Robot off-line programming method and system based on technology knowledge base generation technology

Abstract

The invention relates to a robot offline programming method and system based on a process knowledge base generation process, and the method comprises the following steps: S1, scene modeling: importing a CAD model of a robot operation scene, building a 3D virtual environment, and calibrating the 3D virtual environment; s2, track design and process parameter analysis: selecting a processing task target area, extracting task characteristic parameters, interacting with a process knowledge base in real time, reasoning initial process parameters according to the task characteristic parameters, generating a process procedure, analyzing the process procedure, determining associated parameters of a motion track and generating the motion track; s3, simulation verification: based on the three-dimensional model motion, simulating the machining process through a control technology to verify the correctness of the robot track; and S4, generating an executable program. According to the method, the process knowledge graph is introduced into industrial robot off-line programming, and the problems of efficiency, precision and maintenance of traditional industrial robot off-line programming are systematically solved through intelligent driving and full-process automatic design of the process knowledge graph.

Description

Technical Field

[0001] This invention belongs to the field of industrial robot control technology, specifically relating to a robot offline programming method and system based on a process knowledge base for generating processes. Background Technology

[0002] The path generation technology for industrial robots has gone through three important stages of development. Each generation of technology has made improvements on the basis of the previous generation, but there are still corresponding limitations.

[0003] The first-generation technology employed an embedded programming approach, characterized by its heavy reliance on dedicated hardware platforms. This approach required close collaboration among teams from multiple disciplines, including control theory, software engineering, and process technology, resulting in lengthy development cycles and high costs. Particularly in welding applications, engineers not only needed to independently develop weld trajectory planning algorithms but also implement underlying PID control systems, placing extremely high demands on the R&D team's algorithmic capabilities and programming skills.

[0004] The second-generation technology introduces a teach-programming mode, using a teach pendant to record paths. While this lowers the programming barrier, it has significant shortcomings when dealing with complex 3D trajectories: the teaching process is time-consuming, and the final accuracy is directly limited by the operator's skill level. Furthermore, this technology lacks path visualization support, making path pre-simulation and optimization difficult.

[0005] The current mainstream third-generation offline programming technology achieves path planning through dedicated software. While it addresses some shortcomings of the previous two generations, two major challenges remain: First, engineers need to master professional programming skills and domain knowledge, requiring novices to invest significant learning time; second, in practical applications, whether welding, painting, or polishing, the system requires users to set numerous parameters and mark many path points. More importantly, this path generation process is often "all or nothing," meaning any later modifications require complete replanning, significantly impacting work efficiency.

[0006] In summary, there is an urgent need to provide a robot offline programming method and system based on process knowledge base that is highly efficient, has high quality stability, and is easy to maintain. Summary of the Invention

[0007] The purpose of this invention is to provide a robot offline programming method and system based on a process knowledge base that generates processes, which has high programming efficiency, high quality stability, and is easy to maintain.

[0008] The above objective is achieved through the following technical solution: an offline robot programming method based on a process knowledge base, comprising the following steps:

[0009] S1, Scene Modeling: Import the CAD model of the robot's work scene, build a 3D virtual environment and calibrate it;

[0010] S2, Trajectory Design and Process Parameter Analysis: Select the target area of ​​the processing task, extract the task feature parameters (bevel type, base material, wall thickness, etc.), interact with the process knowledge base in real time, deduce the initial process parameters and generate the process procedure based on the task feature parameters, parse the process procedure, determine the associated parameters of the motion trajectory and generate the motion trajectory.

[0011] S3, Simulation Verification: Based on the motion of the 3D model, the correctness of the robot trajectory is verified by simulating the processing process through control technology;

[0012] S4, Executable program generation: Generates executable code for the robot based on the motion trajectory and its associated parameters.

