A task planning method, system, robot and medium of a construction robot

Abstract

The application provides a task planning method of a construction robot, comprising the following steps: obtaining semantic information of a to-be-constructed area and screening a to-be-constructed object; according to the screened to-be-constructed object, matching appropriate construction types and mechanical construction parameters from a preset process library to generate a construction operation node information list; according to the semantic information and the construction operation node information list, generating a path node information list; according to the path node information list, matching appropriate processes in a preset trajectory action library to generate a specific construction action sequence; and according to the path node information list and the semantic information, judging whether there is interference between the connection lines between adjacent path nodes, if yes, adding a transfer path node, and if not, generating a final construction path. The application can ensure that the construction task of each unit reaches a local optimal solution, can better adapt to the complex environment of a construction site, and can ensure the safety and high efficiency of the construction robot in the construction process.

Description

Technical Field

[0001] This application relates to the field of construction automation technology, specifically to a task planning method, system, robot, and medium for a construction robot. Background Technology

[0002] With the continuous development of the construction industry, intelligent construction has become a key focus for promoting high-quality development in the industry. Construction robots, as a new type of production factor, have become a crucial tool for advancing intelligent construction. However, construction robots face many challenges in practical applications. Currently, there is no mature task planning technology to empower construction robots, and most current control models for construction robots have not yet evolved to the level of automated construction.

[0003] In recent years, reinforcement learning has been applied to task planning in construction robots due to its advantages, such as its ability to handle high-dimensional state and action spaces and its suitability for handling nonlinear and complex relationships. However, reinforcement learning algorithms still face many problems when applied to construction environments, such as the tendency for high-dimensional data input to cause dimensionality explosion, the limited action and state spaces, the sparse reward function, and the lack of successful practical applications in the construction field. Summary of the Invention

[0004] To address the shortcomings of existing technologies, the purpose of this application is to provide a task planning method, system, robot, and medium for construction robots, thereby achieving an efficient, reliable, and highly adaptable task planning method for construction robots and improving the intelligent construction level of construction robots.

[0005] One aspect of this application provides a task planning method for a construction robot, comprising: Obtain semantic information of the area to be constructed and filter the objects to be constructed; Based on the selected objects to be constructed, a list of construction operation node information is generated by matching appropriate construction types and mechanical construction parameters from a preset process library. Based on the semantic information and the list of construction operation node information, a path node information list is generated; Based on the path node information list, within the preset trajectory action library, actions of appropriate processes are matched to generate a specific construction action sequence. Based on the path node information list and the semantic information, it is determined whether there is interference between the connections of adjacent path nodes. If so, a transit path node is added; otherwise, the final construction path is generated.

[0006] Furthermore, the step of obtaining semantic information of the area to be constructed and filtering construction objects includes: Acquire images or point cloud data of the area to be constructed, extract information from the images or point cloud data, and generate semantic information of the area to be constructed. Based on the semantic information of the area to be constructed, determine the orientation of the normal vectors of different information in the semantic information, and identify the construction object; Perform depth calculations on the construction object to determine the distance between the construction object and the building wall; Based on the distance, the objects to be constructed are selected.

[0007] Furthermore, based on the selected objects to be constructed, a suitable construction type and mechanical construction parameters are matched from a preset process library to generate a list of construction operation node information, including: Obtain the filtered objects to be constructed, and determine the size information of the construction area of ​​the objects to be constructed; Based on the size information, determine whether the size information is greater than a set threshold. If so, divide the construction area into multiple construction areas according to the size of the construction area. If not, fix the size process rules. The system matches suitable construction types and mechanical construction parameters from a pre-defined process library to generate a list of construction operation node information.

[0008] Furthermore, the construction operation node information list includes operation node information for multiple construction areas; The set threshold is the size of the construction robot body.

[0009] Further, generating the path node information list based on the semantic information and the construction operation node information list includes: Obtain the list of construction operation node information and the semantic information, and generate information for multiple path nodes corresponding to the operation node information of the multiple construction areas; Based on the information of the multiple path nodes, the correspondence between the job node and the path node is assigned, the three-dimensional coordinate information of the multiple path nodes is generated, and a list of the information of the multiple path nodes is formed.

