A Low-Code Manipulator Task Configuration Parsing Method

Abstract

A low-code robotic arm task configuration parsing method of the present invention reads the task configuration information written by the user, converts it into a structured format file using a parsing engine, and forms a task description in a tree structure; clarifies the logical association between each state module through a semantic analysis and hierarchical relationship recognition algorithm; predicts the behavior of the state module with the help of machine learning, constructs a data model that conforms to the tree topology structure to support the dynamic adjustment of the task sequence; uses an intelligent recommendation system to evaluate the importance of nodes and automatically supplement default values, combines deep learning and natural language processing technologies, automatically detects and corrects errors in the configuration information, and the generated data structure is directly used to guide the robotic arm to execute tasks, allowing the execution status to be monitored through a visualization tool to ensure that the tasks are carried out in a predetermined order and can quickly respond to abnormal situations. This method significantly improves the efficiency and accuracy of task configuration and enhances the intelligent level of the robotic arm system.

Description

Technical Field

[0001] The present invention relates to the technical field of mechanical automation control, and particularly to a low-code robotic arm task configuration parsing method. Background Art

[0002] With the continuous improvement of industrial automation level, robotic arms have been widely used in many fields such as manufacturing, warehousing and logistics. However, traditional robotic arm task configuration methods usually require writing complex programming codes, which not only increases the burden on users, but also restricts non-professional programmers from participating in the design of robotic arm tasks. In addition, traditional methods lack intelligent support during the configuration process, resulting in inflexible task configuration, difficult to adapt to rapidly changing production requirements, and prone to errors, affecting production efficiency.

[0003] Traditional robotic arm task configuration technologies mainly rely on programmers to directly write control logic through programming languages. Although this method is powerful, it is very complex for operators without programming skills. This means that every time the robotic arm task is changed, professional personnel need to intervene, greatly increasing the time and cost. In addition, due to the lack of an effective error detection mechanism, configuration errors are often difficult to detect in time, resulting in an extended debugging cycle and low production efficiency. Although some graphical programming tools have tried to simplify this process in recent years, they still fail to completely solve the difficulties of non-programmers in task configuration.

[0004] Currently, with the development of artificial intelligence technology, some advanced robotic arm control systems have begun to introduce automated configuration and error detection functions to reduce human intervention. These systems usually use natural language processing technology to parse the user's intention and generate control codes with the assistance of artificial intelligence. However, the intelligence of these systems is limited, and they can often only handle relatively simple task configurations, and are still insufficient for complex multi-step task configurations. In addition, when facing the hierarchical relationship and logical association in task configuration, the existing technology often requires a large number of manual settings, which is not only time-consuming and laborious, but also prone to human errors.

[0005] In view of the above problems, the present invention proposes a low-code robotic arm task configuration parsing method. Through a graphical interface and natural language processing technology, this method allows users to design the robotic arm task process in a more intuitive way, significantly reducing the difficulty of task configuration. At the same time, with the help of artificial intelligence and machine learning technologies, this method can automatically identify the hierarchical relationship in the configuration information during the parsing process, and predict the behavior pattern of each state module through a pre-trained model, thereby optimizing the task execution logic. In addition, the present invention also introduces an intelligent recommendation system and deep learning technology, which can detect and correct configuration errors during the parsing process to ensure the accuracy and reliability of task configuration. Generally speaking, through the intelligent configuration parsing method, this technical solution significantly improves the flexibility and robustness of the robotic arm task configuration, reduces the error rate, and improves the production efficiency. Summary of the Invention

[0006] A low-code robotic arm task configuration parsing method includes the following steps:

[0007] S1: Read the low-code task configuration information written by the user in the form of a graphical interface and natural language. The configuration information is automatically converted into a standardized YAML or JSON format file by a parsing engine based on semantic understanding and machine learning models. This file uses a tree structure to describe the robotic arm task, and each node represents a state module;

[0008] S2: Identify semantic units through a semantic analysis module, use a hierarchical relationship recognition algorithm to parse the hierarchical structure in the configuration information, identify the logical association between nodes, and confirm the specific content of the state module;

[0009] S3: Use a machine learning model to predict the behavior pattern of the state module, construct an internal data structure through recursive or iterative methods, map the hierarchical relationship in the configuration information to a software data structure that satisfies the tree topology structure, and at the same time support dynamic adjustment of the task sequence to enhance the flexibility of the configuration;

[0010] S4: During the parsing process, use an intelligent recommendation system to detect the necessity and optionality of nodes. For necessary nodes, ensure that they exist and the corresponding information is set; for optional nodes, the intelligent recommendation system automatically fills in reasonable default values according to the context environment to ensure the robustness of the parsing process;

[0011] S5: Use deep learning technology to detect syntax errors and logical errors in the configuration information, combine natural language processing technology to perform automatic error correction and provide correction suggestions to enhance the reliability of the parsing process;

[0012] S6: After completion of the parsing, the generated internal data structure is passed into the execution logic, supporting real-time monitoring of the task execution status through a visualization tool, ensuring that the robotic arm can execute tasks according to the predetermined sequence and dependency relationships, and quickly responding and taking corresponding handling measures through an intelligent decision-making system in case of abnormal situations.

