from the starting position to a destination position.

[0007] According to another embodiment of the present invention, the present invention provides a robotic object moving system, including an image acquisition module, an image processing module, a picking module, a sensing module, and a control module. The image acquisition module is used to acquire at least one image of an object placed at a starting position. The image processing module is connected to the image acquisition module and is used to perform image processing on the image to determine a pickable area on the first surface of the object and a first position as the picking position, where the first position is the center of the pickable area. The picking module is used to pick up the object at the picking position on the first surface of the object with a preset first-level picking force, where the picking distance and time for picking up the object are a preset distance and a first time period respectively. The sensing module is connected to the picking module and is used to sense sensing data of at least two dimensions of the picking module during the first time period. The control module is connected to the image acquisition module, the image processing module, the picking module, and the sensing module, and is used to control the picking module. Wherein, the control module receives the processed image and the sensing data, and determines whether to adjust the picking position according to the sensing data and by using a first pre-trained model with a first criterion. Wherein, the control module adjusts the picking position according to the sensing data and the pickable area on the first surface of the object, and by using a second pre-trained model with a second criterion, and obtains the adjusted picking position. The picking module picks up the object according to the adjusted picking position to move the object from the starting position to the destination position. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] Figure 1 FIG. is a schematic structural diagram of a robotic object moving system according to a preferred embodiment of the present invention.

[0009] Figure 2 is Figure 1 a schematic structural diagram of the robotic object moving system.

[0010] Figure 3 is Figure 2 a top view of the object, where there is a certain distance between the picking position A and the adjusted picking position B.

[0011] Figure 4A FIG. is a schematic distribution diagram of the torsion value and the force value sensed by the sensing module. Figure 4B FIG. is a schematic distribution diagram of dividing the torsion value and the force value sensed by the sensing module into three intervals by using a critical point.

[0012] Figure 5 FIG. is a flowchart of a robotic object moving method for moving an object with an unknown weight distribution according to a preferred embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0013] Some exemplary embodiments that can embody the features and advantages of the present invention will be described in detail in the following description. It should be understood that the present invention can have various variations in different aspects and all do not depart from the scope of the present invention, and the descriptions and illustrations therein are for illustrative purposes in nature and not intended to limit the present case.

[0014] Figure 1 FIG. is a schematic diagram of the system structure of the robot object moving system 1 according to a preferred embodiment of the present invention. Figure 2 is Figure 1 a schematic diagram of the structure of the robot object moving system 1. Figure 3 is Figure 2 a top view of the object 7, in which there is a certain distance between the picking position A and the adjusted picking position B. As Figure 1 , Figure 2 and Figure 3 shown, the robot object moving system 1 includes an image acquisition module 2, an image processing module 3, a picking module 4, a sensing module 5, and a control module 6. In an embodiment of the present invention, the robot object moving system 1 performs object picking at the initial picking position in the initial picking attempt, then obtains the weight distribution information of the object 7 according to the sensing data obtained by the sensing module 5, and then uses the control module 6 to adjust the picking position to the optimal picking position according to the weight distribution information. Finally, the picking task is performed according to the adjusted picking position. The sensing data obtained by the sensing module 5 can be used not only to adjust the picking position but also to adjust the picking force when picking an object.

[0015] The image acquisition module 2 is used to acquire at least one image of the object 7 placed at the starting position. In some embodiments, the image acquisition module 2 is connected to a camera, and the camera is used to generate and provide one or more images to the image acquisition module 2. The camera can be externally connected to the robot object moving system 1. The image acquisition module 2 can be a wired or wireless interface. For example, the image acquisition module 2 can be a wireless communication interface that uses infrared data association (IrDA), Bluetooth, Wi-Fi, cellular network, or any wireless communication protocol that meets the requirements and is applicable. The camera externally connected to the object moving system 1 can transmit the image of the object 7 to the image acquisition module 2 through the above wireless transmission technology. Alternatively, in another embodiment of the present invention, the camera can be a part of the image acquisition module 2 and is disposed inside the robot object moving system 1.

[0016] The image processing module 3 is connected to the image acquisition module 2 and is used to perform image processing on the image of the object 7 to determine the pickable area of the object 7 on its first surface and the first position as the pick-up position, where the first position is the center of the pickable area. The image processing module 3 can utilize existing image processing technologies in the art, such as object detection and edge detection, to obtain information about the object 7, so the image processing technologies will not be elaborated here. In some embodiments, the pickable area is a square area or a circular area whose edges do not exceed the edges of the object 7. The initial pick-up position based on the center of the pickable area takes Figure 3 the pick-up position A shown as an example. The pickable area of the object 7 takes Figure 3 the dashed rectangle shown as an example. In some embodiments, the image processing module 3 can be independent of the image acquisition module 2 or can be integrated with the image acquisition module 2. For example, the image processing module 3 can be a dedicated image signal processor, a part of a graphics processor, or a general-purpose processor. In some other embodiments, the image acquisition module 2 and the image processing module 3 can be integrated into the image unit of the robotic object moving system 1.

