Body robot double-arm floating grabbing force-position hybrid collaborative control method and system

Abstract

The application provides a force-position hybrid collaborative control method and system for dual-arm floating grabbing of a body robot, which is applied to the technical field of industrial automation robots, and through taking the second mechanical arm as a master arm, the movement of a workpiece is driven according to a target movement track to keep the movement track of the workpiece; the first mechanical arm is taken as a slave arm, and force control is performed according to the force deviation between the actual grabbing force and the target grabbing force to make the actual grabbing force track the target grabbing force; in the movement process of the workpiece, the relative pose deviation between the two end effectors is compensated through passive floating of a composite floating mechanism, through the above control division, the movement track of the workpiece is mainly controlled by the second mechanical arm, and the grabbing force of the first end effector is mainly controlled by the first mechanical arm, so that the rigid pulling force caused by the different movements of the dual arms of the body robot is reduced, and the stability and safety of the dual-arm collaborative grabbing and carrying of the body robot are improved.

Description

Technical Field

[0001] This application relates to the field of industrial automation robot technology, specifically to a force-position hybrid collaborative control method and system for a body-worn robot with floating dual arms for grasping. Background Technology

[0002] With the rapid development of Industry 4.0 and intelligent manufacturing, mobile robots have been widely used in warehousing and logistics, electronics manufacturing, and automotive parts assembly to achieve automated material handling, loading and unloading, and assembly operations. For workpieces that are large in size, heavy in weight, or need to maintain a specific orientation, dual-arm robots are typically used to grip the workpiece with two end effectors and move it along the target trajectory.

[0003] Most existing dual-arm robots employ master-slave position control, with the slave arm following the movement of the master arm. However, the two robotic arms differ in manufacturing precision, installation position, coordinate calibration, and control response, which can lead to asynchronous movements during actual operation, resulting in relative position and attitude deviations between the two end effectors.

[0004] When two end effectors are simultaneously gripping the same workpiece, the aforementioned relative positional deviations can easily translate into rigid tensile forces acting on the workpiece and the robotic arm, leading to fluctuations in gripping force, workpiece deformation, or increased load on the robotic arm. Simply improving the motion synchronization of the two robotic arms through position control is insufficient to eliminate the effects of mechanical structural errors, calibration errors, and motion asynchrony, and it is also difficult to maintain a stable gripping force exerted by the end effectors on the workpiece.

[0005] Furthermore, if both robotic arms rigidly position themselves according to the target motion trajectory, significant rigid pulling forces can easily arise due to relative pose deviations between the two end effectors. Reducing these rigid pulling forces by relaxing position constraints may affect the workpiece's motion trajectory and gripping stability. Therefore, existing dual-arm collaborative control methods struggle to effectively control the gripping force exerted on the workpiece by the end effectors while maintaining the workpiece's motion trajectory.

[0006] Therefore, a new collaborative control scheme for the two arms of an embodied robot is needed. Summary of the Invention

[0007] In view of this, the embodiments of this specification provide a force-position hybrid collaborative control method and system for dual-arm floating grasping of a skeletal robot. By combining the passive floating compensation of the composite floating mechanism with the force-position hybrid control of the dual arms of the skeletal robot, the rigid pulling force caused by the relative posture deviation of the dual arms of the skeletal robot is reduced, while maintaining the stability of the workpiece's motion trajectory and grasping force. This achieves unmanned operation of the entire process from cargo docking, loading, transfer, grasping to placement, and greatly improves the system's operating efficiency, stability and safety.

[0008] The embodiments in this specification provide the following technical solutions: This specification provides a force-position hybrid cooperative control method for dual-arm floating grasping of a robotic body, applied to a dual-arm robot. The dual-arm robot includes a first robotic arm and a second robotic arm. The end of the first robotic arm is connected to a first end effector via a composite floating mechanism, and the end of the second robotic arm is fixedly connected to a second end effector. The method includes: Obtain the target motion trajectory corresponding to the workpiece to be grasped; Obtain the target gripping force corresponding to the workpiece; Control the first end effector and the second end effector to jointly grasp the workpiece; The second robotic arm is controlled to move according to the target motion trajectory, so that the second end effector drives the workpiece to move along the target motion trajectory. During the movement of the workpiece, the first end effector floats relative to the end of the first robotic arm along with the composite floating mechanism to compensate for the relative pose deviation between the first end effector and the second end effector. The actual gripping force exerted by the first end effector on the workpiece is obtained, and the force deviation is obtained based on the target gripping force and the actual gripping force. Based on the force deviation, force control commands for the first robotic arm are generated; The first robotic arm is controlled to move according to the force control command, so that the actual grasping force tracks the target grasping force as the second robotic arm moves along the target motion trajectory.

[0009] This specification also provides a force-position hybrid collaborative control system for a dual-arm floating grasping robot, comprising: a mobile chassis, a first robotic arm, a second robotic arm, a composite floating mechanism, a first end effector, a second end effector, and a central collaborative controller; The first robotic arm and the second robotic arm are mounted on the mobile chassis. The end of the first robotic arm is connected to the first end effector via the composite floating mechanism, and the end of the second robotic arm is fixedly connected to the second end effector. The central collaborative controller is communicatively connected to the first robotic arm and the second robotic arm respectively, and is used to acquire the target motion trajectory and target grasping force corresponding to the workpiece to be grasped, and control the first end effector and the second end effector to jointly grasp the workpiece; The central coordinating controller is also used to control the movement of the second robotic arm according to the target motion trajectory, so that the second end effector drives the workpiece to move along the target motion trajectory; The composite floating mechanism is used to make the first end effector float relative to the end of the first robotic arm during the movement of the workpiece, so as to compensate for the relative pose deviation between the first end effector and the second end effector. The central coordinating controller is also used to acquire the actual gripping force of the first end effector acting on the workpiece, obtain the force deviation based on the target gripping force and the actual gripping force, generate force control commands for the first robotic arm based on the force deviation, and control the movement of the first robotic arm based on the force control commands, so that the actual gripping force tracks the target gripping force during the process of the second end effector driving the workpiece to move along the target motion trajectory.

[0010] Compared with the prior art, the technical effects that can be achieved by at least one of the above-mentioned technical solutions adopted in the embodiments of this specification include at least the following: This application controls the second robotic arm to move the workpiece according to the target motion trajectory, controls the first robotic arm to adjust the actual gripping force according to the gripping force deviation, and uses a composite floating mechanism to compensate for the relative posture deviation between the two end effectors. This can reduce the rigid tension caused by the position error, posture error and asynchronous movement of the two arms of the robot, so that while the workpiece moves along the target motion trajectory, the actual gripping force of the first end effector tracks the target gripping force, thereby improving the stability of the robot's two arms in gripping and transporting the workpiece. Attached Figure Description

[0011] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0012] Figure 1 This is a flowchart of a force-position hybrid collaborative control method for dual-arm floating grasping of a holographic robot according to a specific embodiment of this application; Figure 2 This is a diagram of a force-position hybrid cooperative control system architecture for a body-worn robot with dual-arm floating grasping in a specific embodiment of this application; Figure 3 This is a structural diagram of the composite floating mechanism in a specific embodiment of this application. Detailed Implementation

[0013] The embodiments of this application will now be described in detail with reference to the accompanying drawings.

[0014] The following specific examples illustrate the implementation of this application. Those skilled in the art can easily understand other advantages and effects of this application from the content disclosed in this specification. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. This application can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of this application. It should be noted that, in the absence of conflict, the following embodiments and features in the embodiments can be combined with each other. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0015] It should be noted that the following description covers various aspects of embodiments within the scope of the appended claims. It will be apparent that the aspects described herein can be embodied in a wide variety of forms, and any particular structure and / or function described herein is merely illustrative. Based on this application, those skilled in the art will understand that one aspect described herein can be implemented independently of any other aspect, and two or more of these aspects can be combined in various ways. For example, any number and aspects set forth herein can be used to implement the device and / or practice the method. Additionally, this device and / or method can be implemented using structures and / or functionalities other than one or more of the aspects set forth herein.

[0016] It should also be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of this application. The drawings only show the components related to this application and are not drawn according to the actual number, shape and size of the components. In actual implementation, the shape, quantity and proportion of each component can be arbitrarily changed, and the layout of the components may also be more complex.

[0017] Additionally, specific details are provided in the following description to facilitate a thorough understanding of the examples. However, those skilled in the art will understand that practice can be carried out without these specific details.

