Experimental method and system for combining robot programming control with industrial assembly simulation

Abstract

This invention provides an experimental method and system combining robot programming control and industrial assembly simulation. The method includes: constructing an industrial assembly scenario library by classifying different scenarios and corresponding assembly and programming tasks; building an experimental platform supporting multiple learning stages and scenarios through local large-scale model deployment, programming mode configuration, and collaborative rule settings; executing programming control and real-time simulation for industrial assembly based on the industrial assembly scenario library and experimental platform; collecting and analyzing multimodal experimental data in real time during the execution of programming control and real-time simulation; and generating experimental evaluations and optimization suggestions based on the analysis results. This invention enhances the practicality of teaching; utilizes a local large-scale model to achieve real-time feedback and error warnings, ensuring operational safety and improving teaching efficiency; and through a multi-dimensional evaluation system, helps teachers accurately grasp student learning progress, promoting the standardization and personalization of programming control and assembly simulation in the industrial field.

Description

Technical Field

[0001] This invention relates to the field of AI education technology, specifically to an experimental method and system that combines robot programming control with industrial assembly simulation. Background Technology

[0002] Current industrial assembly teaching often relies on theoretical explanations or static model demonstrations, lacking dynamic practical components. This makes it difficult for students to understand the programming control and mechanical collaboration processes in real industrial scenarios. Existing technologies often disconnect robot programming instruction from industrial assembly scenarios, failing to simulate multi-task collaboration, error adjustment, and efficiency optimization in real production lines. Furthermore, data collection relies on manual recording, resulting in delayed feedback and hindering personalized guidance. In addition, traditional experimental techniques lack multi-dimensional evaluation of students' programming logic, operational procedures, and teamwork abilities, limiting the development of students' comprehensive skills. Summary of the Invention

[0003] To address the aforementioned shortcomings in existing technologies, this invention provides an experimental method and system that combines robot programming control with industrial assembly simulation.

[0004] According to a first aspect of the present invention, an experimental method combining robot programming control and industrial assembly simulation is provided, comprising: An industrial assembly scenario library is constructed by classifying different scenarios and corresponding assembly and programming tasks; By deploying large local models, configuring programming modes, and setting collaborative rules, an experimental platform that can support multiple academic levels and multiple scenarios is built. Based on the industrial assembly scenario library and experimental platform, programmatic control and real-time simulation for industrial assembly are performed. During the execution of the programming control and real-time simulation, multimodal experimental data is collected and analyzed in real time; Based on the analysis results, experimental evaluation and optimization suggestions are generated.

[0005] Preferably, the construction of an industrial assembly scenario library through different scenario classifications and corresponding assembly and programming tasks includes: Based on the complexity of the scenarios at different learning stages, the scenarios are divided into elementary, intermediate, and advanced stages. For each graded stage, corresponding industrial assembly scenarios and assembly task requirements are designed. Each industrial assembly scenario is matched with a dedicated programming task package, and industrial safety operation specifications are built in. Based on programming and assembly tasks, a mapping relationship is established between programming operations and industrial assembly knowledge points.

[0006] Preferably, the construction of an experimental platform capable of supporting multiple educational stages and scenarios through local large-scale model deployment, programming mode configuration, and collaborative rule settings includes: An industrial assembly model is built based on the Transformer architecture and deployed on a local server; a training dataset is constructed, which includes: an industrial assembly operation manual, programming error cases, and teaching feedback corpus; the industrial assembly model is pre-trained using the training dataset; The configuration programming mode includes two types: graphical programming and code programming, each adapted to different learning stages; Multi-robot collaborative logic is defined based on task timing tables.

[0007] Preferably, the step of performing programmatic control and real-time simulation for industrial assembly based on the industrial assembly scenario library and experimental platform includes: Based on the aforementioned industrial assembly scenario library, control code is written for assembly tasks and programming tasks in different industrial assembly scenarios, and parameter values ​​are specified. Based on the local large model deployed on the experimental platform and the set collaborative rules, the control code logic and parameter values ​​are analyzed in real time, and unreasonable operations are given immediate prompts. A simulation environment is built using the Unity3D engine to display the robot's movements and assembly process in real time, and to generate a high-precision heat map after assembly.

[0008] Preferably, the real-time acquisition and analysis of multimodal experimental data includes: By combining embedded point technology with sensor simulation, programming data, operation data, robot data, and voice feedback data are collected during the programming control and real-time simulation process. The collected data undergoes error identification, efficiency statistics, and capability diagnosis to complete data analysis; among which: The error identification utilizes a local large model deployed on the experimental platform to match pre-trained programming error cases and statistically analyze the occurrence frequency and proportion of various types of errors in the programming data. The efficiency statistics are calculated using the following method: Task completion efficiency = (Standard completion time / Actual completion time of student) × 100% The ability diagnosis identifies students' weaknesses by analyzing the distribution of error types.

[0009] Preferably, generating experimental evaluation and optimization suggestions based on the analysis results includes: Construct a multi-dimensional evaluation system, including: programming accuracy dimension, operational standardization dimension, team collaboration efficiency dimension and / or innovative application dimension; Based on the aforementioned multi-dimensional evaluation system, student performance is quantitatively scored. Based on the ability diagnosis results in the analysis, a personalized learning path is generated, and targeted training suggestions are pushed to students; The system outputs two-dimensional data reports to teachers, one for the class as a whole and one for each individual, and generates teaching optimization suggestions based on the overall class data report. The overall class data report includes: average completion rate of tasks in each scenario, distribution of major error types, and mastery rate of knowledge points. The individual data report includes: student scores in each dimension, details of error types, and progress of personalized training suggestions.