[0013] This invention relates to robot offline programming, creatively combining offline programming with a process knowledge base (process knowledge graph). Based on the knowledge graph, it generates processes for robot offline programming. After scene modeling, the processing task range (e.g., weld seam, grinding and polishing) is selected. The system requests inference from the knowledge graph server via HTTP, inputting scene parameters (groove type, base material, wall thickness, etc.) to obtain recommended process parameters (e.g., welding current, voltage, speed), generating a process specification. Based on the process specification, trajectory-related parameters are parsed, and the motion trajectory (e.g., welding arc initiation / outitiation position, welding path) is calculated in real time. Then, based on the process specification, trajectory-related parameters are parsed again, and the motion trajectory (e.g., welding arc initiation / outitiation position, welding path) is calculated in real time. The processing process is simulated using 3D model control technology, collision risks are detected, and the trajectory is adjusted through virtual-real interaction (e.g., manual path point calibration). A post-template is matched based on the robot brand (e.g., Fanuc, KUKA), and executable code (e.g., MOVL instructions, ARCSET parameters) is generated by parsing the template rules.

[0014] The utilization rate of process knowledge in the technical solution of this invention is increased from less than 35% in traditional methods to more than 85%, while improving programming efficiency, reducing the programming time of complex trajectories by 87%, and increasing the accuracy of process parameter selection to 95% ± 2%. In addition, it can support the automatic generation of code for multiple brands of robots, reduce the workload of process knowledge updates by 70%, and greatly reduce maintenance costs.

[0015] A further technical solution is that the process knowledge base is a process knowledge graph, and before step S1, there is also a step of constructing the process knowledge graph: extracting descriptive features of entities, attributes and relationships from scattered unstructured text data of the manufacturing process, and constructing a process knowledge graph containing static knowledge and quasi-static knowledge according to predetermined annotation rules. The static knowledge includes theoretical knowledge and empirical knowledge, and the quasi-static knowledge includes management layer, process layer and process step layer knowledge.

[0016] This invention organizes process knowledge into a graph model, providing intelligent process decision-making for offline programming. In the process of sharing the process knowledge graph, a distributed architecture process knowledge modeling system is adopted, defining 13 types of general entities and one type of potential relation entity. The corpus annotation uses BIOES rules to extract entities (13 types of general entities + 1 type of potential relation entity), attributes, and relationships from scattered unstructured text (such as process parameter tables and PDFs), and annotates the corpus using BIOES rules. Static knowledge (theoretical / empirical knowledge) and quasi-static knowledge (management level, process level, step level) are defined, constructing a graph model containing nodes and edges to provide parameter recommendations for offline programming. In this way, a unified standard for multi-process knowledge systems is established, solving the problems of traditional data being "scattered, redundant, and unstructured," improving the accuracy of knowledge extraction, and providing a standardized data foundation for subsequent reasoning.

[0017] A further technical solution is that, after step S4, a quality assessment and knowledge update step is also included: based on the process characteristics of the processing, data is collected during the robot's operation, the operation quality is analyzed and evaluated, the process execution effect is analyzed, optimization schemes are reasoned, and feedback is given to the process knowledge graph to correct and update its rules.

[0018] This research utilizes machine vision and sensors to collect data during robot operations, including robot motion data, process parameters, and operational quality data (such as weld morphology). In specific applications, based on the process characteristics of welding, spraying, grinding, polishing, hole making, and laser processing, a robot process testing device is developed, and a robot operation testing system is built. Combining sensory data on robot motion, process, and operating devices, process and quality data are extracted to study robot operation quality evaluation methods. A dynamic process evaluation method based on "process mechanism - machine learning - expert knowledge base" is adopted to establish a robot operation quality assessment model. Principal component and factor analysis methods are used to perform correlation analysis on the virtual and real model data of robot operation quality. Combined with robot process standards, the process performance evaluation results and process parameter errors are established. Based on the evaluation results, the knowledge graph rules are optimized to achieve a "detection-diagnosis-optimization" closed loop.

[0019] A further technical solution is that the simulation verification in step S3 includes:

[0020] Collision detection: The interference volume is calculated using a relevant algorithm;

[0021] Virtual-real interaction: Real-time adjustment and verification of robot movement trajectory in a virtual environment;

[0022] Track correction: If a joint overshoot or collision is detected, the trajectory will be automatically shifted and the map safety rules will be updated.