[0017] The pick-up module 4 is used to pick up the object 7 and is connected to the sensing module 5 and the control module 6. The pick-up module 4 can execute the pick-up of the object 7 according to the control instructions / signals / commands of the control module 6. In some embodiments, the pick-up module 4 performs a first pick-up attempt to pick up the object 7 to obtain the weight distribution of the object 7. Among them, when the pick-up module 4 performs the first pick-up attempt, it picks up the object 7 at the initial pick-up position on the first surface of the object 7 with a preset first-level pick-up force, and the pick-up distance and time for picking up the object 7 are the preset distance and the first time period respectively. The pick-up module 4 adjusts the pick-up position of the pick-up task according to the sensing data obtained in the first pick-up attempt to move the object 7 from the starting position to the destination position. The control mechanisms of the control module 6 and the pick-up module 4 will be elaborated below.

[0018] In some embodiments, the picking module 4 includes a picking unit having a contact point or area in contact with the object 7, wherein the area of the picking unit in contact with the object 7 is much smaller than the surface area of the object 7. For a robotic object moving system, if the area of the picking unit in contact with the object 7 is equal to or close to the surface area of the object 7, an unbalanced weight distribution of the object 7 will not affect the picking process of the object 7 because the picking force for picking up the object 7 can be evenly applied to most of the surface area of the object 7. However, if the area of the picking unit in contact with the object 7 is much smaller than the surface area of the object 7, when the picking position is far from the center of gravity of the object 7 or the picking position is close to the edge of the object 7, it will cause instability of the object 7 during the picking process. Therefore, the selection of the picking position is quite important for the picking task. In some embodiments, the picking unit in the picking module 4 in contact with the object 7 is a suction cup fixture or an array of suction cup fixtures, and the picking force is applied to the object 7 in a direction perpendicular to the first surface of the object 7. Please refer to Figure 3 , when the image processing module 3 determines the pickable area of the object 7, the size and / or width of the picking unit also need to be considered simultaneously. In some embodiments, the distance between the edge of the object 7 and the pickable area is at least greater than half of the area width of the picking unit. In some other embodiments, the pickable area can be adjusted according to system requirements and user configurations, as long as it ensures that the picking module 4 does not cause incomplete or insufficient contact between the picking unit and the object 7 during the process of picking up the object 7 at the picking position. In some embodiments, the preset distance can be determined according to the user, the size of the object, or the environmental space, and the length of the first time period can be determined according to the preset distance, the preset first-level picking force, the moving speed of the picking module 4, or other considerations.

[0019] The sensing module 5 is connected to the picking module 4 and is used to sense sensing data corresponding to at least two dimensions of the picking module 4. In some embodiments, the sensing module 5 obtains the sensing data during the first time period of the first picking attempt and provides it to the control module 6. The sensing data can reflect the torsion and force borne by the picking module 4 when picking up the object 7 at the picking position A. Specifically, the sensing data can reflect the posture of the robotic object moving system 1 when the object 7 is picked up by the picking module 4. When picking up the object 7, if the sensing data shows that any one of the sensing values has a large difference compared with the no-load initial value corresponding to when the object is not picked up, it means that the object 7 is not picked up at a better position close to the center of gravity. If the sensing data shows that the sensing value deviates within a preset range relative to the initial value, it can be determined that the object 7 is picked up at a relatively safe position not far from the center of gravity of the object 7. The sensing module 5 can be a two-dimensional or more-dimensional sensor, such as the X-axis, Y-axis, and / or Z-axis, and can correspond to two or more types of sensing data, such as torsion and force.

[0020] In some embodiments, the sensing module 5 may be integrated with the picking module 4 or be independent of the picking module 4. For example, the picking module 4 and the sensing module 5 may be integrated into a vacuum gripper robot with a multi-axis torque sensor inside, or may be integrated into an electrostatic gripper with a torque sensor outside. Without departing from the spirit of the invention of this application, the picking module 4 and the sensing module 5 may be implemented by various types of grippers and sensors respectively, as long as they can pick up an object at a specified position and sense sensing data in at least two dimensions respectively.