[0010] Furthermore, the step of matching appropriate process actions within a preset trajectory action library based on the path node information list to generate a specific construction action sequence includes: Obtain a list of path node information, and in conjunction with the semantic information, the construction type, and the mechanical construction parameters, select a trajectory action that matches the list of path node information from a preset trajectory action library; Based on the trajectory actions, the construction section within the work node information is fine-tuned to generate a specific sequence of construction actions.

[0011] Further, the step of determining whether there is interference between adjacent path nodes based on the path node information list and the semantic information, and if so, adding a transit path node; otherwise, generating the final construction path, includes: Obtain the list of path node information and the semantic information, and convert the semantic information into map information; Based on the list of path node information and the map information, the spatial relationship between two adjacent path node information and the map information is obtained; Connect two adjacent path nodes and determine if there is any interference with the map information. If so, add a transit path node to the specific construction action sequence; otherwise, generate the final construction path.

[0012] A second aspect of this application provides a task planning system for a construction robot, comprising: The construction object identification module is used to receive semantic information of the area to be constructed, identify the construction objects, and determine the type of the construction objects. The process classification module is used to match the appropriate construction type and mechanical construction parameters from the preset process library according to the selected construction object, and generate a list of construction operation node information. The path node module is used to generate a path node information list based on the semantic information and the construction operation node information list; The decision trajectory action module is used to match appropriate process actions within a preset trajectory action library based on the path node information list, and generate a specific construction action sequence. The target planning path determination module is used to determine whether there is interference between adjacent path nodes based on the path node information list and the semantic information. If so, an intermediate path node is added; otherwise, the final construction path is generated.

[0013] A third aspect of this application provides a construction robot, comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program being configured to implement a task planning method for the construction robot.

[0014] A fourth aspect of this application provides a storage medium, which is a computer-readable storage medium, on which a computer program is stored, and which, when executed by a processor, implements the task planning method for the construction robot.

[0015] Compared with the prior art, this application has at least one of the following beneficial effects: 1. The construction robot task planning method of this application significantly improves the accuracy and reliability of construction robot task planning through semantic information identification, path node optimization and obstacle avoidance strategies, enhances the adaptability of construction environment, and improves construction efficiency and quality. It has important practical application value and broad development prospects.

[0016] 2. This application identifies construction objects by receiving semantic information from the construction area and matches appropriate construction types and mechanical construction parameters based on the identification results. This enables the construction robot to better adapt to different construction environments and task requirements, improving the flexibility and adaptability of construction. By appropriately increasing transfer stations during construction, reasonable obstacle avoidance and safe posture are ensured, further enhancing the stability and safety of the construction robot in complex construction environments. By dynamically adjusting the path and transfer stations, the construction robot can effectively avoid obstacles and ensure the smooth progress of the construction process.

[0017] 3. This application, through the optimization of path nodes and the generation of construction action sequences, can generate smooth and efficient construction paths, improving construction efficiency and ensuring the uniformity and consistency of construction quality. During task planning, the information of all path nodes is linked together to generate the final construction path, ensuring the continuity and integrity of the construction process, reducing downtime and adjustment time, and improving overall construction efficiency. Attached Figure Description

[0018] Other features, objects, and advantages of this application will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings: Figure 1 This is a flowchart of a task planning method for a construction robot according to one embodiment of this application.

[0019] Figure 2 This is a flowchart of the process for selecting construction objects in one embodiment of this application.

[0020] Figure 3 This is a flowchart illustrating the process of generating a list of construction operation node information in one embodiment of this application.

[0021] Figure 4 This is an example diagram of a construction operation node structure in one embodiment of this application.

[0022] Figure 5 This is a flowchart illustrating the process of generating a list of path node information in one embodiment of this application.

[0023] Figure 6 A flowchart for generating a specific construction action sequence is provided in one embodiment of this application.