[0023] This setup allows for early detection of trajectory defects, reduces hardware wear and tear during actual debugging, lowers the cost of manual programming and trial and error, and improves efficiency.

[0024] A further technical solution is that, in step S2, the process knowledge base is a process knowledge graph or a large process model, and the interaction process between the process knowledge base and the process includes: matching process parameter cases through semantic similarity algorithm, inferring process parameters and generating process procedures, parsing process procedures and dynamically generating motion trajectories and associated parameters.

[0025] A further technical solution is that, in step S4, a post-template file is matched according to the robot type, and an executable program for the robot is generated based on the CAD model and the determined motion trajectory and its associated parameters by parsing the rules of the post-template file.

[0026] To achieve the above objectives, the present invention also provides a robot offline programming system based on a process knowledge base for generating processes, comprising:

[0027] Process knowledge modeling system: used to define a multi-process manufacturing knowledge system, analyze the descriptive features of entities, attributes and relationships in Chinese processing technology texts, formulate annotation rules, and extract and construct process knowledge graphs;

[0028] Knowledge Graph Cloud Service Cluster: Based on the process rules generated by the process knowledge modeling system, the domain knowledge graph is stored in the database, providing a knowledge reasoning interface to the offline programming system and a process feedback interface to the quality assessment system.

[0029] Offline programming system: Dynamically generates post-programs based on process parameters recommended by CAD models and process knowledge graphs;

[0030] Quality assessment system: Based on the process characteristics of the machining, it collects data during the robot's operation, analyzes and evaluates the quality of the operation, analyzes the effect of process execution, infers optimization schemes, and provides feedback to correct the process knowledge graph rules.

[0031] This invention features an integrated architecture and fully digital processes, solving the problems of traditional "dispersed knowledge bases and lack of dynamic coupling" while improving system collaboration efficiency.

[0032] A further technical solution is that the offline programming system includes:

[0033] Scene building module: Used to construct a 3D virtual scene that matches the real-world processing environment;

[0034] Process planning module: used for intelligent process decision-making based on process knowledge graph, realizing processing path selection, process adaptation and path planning generation;

[0035] Simulation module: used to simulate the machining process, collision detection, and virtual-real interaction, and to virtually verify the generated robot trajectory;

[0036] Post-program module: Used to convert motion trajectories and their associated parameters into control code that can be executed by the robot.

[0037] This setup supports robots from multiple brands (conversion accuracy ≥99.8%, advantage 4), improving equipment compatibility; at the same time, the modular design reduces development complexity and facilitates functional expansion (such as adding laser processing technology).

[0038] A further technical solution is that the knowledge graph cloud service cluster adopts a distributed architecture, including:

[0039] Process Knowledge Graph Management System: Used for the construction, querying, and node management of process knowledge graphs;

[0040] Knowledge graph server: Stores process rules and reasoning logic, including a process reasoning module, and communicates with offline programming software through a knowledge reasoning interface to achieve dynamic recommendation of process parameters;

[0041] Neo4j graph database cluster: knowledge of structured storage technology.

[0042] A further technical solution is that the offline programming system and the process knowledge graph management system achieve data interaction and communication through the process knowledge graph API server, process inference server, and quality assessment server.

[0043] This invention introduces process knowledge graphs into offline programming of industrial robots. Through the intelligent driving force and full-process automated design of process knowledge graphs, it systematically solves the problems of efficiency, accuracy and maintenance in traditional offline programming of industrial robots, and provides a feasible solution for the flexible and intelligent upgrading of intelligent manufacturing. Attached Figure Description

[0044] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an improper limitation of the invention.

[0045] Figure 1 This is a flowchart illustrating a robot offline programming method based on a process knowledge base for generating processes, according to one embodiment of the present invention.

