Task grading method and device for steel grabbing machine, electronic equipment and storage medium

Abstract

The invention provides a steel grabbing machine task grading method and device, electronic equipment and a storage medium, and the method comprises the steps: carrying out the downsampling of the whole operation area of a steel grabbing machine, obtaining an environment map, carrying out the task decomposition according to the real working space of a mechanical arm, and enabling the maximum utilization of the reachable range of the mechanical arm, the defect that a reachable space is ignored in a traditional method is overcome, meanwhile, an overall grabbing task is decomposed into a steel material carrying sub-task and a machine body moving sub-task, parallel processing and optimal scheduling of the tasks are achieved, and therefore the overall operation efficiency is improved; the whole task is decomposed into smaller subtasks and motion primitive sequences, flexible planning can be carried out according to actual working condition operation requirements, various changes in the operation process can be more flexibly dealt with through task grading and refined planning, the reliability and stability of task execution are ensured, and the working efficiency is improved. According to the hierarchical planning architecture in the application, a complex grabbing task is reasonably decomposed, and the robustness and efficiency of planning are improved.

Description

Technical Field

[0001] The present application relates to the technical field of the steel industry, and specifically to a steel grabber task classification method, device, electronic equipment and storage medium. Background Art

[0002] At present, the hierarchical task planning methods for large hydraulic mobile robotic arms in actual scenarios usually set the operating space of the robotic arm as a regular hexahedron with a fixed volume. However, in fact, the reachable space of the robotic arm is an irregularly shaped space, which ignores the actual reachable range of the robotic arm and fails to fully utilize the reachable space of the robotic arm during the planning process.

[0003] In addition, when performing hierarchical planning, most low-level planners only consider the motion primitives of the decoupled form of the mobile chassis and the manipulator, and ignore the motion primitives of the coordinated motion of the mobile chassis and the manipulator. How to design a hierarchical task planner for the mobile manipulator that conforms to common sense and fully utilizes the reachable space of the manipulator and the coordinated motion of the mobile manipulator hand according to the actual scrap steel grabbing and handling work scenes and task requirements is one of the main research contents of this paper. Summary of the invention

[0004] The present application provides a steel grabber task classification method, device and electronic equipment to solve the above-mentioned technical problems.

[0005] An embodiment of the present application provides a method for grading steel grabber tasks, the method comprising: downsampling the overall operating area of ​​the steel grabber to obtain an environmental map, and obtaining a robotic arm operating space based on the overall grasping task of the overall operating area of ​​the steel grabber; decomposing the overall grasping task into subtasks based on the robotic arm operating space and the environmental map to obtain a subtask sequence, and determining the order of steel grabbing areas of the overall grasping task to determine the order of path points for the steel grabber chassis to stop, the subtasks comprising a steel material handling subtask and a machine body movement subtask; decomposing each subtask into a motion primitive sequence, and planning a collision-free path based on the motion primitive sequence and the path point sequence to complete task grading of the overall grasping task.

[0006] In one embodiment of the present application, the overall grasping task is decomposed into a steel handling subtask and a body movement subtask to obtain a subtask sequence, including: inputting the environmental map into the steel grasping task decomposition planner, and decomposing the overall grasping task into a steel handling subtask and a body movement subtask through the steel grasping task decomposition planner; if the subtask type is the steel handling subtask, then the steel handling subtask is planned by the steel handling subtask planner to obtain a steel handling subtask planning result, and the steel handling local map is updated, and the steel handling subtask planning result includes steel handling subtask parameters; if the subtask type is the body movement subtask, then the body movement subtask is planned by the body movement subtask planner to obtain a body movement subtask planning result, and the body movement subtask planning result includes body movement subtask parameters; and a subtask sequence is constructed according to the steel handling subtask planning result and the body movement subtask planning result.

[0007] In one embodiment of the present application, a steel handling subtask is planned by a steel handling subtask planner to obtain a steel handling subtask planning result, including: determining whether the grid operating heights in the steel handling subtask are all less than a preset height threshold; if all grid operating heights are less than the preset height threshold, then planning the steel handling subtask according to the grid operating heights to obtain a steel handling subtask planning result; if any grid operating height is greater than the preset height threshold, then grabbing the clearable grid according to the theoretical maximum grabbing height map of steel grabbing, and updating the grabbing map to obtain a steel handling subtask planning result.

[0008] In one embodiment of the present application, a steel handling subtask is planned according to the grid working height to obtain a steel handling subtask planning result, including: if the working heights of all grids are less than a preset height threshold, then whether the working heights of all grids are the same is determined; if the working heights of all grids are equal, then the grid closest to the steel grabber base is selected for grabbing; if the working heights of all grids are not equal, then the grid with the highest working height is selected for grabbing, and the grabbing map is updated to obtain the steel handling subtask planning result.

[0009] In one embodiment of the present application, each subtask is decomposed into a motion primitive sequence, including: if the subtask type is a steel handling subtask, then the steel handling motion primitive is obtained through the steel handling subtask parameters and the local environment information of steel handling; when the subtask type is a body movement subtask, the body movement motion primitive is obtained according to the body movement subtask parameters and the local environment information of the body movement; and a motion primitive sequence is constructed according to the steel handling motion primitive and the body movement motion primitive.

[0010] In one embodiment of the present application, after constructing the motion primitive sequence, it also includes: executing the body movement action according to the body movement motion primitive, and executing the steel material handling motion action according to the steel material handling motion primitive; if the motion primitive is a gripping claw motion primitive, then after the steel gripper completes the operation, the environmental map is updated to determine the change in the position or state of the steel material.

[0011] In one embodiment of the present application, the order of the steel material grabbing areas of the overall grabbing task is determined to determine the order of the path points for the chassis of the steel grabber to stop, including: for the steel material handling task, the steel material grabbing path points of the steel material grabbing task are obtained according to the chassis position of the steel grabber; for the body movement subtask, the position of the end effector of the robot arm is set above the steel material, and the path points for the chassis of the steel grabber to stop are determined according to the working space of the robot arm and the steel material environment.

