Methods, equipment, electronic devices, and storage media for the coordinated assembly and welding of steel structure components by multiple robots.

Abstract

This invention provides a method, apparatus, electronic device, and storage medium for the coordinated assembly and welding of steel structural components by multiple robots. This coordinated method for the assembly and welding of steel structural components collects steel structural components awaiting assembly and welding, assembles assembly and welding tasks, the assembly and welding task set comprises multiple assembly and welding subtask sets, trains multiple assembly and welding subtask sets, obtains the time sequence of multiple assembly and welding subtask sets, the assembly and welding subtask set includes picking and welding tasks, allocates picking and welding tasks to different robots, and controls different robots to perform assembly and welding of steel structural components awaiting assembly and welding according to the time sequence of multiple assembly and welding subtask sets.By implementing autonomous allocation and coordinated work of multiple robots for multiple tasks, this invention reduces the overall welding process time, improves welding efficiency and quality, and simultaneously reduces labor costs and human labor intensity.

Description

Method, device, electronic equipment and medium for multi-robot collaborative welding of steel structures Technical Field

[0001] The present invention relates to the technical field of steel structure welding, and in particular to a method, device, electronic equipment and medium for collaborative welding of steel structures by multiple robots. Background Art

[0002] In recent years, the proportion of steel structure buildings has gradually increased, the forms of building steel structures have become increasingly diverse, and the demand for steel structural parts has increased. The steel structure includes the main frame and connecting components, and the assembly process is completed through the connection and welding of the main frame and connecting components. Among them, the main frame is welded steel and rolled steel columns, etc., and the connecting components include ribs, supporting brackets and connecting brackets.

[0003] At present, steel structures are mainly welded manually. The general welding order is to weld the corbels and stiffeners first, and then weld them together with the steel columns. Repeated operations such as grabbing, spot welding, and flipping are required. The operation is complicated and difficult, especially for large-sized and large-tonnage steel components. During the welding process, welding at corners and gaps is difficult, and auxiliary lifting and flipping are laborious, which makes the overall welding process time-consuming, resulting in low welding efficiency and insufficient output.

[0004] Therefore, there is an urgent need for a multi-robot collaborative assembly welding method to realize the autonomous allocation and collaborative operation of multiple robots and multiple tasks, shorten the duration of the overall welding process, improve welding efficiency, and reduce labor costs and labor intensity.

[0005] Summary of the Invention

[0006] In view of the above-mentioned shortcomings of the prior art, the present application provides a method, device, electronic product and storage medium for multi-robot collaborative assembly and welding of steel structures to solve the above-mentioned technical problems.

[0007] According to one aspect of an embodiment of the present application, a multi-robot collaborative assembly and welding method for steel structural parts is provided, comprising: obtaining steel structural parts to be assembled and a welding task set, the welding task set including a plurality of welding sub-task sets; training the plurality of welding sub-task sets to obtain a temporal sequence of the plurality of welding sub-task sets, the welding sub-task set including a grasping task and a welding task; and assigning the grasping task and the welding task to different robots to control the different robots to assemble and weld the steel structural parts to be assembled according to the temporal sequence of the plurality of welding sub-task sets.

[0008] In one embodiment of the present application, the process of training multiple welding subtask sets to obtain the time sequence of multiple welding subtask sets includes: obtaining the current welding state of the steel structure to be welded; taking the current welding state of the steel structure to be welded as the initialization welding state, and based on the initialization welding state, respectively calculating the probability of selecting all the welding subtask sets, and determining the next welding subtask set based on the probabilities of all the welding subtask sets, until all the welding subtask sets are selected, and obtaining the time sequence of multiple welding subtask sets.

[0009] In one embodiment of the present application, based on the initialized welding state, the process of respectively calculating the probability of selecting all the welding subtask sets includes: inputting the initialized welding state into a convolutional neural network model to obtain a feature matrix between the initialized welding state and all the welding subtask sets; inputting the feature matrix between the initialized welding state and all the welding subtask sets into a first fully connected network model, and obtaining a one-dimensional feature vector through conversion by the first fully connected network model; inputting the one-dimensional feature vector into a normalized exponential function, and calculating and outputting the probability of selecting all the welding subtask sets through the normalized exponential function.

[0010] In one embodiment of the present application, the process of determining the next welding subtask set based on the probabilities of all the welding subtask sets until the selection of all the welding subtask sets is completed includes: based on the initialization welding state, selecting the maximum probability value among the probabilities of all the welding subtask sets, and using the welding subtask set corresponding to the maximum probability value as the next welding subtask set; using the welding state corresponding to the next welding subtask set as the initialization welding state, and repeatedly selecting the maximum probability value among the probabilities of all the welding subtask sets until the selection of all the welding subtask sets is completed.

[0011] In one embodiment of the present application, after obtaining the timing sequence of multiple welding subtask sets, the method further includes: sorting multiple welding states corresponding to the welding subtask sets according to the timing sequence of the multiple welding subtask sets to form a timing path, and using the timing relationship between the multiple welding states as a state transition link in the timing path; jointly determining the reward value of each state transition link by using the distance value between the end point of the current welding state and the starting point of the next welding state in each state transition link and the weight corresponding to the distance value, the robot joint angle from the current welding state to the next welding state and the weight corresponding to the robot joint angle, the positioner rotation angle from the current welding state to the next welding state and the weight corresponding to the positioner rotation angle, the collision from the current welding state to the next welding state and the weight corresponding to the collision; determining the timing path discount reward according to the reward value of each state transition link and the discount coefficient corresponding to each state transition link; Input a convolutional neural network model, output the feature matrix of the current welding state, input the next welding subtask set corresponding to each state transition link into the second fully connected network model, output the feature matrix of the next welding subtask set, perform feature association between the feature matrix of the current welding state and the feature matrix of the next welding subtask set, obtain the association matrix between the current welding state and the next welding subtask set, input the association matrix into the third fully connected network model, and obtain the value of each state transition link; calculate the difference between the value of the starting state transition link and the value of the ending state transition link in the timing path and the discount reward of the timing path, and take the absolute value of the difference to obtain the timing path value; with the goal of minimizing the timing path value, update the parameters of the convolutional neural network model and the second fully connected network model and the third fully connected network model to obtain the first updated convolutional neural network model, the second updated fully connected network model, and the third updated fully connected network model.