[0010] According to a second aspect of the present invention, an experimental system combining robot programming control and industrial assembly simulation is provided, comprising: The Industrial Assembly Scenario Library Building Module is used to build an industrial assembly scenario library based on different scenario levels and corresponding assembly and programming tasks. The experimental platform construction module is used to build an experimental platform that supports multiple academic levels and multiple scenarios through local large-scale model deployment, programming mode configuration, and collaborative rule settings. The programming control and real-time simulation module, based on the industrial assembly scenario library and experimental platform, performs programming control and real-time simulation for industrial assembly. The data analysis module is used to collect and analyze multimodal experimental data in real time during the execution of the programming control and real-time simulation. The evaluation and suggestion module generates experimental evaluations and optimization suggestions based on the analysis results.

[0011] By adopting the above technical solution, the present invention has at least one of the following beneficial effects compared with the prior art: This invention presents an experimental method and system that combines robot programming control with industrial assembly simulation. By deeply integrating industrial assembly scenarios with programming control, it realizes a concrete industrial assembly simulation environment, enhancing the practicality of teaching and improving students' hands-on skills and logical thinking abilities in a practical training environment.

[0012] This invention presents an experimental method and system that combines robot programming control with industrial assembly simulation. It utilizes a local large-scale model to achieve real-time feedback and error warning, ensuring operational safety and improving teaching efficiency.

[0013] This invention presents an experimental method and system that combines robot programming control with industrial assembly simulation. Through a multi-dimensional evaluation system, it helps teachers accurately grasp students' learning progress and promotes the standardization and personalization of programming control and assembly simulation in the industrial field.

[0014] The experimental method and system of this invention, which combines robot programming control with industrial assembly simulation, more closely resembles the simulated real industrial environment and assembly objects. It is not only suitable for science and technology innovation classroom teaching experiments in primary and secondary schools and vocational education, but can also be applied to industrial design modeling. Attached Figure Description

[0015] Other features, objects, and advantages of the present invention will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings: Figure 1 This is a flowchart illustrating the experimental method combining robot programming control and industrial assembly simulation in a preferred embodiment of the present invention.

[0016] Figure 2 This is a schematic diagram of the composition architecture of an experimental system combining robot programming control and industrial assembly simulation in a preferred embodiment of the present invention.

[0017] Figure 3 This is a flowchart of an experimental method combining robot programming control and industrial assembly simulation in a specific application example of the present invention, illustrating the core process from scenario construction to evaluation feedback. Detailed Implementation

[0018] The embodiments of the present invention are described in detail below: These embodiments are implemented based on the technical solution of the present invention, and provide detailed implementation methods and specific operation processes. It should be noted that those skilled in the art can make several modifications and improvements without departing from the concept of the present invention, and these all fall within the protection scope of the present invention.

[0019] Existing industrial assembly teaching methods often suffer from problems such as lack of tiered scenarios, delayed feedback, simplistic evaluation methods, and a disconnect between simulation and industrial application. To address these issues, one embodiment of this invention provides an experimental method that combines robot programming control with industrial assembly simulation. This method guides students to simulate industrial assembly by programming a robot, achieving the experimental objectives.

[0020] Specifically, such as Figure 1 As shown, the experimental method combining robot programming control and industrial assembly simulation provided in this embodiment may include: S1 constructs an industrial assembly scenario library by classifying different scenarios and corresponding assembly and programming tasks; S2, through local large-scale model deployment, programming mode configuration, and collaborative rule settings, builds an experimental platform that can support multiple academic levels and multiple scenarios; S3, based on an industrial assembly scenario library and experimental platform, performs programmed control and real-time simulation for industrial assembly. S4, during the execution of programming control and real-time simulation, collects and analyzes multimodal experimental data in real time; S5. Based on the analysis results, generate experimental evaluation and optimization suggestions.

[0021] In some preferred embodiments, the above-mentioned S1, which constructs an industrial assembly scenario library through different scenario classifications and corresponding assembly and programming tasks, may further include: S11, based on the complexity of the scenarios at different learning stages, the scenarios are divided into elementary, intermediate and advanced stages; S12 is designed to adapt industrial assembly scenarios and assembly task requirements for each grade stage. The S13 is equipped with a dedicated programming task package for each industrial assembly scenario and includes built-in industrial safety operation specifications. S14. Based on programming tasks and assembly tasks, establish a mapping relationship between programming operations and industrial assembly knowledge points.

[0022] In some preferred embodiments, the above-mentioned S14, establishing the mapping relationship between programming operations and industrial assembly knowledge points, may further include: S141, Construct a structured mapping rule base, label each programming operation element with its corresponding industrial assembly knowledge point, and define its application level under different scenario levels through association rules, thus associating the programming operation element with the industrial assembly knowledge point; wherein, the mapping rule base can be a mapping database or a knowledge graph; S142, based on scenario classification and task requirements, uses a rule engine or machine learning algorithm to dynamically match the correspondence between programming behaviors and related knowledge points; S143 embeds a mapping interface in the experimental platform to perform real-time correlation and retrieval of data, supporting multimodal data acquisition and analysis.

[0023] In this step, the mapping rule base is stored and managed in the form of database tables or knowledge graphs, and can be called in real time by other modules of the experimental platform through the application programming interface, thereby realizing the correlation analysis and diagnosis between programming behavior and knowledge mastery.