[0012] The embodiment of the present application provides a task grading device for a steel grabber, the device comprising: a sampling module, used to downsample the overall operating area of ​​the steel grabber to obtain an environmental map, and obtain a robot arm operating space based on the overall grasping task of the overall operating area of ​​the steel grabber; a task decomposition module, used to decompose the overall grasping task into subtasks according to the robot arm operating space and the environmental map, obtain a subtask sequence, and determine the order of steel material grasping areas of the overall grasping task to determine the path point sequence of the steel grabber chassis docking, the subtasks including steel material handling subtasks and machine body movement subtasks; a motion primitive decomposition module, used to decompose each subtask into a motion primitive sequence, and plan a collision-free path based on the motion primitive sequence and the path point sequence to complete task grading of the overall grasping task.

[0013] An embodiment of the present application provides an electronic device, including a processor, a memory and a communication bus; the communication bus is used to connect the processor and the memory; the processor is used to execute a computer program stored in the memory to implement the above-mentioned steel grabber task classification method.

[0014] The beneficial effects of the present application are as follows: by downsampling the overall operation area of ​​the steel grabber to obtain an environmental map, complex environmental information can be simplified, making subsequent task planning more efficient; by decomposing tasks according to the real working space of the robot arm, the reachable range of the robot arm can be utilized to the greatest extent, avoiding the defect of the traditional method ignoring the reachable space; at the same time, decomposing the overall grasping task into steel material handling subtasks and body movement subtasks, it is helpful to realize parallel processing and optimized scheduling of tasks, thereby improving the overall operation efficiency; introducing motion primitives for the coordinated movement of the mobile chassis and the robot arm, decomposing the overall task into smaller subtasks and motion primitive sequences, enriching the movement mode, and being able to flexibly plan according to the actual working conditions. At the same time, planning a collision-free path based on the motion primitive sequence and the path point sequence can effectively avoid the collision of the steel grabber with obstacles during the operation process, and ensure the safety of the operation; through task classification and refined planning, it is possible to more flexibly respond to various changes in the operation process, such as changes in the position of steel materials, dynamic adjustment of the working environment, etc. At the same time, since each subtask and motion primitive is carefully designed and planned, the reliability and stability of task execution can be ensured. The hierarchical planning architecture in this application enables complex grasping tasks to be reasonably decomposed, improving the robustness and efficiency of planning.

[0015] 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

[0016] The drawings herein are incorporated into the specification and constitute a part of the specification, showing embodiments consistent with the present application, and together with the specification, are used to explain the principles of the present application. Obviously, the drawings described below are only some embodiments of the present application, and for those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative work. In the drawings:

[0017] Figure 1 is a flow chart of a steel grabber task classification method shown in an exemplary embodiment of the present application;

[0018] Figure 2 is a schematic diagram of a research method for hierarchical task planning of a mobile manipulator based on a neural network, shown as an exemplary embodiment of the present application;

[0019] Figure 3 is a network structure diagram of a steel grabber task decomposition planner SGTDP shown in an exemplary embodiment of the present application;

[0020] Figure 4 is a planning flow chart of a steel material handling subtask planner shown in an exemplary embodiment of the present application;

[0021] Figure 5 This is a network structure diagram of SGTDP based on NPI shown in an exemplary embodiment of the present application.

[0022] Figure 6 is a flow chart of steel grabber task classification shown in an exemplary embodiment of the present application;

[0023] Figure 7 is a block diagram of a steel grabber task classification device shown in an exemplary embodiment of the present application;

[0024] Figure 8 It is a structural schematic diagram of an electronic device shown in an exemplary embodiment of the present application. DETAILED DESCRIPTION

[0025] The following will describe the implementation methods of the present application with reference to the accompanying drawings and preferred embodiments. Those skilled in the art can easily understand other advantages and effects of the present application from the contents disclosed in this specification. The present application can also be implemented or applied through other different specific implementation methods, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present application. It should be understood that the preferred embodiments are only for illustrating the present application, not for limiting the scope of protection of the present application.

[0026] It should be noted that the illustrations provided in the following embodiments are only schematic illustrations of the basic concept of the present application, and thus the drawings only show components related to the present application rather than being drawn according to the number, shape and size of components in actual implementation. In actual implementation, the type, quantity and proportion of each component may be changed at will, and the component layout may also be more complicated.

[0027] In the following description, a large number of details are discussed to provide a more thorough explanation of the embodiments of the present application. However, it is obvious to those skilled in the art that the embodiments of the present application can be implemented without these specific details. In other embodiments, well-known structures and devices are shown in the form of block diagrams rather than in detail to avoid making the embodiments of the present application difficult to understand.

[0028] The embodiments of the present application respectively propose a steel grabber task classification method, a steel grabber task classification device, an electronic device and a storage medium, and these embodiments will be described in detail below.

[0029] See also Figure 1 , Figure 1 FIG. 1 is a flow chart of a steel grabber task classification method according to an embodiment of the present application. Figure 1 As shown, the method at least includes steps S110 to S130, which are described in detail as follows:

[0030] Step S110, downsampling the entire operation area of ​​the steel grabber to obtain an environmental map, and obtaining a robot arm operation space according to the overall grasping task of the entire operation area of ​​the steel grabber.

[0031] The existing technology usually sets the operation space of the robot arm as a regular fixed-size hexahedron, ignoring the actual irregular reach of the robot arm. However, this solution introduces the NPI-based Steel Grasping Task Decomposition Planner (SGTDP), which can decompose the overall grasping task according to the actual workspace of the robot arm, thereby making full use of the reachable space of the robot arm.