[0012] In one embodiment of the present application, after obtaining the updated convolutional neural network model, the method further includes: multiplying the value of each state transition link by the probability of the corresponding state transition link to obtain the expected value of each state transition link; adding the expected values ​​of all the state transition links to obtain the total expected value; with the goal of maximizing the total expected value, updating the parameters of the first updated convolutional neural network model and the first fully connected network model to obtain the second updated convolutional neural network model and the first updated fully connected network model; based on the second updated convolutional neural network model, the first updated fully connected network model, the second updated fully connected network model, and the third updated fully connected network model, repeatedly training the multiple assembly and welding subtask sets until the number of training times for the multiple assembly and welding subtask sets reaches a preset number threshold, and outputting the time sequence of the multiple assembly and welding subtask sets.

[0013] In one embodiment of the present application, before obtaining the steel structures to be welded and the welding task set, it includes: obtaining the equipment required in the process of welding the steel structures and the corresponding equipment information; constructing a three-dimensional simulation model based on the equipment and equipment information, so as to control different robots to perform welding operations through the three-dimensional simulation model.

[0014] According to one aspect of an embodiment of the present application, a multi-robot collaborative assembly and welding device for steel structures is provided, including: a steel structure acquisition module, used to acquire steel structures to be assembled and a welding task set, wherein the welding task set includes multiple welding sub-task sets; a timing determination module, used to train multiple welding sub-task sets to obtain the timing sequence of multiple welding sub-task sets, wherein the welding sub-task set includes grasping tasks and welding tasks; a task allocation module, used to allocate the grasping tasks and the welding tasks to different robots, so as to control different robots to assemble and weld the steel structures to be assembled according to the timing sequence of the multiple welding sub-task sets.

[0015] According to one aspect of an embodiment of the present application, an electronic device is provided, comprising: one or more processors; a storage device for storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the electronic device implements the multi-robot collaborative assembly and welding method for steel structures as described above.

[0016] According to one aspect of an embodiment of the present application, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor of a computer, the computer executes the above-mentioned multi-robot collaborative assembly and welding method for steel structures.

[0017] Beneficial effects of the present invention: The present invention obtains steel structures to be welded and welding task sets, trains multiple welding sub-task sets, obtains the time sequence of multiple welding sub-task sets, and assigns grasping tasks and welding tasks to different robots to control different robots to assemble and weld the steel structures to be welded according to the time sequence of multiple welding sub-task sets, thereby realizing the autonomous allocation and collaborative operation of multiple robots and multiple tasks, reducing the duration of the overall welding process, improving welding efficiency and welding quality, and reducing labor costs and labor intensity, thereby further promoting the coordinated development of digitalization, intelligence and greening of the steel structure industry.

[0018] It should be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] The accompanying drawings are incorporated into and constitute a part of the specification, illustrating embodiments consistent with the present application and, together with the specification, serving to explain the principles of the present application. It is obvious that the drawings described below are merely some embodiments of the present application, and a person of ordinary skill in the art can derive other drawings based on these drawings without inventive effort. In the drawings:

[0020] FIG1 is a schematic diagram of an exemplary system architecture shown in an exemplary embodiment of the present application;

[0021] FIG2 is a flow chart of a multi-robot collaborative welding method for steel structures shown in an exemplary embodiment of the present application;

[0022] FIG3 is a schematic diagram of a layout of a robotic assembly welding system according to an exemplary embodiment of the present application;

[0023] FIG4 is a flow chart of a method for training a set of robot welding subtasks according to an exemplary embodiment of the present application;

[0024] FIG5 shows a block diagram of a multi-robot collaborative welding device for steel structures suitable for implementing an embodiment of the present application;

[0025] FIG6 shows a schematic structural diagram of a computer system suitable for implementing an electronic device according to an embodiment of the present application. DETAILED DESCRIPTION

[0026] Exemplary embodiments will be described in detail herein, with examples illustrated in the accompanying drawings. When the following description refers to the drawings, identical numerals in different figures represent identical or similar elements unless otherwise indicated. The embodiments described in the following exemplary embodiments are not intended to represent all embodiments consistent with the present application. Rather, they are merely examples of apparatus and methods consistent with certain aspects of the present application, as detailed in the appended claims.

[0027] The block diagrams shown in the accompanying drawings are merely functional entities and do not necessarily correspond to physically separate entities. That is, these functional entities may be implemented in software, in one or more hardware modules or integrated circuits, or in different networks and / or processor devices and / or microcontroller devices.

[0028] The flowcharts shown in the accompanying drawings are for illustrative purposes only and do not necessarily include all contents and operations / steps, nor must they be executed in the order described. For example, some operations / steps may be decomposed, while others may be combined or partially combined. Therefore, the actual execution order may vary depending on the actual situation.

[0029] In this application, "plurality" refers to two or more. "And / or" describes the relationship between related objects, indicating that three possible relationships exist. For example, "A and / or B" can mean: A exists alone, A and B exist simultaneously, and B exists alone. The character " / " generally indicates that the related objects are in an "or" relationship.

[0030] The technical solutions of the embodiments of the present application involve related technologies such as welding of steel structures, and are specifically described through the following embodiments:

[0031] FIG1 is a schematic diagram of an exemplary system architecture shown in an exemplary embodiment of the present application.

[0032] As shown in FIG1 , the system architecture may include a steel structure storage device 101 and a computer device 102. The computer device 102 may be at least one of a desktop graphics processing unit (GPU) computer, a GPU computing cluster, a neural network computer, and the like. Relevant technicians may use the computer device 102 to train multiple welding subtask sets to obtain the temporal sequence of the multiple welding subtask sets, and assign the grasping tasks and welding tasks to different robots to control the different robots to perform assembly welding according to the temporal sequence of the multiple welding subtask sets. The steel structure storage device 101 is used to store the steel structures to be welded and the welding task sets. In this embodiment, the steel structure storage device 101 uses a read-only memory (ROM) or a random access memory (RAM) to store the steel structures to be welded and the welding task sets, and provides them to the computer device 102 for processing.

