A collaborative control method, device, product and medium for humanoid robot construction

Abstract

This invention relates to a collaborative control method, equipment, product, and medium for humanoid robot construction, within the field of robot control. Addressing the spatial interference and deadlock problems that easily arise when multiple robots collaboratively handle the same rigid workpiece, this method, after acquiring the robot joint states and workpiece pose, divides the links into end-locked links and non-end-floating links. When the distance between the non-end-floating links is lower than a preset safety threshold, the interference vector is calculated and decomposed into an absorbable component within the link and a residual component outside the link while maintaining the workpiece pose. If the residual component outside the link does not exceed the threshold, the non-end-floating links are adjusted to eliminate interference; if it does, the relative displacement and safe adjustment posture are calculated, controlling the two robots to collaboratively move the chassis and adjust the workpiece pose. This application prioritizes utilizing redundant degrees of freedom of the robot body to resolve interference, avoiding workpiece slippage or physical deadlock caused by blind avoidance in confined spaces, thus improving the reliability of collaborative robot construction.

Description

Technical Field

[0001] This application relates to the field of robot control, and in particular to a collaborative control method, equipment, product and medium for humanoid robot construction. Background Technology

[0002] With the development of intelligent construction technology, the use of multiple humanoid robots to collaboratively complete construction tasks (such as collaborative lifting and installation of prefabricated components) has become a trend. When performing multi-step collaborative processes in confined and narrow spaces, multiple robots usually need to perform intensive operations around the same workpiece. Due to the small working space and the high degree of freedom of humanoid robots' limbs, spatial interference and physical collisions may occur between robots during the alternating execution of actions, which may lead to construction interruptions or even equipment damage.

[0003] To address the aforementioned collision issues, relevant technologies typically employ a replanning avoidance control method based on envelope distance detection. Specifically, this method establishes a virtual safety envelope space outside each robot's limb links and calculates the minimum distance between the envelope spaces of multiple robots in real time during collaborative construction. Once interference between the limbs of two robots is predicted, the control system immediately pauses all motion commands, instructing one robot to retreat or move its chassis to create space. A collision-free detour trajectory is then recalculated and planned for the interfering limb of the other robot in a three-dimensional coordinate system. The original construction sequence resumes only after the detour has been completed.

[0004] However, in physically coupled scenarios where multiple robots share the same rigid workpiece, the end effectors (hands) of these robots simultaneously grasp the same workpiece, effectively forming a forced, closed-loop kinematic chain in space. If one robot is instructed to retreat from its chassis or significantly replan its trajectory to avoid obstacles, the absolute spatial coordinates of its end effector will change, potentially causing workpiece damage or slippage. Conversely, if the coordinates of the end effector are forcibly locked to prevent workpiece slippage, in extremely confined spaces, the relevant technical solutions often struggle to calculate any effective detour paths for the interfering limbs, ultimately leading to a deadlock among the multiple robots and reducing the reliability of collaborative robot construction. Summary of the Invention

[0005] This application provides a collaborative control method, equipment, product, and medium for humanoid robot construction, which improves the reliability and continuity of collaborative robot construction in space-constrained environments.

[0006] In a first aspect, this application provides a collaborative control method for humanoid robot construction, applied to a collaborative control device. The collaborative control device is communicatively connected to a first robot and a second robot, respectively. The method includes: acquiring joint state data of the first robot and the second robot and pose data of the rigid workpiece while the first robot and the second robot are jointly gripping the same rigid workpiece; based on the joint state data and pose data, dividing the links between the first robot and the second robot into end-locked links and non-end-floating links, wherein the end-locked links are links that directly contact the rigid workpiece and perform gripping actions; when the distance between the non-end-floating links of the first robot and the second robot is lower than a preset safety threshold, calculating the interference between the non-end-floating links. Interference vector; while keeping the rigid workpiece's pose unchanged, the interference vector is decomposed into an absorbable component within the chain and a residual component outside the chain; the absorbable component within the chain is the amount of interference that can be eliminated by adjusting the joint angles of the non-end floating links while keeping the rigid workpiece's pose unchanged; if the residual component outside the chain does not exceed a preset residual threshold, the corresponding non-end floating links are adjusted according to the absorbable component within the chain to eliminate interference; if the residual component outside the chain exceeds the preset residual threshold, the relative displacement between the chassis of the first robot and the second robot, and the corresponding safe adjustment posture of the rigid workpiece are calculated; based on the relative displacement and the safe adjustment posture, the first robot and the second robot are controlled to move their chassis together and adjust the pose of the rigid workpiece to eliminate interference.

[0007] By adopting the above technical solution, the collaborative control equipment first divides the robot's links into end-capped locking links and non-end-capped floating links when the robots are jointly gripping the same rigid workpiece. This clarifies which limbs are restricted and which have adjustment margins under physical gripping constraints. Second, when the distance is lower than a preset safety threshold, the interference vector is calculated and decomposed into absorbable components within the chain and residual components outside the chain. This quantifies and distinguishes between interference quantities that can be resolved using the robot's own redundant degrees of freedom and those that cannot be resolved autonomously. Finally, based on the residual components outside the chain, a graded strategy is adopted, either by adjusting joint angles or by collaboratively moving the chassis and adjusting the rigid workpiece's pose. Under the premise of ensuring workpiece pose stability, interference is absorbed primarily using the robot's own redundant degrees of freedom. In summary, this solution effectively avoids physical deadlock or workpiece slippage caused by spatial interference when robots are jointly gripping workpieces in confined spaces, improving the reliability and operational continuity of multi-robot collaborative construction.

[0008] In conjunction with some embodiments of the first aspect, in some embodiments, based on joint state data and pose data, the links between the first robot and the second robot are divided into end-locked links and non-end-floating links. Specifically, this includes: determining the kinematic chain topology of the first robot and the second robot according to the joint state data; identifying links in the first robot and the second robot that form a fixed constraint relationship with the rigid workpiece based on the pose data of the rigid workpiece, and determining them as end-locked links; and determining all intermediate links in the kinematic chain topology from the chassis to the end-locked links as non-end-floating links.

[0009] By adopting the above technical solution, the collaborative control equipment determines the kinematic chain topology of each robot based on the joint state data, and grasps the physical connection and hierarchical relationship of the robot's joints. Secondly, based on the workpiece pose data, the end-locking link that forms a fixed constraint relationship with the workpiece is identified, and then the intermediate link between the chassis and the end-locking link is determined as a non-end-floating link. This provides an accurate structured control model for interference decomposition under closed-loop constraints and improves the accuracy of link attribute division in the hierarchical collision avoidance strategy.

[0010] In conjunction with some embodiments of the first aspect, in some embodiments, calculating the relative displacement between the chassis of the first robot and the second robot, and the corresponding safe adjustment posture of the rigid workpiece, specifically includes: removing the constraint condition that the rigid workpiece's pose remains unchanged; obtaining spatial constraint information of the current construction scene, including the movable area of ​​the chassis and the distribution of obstacles; calculating the relative displacement between the chassis of the first robot and the second robot based on the spatial constraint information and the residual components outside the chain; and calculating the safe adjustment posture of the rigid workpiece that matches the relative displacement based on the relative displacement.

[0011] By adopting the above technical solution, the collaborative control equipment removes the constraint of the rigid workpiece's unchanged pose. When the body's degrees of freedom are exhausted, the forcibly bound end-effector constraints are released to find avoidance space. Secondly, based on spatial constraint information and residual components outside the chain, the relative displacement of the dual-chassis is calculated. Under the premise of not colliding with the construction environment, an optimal chassis motion strategy to break the spatial deadlock is obtained. Finally, based on the relative displacement, a matching safe adjustment posture is calculated, ensuring that the workpiece and gripping system remain in a safe and compliant state after the chassis moves. In summary, this solution provides a reliable overall collaborative reconstruction path for deadlock states in extremely confined spaces, improving the global solution capability of collision avoidance planning.