[0032] Step S120, according to the robot arm working space and environmental map, decompose the overall grasping task into subtasks, obtain the subtask sequence, and determine the order of the steel material grasping areas of the overall grasping task to determine the path point sequence for the steel grabber chassis to stop. The subtasks include steel material handling subtasks and machine body movement subtasks.

[0033] Step S130, decomposing each subtask into a motion primitive sequence, and planning a collision-free path based on the motion primitive sequence and the path point sequence, thereby completing the task classification of the overall grasping task.

[0034] exist Figure 1 In the technical solution shown, by downsampling the overall operation area of ​​the steel grabber to obtain an environmental map, complex environmental information can be simplified, making subsequent task planning more efficient. Task decomposition based on the real workspace of the robot arm can maximize the use of the reachable range of the robot arm, avoiding the defect of the traditional method ignoring the reachable space. At the same time, decomposing the overall grasping task into steel material handling subtasks and body movement subtasks helps to achieve parallel processing and optimized scheduling of tasks, thereby improving overall operation efficiency; introducing motion primitives for coordinated movement of the mobile chassis and the robot arm, decomposing the overall task into smaller subtasks and motion primitive sequences, enriching the movement mode, and being able to flexibly plan according to actual working conditions. At the same time, planning a collision-free path based on the motion primitive sequence and the path point sequence can effectively avoid collisions between the steel grabber and obstacles during operation and ensure operation safety; through task classification and refined planning, it can more flexibly respond to various changes in the operation process, such as changes in the position of steel materials and dynamic adjustments to the working environment. At the same time, since each subtask and motion primitive is carefully designed and planned, the reliability and stability of task execution can be ensured. The hierarchical planning architecture in this application enables complex grasping tasks to be reasonably decomposed, improving the robustness and efficiency of planning.

[0035] In one embodiment of the present application, Figure 2 The present application embodiment shows a schematic diagram of a research method for layered task planning of a mobile manipulator based on a neural network, as shown in FIG. Figure 2As shown, for the grabbing task of the whole area, the working space of the robot arm is first approximated, and the overall environmental map is downsampled in combination with the environmental information provided by the environmental perception module. On this basis, a grabbing task decomposition planner based on a neural programmer-interpreter (NPI) is designed in the upper planning layer to decompose the overall grabbing task into a subtask sequence consisting of several steel material handling subtasks and body movement subtasks; then a subtask planner is designed in the subtask planning layer to decompose each subtask into a motion primitive sequence consisting of multiple pre-set motion primitives. In the upper task planning layer, the actual grabbing task has a large operating area, and the steel grabber cannot rely solely on the robot arm to complete the entire grabbing task without moving the chassis, so it is necessary to decompose the target grabbing task to determine the grabbing order of the whole operating area. According to the steel grabbing task, this application names the upper task decomposition planner as the Steel Grabbing Task Division Planner (SGTDP). For different types of grasping tasks, SGTDP uses the environmental map information provided by the environmental perception module as input, and then decomposes the entire grasping task into multiple steel material handling subtasks and body movement subtasks according to the working space of the robot arm. For steel material grasping tasks, SGTDP assumes that the chassis of the steel grasper is fixed during the grasping process, and the working space of the robot arm at a fixed position is also fixed. For the body movement subtask, SGTDP determines the path point where the chassis of the steel grasper stops based on the working space of the robot arm and the steel material environment, and sets the position of the end effector of the robot arm above the steel material. As a specific example, Figure 3The network structure diagram of the steel grabber task decomposition planner SGTDP is shown. For each overall grasping task, SGTDP needs to determine the order of the steel material grasping area to determine the order of the path points where the steel grabber chassis stops. At the subtask planning layer, given the various subtasks output by SGTDP, the subtask planner provides solutions to complete different subtasks. This application names the steel material handling subtask planner (SteelTransport Sub-task Planner, STSP) and the body movement subtask planner (Agent Move Sub-task Planner, AMSP). STSP determines the grasping position of the end effector of the steel grabber and determines whether the chassis needs to be moved again based on the environmental information of the current subtask area and the working space of the robot arm, and outputs a sequence of motion primitives composed of multiple motion primitives. AMSP plans a collision-free path based on the path points where the steel grabber chassis stops and the position of the end effector output by SGTDP and combined with motion constraints. By downsampling the overall operation area of ​​the steel grabber to obtain an environmental map, complex environmental information can be simplified, making subsequent task planning more efficient. Decomposing tasks according to the real working space of the robot arm can maximize the reach of the robot arm and avoid the defect of ignoring the reachable space in the traditional method. At the same time, decomposing the overall grasping task into steel material handling subtasks and body movement subtasks is helpful to realize parallel processing and optimized scheduling of tasks, thereby improving the overall operation efficiency; introducing motion primitives for coordinated movement of the mobile chassis and the robot arm, decomposing the overall task into smaller subtasks and motion primitive sequences, enriching the movement mode, and being able to flexibly plan according to the actual working conditions. At the same time, planning a collision-free path based on the sequence of motion primitives and the sequence of path points can effectively avoid collisions between the steel grabber and obstacles during operation and ensure operation safety; through task classification and refined planning, it can more flexibly respond to various changes in the operation process, such as changes in the position of steel materials, dynamic adjustment of the working environment, etc. At the same time, since each subtask and motion primitive is carefully designed and planned, the reliability and stability of task execution can be ensured. The hierarchical planning architecture in this application enables complex grasping tasks to be reasonably decomposed, improving the robustness and efficiency of planning.