[0033] Schematically, after obtaining the steel structure parts to be welded and the welding task set from the steel structure parts storage device 101, the computer device 102 trains multiple welding sub-task sets to obtain the time sequence of multiple welding sub-task sets, and assigns the grasping tasks and welding tasks to different robots to control different robots to perform assembly and welding according to the time sequence of multiple welding sub-task sets, thereby realizing the autonomous allocation and collaborative operation of multiple robots and multiple tasks, reducing the duration of the overall welding process, improving welding efficiency and welding quality, and reducing labor costs and labor intensity, thereby further promoting the coordinated development of digitalization, intelligence and greening of the steel structure industry.

[0034] It should be noted that the multi-robot collaborative assembly and welding method for steel structures provided in the embodiment of the present application is generally executed by the computer device 102 , and accordingly, the multi-robot collaborative assembly and welding device for steel structures is generally provided in the computer device 102 .

[0035] The following is a detailed description of the implementation details of the technical solution of the embodiment of the present application:

[0036] FIG2 is a flowchart of a multi-robot collaborative welding method for steel structures, shown in an exemplary embodiment of the present application. The multi-robot collaborative welding method for steel structures can be executed by a computing device, which can be the computer device 102 shown in FIG1 . Referring to FIG2 , the multi-robot collaborative welding method for steel structures includes at least steps S210 to S230, which are described in detail as follows:

[0037] In step S210, a set of steel structural parts to be welded and welding tasks are obtained.

[0038] In one embodiment of the present application, the welding task set includes multiple welding sub-task sets, and the multiple welding sub-task sets correspond to multiple welding steel structural parts. The steel structural parts to be welded include ribs and partitions, etc. For example, the welding sub-task set corresponding to the ribs includes grabbing the ribs, spot welding assembly, arc welding, rotation of auxiliary mechanism workstations, etc.

[0039] In step S220 , multiple assembly and welding subtask sets are trained to obtain a temporal sequence of the multiple assembly and welding subtask sets.

[0040] In this embodiment, the process of training multiple welding subtask sets to obtain the temporal sequence of the multiple welding subtask sets includes: (1) obtaining the current welding state of the steel structure to be welded; (2) taking the current welding state of the steel structure to be welded as the initialization welding state, and based on the initialization welding state, respectively calculating the probability of selecting all welding subtask sets, and determining the next welding subtask set based on the probabilities of all welding subtask sets, until all welding subtask sets are selected, thereby obtaining the temporal sequence of the multiple welding subtask sets.

[0041] In this embodiment, the assembly and welding sub-task set includes grasping tasks and welding tasks. When the welding sub-task combination includes grasping the rib plate, spot welding assembly, arc welding, and rotation of the auxiliary mechanism station, grasping the rib plate is the grasping task, spot welding assembly and arc welding are the welding tasks, and the rotation of the auxiliary mechanism station is the rotation task of the positioner.

[0042] In step S230 , the grasping task and the welding task are assigned to different robots to control the different robots to assemble and weld the steel structure to be welded according to the time sequence of the plurality of assembling and welding subtask sets.

[0043] In this embodiment, by training multiple assembly and welding sub-task sets, the time sequence of multiple assembly and welding sub-task sets is obtained, and the grasping tasks and welding tasks are assigned to different robots to control different robots to perform assembly and welding according to the time sequence of multiple assembly and welding sub-task sets, thereby realizing the autonomous allocation and collaborative operation of multiple robots and multiple tasks, reducing the duration of the overall welding process, improving welding efficiency and welding quality, and reducing labor costs and labor intensity, thereby further promoting the coordinated development of digitalization, intelligence and greening of the steel structure industry.

[0044] In one embodiment of the present application, a process of training a plurality of assembly and welding subtask sets to obtain a temporal sequence of the plurality of assembly and welding subtask sets includes:

[0045] Get the current welding status of the steel structure to be welded.

[0046] In this embodiment, the welding status of the steel structure to be welded is set according to the welding sub-task set. For example, when the welding is not performed, the corresponding status of the steel structure to be welded is that the welding task is not implemented, when the rib plate welding is completed, the corresponding status is that the rib plate welding task has been completed, when the partition plate welding is completed, the corresponding status is that the welding plate welding task has been completed, etc. The status such as the welding task not implemented, the rib plate welding task completed or the partition plate welding task completed can be used as the current welding status of the steel structure to be welded.

[0047] The current welding state of the steel structure to be welded is taken as the initial welding state. Based on the initial welding state, the probability of selecting all welding subtask sets is calculated respectively. The next welding subtask set is determined based on the probabilities of all welding subtask sets until all welding subtask sets are selected, and the temporal sequence of multiple welding subtask sets is obtained.

[0048] In this embodiment, when the next welding subtask set is selected for the first time, the welding subtask set with the highest probability is selected as the next welding subtask set among the probabilities of all welding subtask sets, and the selected welding subtask set is marked to distinguish the selected welding subtask set from the unselected welding subtask set by marking. When the probability of selecting all welding subtask sets is calculated again, the probability of the selected welding subtask set is set to 0, and the welding subtask set with the highest probability is continued to be selected as the next welding subtask set among the probabilities of all welding subtask sets until all welding subtask sets are selected.

[0049] In one embodiment of the present application, based on the initial welding state, the process of calculating the probability of selecting all welding subtask sets includes:

[0050] The initial welding state is input into the convolutional neural network model to obtain the feature matrix between the initial welding state and the set of all welding subtasks.