[0012] In some embodiments of the first aspect, after controlling the first robot and the second robot to move the chassis and adjust the position of the rigid workpiece in coordination according to the relative displacement and safety adjustment posture, the method further includes: after the current construction action is completed, controlling the first robot and the second robot to synchronously reset with the rigid workpiece.

[0013] By adopting the above technical solution, after the chassis collaborative movement and posture adjustment for obstacle avoidance, the collaborative control equipment can control the robot to smoothly restore the workpiece and body configuration to the predetermined standard state or adapt to the requirements of the next process. This eliminates the impact of emergency collision avoidance actions on the formation deviation of subsequent construction processes, ensures the continuity of multi-machine collaborative construction operations and seamless connection between processes, and improves the overall construction efficiency.

[0014] In conjunction with some embodiments of the first aspect, in some embodiments, before calculating the relative displacement between the first robot and the second robot chassis based on spatial constraint information and the residual component outside the chain, the method further includes: calculating the maximum movable distance of the first robot and the second robot chassis in each movement direction based on spatial constraint information; when the maximum movable distance in all movement directions is less than the modulus of the residual component outside the chain, adjusting the residual component outside the chain according to the preset attitude tolerance range of the rigid workpiece.

[0015] By adopting the above technical solution, the collaborative control equipment determines the motion limits of the chassis in each direction under the current environment based on spatial constraint information. When the movable distance in all directions cannot meet the requirements for eliminating the residual component modulus outside the chain, the residual component outside the chain is adjusted according to the preset attitude tolerance range of the rigid workpiece. By making reasonable use of the positional deviation space of the rigid workpiece within the allowable range of the specification, the physical displacement required by the chassis is shared and reduced. Thus, in narrow scenarios where the chassis mobility is limited, the success rate of collision avoidance planning under extreme working conditions is improved.

[0016] In conjunction with some embodiments of the first aspect, in some embodiments, the interference vector is decomposed into an in-chain absorbable component and an out-of-chain residual component. Specifically, this includes: calculating the degree-of-freedom redundancy of each joint in the non-end-of-chain floating link based on joint state data and pose data. The degree-of-freedom redundancy refers to the range of angles that the corresponding joint can be adjusted while keeping the pose of the end-of-chain locked link unchanged; based on the degree-of-freedom redundancy, calculating the maximum displacement vector that can be achieved by adjusting the joint angle of the non-end-of-chain floating link with the interference vector as the target direction, and determining it as the in-chain absorbable component; and using the difference vector between the interference vector and the in-chain absorbable component as the out-of-chain residual component.

[0017] By adopting the above technical solution, the collaborative control device first calculates the degree-of-freedom redundancy of each joint of the non-end-of-link floating linkage under the condition of maintaining the end-of-link pose, quantitatively determining the actual available collision avoidance space of each joint. Secondly, using the interference vector as the target direction, the maximum achievable displacement vector based on the redundancy is determined as the absorbable component within the chain, clarifying the interference limit that can be absorbed by self-attitude adjustment alone. Finally, the difference between the interference vector and the absorbable component within the chain is taken as the residual component outside the chain, separating the interference portion that must be resolved through chassis reconstruction. In summary, this solution achieves scientific decoupling of the interference vector under closed-loop motion constraints, providing a reliable quantitative basis for triggering subsequent graded collision avoidance strategies.

[0018] In conjunction with some embodiments of the first aspect, in some embodiments, after calculating the interference vector between the non-end floating links, the method further includes: acquiring the current joint torque data of the first robot and the second robot, and the mass parameters of the rigid workpiece; calculating the deformation compensation vector of the non-end floating links under load based on the joint torque data and the mass parameters, and correcting the interference vector according to the deformation compensation vector.

[0019] By adopting the above technical solution, the collaborative control equipment calculates the deformation compensation vector of the non-end floating link under load using the current joint torque data and the mass parameters of the rigid workpiece. This quantifies the actual spatial deviation caused by the downward pressure of the heavy object and the flexibility of the link. The deformation compensation vector is then used to correct the original interference vector, incorporating mechanical deformation factors into the collision avoidance considerations and improving the accuracy of interference detection and collision avoidance calculation in the collaborative handling scenario of heavy workpieces.

[0020] In a second aspect, this application provides a cooperative control device, which includes: one or more processors and a memory; the memory is coupled to one or more processors, and the memory is used to store computer program code, the computer program code including computer instructions, and the one or more processors call the computer instructions to cause the cooperative control device to perform the method described in the first aspect and any possible implementation thereof.

[0021] Thirdly, this application provides a computer-readable storage medium storing computer instructions that, when executed on a cooperative control device, cause the cooperative control device to perform the method described in the first aspect and any possible implementation thereof.

[0022] Fourthly, this application provides a computer program product, including a computer program or instructions that, when run on a cooperative control device, cause the cooperative control device to perform the method described in the first aspect and any possible implementation thereof.

[0023] It is understood that the collaborative control device provided in the second aspect, the computer-readable storage medium provided in the third aspect, and the computer program product provided in the fourth aspect are all used to execute the methods provided in the embodiments of this application. Therefore, the beneficial effects they can achieve can be referred to the beneficial effects in the corresponding methods, and will not be repeated here.

[0024] One or more technical solutions provided in the embodiments of this application have at least the following technical effects or advantages:

[0025] 1. By adopting a hierarchical strategy that decomposes the interference vector into absorbable components within the chain and residual components outside the chain while keeping the rigid workpiece's pose unchanged, and then adjusting the non-end-mounted floating links or coordinating chassis movement and adjusting the rigid workpiece's pose based on whether the residual components outside the chain exceed a preset residual threshold, the robot can prioritize utilizing the robot's redundant degrees of freedom to absorb interference, and only perform chassis collaborative reconfiguration when necessary. This effectively solves the problems of workpiece slippage caused by blindly replanning the trajectory or deadlock in extremely narrow spaces caused by forcibly locking the end-mounted components when multiple robots are holding rigid workpieces in related technologies. As a result, safe, continuous, and reliable collaborative construction operations of multiple humanoid robots in confined spaces are achieved.

[0026] 2. By adopting the constraint of removing the rigid workpiece's pose invariance, and based on the spatial constraint information including the movable area of ​​the chassis and the distribution of obstacles, as well as the residual components outside the chain, the relative displacement between the chassis of the first robot and the second robot and the safe adjustment posture of the matching rigid workpiece are calculated. Therefore, when the robot's degrees of freedom are exhausted, the avoidance solution space can be expanded and a collision-free overall reconstruction path can be planned under environmental constraints. This effectively solves the problem of task deadlock and stagnation caused by the failure of local trajectory replanning in extremely restricted construction scenarios in related technologies, and thus achieves global safe collision avoidance and working condition continuity under complex obstacle distribution.

[0027] 3. By acquiring the current joint torque data of the first and second robots and the mass parameters of the rigid workpiece, and calculating the deformation compensation vector of the non-end floating link under load based on the joint torque data and mass parameters, and then correcting the interference vector according to the deformation compensation vector, it is possible to quantify and compensate for the physical micro-deformation of the robot link caused by the heavy workpiece. This effectively solves the problem of actual envelope spacing prediction error caused by ignoring the link deformation under heavy load conditions in related technologies, thereby improving the accuracy of interference vector calculation and the reliability of collision avoidance in heavy-load collaborative construction operations. Attached Figure Description

[0028] Figure 1 This is a schematic diagram of a scenario of a collaborative control method for humanoid robot construction in an embodiment of this application;

[0029] Figure 2 This is a flowchart illustrating a collaborative control method for humanoid robot construction in an embodiment of this application;

[0030] Figure 3 This is another flowchart illustrating a collaborative control method for humanoid robot construction in an embodiment of this application;

[0031] Figure 4 This is a schematic diagram of the physical device structure of a collaborative control device in the embodiments of this application. Detailed Implementation

[0032] The terminology used in the following embodiments of this application is for the purpose of describing particular embodiments only and is not intended to be limiting of this application. As used in the specification of this application, the singular expressions “a,” “an,” “the,” “the,” and “this” are intended to include the plural expressions as well, unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used in this application refers to any or all possible combinations including one or more of the listed items.