[0024] Figure 7This is a flowchart illustrating the final construction path determination in one embodiment of this application.

[0025] Figure 8 This is an example diagram illustrating how transit path nodes are added in one embodiment of this application.

[0026] Figure 9 This is an example diagram of depth calculation in one embodiment of this application. Figure 10 This is a flowchart illustrating the identification of beams and bay windows in a preferred embodiment of the present invention; Figure 11 This is a schematic diagram illustrating the process of identifying a single door or window opening in a preferred embodiment of the present invention; Figure 12 This is a schematic diagram of the process of identifying a single bay window in a preferred embodiment of the present invention; Figure 13 This is a schematic diagram of the process of identifying a single beam in a preferred embodiment of the present invention. Detailed Implementation

[0027] The present application will now be described in detail with reference to specific embodiments. These embodiments will help those skilled in the art to further understand the present application, but do not limit the present application in any way. It should be noted that those skilled in the art can make several modifications and improvements without departing from the concept of the present application. These all fall within the protection scope of the present application.

[0028] Reference Figure 1 As shown, this application provides an embodiment of a task planning method for a construction robot, including: S100, obtaining semantic information of the area to be constructed and filtering the objects to be constructed; S200. Based on the selected objects to be constructed, match the appropriate construction type and mechanical construction parameters from the preset process library to generate a list of construction operation node information. S300. Generate a path node information list based on semantic information and the list of construction operation node information; S400: Based on the list of path node information, match the appropriate process actions in the preset trajectory action library to generate a specific construction action sequence. S500: Based on the list of path node information and semantic information, determine whether there is interference between the connections of adjacent path nodes. If so, add a transit path node; otherwise, generate the final construction path.

[0029] The embodiments described above improve the accuracy and efficiency of task planning by acquiring semantic information of the area to be constructed, intelligently filtering construction objects, and automatically matching appropriate construction types and mechanical parameters. Simultaneously, they can automatically generate path node information and specific construction action sequences, and add intermediate path nodes when necessary to avoid interference, thereby ensuring a smooth and safe construction process and significantly improving the automated construction capabilities and operational efficiency of construction robots.

[0030] Specifically, firstly, by acquiring semantic information of the area to be constructed, the objects requiring construction are selected. Next, based on the selected construction objects, suitable construction types and mechanical construction parameters are searched and matched in a pre-set process library to generate a list of construction operation node information. Then, combining the semantic information and the construction operation node information, a list of path node information is generated. Subsequently, suitable process actions are matched for the path nodes in a pre-set trajectory action library to form a specific construction action sequence. Finally, based on the path node information and semantic information, it is determined whether there is interference between adjacent path nodes. If interference exists, intermediate path nodes are added to avoid conflict; if no interference exists, the final construction path is directly generated for the construction robot to execute.

[0031] Among them, the construction action sequence in the task planning of construction robots decomposes complex tasks into a set of instructions for orderly actions, ensuring that the robot can complete construction tasks efficiently and accurately, can dynamically adjust according to the real-time environment, support multi-robot collaborative operation, and transform design intent into actual operation, realizing the effective connection between design and construction. This is the key to realizing automated construction of construction robots and is of great significance to improving construction efficiency, quality and flexibility.

[0032] Reference Figure 2 As shown, in some possible embodiments, obtaining semantic information of the area to be constructed and filtering the objects to be constructed includes: obtaining image or point cloud data of the area to be constructed, extracting information from the image or point cloud data, and generating semantic information of the area to be constructed; determining the orientation of the normal vectors of different information in the semantic information based on the semantic information of the area to be constructed, determining the construction objects, performing depth calculation on the construction objects, and determining the distance between the construction objects and the building wall; and filtering the objects to be constructed based on the distance.

[0033] In this embodiment of the application, the presence of holes and beams in the construction object results in inconsistent semantic information for each wall, which may present as quadrilaterals, hexagons, or even octagons. If these polygons are not pre-processed by clipping, and a specialized algorithm is not applied to each case, the final code will be very complex and redundant, and the program complexity will increase accordingly.

