Method, device and system for automatically generating operation instruction by numerical control machine tool controller based on language large model

Abstract

The invention belongs to the technical field of numerical control machine tools, and particularly relates to a method, device and system for a numerical control machine tool controller to automatically generate an operation instruction based on a language large model.The method comprises the first step of obtaining operation requirements and problems of operators, the second step of preliminarily screening the operation requirements and problem fields, and the third step of setting the numerical control machine tool controller, the method comprises the following steps of 1, establishing a multi-modal model, and obtaining multi-modal information of the numerical control machine tool through a numerical control machine tool controller, 4, constructing the multi-modal large model, and inputting the multi-modal information obtained in the third step into the multi-modal large model for processing, 5, executing a control instruction and an operation prompt, and 6, performing real-time feedback and dynamic adjustment. According to the method, the operation requirement of an operator can be obtained in a voice or character mode, and the corresponding control instruction is automatically generated by utilizing the embedded multi-mode large model in combination with the current state of the machine tool and the operation interface information.

Description

Technical Field

[0001] The present invention belongs to the technical field of numerical control machine tools, and in particular relates to a method, a device and a system for a numerical control machine tool controller to self-generate operation instructions based on a large language model. Background Art

[0002] With the rapid development of the manufacturing industry, CNC machine tools, as the core equipment for precision machining, have a direct impact on the processing efficiency and product quality through their intelligent operation and control technology. However, existing CNC machine tool controllers still face many challenges in the generation and operation of control instructions.

[0003] First of all, the setting of process parameters is the most complex and critical part of CNC machine tool operation. Under different material, tool and processing speed requirements, the reasonable setting of process parameters has a decisive influence on the processing results. For example, when processing cemented carbide, if the cutting speed is too fast or the cutting depth is too large, it is very easy to cause the tool to wear more severely or even break, thus affecting the processing accuracy and efficiency. Most existing controllers rely on the operator's experience to adjust parameters and lack intelligent automatic optimization capabilities.

[0004] Secondly, CNC machine tools support multiple control modes, such as manual, automatic and semi-automatic. Switching and adjusting between these modes requires operators to have high professional skills. In the actual processing process, operators often need to frequently switch control modes according to changes in processing tasks and manually input corresponding control instructions. This operation method is not only inefficient, but also prone to processing accidents due to human errors.

[0005] Furthermore, with the continuous advancement of processing technology, CNC machine tools need to handle more complex control logic and data operations. Especially in advanced applications such as adaptive processing and automated processing, the controller needs to analyze various data in the processing process in real time and make corresponding adjustments and decisions. This puts higher demands on the computing power and intelligence level of the controller, and existing controllers are often unable to cope with such tasks.

[0006] In addition, CNC machine tools of different brands and models have significant differences in operating interface, functions and programming. When operators change equipment, they need to relearn and adapt to new operating methods, which undoubtedly increases the difficulty of operation and training costs. Such differences between equipment limit the popularization and expansion of the application scope of CNC machine tools.

[0007] In summary, existing CNC machine tool controllers have problems in the generation and operation of control instructions, such as complex process parameter settings, difficulty in switching between multiple control modes, complex control logic, and large differences in operation between devices. These problems not only affect processing efficiency and product quality, but also increase the burden and training costs of operators.

[0008] Therefore, developing a CNC machine tool controller that can intelligently generate control instructions has important practical significance and application value. Summary of the invention

[0009] The purpose of the present invention is to provide a method, device and system for a CNC machine tool controller to self-generate operating instructions based on a language large model, which can obtain the operator's operating requirements through voice or text, and combine the current state of the machine tool with the operation interface information, and use the embedded multimodal large model to automatically generate corresponding control instructions.

[0010] The technical solution adopted by the present invention is as follows:

[0011] A method for a numerical control machine tool controller to self-generate operation instructions based on a language large model comprises the following steps:

[0012] Step 1: Obtain the operator's operational requirements and problems;

[0013] Step 2: Preliminary screening of operational needs and problem areas;

[0014] Step 3: Set up the CNC machine tool controller and obtain the multi-modal information of the CNC machine tool through the CNC machine tool controller;

[0015] Step 4: construct a large multimodal model, and input the multimodal information obtained in the third step into the large multimodal model for processing;

[0016] Step 5: Execute control instructions and operation prompts;

[0017] Step 6: Real-time feedback and dynamic adjustment.