[0033] Hereinafter, the terms "first" and "second" are used for descriptive purposes only and should not be construed as implying or suggesting relative importance or implicitly indicating the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature, and in the description of the embodiments of this application, unless otherwise stated, "multiple" means two or more.

[0034] This application provides a collaborative control method for humanoid robot construction, which is described below in conjunction with... Figure 1 This application describes the application scenarios of the embodiments. Figure 1 This is a schematic diagram of a scenario of a collaborative control method for humanoid robot construction in an embodiment of this application.

[0035] In related technologies, collision avoidance among multiple robots can be achieved by employing a replanning avoidance control method based on envelope distance detection. When interference in the envelope space is detected, this method pauses commands and instructs one robot to retreat or move its chassis as a whole, then replans a collision-free trajectory for the other robot in the three-dimensional coordinate system. However, in physically coupled situations where multiple robots share the same rigid workpiece, this unilateral retreat or significant replanning can cause changes in the absolute coordinates of the end effector, leading to workpiece damage or slippage. Forcibly locking the end effector coordinates often makes it impossible to calculate a detour path in confined spaces, resulting in a deadlock among the multiple robots and reducing the reliability of collaborative construction.

[0036] The collaborative control method described in this application acquires joint state and pose data when multiple robots jointly grasp a rigid workpiece, dividing the links into end-locked links and non-end-floating links. When the distance between the non-end-floating links is lower than a preset safety threshold, the interference vector is calculated and decomposed into an absorbable component within the chain and a residual component outside the chain. Depending on whether the residual component outside the chain exceeds the threshold, the non-end-floating links are adjusted in stages to utilize the robot's redundant degrees of freedom to eliminate interference, or the dual-robot collaborative movement of the chassis and adjustment of the rigid workpiece's pose are calculated and controlled to reconstruct the overall formation, thus achieving interference elimination. This not only fully utilizes the robot's redundant degrees of freedom and reduces unnecessary chassis movement, but also provides reliable overall obstacle avoidance planning when the robot's degrees of freedom are exhausted.

[0037] As can be seen, the collaborative control method in this application embodiment can effectively avoid interference when multiple humanoid robots jointly grasp rigid workpieces in a confined space. It can also effectively solve the problems of workpiece slippage caused by blind replanning of trajectories or deadlock in extremely narrow spaces caused by forced locking of the end in related technologies. Thus, it realizes safe, continuous and reliable multi-robot collaborative construction operations in a space-constrained environment.

[0038] To facilitate understanding, the method provided in this implementation will be described in detail below, using the above scenario as an example. Please refer to [link / reference]. Figure 2 This is a flowchart illustrating a collaborative control method for humanoid robot construction in an embodiment of this application.

[0039] S201. In the state where the first robot and the second robot are jointly gripping the same rigid workpiece, acquire the joint state data of the first robot and the second robot and the pose data of the rigid workpiece.

[0040] Among them, the first robot and the second robot refer to two robots that jointly participate in collaborative construction tasks, such as humanoid robots, which include structures such as torso, arms, and legs. The limbs are connected to each other through multi-degree-of-freedom joints and can complete complex construction operations such as grasping, carrying, and installation. Rigid workpieces refer to building components with sufficient structural rigidity, whose deformation under external forces is negligible, such as precast concrete beams, prefabricated steel frame nodes, or wall panels. Joint state data refers to the real-time kinematic state information of all driveable joints of each robot. For humanoid robots, joint state data includes the current angle value (joint angle), angular velocity, and torque of each joint. Pose data refers to the complete state description of the rigid workpiece in three-dimensional space, including the workpiece's three-dimensional position coordinates (translational components) and spatial posture (rotational components, usually represented by quaternions, Euler angles, or rotation matrices), which together constitute the workpiece's six-degree-of-freedom pose, such as the coordinates of the workpiece's center of mass in the world coordinate system and the spatial orientation of the workpiece's major axis. Pose data is the core reference quantity for maintaining the consistency of multi-machine grasping constraints.

[0041] After initiating the multi-machine collaborative control task, the collaborative control device enters the real-time data acquisition phase, continuously providing real-time status input for subsequent link attribute classification, collision detection, and motion planning. Specifically, the collaborative control device reads joint status data fed back by sensors such as joint encoders or inertial measurement units of the two robots at fixed control cycles through communication connections established with the first and second robots. Simultaneously, it acquires the current pose data of the rigid workpiece in the world coordinate system in real time through vision positioning systems, motion capture systems, or force / torque sensors deployed on the workpiece surface at the construction site. The collaborative control device timestamps the above three sets of data (joint status data of the first robot, joint status data of the second robot, and workpiece pose data) to ensure the time synchronization of multi-source data, thereby guaranteeing the accuracy of subsequent kinematic calculations and collision predictions.

[0042] Optionally, in some embodiments, the collaborative control device can also, based on the robot kinematics model, calculate the pose of each link end in real time through forward kinematics according to the acquired joint state data, and perform cross-fusion verification with the workpiece pose data to eliminate the state estimation error caused by single sensor noise.

[0043] S202. Based on joint state data and pose data, the links between the first robot and the second robot are divided into end-locked links and non-end-floating links. The end-locked links are links that directly contact the rigid workpiece and perform gripping actions.

[0044] In robotic limbs, a link is a rigid rod structure between two adjacent joints. It is the basic structural unit constituting the robot's kinematic chain. For humanoid robots, links include upper arm links, forearm links, torso links, thigh links, and lower leg links. Driving joint rotation can change the position and orientation of the corresponding link in space. An end-capturing link is a link located at the end of the robot's kinematic chain that directly contacts a rigid workpiece and performs a grasping action. For humanoid robots, end-capturing links typically correspond to hand links (including wrist and finger structures). Because it forms direct physical contact with the workpiece, in a co-grasping state, the spatial position and orientation of the end-capturing link are strongly constrained by the current orientation of the workpiece. While maintaining the workpiece's orientation, the end-capturing link... The absolute spatial coordinates of a link cannot be changed unilaterally. Non-end-floating links refer to the limb links other than the end-locked links. For humanoid robots, non-end-floating links include upper arm links, forearm links, torso links, etc. Since the total number of degrees of freedom of a humanoid robot's joints usually exceeds the minimum number of degrees of freedom required to complete the end-space positioning task, non-end-floating links still have usable redundant motion degrees of freedom (i.e., self-motion space) while maintaining the pose of the end-locked links. They can change their spatial posture by adjusting the corresponding joint angles, providing usable configuration margin for interference avoidance. For example, while keeping both hands holding the workpiece absolutely still, the forearm can be displaced laterally by abducting the elbow joint to make room for the links of adjacent robots.

[0045] This step is performed after data acquisition in S201 and before interference detection. Its purpose is to clearly define the boundaries of "rigid constraints that cannot be changed" and the range of "freely adjustable ranges with self-motion margins" under physical gripping constraints. Specifically, the collaborative control device establishes a closed kinematic chain model of the multi-machine-workpiece based on the current joint state data of the first and second robots and their kinematic models. In the closed kinematic chain, the contact points between the end effectors of the two robots and the rigid workpiece are used as constraint endpoints. The final link in each robot's kinematic chain that performs the gripping function is identified and marked as an end-locked link; all other links are classified as non-end-floating links. After classification, the collaborative control device calculates the degree-of-freedom redundancy of each non-end-floating link's corresponding joint under the condition of maintaining the end-locked link's pose unchanged, i.e., the adjustable angle range of each joint in the current configuration, based on the current joint state data and by constructing the task Jacobian matrix of each robot regarding the end-locked link.

[0046] Optionally, in some embodiments, the cooperative control device can also prioritize the non-end floating links based on the topological distance (joint level) between the non-end floating links and the end locking links in the kinematic chain, and in subsequent joint adjustments, prioritize driving the links closer to the end (such as the forearm link) to avoid collisions, so as to reduce the disturbance to the overall robot configuration and improve the locality and smoothness of the collision avoidance action.