[0013] Preferably, in S1, the graphical interface allows users to intuitively design the robotic arm task process by means of dragging and selection, understand and convert the natural language description input by the user through a natural language processing engine, and convert it into a standard YAML or JSON format file, reducing the user's need for professional programming knowledge.

[0014] Preferably, S1 specifically includes the following steps:

[0015] S1.1: Start the natural language processing module, perform a preliminary parsing on the natural language description input by the user, extract the keyword actions "move to", "grab", "release" and their parameters in the description, and use them to construct a preliminary task intention model;

[0016] S1.2: Use the semantic understanding model to perform semantic parsing on the extracted keywords, understand the meaning of the vocabulary in a specific context, and map it to the corresponding robotic arm actions; when the semantic understanding model is not precise enough in describing the keywords, infer the correct meaning of the keywords through the context.

[0017] S1.3: Integrate the parsing results into a standardized YAML or JSON format file, and all nodes in the file contain complete status module information, with a clear and definite hierarchical relationship, facilitating further processing by the subsequent parsing engine.

[0018] Preferably, S2 specifically includes the following steps:

[0019] S2.1: Identify the semantic units in the configuration information through the semantic analysis module;

[0020] S2.2: Analyze the hierarchical relationship in the configuration information through the hierarchical relationship recognition algorithm, automatically identify the logical association between the parent node and the child node, and ensure that the relationship between the status module and the upper and lower nodes is correct.

[0021] S2.3: Use the context awareness algorithm to refine the context environment of the status module, making the description of the status module grammatically and logically correct and conforming to the actual application scenario.

[0022] S2.4: Confirm the specific content of the status module, including but not limited to TCP position, pose, speed, acceleration, path planning constraint conditions, peripheral operations, and exception handling logic, to ensure logical consistency and meet the requirements of the robotic arm for task execution;

[0023] S2.5: Generate a semantic analysis report, which contains the detailed information and hierarchical relationships of the status module, enabling subsequent steps to perform accurate parsing based on the semantic analysis report.

[0024] Preferably, the specific steps of the machine learning model in S3 are as follows:

[0025] S3.11: Collect and organize historical task configuration data, which includes successfully executed task configurations and actual operation results; form a dataset of the robotic arm's task behavior patterns through the historical task configuration data, serving as the basis for training the machine learning model.

[0026] S3.12: Train the machine learning model through a supervised learning method; the input of the machine learning model is the description information of the status module, and the output is the predicted behavior pattern; during the training process, the model learns the mapping relationship between different configuration information and actual behavior patterns, and predicts the expected behavior pattern based on the input status module description.

[0027] S3.13: During the parsing process, input the information of the status module into the trained machine learning model. The machine learning model predicts the behavior pattern of the status module according to experience and rules, and converts it into a part of the internal data structure.

[0028] Preferably, the specific steps for S3 to construct the internal data structure are as follows:

[0029] S3.21: Initialize the internal data structure and create a root node as the starting point of the entire task configuration; the root node serves as the top-level container for all subsequent status modules; the creation of the root node marks the start of the construction of the internal data structure and is also the basis for the recursive or iterative process.

[0030] S3.22: For the status module, after predicting its behavior pattern using the machine learning model, create corresponding nodes in the internal data structure according to the prediction results, and add the newly created nodes as child nodes or sibling nodes of the existing nodes to the internal data structure according to the hierarchical relationships defined in the configuration file, achieved through recursive calls; the recursive call is when a new status module is parsed and its behavior pattern is predicted, then a recursive function is called; the recursive function adds new nodes and recursively processes the sub-status modules in the status module until all status modules are parsed and added to the internal data structure.

[0031] S3.23: During the construction of the internal data structure, track the hierarchical relationship to ensure the correct parent-child relationship of nodes; at the same time, support dynamic adjustment of the task sequence, insert or delete status modules according to the actual situation during the construction process, and achieve this by iteratively checking the dependency relationship of status modules and updating the internal data structure when necessary, thereby enhancing the flexibility and adaptability of the configuration.

[0032] Preferably, the specific steps of the necessary nodes in S4 include the following:

[0033] S4.11: Define the necessary node categories, including but not limited to the key attributes of the status module: TCP position, pose, speed, acceleration; the necessary nodes form the basis for an effective task configuration and enable the correct execution of tasks; the intelligent recommendation system checks whether the status module contains the necessary attributes to ensure reasonable attribute settings;

[0034] S4.12: Implement the necessary node detection logic. The system will verify one by one whether the status module contains the necessary nodes during the process of parsing the configuration file, and check whether the necessary nodes have been correctly set; if it is detected that a necessary node is missing or the necessary node is incompletely set, the parsing will be stopped immediately, and the user will be prompted through a visual interface or a message about the specific missing items to guide the user to supplement or correct them.

[0035] Preferably, the specific steps of the optional nodes in S4 include the following:

[0036] S4.21: Define the optional node categories, including but not limited to speed ratio, acceleration ratio, path planning constraint conditions, peripheral operations, and exception handling logic; the optional nodes are the optional existence of the status module, which can enhance the flexibility and robustness of tasks in specific situations;

[0037] S4.22: Implement the optional node detection logic; during the process of parsing the configuration file by the intelligent recommendation system, identify which nodes belong to the optional type and check whether the nodes are defined; if the nodes are not defined, the intelligent recommendation system will automatically fill in reasonable default values according to the context environment;

[0038] S4.23: Using big data analysis and machine learning technologies, the intelligent recommendation system can learn and recommend best practices based on successful cases of past task configurations; when it is detected that an optional node is not defined, the system can fill in the default value and provide optimization suggestions according to the characteristics of the current task.

