Construction robot multi-machine cooperation and man-machine hybrid construction method and system thereof

Abstract

The application relates to the technical field of robot control, and discloses a building robot multi-machine cooperation and man-machine hybrid construction method and system, wherein the building robot multi-machine cooperation and man-machine hybrid construction method comprises the following steps: step S1, calculating an initial end distance; step S2, calculating an end distance increment; step S3, generating a deflection sequence; step S4, calculating an axial internal force estimation; step S5, generating an updated grabbing point target; step S6, calculating a support force; step S7, constructing a deterministic intervention condition; and step S8, driving a operation arm to move and release a component. By constructing a deflection-induced axial geometric locking index, the application realizes the quantification of component bending deformation to axial implicit internal force, determines an end distance correction amount and reasonably distributes the end distance correction amount according to an operation arm stroke, so that internal force accumulation with construction advancement is avoided; the geometric precision of component construction and assembly is ensured; and clamping damage of the component caused by excessive internal force is reduced.

Description

Technical Field

[0001] This invention relates to the field of robot control technology, and more specifically, to a method and system for multi-robot collaborative and human-robot hybrid construction. Background Technology

[0002] In industrialized construction, the demand for installing prefabricated slender components is increasing. Narrow corridors and other confined spaces are the primary construction scenarios for these components. Manual handling is difficult and unsafe, making multi-machine collaborative gripping with mobile manipulators the mainstream construction method. To ensure the stability of component gripping, the industry commonly uses a closed-chain collaborative gripping method with two mobile manipulators at both ends of the component, forming a closed kinematic constraint chain between the robot and the component. While this gripping method can limit the disordered movement of the component, it also creates mutual constraints on the movements of the two ends of the component, potentially leading to the generation of axial internal forces later.

[0003] Slender components are prone to vertical bending deformation due to their own weight during construction. Based on the geometric nonlinear strain characteristics, the deflection caused by bending can trigger an axial geometric locking effect, resulting in equivalent geometric elongation of the component. However, under closed-chain constraints, the change in component end distance is limited by the movement of the manipulator and space, making it impossible to precisely match the geometric elongation. This mismatch translates into a hidden axial internal force, which cannot be detected by conventional visual monitoring. Passive detection methods using force sensors are insufficient to capture this force in a timely manner during construction. The internal force accumulates continuously as the component advances, leading to component attitude drift and affecting the accuracy of construction progress.

[0004] In existing construction methods, human intervention in human-machine hybrid construction is mostly based on experience, lacking clear standards for support positions and forces. The coordination between manual actions and robot end-distance adjustments is poor, failing to specifically address component bending deformation and alleviate geometric locking. Furthermore, construction in narrow spaces lacks path-specific division, and single geometric and mechanical calculations cannot accommodate continuous changes in component posture along the path, further exacerbating internal force accumulation and posture deviations. This ultimately leads to insufficient assembly geometric accuracy after component placement, and in some scenarios, excessive internal forces can cause component clamping damage. Current technologies have yet to provide an effective solution. Summary of the Invention

[0005] This invention provides a method and system for multi-robot collaborative construction and human-robot hybrid construction, which solves the technical problems mentioned in the background.

[0006] This invention provides a method for multi-robot collaborative and human-robot hybrid construction, comprising the following steps:

[0007] Step S1: Control the two moving manipulators to clamp the two ends of the component in a closed-chain cooperative gripping manner, and establish the initial end distance and discrete coordinates of the component axis;

[0008] Step S2: Obtain the current gripping point distance between the two moving manipulators, and calculate the end distance increment based on the initial end distance.

[0009] Step S3: Obtain the coordinates of discrete points on the center line of the component based on the discrete coordinates of the component's axial direction and construct a reference line. Extract the displacement of the discrete point coordinates relative to the reference line along the direction of gravity to generate a deflection sequence.

[0010] Step S4: Calculate the deflection-induced axial geometric locking index through the deflection sequence, convert it into geometric elongation, calculate the difference between the end distance increment and the geometric elongation, and convert it into an axial internal force estimate.

[0011] Step S5: Calculate the end distance correction amount based on the axial internal force estimate, allocate the end distance correction amount according to the ratio of the available axial stroke of the two moving manipulators, and generate the updated gripping point target.

[0012] Step S6: Find the extreme point of deflection in the deflection sequence as the support location, and calculate the support force corresponding to the support location based on the stiffness parameters of the component.

[0013] Step S7: Discretize the construction path into multiple fixed work positions. At each fixed work position, execute steps S2 to S6 in sequence. Use the grab point target to drive the two mobile manipulators to advance, and use the support position and support force as deterministic intervention conditions for human-machine hybrid construction.

[0014] Step S8: After advancing to the installation position, generate the final end distance correction amount according to the final assembly end distance, drive the two moving operating arms to move and release the component.

[0015] This invention provides a multi-robot collaborative and human-robot hybrid construction system, comprising:

[0016] The first module controls two mobile manipulators to grip the two ends of the component in a closed-chain collaborative gripping manner, and establishes the initial end distance and discrete coordinates of the component's axial direction.

[0017] The second module obtains the current gripping point distance between the two mobile manipulators and calculates the end distance increment based on the initial end distance.

[0018] The third module obtains the coordinates of discrete points on the centerline of the component based on the discrete coordinates of the component's axial direction and constructs a reference line. It then extracts the displacement of the discrete point coordinates relative to the reference line along the direction of gravity to generate a deflection sequence.

[0019] The fourth module calculates the deflection-induced axial geometric locking index through the deflection sequence, converts it into geometric elongation, calculates the difference between the end distance increment and the geometric elongation, and converts it into an axial internal force estimate.

[0020] The fifth module calculates the end distance correction based on the axial internal force estimate, allocates the end distance correction according to the ratio of the available axial travel of the two moving manipulators, and generates an updated grab point target.

[0021] The sixth module identifies the extreme deflection points from the deflection sequence as support locations and calculates the support force corresponding to the support locations based on the component stiffness parameters.

[0022] The seventh module discretizes the construction path into multiple fixed work stations. At each fixed work station, the second to sixth modules are executed sequentially. The grab point target drives the two mobile manipulators to advance, and the support position and support force are used as deterministic intervention conditions for human-machine hybrid construction.

[0023] The eighth module, after being advanced to the installation position, generates the final end distance correction amount according to the final assembly end distance, drives the two moving manipulators to move and release the component.

[0024] The beneficial effects of this invention are as follows: Addressing the construction challenges of closed-chain collaborative gripping of slender components in confined spaces using mobile manipulators, this invention quantifies the entire process from component bending deformation to axial latent internal forces by constructing a deflection-induced axial geometric locking index. It allows for the determination of end-distance corrections based on internal force calculations and their rational allocation according to the manipulator's stroke, achieving real-time dissipation of axial internal forces and preventing their accumulation as construction progresses. Simultaneously, it transforms manual intervention into quantified support positions and force parameters, enabling precise human-machine collaboration and specifically improving component bending deformation, thus mitigating the geometric locking effect. Furthermore, this invention discretizes the construction path into fixed workstations, cyclically calculating and adjusting the component's posture along the path to ensure the continuity of geometric calculation benchmarks at each workstation. The final end-distance adjustment uses a unified allocation rule to achieve precise matching between the construction end-distance and the assembly design end-distance, preventing residual internal forces from affecting assembly quality during construction. Overall, it standardizes multi-machine collaboration and human-machine hybrid construction processes, ensuring the geometric accuracy of component construction and assembly, and reducing clamping damage caused by excessive internal forces. Attached Figure Description

[0025] Figure 1 This is a schematic diagram of the calculation process of the present invention;

[0026] Figure 2 This is a schematic diagram of the computational scenario of the present invention. Detailed Implementation

[0027] The subject matter described herein will now be discussed with reference to exemplary embodiments. It should be understood that these embodiments are discussed only to enable those skilled in the art to better understand and implement the subject matter described herein, and changes may be made to the function and arrangement of the elements discussed without departing from the scope of this specification. Various processes or components may be omitted, substituted, or added as needed in the examples. Furthermore, features described in some examples may be combined in other examples.

[0028] It should be noted that, unless otherwise defined, the technical or scientific terms used in one or more embodiments of the present invention should have the ordinary meaning understood by one of ordinary skill in the art to which this invention pertains. The terms "first," "second," and similar terms used in one or more embodiments of the present invention do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Terms such as "comprising" or "including" indicate that the element or object preceding the term encompasses the elements or objects listed following the term and their equivalents, without excluding other elements or objects. Terms such as "connected" or "linked" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. Terms such as "upper," "lower," "left," and "right" are used only to indicate relative positional relationships; when the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0029] like Figures 1-2 As shown, a method for multi-robot collaborative and human-robot hybrid construction includes the following steps:

[0030] Step S1: Control the two moving manipulators to clamp the two ends of the component in a closed-chain cooperative gripping manner, and establish the initial end distance and discrete coordinates of the component axis;

[0031] Step S2: Obtain the current gripping point distance between the two moving manipulators, and calculate the end distance increment based on the initial end distance.

[0032] Step S3: Obtain the coordinates of discrete points on the center line of the component based on the discrete coordinates of the component's axial direction and construct a reference line. Extract the displacement of the discrete point coordinates relative to the reference line along the direction of gravity to generate a deflection sequence.

[0033] Step S4: Calculate the deflection-induced axial geometric locking index through the deflection sequence, convert it into geometric elongation, calculate the difference between the end distance increment and the geometric elongation, and convert it into an axial internal force estimate.

[0034] Step S5: Calculate the end distance correction amount based on the axial internal force estimate, allocate the end distance correction amount according to the ratio of the available axial stroke of the two moving manipulators, and generate the updated gripping point target.

[0035] Step S6: Find the extreme point of deflection in the deflection sequence as the support location, and calculate the support force corresponding to the support location based on the stiffness parameters of the component.