[0047] S203. When the distance between the non-end floating links of the first robot and the second robot is lower than the preset safety threshold, calculate the interference vector between the non-end floating links.

[0048] The preset safety threshold refers to the distance judgment benchmark value that triggers the interference avoidance control process. It is used to reserve sufficient response margin for the collaborative control equipment before an actual physical collision occurs. The setting of the preset safety threshold should comprehensively consider factors such as the dynamic response delay of the robot joint drive, the control cycle, the maximum possible movement speed of each link, and the spatial density of the construction environment. It is usually configured offline after determining a reasonable range through multi-scenario testing during the system integration and calibration phase to ensure that the system can still complete effective avoidance before the collision occurs under the most unfavorable relative motion conditions. The interference vector is a vector that describes the interference state between the non-end-of-arm floating links of two robots. It includes two elements: the direction component and the magnitude component. The direction component is the direction along the line connecting the nearest points of the two envelopes, pointing towards the direction that separates the two. The magnitude component is the minimum separation displacement required to restore the current link spacing to the safety threshold.

[0049] During the real-time monitoring of the spacing between the non-end-of-motion floating links of the two robots by the collaborative control device, once the spacing between any pair of links is detected to be lower than the preset safety threshold, the interference vector calculation process in S203 is triggered. Specifically, the collaborative control device constructs a virtual safety envelope outside each non-end-of-motion floating link (for each link of the humanoid robot, a geometric model such as a capsule or a directed bounding box can be used for approximation), traverses all pairwise combinations of non-end-of-motion floating links of the first and second robots, and uses the nearest point distance algorithm such as GJK (Gilbert–Johnson–Keerthi) to solve for the nearest point and its spacing between the envelopes of each link pair in real time; for link pairs whose spacing is lower than the preset safety threshold, based on the direction of the line connecting the nearest points of the two envelopes and the current penetration depth (or the predicted penetration depth based on the current relative motion speed), the minimum separation displacement required to restore the spacing of the link pair to the safety threshold is calculated, forming an interference vector; finally, the collaborative control device converts the interference vector into a task space representation in the world coordinate system and transmits it to S204 for subsequent decomposition processing.

[0050] Optionally, in some embodiments, the collaborative control device may also set an additional warning distance range before the distance falls below a preset safety threshold. When the distance between the links enters the warning range but has not yet reached the safety threshold, a low-amplitude preventive joint adjustment is initiated in advance, so that the link configuration transitions to a safer direction in advance, thereby avoiding large abrupt adjustment actions when the distance drops sharply, and improving the smoothness of the overall control.

[0051] S204. Under the condition of keeping the position of the rigid workpiece unchanged, the interference vector is decomposed into an absorbable component inside the chain and a residual component outside the chain; the absorbable component inside the chain is the amount of interference that can be eliminated by adjusting the joint angle of the non-end floating link while keeping the position of the rigid workpiece unchanged.

[0052] Among them, the absorbable component within the chain refers to the part of the interference vector that can be resolved by adjusting the joint angles of the non-end-floating links, under the strong constraint that the pose of the end-locked link remains absolutely unchanged. Its physical essence is the available self-motion displacement corresponding to the effective projection of the interference vector into the null space of the task Jacobian matrix corresponding to the current configuration of the robot. The residual component outside the chain refers to the remaining part of the interference vector that exceeds the self-motion range provided by the null space of all non-end-floating links and cannot be resolved by adjusting the body joint pose.

[0053] This step is performed after the interference vector is calculated in S203. Under the constraints of the closed-loop kinematic chain, the amount of interference that the robot body can autonomously resolve and its limits are quantitatively evaluated. Specifically, the cooperative control device constructs the task Jacobian matrix of each robot relative to the end-effector locking link based on the current joint state data of the first and second robots. and The null projection matrix corresponding to each Jacobian matrix is ​​calculated through singular value decomposition. and ,in The pseudo-inverse of the Jacobian matrix is ​​represented; the interference vector d obtained from S203 is transformed into the motion requirements of each robot joint space, and then projected through the null space matrix. and Projecting this onto the null space of each robot's current configuration, and within the constraints of redundancy in each joint's degrees of freedom, solve for the maximum displacement vector that can be generated along the direction of the interference vector. The component was identified as an in-chain absorbable component; subsequently, the vector difference between the original interference vector and the in-chain absorbable component was calculated. The residual components outside the chain are obtained.

[0054] During the above solution process, the collaborative control device simultaneously verifies whether the angle adjustment of each joint to be adjusted exceeds the mechanical limit range, and whether the current configuration is close to the kinematic singularity. If there is a risk of joint exceeding the limit or singularity, the absorbable component in the chain in the corresponding direction is truncated, and the truncated amount is included in the residual component outside the chain to ensure the physical executability of the absorbable component in the chain.

[0055] S205. If the residual component outside the chain does not exceed the preset residual threshold, adjust the corresponding non-end floating link according to the absorbable component inside the chain to eliminate interference.

[0056] Among them, the preset residual threshold is a threshold parameter used to determine whether the residual components outside the chain are negligible within the allowable range of engineering accuracy. The setting of the preset residual threshold needs to take into account factors such as the allowable end pose error tolerance of the workpiece gripping mechanism, the passive compliance range of the robot hand gripping mechanism, and the requirements of the construction task for the installation position accuracy. While maximizing the applicability of the body self-adjustment scheme, it is necessary to ensure that the residual interference will not have a substantial impact on the safety of the workpiece and the construction accuracy. It should be configured offline after comprehensively evaluating various constraints during the system integration and calibration stage.

[0057] When S204 calculates the residual components outside the chain and they do not exceed the preset residual threshold, the body joint adjustment execution stage of S205 is entered. At this time, the scenario is: the interference can be completely eliminated by the redundant degrees of freedom of the body while the position of the end-locked link remains unchanged, or although there are small residuals, they are acceptable within the engineering tolerance range, without the need to move the chassis, and the construction process can be maintained continuously.

[0058] Specifically, the collaborative control device calculates the absorbable components within the chain using S204, and weights them according to the null space dimension and joint redundancy of each robot's current configuration. This weighted pseudo-inverse method, combined with null space projection, is then used to convert these into incremental joint angle adjustment commands for the corresponding non-end-effector floating links of the first and second robots, causing the non-end-effector floating links of both robots to move away from each other. The collaborative control device synchronously sends adjustment commands to the joint drive controllers of both robots using a unified time base, driving the corresponding joints to gradually adjust according to a smooth velocity curve (such as trapezoidal velocity planning). This causes the interfering non-end-effector floating links to move away from each other (e.g., both robots extend their upper arms and raise their elbows, making room for each other), until the distance between all non-end-effector floating links of both robots returns to above the safe threshold, eliminating the interference. During joint adjustment, the collaborative control device continuously monitors the real-time pose changes of the end-effector locking links. If the end-effector pose deviation exceeds the allowable range, it is corrected in real-time via joint compensation commands, ensuring that the pose of the rigid workpiece remains stable and controlled throughout the entire joint adjustment process.

[0059] S206. If the residual component outside the chain exceeds the preset residual threshold, calculate the relative displacement between the chassis of the first robot and the second robot, and the corresponding safe adjustment posture of the rigid workpiece.

[0060] The relative displacement refers to the displacement change between the chassis of the first robot and the second robot during mutual adjustment in the work space, including parameters such as the translation direction, movement distance and deflection angle of each chassis; the safe adjustment posture refers to the target pose to which the rigid workpiece is allowed to transition in three-dimensional space after the dual-machine chassis coordinate movement and overall formation reconstruction are implemented. The safe adjustment posture must meet three conditions at the same time: the gripping constraints of the dual machines on the workpiece under the new formation are still effective, the new posture of the workpiece does not affect the final installation accuracy within the tolerance range specified by the construction task, and the spacing between all non-end floating links of the dual machines under the new formation has been restored to above the safety threshold.