[0039] Preferably, the steps included in the deep learning technology detection in S5 are as follows:

[0040] S5.1: Build an error detection model; use long short-term memory network deep learning technology to train an error detection model for low-code task profiles; the error detection model is trained based on a large number of labeled correct profiles and error profiles, and can identify common syntax errors, logical errors, and other potential problems; the training dataset of the error detection model includes instances of missing necessary fields, unreasonable parameter settings, and incorrect logical order, enabling the error detection model to have comprehensive error recognition capabilities;

[0041] S5.2: Perform error detection; during the process of parsing the configuration file, input the configuration file line by line and segment by segment into the trained long short-term memory network deep learning model. The long short-term memory network deep learning model will analyze the text of the configuration file to detect existing errors. If an error is detected, the error detection model will output the specific error type and its location;

[0042] S5.3: Correct errors through natural language processing technology; when an error is detected, use natural language processing technology to automatically generate correction suggestions and perform correction operations;

[0043] S5.4: Provide a detailed correction report; after completing error detection and correction, generate a correction report, which includes the detected errors and their locations, detailed correction suggestions for the errors, and the executed correction operations.

[0044] Preferably, the intelligent decision-making system of S6 includes an anomaly detection module;

[0045] The anomaly detection module uses machine learning algorithms to analyze the real-time data generated during the execution of the robotic arm task, and can identify and issue an alarm for situations with excessive position deviation, abnormal speed, and external interference from the expected behavior;

[0046] When the anomaly detection module detects an abnormal situation, the intelligent decision-making system automatically selects appropriate countermeasures according to the type and severity of the anomaly, including but not limited to pausing task execution, reverting to the previous stable state, attempting automatic repair, requesting manual intervention, and executing corresponding commands.

[0047] Compared with the prior art, the technical solution of the present invention has the following technical effects:

[0048] The present invention solves the problem that non-programming professionals have difficulty writing robotic arm task configuration codes through a graphical interface and natural language processing technology. Users can intuitively design the robotic arm task process by dragging and selecting, and convert natural language descriptions into standard YAML or JSON format files through a natural language processing engine, thereby reducing the user's need for programming professional knowledge and enabling more users to quickly get started with robotic arm task configuration.

[0049] Through artificial intelligence-assisted semantic analysis means and hierarchical relationship recognition algorithms, the present invention solves the problems of unclear hierarchical relationships and ambiguous logical associations between parent and child nodes in configuration information. By automatically identifying the hierarchical structure in the configuration information and refining the context environment of the status module, it ensures the correct relationship between the status module and its upper and lower nodes, thereby improving the accuracy and consistency of the configuration information and making subsequent task execution more reliable.

[0050] The present invention predicts the behavior patterns of status modules through a pre-trained machine learning model and constructs an internal data structure, solving the problems of lack of intelligent prediction ability and dynamic adjustment of task sequences in the task configuration process. By using historical task configuration data to train the machine learning model, predicting the behavior patterns of each status module, and converting them into a part of the internal data structure, it supports the dynamic adjustment of task sequences, enhances the flexibility and adaptability of the configuration, and makes the robotic arm task execution more efficient and controllable.

[0051] The present invention performs error detection and correction through deep learning technology and natural language processing technology, solving the common syntax errors and logical errors in configuration information. Using long short-term memory network deep learning technology to train the error detection model, it can identify and locate errors in the configuration file, and combine natural language processing technology to automatically generate correction suggestions and perform correction operations, thereby improving the correctness and reliability of the configuration file and reducing the risk of robotic arm task execution failure caused by configuration errors.

[0052] The above description is only an overview of the technical solution of this application. In order to be able to understand the technical means of this application more clearly, so that it can be implemented in accordance with the content of the specification, and in order to make the above and other purposes, features and advantages of this application more obvious and understandable, the following describes the preferred embodiments of this application in detail in conjunction with the accompanying drawings.

[0053] According to the following detailed description of the specific embodiments of this application in conjunction with the accompanying drawings, those skilled in the art will be more clear about the above and other purposes, advantages and features of this application. Brief Description of the Drawings

[0054] In order to more clearly illustrate the technical solutions in the embodiments of this application or in the prior art, the following will briefly introduce the accompanying drawings required for the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description are some embodiments of this application. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings. In all the drawings, similar elements or parts are generally identified by similar reference numerals. In the drawings, the elements or parts are not necessarily drawn to actual scale.

[0055] Figure 1 It is a flow chart for the parsing of the robotic arm;

[0056] Figure 2 It is a flow chart for the reading of configuration information;

[0057] Figure 3 It is a flow chart for the parsing of hierarchical relationships;

[0058] Figure 4 It is a flow chart for the prediction of behavior patterns;

[0059] Figure 5 It is a flow chart for the construction of data structures;

[0060] Figure 6 It is a flow chart for the detection of necessary nodes;

[0061] Figure 7 It is a flow chart for the detection of optional nodes;

[0062] Figure 8 It is a flow chart for error detection. Specific embodiments

[0063] To make the objectives, technical solutions, and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are some, but not all, of the embodiments of the present application. In the following description, specific details such as specific configurations and components are provided only to assist in a comprehensive understanding of the embodiments of the present application. Therefore, those skilled in the art should understand that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of the present application. Additionally, descriptions of known functions and structures are omitted for clarity and conciseness in the embodiments.