[0018] A device for a numerical control machine tool controller to self-generate operation instructions based on a language large model, comprising:

[0019] A controller unit, used to obtain the operator's operational requirements and questions and perform preliminary screening;

[0020] An information acquisition unit, used for acquiring multimodal information of the CNC machine tool;

[0021] A multimodal large model unit, used for processing the acquired multimodal information;

[0022] A control instruction and operation prompt execution unit, used for executing control instructions and operation prompts;

[0023] The real-time feedback and dynamic adjustment unit is used to receive new operational requirements and problems in real time during the operation process and make dynamic adjustments.

[0024] A system for a numerical control machine tool controller to self-generate operation instructions based on a language large model, comprising:

[0025] a processor and a memory for storing instructions executable by the processor;

[0026] The processor is configured to perform the following functions:

[0027] Obtain the operational requirements and problems of operators and conduct preliminary screening;

[0028] Obtain multimodal information of CNC machine tools;

[0029] Processing the acquired multimodal information;

[0030] Execute control instructions and operation prompts;

[0031] Receive new operational requirements and issues in real time during the operation process and make dynamic adjustments.

[0032] The technical effects achieved by the present invention are:

[0033] The present invention can obtain the operator's operating requirements through voice or text, and combine the current state of the machine tool and the operation interface information, and use the embedded multimodal large model to automatically generate corresponding control instructions. This method not only simplifies the operation process and improves the convenience of operation, but also realizes the intelligent generation and optimization of control instructions, providing new ideas and solutions for the intelligent development of CNC machine tools. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 is a flow chart of the present invention;

[0035] Figure 2 It is a unit block diagram of a device for a numerically controlled machine tool controller of the present invention to self-generate operation instructions based on a language large model;

[0036] Figure 3 It is a system schematic diagram of a system in which a numerical control machine tool controller of the present invention self-generates operation instructions based on a large language model. DETAILED DESCRIPTION

[0037] In order to make the purpose and advantages of the present invention more clearly understood, the present invention is specifically described below in conjunction with embodiments. It should be understood that the following text is only used to describe one or several specific embodiments of the present invention, and does not strictly limit the scope of protection of the specific claims of the present invention.

[0038] Embodiment 1:

[0039] A method for a CNC machine tool controller to self-generate operating instructions based on a large language model, which aims to solve the complex problems encountered by operators in the actual operation of CNC machine tools. When operators encounter problems such as unclear process parameter settings, difficult control mode switching, unclear control logic selection, and unfamiliarity with the operating interface of new machine tools when using CNC machine tools, the CNC machine tool controller will activate the intelligent assistance function. By receiving the operator's voice instructions or text input information, the controller can accurately identify operating requirements and problems in the use of the current interface. At the same time, the CNC machine tool controller will obtain the processing status of the machine tool in real time and take a screenshot of the current operating interface to obtain complete display information;

[0040] Next, all the collected information will be input into the embedded multimodal large model for processing and analysis. The multimodal large model has a strong ability to understand and analyze information, and can automatically generate control instructions and operation prompts corresponding to the operation requirements based on the various input information. Finally, these automatically generated operation instructions and interface prompts will be presented on the operation interface to guide the operator to perform accurate operations and complete complex operation tasks. This method not only greatly improves the operating efficiency and intelligence level of CNC machine tools, but also significantly reduces the skill requirements for operators, making machine tool operation simpler, more intuitive and more efficient.

[0041] It includes the following steps:

[0042] S1: Obtain the operation requirements and problems of operators;

[0043] When operators encounter unclear operating situations when using CNC machine tools, such as unclear process parameter settings, difficult control mode switching, unclear control logic selection, or unfamiliarity with the new machine tool operation interface, the system first starts the interactive function of the CNC machine tool controller. At this time, the operator can input requirements and questions in the following two ways:

[0044] S11: Input voice information: The operator expresses the operation requirements through voice, and the voice recognition module of the CNC machine tool controller converts the voice information into text;

[0045] S12: Input text information: The operator can also directly input text on the control interface to describe his or her operation requirements or problems.

[0046] The goal of this step is to ensure that the CNC machine controller accurately captures the operator's intent and question type.