[0051] In this embodiment, the convolutional neural network model includes an input layer, a hidden layer and an output layer. The hidden layer includes a convolution layer, an activation function layer, a pooling layer, a fully connected layer, etc. The input layer, the hidden layer and the output layer extract the features between the initialized welding state and the set of all welding subtasks, and form a feature matrix with the features between the initialized welding state and the set of all welding subtasks.

[0052] In this embodiment, the convolutional neural network model can be a residual network model (Residual Network, ResNet) or a densely connected network model (DenseNet).

[0053] The feature matrix between the initial welding state and all welding subtask sets is input into the first fully connected network model, and is converted into a one-dimensional feature vector through the first fully connected network model.

[0054] In this embodiment, the fully connected network model is also called a multi-layer perceptron, which can map low-dimensional data to high-dimensional data, and make the data linearly separable by increasing the dimensional space where the data is located.

[0055] In this embodiment, different model parameters are set for the first fully connected network model, the second fully connected network model, and the third fully connected network model according to the input and output structures.

[0056] The one-dimensional feature vector is input into the normalized exponential function, and the probability of selecting all the welding subtask sets is calculated and output through the normalized exponential function.

[0057] In this embodiment, the normalized exponential function is used to perform normalization processing on a set of data so that the set of data is within a certain numerical range, generally within the range of [0, 1].

[0058] In one embodiment of the present application, the process of determining the next welding subtask set based on the probabilities of all welding subtask sets until all welding subtask sets are selected includes:

[0059] Based on the initialized welding state, the maximum probability value among the probabilities of all welding subtask sets is selected, and the welding subtask set corresponding to the maximum probability value is used as the next welding subtask set.

[0060] In this embodiment, after the next assembly and welding subtask set is determined for the first time, the selected assembly and welding subtask set needs to be marked. The selected assembly and welding subtask set is distinguished from the unselected assembly and welding subtask set by marking, and the probability of the selected assembly and welding subtask set is set to 0, so as to facilitate the determination of the next assembly and welding subtask set from the unselected assembly and welding subtask set again.

[0061] The welding state corresponding to the next welding subtask set is used as the initial welding state, and the maximum probability value among the probabilities of all welding subtask sets is repeatedly selected until all welding subtask sets are selected.

[0062] In this embodiment, after the next welding subtask set is determined for the first time, the welding state corresponding to the next welding subtask set is used as the initial welding state, and the next welding subtask of the initial welding state is continued to be determined until all welding subtask sets are selected.

[0063] In one embodiment of the present application, after obtaining the temporal sequence of the plurality of welding subtask sets, the multi-robot collaborative welding method for steel structures further includes:

[0064] According to the timing sequence of the multiple welding subtask sets, the multiple welding states corresponding to the welding subtask sets are sorted to form a timing path, and the timing relationship between the multiple welding states is used as a state transition link in the timing path.

[0065] In this embodiment, the timing path includes timing nodes and state transition links. The timing nodes are the welding states corresponding to the welding subtask set, and the state transition links are the timing relationships between multiple welding states.

[0066] The reward value of each state transition link is determined by the distance value between the end point of the current welding state and the starting point of the next welding state in each state transition link and the corresponding weight of the distance value, the robot joint angle from the current welding state to the next welding state and the corresponding weight of the robot joint angle, the positioner rotation angle from the current welding state to the next welding state and the corresponding weight of the positioner rotation angle, and the collision from the current welding state to the next welding state and the corresponding weight of the collision.

[0067] In this embodiment, each welding state corresponds to a starting point and an end point, dis represents the distance between the end point of the current welding state and the starting point of the next welding state in each state transition link, angle1 represents the robot joint angle from the current welding state to the next welding state, angle2 represents the positioner rotation angle from the current welding state to the next welding state, col represents the collision between the current welding state and the next welding state, and taking the weight corresponding to dis as 0.5, angle1 as 0.1, angle2 as 0.2, and col as 0.1 as an example, the calculation formula of the reward value R of each state transition link is as follows: R = -(0.5*dis+0.1*angle1+0.2*angle2+0.1*col) Formula (1)

[0068] Among them, R is the reward value of each state transition link, dis is the distance value between the end point of the current welding state and the starting point of the next welding state in each state transition link, angle1 is the robot joint angle from the current welding state to the next welding state, angle2 is the positioner rotation angle from the current welding state to the next welding state, col is the collision between the current welding state and the next welding state, and when a collision occurs, col is 100, and when no collision occurs, col is 0.

[0069] In this embodiment, as R is negative, the larger dis, angle1, angle2, and col are, the smaller the reward value R is. The reward value for each state transition comprehensively considers factors such as dis, angle1, angle2, and col, enabling the multi-robot to complete all tasks with minimal resource consumption (as reflected in no collisions, smooth robot joint movement, and short idle strokes) while also completing all tasks (the entire assembly welding task) in the shortest possible time.

[0070] The temporal path discount reward is determined based on the reward value of each state transition link and the discount coefficient corresponding to each state transition link.

[0071] In this embodiment, taking the timing path having 5 welding states as an example, the timing path has 4 state transition links R 12 、R 23 、R 34 and R 45 , R 12 The corresponding discount factor is 1, R 23 The corresponding discount factor is r, R 34 The corresponding discount factor is r 2 , R 45 The corresponding discount factor is r 3 , the calculation formula of the temporal path discount reward U is as follows: U=R 12 +r*R 23 +r 2 *R 34 +r 3 *R 45 Formula (2)

[0072] Among them, U is the temporal path discount reward, R 12 、R 23 、R 34 and R 45 They are the four state transition links in the timing path, r is the discount coefficient, and the value range of r is (0,1).

[0073] The current welding state corresponding to each state transition link is input into the convolutional neural network model, and the feature matrix of the current welding state is output. The next welding subtask set corresponding to each state transition link is input into the second fully connected network model, and the feature matrix of the next welding subtask set is output. The feature matrix of the current welding state is associated with the feature matrix of the next welding subtask set to obtain the association matrix between the current welding state and the next welding subtask set. The association matrix is ​​input into the third fully connected network model to obtain the value of each state transition link.