[0018] Regarding existing collaborative control schemes for dual-arm robots, such as the prior art mentioned in the background section, in the process of analyzing related problems and exploring solutions, it was found that: On the one hand, most existing dual-arm robots use master-slave position control, with the slave arm completely following the movement of the master arm. This makes them unable to cope with the movement error of the master arm and the deformation of the workpiece, and they are very prone to mechanical pulling. On the other hand, if both robotic arms use position control, although the movement trajectory of the workpiece can be constrained, it is difficult to keep the force between the two arms stable, and relaxing the position constraint may affect the movement trajectory of the workpiece.

[0019] In view of this, this specification proposes a force-position hybrid collaborative control scheme for dual-arm floating grasping of a robotic body, adopting a "one master, one slave, position-force separation" control architecture. The overall concept is as follows: the second robotic arm 3 serves as the master arm, with a second end effector fixedly connected to its end. It drives the workpiece to move according to the target motion trajectory, maintaining the workpiece's trajectory. The first robotic arm 7 serves as the slave arm, with its end connected to the first end effector via a composite floating mechanism. It controls the force based on the force deviation between the actual grasping force and the target grasping force, ensuring the actual grasping force tracks the target grasping force. During workpiece movement, the composite floating mechanism passively compensates for the relative pose deviation between the two end effectors through floating. Through this division of control, the workpiece's motion trajectory is primarily controlled by the second robotic arm 3, and the grasping force of the first end effector is primarily controlled by the first robotic arm 7. This reduces the rigid pulling force caused by asynchronous movement of the robotic body's two arms, improving the stability and safety of the robotic body's dual-arm collaborative grasping and handling of workpieces.

[0020] The technical solutions provided by the various embodiments of this application are described below with reference to the accompanying drawings.

[0021] like Figure 1 and Figure 2 As shown in the embodiment of this specification, a force-position hybrid cooperative control method for dual-arm floating grasping of a robotic body is provided, applied to a dual-arm robot. The dual-arm robot includes a first robotic arm 7 and a second robotic arm 3. The end of the first robotic arm 7 is connected to a first end effector via a composite floating mechanism, and the end of the second robotic arm is fixedly connected to a second end effector. The composite floating mechanism is the structure in the floating end tool 6 used to achieve translational floating and rotational floating. The method includes: Step S10: Obtain the target motion trajectory and target gripping force corresponding to the workpiece.

[0022] Specifically, the central collaborative controller acquires the initial position, target position, and handling requirements of the workpiece based on the task, obtains the target motion trajectory of the workpiece from the initial position to the target position, and acquires the target gripping force corresponding to the workpiece. The target motion trajectory serves as the motion control reference for the second robotic arm 3, and the target gripping force serves as the gripping force control reference for the first robotic arm 7. In this embodiment, the target motion trajectory and target gripping force can be preset or determined by the central collaborative controller based on the acquired workpiece information.

[0023] Step S20: Control the first end effector and the second end effector to jointly grasp the workpiece.

[0024] Specifically, the central coordinating controller controls the first robotic arm 7 and the second robotic arm 3 to drive their respective end effectors closer to the workpiece, so that the floating end tool 6 and the fixed end tool 4 act on the corresponding gripping positions of the workpiece. After the two end effectors complete gripping, they jointly constrain and support the workpiece, providing a gripping basis for subsequent workpiece movement.

[0025] Step S30: Control the second robotic arm to drive the workpiece to move according to the target motion trajectory.

[0026] Specifically, the central collaborative controller controls the movement of the second robotic arm 3 according to the target motion trajectory. Since the fixed end-effector 4 is fixedly connected to the end of the second robotic arm 3, the movement of the second robotic arm 3 is transmitted to the workpiece through the fixed end-effector 4, thereby driving the workpiece to move along the target motion trajectory. In the collaborative operation of the two arms of the embodied robot, the second robotic arm 3 acts as the main controller of the workpiece's motion trajectory, used to maintain the workpiece's motion direction and process.

[0027] Step S40: Compensate for the relative pose deviation between the first end effector and the second end effector through a composite floating mechanism.

[0028] Specifically, during the movement of the workpiece driven by the second robotic arm 3, the first robotic arm 7 and the second robotic arm 3 may experience relative positional deviations due to manufacturing precision, installation position, coordinate calibration, control response, and motion synchronization errors. Both end effectors grip the same workpiece, and the relative positional deviation transmits external forces to the floating end tool 6 through the workpiece, causing the first end effector to passively float relative to the end of the first robotic arm 7 along with the composite floating mechanism. The composite floating mechanism absorbs the relative positional deviation between the two end effectors through this passive floating, reducing the rigid tensile force directly generated by the relative positional deviation.

[0029] Step S50: Control the movement of the first robotic arm according to the force deviation between the actual gripping force and the target gripping force, so that the actual gripping force tracks the target gripping force.

[0030] Specifically, during the workpiece movement, the actual gripping force exerted on the workpiece by the floating end tool 6 is acquired. In this embodiment, the corresponding force information can be acquired by a force sensor located between the end of the first robotic arm 7 and the floating end tool 6, and the actual gripping force is determined based on the force information. The central coordinating controller compares the actual gripping force with the target gripping force to obtain the force deviation, generates a force control command for the first robotic arm 7 based on the force deviation, and then controls the movement of the first robotic arm 7 according to the force control command. As the workpiece continues to move along the target motion trajectory, the central coordinating controller continuously acquires the actual gripping force and controls the first robotic arm 7 to make the actual gripping force track the target gripping force.

[0031] In this embodiment, the second robotic arm 3 drives the workpiece to move according to the target motion trajectory, the first robotic arm 7 performs force control according to the gripping force deviation, and the composite floating mechanism is used to compensate for the relative posture deviation between the two end effectors. This significantly reduces the rigid pulling force caused by the position error, posture error and asynchronous movement of the two arms of the robot. As the workpiece moves along the target motion trajectory, the actual gripping force of the first end effector tracks the target gripping force, thereby improving the stability of the robot's two arms gripping and transporting the workpiece together.

[0032] In some embodiments, the composite floating mechanism provides the first end effector with translational floating degree of freedom and rotational floating degree of freedom; the translational floating degree of freedom includes vertical floating degree of freedom and in-plane floating degree of freedom, and the rotational floating degree of freedom includes multi-angle swinging degree of freedom.

[0033] Specifically, such as Figure 3 As shown, the first end effector can float vertically along the floating guide 609 (such as a floating guide rod), move in a plane along the horizontal floating member 607 (such as a horizontal floating disk), and also swing at multiple angles through the cooperation of the floating ball head 611 and the circular ball head pressure seat 610. The vertical floating and in-plane movement correspond to the translation, and the multi-angle swing corresponds to the rotation.

[0034] During the workpiece movement, the fixed end tool 4 at the end of the second robotic arm 3 drives the workpiece to move along the target motion trajectory. The floating end tool 6 at the end of the first robotic arm 7 and the fixed end tool 4 jointly grasp the workpiece. When there is a position error, posture error or asynchronous movement between the first robotic arm 7 and the second robotic arm 3, the workpiece transmits external force to the floating end tool 6. Under the external force, the first end actuator translates and rotates relative to the end of the first robotic arm 7 through the composite floating mechanism.

[0035] Specifically, when there is a vertical positional deviation between the two end effectors, the first end effector floats vertically; when there is a planar positional deviation between the two end effectors, the first end effector moves in the planar direction; when there is an attitude deviation between the two end effectors, the first end effector performs a corresponding angular oscillation. The first end effector can also simultaneously generate vertical floating, planar movement and angular oscillation according to the external forces it receives.

[0036] Based on the translation and the rotation, the pose of the first end effector relative to the second end effector is adjusted to reduce the relative pose deviation.

[0037] Specifically, the vertical floating and planar movement are used to adjust the position of the first end effector relative to the second end effector, and the angular swing is used to adjust the posture of the first end effector relative to the second end effector, so that the first end effector can adapt to the relative posture changes generated when the fixed end tool 4 drives the workpiece to move, thereby reducing the relative posture deviation between the two end effectors.

[0038] This embodiment enables the first end effector to float vertically, move in a plane, and swing at multiple angles relative to the end of the first robotic arm. This allows it to tolerate a maximum positional error of ±10mm and an angular error of ±30° between the two robots, avoiding mechanical pulling and significantly improving the adaptability and stability of the robotic arm's dual-arm grasping and handling of workpieces.

[0039] In some implementations, generating force control commands for the first robotic arm based on the force deviation includes: Based on the actual gripping force, information on the change in actual gripping force is obtained; Based on the target grasping force, information on the change in target grasping force is obtained; Based on the actual gripping force change information and the target gripping force change information, the force change deviation is obtained; Based on the force deviation and the force change deviation, force control commands for the first robotic arm are generated; The actual gripping force change information is used to characterize the change of the actual gripping force over time, and the target gripping force change information is used to characterize the change of the target gripping force over time.