[0034] Specifically, firstly, image or point cloud data of the area to be constructed is acquired using image acquisition equipment or a LiDAR system, including structural elements such as holes and beams within the construction area. Next, image processing techniques or point cloud analysis algorithms are used to extract key information from the data, such as the shape, size, and location of holes, generating semantic information for the area to be constructed. Each object or area is assigned a specific semantic label, such as "entrance hall," "ceiling," or "wall." Subsequently, based on the generated semantic information, the orientation of the normal vectors for different structures is calculated. The normal vector is a mathematical tool describing the orientation of an object's surface. By analyzing the orientation of the normal vectors, the area identified as the target construction object is determined.

[0035] This application's embodiments automatically acquire and analyze image or point cloud data of the area to be constructed, and generate detailed semantic information. This allows for a rapid and accurate understanding of the overall situation and specific details of the construction area, enabling the development of a more scientific and reasonable construction plan. Furthermore, determining the construction object based on the orientation of the normal vectors of different information within the semantic information ensures the targeted and effective nature of the construction operations, avoids unnecessary misoperations and resource waste, and improves construction safety.

[0036] One specific implementation of the process for identifying door and window openings involves acquiring the geometric contour of the construction area by obtaining an image or point cloud data of the area to be constructed, and then identifying the door and window openings based on the geometric contour. Specifically, this includes: S111, receive the detected geometric contours of the wall surface; S112, starting from the lower left corner of the wall Figure 11 The red dot in the middle is used to set the traversal starting point in the right and upward directions parallel to the wall. Figure 11 The blue and red points represent the points traversed to the right and upward, respectively; from each starting point, a ray is emitted in the opposite direction to the wall's normal vector. Figure 11 The purple arrow in the middle indicates the emitted ray.

[0037] S113, listen for events where no collision has occurred: The absence of a collision event here is deduced by working backwards from a collision event. For example, given the propagation speed of the ray and the distance to be detected, the approximate time required for a collision event can be calculated. Therefore, if no collision event is detected within this timeframe, it is considered that no collision event has occurred.

[0038] If a collision occurs and no hole is detected, return to continue the traversal; If no collision occurs, the coordinates of the point where no collision occurs will be used as the lower left coordinates of the hole. S114, starting from the lower left corner of the hole, set traversal starting points in the right and up directions parallel to the wall, and emit rays in the opposite direction of the wall normal vector at each starting point; S115, listen for collision events: If no collision occurs, it means that no hole boundary has been detected, and the traversal continues; If a collision occurs, the hole boundary is detected and added to the semantic information of the current wall.

[0039] In some embodiments, it is further determined whether the upper and lower limits of the hole are flush with the wall surface. If they are flush, the hole is an area that was not detected in the contour recognition stage, the wall contour information is updated, and the hole information is deleted. If they are not flush, the hole boundary is further detected and added to the semantic information of the current wall surface.

[0040] Based on the obtained semantic information about the holes, further steps are taken to identify bay windows and beams, such as... Figure 10 and Figure 12 As shown in Figures (a)-(e). Specifically: S115, the geometric profile of the receiving hole; S116, with the lower left corner of the hole as the reference point, Figure 12 The red dot in the middle is shifted a distance 'a' to the left, and the starting point for traversal is set to the right. Figure 12 The blue dots represent the lattice that emits rays (purple arrows); S117, listen for events where no collision has occurred: If a collision occurs, a floating window is detected, and the traversal continues; If no collision occurs, the depth information of the bay window is obtained. Figure 12 (The purple double arrow in diagram (e)). S118, Determine if the bay window depth information is less than the set process parameters: If the value is greater than the value, it indicates that there is a bay window at the hole, and the relevant semantics will be updated to the current wall.