[0036] Step S7: Discretize the construction path into multiple fixed work positions. At each fixed work position, execute steps S2 to S6 in sequence. Use the grab point target to drive the two mobile manipulators to advance, and use the support position and support force as deterministic intervention conditions for human-machine hybrid construction.

[0037] Step S8: After advancing to the installation position, generate the final end distance correction amount according to the final assembly end distance, drive the two moving operating arms to move and release the component.

[0038] In one embodiment of the present invention, two mobile manipulators are controlled to clamp both ends of a component in a closed-chain cooperative gripping manner to determine the first end clamping point. With the second end clamping point ;

[0039] Establish initial end distance It satisfies the following formula:

[0040] ;in The initial end distance, First clamping point With the second end clamping point The resulting displacement vector It is the Euclidean norm;

[0041] Establish the axial unit vector of the component It satisfies the following formula:

[0042] ;in The axial unit vector of the component. The Euclidean norm of the displacement vector is equal to the initial end distance. ;

[0043] Establish the axial discrete coordinates of the component The axial distance of the component is the length of the walk. It satisfies the following formula:

[0044] ; ;in The axial discrete coordinates of the component The axial distance of the component from the walk length, For component length parameters, The number of discrete points is a parameter. is the discrete point index.

[0045] It should be noted that the first clamping point is the spatial position point where the moving manipulator clamps the left end of the component. The second clamping point is the spatial position point where the moving manipulator clamps the right end of the component. This parameter can be collected through a vision sensor, laser positioning sensor, or spatial positioning system at the end of the moving manipulator. The initial end distance is the straight-line distance between the first and second clamping points, reflecting the initial axial geometric constraint scale of the component after it is clamped in a closed chain. The component axial unit vector reflects the axial extension direction of the component. The component length parameter is the total axial length of the component as designed or measured, reflecting the basic geometric dimensions of the component. It is preferably taken as the length marked on the component's engineering drawings, which is the reference dimension for construction and matches the geometric requirements of on-site assembly. The discrete point number parameter is the total number of points used to discretize the axial length of the component. It is preferably an odd number between 21 and 101. An odd number of discrete points can include the axial midpoint of the component, balancing the accuracy and computational efficiency of discrete calculations. 21 points meet the calculation accuracy requirements for conventional slender components, while 101 points are suitable for calculations of components with higher accuracy requirements. Discrete point number is a numerical identifier that sequentially encodes the discrete points along the component's axial direction, reflecting the order in which the discrete points are arranged along the component's axial direction. The preferred value is a consecutive positive integer starting from 1. The component's axial discrete coordinates are the numerical positions of each discrete point along the component's axial direction, reflecting the axial distance of each discrete point relative to the left end of the component. The component's axial discrete step length is the axial distance between two adjacent discrete points, reflecting the interval scale of the component's axial discretization.

[0046] It should be noted that the specific clamping constraints of the closed-chain collaborative gripping method are as follows: the end clamps of both moving manipulators maintain a rigid clamping state with no relative sliding, and the axial direction of the clamped component is consistent with the cooperative movement direction of the manipulators; the clamping force is selected to be 50% to 80% of the allowable clamping force of the component material to avoid slippage due to insufficient force and damage to the component due to excessive force; the clamping point is selected from the process clamping area at the end of the component, which is the clamping area preset during the component design, and the distance from the end of the component is 50 mm to 200 mm. For example, a 100 mm wide process clamping area is preset at the end of a precast steel component, and the manipulator clamp must completely clamp this area. The specific value of the component length parameter should preferably be the design value from the engineering drawings. If there are processing errors in the component, the actual measured value on site should be used; the measurement method is to use a laser rangefinder to measure the distance between the centers of the process clamping areas of the component. During the measurement, the measurement direction of the laser rangefinder is consistent with the axial direction of the component, and the measurement error is no more than 2 mm. The selection principle for the number of discrete points is to balance computational accuracy and efficiency. When the axial length of the component is less than 5 meters, 21 discrete points are selected first, which satisfies the computational accuracy while reducing the computational load on the computer. When the axial length of the component is between 5 and 10 meters, 51 discrete points are selected first. When the axial length of the component is greater than 10 meters, 101 discrete points are selected first to avoid insufficient computational accuracy due to too few discrete points or reduced computational efficiency due to too many discrete points. The hardware collaborative control logic for the closed-chain collaborative gripping of the mobile manipulators involves two mobile manipulators connected to the same main controller. The main controller sends synchronous gripping commands to the two manipulators. After receiving the command, the end grippers of the manipulators complete the gripping action simultaneously within 0.5 seconds. After gripping, the position sensors of the two manipulators feed back the coordinates of the gripping point to the main controller in real time. The main controller synchronously locks the motion state of the two manipulators to ensure no relative movement during gripping, forming a stable closed-chain constraint.

[0047] It should be noted that this invention determines the geometric constraint benchmarks at both ends of the component through closed-chain collaborative gripping, establishing an initial end distance and a unit vector along the component's axis, providing a unified direction and scale benchmark for all subsequent axial geometric calculations. Simultaneously, the component's axial direction is discretized to generate uniform discrete coordinates and discrete step sizes, transforming continuous axial geometric calculations into discrete numerical calculations, adapting to the computational logic of computers in engineering. This allows the initial clamping state of multi-machine collaborative construction to form a stable closed-chain geometric constraint, avoiding axial direction deviations during component clamping and ensuring the directional consistency of subsequent end distance, deflection, and other calculations. The quantification of the initial end distance provides a unified benchmark scale for subsequent incremental end distance calculations, eliminating calculation errors caused by benchmark ambiguity. Uniform axial discretization balances the accuracy and computational efficiency of subsequent shape sampling, making the results of discrete calculations more closely match the actual geometric characteristics of the component.

[0048] In one embodiment of the present invention, the first end gripping point of the two mobile manipulators at the current moment is obtained. With the second end clamping point ;

[0049] Establish current gripping point distance It satisfies the following formula:

[0050] ;in The distance to the current clamping point. Let be the displacement vector of the two clamping points. It is the Euclidean norm;

[0051] Establish end-to-end distance increment It satisfies the following formula:

[0052] ;in For end-to-end distance increments, The distance to the current clamping point. This is the initial end distance.

[0053] It should be noted that the current clamping point distance is the straight-line distance between the first and second clamping points at the current moment, reflecting the axial geometric constraint dimensions of the component under its current construction state. The end distance increment is the difference between the current clamping point distance and the initial end distance, reflecting the real-time change in the axial geometric constraint dimensions of the component during construction. The real-time acquisition frequency of the clamping point position at the current moment is preferably set between 10 Hz and 50 Hz. This frequency range can accurately capture minute positional changes of the clamping point without increasing the computational load on the main controller due to excessively high frequencies. For the collaborative handling of slender components in building construction, 20 Hz is sufficient. The acquisition synchronization requirement is that the acquisition time difference between the clamping point positions of the two moving manipulators should not exceed 10 milliseconds to avoid deviations in displacement vector calculation due to asynchronous acquisition. Acquisition synchronization can be achieved by sending a synchronization acquisition trigger command to the two manipulators through the main controller. A positive value for the end distance increment is defined as the current clamping point distance being greater than the initial end distance, which geometrically means that the current axial constraint dimension of the component is larger than the initial state, and the component is in a stretched state axially; a negative value is defined as the current clamping point distance being less than the initial end distance, which geometrically means that the current axial constraint dimension of the component is smaller than the initial state, and the component is in a compressed state axially; a zero value indicates that the current clamping point distance is exactly the same as the initial end distance, and the axial constraint dimension of the component remains unchanged.

[0054] It should be noted that this invention calculates the current clamping point distance by real-time acquisition of the clamping point position during construction, and then calculates the difference between this distance and the initial end distance to obtain the end distance increment. This quantifies the real-time change in the axial geometric constraint dimensions of the component during construction. This allows for real-time characterization of the component's axial geometric state during construction, promptly capturing axial dimensional changes caused by minute changes in the clamping point position. Furthermore, this calculation method is adapted to the real-time movement of the moving manipulator, providing real-time input parameters for subsequent axial internal force calculations, ensuring that the internal force calculations closely reflect the actual construction state of the component.

[0055] In one embodiment of the present invention, based on the discrete coordinates of the component axial direction... Obtain the coordinates of discrete points on the centerline of the component ;

[0056] Establish the coordinates of discrete points on the reference line by constructing segments using the coordinates of discrete points on the reference line. It satisfies the following formula:

[0057] ;

[0058] in The coordinates of the discrete points on the reference line. The first clamping point, This is the second clamping point. The axial discrete coordinates of the component For component length parameters, Let be the displacement vector of the two clamping points;

[0059] A deflection sequence is established by generating segments from the deflection sequence. It satisfies the following formula:

[0060] ;

[0061] in For each term in the deflection sequence, The unit vector representing the direction of gravity. Represents the unit vector along the direction of gravity. Projection operation, This is the displacement vector of the discrete point coordinates relative to the discrete point coordinates of the reference line.

[0062] It should be noted that the discrete point coordinates of the component centerline are spatial position points corresponding one-to-one with the axial discrete points at the component centerline, reflecting the actual spatial position of the component centerline at each discrete point. The centerline can be extracted by scanning the overall outline of the component using a line laser 3D point cloud scanner, and then this parameter is acquired by matching the component's axial discrete coordinates. The reference line is the line connecting the endpoints of the first and second clamping points, reflecting the ideal centerline position of the component when there is no bending deformation, and serves as the baseline for deflection calculation. The normalized ratio is the ratio of the component's axial discrete coordinates to the component length parameter, reflecting the relative position proportion of each discrete point along the component's axial direction. The discrete point coordinates of the reference line are spatial position points on the reference line corresponding one-to-one with the component's axial discrete points, reflecting the spatial position of the component's ideal centerline at each discrete point. The unit vector of gravity direction is a dimensionless spatial vector representing the direction of gravity, reflecting the actual direction of gravity in the construction site. The preferred value is the unit vector along the negative Z-axis in the unified coordinate system of the site, which is completely consistent with the actual gravity direction of the construction site, conforming to the calculation requirements of vertical geometric quantities. The deflection sequence is a set of projection values ​​of the relevant displacement vectors at each discrete point along the direction of gravity, arranged according to the discrete point number, reflecting the vertical bending deflection distribution characteristics at each axial discrete point of the component.