[0040] Specifically, during the process of the first and second end effectors jointly gripping the workpiece and driving its movement, the actual gripping force exerted by the first end effector on the workpiece is acquired according to the control cycle. The actual gripping force change information is obtained based on the actual gripping force within the continuous control cycle. The actual gripping force can be determined based on the force information collected by the six-dimensional force sensor installed at the end of the first robotic arm. The actual gripping force change information uses the actual gripping force change rate, and the target gripping force change information uses the target gripping force change rate.

[0041] In practice, the control objective of the first robotic arm, i.e. the slave arm (force control), is to keep its gripping force on the workpiece at a set value, regardless of how the main arm moves or how the workpiece shakes.

[0042] The control logic is: The force that the arm should output at the moment = target grasping force + force deviation correction + force change rate correction.

[0043] In one implementation, the slave arm employs a proportional-derivative (PD) control algorithm to generate force control commands based on force deviation and force change deviation. The slave arm force control law is expressed as follows:

[0044] in, This indicates the output force command from the arm at time t; This represents the force applied to the target by the arm at time t; Indicates force proportional gain (adjusted according to workpiece stiffness); Indicates the differential gain of the force (adjusted according to the response speed requirements); This represents the rate of change of the force from the target arm at time t; This represents the rate of change of the actual force on the arm at time t.

[0045] The PD control algorithm generates force control commands for the first robotic arm, and the first robotic arm is controlled to move according to the force control commands. The force deviation correction is used to reduce the current deviation between the actual gripping force and the target gripping force, and the force change rate correction is used to correct the control output of the first robotic arm according to the changing trend of the actual gripping force, so that the actual gripping force tracks the target gripping force as the second robotic arm moves the workpiece.

[0046] This embodiment generates force control commands for the first robotic arm by simultaneously utilizing force deviation and force change deviation. It can adjust the movement of the first robotic arm according to the current deviation and changing trend of the actual grasping force relative to the target grasping force, suppress rapid changes in grasping force and grasping force fluctuations during the control process, improve the responsiveness of the force control of the first robotic arm and the stability of the robotic arm when both arms work together to transport workpieces.

[0047] In some implementations, generating force control commands for the first robotic arm based on the force deviation and the force change deviation includes: The direction of gripping force adjustment is determined based on the relationship between the actual gripping force and the target gripping force. The gripping force adjustment amount is determined based on the force deviation and the force change deviation; The force control command for the first robotic arm is generated based on the gripping force adjustment direction and the gripping force adjustment amount; When the actual gripping force is greater than the target gripping force, the gripping force adjustment direction is to decrease the actual gripping force; when the actual gripping force is less than the target gripping force, the gripping force adjustment direction is to increase the actual gripping force.

[0048] Specifically, during the process of the second robotic arm driving the workpiece to move along the target motion trajectory, the central coordinating controller continuously acquires the actual gripping force of the first end effector acting on the workpiece, compares the actual gripping force with the target gripping force, determines the gripping force adjustment direction based on the comparison result, determines the gripping force adjustment amount based on the force deviation and force change deviation, and then generates force control commands for the first robotic arm based on the gripping force adjustment direction and gripping force adjustment amount.

[0049] When the actual gripping force is greater than the target gripping force, the first robotic arm is controlled to move in the direction that reduces the actual gripping force; when the actual gripping force is less than the target gripping force, the first robotic arm is controlled to move in the direction that increases the actual gripping force.

[0050] It should be noted that the gripping force adjustment amount is determined jointly based on the force deviation and the force change deviation. Specifically, the force deviation correction amount is determined based on the force deviation, the force change rate correction amount is determined based on the force change deviation, and the gripping force adjustment amount is determined based on the force deviation correction amount and the force change rate correction amount.

[0051] The force deviation correction is used to correct the current deviation between the actual gripping force and the target gripping force. For example, if the current gripping force of the arm is 5N less than the set value, an inward force is applied to make it grip tighter; if the current gripping force of the arm is 5N greater than the set value, an outward force is applied to make it looser.

[0052] The force change rate correction is used to adjust the gripping force based on the actual trend of the gripping force. For example, when the gripping force is rapidly increasing (indicating that a pull is about to occur), an outward force is applied in advance to reduce the force; when the gripping force is rapidly decreasing, an inward force is applied in advance to increase the force.

[0053] This embodiment determines the direction of gripping force adjustment based on the relationship between the actual gripping force and the target gripping force, and determines the amount of gripping force adjustment based on force deviation and force change deviation. It can unload when the actual gripping force is too large and load when the actual gripping force is too small. At the same time, it corrects the adjustment amount according to the trend of gripping force change, so that the actual gripping force tracks the target gripping force, thereby ensuring that the gripping force of the slave arm is constant and the reaction speed can reach the millisecond level. It can cancel out the pulling force the instant it is generated, thereby suppressing gripping force overshoot and fluctuation, and improving the responsiveness and stability of the force control of the first robotic arm.

[0054] In some embodiments, controlling the movement of the second robotic arm according to the target motion trajectory includes: Based on the target motion trajectory, determine the target motion state of the second end effector; Obtain the actual motion state of the second end effector; The motion state deviation is obtained based on the target motion state and the actual motion state; Based on the motion state deviation, position control commands for the second robotic arm are generated; The second robotic arm is controlled to move according to the position control command.

[0055] Specifically, the second robotic arm acts as the position control side. The central coordinating controller determines the target motion state of the second end effector within the current control cycle based on the target motion trajectory of the workpiece. The target motion state characterizes the motion state that the second end effector should achieve when moving along the target motion trajectory. The central coordinating controller obtains the actual motion information fed back by the second robotic arm and determines the actual motion state of the second end effector based on the actual motion information. Subsequently, the target motion state is compared with the actual motion state to obtain the motion state deviation of the second end effector relative to the target motion trajectory. Based on the motion state deviation, a position control command for the second robotic arm is generated. The central coordinating controller sends the position control command to the second robotic arm to control the movement of the second robotic arm, so that the second end effector drives the workpiece to move along the target motion trajectory.

[0056] During subsequent control cycles, the central coordinating controller continuously acquires the actual motion state and updates the motion state deviation. Based on the updated motion state deviation, it generates corresponding position control commands to form a closed-loop position control for the second robotic arm.

[0057] This embodiment compares the target motion state of the second end effector with the actual motion state, and performs closed-loop position control on the second robotic arm based on the obtained motion state deviation. This enables timely correction of the motion deviation of the second end effector relative to the target motion trajectory, thereby improving the tracking accuracy and stability of the workpiece motion trajectory.

[0058] In some implementations, the target motion state includes the target position and the target velocity, and the actual motion state includes the actual position and the actual velocity; The position deviation is obtained based on the target position and the actual position, and the speed deviation is obtained based on the target speed and the actual speed; Based on the position deviation and the speed deviation, position control commands for the second robotic arm are generated.

[0059] In practice, the second robotic arm acts as the main arm, using position control to drive the workpiece along the target motion trajectory. The central coordinating controller determines the target position and target speed of the second end effector within the current control cycle based on the target motion trajectory, and obtains the actual position and actual speed fed back by the second end effector. The target position is then compared with the actual position to obtain the position deviation, and the target speed is compared with the actual speed to obtain the speed deviation.

[0060] The goal of controlling the main arm is to make its actual position follow the target position as closely as possible, while ensuring smooth movement without any shaking.

[0061] The control logic of the main arm is: the current position the main arm should go to = target position + position deviation correction + speed deviation correction; The position deviation correction amount is determined based on the position deviation between the target position and the actual position. The larger the position deviation, the greater the correction amount for the position control output. For example, if the main boom is 10mm away from the target position, it will be given a forward force to make it catch up faster; the farther away it is, the greater the force. The speed deviation correction amount is determined based on the speed deviation between the target speed and the actual speed. It is used to suppress speed overshoot or speed lag during the movement of the second robotic arm. For example, if the main arm is moving too fast and exceeds the target speed, a backward force is applied to slow it down; if it is moving too slow, a forward force is applied to accelerate it.

[0062] This implementation uses the PD control algorithm to generate position control commands for the second robotic arm based on position and speed deviations. The position control relationship is expressed as follows: ; in, This indicates the output position command of the main arm at time t; Indicates the target position of the main arm at time t; Indicates the position proportional gain (adjusted according to the workpiece weight); Indicates the position differential gain (adjusted according to the operating speed); This represents the target velocity of the main arm at time t; This represents the actual speed of the main arm at time t.

[0063] The position control command for the second robotic arm is generated based on the target position, position deviation correction amount, and speed deviation correction amount, and the movement of the second robotic arm is controlled according to the position control command.