[0046] Figure 2 This is a structural block diagram of a robot offline programming system based on a process knowledge base that generates processes, according to one embodiment of the present invention. Detailed Implementation

[0047] The present invention will now be described in detail with reference to the accompanying drawings. This description is merely illustrative and explanatory, and should not be construed as limiting the scope of protection of the present invention. Furthermore, those skilled in the art can combine the features in the embodiments described herein and in different embodiments accordingly based on the description in this document.

[0048] The embodiments of the present invention are as follows, with reference to Figure 1 A robot offline programming method based on process knowledge base for generating processes includes the following steps:

[0049] S1, Scene Modeling: Import the CAD model of the robot's work scene, build a 3D virtual environment and calibrate it;

[0050] S2, Trajectory Design and Process Parameter Analysis: Select the target area of ​​the processing task, extract the task feature parameters (bevel type, base material, wall thickness, etc.), interact with the process knowledge base in real time, deduce the initial process parameters and generate the process procedure based on the task feature parameters, parse the process procedure, determine the associated parameters of the motion trajectory and generate the motion trajectory.

[0051] S3, Simulation Verification: Based on the motion of the 3D model, the correctness of the robot trajectory is verified by simulating the processing process through control technology;

[0052] S4, Executable program generation: Generates executable code for the robot based on the motion trajectory and its associated parameters.

[0053] This invention relates to robot offline programming, creatively combining offline programming with a process knowledge base (process knowledge graph). Based on the knowledge graph, it generates processes for robot offline programming. After scene modeling, the processing task range (e.g., weld seam, grinding and polishing) is selected. The system requests inference from the knowledge graph server via HTTP, inputting scene parameters (groove type, base material, wall thickness, etc.) to obtain recommended process parameters (e.g., welding current, voltage, speed), generating a process specification. Based on the process specification, trajectory-related parameters are parsed, and the motion trajectory (e.g., welding arc initiation / outitiation position, welding path) is calculated in real time. Then, based on the process specification, trajectory-related parameters are parsed again, and the motion trajectory (e.g., welding arc initiation / outitiation position, welding path) is calculated in real time. The processing process is simulated using 3D model control technology, collision risks are detected, and the trajectory is adjusted through virtual-real interaction (e.g., manual path point calibration). A post-template is matched based on the robot brand (e.g., Fanuc, KUKA), and executable code (e.g., MOVL instructions, ARCSET parameters) is generated by parsing the template rules.

[0054] The utilization rate of process knowledge in the technical solution of this invention is increased from less than 35% in traditional methods to more than 85%, while improving programming efficiency, reducing the programming time of complex trajectories by 87%, and increasing the accuracy of process parameter selection to 95% ± 2%. In addition, it can support the automatic generation of code for multiple brands of robots, reduce the workload of process knowledge updates by 70%, and greatly reduce maintenance costs.

[0055] Based on the above embodiments, in another embodiment of the present invention, such as Figure 1 The process knowledge base is a process knowledge graph. Before step S1, there is also a step of constructing the process knowledge graph: extracting descriptive features of entities, attributes and relationships from scattered unstructured text data of the manufacturing process, and constructing a process knowledge graph containing static knowledge and quasi-static knowledge according to predetermined annotation rules. The static knowledge includes theoretical knowledge and empirical knowledge, and the quasi-static knowledge includes management layer, process layer and process step layer knowledge.

[0056] This invention organizes process knowledge into a graph model, providing intelligent process decision-making for offline programming. In the process of sharing the process knowledge graph, a distributed architecture process knowledge modeling system is adopted, defining 13 types of general entities and one type of potential relation entity. The corpus annotation uses BIOES rules to extract entities (13 types of general entities + 1 type of potential relation entity), attributes, and relationships from scattered unstructured text (such as process parameter tables and PDFs), and annotates the corpus using BIOES rules. Static knowledge (theoretical / empirical knowledge) and quasi-static knowledge (management level, process level, step level) are defined, constructing a graph model containing nodes and edges to provide parameter recommendations for offline programming. In this way, a unified standard for multi-process knowledge systems is established, solving the problems of traditional data being "scattered, redundant, and unstructured," improving the accuracy of knowledge extraction, and providing a standardized data foundation for subsequent reasoning.