[0036] In one embodiment of the present application, the overall grasping task is decomposed into a steel material handling subtask and a body movement subtask to obtain a subtask sequence, including: inputting the environmental map into the steel grasping task decomposition planner, decomposing the overall grasping task into a steel material handling subtask and a body movement subtask through the steel grasping task decomposition planner; if the subtask type is a steel material handling subtask, the steel material handling subtask is planned by the steel material handling subtask planner to obtain a steel material handling subtask planning result, and the steel material handling local map is updated, and the steel material handling subtask planning result includes the steel material handling subtask parameters; if the subtask type is a body movement subtask, the body movement subtask is planned by the body movement subtask planner to obtain a body movement subtask planning result, and the body movement subtask planning result includes the body movement subtask parameters; construct a subtask sequence according to the steel material handling subtask planning result and the body movement subtask planning result. For grasping operations in large areas, existing methods often cannot reasonably decompose and plan the entire process. This scheme adopts a hierarchical architecture. SGTDP first decomposes the overall task into a subtask sequence, and then the subtask planner plans the specific subtasks, which improves the robustness and efficiency of planning.

[0037] In one embodiment of the present application, a steel material handling subtask planner is used to plan a steel material handling subtask, and a steel material handling subtask planning result is obtained, including: determining whether the grid operation height in the steel material handling subtask is less than a preset height threshold; if all grid operation heights are less than the preset height threshold, the steel material handling subtask is planned according to the grid operation height, and the steel material handling subtask planning result is obtained; if any grid operation height is greater than the preset height threshold, the grid that can be cleared is grabbed according to the maximum grab height map of the steel grab theory, and the grab map is updated to obtain the steel material handling subtask planning result. By setting a preset height threshold, grids that require special treatment (i.e., grids whose height exceeds the threshold) can be quickly screened out. For grids whose heights are all less than the threshold and are equal, the grid closest to the steel grab base is selected for grabbing, which reduces the moving distance and time of the steel grab, and improves the operation efficiency. By setting a height threshold and taking special treatment measures for grids that exceed the threshold, potential accidents caused by the steel grabber trying to grab too high steel can be avoided, thereby enhancing the safety of the operation. Formulating different grabbing strategies according to the different working heights of the grid can more reasonably distribute the workload of the steel grabber and avoid waste of resources.

[0038] In one embodiment of the present application, the steel handling subtask is planned according to the grid working height, and the steel handling subtask planning result is obtained, including: if the working heights of all grids are less than the preset height threshold, then determine whether the working heights of all grids are the same; if the working heights of all grids are equal, then select the grid closest to the base of the steel grabber for grabbing, if the working heights of all grids are not equal, then select the grid with the highest working height for grabbing, and update the grabbing map to obtain the steel handling subtask planning result. By presetting the height threshold, grids that require special treatment (i.e., grids whose height exceeds the threshold) can be quickly screened out. For grids whose heights are less than the threshold and are equal, the grid closest to the base of the steel grabber is selected for grabbing, which reduces the moving distance and time of the steel grabber and improves the working efficiency. Formulating different grabbing strategies according to different grid working heights can more reasonably distribute the workload of the steel grabber and avoid waste of resources.

[0039] As an example, Figure 4 The planning flow chart of the steel handling subtask planner shown in the embodiment of the present application determines whether the grid working heights in the steel handling subtask are all less than a preset height threshold, for example, the preset height threshold is 6 meters; if any grid working height is greater than 6 meters, then the grid that can be cleared is grabbed according to the theoretical maximum grabbing height map of steel grabbing to obtain the planning result of the steel handling subtask, and the map is updated after moving the steel grabber chassis according to the planning result of the steel handling subtask. If all grid working heights are less than 6 meters and all grid working heights are equal, the grid closest to the base of the steel grabber is selected for grabbing; if all grid working heights are less than 6 meters and if all grid working heights are not equal, the grid with the highest working height is selected for grabbing to obtain the planning result of the steel handling subtask and update the grabbing map until the job is completed. By presetting the height threshold, grids that require special treatment (i.e. grids whose height exceeds the threshold) can be quickly screened out. For grids whose heights are all less than the threshold and are equal, the grid closest to the base of the steel grabber is selected for grabbing, which reduces the moving distance and time of the steel grabber and improves work efficiency. For grids with unequal heights, the grid with the highest height is selected for grabbing first, thereby reducing additional adjustments caused by height differences in subsequent operations, further improving work efficiency. By setting a height threshold and taking special treatment measures for grids that exceed the threshold, potential accidents caused by the steel grabber trying to grab too high steel can be avoided, thereby enhancing the safety of the operation. By formulating different grabbing strategies according to the different conditions of the grid operation height, the workload of the steel grabber can be more reasonably distributed to avoid waste of resources.

[0040] In one embodiment of the present application, each subtask is decomposed into a sequence of motion primitives, including: if the subtask type is a steel handling subtask, then the steel handling motion primitive is obtained through the steel handling subtask parameters and the local environmental information of steel handling; when the subtask type is a body movement subtask, the body movement motion primitive is obtained according to the body movement subtask parameters and the local environmental information of body movement; and a motion primitive sequence is constructed according to the steel handling motion primitive and the body movement motion primitive. When performing hierarchical planning, most of the existing low-level planners only consider the motion primitives in the form of decoupling of the mobile chassis and the robot arm, ignoring the possibility of coordinated motion between the two. In addition to considering conventional decoupled motion primitives, the steel handling subtask planner (STSP) of this scheme also introduces motion primitives for the coordinated motion of the mobile chassis and the robot arm, enriching the motion form.

[0041] In one embodiment of the present application, after constructing the motion primitive sequence, it also includes: executing the machine body movement action according to the machine body movement motion primitive, and executing the steel material handling motion action according to the steel material handling motion primitive; if the motion primitive is the gripping claw motion primitive, then after the steel grabber completes the operation, the environment map is updated to determine the change in the position or state of the steel material. By combining the machine body movement motion primitive and the steel material handling motion primitive, the coordinated movement of the steel grabber during the operation is achieved, ensuring that the steel grabber can move to the target position accurately and efficiently.

[0042] In one embodiment of the present application, the order of the steel material grabbing areas of the overall grabbing task is determined to determine the order of the path points for the chassis of the steel grabber to stop, including: for the steel material handling task, the steel material grabbing path points of the steel material grabbing task are obtained according to the chassis position of the steel grabber; for the body movement subtask, the position of the end effector of the robot arm is set above the steel material, and the path points for the chassis of the steel grabber to stop are determined according to the working space of the robot arm and the steel material environment.