[0064] This embodiment uses position deviation to correct the motion position of the second end effector and speed deviation to correct the motion speed of the second end effector. This enables the actual position of the second end effector to track the target position, while suppressing speed fluctuations during the motion process, thereby improving the accuracy and smoothness of the second robotic arm's trajectory tracking.

[0065] In some embodiments, the method further includes: The center of gravity position of the workpiece, and the gripping positions of the first end effector and the second end effector on the workpiece are obtained; Based on the center of gravity position and the grasping position, determine the target load distribution ratio between the first robotic arm and the second robotic arm; Obtain the actual loads of the first robotic arm and the second robotic arm; Based on the actual load, the actual load distribution ratio between the first robotic arm and the second robotic arm is obtained; Based on the deviation between the actual load distribution ratio and the target load distribution ratio, adjust the control output of at least one robotic arm so that the actual load distribution ratio approaches the target load distribution ratio.

[0066] The system pre-establishes a database of different types of workpieces, which includes the workpiece's weight, center of gravity, material, and stiffness parameters. Before gripping the workpiece, the central collaborative controller calls the corresponding workpiece parameters based on the visual recognition results to obtain the workpiece's center of gravity position. Combined with the gripping positions of the first and second end effectors on the workpiece, it determines the target load distribution ratio between the first and second robotic arms.

[0067] In this embodiment, the second robotic arm is the master arm, and the first robotic arm is the slave arm. The central collaborative controller calculates the target load distribution ratio based on the workpiece's center of gravity position and the lever balance principle. Main arm load ratio = distance from the arm gripping position to the center of gravity / total distance between the two gripping positions × 100%; Arm load ratio = distance from the main arm gripping position to the center of gravity / total distance between the two gripping positions × 100%.

[0068] The main arm bears the target load of the second robotic arm, and the slave arm bears the target load of the first robotic arm. The closer the workpiece's center of gravity is to the gripping position of one robotic arm, the larger the target load distribution ratio borne by that robotic arm, in order to maintain torque balance when the robot's two arms grip the workpiece together.

[0069] During the gripping and handling of workpieces, the central collaborative controller continuously acquires the actual load of the first and second robotic arms and calculates the corresponding actual load distribution ratio based on the actual load of the two robotic arms. Real-time feedback from a six-dimensional force sensor can be used during the acquisition and adjustment of the actual load. The central collaborative controller compares the actual load distribution ratio with the target load distribution ratio to obtain the load ratio deviation, and adjusts the control output of at least one robotic arm based on the load ratio deviation. When the actual load distribution ratio of one robotic arm is greater than the corresponding target load distribution ratio, the load borne by that robotic arm is reduced, or the load borne by the other robotic arm is increased; when the actual load distribution ratio of one robotic arm is less than the corresponding target load distribution ratio, the load borne by that robotic arm is increased, or the load borne by the other robotic arm is decreased, so that the actual load distribution ratio of the two robotic arms approaches the corresponding target load distribution ratio.

[0070] This embodiment determines the target load distribution ratio based on the workpiece's center of gravity and the gripping positions of the two end effectors, and makes feedback adjustments based on the actual load during the gripping process. This allows the load borne by the two robotic arms to be adapted to the workpiece's center of gravity, reducing the situation where the load on one side of the robotic arm is too large, and improving the load balance and operational stability when the robotic arm grips and transports the workpiece together.

[0071] In some implementations, obtaining the target motion trajectory corresponding to the workpiece to be grasped includes: obtaining visual information of the workpiece, the visual information including image data and spatial data of the workpiece; determining the pose of the workpiece based on the image data and the spatial data; and determining the grasping position and target motion trajectory of the workpiece based on the pose of the workpiece.

[0072] Specifically, such as Figure 2 As shown, the visual positioning unit can be a working camera 5 located at the end of the second robotic arm 3. The working camera 5 is a 3D industrial camera equipped with a light source, used for photographic recognition and 3D positioning of the workpiece. Before gripping the workpiece, the second robotic arm 3 moves the working camera 5 above the tray to scan the tray and the workpiece on it. The working camera 5 acquires two-dimensional images and three-dimensional point cloud data of the workpiece, wherein the two-dimensional image is used as the image data and the three-dimensional point cloud data is used as the spatial data.

[0073] The central collaborative controller processes the 2D image, using deep learning algorithms to identify the type and approximate location of the workpiece; simultaneously, it processes the 3D point cloud data to reconstruct the 3D model of the workpiece, and calculates its precise position and orientation based on the approximate location, thus obtaining the workpiece's pose. The central collaborative controller generates a grasping task queue according to the identified workpiece type, position, and orientation, and determines the grasping positions of the first and second end effectors on the workpiece based on its pose. The central collaborative controller generates a target motion trajectory based on the grasping positions and the workpiece's pose, enabling the second robotic arm 3 to move to the grasping position along the target trajectory, and then move the workpiece after both arms of the robotic arm grasp it. In this embodiment, the recognition accuracy of the working camera 5 can be ±0.05 mm.

[0074] This embodiment identifies the workpiece type and determines its three-dimensional pose by combining image data and spatial data of the workpiece. This provides more accurate spatial information for determining the grasping position and target motion trajectory, thereby improving the accuracy of the robotic arm in locating, grasping, and planning the trajectory of the workpiece.

[0075] In other implementations, after determining the workpiece's pose based on visual information, the central coordinating controller acquires force feedback information during the grasping process and corrects the workpiece's pose estimation in real time based on the force feedback information. Visual information is used for global workpiece positioning, while force feedback information is used to correct the workpiece's pose estimation after the end effector contacts the workpiece, thus forming a vision-force fusion positioning and grasping process.

[0076] This embodiment uses force feedback information during the grasping process to correct the workpiece pose obtained by visual positioning, which can reduce the impact of visual positioning error on the grasping position determination result and further improve the accuracy of workpiece positioning and the grasping of the robot's dual arms.

[0077] In some implementations, motion state information of the first robotic arm and the second robotic arm is obtained; Based on the motion state information, determine the minimum distance between the first robotic arm and the second robotic arm; When the minimum distance is less than the first safety distance and not less than the second safety distance, the movement speed of at least one robotic arm is reduced, where the first safety distance is greater than the second safety distance. When the minimum distance is less than the second safe distance, the first robotic arm and the second robotic arm are controlled to stop moving and an alarm message is output. Obtain collision detection information between the first robotic arm and the second robotic arm; When the collision detection information of either robotic arm indicates that a collision has occurred, control the first robotic arm and the second robotic arm to stop moving; When the collision force on the robotic arm exceeds a preset force, the power output of the dual-arm robot is cut off.

[0078] Specifically, the central collaborative controller continuously acquires motion status information of the first and second robotic arms during the collaborative operation of the dual-arm robot, and calculates the minimum distance between the two robotic arms in real time based on their motion trajectories. Based on the minimum distance and the comparison between the first and second safe distances, graded safety protection is implemented for the dual-arm robot.

[0079] In the first-level protection phase, the first safety distance is a safety threshold, for example, set to 100 mm, and the second safety distance is a warning threshold, for example, set to 50 mm. When the minimum distance between the two robotic arms is less than 100 mm but not less than 50 mm, it indicates that the two robotic arms are tending to approach each other. The central collaborative controller reduces the movement speed of at least one robotic arm in advance. For example, the movement speed of the two robotic arms is reduced to 50% of their current movement speed to decrease the speed at which the two robotic arms continue to approach each other.

[0080] During the secondary protection phase, when the minimum distance between the two robotic arms is less than 50 mm, the central collaborative controller immediately stops the movement of both robotic arms and outputs an alarm message. For example, an audible and visual alarm is issued to alert the dual-arm robot to enter a warning state.

[0081] In the Level 3 protection phase, the central collaborative controller acquires collision detection information from the first and second robotic arms. In this embodiment, collision forces can be detected using robot joint torque sensors and end effector six-dimensional force sensors. When an abnormal change occurs in the joint torque or end effector force of either robotic arm, and the corresponding collision detection information indicates a collision, the central collaborative controller immediately stops the first and second robotic arms and cuts off the servo power to prevent both robotic arms from continuing their current movement.

[0082] In the Level 4 protection phase, the central collaborative controller compares the detected collision force with a preset force. This preset force can be a safety threshold, such as 50 N. When the collision force on the robotic arm exceeds 50 N, the central collaborative controller immediately cuts off the power supply to the system containing the dual-arm robot and triggers an emergency stop, putting the dual-arm robot into an emergency stop state.