[0074] In this embodiment, the convolutional neural network model includes an input layer, a hidden layer and an output layer. The hidden layer includes a convolution layer, an activation function layer, a pooling layer, a fully connected layer, etc. The input layer, the hidden layer and the output layer extract the feature matrix of the current welding state corresponding to each state transition link, and the fully connected network model extracts the feature matrix of the next welding sub-task set corresponding to each state transition link.

[0075] In this embodiment, the method for performing feature association between the feature matrix of the current welding state and the feature matrix of the next welding subtask set to obtain the association matrix of the current welding state and the next welding subtask set is not specifically limited.

[0076] In this embodiment, the third fully connected network model converts the association matrix between the current welding state and the next welding subtask set into a one-dimensional feature vector, and calculates the value of each state transition link through the one-dimensional feature vector.

[0077] Calculate the difference between the value of the starting state transition link and the value of the ending state transition link in the timing path, and the discounted reward of the timing path, and take the absolute value of the difference to obtain the timing path value.

[0078] In this embodiment, the timing path has 5 welding states, and the value of the starting state transition link is Q 12 , the value of the end state transition link is Q 45 For example, the calculation formula of the timing path value Q is as follows: Q=|Q 12 -r 4 *Q 45 -U| Formula (3)

[0079] Among them, Q is the time path value, Q 45 is the value of the end state transition link, Q 12 is the value of the starting state transition link, U is the discounted reward of the timing path, r is the discount coefficient, and the value range of r is (0,1).

[0080] With the goal of minimizing the temporal path value, the parameters of the convolutional neural network model and the second and third fully connected network models are updated to obtain the first updated convolutional neural network model, the second updated fully connected network model, and the third updated fully connected network model.

[0081] In this embodiment, the process of minimizing the value of the timing path is to calculate the partial derivative of formula (3), update the parameters of the convolutional neural network model and the second fully connected network model and the third fully connected network model according to the parameter values ​​obtained by calculating the partial derivative, and then, according to the preset learning rate, continuously calculate the partial derivative of formula (3) to minimize the value of the timing path.

[0082] In this embodiment, the policy function network model is composed of a convolutional neural network model and a first fully connected network model, and the parameters of the convolutional neural network model and the first fully connected network model in the value function network model are represented by θ.

[0083] In this embodiment, the value function network model is composed of a convolutional neural network model, a second fully connected network model and a third fully connected network model. The parameters of the convolutional neural network model, the second fully connected network model and the third fully connected network model in the value function network model are represented by W, and the convolutional neural network model and the convolutional neural network model in the policy function network model are shared.

[0084] In one embodiment of the present application, after obtaining the updated convolutional neural network model, the multi-robot collaborative welding method for steel structure parts further includes:

[0085] Multiply the value of each state transition link by the probability of the corresponding state transition link to obtain the expected value of each state transition link.

[0086] In this embodiment, the probability corresponding to each state transition link is determined by the initial welding state of the link and the next welding subtask set. The calculation formula for the expected value of each state transition link is as follows: V(s) = π(a|s,θ)·Q(s,a,W) Formula (4)

[0087] Among them, s represents the initial welding state of each state transition link, V(s) represents the expected value of each state transition link, a represents the next welding subtask set, Q(s,a,W) represents the value of each state transition link when the initial welding state is s and the next welding subtask set is a, π(a|s,θ) represents the probability of each state transition link when the initial welding state is s and the next welding subtask set is a, among which W is the parameter of the value function network model Q(s,a,W), and θ is the parameter of the policy function network model π(a|s,θ).

[0088] Add up the expected values ​​of all state transition links to get the total expected value.

[0089] In this embodiment, the calculation formula of the total expected value is as follows: V 总 (s)=∑V(s)=∑π(a|s,θ)·Q(s,a,W) Equation (5)

[0090] Among them, V 总(s) is the total expected value, s represents the initial welding state of each state transition link, V(s) represents the expected value of each state transition link, a represents the next welding subtask set, Q(s,a,W) represents the value of each state transition link when the initial welding state is s and the next welding subtask set is a, π(a|s,θ) represents the probability of each state transition link when the initial welding state is s and the next welding subtask set is a, where W is the parameter of the value function network model Q(s,a,W), and θ is the parameter of the policy network model π(a|s,θ).

[0091] With the goal of maximizing the total expected value, the parameters of the first updated convolutional neural network model and the first fully connected network model are updated to obtain the second updated convolutional neural network model and the first updated fully connected network model.

[0092] In this embodiment, after obtaining the first updated convolutional neural network model and the first updated fully connected network model, the parameter W of the value function network model is used as a known value to maximize the total expected value. The process of maximizing the total expected value is to calculate the partial derivatives of formula (5), update the parameters θ in the policy network model according to the parameter values ​​obtained by calculating the partial derivatives (i.e., update the first updated convolutional neural network model and the first fully connected network model), and then, according to a preset learning rate, continuously calculate the partial derivatives of formula (5) to maximize the total expected value.

[0093] Based on the second updated convolutional neural network model, the first updated fully connected network model, the second updated fully connected network model, and the third updated fully connected network model, multiple welding subtask sets are repeatedly trained until the number of training times for the multiple welding subtask sets reaches a preset threshold, and the temporal sequence of the multiple welding subtask sets is output.

[0094] In this embodiment, the preset number threshold can be set according to actual conditions.

[0095] In one embodiment of the present application, before obtaining the steel structure to be welded and the welding task set, the following steps are included:

[0096] Obtain the equipment and corresponding equipment information required in the process of assembling and welding steel structures.

[0097] In this embodiment, the equipment required for welding steel structures includes multiple robots, robot end-of-line tools, positioners, sensors, and other equipment. The robot end-of-line tools include welding guns, suction cups, and grippers. A welding gun mounted on the end of the robot's manipulator arm enables spot welding and welding, while a suction cup or gripper mounted on the end of the robot's manipulator arm enables grasping and other functions. The positioner flips the steel structure to be welded, and sensors, including cameras, scanners, and contact detection devices, are used to detect the workpiece's position and determine collisions. Equipment information includes the model and size of the equipment, its location, and the coordination between the various devices.