[0064] It should be understood that the "one embodiment" or "the present embodiment" mentioned throughout the specification means that specific features, structures, or characteristics related to the embodiment are included in at least one embodiment of the present application. Therefore, the "one embodiment" or "the present embodiment" that appears throughout the specification does not necessarily refer to the same embodiment. Additionally, these specific features, structures, or characteristics can be combined in one or more embodiments in any suitable manner.

[0065] Furthermore, the present application may repeat reference numerals and / or letters in different instances. This repetition is for the purpose of simplicity and clarity and does not itself indicate the relationship between the various embodiments and / or arrangements discussed.

[0066] In this text, the term "and / or" is merely a description of the relationship between associated objects, indicating that three relationships can exist. For example, A and / or B can represent three situations: A exists alone, B exists alone, and both A and B exist simultaneously. In this text, the term " / and" describes another relationship between associated objects, indicating that two relationships can exist. For example, A / and B can represent two situations: A exists alone, and both A and B exist. Additionally, in this text, the character " / " generally indicates that the associated objects before and after are in an "or" relationship.

[0067] In this text, the term "at least one" is merely a description of the relationship between associated objects, indicating that three relationships can exist. For example, at least one of A and B can represent three situations: A exists alone, both A and B exist simultaneously, and B exists alone.

[0068] It should also be noted that in this text, relational terms such as first and second are merely used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise", or any other variant thereof are intended to cover non-exclusive inclusion.

[0069] Embodiment 1

[0070] This embodiment mainly specifically describes a method for parsing low-code robotic arm task configurations;

[0071] Furthermore, as Figure 1 shown, a method for parsing low-code robotic arm task configurations includes the following steps:

[0072] S1: Read the low-code task configuration information written by the user in a graphical interface and natural language form. The configuration information is automatically converted into a standardized YAML or JSON format file by a parsing engine based on semantic understanding and a machine learning model. This file uses a tree structure to describe the robotic arm task, and each node represents a state module;

[0073] The graphical interface allows the user to intuitively design the robotic arm task flow by dragging and selecting. Through a natural language processing engine, it understands and converts the natural language description input by the user into a standard YAML or JSON format file, reducing the user's need for programming expertise.

[0074] By allowing the user to intuitively design the robotic arm task flow through a graphical interface and then converting the natural language description into a standard format file through a natural language processing engine, the user's need for programming expertise is reduced, enabling non-technical personnel to conveniently configure robotic arm tasks.

[0075] S2: Identify semantic units through the semantic analysis module, use the hierarchical relationship recognition algorithm to parse the hierarchical structure in the configuration information, identify the logical associations between nodes, and confirm the specific content of the status module;

[0076] S3: Use a machine learning model to predict the behavior pattern of the status module, construct an internal data structure in a recursive or iterative manner, map the hierarchical relationship in the configuration information to a software data structure that satisfies a tree topology structure, and at the same time support dynamic adjustment of the task sequence to enhance the flexibility of the configuration;

[0077] S4: During the parsing process, use an intelligent recommendation system to detect the necessity and optionality of nodes. For necessary nodes, ensure that they exist and the corresponding information is set; for optional nodes, the intelligent recommendation system automatically fills in reasonable default values according to the context environment to ensure the robustness of the parsing process;

[0078] S5: Use deep learning technology to detect syntax errors and logical errors in the configuration information, combine natural language processing technology to perform automatic error correction and provide correction suggestions to enhance the reliability of the parsing process;

[0079] S6: After the parsing is completed, pass the generated internal data structure into the execution logic, support real-time monitoring of the task execution status through a visualization tool, ensure that the robotic arm can execute tasks in a predetermined order and dependency relationship, and quickly respond and take corresponding handling measures through an intelligent decision-making system in case of abnormal situations.

[0080] By reading the low-code task configuration information written by the user in the form of a graphical interface and natural language, and using a parsing engine based on semantic understanding and machine learning models to automatically convert it into a standardized YAML or JSON format file, the user can easily define the tasks of the robotic arm without having to deeply understand programming details, greatly reducing the threshold of robotic arm programming.

[0081] Furthermore, as Figure 2 shown, S1 includes the following steps:

[0082] S1.1: Start the natural language processing module, perform a preliminary parsing on the natural language description input by the user, extract the keyword actions "move to", "grab", "release" and their parameters in the description, and use them to construct a preliminary task intention model;

[0083] S1.2: Use a semantic understanding model to perform semantic parsing on the extracted keywords, understand the meaning of the keywords in a specific context, and map them to the corresponding robotic arm actions; when the description of the keywords by the semantic understanding model is not precise enough, infer the correct meaning of the keywords through the context;

[0084] S1.3: Integrate the parsing results into a standardized YAML or JSON format file. All nodes in the file contain complete status module information, with a clear and explicit hierarchical relationship, facilitating further processing by the subsequent parsing engine.

[0085] By starting the natural language processing module to extract the key instructions input by the user, and using the semantic understanding model to map these instructions to the robotic arm actions, even if the description is not precise enough, the correct meaning can be inferred through the context, improving the flexibility and accuracy of task configuration.