[0041] If the bay window depth information is less than the set process parameters, it indicates that there is no bay window at the opening. Further investigation is needed to determine if a beam exists. Figure 11 and Figure 13 As shown in Figures (a)-(e). Specifically: S119, with the lower left corner of the hole as the reference point, Figure 13 The red dot in the middle is shifted a distance 'a' to the left, and the starting point for traversal is set to the right. Figure 13 The blue dots represent the lattice that emits rays (purple arrows); S120, listens for events where no collision has occurred: If a collision occurs and a beam is detected, continue the traversal; If no collision occurs, obtain the distance of the beam protrusion; S121, Determine the distance of the beam protrusion ( Figure 13Is the purple double arrow in diagram (e) less than the set process parameter? If the value is greater than the value, then there is a beam at the hole, and the relevant semantics are updated to the wall surface. If the size is smaller, then there is no beam at the hole; it is just a hole.

[0042] The above embodiments further supplement the semantic information of the object to be identified, including filtering, merging, and cropping holes, as well as identifying features such as beams and bay windows. This makes the semantic information richer and more detailed, improving the accuracy of task planning.

[0043] The above embodiments utilize the ray casting function in the 3D engine, combined with an efficient collision detection algorithm, to quickly obtain the coordinates and normal direction of the collision point. This significantly reduces the computational load and improves the efficiency of information acquisition.

[0044] The process involves calculating the depth of the construction object to determine its distance from the building wall. Based on this distance, objects to be constructed are selected. Specifically, this includes: acquiring the construction object and dividing the planned surface into multiple sub-surfaces based on information such as door and window openings and beams; calculating the depth of each sub-surface to determine the distance between the construction object and the building wall; and classifying objects whose distance from the building wall exceeds a preset threshold as walls. A complete wall is then determined to be either a regular wall or an entryway wall based on its depth information.

[0045] Among them, such as Figure 9 As shown in the figure, distance 1 and distance 2 to the wall refer to the depth of the building wall.

[0046] See attached document Figure 3 As shown, in some possible embodiments, based on the selected objects to be constructed, a suitable construction type and mechanical construction parameters are matched from a preset process library to generate a list of construction operation node information, including: obtaining the type of the objects to be constructed and determining the size information of the construction area of ​​the objects to be constructed; based on the size information, determining whether the size information is greater than a set threshold; if so, dividing the construction area into multiple construction areas according to the size of the construction area; if not, fixing the size process rules; and matching a suitable construction type and mechanical construction parameters from a preset process library to generate a list of construction operation node information.

[0047] By intelligently matching construction types and mechanical construction parameters, construction efficiency is effectively improved. At the same time, based on the type and size information of the construction object, the most suitable construction plan is automatically selected from the preset process library, avoiding the errors and inefficiencies that may be caused by manual selection, thereby ensuring construction quality and progress.

[0048] Reference Figure 4As shown, specifically, the process begins by acquiring the type of the construction object and determining its dimensions. Next, based on this dimensional information, the system determines whether it exceeds a set threshold. If the dimensions exceed the threshold, the system calculates the number of construction zones to be divided based on the size of the planned area, in order to rationally arrange the construction. If the dimensions do not exceed the threshold, fixed dimensional process rules are used for construction. Finally, the system automatically matches the most suitable construction type and mechanical construction parameters from a preset process library and generates a list of construction operation node information for the construction robot to execute.

[0049] In the construction area, the objects of construction are actually the walls and ceilings, which have become relatively independent quadrilaterals after pretreatment. For the quadrilaterals, the dimensional information is the length and width.

[0050] In particular, the ceiling has an additional ceiling height, and the walls have additional bay depth dimensions. Bay depth is a very common technical term in the construction field and has been a long-standing construction indicator, belonging to national standard information.

[0051] Decision-making on process type based on dimensional information is a complete and self-developed mapping function. For example, when the wall width is between 700 and 1100 mm, process A is used; when it is between 1500 and 3000 mm, process B is used.

[0052] In this application, the fixed-size process rules are the preset process library. Based on this, for example, 700~1100mm and 1500~3000mm are fixed sizes, meaning that only one type of process can correspond to this size. Variable sizes can be 3000~+∞.