[0098] A three-dimensional simulation model is constructed based on the equipment and equipment information, so as to control different robots to perform welding operations through the three-dimensional simulation model.

[0099] In this embodiment, after establishing a three-dimensional simulation model in proportion to the equipment model and size, equipment location information, and coordination relationship of each equipment, a system environment simulation platform is built according to the layout of the equipment. The simulation platform can control the movement of moving parts (such as robots, guide rails) in the assembly welding system in the form of instructions.

[0100] This application obtains the steel structures to be welded and the welding task sets, trains multiple welding sub-task sets, obtains the time sequence of multiple welding sub-task sets, and assigns the grasping tasks and welding tasks to different robots to control different robots to assemble and weld the steel structures to be welded according to the time sequence of multiple welding sub-task sets, realizing the autonomous allocation and collaborative operation of multiple robots and multiple tasks, reducing the duration of the overall welding process, improving welding efficiency and welding quality, and reducing labor costs and labor intensity, thereby further promoting the coordinated development of digitalization, intelligence and greening of the steel structure industry.

[0101] Please refer to Figure 3, which is a schematic diagram of the layout of a robot assembly welding system shown in an exemplary embodiment of the present application, including a robot module, an end tool module, a positioner, a sensor module and a sensor module, wherein the robot module includes 2-3 robots that cooperate with each other (for example, robot 1 and robot 2), and a robot external axis for linear motion (moving guide rails, robot internal system control), which is used to move and perform assembly welding of steel structures. The end tool module includes robot end tools such as welding guns, suction cups, and grippers, which are used to achieve the grasping and welding of steel structures. The positioner is used to rotate steel structures so that the robot can assemble and weld various components to the workpiece. The sensor module mainly includes cameras, scanners, and contact detection, which are used to obtain the position of the workpiece and determine whether a collision occurs. The control module is used to assign assembly welding sequence tasks, issue robot walking paths, and control the robot end tools at the same time to ultimately achieve assembly welding.

[0102] In this embodiment, for the steel structure welding tasks that are about to be completed in subsequent production, the name and type of the steel structure to be welded are input, and then according to the multi-robot collaborative welding method of steel structures, the global assembly welding task set is divided into a sub-task set with a time sequence, and then allocated and sent to the task storage area of ​​each designated robot in the robot assembly welding system, and the robots are controlled to cooperate with each other to realize the assembly welding of steel structures.

[0103] Please refer to Figure 4, which is a flowchart of a robot welding subtask set training method shown in an exemplary embodiment of the present application. The method includes: (1) initializing the simulation experimental environment, that is, obtaining the equipment and corresponding equipment information required in the process of welding steel structures, and constructing a three-dimensional simulation model based on the equipment and equipment information to control different robots to perform welding operations through the three-dimensional simulation model; (2) based on the initialized welding state, respectively calculating the probability of selecting all welding subtask sets, and determining the next welding subtask set based on the probability of all welding subtask sets to form a timing path; (3) calculating the timing path value, minimizing the timing path value, and updating the parameters of the value function network model; (4) calculating the total expected value, and maximizing the total expected value. When the parameters of the updated value function network model are known, the parameters in the strategy function network model are updated; (5) executing in a loop until the training is completed.

[0104] The following describes an embodiment of the apparatus of the present application, which can be used to implement the multi-robot collaborative welding method for steel structures described in the above-mentioned embodiment of the present application. For details not disclosed in the apparatus embodiment of the present application, please refer to the embodiment of the multi-robot collaborative welding method for steel structures described in the above-mentioned embodiment of the present application.

[0105] FIG5 is a block diagram of a multi-robot collaborative welding apparatus for steel structures, illustrating an exemplary embodiment of the present application. This apparatus can be applied to the implementation environment shown in FIG1 and specifically configured in computer device 102. This apparatus can also be applied to other exemplary implementation environments and specifically configured in other devices. This embodiment does not limit the applicable implementation environment of this apparatus.

[0106] As shown in FIG5 , the exemplary multi-robot collaborative welding device for steel structures includes:

[0107] The steel structure acquisition module 501 is used to acquire steel structures to be welded and a welding task set, where the welding task set includes multiple welding subtask sets.

[0108] The timing determination module 502 is used to train multiple assembly and welding subtask sets to obtain the timing sequence of the multiple assembly and welding subtask sets, where the assembly and welding subtask sets include grasping tasks and welding tasks.

[0109] The task allocation module 503 is used to allocate the grasping task and the welding task to different robots, so as to control the different robots to assemble and weld the steel structure to be assembled according to the time sequence of multiple assembly and welding subtask sets.

[0110] In one embodiment of the present application, the welding task set includes multiple welding sub-task sets, and the multiple welding sub-task sets correspond to multiple welding steel structural parts. The steel structural parts to be welded include ribs and partitions, etc. For example, the welding sub-task set corresponding to the ribs includes grabbing the ribs, spot welding assembly, arc welding, rotation of auxiliary mechanism workstations, etc.

[0111] In this embodiment, the process of training multiple welding subtask sets to obtain the temporal sequence of the multiple welding subtask sets includes: (1) obtaining the current welding state of the steel structure to be welded; (2) taking the current welding state of the steel structure to be welded as the initialization welding state, and based on the initialization welding state, respectively calculating the probability of selecting all welding subtask sets, and determining the next welding subtask set based on the probabilities of all welding subtask sets, until all welding subtask sets are selected, thereby obtaining the temporal sequence of the multiple welding subtask sets.

[0112] In this embodiment, the assembly and welding sub-task set includes grasping tasks and welding tasks. When the welding sub-task combination includes grasping the rib plate, spot welding assembly, arc welding, and rotation of the auxiliary mechanism station, grasping the rib plate is the grasping task, spot welding assembly and arc welding are the welding tasks, and the rotation of the auxiliary mechanism station is the rotation task of the positioner.