[0083] This embodiment achieves comprehensive safety protection from prediction to emergency stop through a four-level hierarchical safety protection mechanism, avoiding secondary accidents caused by the failure of a single robot. All safety functions comply with the ISO 10218-1 and ISO / TS 15066 collaborative robot safety standards, forming a multi-level safety protection from collision prediction, proximity warning, collision response to emergency stop, which significantly improves the safety of the robot's dual-arm collaborative operation process.

[0084] Below is an example of a master-slave force-position hybrid control system for collaborative operation of embodied robot dual arms, combining the aforementioned implementation methods such as collaborative control of embodied robot dual arms, vision-force fusion positioning, dynamic load distribution, and hierarchical safety protection. Employing a collaborative operation algorithm framework that combines passive floating decoupling with active force-position hybrid control, the rigid fixed gripper is responsible for "leading the way" (ensuring positional accuracy), while the floating gripper is responsible for "borrowing force" (ensuring gentle force). With a clear division of labor, and further enhanced by the mechanical buffering of the floating structure, the pulling force between the two arms of the embodied robot is fundamentally eliminated.

[0085] The main arm is only responsible for position control. It is the "navigator" of the entire collaborative operation, moving strictly according to the planned trajectory to ensure that the workpiece can be accurately delivered to the target position. The slave arm is only responsible for force control. It is the "buffer" of the entire collaborative operation. It does not strictly follow the position trajectory, but only ensures that its gripping force on the workpiece is always kept at the set value. The two do not interfere with each other. Even if there is a slight asynchrony between the movements of the two robots, it will only manifest as a slight change in the gripping force of the slave arm, and will not cause rigid pulling.

[0086] During the coordinated movement of the two robotic arms, there may be repetitive positioning errors of the robot body, visual positioning errors, asynchronous motion control errors, and workpiece deformation errors between the two robotic arms. For example, the repetitive positioning error of the robot body is ±0.02mm, the visual positioning error is ±0.05mm, the asynchronous motion control error between the two robots is ±10ms, and the deformation error of the workpiece itself may cause relative positional and attitude deviations between the first and second end effectors, and further generate rigid tensile forces acting on the workpiece and the two robotic arms.

[0087] This implementation method decouples these errors through a dual mechanism combining mechanical floating and active power control. Specifically: The first decoupling is achieved by a composite floating end effector: In one test example, the composite floating end effector can absorb 80%-90% of the position and angle errors; when the position difference between the master arm and the slave arm is within ±10mm or the angle difference is within ±30°, the composite floating mechanism adapts to the relative posture deviation between the two robotic arms by passive floating, reducing the conversion of the relative posture deviation into rigid tensile force.

[0088] The second decoupling is achieved by the active gripping force control of the slave arm: for the remaining 10%-20% of minute errors that the composite floating mechanism cannot absorb, the force control algorithm of the slave arm will immediately adjust the output force to pull the gripping force back to the set value.

[0089] These two mechanisms work together like a "super shock absorber" for the slave arm. No matter how the main arm moves, the slave arm can gently grip the workpiece like a hand, without any pulling.

[0090] A six-dimensional force sensor is mounted at the end of the slave arm (with a floating gripper). The sensor has a sampling frequency of 1000 Hz and is used to detect the actual gripping force acting on the workpiece at the end of the slave arm. When there is no tensile force between the master and slave arms, the force measured by the sensor is equal to the set target gripping force. For example, if the target gripping force is set to 50 N, the six-dimensional force sensor reading will be 50 N. When the actual gripping force detected by the six-dimensional force sensor deviates from 50 N, for example, when the actual gripping force increases to 55 N or decreases to 45 N, this force deviation can be used to characterize the presence of forces between the two robotic arms that have not been fully absorbed by the composite floating mechanism. The larger the absolute value of the force deviation, the greater the force that needs to be adjusted by the slave arm.

[0091] The central coordinating controller can preset the force tolerance deviation threshold, for example, setting the force tolerance deviation threshold to ±5 N. When the difference between the force measured by the sensor and the target force exceeds 5 N, the system immediately determines that there is a "residual small error" and calculates the specific force value that needs to be compensated (for example, if the current force is 56 N, then 6 N of force needs to be actively released).

[0092] During force control, the main arm performs position control according to the target motion trajectory, while the slave arm does not use the absolute position of its end effector as the primary control target, but adjusts its movement according to force control commands. When the actual gripping force increases, the central coordinating controller controls the slave arm to move in the direction that reduces the actual gripping force to release the increased force; when the actual gripping force decreases, the central coordinating controller controls the slave arm to move in the direction that increases the actual gripping force to compensate for the insufficient force.

[0093] The six-dimensional force sensor can read the actual gripping force every 1 ms. In addition to determining the adjustment amount based on the force deviation, the central coordinating controller also determines the force change rate based on the change in actual gripping force over time. When the actual gripping force increases rapidly, the central coordinating controller generates a correction amount to reduce the actual gripping force based on the force change rate, and performs unloading adjustment before the actual gripping force increases further; when the actual gripping force decreases rapidly, the central coordinating controller generates a correction amount to increase the actual gripping force based on the force change rate, thereby improving the response speed of the arm force control.

[0094] The main boom position control logic flow may include: obtaining the target position X_d and target velocity V_d, reading the current actual position X_b and actual velocity V_b of the main boom, and calculating the position deviation based on the target position and actual position: ΔX = X_d - X_b; Calculate the speed deviation based on the target speed and the actual speed: ΔV = V_d - V_b; Then, the position control output of the main boom is calculated based on the target position, position deviation, and velocity deviation: X_b_out = X_d + K_p*ΔX + K_d*ΔV; In the formula, X_b_out represents the position control output of the main arm, K_p represents the proportional coefficient corresponding to the position deviation, and K_d represents the differential coefficient corresponding to the speed deviation. The central coordination controller sends the position control command generated based on the position control output to the main arm servo driver to control the main arm to move along the target trajectory.

[0095] In another implementation, the slave arm corrects the force control output based on the rate of change of the actual gripping force. The slave arm control process may include: acquiring the target gripping force F_d, reading the current actual force F_a from the six-dimensional force sensor, and calculating the force deviation: ΔF = F_d - F_a; Based on the actual gripping force F_a in the current control cycle, the actual gripping force F_a_prev in the previous control cycle, and the control cycle Δt, the force change rate is calculated as: dF / dt = (F_a - F_a_prev) / Δt; Then, the force control output from the arm is calculated based on the target grasping force, force deviation, and force change rate: F_a_out = F_d + K_fp*ΔF + K_fd*dF / dt; In the formula, F_a_out represents the force control output of the slave arm, K_fp represents the proportional coefficient corresponding to the force deviation, and K_fd represents the differential coefficient corresponding to the force change rate. The central coordinating controller sends the force control command generated based on the force control output to the slave arm servo driver to adjust the movement of the slave arm. The master arm position control process and the slave arm force control process are executed cyclically in each control cycle until the dual-arm collaborative operation task of the embodied robot is completed.

[0096] During the workpiece positioning stage, a high-precision positioning and grasping algorithm that integrates vision and force is adopted.

[0097] In the coarse positioning stage, an industrial camera acquires a two-dimensional image of the workpiece, and a deep learning algorithm is used to identify the type and approximate location of the workpiece, obtain the three-dimensional point cloud data of the workpiece, reconstruct the three-dimensional model of the workpiece, and calculate the precise pose of the workpiece with a positioning accuracy of ±0.05mm. During the force-guided phase, when the gripper is 5mm away from the workpiece, it switches to force control mode and uses a six-dimensional force sensor to detect the contact force, achieving compliant contact and precise gripping. During the grasping process, based on the force feedback information, a pose correction algorithm is used to correct the workpiece's pose estimation in real time, further improving the grasping accuracy.

[0098] During the load distribution phase, the system pre-establishes a database of different types of workpieces, including parameters such as workpiece weight, center of gravity, material, and stiffness. Before gripping, the system calls the corresponding workpiece parameters based on the visual recognition results and automatically calculates the optimal load distribution ratio between the two robots. During the gripping process, the system dynamically adjusts the load distribution ratio based on real-time feedback from the six-dimensional force sensor to ensure balanced load between the two robots. When the workpiece weight exceeds the load capacity of a single robot, the collaborative gripping mode is automatically triggered, with a maximum collaborative gripping weight of 50kg.

[0099] After visually identifying the workpiece's center of gravity coordinates, the baseline load distribution ratio between the two arms is determined according to the lever balance principle. The main arm's load-bearing ratio is the ratio of the distance from its gripping position to the workpiece's center of gravity to the total distance between its two ends, where the total distance is the distance between the main arm's gripping position and the secondary arm's gripping position. The secondary arm's load-bearing ratio is the ratio of the distance from its gripping position to the workpiece's center of gravity to the total distance between its two ends. The closer the workpiece's center of gravity is to the gripping position of one side of the robotic arm, the greater the load-bearing ratio of that side's robotic arm, in order to maintain torque balance.