[0053] The pre-set process library includes the correspondence between processes and construction materials, as well as matching construction parameters, such as spray gun movement speed, spray gun distance from the wall, and effective spray width. For example, if process A is to be performed in a certain construction scenario, then the material α (material brand, model, etc.) suitable for process A recorded in the process library must be used, and the corresponding parameters must be extracted.

[0054] In some specific embodiments, the construction operation node information list includes operation node information for multiple construction areas. The set threshold is the size of the construction robot body.

[0055] Specifically, for the construction robot itself, there are the length and width of the chassis, the height of the body, the safe rotation diameter of the robotic arm in its maximum extended position, and the safe rotation diameter of the construction robot in its retracted position. These values ​​will serve as threshold references for subsequent decisions. For example, the body width of a normal adult is 40-50cm, so 40cm is used as a common-sense threshold. A normal adult cannot cross a narrow passage with a width of less than 40cm sideways, so it is determined that "it is necessary to pass sideways".

[0056] When generating a list of construction task node information, it is necessary to create a list of task node information for each construction area. Alternatively, a number of task node information can be calculated for each construction area and then integrated into a list. For example, if the list of task node information has 100 nodes, nodes 1-20 belong to area A, nodes 20-50 belong to area B, and so on. Each task node will be marked with a corresponding label.

[0057] Reference Figure 5 As shown, in some possible embodiments, generating a path node information list based on semantic information and a list of construction operation node information includes: obtaining the list of construction operation node information and semantic information, generating information on multiple path nodes corresponding to the operation node information of multiple construction areas; allocating the correspondence between operation nodes and path nodes based on the information of multiple path nodes, generating three-dimensional coordinate information of multiple path nodes, and forming a list of multiple path node information.

[0058] Specifically, the process involves obtaining a list of construction operation nodes, setting up several path node information based on information such as machinery type, process type, and machinery parameters, assigning the correspondence between operation nodes and path nodes, setting the three-dimensional coordinate information of path nodes, and creating a list of path node information.

[0059] Reference Figure 6 As shown, in some possible embodiments, based on the list of path node information, actions of suitable processes are matched within a preset trajectory action library to generate a specific construction action sequence. This includes: obtaining a list of path node information; combining semantic information, construction type, and the mechanical construction parameters; selecting a trajectory action that matches the list of path node information within the preset trajectory action library; and fine-tuning the construction interval within the work node information based on the trajectory action to generate a specific construction action sequence.

[0060] One specific implementation method is as follows: For example, there is a wall in a construction area, which is 4m wide, 3m high, and has a width and depth of 4m and 5m respectively. These data are the initial semantic information of the wall that needs to be input from the interface.

[0061] The safe rotation diameter of the construction robot when the robotic arm is extended to its maximum posture is 2.6m. This value is a threshold. If 2.6m is less than the depth of the current construction area of ​​4m, the construction robot can be allowed to work in this area. The construction type is process A.

[0062] Under process A, the radius of a single operation by the construction robot is 1m. Therefore, four operation nodes need to be set up to completely cover a 4m wide wall. The relative positions of the path nodes (taking the leftmost side of the wall as an example) are 0.5m, 1.5m, 2.5m and 3.5m respectively.

[0063] For each path node, the trajectory action of process A from 0 to 3m is executed sequentially. The relative position information contained in the above four path nodes, for the process A to be executed, and the starting point (0~3m), form a list with information for 4 nodes.

[0064] Of course, the above are only some specific implementations of this application. In other embodiments, there may be other parameter values ​​and construction action sequences.

[0065] Reference Figure 7 As shown, in some possible embodiments, based on the path node information list and semantic information, it is determined whether there is interference between the connections of adjacent path nodes. If so, a transit path node is added; otherwise, the final construction path is generated. This includes: obtaining the path node information list and semantic information, and converting the semantic information into map information; obtaining the spatial relationship between the path node information list and map information and the map information based on the path node information list and map information; connecting the adjacent path node information and determining whether there is interference with the map information. If so, a transit path node is added to the specific construction action sequence; otherwise, the final construction path is generated.