[0063] It should be noted that during construction, components will undergo overall translation or tilting as the operating arm moves. This type of movement is not a bending deformation of the component itself. When the line connecting the two clamping points is used as a reference line, the reference line will change synchronously with the overall translation and tilt of the component. When calculating the displacement of the component's centerline relative to the reference line, the influence of the overall movement will be naturally offset, retaining only the bending deformation of the component itself. For example, if a slender precast beam tilts 30 degrees with the operating arm, the reference line connecting the endpoints will also tilt 30 degrees synchronously. The calculated displacement only reflects the bending of the beam itself, without any numerical deviation caused by the tilt. The discrete axial coordinates of the component are absolute position values ​​that change with the component's length. After comparing them with the component's length parameter, the result will be converted into a dimensionless value between 0 and 1, eliminating the influence of the component's length scale and achieving a unified quantification of the axial position of components of different lengths. In addition, the bending deformation of slender components in building construction is mainly reflected in the vertical offset in the direction of gravity. The projected displacement in this direction directly corresponds to the degree of vertical bending of the component due to its own weight or external forces, and is the core physical quantity characterizing the bending deformation of the component. The influence of displacement in other directions on the axial geometric locking of the component is negligible.

[0064] It should be noted that the method for acquiring the coordinates of discrete points on the component's centerline is as follows: first, a 3D point cloud scanner is used to scan the entire outline of the component, generating a 3D point cloud model of the component; then, a point cloud processing algorithm is used to extract the centerline outline of the component; finally, the centerline outline is matched with the component's axial discrete coordinates according to their sequence numbers to obtain the coordinates of each discrete point. The acquisition equipment used is a line laser 3D point cloud scanner, accompanied by point cloud processing software; the acquisition accuracy requirement is that the acquisition error of the spatial coordinates is no greater than 1 mm, matching the acquisition accuracy of the clamping point position. The coordinate system of the unit vector in the direction of gravity is defined as the unified coordinate system of the construction site. This coordinate system takes the construction site reference point as the origin, the building axis as the X-axis and Y-axis, and the opposite direction of gravity as the positive direction of the Z-axis; the direction value rule is along the negative direction of the Z-axis of the unified coordinate system of the site, and the magnitude of the vector is fixed at 1, ensuring that the vector is a dimensionless quantity of pure direction. In addition, the numerical storage format of the deflection sequence is a one-dimensional array format; the encoding rule is to arrange the discrete points along the axial direction of the component in ascending order, with the deflection value of the first discrete point being the first element of the array and the deflection value of the last discrete point being the last element of the array; the criterion for determining valid values ​​is that the absolute value of the deflection value is greater than 0.001 mm, and values ​​less than this are considered to be minor errors caused by acquisition or calculation and are not included in the deflection sequence to avoid interfering with subsequent calculations.

[0065] It should be noted that this invention, based on the established axial discrete coordinates of the component, collects the corresponding discrete coordinates of the centerline. Using the line connecting the two clamping points as a reference line, it establishes the discrete coordinates of the reference line's points. The vertical projection displacement of the centerline relative to the reference line is extracted using a unit vector in the direction of gravity, and a deflection sequence is generated sequentially to quantify the vertical bending deformation of the component. This eliminates the interference of overall component translation and tilting on deflection calculations, ensuring that the deflection value reflects only the component's own bending deformation. The use of a normalization ratio ensures accurate matching of the positions of the reference line discrete points and the centerline discrete points, and the projection operation achieves accurate extraction of the vertical deflection components. The deflection sequence is arranged by discrete point number, clearly reflecting the component's deflection distribution characteristics and providing fundamental data for subsequent calculations of deflection slope and geometric locking index.

[0066] In one embodiment of the present invention, the deflection discrete slope is calculated. Integral with the square of the slope It satisfies the following formula:

[0067] ;

[0068] ;

[0069] in The slope of the deflection discrete curve. For each term in the deflection sequence, The axial distance of the component from the walk length, It is the integral of the square of the slope;

[0070] Establish deflection-induced axial geometric locking index It satisfies the following formula:

[0071] ;

[0072] in The deflection-induced axial geometric locking index. This refers to the axial stiffness parameter of the component. The bending stiffness parameter of the component. For component length parameters, Pi is a constant.

[0073] Establish geometric elongation It satisfies the following formula:

[0074] ;

[0075] in For geometric elongation, This is the conversion scaling factor;

[0076] Establish axial internal force estimation It satisfies the following formula:

[0077] ;

[0078] in For axial internal force estimation, For end-to-end distance increments, This is a proportionality factor used to convert the difference into axial force.

[0079] It should be noted that the deflection discrete slope reflects the rate of change of deflection and the steepness of bending within each discrete interval of the component. The integral of the square of the slope reflects the overall bending deformation and potential tendency of geometric elongation of the component. The axial stiffness parameter of the component is an indicator of its ability to resist axial tensile or compressive deformation, reflecting the axial stiffness characteristics of the component. The preferred value is the theoretical value calculated using material mechanics formulas based on the component's material and cross-sectional dimensions. This theoretical value closely matches the mechanical properties of the component design and meets the conventional requirements of engineering mechanics calculations. The bending stiffness parameter of the component is an indicator of its ability to resist bending deformation, reflecting the bending stiffness characteristics of the component. The preferred value is the theoretical value calculated based on the elastic modulus and the second moment of the cross-section of the component's material. This value accurately characterizes the bending resistance of the component and is a fundamental mechanical parameter for calculations related to bending deformation. Pi constant is the fixed ratio of the circumference to the diameter of a circle. The first product result reflects the quantitative value combining the axial stiffness of the component and the overall bending deformation. The second product result reflects the normalized quantitative value related to the bending stiffness of the component. The deflection-induced axial geometric locking exponent is the ratio of the first product result to the second product result, reflecting the strength of axial geometric locking caused by bending deformation of the member, and is a dimensionless index. The conversion proportionality coefficient is the product of the square of pi and the member's bending stiffness parameter, divided by the product of the member's axial stiffness parameter and the member's length parameter, reflecting the dimensional and numerical conversion relationship from the geometric locking exponent to geometric elongation. Geometric elongation is the product of the deflection-induced axial geometric locking exponent and the conversion proportionality coefficient, reflecting the equivalent axial geometric elongation of the member due to bending deformation. The axial internal force estimate is the value obtained by correcting the difference between the end-distance increment and the geometric elongation for the axial stiffness parameter and the member's length parameter, reflecting the magnitude and direction of the axial internal force caused by the mismatch between the end-distance increment and the geometric elongation.

[0080] It should be noted that axial stiffness reflects a component's resistance to axial deformation, while bending stiffness reflects its resistance to bending deformation. The ratio of the two directly reflects the component's mechanical property bias. A large ratio indicates that the component is not easily deformed axially but is easily bent, while a small ratio indicates that the component is easily deformed axially but not easily bent. This is an important quantitative indicator for distinguishing the mechanical properties of components. For example, slender, thin-walled steel components have high axial stiffness and low bending stiffness, resulting in a large ratio; conversely, short, thick concrete components have both high axial stiffness and high bending stiffness, resulting in a small ratio. By normalizing the integral of the square of the slope using parameters such as axial stiffness, bending stiffness, and component length to eliminate the influence of dimensions, the resulting dimensionless exponent can directly quantify the strength of axial geometric locking caused by bending deformation. A larger exponent indicates a more pronounced axial locking caused by bending, achieving a direct conversion from the amount of bending deformation to the degree of locking. The geometric locking index is dimensionless and cannot directly reflect the actual geometric elongation. The conversion ratio factor integrates mechanical and geometric parameters such as the axial stiffness, bending stiffness, and length of the component. By multiplying it with the locking index, the dimensionless locking index can be converted into an actual geometric elongation value with a length dimension, achieving dimensional restoration and numerical matching. For example, if the conversion ratio factor is 0.002 mm and the locking index is 5, the resulting product is a geometric elongation of 0.01 mm. Furthermore, according to Hooke's Law in mechanics of materials, axial internal force is directly proportional to axial deformation and axial stiffness, and inversely proportional to the component length. The difference in axial geometric dimensions is the axial deformation. Multiplying this difference by the axial stiffness and dividing by the component length directly calculates the corresponding axial internal force, a classic method in engineering for calculating internal force from deformation. For example, if the axial stiffness is 1000 N / mm, the component length is 5000 mm, and the geometric dimension difference is 3 mm, the calculated axial internal force is 0.6 N, achieving the conversion from deformation to internal force.