[0061] When the residual components outside the chain calculated by S204 exceed the preset residual threshold, the relative displacement between the chassis of the first robot and the second robot, as well as the corresponding safe adjustment posture of the rigid workpiece, are calculated. At this point, the redundant degrees of freedom of the robot body have been exhausted and the remaining interference cannot be eliminated. The spatial deadlock must be broken by changing the relative position of the two chassis.

[0062] Specifically, the collaborative control equipment takes the overall closed system consisting of the first robot, the second robot, and the rigid workpiece as the unified planning object, and uses the residual interference direction and magnitude indicated by the residual components outside the chain as the optimization driving target. Under the following constraints, it solves for the relative displacement of the dual-chassis and the safe adjustment posture of the rigid workpiece:

[0063] Firstly, the new overall configuration of the two engines after the chassis is moved should restore the spacing between all non-end floating linkage pairs to above the safety threshold.

[0064] Secondly, the movement trajectory of the dual-chassis should not collide with fixed components or other obstacles at the construction site;

[0065] Third, the safe adjustment posture of the workpiece should be within the position tolerance range specified in the construction task, and the end locking linkage of the dual machines under the new configuration should still be able to effectively grip the workpiece.

[0066] The aforementioned multi-constraint optimization problem can be solved using numerical methods such as quadratic programming or gradient projection, and the calculated relative displacement and safe adjustment attitude are passed to S207 for execution.

[0067] Optionally, in some embodiments, if the residual component outside the chain exceeds a preset residual threshold, the corresponding non-end floating link is adjusted according to the absorbable component inside the chain, and the relative displacement between the chassis of the first robot and the second robot, as well as the safe adjustment posture of the corresponding rigid workpiece, are calculated.

[0068] S207. Based on the relative displacement and safety adjustment posture, control the first robot and the second robot to move the chassis together and adjust the position of the rigid workpiece to eliminate interference.

[0069] The chassis refers to the mobile base structure of a humanoid robot. For humanoid robots, the chassis can be a wheeled drive chassis or a legged walking mechanism, etc.

[0070] After planning the relative displacement and safe posture adjustment in S206, the execution phase begins in S207, where spatial deadlock is broken through the coordinated reconstruction of the overall formation. Specifically, based on the relative displacement planned in S206, the collaborative control equipment generates chassis motion trajectories for both the first and second robots, and synchronously issues chassis motion commands to both robots using a unified time base. During the coordinated movement of the two chassis, the collaborative control equipment synchronously generates real-time follow-up compensation commands for each joint of the upper limbs of both robots. Through real-time inverse kinematics solving, it ensures that the position of the end-effector locking link relative to the rigid workpiece is always maintained within the constraint range. Simultaneously, the collaborative control equipment uses admittance control or impedance control strategies, based on real-time force / torque sensing data of the end-effector locking link, to... Feedback is provided to monitor the interaction force exerted on the workpiece by the dual-machine end effectors in real time. When the internal force exceeds the preset safety threshold, the collaborative control equipment dynamically compensates for the internal force fluctuations caused by synchronization errors during the chassis movement by finely adjusting the moving speed of the dual-machine chassis or the compensation angle of the arm joints, ensuring that the resultant force on the workpiece is always controlled within the safe range allowed by the structure. As the dual-machine chassis gradually completes the displacement, the workpiece position smoothly transitions to a safe adjustment posture. After the spacing between all non-end floating links of the dual machines returns to above the safety threshold, the collaborative control equipment confirms that the interference has been eliminated and resumes the execution of the original construction procedure.

[0071] In this embodiment, by employing a hierarchical strategy that decomposes the interference vector into an absorbable component within the chain and a residual component outside the chain while maintaining the pose of the rigid workpiece, and by adjusting the non-end-floating link or coordinating the chassis movement and adjusting the pose of the rigid workpiece based on whether the residual component outside the chain exceeds a preset residual threshold, the redundant degrees of freedom of the robot body can be used to absorb the interference first, and the chassis can be reconstructed only when necessary. This effectively solves the problems of workpiece slippage caused by blindly replanning the trajectory or deadlock in extremely narrow spaces caused by forcibly locking the end-floating link when multiple robots are holding a rigid workpiece in related technologies. Thus, it enables safe, continuous, and reliable collaborative construction operations of multiple humanoid robots in a confined space.

[0072] In the above embodiment, by decomposing the interference vector into an absorbable component within the chain and a residual component outside the chain, and adopting a hierarchical avoidance strategy accordingly, the interference avoidance and deadlock problems when multiple robots share a workpiece in a confined space are initially solved. In practical applications, construction sites often have a dense distribution of obstacles, resulting in extremely limited actual movable space for the robot chassis. The relative displacement of the basic chassis alone may not be enough to eliminate all interference without environmental collisions. Furthermore, after completing the avoidance action, it is also necessary to ensure that the system can smoothly connect to subsequent construction procedures.

[0073] Based on the above embodiments, the method provided in this embodiment will be described in further detail below. Please refer to... Figure 3 This is another flowchart illustrating a collaborative control method for humanoid robot construction in an embodiment of this application.

[0074] S301. When the first robot and the second robot are jointly gripping the same rigid workpiece, acquire the joint state data of the first robot and the second robot and the pose data of the rigid workpiece.

[0075] Step S301 is similar to step S201 in the above embodiment, and will not be repeated here.

[0076] S302. Based on joint state data and pose data, the links between the first robot and the second robot are divided into end-locked links and non-end-floating links. The end-locked links are links that directly contact the rigid workpiece and perform gripping actions.

[0077] This step specifically includes:

[0078] Based on the joint state data, the kinematic chain topology of the first robot and the second robot is determined.

[0079] Based on the pose data of the rigid workpiece, the links in the first and second robots that form a fixed constraint relationship with the rigid workpiece are identified and determined to be end-locking links.

[0080] All intermediate links in the kinematic chain topology from the chassis to the end-locking link are identified as non-end-floating links.

[0081] The kinematic chain topology refers to the hierarchical relationship and sequence structure of the connections between the links and joints of the robot. It describes the physical connection path from the chassis base to the end effector and the upstream and downstream relationships of each node. For humanoid robots, the kinematic chain topology includes multiple parallel branch chains such as the torso chain (waist → chest), the arm chain (shoulder → upper arm → forearm → wrist → hand), and the leg chain (hip → thigh → lower leg → ankle → foot). Each branch chain is composed of several rigid links and drive joints connected in series. The fixed constraint relationship refers to the rigid pose binding state formed between the robot's end effector and the rigid workpiece. It indicates that a specific link of the robot has established physical contact with the workpiece through the grasping action, and the relative pose between the two remains fixed in the grasping state. The intermediate links refer to all the links located between the chassis and the end locking link in the kinematic chain topology. For humanoid robots, the intermediate links usually include waist links, chest links, upper arm links, forearm links, etc.

[0082] S302 executes immediately after acquiring joint state and pose data, and clarifies the boundary between "constrained and unable to move freely" and "having adjustment margin and being actively controllable" under physical gripping coupling constraints. Specifically, the collaborative control device first calculates the complete kinematic topology of each robot from the chassis coordinate system to the end effector through forward kinematics, based on the kinematic models (such as URDF description files or equivalent parameterized models) of the first and second robots and the current joint state data. This is done by combining these kinematic models with the current joint state data. The topology is recorded in a directed graph format, showing the drive axis type and current pose transformation of each link node and its corresponding joint. Then, using the current pose data of the rigid workpiece as a reference, the collaborative control device verifies the connection in the first and second robots through inverse kinematics. It identifies links in the first and second robots whose end effector pose matches the workpiece gripping point pose. The validity of the fixed constraint relationship is confirmed by combining the end effector force / torque sensor data, and the link is marked as an end effector locked link. Finally, starting from the chassis node, the collaborative control device performs a topology traversal along each kinematic branch, marking all intermediate links along the path from the chassis to the end effector locked link as non-end effect floating links, thus completing the full classification of link attributes.