[0047] S2: Initial screening of operational needs and problem areas;

[0048] First, a small language model is established, and the operator's voice information or text information is input into the small language model to determine whether the operation demand or problem belongs to the field of CNC machine tool control. The specific process is as follows:

[0049] S21: Input processing:

[0050] When voice information V is input, the voice information V is converted into text T, and the text T is sent to the small language model for processing;

[0051] When inputting text information, the input text T is directly sent to the small language model for processing;

[0052] S22: Determine the application of the model:

[0053] Use a small language model M s Classify the input text T to determine whether it belongs to the field of CNC machine tool control. The classification output is a binary classification problem:

[0054] y=M s (T)

[0055] Among them, y takes the value of 1 or 0. When y takes the value of 1, it means that it belongs to the field of CNC machine tools, and when y takes the value of 0, it means that it does not belong to the field of CNC machine tools;

[0056] S23: Field judgment result:

[0057] If y=1: ​​it means the operation requirement and problem belong to the field of CNC machine tools, and go to the next step;

[0058] If y=0: the system will give the operator feedback that “the requirement cannot be resolved or answered”.

[0059] This process ensures that the system only provides support for issues related to CNC machine operation, avoiding incorrect operations and waste of resources.

[0060] S3: Setting a CNC machine tool controller and obtaining multi-modal information of the CNC machine tool through the CNC machine tool controller;

[0061] If y=1 is confirmed in S3, and then it is confirmed that the operation requirements and problems belong to the field of CNC machine tools, the CNC machine tool controller obtains the current multi-modal information, which includes the processing state information and operation interface information of the CNC machine tool. The specific process is as follows:

[0062] S31: Real-time status collection:

[0063] The real-time machining status information matrix S of the machine tool is obtained through the CNC machine tool controller, which contains the following contents:

[0064] Current tool position: spatial coordinate information of the tool (x, y, z);

[0065] Processing speed: the speed v of the tool or workpiece;

[0066] Process parameters: such as feed speed, cutting depth, processing temperature, etc.;

[0067] Wherein, the CNC machine tool controller includes a sensor, through which the above information can be obtained;

[0068] For details, please refer to the following table:

[0069]

[0070]

[0071] S32: Operation interface screenshot:

[0072] The CNC machine tool controller takes a screenshot of the current operation interface and stores the screenshot image information as a matrix I t , which contains all parameters, buttons, menus and other information displayed on the current interface.

[0073] This information can provide complete status data for subsequent processing of large multimodal models.

[0074] Among them, the CNC machine tool controller can obtain the operator's operating requirements through voice and text, obtain the processing status of the machine tool in real time, and take screenshots of the operation interface to fully collect multi-modal information. This information collection method ensures that the controller can fully understand the operator's intentions and the current status of the machine tool;

[0075] S4: construct a multimodal large model, and input the multimodal information obtained in S3 (including text input by the operator, machine tool status information, operation interface screenshots, etc.) into the multimodal large model for processing;

[0076] Among them, the multimodal big model is an artificial intelligence model that can process and understand many different types of data input, such as text, images, sounds, etc. This model usually uses deep learning technology to learn the associations and interactions between different modalities through a large amount of data training. The multimodal big model has a wide range of applications in natural language processing, computer vision, speech recognition and other fields, and can provide richer and more accurate information processing capabilities. The big model has been widely and maturely applied in some technical fields, so I will not go into details here.

[0077] The multimodal information obtained in S3 is input into the multimodal large model M m The processing steps are:

[0078] S41: Information input:

[0079] Text information T, machine status information matrix S, interface screenshot matrix I t As input, construct the multimodal information input X vector as:

[0080] X={T,S,I t}

[0081] For details, please refer to the following table:

[0082]

[0083] S42: Model processing: multimodal large model M m After comprehensive analysis of the input information X, the model output is:

[0084] {C,P,R}=M m (X)

[0085] Among them, C is the control instruction, P is the prompt information, and R is the expected result.

[0086] and:

[0087] Control instructions C: Operation instructions generated by the model, used to perform operation tasks.

[0088] Operation step prompt information P: guides the operator on how to perform specific steps on the interface.

[0089] Expected result R: describes the result after each step is completed.