[0043] In one embodiment of the present application, the steel grabbing task decomposition planner (SGTDP) based on NPI is used to classify the steel grabbing tasks, which specifically includes: first, at time t, the environmental information o(t) of the overall grabbing task is input into the steel grabbing task decomposition planner SGTDP, and SGTDP will output the subtask program that needs to be executed at the next step at the current moment and the input parameters of the subtask program. According to the subtask program type, the next subtask planning is executed. If the subtask program type is a steel material handling subtask, the steel material handling subtask planner STSP is required to perform the next steel material handling subtask planning, and the grabbing area map G is generated. block If the subtask program type is a body movement subtask, the body movement subtask planner AMSP is needed to plan the next body movement subtask. Finally, the corresponding subtask program and its input parameters are combined and added to the subtask sequence Φtask In the steel grabbing task decomposition and planning stage, for a given grabbing area map G block , the steel grasping task decomposition planner SGTDP needs to determine the subtask sequence that can complete the overall grasping task. SGTDP is based on the latest grasping area map G block As well as the subtask program trajectories executed in the past period of time, determine the subtask program that the steel grabber needs to execute next. In the training phase, SGTDP needs to map the grasping area G at the current time step t. block (t) and the subtask program trajectory executed in the past k time steps As input, the output is the subtask program at the current time t, the input parameters of the subtask program, and the planning termination identifier {s(t), a(t), t e After obtaining the handling subtask output by the steel grabbing task decomposition planner SGTDP, the steel handling subtask planner STSP obtains local environment information according to the handling subtask program parameters. It is represented as a work unit map G region , and proceed to the next step of planning. When the subtask program is grasp_region, STSP uses local environment information And the subtask program parameters As input, STSP outputs appropriate motion primitives and parameters and determines whether to terminate the mission planner. First, when the steel handling subtask program parameters output by the steel grabbing task decomposition planner are obtained, the steel handling subtask planner STSP obtains the local environment information based on the parameters. Then the motion primitive and its parameters required for the next step at the current moment are output. The steel grabber performs corresponding actions according to the motion primitive and the motion primitive parameters. If the motion primitive is a gripper grab motion primitive, the work unit map G region Update, and finally add the motion primitive sequence of motion primitives and motion primitive parameters output by STSP to the motion primitive sequence Φ primitive As an example, Figure 5The network structure diagram of the NPI-based SGTDP shown in the embodiment of the present application, because the prior art usually sets the operating space of the robot arm as a regular fixed-size hexahedron, ignoring the actual irregular reachable range of the robot arm. By introducing the NPI-based steel grasping task decomposition planner (SGTDP), the overall grasping task can be decomposed according to the actual workspace of the robot arm, thereby making full use of the reachable space of the robot arm. In addition to considering conventional decoupled motion primitives, the steel handling subtask planner (STSP) also introduces motion primitives for the coordinated motion of the mobile chassis and the robot arm, enriching the motion form, and adopting a hierarchical architecture. SGTDP first decomposes the overall task into a subtask sequence, and then the subtask planner plans the specific subtasks, which improves the robustness and efficiency of the planning.

[0044] In one embodiment of the present application, Figure 6 The flowchart of the steel grabber task classification shown in the embodiment of the present application is as follows: Figure 6As shown in the figure, for the grasping task of the whole area, the working space of the robot arm is first approximated, and the overall environmental map is downsampled in combination with the environmental information provided by the environmental perception module to obtain the environmental map. The grasping task is planned by the steel grasping task decomposition planner SGTDP. If the task planning is not completed, the environmental map is updated, and the grasping task planning is performed again for the updated environmental map until the task planning is completed. If the task planning is completed, the grasping order of the whole area is determined, and the subtask sequence is output. The subtask motion primitives in the subtask sequence are decomposed, and it is determined whether the subtask sequence is empty. If the subtask sequence is empty, the running primitive sequence is output. If the subtask sequence is not empty, the first subtask is popped up, and it is determined whether the first subtask is If it is not a grabbing subtask, if the first subtask is a grabbing subtask, the steel handling subtask planner is used to plan the steel handling subtask, obtain the steel handling subtask planning result, and update the map information. The map information includes the local map of steel handling. The steel handling subtask planner plans the steel handling subtask for the updated map information until the subtask sequence is empty, and outputs the steel handling motion primitive; if the first subtask is not a grabbing subtask, the body movement subtask planner is used to plan the body movement subtask, obtain the body movement subtask planning result, until the subtask sequence is empty, and output the body movement motion primitive; construct a motion primitive sequence based on the steel handling motion primitive and the body movement motion primitive, and output the motion primitive sequence. By downsampling the overall operation area of ​​the steel grabber to obtain an environmental map, complex environmental information can be simplified, making subsequent task planning more efficient. Decomposing tasks according to the real working space of the robot arm can maximize the reach of the robot arm and avoid the defect of ignoring the reachable space in the traditional method. At the same time, decomposing the overall grasping task into steel material handling subtasks and body movement subtasks is helpful to realize parallel processing and optimized scheduling of tasks, thereby improving the overall operation efficiency; introducing motion primitives for coordinated movement of the mobile chassis and the robot arm, decomposing the overall task into smaller subtasks and motion primitive sequences, enriching the movement mode, and being able to flexibly plan according to the actual working conditions. At the same time, planning a collision-free path based on the sequence of motion primitives and the sequence of path points can effectively avoid collisions between the steel grabber and obstacles during operation and ensure operation safety; through task classification and refined planning, it can more flexibly respond to various changes in the operation process, such as changes in the position of steel materials, dynamic adjustment of the working environment, etc. At the same time, since each subtask and motion primitive is carefully designed and planned, the reliability and stability of task execution can be ensured. The hierarchical planning architecture in this application enables complex grasping tasks to be reasonably decomposed, improving the robustness and efficiency of planning.