[0113] In this embodiment, by training multiple assembly and welding sub-task sets, the time sequence of multiple assembly and welding sub-task sets is obtained, and the grasping tasks and welding tasks are assigned to different robots to control different robots to perform assembly and welding according to the time sequence of multiple assembly and welding sub-task sets, thereby realizing the autonomous allocation and collaborative operation of multiple robots and multiple tasks, reducing the duration of the overall welding process, improving welding efficiency and welding quality, and reducing labor costs and labor intensity, thereby further promoting the coordinated development of digitalization, intelligence and greening of the steel structure industry.

[0114] It should be noted that the multi-robot collaborative assembly and welding device for steel structures provided in the above embodiment and the multi-robot collaborative assembly and welding method for steel structures provided in the above embodiment are of the same concept, wherein the specific manner in which each module and unit performs operations has been described in detail in the method embodiment and will not be repeated here. In actual applications, the multi-robot collaborative assembly and welding device for steel structures provided in the above embodiment can allocate the above functions to different functional modules as needed, that is, divide the internal structure of the device into different functional modules to complete all or part of the functions described above, and this is not limited here.

[0115] An embodiment of the present application also provides an electronic device, comprising: one or more processors; a storage device for storing one or more programs, wherein when the one or more programs are executed by one or more processors, the electronic device implements the multi-robot collaborative assembly and welding method for steel structures provided in the above-mentioned embodiments.

[0116] Figure 6 shows a schematic diagram of the structure of a computer system of an electronic device suitable for implementing an embodiment of the present application. It should be noted that the computer system 600 of the electronic device shown in Figure 6 is only an example and should not bring any limitation to the functions and scope of use of the embodiment of the present application.

[0117] As shown in Figure 6, computer system 600 includes a central processing unit (CPU) 601, which can perform various appropriate actions and processes according to the program stored in read-only memory (ROM) 602 or the program loaded from storage portion 608 into random access memory (RAM) 603, such as executing the method in the above embodiment. Various programs and data required for system operation are also stored in RAM 603. CPU 601, ROM 602 and RAM 603 are connected to each other via bus 604. Input / output (I / O) interface 605 is also connected to bus 604.

[0118] The following components are connected to the I / O interface 605: an input section 606 including a keyboard, a mouse, and the like; an output section 607 including devices such as a cathode ray tube (CRT), a liquid crystal display (LCD), and a speaker; a storage section 608 including a hard disk; and a communication section 609 including a network interface card such as a LAN (Local Area Network) card or a modem. The communication section 609 performs communication processing via a network such as the Internet. A drive 610 is also connected to the I / O interface 605 as needed. Removable media 611, such as a magnetic disk, an optical disk, a magneto-optical disk, or a semiconductor memory, is installed in the drive 610 as needed, so that computer programs read therefrom can be installed into the storage section 608 as needed.

[0119] In particular, according to an embodiment of the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present application includes a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a computer program for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network via the communication section 609, and / or installed from a removable medium 611. When the computer program is executed by the central processing unit (CPU) 601, the various functions defined in the system of the present application are executed.

[0120] It should be noted that the computer-readable medium shown in the embodiments of the present application may be a computer-readable signal medium or a computer-readable storage medium or any combination of the two. The computer-readable storage medium may be, for example, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or component, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM), a flash memory, an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present application, a computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, which carries a computer-readable computer program. This propagated data signal may take a variety of forms, including but not limited to an electromagnetic signal, an optical signal, or any suitable combination of the above. A computer-readable signal medium may also be any computer-readable medium other than a computer-readable storage medium that can transmit, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device. A computer program embodied on a computer-readable medium may be transmitted using any suitable medium, including but not limited to wireless, wired, or any suitable combination thereof.

[0121] The flowcharts and block diagrams in the accompanying drawings illustrate the possible implementation architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present application. Among them, each box in the flowchart or block diagram can represent a module, program segment, or part of the code, and the above-mentioned module, program segment, or part of the code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in an order different from that marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram or flowchart, and the combination of boxes in the block diagram or flowchart, can be implemented with a dedicated hardware-based system that performs the specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0122] The units involved in the embodiments described in this application may be implemented by software or hardware, and the units described may also be set in a processor. In some cases, the names of these units do not constitute limitations on the units themselves.

[0123] Another aspect of the present application provides a computer-readable storage medium having computer-readable instructions stored thereon. When executed by a computer processor, the computer executes the multi-robot collaborative welding method for steel structures provided in the aforementioned embodiments. The computer-readable storage medium may be included in the electronic device described in the aforementioned embodiments, or may exist independently and not be incorporated into the electronic device.

[0124] It should be noted that, although several modules or units of the device for action execution are mentioned in the above detailed description, this division is not mandatory. In fact, according to the embodiment of the application, the features and functions of two or more modules or units described above can be concretized in one module or unit. On the contrary, the features and functions of one module or unit described above can be further divided into multiple modules or units to be concretized.

[0125] Through the description of the above embodiments, it is easy for those skilled in the art to understand that the example embodiments described herein can be implemented by software or by combining software with necessary hardware. Therefore, the technical solution according to the embodiments of the present application can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.) or on a network, and includes several instructions to enable a computing device (which can be a personal computer, a server, a touch terminal, or a network device, etc.) to execute the method according to the embodiments of the present application.

[0126] Those skilled in the art will readily conceive of other embodiments of the present application after considering the specification and practicing the embodiments disclosed herein. This application is intended to cover any variations, uses, or adaptations of the present application that follow the general principles of this application and include common knowledge or customary techniques in the art that are not disclosed herein.

[0127] It should be understood that the above content is only a preferred exemplary embodiment of the present application and is not intended to limit the implementation scheme of the present application. Ordinary technicians in this field can easily make corresponding changes or modifications based on the main concept and spirit of the present application. Therefore, the scope of protection of the present application should be based on the scope of protection required by the claims.