[0066] Specifically, the process begins by acquiring a list of path node information and related semantic information. This semantic information is then converted into map information for spatial relationship analysis. Next, based on the path node information list and map information, the spatial relationships between adjacent path nodes and map elements are determined. Then, adjacent path nodes are connected, and the connection is checked for interference with obstacles or other impassable areas on the map. If interference is detected, one or more intermediate path nodes are automatically inserted into the construction action sequence to avoid obstacles. If there is no interference, the final construction path is directly generated.

[0067] Reference Figure 8As shown, by intelligently analyzing the spatial relationship between path nodes and map information, potential interference problems in the construction path are effectively avoided. When interference is detected, intermediate path nodes are automatically added to optimize the construction path, ensuring the smooth progress of construction, improving construction efficiency, and reducing the risk of construction delays and increased costs caused by improper path planning.

[0068] Based on the same technical concept, in another embodiment of this application, a task planning system for a construction robot is provided, comprising: a construction object identification module, used to receive semantic information of the area to be constructed, identify the construction object, and determine the type of the construction object; a process classification module, used to match suitable construction types and mechanical construction parameters from a preset process library according to the selected construction object, and generate a list of construction operation node information; a path node module, used to generate a path node information list according to the semantic information and the list of construction operation node information; a decision trajectory action module, used to match suitable process actions from a preset trajectory action library according to the path node information list, and generate a specific construction action sequence; and a target planning path determination module, used to determine whether there is interference between the connections of adjacent path nodes according to the path node information list and the semantic information, and if so, add a transfer path node, otherwise generate the final construction path.

[0069] Specifically, firstly, the construction object identification module receives semantic information of the area to be constructed, identifies the construction object, determines whether the object is a ceiling, wall, or entrance hall, and passes the identification result to the process classification module.

[0070] The acquired data consists of semantic information about the area to be constructed, including images and point cloud data.

[0071] Next, the process classification module matches the appropriate construction type and mechanical construction parameters from the preset process library based on the type of construction object. The matched construction type and parameters are then passed to the path node module. The data obtained is the type of the construction object.

[0072] Then, the path node module obtains the construction type and mechanical construction parameters, wall semantic information, sets path nodes to ensure the rationality and integrity of the construction path, and transmits the path node information to the decision trajectory action module.

[0073] Next, the decision trajectory action module obtains path node information and construction type, matches appropriate actions for the current process in the preset trajectory action library, generates a specific construction action sequence, and passes the generated construction action sequence to the target planning path determination module.

[0074] Finally, the target planning and path determination module acquires the construction action sequence, path node information, connects all path node information, appropriately adds transfer stations, ensures reasonable obstacle avoidance and safe posture, and generates the final construction path. This final construction path is then passed to the construction robot execution module.

[0075] Through the collaborative work of the above modules and their sub-modules, the construction robot task planning system of this application can achieve efficient and accurate task planning and path generation, ensuring the safety and efficiency of construction robots in complex construction environments.

[0076] It should be noted that the modules in the task planning system for construction robots provided in the above embodiments of this application correspond to the steps of the task planning method for construction robots in any of the above embodiments. Those skilled in the art can refer to the step features of the task planning method for construction robots to implement the corresponding modules in the task planning system for construction robots, which will not be elaborated here.

[0077] In another embodiment of this application, a construction robot is provided, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program being configured to implement a task planning method for the construction robot.

[0078] In another embodiment of this application, a storage medium is provided, which is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, it implements a task planning method for a construction robot.

[0079] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0080] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0081] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0082] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0083] Although preferred embodiments of this application have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of this application.

[0084] The specific embodiments of this application have been described above. It should be understood that this application is not limited to the specific embodiments described above, and those skilled in the art can make various modifications or variations within the scope of the claims, which do not affect the substantive content of this application. The above-described preferred features can be used in any combination without conflict.