[0045] Figure 7 FIG. 1 is a block diagram of a steel grabber task classification device according to an exemplary embodiment of the present application. Figure 7 As shown, according to an embodiment of the present application, a steel grabber task grading device 700 includes: a sampling module 710, a task decomposition module 720 and a motion primitive module 730. The sampling module is used to downsample the overall operation area of ​​the steel grabber to obtain an environmental map, and obtain the mechanical arm operation space according to the overall grasping task of the overall operation area of ​​the steel grabber; the task decomposition module is used to decompose the overall grasping task into subtasks according to the mechanical arm operation space and the environmental map, obtain a subtask sequence, and determine the order of the steel material grasping area of ​​the overall grasping task to determine the path point sequence of the steel grabber chassis, and the subtasks include steel material handling subtasks and body movement subtasks; the motion primitive module is used to decompose each subtask into a motion primitive sequence, and plan a collision-free path based on the motion primitive sequence and the path point sequence to complete the task grading of the overall grasping task. By downsampling the overall operation area of ​​the steel grabber to obtain an environmental map, complex environmental information can be simplified, making subsequent task planning more efficient. Decomposing tasks according to the real working space of the robot arm can maximize the reach of the robot arm and avoid the defect of ignoring the reachable space in the traditional method. At the same time, decomposing the overall grasping task into steel material handling subtasks and body movement subtasks is helpful to realize parallel processing and optimized scheduling of tasks, thereby improving the overall operation efficiency; introducing motion primitives for coordinated movement of the mobile chassis and the robot arm, decomposing the overall task into smaller subtasks and motion primitive sequences, enriching the movement mode, and being able to flexibly plan according to the actual working conditions. At the same time, planning a collision-free path based on the sequence of motion primitives and the sequence of path points can effectively avoid collisions between the steel grabber and obstacles during operation and ensure operation safety; through task classification and refined planning, it can more flexibly respond to various changes in the operation process, such as changes in the position of steel materials, dynamic adjustment of the working environment, etc. At the same time, since each subtask and motion primitive is carefully designed and planned, the reliability and stability of task execution can be ensured. The hierarchical planning architecture in this application enables complex grasping tasks to be reasonably decomposed, improving the robustness and efficiency of planning.

[0046] In one embodiment of the present application, the task decomposition module is used to input the environmental map into the steel grasping task decomposition planner, and the overall grasping task is decomposed into a steel material handling subtask and a body movement subtask through the steel grasping task decomposition planner; if the subtask type is a steel material handling subtask, the steel material handling subtask is planned by the steel material handling subtask planner to obtain the steel material handling subtask planning result, and the steel material handling local map is updated, and the steel material handling subtask planning result includes the steel material handling subtask parameters; if the subtask type is a body movement subtask, the body movement subtask is planned by the body movement subtask planner to obtain the body movement subtask planning result, and the body movement subtask planning result includes the body movement subtask parameters; a subtask sequence is constructed according to the steel material handling subtask planning result and the body movement subtask planning result. For grasping operations in large areas, existing methods often cannot reasonably decompose and plan the entire process. However, this solution adopts a hierarchical architecture, firstly SGTDP decomposes the overall task into a subtask sequence, and then the subtask planner plans the specific subtasks, which improves the robustness and efficiency of the planning.

[0047] In one embodiment of the present application, the task decomposition module is used to determine whether the grid operating heights in the steel handling subtask are all less than a preset height threshold; if all grid operating heights are less than the preset height threshold, the steel handling subtask is planned according to the grid operating heights to obtain the steel handling subtask planning results; if any grid operating height is greater than the preset height threshold, the clearable grid is grabbed according to the theoretical maximum grab height map for steel grabbing, and the grab map is updated to obtain the steel handling subtask planning results. By setting a height threshold and taking special treatment measures for grids that exceed the threshold, potential accidents caused by the steel grabber attempting to grab too high steel can be avoided, thereby enhancing the safety of the operation. By formulating different grabbing strategies according to different grid operating heights, the workload of the steel grabber can be more reasonably distributed to avoid waste of resources.

[0048] In one embodiment of the present application, the task decomposition module is used to determine whether the working heights of all grids are the same if they are less than a preset height threshold; if the working heights of all grids are equal, the grid closest to the base of the steel grabber is selected for grabbing; if the working heights of all grids are not equal, the grid with the highest working height is selected for grabbing, and the grabbing map is updated to obtain the planning result of the steel handling subtask. By presetting the height threshold, grids that require special treatment (i.e., grids whose height exceeds the threshold) can be quickly screened out. For grids whose heights are less than the threshold and are equal, the grid closest to the base of the steel grabber is selected for grabbing, which reduces the moving distance and time of the steel grabber and improves the working efficiency. By formulating different grabbing strategies according to different grid working heights, the workload of the steel grabber can be more reasonably distributed to avoid waste of resources.

[0049] In one embodiment of the present application, the motion primitive module is used to obtain the steel handling motion primitive through the steel handling subtask parameters and the local environmental information of steel handling if the subtask type is a steel handling subtask; when the subtask type is a body movement subtask, the body movement motion primitive is obtained according to the body movement subtask parameters and the local environmental information of body movement; and a motion primitive sequence is constructed according to the steel handling motion primitive and the body movement motion primitive. When performing hierarchical planning, most of the existing low-level planners only consider the motion primitives in the form of decoupling of the mobile chassis and the robot arm, ignoring the possibility of coordinated motion between the two. In addition to considering conventional decoupled motion primitives, the steel handling subtask planner (STSP) of this scheme also introduces motion primitives for the coordinated motion of the mobile chassis and the robot arm, enriching the motion form.