[0081] It should be noted that the values ​​of the axial stiffness and bending stiffness parameters of the components are preferentially determined based on the theoretical values ​​calculated from the elastic modulus and cross-sectional area of ​​the component material for axial stiffness, and from the elastic modulus and the second moment of the section for bending stiffness. If there are processing errors in the components, the on-site measured values ​​are used. The measurement method is to sample components from the same batch, conduct axial tensile and compression tests using a universal testing machine to measure axial stiffness, and conduct three-point bending tests to measure bending stiffness. The accuracy requirement is that the relative error of the parameter calculation or measurement is no greater than 5%, meeting the accuracy requirements of building construction mechanics calculations. The effective numerical range of the deflection-induced axial geometric locking index is 0 to 10. A value greater than 10 is considered a calculation error, and data needs to be collected and recalculated. The engineering significance criteria are as follows: a value less than 1 indicates a weak degree of axial geometric locking, a value of 1 to 5 indicates a moderate degree of locking, and a value of 5 to 10 indicates a strong degree of locking. Different degrees of locking correspond to different subsequent end distance correction strategies. The dimension of geometric elongation is defined as length, with the unit being millimeters, consistent with the dimensions of component length and deflection. The criterion for determining a valid value is that an absolute value greater than 0.001 mm is considered valid geometric elongation; values ​​less than this are considered calculation errors and are not included in subsequent axial internal force calculations. Furthermore, a positive value for axial internal force estimation is defined as an end-distance increment greater than the geometric elongation, corresponding to axial tension, meaning the component is subjected to axial stretching; a negative value is defined as an end-distance increment less than the geometric elongation, corresponding to axial compression, meaning the component is subjected to axial compressive stress; a zero value indicates that the end-distance increment perfectly matches the geometric elongation, and the component has no axial internal force.

[0082] It should be noted that this invention calculates the deflection discrete slope and the integral of the square of the slope using a discretization method, integrating two core mechanical parameters—axial and bending stiffness—to construct a deflection-induced axial geometric locking index. This achieves dimensionless quantification of the transformation from bending deformation to axial locking. Then, by converting a proportionality coefficient, the index is transformed into a geometric elongation with practical engineering significance. Finally, combined with the end-distance increment, the axial internal force is estimated based on Hooke's law in mechanics of materials, realizing a full-link quantitative derivation from component bending deformation to axial internal force. This invention achieves engineering-based solutions for continuous bending characteristics through discretization calculations, making the quantification of bending deformation compatible with computer computing logic. The geometric locking index eliminates dimensional interference, making the determination of the degree of axial locking more intuitive. The proportionality coefficient completes the dimensionless to dimensional transformation, ensuring the practical reference value of the geometric elongation. The calculation of axial internal force estimation relies on classical mechanics laws, providing theoretical support for the quantification of internal forces. All calculation steps are clearly defined and error control requirements are clear, which can provide accurate internal force input parameters for subsequent end distance correction, making the internal force resolution in multi-machine collaborative construction more targeted, and the calculation results are more in line with the actual mechanical state of the components.

[0083] In one embodiment of the present invention, the end distance correction amount is established through an end distance correction amount calculation segment. It satisfies the following formula:

[0084] ;

[0085] in This is the end distance correction amount. For component length parameters, This refers to the axial stiffness parameter of the component. For axial internal force estimation, This is the proportionality coefficient for the change in axial internal force to end distance;

[0086] The allocation coefficient is established by allocating the end distance correction segment. End distance correction amount allocated to the first end clamping point And the end distance correction amount allocated to the second end clamping point It satisfies the following formula:

[0087] ;

[0088] ; ;

[0089] in For allocation coefficients, For the first available axial stroke, For the second available axial stroke, The end distance correction amount allocated to the first end clamping point. This is the end distance correction amount allocated to the second end clamping point;

[0090] The first updated capture point target is established through the capture point target generation segment. With the second updated grab point target It satisfies the following formula:

[0091] ; ;

[0092] in The first updated target for grabbing points. For the second updated grab point target, The first clamping point, This is the second clamping point. This is the unit vector along the axial direction of the component.

[0093] It should be noted that the proportionality coefficient of axial internal force to end distance change is the ratio of the component length parameter to the component axial stiffness parameter, reflecting the end distance change corresponding to a unit axial internal force, establishing a linear relationship between axial internal force and end distance change. The end distance correction is the product of the proportionality coefficient of axial internal force to end distance change and the negative value of the estimated axial internal force, reflecting the total end distance adjustment required to offset the axial internal force of the component. The first available axial travel is the maximum distance that the first moving manipulator can move along the component's axial direction (unit vector direction), reflecting the axial movement capability of the first moving manipulator. The second available axial travel is the maximum distance that the second moving manipulator can move along the component's axial direction (unit vector direction), reflecting the axial movement capability of the second moving manipulator; this parameter can be acquired through the servo control system and axial travel sensor of the moving manipulator, reading the measured travel limit values ​​of the manipulator's axial mechanism. The allocation coefficient is the ratio of the first available axial travel to the sum of the available axial travels of the two moving manipulators, reflecting the proportion of the end distance correction allocated to the first moving manipulator. The end-distance correction amount allocated to the first end-grip point reflects the specific value of the end-distance adjustment that needs to be performed on the first end-grip point. The end-distance correction amount allocated to the second end-grip point reflects the specific value of the end-distance adjustment that needs to be performed on the second end-grip point. The first updated gripping point target is the spatial position point of the first end-grip point after axial end-distance correction, reflecting the gripping point target position that the first moving manipulator needs to reach. The second updated gripping point target is the spatial position point of the second end-grip point after axial end-distance correction, reflecting the gripping point target position that the second moving manipulator needs to reach.

[0094] It should be noted that, according to Hooke's Law in mechanics of materials, axial internal force and axial deformation are linearly positively correlated. The change in end distance is the axial deformation of the component. The ratio of the component length to the axial stiffness is the proportionality coefficient of this linear relationship, which can directly establish the correspondence between unit axial internal force and end distance change. For example, if the component length is 5000 mm and the axial stiffness is 1000 N / mm, the proportionality coefficient is 5, meaning that 1 N of axial internal force corresponds to 5 mm of end distance change. Different moving manipulators have different usable axial strokes due to differences in mechanism posture and spatial constraints. Allocating correction amounts according to the stroke proportion ensures that the adjustment amount of each manipulator is within its own stroke range, avoiding situations where a single manipulator exceeds its stroke and cannot perform adjustments. For example, if the first usable axial stroke is 20 mm and the second usable axial stroke is 30 mm, with an allocation coefficient of 0.4 and a correction amount of 50 mm, adjusting the first end by 20 mm and the second end by 30 mm is both within the stroke range. The first clamping point performs a subtraction operation, while the second clamping point performs an addition operation. These two correction actions occur in opposite directions along the component's axial direction, directly changing the straight-line distance between the two clamping points and achieving precise adjustment of the end distance. If operations were performed in the same direction, the end distance could not be changed. Furthermore, the correction amount allocated to both ends is the product of the allocation coefficient and the unit minus the allocation coefficient, respectively, with the total correction amount remaining constant after summing. This ensures that the actual adjusted end distance perfectly matches the theoretical correction amount required to counteract the axial internal force, avoiding deviations in adjustment due to allocation.

[0095] It should be noted that the first and second available axial travel are acquired by reading the axial travel limit parameters of the motor encoder through the moving manipulator servo system, and then verifying them in conjunction with the real-time detection values ​​of the axial travel sensor. The acquisition timing is 100 milliseconds before each end-distance correction calculation to ensure that the acquired values ​​closely match the current actual state of the manipulator. The effective travel criterion is that a acquired value greater than 0 mm is considered effective travel; a value less than or equal to 0 mm is considered a limited axial travel of the manipulator, requiring the end-distance correction to be paused and the manipulator posture replanned. A positive end-distance correction value is defined as needing to increase the end-distance of the component, with the corresponding adjustment direction being the movement of the two clamping points in opposite directions along the axial direction; a negative value is defined as needing to decrease the end-distance of the component, with the corresponding adjustment direction being the movement of the two clamping points in opposite directions along the axial direction; a zero value indicates that no end-distance adjustment is needed, and the two manipulators can maintain their current clamping point positions. The motion control accuracy of the mobile manipulator at the target gripping point is such that the spatial position error of the end gripping point reaching the target position is no greater than 1 mm, meeting the geometric accuracy requirements of building construction. The criteria for determining whether the manipulator is in position are: the spatial distance between the actual position of the end gripping point and the target position is less than 1 mm, and the manipulator's motion mechanism is stationary for at least 500 milliseconds. Furthermore, the travel limit protection rule for the mobile manipulator along the unit vector direction of the component's axial direction is as follows: when the manipulator moves towards the target gripping point, the servo system monitors the axial travel distance in real time. If the actual travel distance reaches 95% of the available axial travel but the target position is not reached, the travel limit protection is immediately triggered, stopping the manipulator's movement and simultaneously sending a travel limit alarm signal to the main controller, which then re-plans the end distance correction scheme.

[0096] It should be noted that, based on the linear relationship between axial internal force and end distance change in mechanics of materials, the end distance correction amount that can offset the axial internal force is calculated. Then, the correction amount is allocated according to the actual available axial travel ratio of the two moving manipulators. Combining the component's axial unit vector, the one-dimensional end distance correction amount is transformed into a three-dimensional spatial displacement. An updated gripping point target is generated through inverse position calculations of the two clamping points, achieving precise adjustment of the end distance and resolution of the axial internal force. This directly links the end distance correction amount to the axial internal force, specifically addressing the axial internal force problem caused by the mismatch between the end distance increment and geometric elongation. Allocating the correction amount according to the available travel adapts to the actual motion capability of the manipulators, avoiding overtravel malfunctions. Transforming a one-dimensional correction amount into a three-dimensional gripping point target ensures the accuracy of spatial position adjustment. The inverse calculation method at both ends directly achieves end distance adjustment, and the total verification rule after the correction amount allocation ensures that the adjusted amount is consistent with the theoretical value; ensuring the consistency and effectiveness of multi-machine collaborative end distance adjustment.