[0024] In some preferred embodiments, the above-mentioned S2, through local large-scale model deployment, programming mode configuration, and collaborative rule setting, builds an experimental platform that can support multiple learning stages and multiple scenarios, and may further include: S21. Build an industrial assembly model based on the Transformer architecture and deploy the model on a local server; construct a training dataset, which includes: industrial assembly operation manual, programming error cases, and teaching feedback corpus; and pre-train the industrial assembly model using the training dataset. S22, the configuration programming mode includes two types: graphical programming and code programming, which are adapted to different learning stages; S23, define multi-robot collaborative logic based on task timing table.

[0025] In some preferred embodiments, the above-mentioned S22 may further include: After modifications are made to the graphical programming interface, the code programming interface will be automatically updated, achieving real-time synchronization between graphical programming and code programming.

[0026] In some preferred embodiments, the above-mentioned S23, which defines multi-robot collaborative logic based on the task timing table, may further include: The overall assembly process is broken down into multiple sequential or parallel sub-task units. For each sub-task unit, the robot role to be executed, the action instructions, the time nodes, and the completion trigger conditions are clearly specified, thereby generating a structured task sequence table. Based on this, a structured task timing table is generated, and the central scheduler of the experimental platform coordinates subtasks in a unified manner according to the table. By monitoring the "status completion signal" of each robot subtask in real time, subsequent tasks are triggered sequentially, and access permissions of multiple robots to the shared workspace and tool resources are managed according to preset rules, thereby ensuring the timing correctness and logical coordination of the entire assembly process.

[0027] In some preferred embodiments, the above-mentioned S3, based on the industrial assembly scenario library and experimental platform, performs programmed control and real-time simulation for industrial assembly, and may further include: S31, based on the industrial assembly scenario library, writes control code for assembly tasks and programming tasks in different industrial assembly scenarios, and specifies parameter values; S32, based on the local large model deployed on the experimental platform and the set collaborative rules, analyzes the control code logic and parameter values ​​in real time and provides immediate prompts for unreasonable operations; S33 uses the Unity3D engine to build a simulation environment that displays the robot's movements and assembly process in real time, and generates a high-precision heatmap after assembly; specifically: S331 builds a standardized industrial assembly scene model in Unity3D, imports STEP / STL format industrial parts models and performs lightweight processing to ensure a balance between model accuracy and rendering efficiency. S332 sets physical properties such as part collision body, friction coefficient, and gravity parameters to simulate the real part contact, gripping, assembly force process. S333 uses a script to write the robot motion control interface, receives external programming instructions, and drives the virtual robot to complete real-time updates of joint angles and end effector poses. S334, the experimental platform parses the written control code into robot motion instructions and transmits them to the Unity3D simulation environment in real time via TCP / IP protocol; S335 uses a high refresh rate to render robot motion, simultaneously displaying key assembly steps such as the end effector grasping parts, precise alignment, and bolt tightening. At the same time, it displays key parameters such as robot joint angles, movement speed, and force feedback values ​​in real time on the interface. S336. If abnormalities such as part assembly deviation or collision risk are detected, the Unity3D simulation interface will provide real-time alerts through text annotations and other means, and pause the simulation process. S337. After assembly, collect the deviation between the actual coordinates and theoretical coordinates of each assembly point, as well as the contact stress distribution data of the part surface. S338 uses color-coded rendering of assembled parts in a Unity3D scene based on the stress value to generate a high-precision heatmap.

[0028] In some preferred embodiments, the above-mentioned S4, which involves real-time acquisition and analysis of multimodal experimental data, may further include: S41 uses a combination of embedded point technology and sensor simulation to collect programming data, operation data, robot data and voice feedback data during programming control and real-time simulation. S42, performs error identification, efficiency statistics, and capability diagnosis on the collected data to complete data analysis; among which: S421, Error Identification: Using a large local model deployed on the experimental platform, matching pre-trained programming error cases, and statistically analyzing the frequency and proportion of various types of errors in the programming data. S422, Efficiency Statistics, calculates efficiency values ​​using the following methods: Task completion efficiency = (Standard completion time / Actual completion time of student) × 100% S423, Ability Diagnosis, identifies students' weaknesses by analyzing the distribution of error types.

[0029] In some preferred embodiments, the above-mentioned S41, the embedding technology and sensor simulation, may further include: S411, data acquisition points are embedded in the user interface layer, business logic layer and simulation engine layer of the experimental platform respectively; S412 records the sequence of interface operation events and programming behaviors at the user interface layer; S413 captures program execution status and task progress data in the business logic layer; S414 acquires robot motion data, end effector status data, assembly accuracy data, and environmental interaction data in real time through a constructed virtual sensor model at the simulation engine layer.

[0030] In some preferred embodiments, the programming data in S41 above includes: code modification frequency, error type, and instruction integrity; the acquisition frequency is real-time acquisition. Operational data includes: task completion time, number of pauses, and number of parameter adjustments; operational data is automatically summarized after the task ends. Robot data includes: motion trajectory accuracy, coordination and synchronization error, and task success rate; the data collection frequency is based on a set time interval. Voice feedback data includes: student questions and team communication scripts; the voice feedback data is recorded in its entirety through the platform's built-in recording function and converted into text data using speech-to-text technology.