[0090] Moreover, the embedded multimodal large model is the core technology of this method. It can comprehensively analyze different forms of information (such as voice, text, interface screenshots, machine tool status) and generate accurate control instructions. The introduction of this model enables the system to have the ability of self-learning and autonomous decision-making, and can provide the most suitable solution according to operational requirements. Through the analysis and processing of the multimodal large model, the system can automatically generate corresponding control instructions and operation interface prompts. This allows operators to get clear guidance when encountering unclear operational problems, ensuring the accuracy and efficiency of operations;

[0091] S5: Execute control instructions and operation prompts;

[0092] According to the prompts and instructions obtained in S42, actual operations are performed according to the following steps:

[0093] S51: Interface prompt display: the system displays the operation step prompt information P generated by the model on the interface, including the operation method of each step, the button to be clicked, the input parameters, etc.;

[0094] S52: Execute automatically generated instructions: The operator can execute instructions manually according to the prompt or select the system to automatically execute the control instructions C generated by the model;

[0095] S53: Comparison of expected results: The operator can compare the actual execution results with the expected results R given by the model to ensure the accuracy of the operation.

[0096] The purpose of this step is to ensure that operators have clear instructions during complex CNC tasks and avoid operating errors.

[0097] S6: Real-time feedback and dynamic adjustment.

[0098] During operation, the CNC machine tool controller can receive new operation requirements and problems in real time and make dynamic adjustments to ensure flexibility and real-time performance. The steps are as follows:

[0099] S61: Real-time monitoring:

[0100] The system continuously monitors the operator's input information and the processing status of the machine tool to form new information input X';

[0101] S62: Information update and model reprocessing:

[0102] When new information input X' appears, the system will call the multimodal large model M again m , generate new control instructions C', operation prompt information P' and expected results R', and perform the following equation operations:

[0103] {C',P',R'}=M m (X')

[0104] S63: Dynamic Execution:

[0105] The latest instructions and prompts output by the model will be immediately updated on the operation interface, and operators can make adjustments based on the new information.

[0106] Embodiment 2:

[0107] This embodiment is used to propose a device for a numerical control machine tool controller to self-generate operation instructions based on a language large model, including:

[0108] The controller unit is responsible for interacting with the operator and obtaining the operator's operational requirements and questions. It can be a user interface that allows the operator to input instructions in text, voice or other forms. The controller unit is also responsible for preliminary screening of these requirements and questions to determine whether they are clear, complete, and whether further processing is required;

[0109] An information acquisition unit, which is responsible for acquiring multimodal information from the CNC machine tool, including but not limited to machine tool status data, sensor readings, visual images, sound signals, etc. This information can be real-time or pre-stored to provide the current and historical operating environment and status of the machine tool;

[0110] A multimodal large model unit, which uses an advanced language large model to process the multimodal information obtained from the information acquisition unit. It can be a deep learning model, such as Transformer or BERT, which is trained to understand and interpret multimodal data and generate corresponding operation instructions;

[0111] The control instruction and operation prompt execution unit is responsible for executing the control instructions and operation prompts generated by the multi-modal large model unit. It can be a control logic processor that converts the instructions into control signals that the CNC machine tool can understand and guides the operator to perform the next operation;

[0112] Real-time feedback and dynamic adjustment unit, which receives new needs and questions from operators in real time during operation and dynamically adjusts the operation instructions based on this information. It can be a feedback loop system to ensure that the operation instructions are always consistent with the operator's intentions and the real-time status of the machine tool.

[0113] Embodiment 3:

[0114] This embodiment is used to propose a system for a numerical control machine tool controller to self-generate operation instructions based on a language large model, including:

[0115] a processor and a memory for storing instructions executable by the processor;

[0116] Furthermore, the processor may be a central processing unit (CPU) or a graphics processing unit (GPU), which is responsible for executing program instructions and processing data, and the memory is used to store instructions and data executable by the processor, and may be a random access memory (RAM), a read-only memory (ROM) or other forms of storage media;

[0117] The processor is configured to perform the following functions:

[0118] a. Obtain the operational requirements and problems of operators and conduct preliminary screening to ensure the accuracy and operability of the requirements and problems;

[0119] b. Obtain multimodal information of CNC machine tools, including machine tool status, sensor data, visual images, etc., in order to fully understand the working environment and status of the machine tool;

[0120] c. Process the acquired multimodal information, use the language model to analyze and understand the information, and generate corresponding operation instructions;

[0121] d. Execute control instructions and operation prompts, convert the generated instructions into operations that can be performed by the machine tool, and provide necessary operation prompts to the operator;

[0122] e. Receive new operational requirements and issues in real time during the operation process, and make dynamic adjustments to ensure the flexibility and adaptability of operational instructions.