[0086] Furthermore, as Figure 3 shown, S2 includes the following steps:

[0087] S2.1: Identify the semantic units in the configuration information through the semantic analysis module;

[0088] S2.2: Analyze the hierarchical relationship in the configuration information through the hierarchical relationship recognition algorithm, automatically identify the logical association between the parent node and the child node, and ensure that the relationship between the status module and the upper and lower nodes is correct.

[0089] S2.3: Use the context awareness algorithm to refine the context environment of the status module, making the description of the status module grammatically and logically correct and conforming to the actual application scenario.

[0090] S2.4: Confirm the specific content of the status module, including but not limited to TCP position, pose, speed, acceleration, path planning constraints, peripheral operations, and exception handling logic, ensuring logical consistency and meeting the requirements of the robotic arm to execute tasks.

[0091] S2.5: Generate a semantic analysis report, which contains the detailed information and hierarchical relationship of the status module, enabling accurate parsing in subsequent steps based on the semantic analysis report.

[0092] Analyze the hierarchical structure in the configuration information through the hierarchical relationship recognition algorithm, automatically identify the logical association between the parent node and the child node, ensure that the relationship between the status modules is correct, and thus establish a complete task execution flow chart, enhancing the logical consistency of task configuration.

[0093] Furthermore, as Figure 4 shown, the specific steps of the machine learning model in S3 are as follows:

[0094] S3.11: Collect and organize historical task configuration data, which includes successfully executed task configurations and actual operation results; form a dataset of the robotic arm task behavior patterns through the historical task configuration data, serving as the basis for training the machine learning model.

[0095] S3.12: Train a machine learning model through a supervised learning method; the input of the machine learning model is the description information of the status module, and the output is the predicted behavior pattern; during the training process, the model learns the mapping relationship between different configuration information and the actual behavior pattern, and predicts the expected behavior pattern according to the input description of the status module.

[0096] S3.13: During the parsing process, input the information of the status module into the trained machine learning model. The machine learning model predicts the behavior pattern of the status module according to experience and rules, and converts it into a part of the internal data structure.

[0097] Predict the behavior pattern of each status module through a pre-trained machine learning model, and construct an internal data structure according to the prediction results, making the generated data structure more efficient and easy to expand, supporting dynamic adjustment of the task sequence, and enhancing the flexibility of the configuration.

[0098] Further, as Figure 5 shown, the specific steps for constructing the internal data structure in S3 are as follows:

[0099] S3.21: Initialize the internal data structure and create a root node as the starting point of the entire task configuration; the root node serves as the top-level container for all subsequent status modules; the creation of the root node marks the start of the construction of the internal data structure and is also the basis for the recursive or iterative process.

[0100] S3.22: For the status module, after predicting its behavior pattern using the machine learning model, create corresponding nodes in the internal data structure according to the prediction results, and add the newly created nodes as child nodes or sibling nodes of the existing nodes to the internal data structure according to the hierarchical relationship defined in the configuration file, which is achieved through recursive calls; the recursive call is when a new status module is parsed and its behavior pattern is predicted, a recursive function is called; the recursive function adds new nodes and recursively processes the sub-status modules in the status module until all status modules are parsed and added to the internal data structure.

[0101] During the construction process of the internal data structure, track the hierarchical relationship to ensure the correct parent-child relationship of the nodes; at the same time, support dynamic adjustment of the task sequence, insert or delete status modules according to the actual situation during the construction process, and achieve this by iteratively checking the dependency relationships of the status modules and updating the internal data structure when necessary, thereby enhancing the flexibility and adaptability of the configuration.

[0102] By initializing the internal data structure, creating the root node, and creating corresponding nodes based on the prediction results and hierarchical relationships, ensuring the correct parent-child relationship of each node, supporting dynamic adjustment of the task sequence, thereby enhancing the flexibility and adaptability of the configuration.

[0103] Furthermore, as Figure 6 shown, the specific steps of S4 necessary nodes are as follows:

[0104] S4.11: Define the necessary node categories, including but not limited to the key attributes of the status module: TCP position, pose, speed, acceleration; the necessary nodes form the basis for an effective task configuration for the correct execution of tasks; the intelligent recommendation system checks whether the status module contains the necessary attributes to ensure reasonable attribute settings;

[0105] S4.12: Implement the necessary node detection logic. The system will verify whether the status module contains the necessary nodes one by one during the process of parsing the configuration file, and check whether the necessary nodes have been correctly set; if it is detected that a necessary node is missing or the necessary node setting is incomplete, the parsing will be stopped immediately, and the user will be prompted through the visual interface or message about the specific missing items, guiding the user to supplement or correct them.

[0106] By defining the necessary node categories, ensuring that each status module contains the necessary attributes, and checking the reasonableness of the settings of these attributes through the intelligent recommendation system, the effectiveness and accuracy of the task configuration are improved.