Claims

1. A multi-robot collaborative welding method for steel structure parts, characterized in that: include: Obtain a steel structure to be welded and a welding task set, wherein the welding task set includes a plurality of welding subtask sets; Training a plurality of the welding subtask sets to obtain a time sequence of the plurality of welding subtask sets, wherein the welding subtask sets include a grasping task and a welding task; The grasping task and the welding task are assigned to different robots, so as to control the different robots to perform assembly welding on the steel structure to be assembled and welded according to the time sequence of the plurality of assembly and welding subtask sets.

2. The multi-robot collaborative welding method for steel structure parts according to claim 1 is characterized in that: The process of training the plurality of welding subtask sets to obtain the time sequence of the plurality of welding subtask sets includes: Obtaining the current welding status of the steel structure to be welded; The current welding state of the steel structure to be welded is taken as the initialization welding state. Based on the initialization welding state, the probabilities of selecting all the welding subtask sets are calculated respectively. The next welding subtask set is determined based on the probabilities of all the welding subtask sets, until all the welding subtask sets are selected, and the time sequence of the multiple welding subtask sets is obtained.

3. The multi-robot collaborative welding method for steel structure parts according to claim 2 is characterized in that: The process of respectively calculating the probability of selecting all the welding subtask sets based on the initial welding state includes: Inputting the initial welding state into a convolutional neural network model to obtain a feature matrix between the initial welding state and all the welding subtask sets; Inputting the feature matrix between the initial welding state and all the welding subtask sets into a first fully connected network model, and converting the first fully connected network model to obtain a one-dimensional feature vector; The one-dimensional feature vector is input into a normalized exponential function, and the normalized exponential function is used to calculate And output the probability of selecting all the welding subtask sets.

4. The multi-robot collaborative welding method for steel structure parts according to claim 2, characterized in that: The process of determining the next welding subtask set based on the probabilities of all the welding subtask sets until all the welding subtask sets are selected includes: Based on the initial welding state, selecting the maximum probability value among all the welding subtask sets, and taking the welding subtask set corresponding to the maximum probability value as the next welding subtask set; The welding state corresponding to the next welding subtask set is used as the initial welding state, and the maximum probability value among the probabilities of all the welding subtask sets is repeatedly selected until all the welding subtask sets are selected.

5. The multi-robot collaborative welding method for steel structure parts according to claim 3 is characterized in that: After obtaining the time sequence of the plurality of welding subtask sets, the method further comprises: According to the time sequence of the plurality of welding subtask sets, a plurality of welding states corresponding to the welding subtask sets are sorted to form a time sequence path, and the time sequence relationship between the plurality of welding states is used as a state transition link in the time sequence path; The reward value of each state transition link is determined by the distance value between the end point of the current welding state and the starting point of the next welding state and the corresponding weight of the distance value, the robot joint angle from the current welding state to the next welding state and the corresponding weight of the robot joint angle, the positioner rotation angle from the current welding state to the next welding state and the corresponding weight of the positioner rotation angle, and the collision from the current welding state to the next welding state and the corresponding weight of the collision. Determine the time sequence path discount reward according to the reward value of each state transition link and the discount coefficient corresponding to each state transition link; Input the current welding state corresponding to each state transition link into the convolutional neural network model, output the feature matrix of the current welding state, input the next welding subtask set corresponding to each state transition link into the second fully connected network model, output the feature matrix of the next welding subtask set, Performing feature association on the feature matrix of the current welding state and the feature matrix of the next welding subtask set to obtain a correlation matrix between the current welding state and the next welding subtask set, inputting the correlation matrix into a third fully connected network model to obtain the value of each state transition link; Calculate the difference between the value of the start state transition link and the value of the end state transition link in the timing path and the discount reward of the timing path, and take the absolute value of the difference to obtain the timing path value; With the goal of minimizing the temporal path value, the parameters of the convolutional neural network model and the second fully connected network model and the third fully connected network model are updated to obtain a first updated convolutional neural network model, a second updated fully connected network model, and a third updated fully connected network model.

6. The multi-robot collaborative welding method for steel structure parts according to claim 5, characterized in that: After obtaining the updated convolutional neural network model, the method further includes: Multiplying the value of each state transition link by the probability of the corresponding state transition link to obtain the expected value of each state transition link; Adding the expected values ​​of all the state transition links together to obtain a total expected value; With the goal of maximizing the total expected value, the parameters of the first updated convolutional neural network model and the first fully connected network model are updated to obtain a second updated convolutional neural network model and a first updated fully connected network model; Based on the second updated convolutional neural network model, the first updated fully connected network model, the second updated fully connected network model and the third updated fully connected network model, the multiple welding subtask sets are repeatedly trained until the number of training times for the multiple welding subtask sets reaches a preset number threshold, and the timing order of the multiple welding subtask sets is output.

7. The multi-robot collaborative welding method for steel structure parts according to any one of claims 1 to 6, characterized in that: Before obtaining the steel structure parts to be welded and the welding task set, it includes: Obtain the equipment and corresponding equipment information required in the process of welding steel structures; A three-dimensional simulation model is constructed based on the device and the device information to control the Different robots are made to perform welding operations.

8. A multi-robot collaborative welding device for steel structure parts, characterized in that: include: A steel structure acquisition module, used to acquire steel structures to be welded and a welding task set, wherein the welding task set includes a plurality of welding subtask sets; A timing determination module is used to train a plurality of the welding subtask sets to obtain a timing sequence of the plurality of welding subtask sets, wherein the welding subtask sets include a grasping task and a welding task; The task allocation module is used to allocate the grasping task and the welding task to different robots, so as to control the different robots to assemble and weld the steel structure to be assembled and welded according to the time sequence of the plurality of assembly and welding subtask sets.

9. An electronic device, characterized in that: include: one or more processors; A storage device for storing one or more programs, which, when executed by the one or more processors, enables the electronic device to implement the multi-robot collaborative welding method for steel structures as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: Computer-readable instructions are stored thereon, and when the computer-readable instructions are executed by a processor of a computer, the computer is enabled to execute the multi-robot collaborative welding method for steel structure parts according to any one of claims 1 to 7.

Images