[0031] In some preferred embodiments, the above-mentioned S5, which generates experimental evaluation and optimization suggestions based on the analysis results, may further include: S51, construct a multi-dimensional evaluation system, including: programming accuracy dimension, operational standardization dimension, team collaboration efficiency dimension and / or innovative application dimension; S52, based on a multi-dimensional evaluation system, quantifies and scores student performance. S53, based on the ability diagnosis results in the analysis, generates a personalized learning path and pushes targeted training suggestions to students; specifically: S531, based on multi-dimensional scoring and multimodal data analysis, identifies students' weaknesses in dimensions such as programming accuracy, operational standardization, teamwork efficiency, and innovative application. S532 provides tiered learning resources tailored to weaknesses, from basic knowledge points and specialized practical training to advanced application scenarios, forming a progressive improvement path. S533 pushes personalized learning tasks, reference cases and improvement directions to students through the experimental platform, and clarifies training objectives and assessment standards; S554: After students complete the training, the platform re-collects data, updates the evaluation results, and dynamically adjusts the learning path. S54 outputs a dual-dimensional data report to teachers, covering both the overall class and individual data, and generates teaching optimization suggestions based on the overall class data report; specifically: S541 outputs individual reports on students' abilities, weaknesses, and training status, as well as overall reports on class average scores, pass rates, and common problems. S542, based on the common weaknesses of the class, adjust the teaching focus, optimize the difficulty and duration of experiments, and supplement targeted teaching content and error correction cases to form a teaching optimization plan that is suitable for the class's learning situation; The class-wide data report includes: average completion rate of tasks in each scenario, distribution of major error types, and mastery rate of knowledge points; the individual data report includes: student scores in each dimension, details of error types, and progress of personalized training suggestions.

[0032] In some preferred embodiments, the above-mentioned S51 includes the following dimensions: programming accuracy, which is scored by combining the error rate of programming data and the completeness of programming instructions; operational standardization, which is scored based on the compliance rate of industrial safety operation standards and the rationality of parameter settings; team collaboration efficiency, which is scored by communication frequency and task division completion rate; and innovative application, which is scored based on whether students have optimized programming logic or proposed new assembly schemes.

[0033] Based on the same inventive concept, one embodiment of the present invention also provides an experimental system that combines robot programming control with industrial assembly simulation.

[0034] Specifically, such as Figure 2 As shown, the experimental system combining robot programming control and industrial assembly simulation provided in this embodiment may include: The Industrial Assembly Scenario Library Building Module is used to build an industrial assembly scenario library based on different scenario levels and corresponding assembly and programming tasks. The experimental platform construction module is used to build an experimental platform that supports multiple academic levels and multiple scenarios through local large-scale model deployment, programming mode configuration, and collaborative rule settings. The programming control and real-time simulation module, based on an industrial assembly scenario library and experimental platform, performs programming control and real-time simulation for industrial assembly. The data analysis module is used to collect and analyze multimodal experimental data in real time during the execution of programming control and real-time simulation. The evaluation and suggestion module generates experimental evaluations and optimization suggestions based on the analysis results.

[0035] The specific implementation methods of each functional module constituting the system provided in the above embodiments of the present invention will be further described in detail below with reference to specific examples.

[0036] The industrial assembly scenario library construction module ensures that the scenarios are both appropriate for students' cognitive levels and possess industrial realism through "scenario grading adapted to different learning stages" and "multi-dimensional resource association." It includes the following units: The scenario complexity grading unit is designed for elementary school students with a "parts sorting and handling" scenario. This scenario uses cylindrical parts with a diameter of 20-30mm and cube parts with a diameter of 20×20×20mm, and a small educational robot with a load capacity of ≤500g. The robot's maximum moving speed is set to 80mm / s. Students are required to programmatically control the robot to move parts to designated workstations spaced 300mm apart according to their shape and color. The task completion time is 10 minutes. For middle school students, a "simple component assembly" scenario is designed, with "basic gearbox assembly" as the core task. This scenario uses parts with 12 teeth, a module of 2, and made of plastic. The system uses gears and bushings with an inner diameter of 4mm, and is equipped with a desktop robotic arm with a load capacity of ≤1kg. The robotic arm's gripping force range is set to 5-10N, and the allowable error for assembly gaps is ±0.5mm. The task focuses on the basic process of "positioning-gripping-assembly-gap detection". For high school and vocational education, a "multi-robot collaborative assembly" scenario is designed, with "automobile wheel hub assembly" as the core task. 2-3 industrial-grade simulation robots (model ABBIRB120) are configured. Robot A is responsible for gripping a 380mm diameter, 2kg wheel hub; Robot B is responsible for gripping an M12×30mm bolt; and Robot C is responsible for tightening the bolt (tightening torque 80N). The task requires multi-robot collaborative synchronization error ≤ 1s, and adds advanced requirements such as "precision calibration" and "fault diagnosis".

[0037] The scene resource association unit matches a dedicated programming task package for each scene. The basic instruction library includes core instructions such as "robot arm movement (move(x,y,z))", "grab (grab(force))", and "release (release())". The advanced challenge library adds the "assembly gap detection (check_gap())" instruction for junior high school scenes and the "cooperative signal sending (send_signal())" instruction for senior high school scenes. It also incorporates industrial safety operation specifications, such as the rules set in the senior high school scenes that "tightening action is prohibited when bolts are not aligned with shaft holes" and "the overlap of working radii of adjacent robots is ≤50mm". It establishes a mapping relationship between "programming operation - industrial assembly knowledge points", for example, "robot arm gripping force setting" maps to "material mechanical load calculation (F=μmg, where μ is the friction coefficient and m is the mass of the part)", "multi-robot speed matching" maps to "production line cycle optimization", and "bolt tightening torque control" maps to "mechanical connection strength design".