[0083] Optionally, in some embodiments, after completing the link attribute division, the collaborative control device can also assign a priority weight to each non-end floating link based on the topological distance (joint level) of each non-end floating link from the end locking link in the kinematic chain. In subsequent joint adjustments, the link closer to the end (such as the forearm link) will be driven first to perform avoidance, so as to localize the impact range of the collision avoidance action and reduce the disturbance to the stability of the overall mechanism.

[0084] S303. When the distance between the non-end floating links of the first robot and the second robot is lower than the preset safety threshold, calculate the interference vector between the non-end floating links.

[0085] Step S303 is similar to step S203 in the above embodiment, and will not be repeated here.

[0086] Optionally, in some embodiments, the collaborative control device can also acquire the current joint torque data of the first robot and the second robot, the mass parameters of the rigid workpiece, and then calculate the deformation compensation vector of the non-end floating link under load based on the joint torque data and mass parameters, and correct the interference vector according to the deformation compensation vector.

[0087] S304. While keeping the pose of the rigid workpiece unchanged, decompose the interference vector into an absorbable component within the chain and a residual component outside the chain.

[0088] This step specifically includes:

[0089] Based on joint state data and pose data, the degree-of-freedom redundancy of each joint in the non-end-of-life floating link is calculated. The degree-of-freedom redundancy refers to the range of angles that the corresponding joint can be adjusted while keeping the pose of the end-of-life locking link unchanged.

[0090] Based on the degree of freedom redundancy, with the interference vector as the target direction, the maximum displacement vector that can be achieved by adjusting the joint angle of the non-end floating link is calculated and determined as the absorbable component within the chain.

[0091] The difference vector between the interference vector and the absorbable component within the chain is taken as the residual component outside the chain.

[0092] Among them, the degree of freedom redundancy refers to the remaining adjustable angle range of each joint in the non-end-of-line floating link under the constraint of keeping the end-of-line locking link's pose unchanged. It is a quantitative indicator of the available motion margin of the joint for interference avoidance under the current configuration. The maximum displacement vector refers to the maximum feasible avoidance displacement vector that can be achieved in zero space by combining joint adjustments with the direction of the interference vector as the target direction, by comprehensively utilizing the degree of freedom redundancy of all joints in the non-end-of-line floating link. The difference vector refers to the vector difference between the interference vector and the absorbable component in the chain. It quantitatively characterizes the degree of inadequacy of the body's self-motion adjustment capability, that is, the remaining interference that must be resolved by external means (such as chassis movement).

[0093] S304 is executed immediately after the interference vector is calculated in S303. By performing null space decomposition on the interference vector under closed-loop constraints, the interference limit that can be eliminated by the self-motion of the body is quantified. Specifically, based on the kinematic chain topology and current joint state data determined by S302, the collaborative control device constructs task Jacobian matrices for the first and second robots, respectively, from each non-end-floating link joint to the end-locking link. It then solves for the null-space basis vectors of each Jacobian matrix using singular value decomposition (SVD), quantifying the available adjustment range of each joint under the condition that the end-locking link pose remains unchanged, thus obtaining the degree-of-freedom redundancy of each joint. Subsequently, the collaborative control device maps the interference vector to the null space of each robot's current configuration. Using the direction of the interference vector as the target, it solves for the maximum displacement vector that can be generated along the direction of the interference vector within the constraints of the degree-of-freedom redundancy of each joint using weighted least squares optimization, identifying it as the absorbable component within the chain. During the solution process, if the joint adjustment approaches the physical angle limit or the Jacobian matrix approaches singularity, the absorbable component in the corresponding direction is truncated to the boundary to ensure that the absorbable component within the chain is physically executable. Finally, the collaborative control device calculates the difference vector between the interference vector and the absorbable component within the chain, outputting it as the residual component outside the chain, thus completing the vector decomposition.

[0094] S305. If the residual component outside the chain does not exceed the preset residual threshold, adjust the corresponding non-end floating link according to the absorbable component inside the chain to eliminate interference.

[0095] Step S305 is similar to step S205 in the above embodiment, and will not be repeated here.

[0096] S306. If the residual component outside the chain exceeds the preset residual threshold, the constraint condition that the rigid workpiece pose remains unchanged is released.

[0097] When the residual components outside the chain calculated in S304 exceed the preset residual threshold, it indicates that the redundant degrees of freedom of the body within the current closed-loop kinematic chain can no longer eliminate the remaining interference. If the rigid workpiece's pose is still forcibly constrained to remain unchanged, the available solution space will become unsolvable due to the dual constraints of end-point locking and spatial geometric limitations. Therefore, the cooperative control device executes S306 to remove the constraint on the rigid workpiece's pose remaining unchanged, allowing subsequent planning steps to make limited adjustments to the workpiece's spatial pose within the workpiece's pose tolerance range, thus expanding the available collision avoidance solution space.

[0098] S307. Obtain spatial constraint information of the current construction scene. The spatial constraint information includes the movable area of ​​the chassis and the distribution of obstacles.

[0099] Among them, spatial constraint information refers to a comprehensive dataset describing all spatial constraints related to chassis motion planning in the current construction site, which serves as the environmental constraint input for subsequent chassis motion planning; the movable area of ​​the chassis refers to the effective range within the construction site that allows the robot chassis to move, and its boundaries are usually determined by fixed physical boundaries such as walls, installed components, ground openings, and work fences; obstacle distribution refers to the spatial location and geometric information of fixed or temporary obstacles in the construction site that may affect the chassis motion path, including installed building structures, scaffolding, construction equipment, and other robot bodies, such as door frames, wall corners, and other structural components, which should all be included as obstacles in the planning constraints. Obstacle distribution provides collision avoidance constraints for chassis motion path planning, preventing new collisions with the environment during the chassis's cooperative movement.

[0100] After the workpiece pose constraints are released by S306, the collaborative control equipment needs to obtain complete spatial constraint information of the construction scene before executing chassis motion planning to ensure that the subsequently planned chassis path is physically feasible. Specifically, the collaborative control equipment retrieves the geometric information of fixed boundaries and obstacles in the current construction area from the 3D building information model (BIM) or scene point cloud map of the construction site as a static constraint layer. At the same time, it integrates the scanning data of real-time perception sensors such as LiDAR and depth camera on the robot body to identify and supplement the real-time distribution information of temporary obstacles (such as newly added components in the construction progress, tool stacking, etc.) in the scene at the current moment as a dynamic constraint layer. The collaborative control equipment performs data fusion of the static constraint layer and the dynamic constraint layer to construct a 3D occupancy grid map of the construction scene at the current moment, and extracts the boundary polygons of the movable areas of the chassis of the first and second robots and the distribution of obstacles inside them as environmental constraint inputs for calculating the maximum movable distance of the chassis and optimizing the relative displacement.

[0101] S308. Based on spatial constraint information, calculate the maximum movable distance of the first robot and the chassis of the second robot in each direction of movement.

[0102] Among them, the maximum movable distance refers to the maximum collision-free straight distance that the robot chassis can travel in a specific direction from its current position under the constraints of the movable area of ​​the chassis and the distribution of obstacles. It is a quantitative indicator for evaluating the feasibility of each movement direction and the margin of the chassis's movement space. For example, in the forward direction, the maximum movable distance is the net distance between the current position of the chassis and the boundary of the nearest obstacle in that direction.

[0103] S308 is executed after acquiring spatial constraint information. Its purpose is to comprehensively quantify the distribution of the degrees of freedom of motion of the current dual-machine chassis in all directions, providing a quantitative basis for determining whether to introduce workpiece posture tolerance extended feasible solutions (S309) and calculating the optimal relative displacement (S310). Specifically, the cooperative control device takes the current chassis positions of the first and second robots as the starting point and performs uniform discretization sampling of the movement directions in the horizontal plane. For each sampling direction, a ray is projected forward from the current position of the chassis of the two robots along that direction, and the first intersection distance between the ray and the obstacle object or movable area boundary in the three-dimensional occupied grid map is detected to obtain the maximum movable distance of each robot in that direction. The cooperative control device takes the smaller value of the maximum movable distance of the two robots in the same direction as the maximum feasible distance of the dual-machine cooperative movement in that direction. After traversing all sampling directions, the cooperative control device summarizes the maximum movable distances in each direction to form a chassis motion feasibility distribution map.