Claims

1. A task planning method for a construction robot, characterized in that, include: Obtain semantic information of the area to be constructed and filter the objects to be constructed; Based on the selected objects to be constructed, a list of construction operation node information is generated by matching appropriate construction types and mechanical construction parameters from a preset process library. Based on the semantic information and the list of construction operation node information, a path node information list is generated; Based on the path node information list, within the preset trajectory action library, actions of appropriate processes are matched to generate a specific construction action sequence. Based on the path node information list and the semantic information, determine whether there is interference between the connections between adjacent path nodes. If so, add a transit path node; otherwise, generate the final construction path. The process of obtaining semantic information about the area to be constructed and filtering construction objects includes: Acquire images or point cloud data of the area to be constructed, extract information from the images or point cloud data, and generate semantic information of the area to be constructed. Based on the semantic information of the area to be constructed, determine the orientation of the normal vectors of different information in the semantic information, and identify the construction object; Perform depth calculations on the construction object to determine the distance between the construction object and the building wall; Based on the distance, the objects to be constructed are selected; The step involves matching suitable construction types and mechanical construction parameters from a preset process library based on the selected objects to be constructed, and generating a list of construction operation node information, including: Obtain the filtered objects to be constructed, and determine the size information of the construction area of ​​the objects to be constructed; Based on the size information, determine whether the size information is greater than a set threshold. If yes, divide the construction area into multiple construction areas according to the size of the construction area. If no, fix the size process rules. The system matches suitable construction types and mechanical construction parameters from a pre-defined process library to generate a list of construction operation node information. The list of construction operation node information includes operation node information for multiple construction areas; The set threshold is the size of the construction robot body; The step of generating a path node information list based on the semantic information and the construction operation node information list includes: Obtain the list of construction operation node information and the semantic information, and generate information for multiple path nodes corresponding to the operation node information of the multiple construction areas; Based on the information of the multiple path nodes, the correspondence between the job node and the path node is assigned, the three-dimensional coordinate information of the multiple path nodes is generated, and a list of the information of the multiple path nodes is formed.

2. The task planning method for a construction robot according to claim 1, characterized in that, The step of matching appropriate process actions within a preset trajectory action library based on the path node information list to generate a specific construction action sequence includes: Obtain a list of path node information, and in conjunction with the semantic information, the construction type, and the mechanical construction parameters, select a trajectory action that matches the list of path node information from a preset trajectory action library; Based on the trajectory actions, the construction section within the work node information is fine-tuned to generate a specific sequence of construction actions.

3. The task planning method for a construction robot according to claim 1, characterized in that, The step of determining whether there is interference between adjacent path nodes based on the path node information list and the semantic information, and if so, adding a transit path node; otherwise, generating the final construction path, includes: Obtain the list of path node information and the semantic information, and convert the semantic information into map information; Based on the list of path node information and the map information, the spatial relationship between two adjacent path node information and the map information is obtained; Connect two adjacent path nodes and determine if there is any interference with the map information. If so, add a transit path node to the specific construction action sequence; otherwise, generate the final construction path.

4. A task planning system for a construction robot, employing the task planning method for a construction robot as described in any one of claims 1-3, characterized in that, include: The construction object identification module is used to receive semantic information of the area to be constructed, identify the construction objects, and determine the type of the construction objects. The process classification module is used to match the appropriate construction type and mechanical construction parameters from the preset process library according to the selected construction object, and generate a list of construction operation node information. The path node module is used to generate a path node information list based on the semantic information and the construction operation node information list; The decision trajectory action module is used to match appropriate process actions within a preset trajectory action library based on the path node information list, and generate a specific construction action sequence. The target planning path determination module is used to determine whether there is interference between adjacent path nodes based on the path node information list and the semantic information. If so, a transit path node is added; otherwise, the final construction path is generated.

5. A construction robot, characterized in that, include: A memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program being configured to implement the task planning method for a construction robot as described in any one of claims 1 to 3.

6. A storage medium, characterized in that, The storage medium is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, it implements the task planning method for the construction robot as described in any one of claims 1 to 3.

Images