[0097] In one embodiment of the present invention, the support position is established by a support position determination segment. It satisfies the following formula:

[0098] ;

[0099] ;

[0100] in The parameter corresponding to the term with the largest absolute value of deflection. For each term in the deflection sequence, It is the absolute value symbol. To support the position, The coordinate value corresponding to the parameter with the largest absolute value of deflection in the discrete axial coordinates of the component;

[0101] Establish the support force corresponding to the support location through the support force calculation segment. It satisfies the following formula:

[0102] ;

[0103] ;

[0104] in To provide the supporting force corresponding to the support position, The bending stiffness parameter of the component. For component length parameters, This is the proportionality coefficient. This represents the deflection value at the support location. This is the deflection term in the deflection sequence corresponding to the support location.

[0105] It should be noted that the term with the largest absolute deflection value is the single deflection value with the largest absolute value in the deflection sequence, reflecting the quantitative value of the deflection corresponding to the position with the most significant bending deformation among the discrete points in the axial direction of the component. The parameter corresponding to the term with the largest absolute deflection value is the coded sequence number of the term with the largest absolute deflection value in the deflection sequence, reflecting the index information of the discrete point in the component's axial direction for this extreme deflection value. The support position is the axial coordinate value in the component's axial discrete coordinates that matches the parameter corresponding to the term with the largest absolute deflection value, reflecting the optimal axial position for implementing artificial support on the component. The deflection value at the support position is the actual value of the term with the largest absolute deflection value corresponding to the support position in the deflection sequence, reflecting the actual degree of vertical bending deformation of the component at the support position. The proportionality coefficient is the ratio of 48 times the component's bending stiffness parameter to the cube of the component's length parameter, reflecting the magnitude of the artificial support force corresponding to a unit deflection value at the support position, establishing a quantitative conversion relationship between deflection and support force. The support force is the product of the proportionality coefficient and the deflection value at the support position, reflecting the quantitative value of the artificial support force that needs to be applied at the support position to improve the bending deformation at the support position.

[0106] It should be noted that the bending deformation of the component is unevenly distributed along the axial direction, with the bending deformation being most prominent at the point of maximum absolute deflection. Applying an upward supporting force at this location can directly counteract the bending tendency caused by gravity, thereby reducing the overall bending deformation of the component, which is far superior to the effect of applying force at locations with small deflections. Within the elastic deformation range of the material, the bending deflection of the component is linearly positively correlated with the applied load. The artificial supporting force, as a reverse applied load, can counteract the deflection caused by bending. Therefore, the larger the deflection value at the supporting location, the larger the supporting force required, and the two maintain a strict linear relationship. Furthermore, the calculation of the supporting force only incorporates the component's own stiffness parameters such as bending stiffness and length, as well as the actual deflection value at the supporting location. These parameters are determined by the component's own characteristics and actual bending state, without considering other external loads during construction; targeted support is implemented only for the bending deformation of the component itself.

[0107] It should be noted that the vertical spatial coordinates of the support position, excluding the axial coordinates, are the vertical values ​​of the discrete points on the reference line at the support position plus the deflection value at the support position. The lateral spatial coordinates are consistent with the lateral coordinates of the component's centerline at that axial position, ensuring that the spatial position of the artificial support coincides with the component's centerline and avoiding additional bending caused by eccentric loading. The unit of support force is defined as Newtons; the criterion for determining a valid value is that the calculated support force is greater than 0 Newtons, and if it is equal to 0 Newtons, no support force needs to be applied; the accuracy requirement for the force direction is along the positive Z-axis of the unified coordinate system of the site, that is, vertically upward, completely opposite to the direction of gravity, with a deviation of no more than 5 degrees, to avoid component displacement caused by lateral force components. The support force is applied using a manual hydraulic jack or portable support device for point support. The contact area between the contact end of the support device and the component surface is no less than 100 square centimeters to avoid stress concentration damage to the component. The accuracy of the applied force is that the deviation between the actual applied support force and the calculated value is no more than 5%, which can be monitored in real time by the force sensor on the support device. The continuous application time requirement is from the completion of the support force calculation at this work station until the component is moved to the next work station and the new support position and support force calculation are completed, ensuring the continuity of support during construction. In addition, the deflection sequence is re-measured every 5 seconds after the support force is applied to re-acquire and calculate the deflection sequence and monitor the improvement of bending deformation in real time. The dynamic adjustment rule is that if the deflection value at the re-measured support position is reduced by more than 50% compared with the initial value, the support force is recalculated according to the new deflection value and adjusted to the new value. If the deflection value does not change significantly, the direction of force application and the force application device are checked, and the original support force is kept unchanged after confirmation.

[0108] It should be noted that this invention identifies the term with the largest absolute deflection value by traversing the deflection sequence, locates the optimal support position based on the sequence's matching relationship with the axial discrete coordinates, extracts the actual deflection value at that position, and then derives a proportional coefficient by combining the component's bending stiffness and length parameters. This quantifies the deflection value into the required support force, achieving parameterization and determinism of manual intervention in human-machine hybrid construction. This allows for the placement of manual support at the point of most significant bending deformation in the component, enabling direct intervention at the core bending location and improving the efficiency of improving component bending deformation. The support force is quantitatively calculated using material mechanics parameters, avoiding the arbitrariness of empirical force application and providing a clear numerical basis for manual force application. The parameterized definition of support position and support force enables standardized intervention procedures in human-machine collaborative construction, facilitating execution by on-site construction personnel. The dynamic adjustment rules for the support force can adapt to real-time changes in component bending deformation, ensuring the support effect continuously matches the actual state of the component. Simultaneously, the improved bending deformation reduces the deflection-induced axial geometric locking index, providing a better component morphology basis for subsequent end-distance correction and internal force resolution.

[0109] In one embodiment of the present invention, the construction path is discretized into fixed workstation discrete segments. A fixed workstation, satisfying the following formula:

[0110] ;

[0111] in For the number of fixed workstations, For fixed workstation variables;

[0112] At a fixed workstation At this point, steps S102 to S106 are executed sequentially through the step call segment within the fixed workstation, satisfying the following calculation link relationship:

[0113] ;

[0114] ;

[0115] in The first updated target for grabbing points. For the second updated grab point target, For end-to-end distance increments, For deflection sequence, The deflection-induced axial geometric locking index. For geometric elongation, For axial internal force estimation, This is the end distance correction amount. For the first available axial stroke, For the second available axial stroke, To support the position, For support, The bending stiffness parameter of the component. This refers to the component length parameter;

[0116] The two mobile manipulators are advanced and their states are updated by driving the advance and deterministic intervention segments, satisfying the following formula:

[0117] ; ;

[0118] in The first clamping point, This is the second clamping point.

[0119] It should be noted that the total number of fixed workstations is the total number of workstations obtained after discretizing the construction path, reflecting the overall number of stages in the phased advancement of the construction path. The preferred value is a positive integer rounded up after dividing the construction path length at intervals of 1 to 2 meters, and the total number is not less than 2. This method balances the timeliness of adjustments and computational efficiency in construction in narrow spaces; too small an interval will increase construction time, while too large an interval will fail to promptly dissipate axial internal forces. The fixed workstation variable is a numerical identifier representing the current fixed workstation position, reflecting the real-time stage position during construction. The preferred value is a continuous positive integer starting from 1, ranging up to the total number of fixed workstations. This coding method conforms to the conventional identification rules for phased construction in the engineering field, facilitating workstation indexing and switching. The first end clamping point of the next fixed workstation is the position of the first moving manipulator clamping point after the current workstation has advanced, serving as the calculation reference for the next workstation, reflecting the left-hand reference dimension for the geometric calculation of the next workstation. The second end clamping point of the next fixed station is the position of the second moving manipulator clamping point after the current station has been advanced, which serves as the calculation benchmark for the next station and reflects the right-end benchmark scale for the geometric calculation of the next station.

[0120] It should be noted that the specific basis for the discretization of the construction path into fixed workstations is the actual length of the construction path and the degree of spatial constraints. Areas with high spatial constraints, such as corridor turns and narrow passages, use smaller spacing values, while straight sections with less constraint use larger spacing values. The rule for determining the workstation spacing is that straight sections are prioritized at 2 meters, and constrained areas are prioritized at 1 meter. The partitioning tools are a construction laser rangefinder and path planning software. First, the length of each segment of the path is measured using the laser rangefinder, and then the discretization is completed in the path planning software. The selection principle for the total number of fixed workstations is to round up after discretization based on the workstation spacing to ensure that the entire path is covered by workstations. The matching relationship is that the longer the construction path, the more fixed workstations there are, and the two are positively correlated. The adjustment method is that if unforeseen spatial obstacles are found in the construction path, one to two workstations are added before and after the obstacle, reducing the workstation spacing around the obstacle to 0.5 meters. The preferred motion speed for the moving manipulator driven by the grasping point target is 0.05 m / s to 0.1 m / s, with 0.05 m / s in narrow, constrained spaces and 0.1 m / s in straight sections. The collaborative control rule involves the main controller synchronously sending motion commands to both moving manipulators, ensuring complete synchronization of their start-up, speed, and stop. The positioning criterion is that the spatial error between the actual position of the manipulator's end gripping point and the grasping point target is less than 1 mm, and the manipulator's motion mechanism is stationary for at least 500 milliseconds. The timing for deterministic intervention in human-machine hybrid construction is as follows: when the moving manipulator begins to move towards the grasping point target, the manual operator places the support device in the preset support position. Once the robot is in position and stationary, the support force is immediately applied. The synchronization requirement is that the time difference between the manual application of the support force and the robot's arrival at position is no more than 1 second to avoid sudden changes in component posture due to asynchronous actions. The criteria for determining the completion of station advancement are: the moving manipulator is in position and stationary; manual support force is applied and the deviation between the measured and calculated values ​​is less than 5%; and the deflection sequence retest results are valid. All three conditions must be met simultaneously for advancement to be considered complete. The trigger condition for updating the clamping point reference is that after the determination result of station advancement completion is generated, the main controller automatically triggers a reference update, setting the current first and second updated gripping point targets as the first and second end clamping points of the next fixed station. Furthermore, during continuous advancement across multiple stations, the monitoring frequency of deflection and internal force is as follows: after each fixed station advancement and reference update, a comprehensive retest of the component's deflection sequence and axial internal force estimation is performed. The abnormal warning threshold is when the absolute value of the deflection is greater than 5 mm or the absolute value of the estimated axial internal force is greater than 500 N. Upon triggering the warning, construction is immediately suspended, the end distance correction and support force are recalculated, and advancement continues only after adjustments are completed and the retest is passed.