[0107] Furthermore, as Figure 7 shown, the specific steps of S4 optional nodes are as follows:

[0108] S4.21: Define the optional node categories, including but not limited to speed ratio, acceleration ratio, path planning constraints, peripheral operations, and exception handling logic; the optional nodes are optional existences in the status module, which can enhance the flexibility and robustness of tasks in specific situations;

[0109] S4.22: Implement the optional node detection logic; during the process of the intelligent recommendation system parsing the configuration file, it identifies which nodes belong to the optional type and checks whether the nodes are defined; if the checked nodes are not defined, the intelligent recommendation system will automatically fill in reasonable default values according to the context environment;

[0110] S4.23: Utilize big data analysis and machine learning technologies. The intelligent recommendation system can learn and recommend best practices based on the successful cases of past task configurations; when it is detected that a certain optional node is not defined, the system can fill in the default value and provide optimization suggestions according to the characteristics of the current task.

[0111] By defining optional node categories, the intelligent recommendation system can identify optional type nodes when parsing the configuration file, automatically fill in reasonable default values according to the context environment, reduce the time for users to manually debug, and enhance the robustness of the configuration process.

[0112] Furthermore, as Figure 8 shown, the steps of the S5 deep learning technology detection include the following:

[0113] S5.1: Build an error detection model; Use long short-term memory network deep learning technology to train the error detection model of the low-code task configuration file; The error detection model is trained based on a large number of labeled correct configuration files and error configuration files, and can identify common syntax errors, logical errors, and other potential problems; The training dataset of the error detection model includes instances of missing necessary fields, unreasonable parameter settings, and incorrect logical order, enabling the error detection model to have comprehensive error recognition capabilities;

[0114] S5.2: Perform error detection; During the process of parsing the configuration file, input the configuration file line by line and section by section into the trained long short-term memory network deep learning model. The long short-term memory network deep learning model will analyze the text of the configuration file to detect existing errors. If an error is detected, the error detection model will output the specific error type and its location;

[0115] S5.3: Correct errors through natural language processing technology; When an error is detected, use natural language processing technology to automatically generate correction suggestions and perform correction operations;

[0116] S5.4: Provide a detailed correction report; After completing error detection and correction, generate a correction report. The correction report includes the detected errors and their locations, detailed correction suggestions for the errors, and the executed correction operations.

[0117] By building an error detection model, using deep learning technology to detect syntax and logical errors in the configuration information, and combining natural language processing technology to provide detailed correction suggestions or directly perform correction operations, the robustness of the system and the reliability of the configuration process are improved.

[0118] Furthermore, the intelligent decision-making system of S6 includes an anomaly detection module;

[0119] The anomaly detection module uses machine learning algorithms to analyze the real-time data generated during the execution of the robotic arm task, and can identify and issue alarms for situations where the position deviation from the expected behavior is too large, the speed is abnormal, or there is external interference;

[0120] When the anomaly detection module detects an abnormal situation, the intelligent decision-making system automatically selects appropriate countermeasures according to the type and severity of the anomaly, including but not limited to pausing task execution, reverting to the previous stable state, attempting automatic repair, requesting manual intervention, and executing corresponding commands.

[0121] By integrating the anomaly detection module, analyzing real-time data using machine learning algorithms, identifying abnormal situations, and automatically taking appropriate countermeasures, the reliability and controllability of the robotic arm task execution are enhanced.

[0122] This embodiment describes a method for parsing low-code robotic arm task configurations. It simplifies the robotic arm task configuration through a graphical interface and natural language processing technology, predicts behavior patterns and constructs internal data structures through a machine learning model, making the configuration process both flexible and efficient. Combining an intelligent recommendation system and deep learning technology, the present invention not only ensures the accuracy and integrity of the configuration information but also provides a real-time monitoring and anomaly handling mechanism, greatly improving the reliability and execution efficiency of the robotic arm task configuration.

[0123] Embodiment 2

[0124] This embodiment is based on Embodiment 1 and mainly describes the specific implementation methods:

[0125] The task flow of the robotic arm is set by factory engineers through simple graphical interfaces and natural language descriptions to achieve the automated installation of car door panels.

[0126] Specific steps:

[0127] Task design and configuration:

[0128] Engineers drag and drop basic action modules such as "move to", "grab", and "place" through a graphical interface and describe task details in natural language. The specific description is as follows:

[0129] "The robotic arm moves to the door panel storage area, grabs the door panel, then moves to the body assembly area, and installs the door panel on the vehicle body."

[0130] The natural language processing engine identifies and extracts action instructions such as "move to", "grab", "place" and their parameters, and constructs a preliminary task intention model.

[0131] Semantic parsing and standardization:

[0132] The semantic understanding model converts the natural language description into specific action instructions for the robotic arm. The specific conversion is as follows:

[0133] "Move to the door panel storage area" is parsed as the target position (x = 50, y = 100, z = 20) of the TCP of the robotic arm, with the pose (roll = 0, pitch = 0, yaw = 0).

[0134] Generate a standardized YAML configuration file, where each node in the file contains complete status module information, TCP position, and pose.

[0135] Machine learning model prediction and internal data structure construction:

[0136] The machine learning model trained with historical task configuration data predicts the behavior patterns of each status module.

[0137] The specific predictions are as follows:

[0138] Based on historical data, the model predicts the optimal path and speed for "grasping the door panel".

[0139] According to the prediction results, corresponding nodes are created in the internal data structure to construct a data model that satisfies the tree topology structure, supporting dynamic adjustment of the task sequence.

[0140] Node detection and intelligent recommendation:

[0141] The necessary node detection logic checks whether each status module contains the key attributes of TCP position and pose. If missing, the user is prompted to supplement them.