[0050] As an example, Figure 4 The planning flow chart of the steel handling subtask planner shown in the embodiment of the present application determines whether the grid working heights in the steel handling subtask are all less than a preset height threshold, for example, the preset height threshold is 6 meters; if any grid working height is greater than 6 meters, then the grid that can be cleared is grabbed according to the theoretical maximum grabbing height map of steel grabbing to obtain the planning result of the steel handling subtask, and the map is updated after moving the steel grabber chassis according to the planning result of the steel handling subtask. If all grid working heights are less than 6 meters and all grid working heights are equal, the grid closest to the base of the steel grabber is selected for grabbing; if all grid working heights are less than 6 meters and if all grid working heights are not equal, the grid with the highest working height is selected for grabbing to obtain the planning result of the steel handling subtask and update the grabbing map until the job is completed. By presetting the height threshold, grids that require special treatment (i.e. grids whose height exceeds the threshold) can be quickly screened out. For grids whose heights are all less than the threshold and are equal, the grid closest to the base of the steel grabber is selected for grabbing, which reduces the moving distance and time of the steel grabber and improves work efficiency. For grids with unequal heights, the grid with the highest height is selected for grabbing first, thereby reducing additional adjustments caused by height differences in subsequent operations, further improving work efficiency. By setting a height threshold and taking special treatment measures for grids that exceed the threshold, potential accidents caused by the steel grabber trying to grab too high steel can be avoided, thereby enhancing the safety of the operation. By formulating different grabbing strategies according to the different conditions of the grid operation height, the workload of the steel grabber can be more reasonably distributed to avoid waste of resources.

[0051] In one embodiment of the present application, the motion primitive module is used to perform the machine body movement action according to the machine body movement motion primitive, and to perform the steel material handling motion action according to the steel material handling motion primitive; if the motion primitive is the gripper gripping motion primitive, the environment map is updated after the steel grabber completes the operation to determine the change in the position or state of the steel material. By combining the machine body movement motion primitive and the steel material handling motion primitive, the coordinated movement of the steel grabber during the operation is achieved, ensuring that the steel grabber can move to the target position accurately and efficiently.

[0052] In one embodiment of the present application, the task decomposition module is used to obtain the steel material grabbing path points of the steel material handling task according to the chassis position of the steel grabber; for the machine body movement subtask, the position of the end effector of the robot arm is set above the steel material, and the path point for the chassis of the steel grabber to stop is determined according to the working space of the robot arm and the steel material environment.

[0053] As an example, for the grasping task of the entire area, the working space of the robot arm is first approximated, and the overall environmental map is downsampled in combination with the environmental information provided by the environmental perception module to obtain the environmental map. The grasping task is planned by the steel grasping task decomposition planner SGTDP. If the task planning is not completed, the environmental map is updated, and the grasping task planning is performed again for the updated environmental map until the task planning is completed. If the task planning is completed, the grasping order of the entire area is determined, and the subtask sequence is output. The subtask motion primitives in the subtask sequence are decomposed, and it is determined whether the subtask sequence is empty. If the subtask sequence is empty, the running primitive sequence is output. If the subtask sequence is not empty, the first subtask is popped up, and the first subtask is determined. Whether it is a grabbing subtask, if the first subtask is a grabbing subtask, the steel handling subtask planner is used to plan the steel handling subtask, obtain the steel handling subtask planning result, and update the map information, which includes the local map of steel handling. The steel handling subtask planner plans the steel handling subtask for the updated map information until the subtask sequence is empty, and outputs the steel handling motion primitive; if the first subtask is not a grabbing subtask, the body movement subtask planner is used to plan the body movement subtask, obtain the body movement subtask planning result, until the subtask sequence is empty, and output the body movement motion primitive; construct a motion primitive sequence based on the steel handling motion primitive and the body movement motion primitive, and output the motion primitive sequence. By downsampling the overall operation area of ​​the steel grabber to obtain an environmental map, complex environmental information can be simplified, making subsequent task planning more efficient. Decomposing tasks according to the real working space of the robot arm can maximize the reach of the robot arm and avoid the defect of ignoring the reachable space in the traditional method. At the same time, decomposing the overall grasping task into steel material handling subtasks and body movement subtasks is helpful to realize parallel processing and optimized scheduling of tasks, thereby improving the overall operation efficiency; introducing motion primitives for coordinated movement of the mobile chassis and the robot arm, decomposing the overall task into smaller subtasks and motion primitive sequences, enriching the movement mode, and being able to flexibly plan according to the actual working conditions. At the same time, planning a collision-free path based on the sequence of motion primitives and the sequence of path points can effectively avoid collisions between the steel grabber and obstacles during operation and ensure operation safety; through task classification and refined planning, it can more flexibly respond to various changes in the operation process, such as changes in the position of steel materials, dynamic adjustment of the working environment, etc. At the same time, since each subtask and motion primitive is carefully designed and planned, the reliability and stability of task execution can be ensured. The hierarchical planning architecture in this application enables complex grasping tasks to be reasonably decomposed, improving the robustness and efficiency of planning.

[0054] It should be noted that the device provided in the above embodiment and the method provided in the above embodiment belong to the same concept, wherein the specific manner in which each module and unit performs the operation has been described in detail in the method embodiment, and will not be repeated here. In practical applications, the device provided in the above embodiment can distribute 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.

[0055] See also Figure 8 , Figure 8 is a schematic diagram of a structure of an electronic device shown in an exemplary embodiment of the present application. It should be noted that: Figure 8 The electronic device 800 shown is only an example and should not bring any limitation to the functions and usage area of ​​the embodiments of the present application.

[0056] like Figure 8 As shown, the electronic device 800 includes a processor 801, a memory 802 and a communication bus 803; the communication bus 803 is used to connect the processor 801 and the memory 802; the processor 801 is used to execute the computer program stored in the memory 802 to implement one or more methods in the above embodiments.