[0021] The control module 6 is connected to the image acquisition module 2, the image processing module 3, the picking module 4, and the sensing module 5, and is used to control the picking module 4. The control module 6 receives the processed image and the sensing data to determine the initial picking position, the pickable area, the sensing range, and the adjusted picking position according to the criteria of various judgment methods. In addition, the control module 6 determines whether to perform the picking task at the adjusted picking position and with the adjusted picking force, or whether it is safe to continue performing the picking task, or the picking task should be aborted. In some embodiments, the control module 6 determines whether to adjust the picking position according to the sensing data and by using a first pre-trained model with a first criterion. In addition, the control module 6 adjusts the picking position according to the sensing data and the pickable area of the first surface of the object 7, and by using a second pre-trained model with a second criterion, thereby obtaining the adjusted picking position. In some embodiments, the adjusted picking position determined according to the sensing data and the pickable area is regarded as approximate to or close to the center of gravity of the object 7. The picking position adjusted based on the sensing data and the pickable area of the first surface of the object 7 and by using a second pre-trained model with a second criterion is Figure 3 shown by the picking position B as an example. The picking module 4 picks up the object 7 at the adjusted picking position B and moves the object 7 from the starting position to the destination position.

[0022] In the robotic object moving system 1 of the present invention, the picking position of the object 7 is adjusted by the sensing data obtained during the initial picking attempt, thereby improving the stability of picking objects with non-fixed sizes and uneven weights. In addition, compared with the traditional method of using the position at the center of the surface of the object 7 as the picking position, when picking up the object 7 at the adjusted picking position in the present invention, the object 7 can be picked up with a smaller picking force, and the adjusted picking position is closer to the center of gravity of the object 7.

[0023] In some embodiments, the robotic object moving system 1 is not limited to adjusting the picking position when picking up the object 7, but can also adjust the picking force when picking up the object 7. The control module 6 adjusts the picking force to a second level according to the sensing data and by using the second pre-trained model, and the picking module 4 picks up the object 7 at the adjusted picking position with the picking force at the second level.

[0024] In some embodiments, the sensed data includes at least torsional values and force values in two or more dimensions. In some embodiments, the torsional values include six-degrees-of-freedom (6DoF) information. The 6DoF information of the torsional values includes three translational degrees of freedom and three rotational degrees of freedom. The three translational degrees of freedom include moving forward and backward along the X-axis, moving left and right along the Y-axis, and moving up and down along the Z-axis. The three rotational degrees of freedom include rotating about the X-axis, rotating about the Y-axis, and rotating about the Z-axis. The force values include 6DoF information. The 6DoF information of the force values includes three translational degrees of freedom and three rotational degrees of freedom.

[0025] The first criterion and the second criterion are exemplified by the relationship between the torsional values and force values shown by Figure 4A and Figure 4B and the first, second, and third intervals. It should be particularly noted that the number of dimensions of the torsional values and force values in this case is not limited. In the embodiments shown by Figure 4A and Figure 4B , the torsional values include two dimensions, and the force values include one dimension. Among them, the torsional value of rotating along the X-axis is represented by Tx, the torsional value of rotating along the Y-axis is represented by Ty, and the force value of moving along the Z-axis is represented by Fz.

[0026] First, the first criterion includes: (a) when all the torsional values fall within the first interval, the picking position is maintained at the first position; (b) when any of the torsional values falls within the second interval, the picking position is adjusted to the adjusted picking position; and (c) when any of the torsional values and force values falls within the third interval, the picking task of the object 7 is aborted. In the first criterion, the torsional values are divided into three intervals, and the control module 6 determines whether the picking position needs to be adjusted according to the interval in which the torsional values fall. Specifically, if the torsional value falls within the first interval, it means that the torsional value has only a slight deviation. Therefore, the first position can be used as the final picking position without adjustment. If the torsional value falls within the second interval, it means that the torsional value has a relatively large deviation, that is, the object 7 is prone to instability during the picking process. Therefore, the picking position needs to be adjusted from the first position to the second position to provide better stability for the object 7 during the picking process. If the torsional value falls within the third interval, it means that the torsional value has a drastic deviation, that is, the distance between the center of the pickable area and the center of gravity of the object 7 is relatively far. In this case, picking up and moving the object 7 may cause danger or damage to the object. Therefore, it is recommended to abort the picking of the object 7, otherwise the object 7 may fall or be damaged due to the unbalanced weight distribution during the picking process.