[0121] It should be noted that this invention discretizes the construction path in a narrow space into fixed workstations. These fixed workstation variables identify the construction stage positions. At each workstation, the calculation of the gripping point target and human-machine intervention parameters is completed sequentially according to a predetermined calculation link. This drives two mobile manipulators to collaboratively advance and perform deterministic manual support. After advancement, the gripping point reference is updated as the calculation basis for the next workstation, realizing phased, programmed multi-machine collaboration and human-machine hybrid construction. This transforms continuous construction in a narrow space into phased operations, adapting to the programmed control logic of the main controller and facilitating standardized management of the construction process. The cyclical calculation and adjustment at each workstation can resolve axial internal forces in real time, avoiding the accumulation of internal forces and calculation errors. The calculation link ensures the correlation of parameter derivation, allowing the gripping point target and intervention parameters to closely match the actual geometric and mechanical state of the component. The actions of the robot and the human are synchronized at each workstation, giving deterministic human-machine collaborative construction. The continuous updating of the gripping point reference ensures the consistency of the geometric calculation reference at each workstation, eliminating errors caused by sudden changes in the reference. By defining intervention conditions at each work station, the system can accurately adapt to changes in the posture of components along the path, making construction adjustments more targeted.

[0122] In one embodiment of the present invention, the final assembly end distance is established through the final end distance correction generation segment. Current clamping point distance and the last end distance correction amount It satisfies the following formula:

[0123] ;

[0124] ;

[0125] ;

[0126] in For the final assembly end distance, Assemble the target clamping point at the first end. Assemble the target clamping point at the second end. The distance to the current clamping point. The first clamping point, This is the second clamping point. It is the Euclidean norm. This is the final end distance correction amount;

[0127] The allocation coefficient is established using the last end distance correction allocation segment. The final end distance correction amount allocated to the first end clamping point and the final end distance correction amount allocated to the second end clamping point It satisfies the following formula:

[0128] ;

[0129] ; ;

[0130] in For allocation coefficients, For the first available axial stroke, For the second available axial stroke, This is the final end distance correction amount allocated to the first end clamping point. This is the final end distance correction amount allocated to the second end clamping point;

[0131] The first final grab target is established through the final grab target generation and release segment. With the second final grab target It satisfies the following formula:

[0132] ; ;

[0133] in For the first and last capture point target, For the second and final capture point target, The first clamping point, This is the second clamping point. This is the unit vector along the axial direction of the component.

[0134] It should be noted that the first end assembly target clamping point is the preset clamping positioning point on the left end of the final installation position of the component. The second end assembly target clamping point is the preset clamping positioning point on the right end of the final installation position of the component. The final assembly end distance is the straight-line distance between the first end assembly target clamping point and the second end assembly target clamping point, reflecting the axial geometric dimension requirements required for the final installation of the component. The final end distance correction is the difference between the final assembly end distance and the current clamping point distance, reflecting the final axial end distance correction value required for the component to be adjusted from the construction state to the assembly state. The final end distance correction allocated to the first end clamping point reflects the specific end distance correction value that needs to be performed on the left end of the component to achieve the assembly end distance. The final end distance correction allocated to the second end clamping point reflects the specific end distance correction value that needs to be performed on the right end of the component to achieve the assembly end distance. The first final gripping point target is the spatial position point of the first end clamping point after the final end distance correction, reflecting the final clamping target position of the first moving manipulator pushing the component to the assembly position. The second final gripping point target is the spatial position of the second end gripping point after the final end distance correction, reflecting the final gripping target position of the component pushed by the second moving manipulator to the assembly position.

[0135] It should be noted that the determination of the first and second end assembly target clamping points is based on the installation reference points, assembly bracket positioning points, or component mounting hole centers marked on the building construction and installation design drawings. The data acquisition method involves scanning the installation position reference marks with a 3D laser positioning instrument and performing coordinate calculations using a unified site coordinate system to obtain the 3D spatial coordinates of the clamping points. The spatial accuracy requirement is that the spatial error of the acquired coordinate values ​​is no greater than 1 mm, meeting the accuracy requirements for building component assembly. A positive value for the final end distance correction is defined as the final assembly end distance being greater than the current clamping point distance, with the corresponding adjustment direction being that the two clamping points move away from each other along the component axis to increase the end distance. A negative value is defined as the final assembly end distance being less than the current clamping point distance, with the corresponding adjustment direction being that the two clamping points move towards each other along the component axis to decrease the end distance. A zero value indicates that the current clamping point distance is consistent with the final assembly end distance, and no end distance adjustment is required. The specific timing for component release is as follows: release is performed immediately after both moving manipulators reach the final gripping point and meet the positioning criteria. The release sequence is simultaneous release of the end clamps of both manipulators, without any sequential order, to avoid component attitude deviation caused by unilateral release. Temporary positioning protection measures include using assembled positioning clamps to fix the component to the mounting bracket after release. The clamping force is 10% to 20% of the component's weight to prevent slippage and avoid generating additional clamping forces. If the final end-distance correction exceeds the available axial travel of the moving manipulators, the emergency adjustment plan is as follows: first, re-acquire the available axial travel of the moving manipulators to confirm whether the judgment deviation is caused by acquisition error. If there is indeed no travel margin, adjust the base attitude of both manipulators through the main controller to increase the available axial travel. After adjustment, recalculate the allocation coefficient and the allocation value of the final end-distance correction. If attitude adjustment still cannot meet the requirements, use manual assisted pushing to assist the manipulators in completing the final end-distance adjustment. The force value of manual pushing is monitored by a force sensor and does not exceed the allowable thrust of the component material. In addition, after the mobile operating arm releases the component, the safe avoidance distance between the component and the component is that the minimum straight-line distance between the end clamp of the operating arm and the surface of the component is not less than 50 mm, so as to avoid collision between the operating arm and the component when the operating arm resets; the action requirements are that after the release is completed, the two operating arms move away from the component radially along the component under the command of the main controller. After reaching the safe avoidance distance, they reset along the construction path towards the construction starting point. The reset speed is 0.1 m / s to 0.2 m / s.

[0136] It should be noted that this invention determines the assembly target clamping point based on the design geometry requirements of building assembly, calculates the final assembly end distance, obtains the final end distance correction amount by the difference between the current clamping point and the original distance, distributes the correction amount to the clamping points at both ends using the allocation coefficient from the construction process, and transforms a maintenance positive amount into the final gripping point target in three-dimensional space by combining the component's axial unit vector. This drives the moving manipulator to precisely adjust to the assembly position before releasing the component, achieving a smooth transition from the construction state to the assembly state. Using the assembly design benchmark as the core ensures that the final installation end distance of the component precisely matches the design requirements of the building construction, guaranteeing assembly geometric accuracy. Using the previous allocation coefficient ensures that the final end distance adjustment is consistent with previous actions, avoiding fluctuations in the component's posture due to sudden changes in the adjustment ratio. The generation of the three-dimensional gripping point target adapts to the positional requirements of spatial assembly, making the manipulator adjustment more precise. Releasing the component after adjustment avoids component rebound caused by residual axial internal forces, ensuring the stability of the assembly position. Synchronous release and temporary positioning protection measures can prevent slippage of components after release. All operation and verification rules are clear, which is convenient for on-site implementation. Accuracy verification can also ensure that the assembly quality meets the engineering requirements.

[0137] In one embodiment of the present invention, a multi-robot collaborative and human-robot hybrid construction system for building robots includes: a first module, controlling two mobile manipulators to clamp both ends of a component in a closed-chain collaborative gripping manner, and establishing an initial end distance and discrete axial coordinates of the component; a second module, acquiring the current gripping point distance between the two mobile manipulators, and calculating the end distance increment in conjunction with the initial end distance; a third module, acquiring the discrete coordinates of the component's centerline based on the discrete axial coordinates of the component and constructing a reference line, and extracting the displacement of the discrete point coordinates relative to the reference line along the gravity direction to generate a deflection sequence; and a fourth module, calculating the deflection-induced axial geometric locking index through the deflection sequence, converting it into equivalent geometric elongation, calculating the difference between the end distance increment and the geometric elongation, and converting it into an axial internal force estimate. The system is divided into several modules: Module 5, Module 6, Module 7, Module 8, Module 9, Module 10, Module 10, Module 11, Module 21, Module 12, Module 13, Module 14, Module 15, Module 16, Module 17, Module 18, Module 19, Module 12, Module 16, Module 18, Module 19, Module 12, Module 13, Module 14, Module 15, Module 16, Module 17, Module 18, Module 19 ...