[0100] During the handling process, a six-dimensional force sensor monitors the actual force on both arms at a period of 50 ms and determines the actual load distribution ratio based on the actual force on both arms. When the actual load distribution ratio deviates from the target load distribution ratio by more than ±3%, the system gradually corrects the control output of both arms in 1% increments, reducing the control output of the overloaded robotic arm and increasing the control output of the underloaded robotic arm, until the actual load distribution ratio converges to within ±3% of the target load distribution ratio.

[0101] During the security protection phase, a tiered security protection algorithm is executed.

[0102] Level 1 protection is a predictive protection system. Based on the movement trajectory of the robot's two arms, it calculates the minimum distance between them in real time. When the distance is less than the safety threshold (100mm) but not less than the warning threshold (50mm), it decelerates to 50% in advance. Level 2 protection is the early warning level protection. When the distance is less than the early warning threshold (50mm), the movement will stop immediately and an audible and visual alarm will be issued. Level 3 protection is collision-level protection. When a minor collision occurs, the collision force is detected by the robot's joint torque sensor and the end effector six-dimensional force sensor, and the servo power is immediately cut off. Level 4 protection is an emergency protection level. When the collision force exceeds the safety threshold (50N), the power supply of the entire system is immediately cut off, triggering an emergency stop.

[0103] When either robotic arm triggers its safety protection, the other robot will also stop moving simultaneously to avoid causing secondary injury.

[0104] This example utilizes a composite floating mechanism to absorb some of the relative pose deviations between the two arms of the embodied robot. It actively adjusts the movement of the slave arm based on the gripping force feedback, while the main arm moves the workpiece according to the target trajectory. This reduces the rigid tension caused by positional errors, posture errors, and asynchronous movements of the embodied robot's two arms, while balancing the workpiece's trajectory and the slave arm's gripping force. Vision-force fusion positioning, dynamic load distribution, and tiered safety protection supplement workpiece positioning, load balancing between the embodied robot's two arms, and collaborative operation safety, respectively.

[0105] In one prototype test example, compared with the corresponding purely passive floating scheme or purely active control scheme, the above combined scheme improves error compensation capability by more than 300%, reduces downtime by more than 95%, and increases mean time between failures (MTBF) from 200 hours to more than 4000 hours. The above test results correspond to specific prototype configurations and test conditions, and are used to illustrate one implementation effect of the combined scheme, and do not constitute a limitation on other implementation methods.

[0106] Based on the same inventive concept, this application also provides a force-position hybrid cooperative control system for a dual-arm floating grasping system of a body robot, including: a mobile chassis 1, a first robotic arm 7, a second robotic arm 3, a composite floating mechanism, a first end effector, a second end effector, and a central cooperative controller. The first robotic arm 7 and the second robotic arm 3 are mounted on the mobile chassis. The end of the first robotic arm 7 is connected to the first end effector through the composite floating mechanism, and the end of the second robotic arm 3 is fixedly connected to the second end effector. The central collaborative controller is communicatively connected to the first robotic arm 7 and the second robotic arm 3 respectively, and is used to acquire the target motion trajectory and target grasping force corresponding to the workpiece to be grasped, and control the first end effector and the second end effector to jointly grasp the workpiece; The central coordinating controller is also used to control the movement of the second robotic arm according to the target motion trajectory, so that the second end effector drives the workpiece to move along the target motion trajectory; The composite floating mechanism is used to make the first end effector float relative to the end of the first robotic arm during the movement of the workpiece, so as to compensate for the relative pose deviation between the first end effector and the second end effector. The central coordinating controller is also used to acquire the actual gripping force of the first end effector acting on the workpiece, obtain the force deviation based on the target gripping force and the actual gripping force, generate force control commands for the first robotic arm based on the force deviation, and control the movement of the first robotic arm based on the force control commands, so that the actual gripping force tracks the target gripping force during the process of the second end effector driving the workpiece to move along the target motion trajectory.

[0107] Specifically, such as Figure 2 As shown, the mobile chassis 1 can be an autonomous mobile robot (AMR) chassis. The AMR chassis adopts a submerged lifting AMR structure, with a maximum load capacity of 1500kg. It is equipped with LiDAR, depth camera and ultrasonic sensor to achieve 360° omnidirectional obstacle avoidance.

[0108] Preferably, the AMR chassis 1 employs a navigation method that combines Simultaneous Localization and Mapping (SLAM) with visual fusion. It performs positioning and navigation based on environmental information collected by the lidar and visual information collected by the depth camera, achieving a navigation accuracy of ±3 mm. The AMR chassis 1 also includes an edge computing unit, which performs local path planning and obstacle avoidance control based on the environmental information collected by the AMR chassis 1, enabling the AMR chassis 1 to autonomously travel to the production line docking position or workstation.

[0109] The first robotic arm 7 and the second robotic arm 3 are both collaborative robots. The first robotic arm 7 and the second robotic arm 3 are respectively installed on both sides of the upper surface of the AMR chassis 1. The composite floating mechanism and the first end effector together constitute the floating end tool 6. The end of the first robotic arm 7 is connected to the first end effector through the composite floating mechanism. The second end effector is a fixed end tool 4. The end of the second robotic arm 3 is fixedly connected to the fixed end tool 4.

[0110] In this embodiment, both the first robotic arm 7 and the second robotic arm 3 are six-axis collaborative robots. The payload of a single collaborative robot can be 30 kg, the working radius can be 1800 mm, and the repeatability can be ±0.02 mm. Each joint integrates a torque sensor. The end effector of the first robotic arm 7 is equipped with a six-dimensional force sensor and a floating end effector 6. The six-dimensional force sensor is located between the end effector of the first robotic arm 7 and the composite floating mechanism, and is used to detect the actual gripping force of the first end effector on the workpiece.

[0111] The second robotic arm 3 is equipped with a fixed end effector 4 at its end. The first and second end effectors grip different positions of the workpiece to jointly carry and transport it.

[0112] Preferably, the AMR chassis 1 is also provided with a powered roller conveyor 2, which is used to convey pallets when the AMR chassis 1 is connected to the production line conveyor. In one example, the effective conveying width of the powered roller conveyor 2 is 800 mm and the roller spacing is 50 mm, which can convey European standard pallets as well as customized pallets set according to the workpiece size.

[0113] The central coordinating controller employs a multi-core industrial processor and runs a real-time operating system, communicating with the AMR chassis 1, the powered roller conveyor 2, the first robotic arm 7, the second robotic arm 3, the working camera 5, and the six-dimensional force sensor. The central coordinating controller acquires the force information output by the six-dimensional force sensor and determines the actual gripping force exerted on the workpiece by the first end effector based on this force information. The central coordinating controller also acquires the target motion trajectory and target gripping force corresponding to the workpiece to be gripped, and controls the first and second end effectors to grip the workpiece together. Furthermore, the central coordinating controller acquires image data and spatial data collected by the working camera 5, and determines the workpiece's pose based on the image data and spatial data. It also controls the operating speed of the powered roller conveyor 2 according to the conveying speed of the production line, ensuring that the powered roller conveyor 2 operates synchronously with the production line.

[0114] After the robot's two arms jointly grasp the workpiece, the central collaborative controller uses the second robotic arm 3 as the position control side, generates position control commands for the second robotic arm 3 according to the target motion trajectory, and controls the movement of the second robotic arm 3 according to the position control commands, so that the fixed end tool 4 drives the workpiece to move along the target motion trajectory.

[0115] During workpiece movement, the first robotic arm 7 and the second robotic arm 3 may experience relative pose deviations due to manufacturing precision, installation position, coordinate calibration, control response, and motion synchronization errors. When both end effectors jointly grasp the same workpiece, the workpiece transmits the external force corresponding to the relative pose deviation to the first end effector, causing the first end effector to passively float relative to the end of the first robotic arm 7 along with the composite floating mechanism. The composite floating mechanism absorbs at least part of the relative pose deviation through this passive floating, reducing the conversion of the relative pose deviation into rigid pulling forces acting on the workpiece and the two collaborative robots.

[0116] The central coordinating controller uses the first robotic arm 7 as the force control side. It continuously acquires the actual gripping force exerted on the workpiece by the first end effector through a six-dimensional force sensor. The controller compares the actual gripping force with the target gripping force to obtain the force deviation. Based on this force deviation, the central coordinating controller generates force control commands for the first robotic arm 7 and controls its movement accordingly. When the actual gripping force is greater than the target gripping force, the controller moves the first robotic arm 7 in the direction that decreases the actual gripping force; when the actual gripping force is less than the target gripping force, the controller moves the first robotic arm 7 in the direction that increases the actual gripping force, ensuring that the actual gripping force tracks the target gripping force.