[0057] The electronic device provided in the present application includes a processor, a memory, a transceiver and a communication interface. The memory and the communication interface are connected to the processor and the transceiver and complete communication with each other. The memory is used to store computer programs, the communication interface is used to communicate, and the processor and the transceiver are used to run the computer program so that the electronic device executes each step of the above method.

[0058] In this embodiment, the memory may include a random access memory (RAM), and may also include a non-volatile memory (non-volatile memory), such as at least one disk memory.

[0059] The above-mentioned processor can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

[0060] The above embodiments are merely illustrative of the principles and effects of the present application, and are not intended to limit the present application. Anyone familiar with the technology may modify or change the above embodiments without violating the spirit and scope of the present application. Therefore, all equivalent modifications or changes made by a person of ordinary skill in the art without departing from the spirit and technical ideas disclosed in the present application shall still be covered by the claims of the present application.

Claims

1. A steel grabber task classification method, characterized in that: The method comprises: Downsample the overall operation area of ​​the steel grabber to obtain an environmental map, and obtain the robot arm operation space based on the overall grasping task of the overall operation area of ​​the steel grabber; According to the working space of the robot arm and the environmental map, the overall grasping task is decomposed into subtasks to obtain a subtask sequence, and the order of the steel material grasping areas of the overall grasping task is determined to determine the path point sequence of the steel grabber chassis docking, and the subtasks include a steel material handling subtask and a machine body movement subtask; Each subtask is decomposed into a sequence of motion primitives, and a collision-free path is planned based on the sequence of motion primitives and the sequence of path points, thereby completing the task classification of the overall grasping task.

2. The steel grabber task classification method according to claim 1, characterized in that: The overall grasping task is decomposed into a steel material handling subtask and a machine body movement subtask, and a subtask sequence is obtained, including: The environmental map is input into a steel grabbing task decomposition planner, and the overall grabbing task is decomposed into a steel material handling subtask and a machine body movement subtask by the steel grabbing task decomposition planner; If the subtask type is the steel material handling subtask, the steel material handling subtask is planned by a steel material handling subtask planner to obtain a steel material handling subtask planning result, and the steel material handling local map is updated, wherein the steel material handling subtask planning result includes steel material handling subtask parameters; If the subtask type is the body movement subtask, the body movement subtask is planned by the body movement subtask planner to obtain a body movement subtask planning result, wherein the body movement subtask planning result includes body movement subtask parameters; A subtask sequence is constructed according to the planning results of the steel material handling subtask and the planning results of the machine body movement subtask.

3. The steel grabber task classification method according to claim 2, characterized in that: The steel material handling subtask planner is used to plan the steel material handling subtask, and the steel material handling subtask planning results are obtained, including: Determine whether the grid operation height in the steel material handling subtask is less than the preset height threshold; If all grid operation heights are less than the preset height threshold, the steel material handling subtask planning is performed according to the grid operation height to obtain the steel material handling subtask planning result; If the working height of any grid is greater than the preset height threshold, the grid that can be cleared is grabbed according to the theoretical maximum grab height map of steel grabbing, and the grab map is updated to obtain the planning result of the steel handling subtask.

4. The steel grabber task classification method according to claim 3, characterized in that: The steel material handling subtask is planned according to the grid operation height, and the steel material handling subtask planning results are obtained, including: If the heights of all grid operations are less than the preset height threshold, then determine whether the heights of all grid operations are the same; If the working heights of all grids are equal, the grid closest to the steel grabber base will be selected for grabbing. If the working heights of all grids are not equal, the grid with the highest working height will be selected for grabbing, and the grabbing map will be updated to obtain the planning results of the steel handling subtask.

5. The steel grabber task classification method according to claim 2, characterized in that: Decompose each subtask into a sequence of motion primitives, including: If the subtask type is a steel material handling subtask, a steel material handling motion primitive is obtained through the steel material handling subtask parameters and the steel material handling local environment information; When the subtask type is a body movement subtask, obtaining a body movement motion primitive according to the body movement subtask parameter and the body movement local environment information; A motion primitive sequence is constructed according to the steel material handling motion primitive and the machine body movement motion primitive.

6. The steel grabber task classification method according to claim 5, characterized in that: After constructing the sequence of motion primitives, also include: Executing a machine body movement action according to the machine body movement motion primitive, and executing a steel material handling motion action according to the steel material handling motion primitive; If the motion primitive is a gripper gripping motion primitive, the environment map is updated after the steel gripper completes the operation to determine changes in the position or state of the steel material.

7. The steel grabber task classification method according to claim 1, characterized in that: Determine the order of the steel material grabbing areas in the overall grabbing task to determine the path point order for the steel grabber chassis to stop, including: For the steel material handling task, the steel material grabbing path points of the steel material grabbing task are obtained according to the chassis position of the steel grabber; For the machine body movement subtask, the position of the end effector of the robot arm is set above the steel material, and the path point where the chassis of the steel grabber stops is determined according to the working space of the robot arm and the steel material environment.

8. A task grading device for a steel grabber, characterized in that: The device comprises: The sampling module is used to downsample the overall operation area of ​​the steel grabber to obtain an environmental map, and obtain the robot arm operation space according to the overall grasping task of the overall operation area of ​​the steel grabber; A task decomposition module is used to decompose the overall grasping task into subtasks according to the robot arm working space and the environment map, obtain a subtask sequence, and determine the order of steel material grasping areas of the overall grasping task to determine the path point sequence of the steel grasping machine chassis parking, wherein the subtasks include a steel material handling subtask and a machine body movement subtask; The motion primitive module is used to decompose each subtask into a motion primitive sequence, and plan a collision-free path based on the motion primitive sequence and the path point sequence, so as to complete the task classification of the overall grasping task.

9. An electronic device, characterized in that: The electronic device comprises: 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 steel grabber task classification method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: A computer program is stored thereon, and the computer program is used to enable a computer to execute the steel grabber task classification method as described in any one of claims 1 to 7.

Images