[0057] Static knowledge includes theoretical and empirical knowledge; it is axiomatic and remains unchanged regardless of the production task. It represents processing principles and technical details. Quasi-static knowledge refers to knowledge with a fixed organizational form, but whose values ​​depend on the specific task, mainly involving production processes. A quasi-static knowledge system comprises three levels: the management level includes knowledge of production resource management, schedule management, demand management, and process management; the process level includes understanding of processes, equipment, working hours, personnel, and steps; and the step level contains the steps within each process, including knowledge of tools, processing parameters, working hours, etc., which are closely related to product quality. Define a general entity set A and a potential relation entity set B. A general entity refers to a knowledge triple.<S,P,O> The text defines a subject entity S and an object entity O. A potential relational entity is an entity that may represent a predicate P. Thirteen types of general entities and one type of potential relational entity are defined.

[0058] Based on the above embodiments, in another embodiment of the present invention, such as Figure 1 The step S4 is followed by a quality assessment and knowledge update step: based on the process characteristics of the processing, data is collected during the robot's operation, the operation quality is analyzed and evaluated, the process execution effect is analyzed, optimization schemes are deduced, and feedback is given to the process knowledge graph to correct and update its rules.

[0059] This research utilizes machine vision and sensors to collect data during robot operations, including robot motion data, process parameters, and operational quality data (such as weld morphology). In specific applications, based on the process characteristics of welding, spraying, grinding, polishing, hole making, and laser processing, a robot process testing device is developed, and a robot operation testing system is built. Combining sensory data on robot motion, process, and operating devices, process and quality data are extracted to study robot operation quality evaluation methods. A dynamic process evaluation method based on "process mechanism - machine learning - expert knowledge base" is adopted to establish a robot operation quality assessment model. Principal component and factor analysis methods are used to perform correlation analysis on the virtual and real model data of robot operation quality. Combined with robot process standards, the process performance evaluation results and process parameter errors are established. Based on the evaluation results, the knowledge graph rules are optimized to achieve a "detection-diagnosis-optimization" closed loop.

[0060] Based on the above embodiments, in another embodiment of the present invention, the simulation verification in step S3 includes:

[0061] Collision detection: The interference volume is calculated using a relevant algorithm;

[0062] Virtual-real interaction: Real-time adjustment and verification of robot movement trajectory in a virtual environment;

[0063] Track correction: If a joint overshoot or collision is detected, the trajectory will be automatically shifted and the map safety rules will be updated.

[0064] This setup allows for early detection of trajectory defects, reduces hardware wear and tear during actual debugging, lowers the cost of manual programming and trial and error, and improves efficiency.

[0065] Based on the above embodiments, in another embodiment of the present invention, in step S2, the process knowledge base is a process knowledge graph or a large process model, and the process knowledge base interacts with the process including: matching process parameter cases through semantic similarity algorithm, inferring process parameters and generating process procedures, parsing process procedures and dynamically generating motion trajectories and associated parameters.

[0066] Based on the above embodiments, in another embodiment of the present invention, in step S4, a post-template file is matched according to the robot type, and a robot executable program is generated based on the CAD model and the determined motion trajectory and its associated parameters by parsing the rules of the post-template file.

[0067] To achieve the above objectives, the present invention also provides a robot offline programming system based on a process knowledge base for generating processes, such as... Figure 2 ,include:

[0068] 1. Process Knowledge Modeling System: Used to define a multi-process manufacturing knowledge system, analyze the descriptive features of entities, attributes and relationships in Chinese processing technology texts, formulate annotation rules, and extract and construct process knowledge graphs;

[0069] 2. Knowledge Graph Cloud Service Cluster: Based on the process rules generated by the process knowledge modeling system, the domain knowledge graph is stored in the database via the knowledge graph generation module server of the knowledge graph cloud service cluster, using online / offline methods. It provides a knowledge reasoning interface to offline programming software for recommending process parameters; and a process feedback interface to the quality assessment model for updating the knowledge graph content.