[0104] S309. When the maximum movable distance in all directions of movement is less than the modulus of the residual component outside the chain, adjust the residual component outside the chain according to the preset attitude tolerance range of the rigid workpiece.

[0105] The preset attitude tolerance range refers to the maximum allowable positional deviation of a rigid workpiece according to the construction specifications, including positional tolerances in three translational directions and attitude tolerances in three rotational directions. The preset attitude tolerance range is usually set as follows, based on building construction specifications (such as the "Code for Acceptance of Construction Quality of Concrete Structures" GB50204) and the dimensional characteristics of the rigid workpiece: positional tolerance ±5mm to ±15mm, attitude tolerance ±1° to ±3°. The specific values ​​should be determined according to the accuracy specifications of the construction task and the structural connection method. Its core function is to reduce the residual components outside the chain that originally exceeded the chassis movement limit to within the feasible range of the chassis by allowing the workpiece to undergo limited positional deviations within the tolerance range when the chassis movement capability is limited, thereby achieving a balance between workpiece installation compliance and collision avoidance feasibility.

[0106] When S308 detects that the maximum movable distance of the dual-machine chassis in all sampling directions is less than the modulus of the residual component outside the chain, it indicates that the spatial movement provided by the chassis moving to its physical limit in any direction is still insufficient to completely eliminate the remaining interference. At this point, using the original residual component outside the chain as the planning target will not be feasible. S309 is needed to adjust the residual component outside the chain by utilizing the workpiece attitude tolerance space. Specifically, the collaborative control device uses the preset attitude tolerance range as the constraint boundary. Within the workpiece's allowable pose adjustment space, it optimizes by minimizing the required movement distance of the chassis and remaps the residual component outside the chain: it calculates the interference elimination amount that the workpiece can "share" for the chassis movement by making limited adjustments within the attitude tolerance range, subtracts this shared amount from the original residual component outside the chain, and obtains the corrected residual component outside the chain, reducing the corresponding chassis required movement distance to within the feasible range of the chassis. If the chassis required movement distance corresponding to the adjusted residual component outside the chain still exceeds the maximum movable distance in all directions, the collaborative control device should output a construction deadlock alarm, prompting manual intervention to replan the construction sequence layout.

[0107] S310. Based on spatial constraint information and residual components outside the chain, calculate the relative displacement between the chassis of the first robot and the chassis of the second robot.

[0108] After S308 and S309 complete the feasibility analysis of chassis motion and the correction of residual components outside the chain, the collaborative control equipment calculates the relative displacement between the chassis of the first robot and the second robot through multi-constraint optimization based on spatial constraint information and the corrected residual components outside the chain. Specifically, the collaborative control device uses the residual interference direction and magnitude indicated by the residual components outside the chain as the optimization driving target, the maximum movable distance in each direction as the upper limit of the constraint, and the obstacle distribution as the collision avoidance constraint to construct a chassis relative displacement optimization problem: restoring the spacing between all non-end floating link pairs to above the safety threshold is the hard constraint, and minimizing the total moving distance of the dual chassis is the optimization objective. In the feasible solution space, the displacement vectors (including translation direction, moving distance, and deflection angle) of the chassis of the first and second robots are solved, and the difference between the displacement vectors of the two robots is the relative displacement. During the solution process, the collaborative control device also needs to perform inverse kinematics verification on the reachability of the arm joints of the two robots at the target chassis position to ensure that the end locking links of the two robots are still within their respective arm working spaces after the chassis is in place. Only feasible solutions that simultaneously satisfy the environmental collision-free and joint reachability constraints are accepted.

[0109] S311. Based on the relative displacement, calculate the safe adjustment posture of the rigid workpiece that matches the relative displacement.

[0110] After calculating the relative displacement, the collaborative control device uses forward kinematics to calculate the target pose of the end-locking links of the two robots in the world coordinate system after the chassis moves to the target position. Based on the gripping constraint equations of the end-locking links of the two robots on the rigid workpiece, the target pose of the rigid workpiece that can simultaneously satisfy the new pose constraints of the end-locking links of the two robots is solved. The target pose is compared with the current pose of the workpiece, and after confirming that its position deviation and attitude deviation are both within the preset attitude tolerance range, it is determined as the safe adjustment posture.

[0111] Optionally, in some embodiments, after determining the safe adjustment posture, the collaborative control device can also generate a continuous time trajectory sequence for the transition of the workpiece from the current posture to the safe adjustment posture by using B-spline or cubic polynomial interpolation, and coordinate and synchronize this time trajectory with the dual-machine chassis movement trajectory on the time axis to ensure that the change of workpiece posture is always kinematically consistent with the actual movement of the dual-machine end locking linkage throughout the chassis movement, thereby eliminating the risk of internal force impact in the closed-loop motion chain caused by the asynchronous timing of chassis movement and workpiece posture adjustment.

[0112] S312. Based on the relative displacement and safety adjustment posture, control the first robot and the second robot to move the chassis together and adjust the position of the rigid workpiece to eliminate interference.

[0113] Step S312 is similar to step S207 in the above embodiment, and will not be repeated here.

[0114] S313. After the current construction action is completed, control the first robot and the second robot to synchronously reset with the rigid workpiece.

[0115] Among them, reset refers to the process in which, after the current construction action is completed, the collaborative control equipment controls the first robot and the second robot to carry the rigid workpiece from the intermediate position under the safe adjustment posture and synchronously return to the standard position predetermined by the construction task or the initial position required by the next construction procedure. Synchronous reset refers to the two robots coordinating and synchronously performing the reset action under a unified time reference, and maintaining the gripping constraint of the rigid workpiece by the two robots without being broken throughout the entire reset process.

[0116] After the current construction action is completed, if the positions of the dual-machine chassis and the workpiece posture are coordinated and adjusted in S306 to S312, the current actual position of the dual machines may deviate from the standard position predetermined by the construction task or not meet the initial conditions of the next construction procedure. Therefore, S313 needs to be executed for synchronous reset. Specifically, after confirming that the current construction action has been completed and the workpiece installation status is stable, the collaborative control equipment plans the collaborative motion trajectory from the current configuration to the target reset configuration for the first and second robots, based on the predefined reset target in the construction task plan (including the target position of the dual-machine chassis and the target pose of the workpiece). This includes the chassis movement trajectory and the arm joint follow-up trajectory. The collaborative control equipment synchronously issues reset commands to the two robots with a unified time reference, driving them to move synchronously along the planned trajectory while carrying the rigid workpiece. During the reset execution, the collaborative control equipment continuously monitors the gripping constraint status between the dual-machine end locking linkage and the workpiece, as well as the workpiece pose deviation, and maintains the smooth transition of the workpiece through real-time compensation commands. Once the dual-machine chassis and the workpiece pose have both reached the target reset state and stabilized, the collaborative control equipment outputs a reset completion confirmation signal, allowing the construction task scheduling system to trigger the execution of the next construction procedure.

[0117] In this embodiment, the constraint condition of releasing the rigid workpiece's unchanged pose when the body's degrees of freedom are exhausted is adopted. Combined with spatial constraint information including the movable area of ​​the chassis and the distribution of obstacles, as well as the preset attitude tolerance range, the residual components outside the chain are evaluated and adjusted. Finally, the relative displacement and safe adjustment attitude are calculated to perform collaborative chassis reconstruction and synchronous reset after the action is completed. Therefore, the avoidance space can be maximized while ensuring compliance with working conditions and environmental safety. This effectively solves the problem of task deadlock and stagnation caused by the failure of local trajectory replanning in extremely restricted construction scenarios in related technologies. Thus, global safe collision avoidance and seamless connection between processes are achieved in continuous construction operations under complex obstacle distribution.

[0118] The collaborative control device in the embodiments of this application is described below from the perspective of hardware processing. Please refer to [link / reference]. Figure 4 This is a schematic diagram of the physical device structure of the collaborative control equipment in the embodiments of this application.