[0038] The experimental platform construction module, through "local large-scale model deployment," "programming mode adaptation," and "collaborative rule settings," builds an experimental platform that supports multiple educational levels and scenarios. It includes the following units: The large-scale model deployment and programming mode configuration unit is built on a finely tuned version of the MiniGPT-4 based on the Transformer architecture to create an industrial assembly model. Serving as an "industrial assembly AI mentor," it is deployed on a local server equipped with an Intel Xeon E3-1230v6 CPU and 32GB of memory. The training data includes 500+ industrial assembly operation manuals, 1000+ programming error cases (including loop logic errors, coordinate offsets, missing parameters, etc.), and 800+ teaching feedback corpora, ensuring that the AI ​​mentor has the ability to identify errors, provide instruction prompts, and issue safety warnings. In terms of programming mode configuration, this unit supports both graphical and code-based programming. The mode automatically adapts via the "student level selection button." For elementary school students, the default is Blockly graphical programming (students can generate corresponding code by dragging and dropping modules such as "move" and "grab"). For middle school and above students, they can switch to Python code programming (which supports syntax highlighting and auto-completion). The two modes are synchronized in real time—when the graphical module is modified, the Python code will be automatically updated, ensuring that students can gradually transition from graphical programming to code-based programming.

[0039] The collaborative rule setting unit defines collaborative logic based on a "task sequence table" for multi-robot collaborative rule settings. Taking the high school "car wheel assembly" scenario as an example, robot A first completes wheel hub positioning and sends a "ready_A" signal to the system. After receiving the signal, the system sends a "start gripping bolt" command to robot B after a delay of ≤0.5s. After robot B completes bolt placement, it sends a "ready_B" signal to the system, triggering robot C to start tightening. If a robot fails to send a signal for more than 2s, the system automatically pauses the experiment and displays a "Robot B response timed out, check communication connection" message to ensure the stability of multi-robot collaboration.

[0040] The programming control and real-time simulation module enables student interaction with the experimental system through "programming instruction execution," "real-time AI tutor alerts," and "dynamic simulation demonstrations," ensuring safe and controllable programming operations. It includes the following units: The programming instruction execution unit involves students writing control code based on the instruction library provided by the platform. Taking the junior high school "gearbox primary assembly" scenario as an example, the Python code written by students must include core steps such as robotic arm initialization (arm=RobotArm(model="Desktop_Arm_01")), moving to the gear storage position (arm.move(x=200,y=150,z=50,speed=60)), grabbing the gear (arm.grab(force=8)), moving to the shaft assembly position (arm.move(x=300,y=150,z=50,speed=40)), releasing the gear (arm.release()), and checking the assembly gap (gap=arm.check_gap()). The code must specify the parameter values ​​(such as the unit of movement coordinates being mm, the unit of speed being mm / s, and the unit of gripping force being N).

[0041] The large-scale model early warning unit, within the AI ​​tutor's real-time warning process, provides immediate alerts for unreasonable operations by analyzing code logic and parameter values ​​in real time. For example, if a student sets the robotic arm's movement speed to exceed the corresponding scenario threshold (e.g., setting the speed to 85mm / s in a middle school scenario, exceeding the safety limit of 80mm / s), the system will display a message: "Current speed 85mm / s, exceeding the gear assembly safety threshold (80mm / s), it is recommended to reduce to 60-70mm / s (based on: the gear mass is 500g, excessive speed can easily cause centrifugal force to exceed the gripping force, causing parts to fall off)." If the "grab command" in the code does not set the gripping force parameter (e.g., only arm.grab() is written), the AI ​​tutor will mark the error and provide an example: "Gripping force parameter is missing. Refer to the correct format: arm.grab(force=5-10N) (Based on: the gear weight is 5N, the gripping force needs to be 1.2-2 times the weight to ensure stable gripping)". If the multi-robot collaborative code does not set the signal interaction logic, it will prompt: "Collaborative signal command is missing. You need to add send_signal() or receive_signal() to realize inter-robot communication".

[0042] The dynamic simulation unit uses the Unity3D engine to build a simulation environment to demonstrate the robot's movements and assembly process in real time. The robot arm's end effector trajectory is marked with red dotted lines, and playback at 0.5-2x speed is supported, allowing students to easily observe whether the movement path is reasonable. Different colored indicator lights indicate the robot's status (green for running, yellow for waiting for a signal, and red for a fault), visually presenting the progress of multi-robot collaboration. After assembly, a "precision heatmap" is generated, with red areas indicating assembly gaps exceeding tolerance (>0.5mm) and green areas indicating acceptable gaps, helping students quickly identify assembly problems.

[0043] The data analysis module, through "full-process data collection" and "AI intelligent analysis," achieves quantitative recording and problem identification of the experimental process, providing data support for subsequent evaluation. It includes the following units: The multi-source data acquisition unit employs "embedded point technology + sensor simulation" to collect four types of core data in the multi-source data acquisition stage: Programming data includes code modification frequency (times / minute), error type (e.g., loop logic errors manifested as a for loop without a termination condition, coordinate offset errors manifested as a deviation between actual and target coordinates > 1mm), and instruction completeness (e.g., whether it includes necessary parameters such as gripping force and movement speed). Data acquisition is real-time, recording once per second. Operational data includes task completion time (total duration from programming start to assembly end), number of pauses (experiment pauses due to code errors or unreasonable parameters), and number of parameter adjustments (e.g., modification). The data (number of times the movement speed and gripping force were adjusted) is automatically summarized after the task ends; robot data includes motion trajectory accuracy (deviation between actual coordinates and target coordinates), collaborative synchronization error (time difference between multiple robot actions), and gripping success rate (ratio of successful part gripping to total gripping), collected once every 0.5 seconds; voice feedback data includes student questions (such as "how to adjust assembly gaps") and team communication scripts (such as "you are responsible for programming robot A, and I am responsible for setting the parameters of robot B"), which are recorded in the whole process through the platform's built-in recording function and converted into text data using speech-to-text technology (recognition accuracy ≥98%), facilitating subsequent analysis of team collaboration.