[0142] The optional node detection logic automatically fills in default values, such as path planning constraint conditions, to ensure the flexibility and robustness of the task.

[0143] Error detection and correction:

[0144] Use deep learning technology to detect syntax and logical errors in the configuration file and provide correction suggestions. The specific detections are as follows:

[0145] If it is found that the "grasp" action between "move to" and "place" is missing, the user is prompted to supplement this step.

[0146] Combine natural language processing technology to automatically generate correction suggestions and perform correction operations.

[0147] Task execution and monitoring:

[0148] Pass the generated internal data structure into the execution logic, and the robotic arm executes the task according to the predetermined order and dependencies.

[0149] The anomaly detection module uses machine learning algorithms to analyze the real-time data generated during the execution of the robotic arm task. If a large position deviation or abnormal speed is found, an alarm is triggered.

[0150] The intelligent decision-making system automatically selects appropriate countermeasures according to the anomaly type and severity, and suspends task execution.

[0151] Furthermore, the following operation mode is adopted:

[0152]

[0153] Through the above implementation steps, this embodiment easily configures complex assembly tasks for the robotic arm without the need to deeply understand programming knowledge, thereby improving the automation level and work efficiency of the production line.

[0154] The above are only the preferred embodiments of the present invention, and it does not thereby limit the protection scope of the present invention. For those skilled in the art, the present invention can have various changes and modifications; all changes, modifications, substitutions, integrations, and parameter changes made to these embodiments within the spirit and principle of the present invention, through conventional substitutions or capable of achieving the same functions without departing from the principle and spirit of the present invention, fall within the protection scope of the present invention.

Claims

1. A low-code robotic arm task configuration parsing method, characterized in that: The following steps are involved: S1: Read the low-code task configuration information written by the user through a graphical interface and natural language. The configuration information is automatically converted into a standardized YAML or JSON format file through a parsing engine based on semantic understanding and machine learning models. The file uses a tree topology structure to describe the robot arm task, and each node represents a state module; S2: Identify semantic units through the semantic analysis module, parse the hierarchical structure in the configuration information using the hierarchical relationship recognition algorithm, identify the logical association between nodes, and confirm the specific content of the status module; S3: Use machine learning models to predict the behavior patterns of state modules. The machine learning model trained with historical task configuration data predicts the behavior patterns of each state module. Based on historical data, the model predicts the optimal path and speed for "grabbing the door panel" and constructs an internal data structure in a recursive or iterative manner to map the hierarchical relationship in the configuration information to a software data structure that satisfies the tree topology structure. It also supports dynamic adjustment of task sequences to enhance configuration flexibility. S4: During the parsing process, an intelligent recommendation system is used to detect the necessity and optionality of nodes. For necessary nodes, the corresponding information is ensured to exist and be set. For optional nodes, the intelligent recommendation system automatically fills in reasonable default values ​​according to the context to ensure the robustness of the parsing process. S5: Use deep learning technology to detect syntax errors and logical errors in configuration information, and combine natural language processing technology to automatically correct errors and provide correction suggestions to enhance the reliability of the parsing process; S6: After the analysis is completed, the generated internal data structure is passed to the execution logic, which supports real-time monitoring of the task execution status through visualization tools to ensure that the robot can execute tasks according to the predetermined order and dependencies. When encountering abnormal situations, the intelligent decision-making system can quickly respond and take corresponding processing measures; The reading of the low-code task configuration information written by the user through a graphical interface and natural language specifically includes the following steps: S1.1: Start the natural language processing module to perform preliminary analysis on the natural language description input by the user, extract the key words "move to", "grab", "release" action instructions and their parameters in the description, and use them to build a preliminary task intention model; S1.2: Using the semantic understanding model, the extracted key words are semantically analyzed to understand the meaning of the words in a specific context and map them to the corresponding robot arm movements; if the semantic understanding model does not accurately describe the key words, the correct meaning of the key words is inferred from the context; S1.3: Integrate the parsing results into a standardized YAML or JSON format file. The nodes in the file contain complete state module information, and the hierarchical relationship is clear and explicit, which is convenient for subsequent parsing engines to perform further processing; The machine learning model in S3 specifically includes the following steps: S3.11: Collect and organize historical task configuration data, wherein the historical task configuration data includes successfully executed task configurations and actual operation results; form a data set of robot arm task behavior patterns through the historical task configuration data as a basis for training a machine learning model; S3.12: training a machine learning model by a supervised learning method; the machine learning model input is the description information of the state module, and the output is the predicted behavior pattern; during the training process, the model learns the mapping relationship between different configuration information and the actual behavior pattern, and predicts the predicted behavior pattern based on the input description information of the state module; S3.13: During the parsing process, the description information of the state module is input into the trained machine learning model. The machine learning model predicts the behavior pattern of the state module based on experience and rules and converts it into part of the internal data structure.

2. A low-code robotic arm task configuration parsing method according to claim 1, characterized in that: In S1, the graphical interface allows the user to intuitively design the robot arm task process by dragging and selecting, and the natural language processing engine understands and converts the natural language description input by the user, converting it into a standard YAML or JSON format file, reducing the user's need for programming expertise.