[0123] This method realizes efficient interaction between operators and CNC machine tool controllers through voice and text input, improving the efficiency and intelligence level of human-machine collaboration. This humanized interaction method not only reduces the difficulty of operation, but also makes complex CNC operations more intuitive and convenient.

[0124] Real-time adaptability and guidance: The system can provide highly targeted operation guidance based on the real-time processing status and operation interface display information. This ensures that the system can quickly generate and execute the most appropriate instructions in a variety of complex operation scenarios, thereby improving the efficiency of completing operation tasks;

[0125] This implementation plan gradually obtains the needs of operators, uses small models for field judgment, and adopts multi-modal large models for information analysis and instruction generation, thus realizing intelligent CNC machine tool operation assistance. The entire solution makes full use of advanced technologies such as speech recognition, natural language processing, multi-modal information fusion, and large model reasoning to ensure the efficiency, accuracy, and flexibility of the system in complex operating scenarios.

[0126] The technical advantage of this method is that it can effectively cope with various complex operation scenarios, provide timely and accurate operation guidance and instruction generation, and significantly improve the intelligence level and operation efficiency of CNC machine tools.

[0127] From a technical point of view, this method has the following advantages:

[0128] Multimodal information fusion: Using multiple information sources such as voice, text, machine status, and screenshots of the operation interface, a comprehensive perception of the operator's intention and machine status is achieved. Multimodal information fusion technology can more accurately understand operational requirements, avoid errors that may be caused by a single information source, and ensure the accuracy and reliability of instruction generation.

[0129] Embedded multimodal large model: The embedded multimodal large model processes complex information, achieving efficient operation and real-time response of the model. Compared with traditional methods, the embedded model can operate efficiently within the machine tool controller, reducing dependence on external computing resources, while providing powerful information understanding and decision-making capabilities.

[0130] High adaptability and automation: This method can automatically generate corresponding control instructions and operation prompts based on the real-time collected operation information and machine tool status, realizing highly automated operation guidance. This adaptive capability enables the system to adapt to various complex operation scenarios, reduce human intervention, and improve work efficiency.

[0131] Intelligent human-computer interaction: Through voice and text input, a more natural and efficient interaction method is achieved, which reduces the skill requirements for operators. Compared with the traditional CNC machine tool operation method, this method significantly improves the intelligent level of human-computer interaction and makes the operation more intuitive and simple.

[0132] Real-time response and operation guidance: This method can obtain the processing status and operation interface information of the machine tool in real time, and generate operation prompts and control instructions in time according to the analysis of the multi-modal large model. This real-time performance ensures that the system can respond quickly in complex operation processes, reduces operation waiting time, and improves work efficiency.

[0133] Reduce operation complexity: By automatically generating operation instructions and prompts, it helps operators make correct decisions in complex operation tasks, reduces the requirements for CNC machine tool expertise, and reduces the possibility of operation errors. This not only improves overall operation efficiency, but also provides good technical support for novice operators.

[0134] Improve operational safety and accuracy: The control instructions generated by the multimodal large model after comprehensive analysis of information can accurately guide machine tool operations, reduce the risk of misoperation, and ensure the safety and accuracy of machine tool processing.

[0135] In summary, this method makes full use of advanced technologies such as multimodal information processing, embedded large model technology and intelligent human-computer interaction, and significantly improves the intelligence, automation and safety of CNC machine tool operation.

[0136] The above 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 principles of the present invention, and these improvements and modifications should also be considered as the protection scope of the present invention. The structures, devices and operating methods not specifically described and explained in the present invention shall be implemented according to the conventional means in the art unless otherwise specified and limited.

Claims

1. A method for a CNC machine tool controller to self-generate operation instructions based on a language large model, characterized in that: The steps include: S1: Obtain the operation requirements and problems of operators; S2: Initial screening of operational needs and problem areas; S3: Setting a CNC machine tool controller and obtaining multi-modal information of the CNC machine tool through the CNC machine tool controller; S4: constructing a large multimodal model, and inputting the multimodal information obtained in S3 into the large multimodal model for processing; S5: Execute control instructions and operation prompts; S6: Real-time feedback and dynamic adjustment.

2. The method according to claim 1, characterized in that: In S1, requirements and questions are input in the following two ways: S11: input voice information; S12: Input text information.