[0044] The data analysis unit processes data in a three-step process: error identification, efficiency statistics, and capability diagnosis. In the error identification phase, a pre-defined error case library is used for matching. For example, a "loop logic error" is defined as "a for loop without a range parameter (e.g., for i in range()) or repeated actions within the loop causing the robot to remain stationary," and a "coordinate offset error" is defined as "the actual movement coordinate deviates from the target coordinate by more than 1mm and lasts for more than 3 seconds." The frequency and percentage of each type of error are statistically analyzed. In the efficiency statistics phase, the efficiency value is calculated using the formula: "Task completion efficiency = (Standard completion time for the scenario / Actual completion time for the student) × 100%." ​​The standard completion time for elementary school scenarios is 5 minutes, for middle school it is 8 minutes, and for high school it is 12 minutes. An efficiency value ≥ 90% is considered highly efficient. In the capability diagnosis phase, the distribution of error types identifies students' weaknesses. For example, "coordinate offset error percentage ≥ 40%" indicates "weak spatial coordinate cognition," "team communication frequency < 2 times / minute" indicates "collaboration ability needs improvement," and "number of safety rule violations ≥ 2 times" indicates "insufficient industrial safety awareness."

[0045] The evaluation and suggestions module, through "multi-dimensional evaluation," "personalized learning paths," and "teaching optimization suggestions," achieves a two-way empowerment of "student ability improvement and teacher curriculum adjustment." It includes the following units: The multi-dimensional evaluation unit quantifies student performance based on four core dimensions: Programming accuracy, which combines error rate (number of errors / total number of instructions) and instruction completeness. An error rate ≤5% with complete instructions earns 90-100 points; an error rate of 6-15% earns 70-89 points; and an error rate >15% earns <70 points. Operational compliance is based on safety rule adherence rate (number of compliant operations / total number of operations) and parameter setting rationality. A compliance rate of 100% with reasonable parameters earns 90-100 points; a compliance rate of 80-99% earns 70-89 points; and a compliance rate <80% earns <70 points. 0% gets <70 points; In the team collaboration efficiency dimension, scores are based on communication frequency (effective communication times / minute) and task completion rate. A completion rate of 100% and communication frequency ≥2 times / minute earns 90-100 points, a completion rate of 80-99% earns 70-89 points, and a completion rate of <80% earns <70 points; In the innovative application dimension, bonus points are awarded based on whether students have optimized programming logic (e.g., simplifying multi-part assembly steps through loop instructions) and proposed new assembly solutions (e.g., adjusting the robot's action sequence to shorten assembly time). Optimization with a feasible solution earns 10 points, optimization with a solution that needs improvement earns 5 points, and no optimization earns 0 points.

[0046] The targeted suggestion push unit provides personalized learning path generation based on capability diagnosis results, offering targeted training suggestions: If "coordinate offset error rate is 45%", it recommends completing the "Spatial Coordinate Transformation Special Training", which includes three sub-tasks: "2D coordinate mapping (e.g., converting planar coordinates to robot joint coordinates)," "3D coordinate calibration (e.g., correcting coordinate deviations through reference points)," and "error compensation programming (e.g., writing code to automatically correct offsets within ±0.5mm)," accompanied by the "Basic Learning Materials for Robotic Arm Coordinate Systems" (downloadable directly from the platform); If "team collaboration efficiency score < 70 points", it recommends participating in the "Multi-robot Collaborative Simulation Group Task", which requires two people to work in a group, each responsible for programming different robots, and to synchronize programming progress and parameter settings through the platform's built-in chat function to train collaboration and communication skills; If "the number of violations of industrial safety rules is ≥ 2 times", it recommends a "Micro-course on Industrial Assembly Safety Specifications" (10 minutes long), which includes core content such as robotic arm safe operation and multi-device collaborative protection.

[0047] The optimization suggestion generation unit provides teachers with a two-dimensional report: "class-wide + individual." The class-wide progress report includes the average completion rate of tasks in each scenario (e.g., the average completion rate of the "gear assembly" task in junior high school is 82%), the distribution of major error types (e.g., "missing assembly gap detection instructions" accounts for 60% of all errors), and the mastery rate of knowledge points (e.g., the mastery rate of the "robotic arm coordinate calibration" knowledge point is only 65%). Based on the overall data, teaching adjustment suggestions are proposed, such as "the mastery rate of the 'robotic arm coordinate calibration' knowledge point is low, it is recommended to add a 15-minute physical demonstration (using a desktop robotic arm to demonstrate the impact of coordinate deviation on assembly accuracy)" and "the class team collaboration score pass rate is only 75%, it is recommended to adopt a fixed group model for subsequent experiments to strengthen division of labor and communication training." The individual report includes students' scores in various dimensions, details of error types, and personalized training completion progress, which facilitates teachers to conduct targeted tutoring.

[0048] It should be noted that the steps in the method provided by the present invention can be implemented using the corresponding components in the system. Those skilled in the art can refer to the technical solution of the system to implement the steps of the method, and can also refer to the technical solution of the method to implement the composition of the system. That is, the embodiments in the system and the embodiments in the method can be understood as preferred examples of each other, which will not be elaborated here.