[0119] It should be noted that, Figure 4 The structure of the collaborative control device shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of this application.

[0120] like Figure 4As shown, the cooperative control device includes a CPU 401, which can perform various appropriate actions and processes based on a program stored in the read-only memory ROM 402 or a program loaded from the storage section 408 into the random access memory RAM 403, such as performing the methods described in the above embodiments. The RAM 403 also stores various programs and data required for system operation. The CPU 401, ROM 402, and RAM 403 are interconnected via a bus 404. An I / O interface 405 is also connected to the bus 404.

[0121] The following components are connected to I / O interface 405: input section 406 including audio input devices, push-button switches, etc.; output section 407 including a liquid crystal display (LCD) and audio output devices, indicator lights, etc.; storage section 408 including a hard disk, etc.; and communication section 409 including a network interface card such as a LAN (Local Area Network) card, modem, etc. Communication section 409 performs communication processing via a network such as the Internet. Drive 410 is also connected to I / O interface 405 as needed. Removable media 411, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., are installed on drive 410 as needed so that computer programs read from them can be installed into storage section 408 as needed.

[0122] Specifically, according to embodiments of this application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of this application include a computer program product comprising a computer program / instructions carried on a computer-readable medium, the computer program / instructions containing computer program / instructions for performing the methods shown in the flowcharts. In such embodiments, the computer program / instructions can be downloaded and installed from a network via communication section 409, and / or installed from removable medium 411. When the computer program is executed by CPU 401, it performs the various functions defined in this application.

[0123] It should be noted that specific examples of computer-readable storage media may include, but are not limited to: electrical connections having one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), flash memory, optical fiber, portable compact disc read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing. In this application, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0124] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. Each block in a flowchart or block diagram may represent a module, segment, or portion of code, which contains one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those shown in the drawings.

[0125] Specifically, the collaborative control device in this embodiment includes a processor and a memory. The memory stores a computer program, and when the computer program is executed by the processor, it implements the collaborative control method for humanoid robot construction provided in the above embodiment.

[0126] In another aspect, this application also provides a computer-readable storage medium, which may be included in the collaborative control device described in the above embodiments; or it may exist independently and not assembled into the collaborative control device. The storage medium carries one or more computer programs, which, when executed by a processor of the collaborative control device, cause the collaborative control device to implement the collaborative control method for humanoid robot construction provided in the above embodiments.

[0127] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit it. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application.

[0128] As used in the above embodiments, depending on the context, the term "when..." can be interpreted as meaning "if...", "after...", "in response to determining...", or "in response to detecting...". Similarly, depending on the context, the phrase "when determining..." or "if (the stated condition or event) is interpreted as meaning "if determining...", "in response to determining...", "when (the stated condition or event) is detected", or "in response to detecting (the stated condition or event)".

[0129] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. This program can be stored in a computer-readable storage medium, and when executed, it can include the processes described in the above method embodiments. The aforementioned storage medium includes various media capable of storing program code, such as ROM or random access memory (RAM), magnetic disks, or optical disks.

Claims

1. A method for collaborative control of human-shaped robot construction, applied to a collaborative control device, the collaborative control device being in communication connection with a first robot and a second robot respectively, and characterized in that, The method includes: When the first robot and the second robot are jointly gripping the same rigid workpiece, acquire the joint state data of the first robot and the second robot and the pose data of the rigid workpiece; Based on the joint state data and the pose data, the link between the first robot and the second robot is divided into end-locked link and non-end-floating link. The end-locked link is the link that directly contacts the rigid workpiece and performs a gripping action. When the distance between the non-end floating links of the first robot and the second robot is lower than a preset safety threshold, the interference vector between the non-end floating links is calculated. While keeping the pose of the rigid workpiece unchanged, the interference vector is decomposed into an absorbable component within the chain and a residual component outside the chain; the absorbable component within the chain is the amount of interference that can be eliminated by adjusting the joint angle of the non-end floating link while keeping the pose of the rigid workpiece unchanged. If the residual component outside the chain does not exceed the preset residual threshold, the corresponding non-end floating link is adjusted according to the absorbable component inside the chain to eliminate interference. If the residual component outside the chain exceeds the preset residual threshold, then calculate the relative displacement between the chassis of the first robot and the second robot, and the corresponding safe adjustment posture of the rigid workpiece. Based on the relative displacement and the safety adjustment posture, the first robot and the second robot are controlled to move the chassis together and adjust the position of the rigid workpiece to eliminate interference.

2. The method of claim 1, wherein, Based on the joint state data and the pose data, the linkage between the first robot and the second robot is divided into end-capped locking linkages and non-end-capped floating linkages, specifically including: Based on the joint state data, the kinematic chain topology of the first robot and the second robot is determined respectively; Based on the pose data of the rigid workpiece, the links in the first robot and the second robot that form a fixed constraint relationship with the rigid workpiece are identified and determined to be end-locking links; All intermediate links in the kinematic chain topology from the chassis to the end locking link are identified as non-end floating links.

3. The method according to claim 1 or 2, characterized in that, The calculation of the relative displacement between the chassis of the first robot and the second robot, and the corresponding safe adjustment posture of the rigid workpiece, specifically includes: Release the constraint condition that the rigid workpiece's pose remains unchanged; Obtain spatial constraint information of the current construction scene, including the movable area of ​​the chassis and the distribution of obstacles; Based on the spatial constraint information and the residual components outside the chain, the relative displacement between the chassis of the first robot and the second robot is calculated. Based on the relative displacement, calculate the safe adjustment posture of the rigid workpiece that matches the relative displacement.

4. The method of claim 3, wherein, After controlling the first robot and the second robot to collaboratively move the chassis and adjust the pose of the rigid workpiece based on the relative displacement and the safety adjustment posture, the method further includes: After the current construction action is completed, control the first robot and the second robot to synchronously reset the rigid workpiece.

5. The method of claim 3, wherein, Before calculating the relative displacement between the chassis of the first robot and the second robot based on the spatial constraint information and the off-chain residual components, the method further includes: Based on the spatial constraint information, calculate the maximum movable distance of the first robot and the second robot chassis in each direction of movement; When the maximum movable distance in all directions is less than the modulus of the residual component outside the chain, the residual component outside the chain is adjusted according to the preset attitude tolerance range of the rigid workpiece.

6. The method of claim 1, wherein, The process of decomposing the interference vector into an in-chain absorbable component and an out-of-chain residual component specifically includes: Based on the joint state data and the pose data, the degree of freedom redundancy of each joint in the non-end floating link is calculated. The degree of freedom redundancy refers to the range of angles that the corresponding joint can be adjusted while keeping the pose of the end locking link unchanged. Based on the aforementioned degree of freedom redundancy, and with the interference vector as the target direction, the maximum displacement vector that can be achieved by adjusting the joint angle of the non-end floating link is calculated and determined as the absorbable component within the chain. The difference vector between the interference vector and the absorbable component within the chain is taken as the residual component outside the chain.

7. The method according to claim 1, characterized in that, After calculating the interference vector between the non-end floating links, the method further includes: Obtain the current joint torque data of the first robot and the second robot, and the mass parameters of the rigid workpiece; Based on the joint torque data and the mass parameters, the deformation compensation vector of the non-end floating link under load is calculated, and the interference vector is corrected according to the deformation compensation vector.

8. A cooperative control apparatus characterized by comprising: The cooperative control device includes: one or more processors and a memory; the memory is coupled to the one or more processors, the memory is used to store computer program code, the computer program code including computer instructions, and the one or more processors call the computer instructions to cause the cooperative control device to perform the method as described in any one of claims 1-7.

9. A computer readable storage medium storing computer instructions, characterized in that, When the computer instructions are executed on the cooperative control device, the cooperative control device causes the cooperative control device to perform the method as described in any one of claims 1-7.

10. A computer program product comprising computer programs or instructions, characterized in that, When the computer program or instructions are run on the cooperative control device, the cooperative control device causes the cooperative control device to perform the method as described in any one of claims 1-7.

Images