[0117] While the second robotic arm 3 moves the workpiece along the target trajectory, the first robotic arm 7 adjusts its motion according to force control commands. The composite floating mechanism adapts to the relative pose deviation between the two end effectors through passive floating. The position control of the second robotic arm 3, the force control of the first robotic arm 7, and the passive floating of the composite floating mechanism work together to make the workpiece move along the target trajectory and reduce the fluctuation of gripping force caused by the relative pose deviation and asynchronous movement of the two arms of the robot.

[0118] In this embodiment, the motion trajectory of the workpiece is controlled by the second robotic arm 3 and the fixed end effector 4. The relative posture deviation between the two end effectors is absorbed by the composite floating mechanism, and the force control of the first robotic arm 7 makes the actual grasping force track the target grasping force. This can reduce the rigid pulling force generated when two collaborative robots grasp the same workpiece together, and improve the stability of the robot's dual-arm grasping and handling of workpieces.

[0119] In some embodiments, the composite floating mechanism includes a mounting base, a planar floating component, a floating guide component, a swing connection assembly, and a reset assembly; The mounting base is connected to the end of the first robotic arm; The planar floating component is movably disposed on the mounting base and is capable of moving relative to the mounting base along a plane; The floating guide is inserted through the planar floating member and can move axially relative to the planar floating member; The swing connection assembly is disposed between the floating guide and the first end effector, enabling the first end effector to swing relative to the floating guide. The reset component acts on the planar floating component, the floating guide component, and the swing connection component respectively, and is used to reset the planar floating component, the floating guide component, and the swing connection component when the external force on the first end effector decreases or disappears.

[0120] Specifically, such as Figure 3 As shown, the mounting base includes a circular connecting flange 601 for the robot end effector, a first circular transition frame 602, and a second circular transition frame 605; the planar floating component is a horizontal floating disk 607; the floating guide component is a floating guide rod 609; the swing connection assembly includes a circular ball head pressure seat 610, a floating ball head 611, and a floating ball head connecting plate 613; the reset assembly includes a buffer spring pressure plate 603, a first buffer spring 604, a center-aligning ball head plunger 606, a second buffer spring 608, and a planar centering ball head plunger 612; and the first end effector can be an end gripper 614.

[0121] In practice, the mounting base is connected to the end of the first robotic arm via a six-dimensional force sensor. Specifically, the circular connecting flange 601 at the robot end is connected to the six-dimensional force sensor via bolts, and the six-dimensional force sensor is connected to the end of the first robotic arm. A first circular transition frame 602 is fixed below the connecting flange 601, and a first buffer spring plate 603 is fixed to the inner center of the first circular transition frame 602 via bolts. A first buffer spring 604 is sleeved on the upper part of the floating guide rod 609, with its upper end contacting the first buffer spring plate 603 and its lower end contacting the second circular transition frame 605.

[0122] The outer side of the second circular transition frame 605 is connected to the horizontal floating disk 607 via four sets of center-aligning ball-head plungers 606. The horizontal floating disk 607 can achieve ±8mm translational floating relative to the second circular transition frame 605 in the XY plane. The floating guide rod 609 passes through the central hole of the horizontal floating disk 607, and its lower end is fixedly connected to the floating ball head 611. The buffer spring 608 is sleeved on the lower part of the floating guide rod 609, with its upper end contacting the horizontal floating disk 607 and its lower end contacting the circular ball head pressure seat 610.

[0123] A circular ball head pressure seat 610 is fixed below the horizontal floating disk 607. A floating ball head 611 is installed inside the circular ball head pressure seat 610, allowing for swinging in any direction within ±30°. A floating ball head connecting plate 613 is connected to the circular ball head pressure seat 610 via four sets of planar centering ball head plungers 612, automatically centering the floating ball head 611 when no external force is applied. An end gripper 614 is fixed below the floating ball head connecting plate 613.

[0124] The core control flow of this embodiment is described below: System initialization: After the system is powered on, the central coordinating controller performs a self-test on all hardware and initializes the floating end effector to the center position.

[0125] Task reception: The central collaborative controller receives job tasks through the MES system, including information such as workpiece type, quantity, and target location.

[0126] Production line docking: The AMR chassis autonomously navigates to the docking position on the production line, communicates with the production line PLC, and obtains the production line status and cargo information.

[0127] Cargo loading: The powered roller conveyor starts and operates synchronously with the production line to load pallets of cargo onto the AMR chassis.

[0128] Visual scanning: The second collaborative robot 3 (i.e. the main arm) moves the working camera above the pallet to scan the pallet and identify the workpiece information.

[0129] Parameter calculation: The central collaborative controller calculates the optimal gripping point, load distribution ratio, and control gain parameters based on the workpiece parameters.

[0130] Collaborative gripping: The main arm moves to the gripping position according to the planned trajectory; the main arm moves according to the planned target motion trajectory, and the slave arm performs force control based on the force deviation between the actual gripping force and the target gripping force, and moves in coordination with the main arm. At the same time, the six-dimensional force sensor detects force information in real time; when the gripper contacts the workpiece, it switches to force control mode to keep the gripping force constant (that is, the actual gripping force tracks the target gripping force); after gripping is completed, the two robots work together to transport the workpiece to the target position.

[0131] Workpiece placement: After reaching the target position, the two robots work together to place the workpiece in place and release the gripper.

[0132] Task completed: Once all workpieces have been processed, the AMR chassis will transport the empty pallet to the recycling area and then return to the loading area to await the next task.

[0133] The working principle of this embodiment is explained below: 1. Automated production line docking and cargo loading: The AMR chassis autonomously navigates to the docking position of the production line conveyor and interacts with the production line's programmable logic controller (PLC) via the Modbus TCP protocol to obtain the production line's conveyor speed and cargo information.

[0134] During autonomous navigation, the AMR chassis 1 collects surrounding environmental information through LiDAR and depth camera, and determines its own position using a navigation method that combines LiDAR SLAM and vision. The edge computing unit performs local path planning and obstacle avoidance control based on the environmental information and the current position of the AMR chassis 1, so that the AMR chassis 1 can travel to the docking position of the production line conveyor.

[0135] The central coordinating controller adjusts the operating speed of the powered roller conveyor 2 according to the conveying speed, so that the powered roller conveyor 2 runs synchronously with the production line conveyor line and smoothly loads the entire pallet of goods onto the AMR chassis.

[0136] It should be noted that the powered roller conveyor 2 can carry and transport European standard pallets or customized pallets set according to the workpiece size.

[0137] After loading of goods, the AMR chassis automatically drives away from the production line and is transferred to the work station.

[0138] 2. Visual-mechanical fusion positioning: After the AMR chassis 1 arrives at the work station, the second collaborative robot 3 (i.e., the main arm) moves the working camera 5 above the pallet to scan the pallet and the workpiece on it. The working camera 5 can be a 3D industrial camera, used to acquire two-dimensional image data and three-dimensional point cloud data of the workpiece.

[0139] The central collaborative controller processes the image and point cloud data, identifies the type, location, and orientation of each workpiece, and generates a grasping task queue.

[0140] The central collaborative controller sequentially determines the workpieces to be grasped according to the grasping task queue, and automatically calculates the optimal grasping point, load distribution ratio and control gain parameters based on the type, position, posture and corresponding workpiece parameters of the workpieces to be grasped.

[0141] During the grasping process, a six-dimensional force sensor located at the end of the first collaborative robot 7 (i.e., the first robotic arm 7) detects the contact force between the first end effector and the workpiece. The central collaborative controller can correct the workpiece's pose estimation based on the contact force to combine visual positioning with force feedback.

[0142] 3. Collaborative grasping operation using two arms of a hymenoplasty: The central coordinating controller sends motion commands to both robots simultaneously. The main arm moves along the planned trajectory, and the secondary arm moves along the planned target trajectory. The secondary arm performs force control based on the force deviation between the actual gripping force and the target gripping force, and moves in coordination with the main arm. When the gripper approaches the workpiece, it switches to force control mode to achieve compliant contact. After gripping, the two robots work together to transport the workpiece to the target position.

[0143] Throughout the entire process, the composite floating mechanism absorbs most of the positional error, and the active force control algorithm compensates for the remaining error, ensuring that the gripping force is constant and there is no pulling force, and making the actual gripping force approach the target gripping force.