[0027] In some embodiments, the first criterion further includes: (d) when the force value falls within the first range, the picking force is maintained at the first level; and (e) when the force value falls within the second range, the picking force is adjusted to the second level. In the first criterion, the force value is divided into three ranges, and the control module 6 determines whether to adjust the picking force according to the range in which the force value falls. Specifically, if the force value falls within the first range, it means that the force value has only a slight deviation. Therefore, the picking force can be used as the final picking force without adjustment. If the force value falls within the second range, it means that the force value has a relatively large deviation. Therefore, the picking force is adjusted to the second level according to the sensing data obtained from the first picking attempt to reduce the deviation of the force value.

[0028] In some embodiments, the maximum absolute value of the first range is less than the maximum absolute value of the second range, and the maximum absolute value of the second range is less than the maximum absolute value of the third range. In other words, the smaller the sensed value of the sensing data, the closer the picking position is to the center of gravity of the object 7, so it will not cause a drastic impact on the posture of the robot object moving system 1. If the sensed value changes drastically, it is recommended to adjust the picking position to make the picking position closer to the center of gravity of the object 7.

[0029] In some embodiments, the first, second, and third ranges are defined by a plurality of critical points. As Figure 4B shown, relative to the initial value, the first range is the region where the offset value is less than the first critical point, the second range is the region where the offset value is between the first critical point and the second critical point, and the third range is the region where the offset value is greater than the second critical point, and the second critical point is greater than the first critical point. The first critical point and the second critical point of the torsion value are determined by at least one of the following: the characteristics of the picking module 4 and the pickable region. The first critical point and the second critical point of the force value can be determined by the maximum force level of the picking module 4.

[0030] As for the second criterion, please refer to Figure 4A , the second criterion includes: (a) determining the offset value of the torsion value relative to the first range; and (b) determining the adjusted picking position in the pickable region according to the offset value. In some embodiments, the second criterion further includes: (c) if the adjusted picking position cannot be obtained in the pickable region, the picking task for the object 7 is aborted. In the second criterion, the offset value of the torsion value and the first range can be determined, and the control module 6 can therefore determine the degree of adjustment of the adjusted picking position according to the offset value.

[0031] In some embodiments, the object 7 is a rectangular box with six faces. The picking module 4 picks up the box through one of the six faces of the box. Specifically, the control module 6 adjusts the posture of the robot object moving system 1 so that the picking module 4 is aligned with one of the six faces of the box and applies force to that face, and thus the picking module 4 can pick up the box in a better way. In some embodiments, the area of contact between the picking module 4 and the box is smaller than the area of the corresponding face on the box.

[0032] Figure 5 The flowchart of the robot object moving method for moving an object with unknown weight distribution in the preferred embodiment of the present invention. The robot object moving method of the present invention is applicable to the aforementioned robot object moving method system 1. As Figure 5 shown, the method of the present invention includes steps S1, S2, S3, S4, S5, S6, and S7. In step S1, at least one image of the object 7 placed at the starting position is acquired. In step S2, image processing is performed on the image to determine the pickable area on the first surface of the object 7 and the first position as the picking position, where the first position is the center of the pickable area. In step S3, the picking module 4 picks up the object 7 at the picking position on the first surface of the object 7 with a preset first-level picking force, and the picking distance and time for picking up the object 7 are a preset distance and a first time period, respectively. In step S4, at least two-dimensional sensing data of the picking module 4 in the first time period is acquired. In step S5, according to the sensing data, it is determined whether the picking position needs to be adjusted by using a first pre-trained model with a first criterion. In step S6, according to the sensing data and the pickable area on the first surface of the object 7, the picking position is adjusted by using a second pre-trained model with a second criterion, and the adjusted picking position is obtained. In step S7, the object 7 is picked up at the adjusted picking position, and the object 7 is moved from the starting position to the destination position.

[0033] In some embodiments, the method of the present invention further includes the steps of: adjusting the picking force to a second level according to the sensing data and using the second pre-trained model; and picking up the object 7 with the second-level picking force at the adjusted picking position.

[0034] In summary, the present invention provides a robot object moving system and method for moving an object with unknown weight distribution. In the robot object moving system and method of the present invention, the picking position of the object is adjusted by using the sensing data obtained during the initial picking attempt, so as to improve the stability of picking objects with unfixed sizes and uneven weights. In addition, compared with the traditional method of using the position at the center of the object surface as the picking position, when picking up the object at the adjusted picking position in the present invention, the object can be picked up with a smaller picking force, and the adjusted picking position is closer to the center of gravity of the object.

[0035] It should be noted that the above are only preferred embodiments proposed for explaining the present invention. The present invention is not limited to the described embodiments, and the scope of the present invention is determined by the scope of the appended claims. And this case can be variously modified by those skilled in this technology, but all are not beyond what is intended to be protected by the appended claims of the present invention.