[0049] The technical solution provided by the above embodiments of the present invention will be further described in detail below with reference to a specific verification example.

[0050] like Figure 3 As shown, in this specific application example, guiding students to simulate the industrial assembly process by controlling a robot through programming mainly includes the following five core steps: Step 1: Build an industrial assembly scenario library; Step 2: Configure the programming control platform and robot simulation environment; Step 3: Perform programmed control and real-time simulation; Step 4: Collect and analyze multimodal experimental data; Step 5: Generate experimental evaluation and optimization suggestions.

[0051] First, in the "Building an Industrial Assembly Scenario Library" phase, the system establishes a hierarchical scenario system based on teaching needs. Taking the junior high school "Simple Component Assembly" course as an example, teachers can select the "Gearbox Assembly" scenario. This scenario relates to knowledge points such as "Mechanical Transmission Principles" and "Tolerance Fits," and provides digital resources such as 3D gear models and assembly process cards. The system automatically adjusts the scenario complexity according to the characteristics of the grade level. The junior high school stage focuses on the basic process of "positioning-assembly-inspection," while the senior high school stage adds advanced requirements such as "precision calibration" and "fault diagnosis."

[0052] Next, we proceeded to the "Configuring the Programming Control Platform and Robot Simulation Environment" section. The platform deploys a web-based integrated development environment, supporting both Blockly graphical programming and Python code programming modes. In the "Automotive Wheel Assembly" experiment, the AI ​​tutor module pre-sets industrial standard knowledge such as bolt tightening sequence and torque control. The simulation engine loads a physical model, including robot working range limitations and component weight parameters, ensuring the virtual simulation closely resembles a real industrial scenario.

[0053] In the "Programming Control and Real-Time Simulation" step, students write programs to control the robot to complete designated tasks. For example, in the "Circuit Board Assembly" scenario, students need to program a suction cup robotic arm to pick up electronic components and accurately place them on the PCB board. When a logical error occurs in the program, leading to a risk of component collision, the system provides a real-time prompt: "Insufficient spacing detected for components; adjustment of the placement path is recommended." The simulation interface synchronously displays dynamic effects such as the robot's movement trajectory and component assembly status.

[0054] The step of "collecting and analyzing multimodal experimental data" is integrated throughout the entire experimental process. The system records students' programming behavior data (code modification frequency, debugging strategies), operation process data (task completion time, error type), and robot status data (motion trajectory accuracy, collaboration efficiency). In the high school "multi-robot collaborative assembly" project, the system evaluates the efficiency of the scheduling algorithm programmed by students by analyzing the idle rate and task waiting time of each robot.

[0055] Finally, the "Generate Experiment Evaluation and Optimization Suggestions" step integrates all data to generate a multi-dimensional evaluation report. Taking the group project as an example, the report includes indicators such as: task completion rate (85%), programming standardization (4.2 / 5 points), collaboration efficiency (78%), and innovation. The system pushes personalized exercises based on weak areas, such as "Coordinate transformation accuracy needs improvement; it is recommended to complete specialized training in positioning accuracy."

[0056] The experimental method and system combining robot programming control and industrial assembly simulation provided in the above embodiments of the present invention design graded industrial scenarios according to academic stages, reducing the learning threshold and adapting to students with different cognitive levels; it realizes real-time early warning of programming errors and automatic data analysis based on a local large model, shortening the feedback cycle; it constructs a multi-dimensional evaluation system of "programming-operation-collaboration" to improve the cultivation of students' engineering literacy; and it links industrial standards and knowledge points to ensure that the simulation scenario closely resembles the real industrial assembly process.

[0057] Any matters not covered in the above embodiments of the present invention are well-known in the art.

[0058] The specific embodiments of the present invention have been described above. It should be understood that the present invention is not limited to the specific embodiments described above, and those skilled in the art can make various modifications or variations within the scope of the claims, which do not affect the essence of the present invention.

Claims

1. An experimental method combining robot programming control and industrial assembly simulation, characterized in that, include: An industrial assembly scenario library is constructed by classifying different scenarios and corresponding assembly and programming tasks; By deploying large local models, configuring programming modes, and setting collaborative rules, an experimental platform that can support multiple academic levels and multiple scenarios is built. Based on the industrial assembly scenario library and experimental platform, programmatic control and real-time simulation for industrial assembly are performed. During the execution of the programming control and real-time simulation, multimodal experimental data is collected and analyzed in real time; Based on the analysis results, experimental evaluation and optimization suggestions are generated.

2. The experimental method combining robot programming control and industrial assembly simulation according to claim 1, characterized in that, The process involves constructing an industrial assembly scenario library through different scenario classifications and corresponding assembly and programming tasks, including: Based on the complexity of the scenarios at different learning stages, the scenarios are divided into elementary, intermediate, and advanced stages. For each graded stage, corresponding industrial assembly scenarios and assembly task requirements are designed. Each industrial assembly scenario is matched with a dedicated programming task package, and industrial safety operation specifications are built in. Based on programming and assembly tasks, a mapping relationship is established between programming operations and industrial assembly knowledge points. This mapping relationship includes: Build a structured mapping rule base to associate programming operation elements with industrial assembly knowledge points; Based on scenario classification and task requirements, a rule engine or machine learning algorithm is used to dynamically match the correspondence between programming behaviors and relevant knowledge points. An embedded mapping interface enables real-time data association and retrieval, supporting multimodal data acquisition and analysis.