[0144] 4. Workpiece placement and empty pallet recycling: Once the workpiece reaches the target position, the two robots work together to place it in place and release the gripper. After placing the current workpiece, the central collaborative controller determines the next workpiece to be gripped from the gripping task queue, and repeats the above steps until all workpieces on the pallet have been processed.

[0145] After the workpieces on the pallet are processed, the AMR chassis drives the power roller conveyor to automatically transport the empty pallet to the empty pallet recycling area of ​​the production line, and then returns to the loading area to continue the operation.

[0146] This application adopts a dual-arm collaborative operation mode of embodied robot, which improves the operation efficiency by more than 120% compared with single-arm robot; This application improves the grasping accuracy from ±0.5mm to ±0.1mm through vision-force fusion positioning and grasping technology; This application utilizes a dynamic load distribution algorithm to increase the maximum gripping weight of the system from 30kg to 50kg per robot, thereby increasing the load capacity by 67%.

[0147] The powered roller conveyor in this application can be seamlessly integrated with any standard production line, with a 100% success rate.

[0148] In this specification, the same or similar parts between the various embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the descriptions of the embodiments described later are relatively simple, and relevant parts can be referred to the descriptions of the foregoing embodiments.

[0149] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A force-position hybrid cooperative control method for dual-arm floating grasping of an embodied robot, characterized in that, Applied to a dual-arm robot, the dual-arm robot includes a first robotic arm and a second robotic arm, the end of the first robotic arm is connected to a first end effector via a composite floating mechanism, and the end of the second robotic arm is fixedly connected to a second end effector; the method includes: Obtain the target motion trajectory corresponding to the workpiece to be grasped; Obtain the target gripping force corresponding to the workpiece; Control the first end effector and the second end effector to jointly grasp the workpiece; The second robotic arm is controlled to move according to the target motion trajectory, so that the second end effector drives the workpiece to move along the target motion trajectory. During the movement of the workpiece, the first end effector floats relative to the end of the first robotic arm along with the composite floating mechanism to compensate for the relative pose deviation between the first end effector and the second end effector. The actual gripping force exerted by the first end effector on the workpiece is obtained, and the force deviation is obtained based on the target gripping force and the actual gripping force. Based on the force deviation, force control commands for the first robotic arm are generated; The first robotic arm is controlled to move according to the force control command, so that the actual grasping force tracks the target grasping force as the second robotic arm moves along the target motion trajectory.

2. The force-position hybrid cooperative control method for dual-arm floating grasping of a unibody robot according to claim 1, characterized in that, The composite floating mechanism provides the first end effector with translational and rotational floating degrees of freedom. During the movement of the workpiece, the first end effector, under the action of the workpiece, translates and rotates relative to the end of the first robotic arm through the composite floating mechanism; Based on the translation and the rotation, the pose of the first end effector relative to the second end effector is adjusted to reduce the relative pose deviation.

3. The force-position hybrid cooperative control method for dual-arm floating grasping of a unibody robot according to claim 1, characterized in that, The step of generating force control commands for the first robotic arm based on the force deviation includes: Based on the actual gripping force, information on the change in actual gripping force is obtained; Based on the target grasping force, information on the change in target grasping force is obtained; Based on the actual gripping force change information and the target gripping force change information, the force change deviation is obtained; Based on the force deviation and the force change deviation, force control commands for the first robotic arm are generated; The actual gripping force change information is used to characterize the change of the actual gripping force over time, and the target gripping force change information is used to characterize the change of the target gripping force over time.

4. The force-position hybrid cooperative control method for dual-arm floating grasping of a unibody robot according to claim 3, characterized in that, The step of generating force control commands for the first robotic arm based on the force deviation and the force change deviation includes: The direction of gripping force adjustment is determined based on the relationship between the actual gripping force and the target gripping force. The gripping force adjustment amount is determined based on the force deviation and the force change deviation; The force control command for the first robotic arm is generated based on the gripping force adjustment direction and the gripping force adjustment amount; When the actual gripping force is greater than the target gripping force, the gripping force adjustment direction is to decrease the actual gripping force; when the actual gripping force is less than the target gripping force, the gripping force adjustment direction is to increase the actual gripping force.

5. The force-position hybrid cooperative control method for dual-arm floating grasping of a unibody robot according to claim 1, characterized in that, The step of controlling the movement of the second robotic arm according to the target motion trajectory includes: Based on the target motion trajectory, determine the target motion state of the second end effector; Obtain the actual motion state of the second end effector; The motion state deviation is obtained based on the target motion state and the actual motion state; Based on the motion state deviation, position control commands for the second robotic arm are generated; The second robotic arm is controlled to move according to the position control command.

6. The force-position hybrid cooperative control method for dual-arm floating grasping of an embodied robot according to claim 5, characterized in that, The target motion state includes the target position and the target velocity, and the actual motion state includes the actual position and the actual velocity; The position deviation is obtained based on the target position and the actual position, and the speed deviation is obtained based on the target speed and the actual speed; Based on the position deviation and the speed deviation, position control commands for the second robotic arm are generated.

7. The force-position hybrid cooperative control method for dual-arm floating grasping of a unibody robot according to claim 1, characterized in that, The method further includes: The center of gravity position of the workpiece, and the gripping positions of the first end effector and the second end effector on the workpiece are obtained; Based on the center of gravity position and the grasping position, determine the target load distribution ratio between the first robotic arm and the second robotic arm; Obtain the actual loads of the first robotic arm and the second robotic arm; Based on the actual load, the actual load distribution ratio between the first robotic arm and the second robotic arm is obtained; Based on the deviation between the actual load distribution ratio and the target load distribution ratio, adjust the control output of at least one robotic arm so that the actual load distribution ratio approaches the target load distribution ratio.

8. The force-position hybrid cooperative control method for dual-arm floating grasping of a unibody robot according to claim 1, characterized in that, The method further includes: Obtain motion state information of the first robotic arm and the second robotic arm; Based on the motion state information, determine the minimum distance between the first robotic arm and the second robotic arm; When the minimum distance is less than the first safety distance and not less than the second safety distance, the movement speed of at least one robotic arm is reduced, where the first safety distance is greater than the second safety distance. When the minimum distance is less than the second safe distance, the first robotic arm and the second robotic arm are controlled to stop moving and an alarm message is output. Obtain collision detection information between the first robotic arm and the second robotic arm; When the collision detection information of either robotic arm indicates that a collision has occurred, control the first robotic arm and the second robotic arm to stop moving; When the collision force on the robotic arm exceeds a preset force, the power output of the dual-arm robot is cut off.

9. A force-position hybrid cooperative control system for a unibody robot with floating dual arms for grasping, characterized in that, include: Mobile chassis, first robotic arm, second robotic arm, composite floating mechanism, first end effector, second end effector, and central coordinating controller; The first robotic arm and the second robotic arm are mounted on the mobile chassis. The end of the first robotic arm is connected to the first end effector via the composite floating mechanism, and the end of the second robotic arm is fixedly connected to the second end effector. The central collaborative controller is communicatively connected to the first robotic arm and the second robotic arm respectively, and is used to acquire the target motion trajectory and target grasping force corresponding to the workpiece to be grasped, and control the first end effector and the second end effector to jointly grasp the workpiece; The central coordinating controller is also used to control the movement of the second robotic arm according to the target motion trajectory, so that the second end effector drives the workpiece to move along the target motion trajectory; The composite floating mechanism is used to make the first end effector float relative to the end of the first robotic arm during the movement of the workpiece, so as to compensate for the relative pose deviation between the first end effector and the second end effector. The central coordinating controller is also used to acquire the actual gripping force of the first end effector acting on the workpiece, obtain the force deviation based on the target gripping force and the actual gripping force, generate force control commands for the first robotic arm based on the force deviation, and control the movement of the first robotic arm based on the force control commands, so that the actual gripping force tracks the target gripping force during the process of the second end effector driving the workpiece to move along the target motion trajectory.

10. The force-position hybrid cooperative control system for the dual-arm floating grasping of the embodied robot according to claim 9, characterized in that, The composite floating mechanism includes a mounting base, a planar floating component, a floating guide component, a swing connection assembly, and a reset assembly; The mounting base is connected to the end of the first robotic arm; The planar floating component is movably disposed on the mounting base and is capable of moving relative to the mounting base along a plane; The floating guide is inserted through the planar floating member and can move axially relative to the planar floating member; The swing connection assembly is disposed between the floating guide and the first end effector, enabling the first end effector to swing relative to the floating guide. The reset component acts on the planar floating component, the floating guide component, and the swing connection component respectively, and is used to reset the planar floating component, the floating guide component, and the swing connection component when the external force on the first end effector decreases or disappears.

Images