[0070] The knowledge graph cloud service cluster adopts a distributed architecture, including:

[0071] Process Knowledge Graph Management System: Used for the construction, querying, and node management of process knowledge graphs;

[0072] Knowledge graph server: Stores process rules and reasoning logic, including a process reasoning module, and communicates with offline programming software through a knowledge reasoning interface to achieve dynamic recommendation of process parameters;

[0073] Neo4j graph database cluster: knowledge of structured storage technology.

[0074] In one specific embodiment, the offline programming system sends scenario parameters (such as beveling shape, base material, wall thickness, processing type, etc.) to the cloud service cluster via the HTTP protocol. Based on the input parameters, the knowledge graph server traverses nodes and relationships in the Neo4j graph database, uses graph algorithms to generate matching combinations of process parameters, and returns recommended process parameters to the offline programming software through the knowledge reasoning interface to form an initial process specification.

[0075] The manufacturing process involves diverse knowledge types and a large number of nodes and relationships. In actual production, it is often necessary to simultaneously conduct queries and matching using multiple knowledge bases. A process knowledge graph management system is used for this purpose. Furthermore, small-batch customized production requires the knowledge graph to be real-time and flexible, enabling convenient addition, deletion, modification, and query operations. Therefore, a process knowledge data management software suitable for small-batch customized production was developed. The software adopts traditional software development lifecycle methods, using a top-down, stepwise refinement structured software design approach.

[0076] The process knowledge graph management system mainly has the following functions:

[0077] (1) Construction of process knowledge graph;

[0078] (2) Process knowledge inquiry;

[0079] (3) Adding, deleting and modifying process knowledge nodes;

[0080] (4) Adding, deleting and modifying the relationships between process knowledge nodes;

[0081] (5) Storage of process knowledge.

[0082] The requirements for software development are a regular PC and the corresponding system and software. The system used in this development is Windows 10, the development environment is Python 3.7+, and the dependent libraries include PyQt 5.15.7, Pandas 1.5.3, NetworkX 3.0, Matplotlib 3.7.1 and PyQtGraph 0.13.1.

[0083] 3. Offline programming system: Dynamically generates post-programs based on process parameters recommended by CAD models and process knowledge graphs; the offline programming system includes:

[0084] Scene building module: Used to construct a 3D virtual scene that matches the real-world processing environment;

[0085] Process planning module: used for intelligent process decision-making based on process knowledge graph, realizing processing path selection, process adaptation and path planning generation;

[0086] Simulation module: used to simulate the machining process, collision detection, and virtual-real interaction, and to virtually verify the generated robot trajectory;

[0087] Post-program module: Used to convert motion trajectories and their associated parameters into control code that can be executed by the robot.

[0088] This setup supports robots from multiple brands (conversion accuracy ≥99.8%, advantage 4), improving equipment compatibility; at the same time, the modular design reduces development complexity and facilitates functional expansion (such as adding laser processing technology).

[0089] The core functions of the scene building module include:

[0090] CAD Model Import: Supports importing CAD models in various formats (such as STEP and STL) to create virtual entities such as workpieces, robots, and tooling fixtures;

[0091] 3D calibration: The coordinates of each entity in the virtual scene are calibrated (such as the workpiece coordinate system and the robot base coordinate system) to ensure the consistency of the position between the virtual space and the actual physical space;

[0092] Kinematic parameter settings: Configure the kinematic parameters of the robot and related mechanisms (such as joint range of motion and end effector size), and define the motion logic and constraints of the device;

[0093] Scene environment configuration: Add auxiliary elements required for processing (such as welding power source, spraying equipment) to build a complete virtual processing workstation.

[0094] The core functions of the process planning module include:

[0095] Processing task acquisition: Identify processing features (such as welds, polishing surfaces, and hole-making areas) based on the CAD model and delineate the robot's operating range;

[0096] Process parameter reasoning: Communicates with the knowledge graph server via HTTP protocol, inputs scenario parameters (such as bevel type, base material, wall thickness, and processing type), and calls the process reasoning module to generate initial process parameters (such as welding current, voltage, wire feed speed; spraying pressure, and nozzle distance).