3. The experimental method combining robot programming control and industrial assembly simulation according to claim 1, characterized in that, The aforementioned method, through local large-scale model deployment, programming mode configuration, and collaborative rule settings, establishes an experimental platform capable of supporting multiple educational stages and scenarios, including: An industrial assembly model is built based on the Transformer architecture and deployed on a local server; a training dataset is constructed, which includes: an industrial assembly operation manual, programming error cases, and teaching feedback corpus; the industrial assembly model is pre-trained using the training dataset; The configuration programming mode includes two types: graphical programming and code programming, each adapted to different learning stages; Multi-robot collaborative logic is defined based on task timing tables.

4. The experimental method combining robot programming control and industrial assembly simulation according to claim 3, characterized in that, The configuration programming mode includes two types: graphical programming and code programming, and also includes: After the graphical programming is modified, the code programming will be automatically updated, realizing real-time synchronization between graphical programming and code programming; The multi-robot collaborative logic defined based on the task timing table includes: The overall assembly task is broken down into multiple ordered sub-task nodes, and each sub-task node is assigned an execution role, time constraints and triggering conditions are set to generate a structured task sequence table. Subtasks are scheduled according to the task timing table, and the status completion signals of each robot subtask are monitored to trigger the start command of the next stage task.

5. The experimental method combining robot programming control and industrial assembly simulation according to claim 1, characterized in that, The process of performing programmed control and real-time simulation for industrial assembly based on the industrial assembly scenario library and experimental platform includes: Based on the aforementioned industrial assembly scenario library, control code is written for assembly tasks and programming tasks in different industrial assembly scenarios, and parameter values ​​are specified. Based on the local large model deployed on the experimental platform and the set collaborative rules, the control code logic and parameter values ​​are analyzed in real time, and unreasonable operations are given immediate prompts. A simulation environment is built using the Unity3D engine to display the robot's movements and assembly process in real time, and to generate a high-precision heat map after assembly.

6. The experimental method combining robot programming control and industrial assembly simulation according to claim 1, characterized in that, The real-time acquisition and analysis of multimodal experimental data includes: By combining embedded point technology with sensor simulation, programming data, operation data, robot data, and voice feedback data are collected during the programming control and real-time simulation process. The collected data undergoes error identification, efficiency statistics, and capability diagnosis to complete data analysis; among which: The error identification utilizes a local large model deployed on the experimental platform to match pre-trained programming error cases and statistically analyze the occurrence frequency and proportion of various types of errors in the programming data. The efficiency statistics are calculated using the following method: Task completion efficiency = (Standard completion time / Actual completion time of student) × 100% The ability diagnosis identifies students' weaknesses by analyzing the distribution of programming error types.

7. The experimental method combining robot programming control and industrial assembly simulation according to claim 6, characterized in that, The embedded point technology and sensor simulation include: Non-intrusive data acquisition points are preset in the graphical programming interface, code editor, and simulation control logic to actively capture the user's operation sequence, code modification behavior, and program running status; A virtual sensor model is built in the Unity3D simulation environment to simulate and collect the robot's joint angles, end-effector poses, motion trajectories, collision information, and interaction force data between the robot and the assembled workpiece in real time. The programming data includes: code modification frequency, error type, and instruction integrity; the data is collected in real time. The operation data includes: task completion time, number of pauses, and number of parameter adjustments; the operation data is automatically summarized after the task ends. The robot data includes: motion trajectory accuracy, collaborative synchronization error, and task success rate; the collection frequency is based on a set time. The voice feedback data includes: student questions and team communication scripts; the voice feedback data is recorded in its entirety by the platform's built-in recording function and converted into text data using speech-to-text technology.

8. The experimental method combining robot programming control and industrial assembly simulation according to claim 1, characterized in that, Based on the analysis results, experimental evaluation and optimization suggestions are generated, including: Construct a multi-dimensional evaluation system, including: programming accuracy dimension, operational standardization dimension, team collaboration efficiency dimension and / or innovative application dimension; Based on the aforementioned multi-dimensional evaluation system, student performance is quantitatively scored. Based on the ability diagnosis results in the analysis, a personalized learning path is generated, and targeted training suggestions are pushed to students; The system outputs two-dimensional data reports to teachers, one for the class as a whole and one for each individual, and generates teaching optimization suggestions based on the overall class data report. The overall class data report includes: average completion rate of tasks in each scenario, distribution of major error types, and mastery rate of knowledge points. The individual data report includes: student scores in each dimension, details of error types, and progress of personalized training suggestions.

9. The experimental method combining robot programming control and industrial assembly simulation according to claim 8, characterized in that, The programming accuracy dimension is scored by combining the error rate of programming data and the completeness of programming instructions; the operational standardization dimension is scored based on the compliance rate of industrial safety operation standards and the rationality of parameter settings; the team collaboration efficiency dimension is scored by communication frequency and task division completion rate; the innovative application dimension is scored based on whether students have optimized programming logic and proposed new assembly solutions.

10. An experimental system combining robot programming control and industrial assembly simulation, characterized in that, include: The Industrial Assembly Scenario Library Building Module is used to build an industrial assembly scenario library based on different scenario levels and corresponding assembly and programming tasks. The experimental platform construction module is used to build an experimental platform that supports multiple academic levels and multiple scenarios through local large-scale model deployment, programming mode configuration, and collaborative rule settings. The programming control and real-time simulation module, based on the industrial assembly scenario library and experimental platform, performs programming control and real-time simulation for industrial assembly. The data analysis module is used to collect and analyze multimodal experimental data in real time during the execution of the programming control and real-time simulation. The evaluation and suggestion module generates experimental evaluations and optimization suggestions based on the analysis results.

Images