[0138] It should be noted that, for example, when transporting and installing a 5-meter-long prefabricated steel component in a narrow corridor at a construction site, the basic parameters set include: component length parameter 5000 mm, number of discrete points parameter 21, component axial stiffness parameter 2000 N / mm, component bending stiffness parameter 1.2 × 10⁷ N × mm², total construction path length 8 meters, discrete into 4 fixed work positions at 2-meter intervals, and the final assembly end distance of the installation position design is 5000 mm. The final output results are as follows: Physical placement results: The 5-meter-long precast steel component was smoothly moved to the preset position of the installation bracket. Temporary fixation of the component to the installation bracket was achieved using prefabricated positioning clamps. The clamping force was 15% of the component's self-weight. Both ends of the component were perfectly aligned with the preset reference holes at the installation position without any offset. There was no significant axial springback in the component, and bending deformation only occurred in the middle of the component, with no localized bending. Subsequent permanent fixing construction can proceed directly. Assembly accuracy verification results: Verification was performed using a three-dimensional laser positioning instrument and a laser rangefinder. The actual assembly end distance of the component was measured to be 4999 mm, and the difference from the design value was within the allowable deviation range. The absolute value of the three-dimensional spatial coordinate deviation of the actual installation position of the component was less than 3 mm. The maximum deflection value at the middle of the component was measured to be 1.2 mm, and the axial internal force was measured to be 0 N. All verification indicators met the building requirements. General requirements for the assembly of precast components; quantitative data results for the entire construction process: generating complete parameter record ledgers for 4 fixed workstations, including the three-dimensional spatial coordinates of the first and second end clamping points of each workstation, the specific values ​​of the deflection sequence of 21 discrete points, the estimated values ​​of the axial internal forces and end distance corrections for each workstation, the support positions and corresponding support force values ​​for each workstation, such as the support position of workstation 2 being 2500 mm and the support force being 230 N, and the support position of workstation 3 being 2450 mm and the support force being 210 N; at the same time, the ledger includes the final end distance correction of 1 mm and the distribution values ​​at both ends, with the final end distance correction of 0.4 mm for the first end and 0.6 mm for the second end, as well as the three-dimensional spatial coordinates of the first, second, and final grabbing point targets; all parameters are labeled with the corresponding acquisition time, acquisition equipment, and calculation process, and the ledger can be directly used for the quality acceptance of this construction process.

[0139] It should be noted that the data acquisition of this invention relies on various on-site sensing devices and measuring instruments. All acquisition devices are connected to a unified industrial control network at the construction site. The acquisition actions are uniformly triggered and scheduled by the main controller, with no independent acquisition operations. Visual sensors, laser positioning sensors, and spatial positioning systems acquire the three-dimensional spatial coordinates of the clamping points at both ends of the moving manipulator; axial travel sensors acquire the available axial travel of the two moving manipulators along the unit vector direction of the component's axial direction; a line laser three-dimensional point cloud scanner acquires the three-dimensional spatial coordinates of discrete points on the component's centerline; force sensors acquire the actual support force values ​​during manual support; and laser rangefinders and three-dimensional laser positioning instruments acquire the construction path dimensions, the coordinates of the assembly target clamping points, and the geometric verification parameters after assembly. The main controller sends synchronous acquisition trigger commands to each acquisition device to achieve synchronous acquisition of multiple devices. The acquisition frequency of dynamic parameters such as clamping point coordinates is 10 Hz to 50 Hz, which is adjusted according to the degree of constraint in narrow space construction. The high frequency is used in the constrained area and the low frequency is used in the straight area. Static parameters such as the discrete point coordinates of the component centerline and the clamping point coordinates of the assembly target are acquired once before construction. If the posture of the component changes unexpectedly during construction, the main controller triggers a re-acquisition. After the support force, deflection sequence and other parameters are calculated at each fixed station, they are re-measured and acquired at a frequency of once every 5 seconds.

[0140] It should be noted that the data transmission of this invention adopts the Industrial Ethernet protocol to achieve end-to-end data interaction. The transmission process uses the main controller as the relay core and is divided into two transmission directions: from field devices to the main controller and from the main controller to the execution terminal. All transmitted data adopts a standardized numerical format without custom encoding formats, ensuring the stability and compatibility of transmission. The transmission from field devices to the main controller involves various acquisition devices and measuring instruments uploading pre-processed valid data to the main controller according to the synchronous transmission command of the main controller. The transmission delay is controlled within 10 milliseconds to avoid asynchrony between calculation and construction actions due to transmission delay. The clamping point coordinates are transmitted in a three-dimensional spatial numerical format, the deflection sequence is transmitted in a one-dimensional array format, and stiffness parameters, stroke parameters, etc. are transmitted in a single numerical format. All data is accompanied by the acquisition time and device identification information. The transmission from the main controller to the execution terminal involves the main controller transmitting calculated parameters such as the target grasping point, support position, support force, and end distance correction to the servo control system of the mobile manipulator and the manual intervention terminal, respectively. The mobile manipulator receives the geometric and motion parameters, while the manual intervention terminal receives the support position and support force parameters. During transmission, key motion parameters are repeatedly sent for verification to ensure accurate reception by the execution terminal. The main controller monitors the transmission status of each device in real time. If data loss or transmission errors occur, a retransmission command is immediately sent to the corresponding acquisition device or execution terminal. If three consecutive retransmissions fail, the main controller will trigger an audible and visual alarm on-site, suspending the construction process until the transmission fault is resolved.

[0141] It should be noted that data transmission is uniformly executed by the main controller. All transmitted parameters and instructions are the results of the main controller's end-to-end calculations based on the collected data. The control of the mobile manipulator is solely based on the instructions transmitted by the main controller, achieving closed-loop coordinated movement of the two manipulators. The manual intervention terminal only receives reference parameters and does not participate in the control logic of the manipulator. The controller transmits motion control parameters such as the target coordinates of the grasping point, the end distance correction allocation value, the axial unit vector, and the movement speed to the mobile manipulator servo control system, and also transmits the numerical standard for determining the position. It transmits intervention reference parameters such as the support position, support force value, and force direction of each fixed station to the manual intervention terminal. The transmitted actions are triggered according to the progress rhythm of the fixed stations, and a data transmission is completed before the construction of each station. After receiving instructions from the main controller, the servo control system of the mobile manipulator drives the joint motor of the manipulator to move at a preset speed. The direction of movement is strictly along the unit vector direction of the component's axial direction, achieving precise adjustment of the gripping point position. The movement of the two manipulators is constrained by the synchronous movement command sent by the main controller, ensuring that they start simultaneously, run at the same speed, and stop synchronously, avoiding component posture deviation caused by single-arm movement. During the movement, the position sensor of the manipulator feeds back the actual coordinates of the gripping point collected in real time to the main controller, which continuously compares the actual coordinates with the target coordinates of the gripping point. When the spatial error between the actual coordinates of the gripping point and the target coordinates is less than 1 mm, and the duration of the manipulator's motion mechanism in a stationary state is not less than 500 milliseconds, the main controller sends an action termination command to the servo control system, and the manipulator stops moving and maintains the current gripping state. If the manipulator moves 95% of the available axial travel distance but still does not reach the target position, the main controller immediately triggers a travel limit protection command, the manipulator stops moving and sends a travel limit signal back to the main controller. After being advanced to the installation position, the main controller issues the target coordinates of the last gripping point and the allocation value of the last end distance correction. The operating arm completes the last position adjustment according to the above motion control logic. After the positioning judgment criteria are met, the main controller issues a release command to both operating arms simultaneously. The end clamps of the operating arms complete the release action at the same time. After release, an avoidance motion command is issued to drive the operating arm to move away from the component radially until a safe avoidance distance is reached.

[0142] It should be noted that the data management of the present invention is realized by relying on the industrial-grade local storage module supporting the main controller and the cloud data server. It adopts the combination of local storage and cloud storage. All construction-related data are stored in a structured classification manner to ensure that the data can be retrieved, traced, and reused. There is no manual entry during the management process, and all are automatically completed by the system. The local storage module is deployed beside the main controller at the construction site, storing the real-time collected data during this construction process, the calculation parameters of each fixed work station, the motion control parameters of the robotic arm, and the status monitoring data during the construction process, ensuring the real-time retrieval of data during the construction process; the cloud data server is deployed in a remote computer room, receiving the full-process construction data uploaded by the local storage module, and at the same time storing the basic parameters of the components before construction, the assembly verification data after construction, and the historical construction data of the same type of components, realizing the long-term retention of data. All data is divided into four categories according to the construction process: initial clamping data, station advancement data, final assembly data, and assembly verification data, and into four categories according to the data type: geometric parameters, mechanical parameters, motion parameters, and status parameters; the data is stored in the format of a structured data table, and the data table contains information such as parameter name, value, dimension, collection time, collection device, calculation process, execution result, etc. The parameters of each fixed work station are separately archived according to the station number, forming an independent station data ledger. Data access is divided into three levels: on-site operation permission, technical review permission, and management viewing permission; the on-site operation permission can only retrieve the real-time data and station parameters of this construction, without modification permission; the technical review permission can retrieve the historical construction data of the same type of components and the full-process data of this construction, and can export the data, without deletion permission; the management viewing permission can access all stored data, has the permission to export and archive and organize the data, without the permission to modify and delete randomly. The local storage module completes data incremental backup at a frequency of once every 30 minutes, and completes full backup after the construction is completed; after receiving the local backup data, the cloud data server automatically completes off-site disaster recovery backup to avoid data loss caused by hardware failures.

[0143] It should be noted that the setting of the interval and the threshold size is for the convenience of comparison. Among them, the size of the threshold depends on the amount of sample data and the base quantity set by those skilled in the art for each group of sample data, as long as it does not affect the proportional relationship between the parameter and the quantified value. And the above formulas are all calculations that remove the dimension and take their numerical values. The formulas are all obtained by software simulation of a large amount of collected data to get a formula closest to the real situation. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0144] The embodiments of this example have been described above. However, this example is not limited to the specific implementation methods described above. The specific implementation methods described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms based on the guidance of this example, and all of them are within the protection scope of this example.