3. A low-code robotic arm task configuration parsing method according to claim 1, characterized in that: The S2 specifically includes the following steps: S2.1: Identify semantic units in the configuration information through the semantic analysis module; S2.2: Analyze the hierarchical relationship in the configuration information through the hierarchical relationship recognition algorithm, automatically identify the logical relationship between the parent node and the child node, and ensure that the relationship between the state module and the upper and lower nodes is correct; S2.3: Use context-aware algorithms to refine the context of the state module so that the description of the state module is grammatically and logically correct and conforms to the actual application scenario; S2.4: Confirm the specific contents of the status module, including TCP position, posture, speed, acceleration, path planning constraints, peripheral operations, and exception handling logic, to ensure that they are logically consistent and meet the requirements for the robot to perform tasks; S2.5: Generate a semantic analysis report, which contains description information and hierarchical relationships of the state modules, so that subsequent steps can be accurately parsed based on the semantic analysis report.

4. A low-code robotic arm task configuration parsing method according to claim 1, characterized in that: The S3 internal data structure construction specifically includes the following steps: S3.21: Initialize the internal data structure and create a root node as the starting point of the entire task configuration; the root node serves as the top-level container for all subsequent state modules; the creation of the root node marks the beginning of the construction of the internal data structure and is also the basis of the recursive or iterative process; S3.22: For the state module, after using the machine learning model to predict its behavior pattern, create a corresponding node in the internal data structure according to the prediction result, and add the newly created node as a child node or sibling node of the existing node to the internal data structure according to the hierarchical relationship defined in the configuration information, and implement it through recursive calls; the recursive call is to call a recursive function when a new state module is parsed and its behavior pattern is predicted; the recursive function is used to add new nodes and recursively process the sub-state modules in the state module until all state modules are parsed and added to the internal data structure; S3.23: During the construction of the internal data structure, the hierarchical relationship is tracked to ensure that the parent-child relationship of the nodes is correct; at the same time, it supports dynamic adjustment of the task sequence, inserting or deleting state modules according to actual conditions during the construction process, and iteratively checking the dependencies of the state modules.

5. A low-code robotic arm task configuration parsing method according to claim 1, characterized in that: The S4 comprises the following steps: S4.11: Define the necessary node categories, including the key attributes of the state module: TCP position, posture, velocity, acceleration; the necessary nodes form the basis for effective task configuration and correct execution of tasks; the intelligent recommendation system checks whether the state module contains necessary attributes to ensure that the attribute settings are reasonable; S4.12: Implement the necessary node detection logic. During the process of parsing the configuration information, the system will verify whether the status module contains the necessary nodes one by one and check whether the necessary nodes have been correctly set. If it is detected that the necessary nodes are missing or the necessary nodes are incompletely set, the parsing will be stopped immediately, and the user will be prompted with the specific missing items through a visual interface or message to guide the user to supplement or correct them.

6. A low-code robotic arm task configuration parsing method according to claim 1, characterized in that: The S4 specifically includes the following steps: S4.21: define optional node categories, including speed ratio, acceleration ratio, path planning constraints, peripheral operations and exception handling logic; the optional nodes are optional existence of the state module; S4.22: Implement optional node detection logic; the intelligent recommendation system identifies which nodes are optional during the process of parsing the configuration information and checks whether the nodes are defined; if the nodes are not defined, the intelligent recommendation system will automatically fill in reasonable default values ​​based on the context; S4.23: Using big data analysis and machine learning techniques, the intelligent recommendation system can learn and recommend best practices based on successful cases of past task configurations. When it detects that an optional node is not defined, the system can fill in the default value and provide optimization suggestions based on the characteristics of the current task.

7. A low-code robotic arm task configuration parsing method according to claim 1, characterized in that: In S5, deep learning technology is used to detect syntax errors and logic errors in the configuration information, which specifically includes the following steps: S5.1: Build an error detection model; use long short-term memory network deep learning technology to train the error detection model of low-code task configuration information; the error detection model is trained based on a large amount of annotated correct configuration information and incorrect configuration information, and can identify common syntax errors and logical errors; the training data set of the error detection model includes instances of missing necessary fields, unreasonable parameter settings, and logical sequence errors, so that the error detection model has comprehensive error recognition capabilities; S5.2: Execute error detection. During the process of parsing the configuration information, the configuration information is input into the trained long short-term memory network deep learning model line by line and segment by segment. The long short-term memory network deep learning model will analyze the text of the configuration information and detect the existing errors. If an error is detected, the error detection model will output the specific error type and its location. S5.3: Perform error correction using natural language processing technology; when an error is detected, use natural language processing technology to automatically generate correction suggestions and perform correction actions; S5.4: Provide a detailed correction report; after completing error detection and correction, generate a correction report that includes the detected errors and their locations, detailed correction suggestions for the errors, and the corrective actions that have been performed.

8. A low-code robotic arm task configuration parsing method according to claim 1, characterized in that: The intelligent decision-making system of S6 includes an abnormality detection module; The anomaly detection module uses machine learning algorithms to analyze real-time data generated during the execution of the robot arm's tasks, and can identify excessive position deviations from expected behavior, abnormal speeds, and external interference, and issue an alarm; When the anomaly detection module detects an abnormal situation, the intelligent decision-making system automatically selects appropriate response measures based on the type and severity of the anomaly, including suspending task execution, rolling back to the last stable state, attempting automatic repair, requesting manual intervention, and executing corresponding commands.

Images