[0097] Process specification generation: Integrate the process parameters obtained through reasoning into structured process specifications (such as welding process cards, grinding and polishing process cards), clarifying details such as processing steps, tool selection, and parameter settings.

[0098] Path planning generation: Based on the process specifications and processing task scope, automatically calculate the robot's motion trajectory (such as straight line or curved path), determine the coordinates of trajectory points, motion sequence, and timing logic.

[0099] Process adaptation: Match the corresponding process rules according to the processing type (welding, spraying, grinding and polishing, etc.) and adjust the trajectory parameters (such as the welding amplitude and the grinding head speed).

[0100] The core functions of the simulation module include:

[0101] Process simulation: Dynamically simulate the robot's motion trajectory in a 3D virtual scene and visualize the relative motion process between the tool (such as welding gun, grinding head) and the workpiece;

[0102] Collision detection: Real-time monitoring of collision risks between robots, tools and workpieces, and tooling fixtures; automatic marking of interference points and generation of collision reports;

[0103] Virtual-real interaction function: It allows users to manually adjust trajectory points (such as modifying position coordinates, movement speed, and attitude angle) through the human-machine interface, and verify in real time whether the adjusted trajectory meets the process requirements;

[0104] Performance verification: Simulate the robot's motion under load conditions, joint torque and other parameters to evaluate the kinematic feasibility of the trajectory (e.g., whether it exceeds the joint range of motion);

[0105] The core functions of the post-processor module include

[0106] Robot type adaptation: Built-in multi-brand robot backend template library, supports automatic matching of corresponding code format and instruction set according to robot model;

[0107] Post-template parsing: Parses the rules of the post-template file, converting trajectory point coordinates, process parameters, and motion instructions into robot-specific code syntax;

[0108] Program generation and export: Generate executable programs (such as .SRC files and .LST files) that can be directly transferred to the robot controller, and support code editor functions (such as syntax highlighting and error prompting).

[0109] Program verification: Perform syntax verification on the generated code to ensure that the instruction format is correct and the parameter range is compliant.

[0110] 4. Quality Assessment System: Based on the process characteristics of the machining, the system collects data during the robot's operation, analyzes and evaluates the quality of the operation, analyzes the effect of process execution, infers optimization schemes, and provides feedback to correct the process knowledge graph rules.

[0111] This invention features an integrated architecture and fully digital processes, solving the problems of traditional "dispersed knowledge bases and lack of dynamic coupling" while improving system collaboration efficiency.

[0112] Based on the above embodiments, in another embodiment of the present invention, the offline programming system and the process knowledge graph management system realize data interaction and communication through the process knowledge graph API server, the process inference server, and the quality assessment server.

[0113] This invention introduces process knowledge graphs into offline programming of industrial robots. Through the intelligent driving force and full-process automated design of process knowledge graphs, it systematically solves the problems of efficiency, accuracy and maintenance in traditional offline programming of industrial robots, and provides a feasible solution for the flexible and intelligent upgrading of intelligent manufacturing.

[0114] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A robot offline programming method based on process knowledge base for generating processes, characterized in that, Includes the following steps: S1, Scene Modeling: Import the CAD model of the robot's work scene, build a 3D virtual environment and calibrate it; S2, Trajectory Design and Process Parameter Analysis: Select the target area of ​​the processing task, extract the task feature parameters, interact with the process knowledge base in real time, infer the initial process parameters and generate the process procedure based on the task feature parameters, parse the process procedure, determine the associated parameters of the motion trajectory and generate the motion trajectory. S3, Simulation Verification: Based on the motion of the 3D model, the correctness of the robot trajectory is verified by simulating the processing process through control technology; S4, Executable program generation: Generates executable code for the robot based on the motion trajectory and its associated parameters.