3. The method according to claim 2, characterized in that: In S2, a small language model is first established, and voice information or text information is input into the small language model to determine whether the operation requirement or problem belongs to the field of CNC machine tool control. The specific process is: S21: Input processing: When voice information V is input, the voice information V is converted into text T, and the text T is sent to the small language model for processing; When inputting text information, the input text T is directly sent to the small language model for processing; S22: Determine the application of the model: Use a small language model M s Classify the input text T to determine whether it belongs to the field of CNC machine tool control. The classification output is a binary classification problem: y=M s (T) Among them, y takes the value of 1 or 0. When y takes the value of 1, it means that it belongs to the field of CNC machine tools, and when y takes the value of 0, it means that it does not belong to the field of CNC machine tools; S23: Field judgment result: If y=1: ​​it means the operation requirement and problem belong to the field of CNC machine tools, and go to the next step; If y=0: the system will give the operator feedback that "this requirement cannot be resolved or answered".

4. The method according to claim 3, characterized in that: If y=1 is confirmed in S3, and then it is confirmed that the operation requirements and problems belong to the field of CNC machine tools, the CNC machine tool controller obtains the current multi-modal information, which includes the processing state information and operation interface information of the CNC machine tool. The specific process is as follows: S31: Real-time status collection: The real-time machining status information matrix S of the machine tool is obtained through the CNC machine tool controller, which contains the following contents: Current tool position: spatial coordinate information of the tool (x, y, z); Processing speed: the speed v of the tool or workpiece; Process parameters; S32: Screenshot of the operation interface: The CNC machine tool controller takes a screenshot of the current operation interface and stores the screenshot image information as a matrix I t .

5. The method according to claim 4, characterized in that: The S4 can input the multimodal information obtained in the S3 into the multimodal large model M m The processing steps are: S41: Information input: Text information T, machine status information matrix S, interface screenshot matrix I t As input, construct the multimodal information input X vector as: X={T,S,I t } S42: Model processing: multimodal large model M m Perform a comprehensive analysis on the input information X and output the model.

6. The method according to claim 5, characterized in that: The output of the model in S42 is: {C,P,R}=M m (X) Among them, C is the control instruction, P is the prompt information, and R is the expected result.

7. The method according to claim 6, characterized in that: In S5, according to the prompts and instructions obtained in S42, actual operations are performed according to the following steps: S51: Interface prompt display: the system displays the operation step prompt information P generated by the model on the interface; S52: Execute the automatically generated instruction: according to the prompt, execute the instruction or select the control instruction C generated by the system to automatically execute the model; S53: Comparison of expected results: Compare the actual execution results with the expected results R given by the model to ensure the accuracy of the operation.

8. The method according to claim 7, characterized in that: During the operation of S6, the CNC machine tool controller can receive new operation requirements and problems in real time and make dynamic adjustments to ensure flexibility and real-time performance. The steps are as follows: S61: Real-time monitoring: The system continuously monitors the operator's input information and the processing status of the machine tool to form new information input X'; S62: Information update and model reprocessing: When new information input X' appears, the system will call the multimodal large model M again m , generate new control instructions C', operation prompt information P' and expected results R', and perform the following equation operations: {C′,P′,R′}=M m (X′) S63: Dynamic Execution: The latest instructions and prompts output by the model will be updated immediately on the operation interface.

9. A device for a CNC machine tool controller to self-generate operation instructions based on a large language model, characterized in that: include: A controller unit, used to obtain the operator's operational requirements and questions and perform preliminary screening; An information acquisition unit, used for acquiring multimodal information of the CNC machine tool; A multimodal large model unit, used for processing the acquired multimodal information; A control instruction and operation prompt execution unit, used for executing control instructions and operation prompts; The real-time feedback and dynamic adjustment unit is used to receive new operational requirements and problems in real time during the operation process and make dynamic adjustments.

10. A system for a CNC machine tool controller to self-generate operation instructions based on a large language model, characterized in that: Included are: a processor and a memory for storing instructions executable by the processor; The processor is configured to perform the following functions: Obtain the operational requirements and problems of operators and conduct preliminary screening; Obtain multimodal information of CNC machine tools; Processing the acquired multimodal information; Execute control instructions and operation prompts; Receive new operational requirements and issues in real time during the operation process and make dynamic adjustments.

Images