Claims

1. A method for a robot to move an object with unknown weight distribution, comprising the following steps: Obtain at least one image of an object placed at an initial position; Image processing is performed on the image to determine a pickable area of the object on one of its first surfaces and a first position as a pick-up position, wherein, The first position is the center of a pickable area; Use a pick-up module to pick up the object at a pick-up position on the first surface of the object with a preset first-level pick-up force, wherein the pick-up distance and time for picking up the object are a preset distance and a first time period respectively; Obtain sensing data of at least two dimensions of the pick-up module during the first time period; According to the sensing data, use a first pre-trained model with a first criterion to determine whether to adjust the pick-up position; According to the sensing data and the pickable area on the first surface of the object, use a second pre-trained model with a second criterion to adjust the pick-up position to obtain the adjusted pick-up position; and Pick up the object at the adjusted pick-up position and move the object from the initial position to a target position.

2. The method for a robot to move an object according to claim 1, further comprising the following steps: According to the sensing data, use the second pre-trained model to adjust the pick-up force to a second level; and Pick up the object at the adjusted pick-up position with the pick-up force at the second level.

3. The method for a robot to move an object according to claim 1, wherein, The sensing data at least includes torsion values and force values in two or more dimensions, and the first criterion includes: When all the torsion values fall within a first interval, the pick-up position remains at the first position; When any one of the torsion values falls within a second interval, the pick-up position is adjusted to the adjusted pick-up position; and Wherein, the maximum absolute value of the first interval is less than the maximum absolute value of the second interval, and the maximum absolute value of the second interval is less than the maximum absolute value of a third interval.

4. The method for moving a robotic object according to claim 3, wherein, The second criterion includes: Judge the offset value of the torsion value relative to the first interval; Determine the adjusted pick-up position in the pickable area according to the offset value; and If the adjusted pick-up position cannot be obtained in the pickable area, abort picking up the object.

5. The method for moving a robotic object according to claim 3, wherein, The first criterion further includes: When the force value falls within the first interval, the pick-up force remains at the first level; and When the force value falls within the second interval, the pick-up force is adjusted to a second level, Wherein, the maximum absolute value of the first interval is less than the maximum absolute value of the second interval.

6. A robot object moving system, comprising: An image acquisition module for obtaining at least one image of an object placed at an initial position; An image processing module connected to the image acquisition module and used for performing image processing on the image to determine a pickable area on a first surface of the object and a first position as a pick-up position, wherein the first position is the center of the pickable area; A picking module is configured to pick up the object at a picking position on the first surface of the object with a preset picking force at a first level, where The pick-up distance and time for picking up the object are a preset distance and a first time period respectively; A sensing module connected to the pick-up module and used for sensing sensing data of at least two dimensions of the pick-up module during the first time period; and A control module is connected to the image acquisition module, the image processing module, the picking module, and the sensing module, and is used to control the picking module. The control module receives the processed image and the sensing data, and determines whether it is necessary to adjust the picking position according to the sensing data and by using a first pre-trained model with a first criterion. The control module adjusts the picking position according to the sensing data and the pickable area of the first surface of the object, and by using a second pre-trained model with a second criterion, and obtains the adjusted picking position. The picking module picks up the object according to the adjusted picking position to move the object from the starting position to a target position.

7. The robot object moving system according to claim 6, wherein, The control module adjusts the picking force to a second level according to the sensing data and by using the second pre-trained model. The picking module picks up the object at the adjusted picking position with the picking force at the second level.

8. The robot object moving system according to claim 6, wherein The sensing data at least includes torsion values and force values in two or more dimensions. The first criterion includes: When all the torsion values fall within a first interval, the picking position remains at the first position; and When any one of the torsion values falls within a second interval, the picking position is adjusted to the adjusted picking position; And When any one of the torsion values and the force values falls within a third interval, the picking of the object is aborted; Wherein, the maximum absolute value of the first interval is less than the maximum absolute value of the second interval, and the maximum absolute value of the second interval is less than the maximum absolute value of the third interval.

9. The robot object moving system according to claim 8, wherein, The second criterion includes: Judging an offset value of the torsion value offset relative to the first interval; and Determining the adjusted picking position in the pickable area according to the offset value; and When the adjusted picking position cannot be obtained in the pickable area, the picking of the object is aborted.

10. The robot object moving system according to claim 8, wherein, The first criterion further includes: When the force value falls within the first interval, the picking force remains at the first level; and When the force value falls within the second interval, the picking force is adjusted to a second level. Wherein, the maximum absolute value of the first interval is less than the maximum absolute value of the second interval.

Images