Claims

1. A method for multi-robot collaborative and human-robot hybrid construction, characterized in that, Includes the following steps: Step S1: Control the two moving manipulators to clamp the two ends of the component in a closed-chain cooperative gripping manner, and establish the initial end distance and discrete coordinates of the component axis; Step S2: Obtain the current gripping point distance between the two moving manipulators, and calculate the end distance increment based on the initial end distance. Step S3: Obtain the coordinates of discrete points on the center line of the component based on the discrete coordinates of the component's axial direction and construct a reference line. Extract the displacement of the discrete point coordinates relative to the reference line along the direction of gravity to generate a deflection sequence. Step S4: Calculate the deflection-induced axial geometric locking index through the deflection sequence, convert it into geometric elongation, calculate the difference between the end distance increment and the geometric elongation, and convert it into an axial internal force estimate. Step S5: Calculate the end distance correction amount based on the axial internal force estimate, allocate the end distance correction amount according to the ratio of the available axial stroke of the two moving manipulators, and generate the updated gripping point target. Step S6: Find the extreme point of deflection in the deflection sequence as the support location, and calculate the support force corresponding to the support location based on the stiffness parameters of the component. Step S7: Discretize the construction path into multiple fixed work positions. At each fixed work position, execute steps S2 to S6 in sequence. Use the grab point target to drive the two mobile manipulators to advance, and use the support position and support force as deterministic intervention conditions for human-machine hybrid construction. Step S8: After advancing to the installation position, generate the final end distance correction amount according to the final assembly end distance, drive the two moving operating arms to move and release the component; Step S1 is as follows: Control two mobile manipulators to grip both ends of the component in a closed-chain collaborative gripping manner, and determine the first end gripping point and the second end gripping point; The initial end distance is obtained by calculating the Euclidean norm of the displacement vector formed by the first end clamping point and the second end clamping point; the axial unit vector of the component is obtained by dividing the displacement vector formed by the second end clamping point and the first end clamping point by the initial end distance. Determine the component length parameter and the discrete point number parameter, calculate the ratio of the value after subtracting one from the discrete point index to the value after subtracting one from the discrete point number parameter, multiply the ratio by the component length parameter to obtain the component's axial discrete coordinates; calculate the ratio of the component length parameter to the value after subtracting one from the discrete point number parameter to obtain the component's axial discrete step length. Step S4 is as follows: The difference between adjacent deflection values ​​is calculated and divided by the axial deviation of the component to obtain the deflection dispersion slope; the weighted sum of the square of the deflection dispersion slope and the axial deviation of the component is calculated to obtain the slope square integral. The first product result is obtained by multiplying the axial stiffness parameter and the length parameter of the component with the integral of the square of the slope. The second product result is obtained by multiplying the square of twice the constant of pi with the bending stiffness parameter of the component. The deflection-induced axial geometric locking index is obtained by dividing the first product result by the second product result. The conversion ratio coefficient is obtained by multiplying the square of the pi constant by the bending stiffness parameter of the component, and then dividing by the product of the axial stiffness parameter and the length parameter of the component. The geometric elongation is obtained by multiplying the deflection-induced axial geometric locking index by the conversion scaling factor; The difference between the end distance increment and the geometric elongation is calculated. The difference is then multiplied by the axial stiffness parameter of the member and divided by the member length parameter to obtain the axial internal force estimate.

2. The construction robot multi-robot collaborative and human-robot hybrid construction method according to claim 1, characterized in that, Obtain the first and second clamping points of the two mobile manipulators at the current moment; calculate the Euclidean norm of the displacement vectors of the two clamping points formed by the first and second clamping points to obtain the current clamping point distance; The end distance increment is obtained by calculating the difference between the current clamping point distance and the initial end distance.

3. The construction robot multi-robot collaborative and human-robot hybrid construction method according to claim 1, characterized in that, The coordinates of discrete points on the component's centerline are obtained based on the component's axial discrete coordinates. Using the line connecting the endpoints determined by the first clamping point and the second clamping point as a reference line, the ratio of the component's axial discrete coordinates to the component's length parameter is calculated to generate a normalized ratio. The normalized ratio is then multiplied by the displacement vector of the two clamping points formed by the first and second clamping points, and the product result is superimposed onto the coordinates of the discrete point of the reference line established by the first clamping point. Determine the unit vector of gravity direction, calculate the displacement vector of the discrete point coordinates of the component centerline relative to the discrete point coordinates of the reference line, and generate the deflection sequence by projecting the numerical values ​​along the unit vector of gravity direction.

4. A construction robot multi-robot collaborative and human-robot hybrid construction method according to claim 1, characterized in that, The proportionality coefficient of the change in axial internal force to end distance is obtained by dividing the component length parameter by the component axial stiffness parameter. The proportionality coefficient of the change in axial internal force to end distance is multiplied by the axial internal force estimate and the negative value is obtained to obtain the end distance correction amount. Obtain the first available axial travel and the second available axial travel of the two moving manipulators along the unit vector direction of the component axis, and obtain the allocation coefficient by dividing the first available axial travel by the sum of the first available axial travel and the second available axial travel; Multiply the allocation coefficient by the end distance correction amount to obtain the end distance correction amount allocated to the first end clamping point; The difference between the unit value and the allocation coefficient is multiplied by the end distance correction amount to obtain the end distance correction amount allocated to the second end clamping point; The first updated gripping point target is obtained by subtracting the end distance correction amount allocated to the first end gripping point from the product of the component axial unit vector. The second updated gripping point target is obtained by adding the end distance correction amount assigned to the second end gripping point to the product of the component axial unit vector.

5. A construction robot multi-robot collaborative and human-robot hybrid construction method according to claim 1, characterized in that, In the deflection sequence, identify the term with the largest absolute deflection value and obtain the parameters corresponding to the term with the largest absolute deflection value; The support position is obtained by extracting the coordinate value corresponding to the parameter with the largest absolute value of deflection from the discrete axial coordinates of the component. Obtain the deflection term corresponding to the support position in the deflection sequence as the deflection value at the support position; Calculate the product of forty-eight and the bending stiffness parameter of the member, and divide the product by the cube of the member length parameter to obtain the proportionality coefficient; The supporting force at the support location is obtained by multiplying the proportional coefficient by the deflection value at the support location.

6. A construction robot multi-robot collaborative and human-robot hybrid construction method according to claim 1, characterized in that, The construction path is discretized into multiple fixed workstations, and the number of fixed workstations is recorded as the total number of fixed workstations. The current fixed workstation is represented by the fixed workstation variable. At the location determined by the fixed workstation variable, calculations are performed sequentially to obtain the first updated gripping point target, the second updated gripping point target, the support position, and the support force; The first and second updated grab point targets are generated by a computational link formed by the end distance increment, deflection sequence, deflection-induced axial geometric locking index, geometric elongation, axial internal force estimation, end distance correction, and the first and second available axial travel. The support location and support force are generated by a computational link formed by the deflection sequence, the bending stiffness parameters of the component, and the length parameters of the component. At the fixed workstation location, the first updated gripping point target and the second updated gripping point target are used to drive the two mobile manipulators to advance. The support position and support force are used to execute the deterministic intervention conditions for human-machine hybrid construction. After the advancement is completed, the first end gripping point of the next fixed workstation is updated to the first updated gripping point target, and the second end gripping point of the next fixed workstation is updated to the second updated gripping point target.

7. A construction robot multi-robot collaborative and human-robot hybrid construction method according to claim 1, characterized in that, Determine the first end assembly target clamping point and the second end assembly target clamping point; calculate the Euclidean norm of the three-dimensional spatial displacement vector formed by the second end assembly target clamping point and the first end assembly target clamping point to obtain the final assembly end distance; The distance to the current clamping point is obtained by calculating the Euclidean norm of the three-dimensional spatial displacement vector formed by the second clamping point and the first clamping point; The final end distance correction amount is obtained by calculating the difference between the final assembly end distance and the current clamping point distance. Obtain the first and second available axial travel distances of the two moving manipulators along the unit vector direction of the component axis, and calculate the allocation coefficient by dividing the first available axial travel distance by the sum of the first and second available axial travel distances. Multiply the allocation coefficient by the final end distance correction to obtain the final end distance correction to be allocated to the first end clamping point; multiply the value obtained by subtracting the allocation coefficient from the unit value by the final end distance correction to obtain the final end distance correction to be allocated to the second end clamping point. The first final gripping point target is obtained by subtracting the final end distance correction amount allocated to the first end gripping point from the product of the component axial unit vector. The second final gripping point target is obtained by adding the final end distance correction amount allocated to the second end gripping point to the product of the component axial unit vector. Drive two mobile manipulators to move and release the component.

8. A multi-robot collaborative and human-robot hybrid construction system, characterized in that, The method for multi-robot collaborative and human-robot hybrid construction as described in any one of claims 1 to 7 includes: The first module controls two mobile manipulators to grip the two ends of the component in a closed-chain collaborative gripping manner, and establishes the initial end distance and discrete coordinates of the component's axial direction. The second module obtains the current gripping point distance between the two mobile manipulators and calculates the end distance increment based on the initial end distance. The third module obtains the coordinates of discrete points on the centerline of the component based on the discrete coordinates of the component's axial direction and constructs a reference line. It then extracts the displacement of the discrete point coordinates relative to the reference line along the direction of gravity to generate a deflection sequence. The fourth module calculates the deflection-induced axial geometric locking index through the deflection sequence, converts it into geometric elongation, calculates the difference between the end distance increment and the geometric elongation, and converts it into an axial internal force estimate. The fifth module calculates the end distance correction based on the axial internal force estimate, allocates the end distance correction according to the ratio of the available axial travel of the two moving manipulators, and generates an updated grab point target. The sixth module identifies the extreme deflection points from the deflection sequence as support locations and calculates the support force corresponding to the support locations based on the component stiffness parameters. The seventh module discretizes the construction path into multiple fixed work stations. At each fixed work station, the second to sixth modules are executed sequentially. The grab point target drives the two mobile manipulators to advance, and the support position and support force are used as deterministic intervention conditions for human-machine hybrid construction. The eighth module, after being advanced to the installation position, generates the final end distance correction amount according to the final assembly end distance, drives the two moving manipulators to move and release the component.

Images