2. The robot offline programming method based on process knowledge base to generate processes according to claim 1, characterized in that, The process knowledge base is a process knowledge graph. Before step S1, there is also a step of constructing the process knowledge graph: extracting descriptive features of entities, attributes and relationships from scattered unstructured text data of the manufacturing process, and constructing a process knowledge graph containing static knowledge and quasi-static knowledge according to predetermined annotation rules. The static knowledge includes theoretical knowledge and empirical knowledge, and the quasi-static knowledge includes management layer, process layer and process step layer knowledge.

3. The robot offline programming method based on process knowledge base to generate processes according to claim 2, characterized in that, Step S4 is followed by a quality assessment and knowledge update step: based on the process characteristics of the processing, data is collected during the robot's operation, the operation quality is analyzed and evaluated, the process execution effect is analyzed, optimization schemes are deduced, and feedback is given to the process knowledge graph to correct and update its rules.

4. The robot offline programming method based on process knowledge base to generate processes according to claim 2, characterized in that, The simulation verification in step S3 includes: Collision detection: The interference volume is calculated using a relevant algorithm; Virtual-real interaction: Real-time adjustment and verification of robot movement trajectory in a virtual environment; Track correction: If a joint overshoot or collision is detected, the trajectory will be automatically shifted and the map safety rules will be updated.

5. The robot offline programming method based on process knowledge base to generate processes according to claim 1, characterized in that, In step S2, the process knowledge base is a process knowledge graph or a large process model. The interaction process between the process knowledge base and the process includes: matching process parameter cases through semantic similarity algorithm, inferring process parameters and generating process procedures, parsing process procedures and dynamically generating motion trajectories and associated parameters.

6. The robot offline programming method based on process knowledge base to generate processes according to claim 1, characterized in that, In step S4, a post-template file is matched according to the robot type, and an executable program for the robot is generated based on the CAD model, the determined motion trajectory and its associated parameters by parsing the rules of the post-template file.

7. A robot offline programming system based on a process knowledge base for generating processes, characterized in that, include: Process knowledge modeling system: used to define a multi-process manufacturing knowledge system, analyze the descriptive features of entities, attributes and relationships in Chinese processing technology texts, formulate annotation rules, and extract and construct process knowledge graphs; Knowledge Graph Cloud Service Cluster: Based on the process rules generated by the process knowledge modeling system, the domain knowledge graph is stored in the database, providing a knowledge reasoning interface to the offline programming system and a process feedback interface to the quality assessment system; Offline programming system: Dynamically generates post-programs based on process parameters recommended by CAD models and process knowledge graphs; Quality assessment system: Based on the process characteristics of the machining, it collects data during the robot's operation, analyzes and evaluates the quality of the operation, analyzes the effect of process execution, infers optimization schemes, and provides feedback to correct the process knowledge graph rules.

8. The robot offline programming system based on process knowledge base for generating processes according to claim 7, characterized in that, The offline programming system includes: Scene building module: Used to construct a 3D virtual scene that matches the real-world processing environment; Process planning module: used for intelligent process decision-making based on process knowledge graph, realizing processing path selection, process adaptation and path planning generation; Simulation module: used to simulate the machining process, collision detection, and virtual-real interaction, and to virtually verify the generated robot trajectory; Post-program module: Used to convert motion trajectories and their associated parameters into control code that can be executed by the robot.

9. The robot offline programming system based on process knowledge base for generating processes according to claim 8, characterized in that, The knowledge graph cloud service cluster adopts a distributed architecture, including: Process Knowledge Graph Management System: Used for the construction, querying, and node management of process knowledge graphs; Knowledge graph server: Stores process rules and reasoning logic, including a process reasoning module, and communicates with offline programming software through a knowledge reasoning interface to achieve dynamic recommendation of process parameters; Neo4j graph database cluster: knowledge of structured storage technology.

10. The robot offline programming system based on process knowledge base for generating processes according to claim 9, characterized in that, The offline programming system and the process knowledge graph management system achieve data interaction and communication through the process knowledge graph API server, process reasoning server, and quality assessment server.

Images