Power equipment operation generation method based on semantic understanding and intention recognition

Abstract

The invention discloses a power equipment operation generation method based on semantic understanding and intention recognition, and relates to the technical field of power equipment operation, and the method comprises the steps: collecting information inputted by a user; performing semantic analysis on the information by using a natural language processing technology, and identifying the intention of the user; generating a corresponding power equipment operation instruction according to the identified intention; and sending the operation instruction to target power equipment so as to execute corresponding operation. The method can effectively improve the management efficiency, operation convenience and safety of the power equipment, is especially suitable for scenes such as a smart power grid, smart power equipment and a household and industrial automation system, and achieves the intelligent management of the power equipment through the fusion of technologies such as a multi-mode input mode, natural language processing, voice recognition and gesture recognition. A more intelligent, efficient and user-friendly solution is provided for the operation of the power equipment, and the safety and reliability of the operation are ensured through intelligent feedback and state monitoring.

Description

Technical Field

[0001] The present invention relates to the technical field of power equipment operation, and in particular to a method for generating power equipment operation based on semantic understanding and intent recognition. Background Art

[0002] With the development of information technology, artificial intelligence and automatic control technology, the management and control of power equipment are gradually moving towards intelligence and automation. Traditional power equipment control methods mainly rely on manual operation, remote control systems and monitoring systems. Although they can realize the basic functions of equipment management, they have certain limitations in terms of operational efficiency, flexibility and user experience.

[0003] At present, the operating system of power equipment usually adopts specific command input methods, such as controlling the equipment through computer interface, remote control or on-site operation. These traditional methods often lack intelligent interaction methods, and cannot efficiently integrate different input channels, and cannot automatically understand user intentions, resulting in complex operations and lack of user friendliness. In recent years, semantic understanding and intent recognition technologies have made significant progress, especially in the fields of natural language processing, speech recognition and computer vision. Through these technologies, intelligent interaction between people and equipment can be better realized, which can not only improve the operating efficiency of power equipment, but also provide users with a more convenient and intuitive operating experience. Summary of the invention

[0004] In order to solve the above technical problems, a method for generating power equipment operations based on semantic understanding and intent recognition is provided. This technical solution solves the above problems.

[0005] In order to achieve the above purpose, the technical solution adopted by the present invention is:

[0006] The method for generating power equipment operation based on semantic understanding and intention recognition includes:

[0007] Collect user input information, including text instructions, voice commands, and gesture operation information;

[0008] Use natural language processing technology to perform semantic analysis on the information and identify the user's intention;

[0009] Generate corresponding power equipment operation instructions according to the recognized intention;

[0010] The operation instruction is sent to the target power device to execute the corresponding operation.

[0011] Preferably, the information collected from the user input, including text instructions, voice commands and gesture operation information, specifically includes:

[0012] Users input text commands through input boxes, mobile applications, smart devices and computer terminals;

[0013] The system receives and records the text input from the user, the text content including the control instructions of the power equipment;

[0014] The system performs preliminary preprocessing on the text, removing irrelevant characters, punctuation marks, and word segmentation;

[0015] The user issues a voice command through a voice input device, and voice recognition technology is used to convert the voice signal into text data;

[0016] The system evaluates the speech quality of the recognized text, handles noise and dialects, and performs unified subsequent processing of the converted text instructions and other types of input information, including gestures and touch information;

[0017] Users operate through gesture input devices. Gestures include pointing, waving, clicking and other touch action commands. The user's gestures are captured and recognized by a convolutional neural network algorithm.

[0018] The system converts the recognized gesture data into standard operation commands. When using a device with a touch screen, the system directly collects the user's touch operation information and maps it to specific operation commands;

[0019] The data from different input devices are uniformly formatted and integrated into standardized instruction data, and the collected multi-modal information is synchronized and matched in time;

[0020] The user confirms the input content. After the user confirms, the system will continue to generate the next operation instruction. If there is no confirmation or the input is wrong, the system prompts the user to re-enter.

[0021] Preferably, the using of natural language processing technology to perform semantic analysis on the information to identify the user's intention specifically includes:

[0022] Convert text into vector representation, using word vector context representation;

[0023] Use classification algorithms to classify the intent of the text and identify the user's intent type;

[0024] Identify important entities in the text, including time, place, and person, to help the system understand the context;

[0025] Integrate context information, generate corresponding system responses based on the recognized intent, and execute corresponding actions.

[0026] Preferably, the identifying of important entities in the text, including time, place and person, and helping the system understand the context specifically includes:

[0027] Split the text into small units, remove stop words, and convert word variants to their base form;

[0028] Extract specific named entities from the text and train the NER model using annotated datasets;

[0029] Entity classification based on the trained NER model;

[0030] Identify and analyze the relationship between entities, and distinguish entities with the same name but different meanings by analyzing the text that appears before and after the entity;

[0031] Analyze sentence structure through syntax tree, identify grammatical relations between entities, and store the identified entities and their relations in the knowledge base;

[0032] Organize the extracted entities and their categories into structured information.

[0033] Preferably, extracting specific named entities from the text and training a NER model using a labeled data set specifically includes:

[0034] Among them, the NER model formula is:

[0035]

[0036] Where P is the conditional probability, y 1 ,y 2 ,…y n is the target sequence, x 1 ,x 2 ,…x n is the input sequence, f k (x i ,y i ) is the characteristic function, γ k is the weight of the feature, f' k (y i ,y i+1 ) is the transition characteristic function, γ” k is the weight associated with the transition feature and Z(x) is the normalization factor.

[0037] Preferably, the fusing context information, generating a corresponding system response according to the recognized intent, and executing a corresponding action specifically include:

[0038] Maintain the state of the conversation, including the user's preferences, known information, and historical conversations, and determine the most relevant information in the current conversation through context weighting technology;

[0039] Adjust the system’s response based on the user’s tone and sentiment, select response templates based on the identified intents and entities, and use generative models to generate natural language responses;

[0040] Adjust the response based on the user's historical behavior preferences. After generating the response, the system verifies it.

[0041] Based on the recognized intents and entities, they are mapped to system operations. After the actions are performed, the results are fed back to the user. The context state is updated based on the current conversation and execution results, and the conversation history is retained.

[0042] Preferably, the generating of corresponding power equipment operation instructions according to the identified intention specifically includes:

[0043] Identify the user's operation intention, identify the device information parameters related to the intention, and map it to specific power equipment operations according to the identified intention, wherein the power equipment operations include switch operations, parameter adjustment, and status query;

[0044] According to the entities in the user input, the type of power equipment involved is identified, and corresponding operation instructions are generated based on the intent;

[0045] According to the identified equipment and operation type, the corresponding command template is selected, the specific parameters in the command are filled in, and the generated instructions are converted into actual operations for controlling the power equipment;

[0046] Through the control system interface with the power equipment, the actual control instructions are sent, the equipment operation is executed, the system returns the operation result and confirms the operation status to the user, and performs status updates and error handling based on the response of the equipment.

[0047] Preferably, the identifying the user's operation intention and identifying the device information parameters related to the intention specifically include:

[0048] Among them, the intent recognition formula is:

[0049]

[0050] In the formula, P(I k |T) is the event I when the observation data T is known. k The conditional probability of occurrence, P(T|I k ) is assumed to be the intention k If it has already happened, the probability of the observed data T appearing, P(I K ) is intention I K The prior probability of itself, P(T) is the total probability of observing data T.

[0051] Preferably, the step of identifying the type of electric equipment involved according to the entity in the user input and generating the corresponding operation instruction based on the intention specifically includes:

[0052] Based on the results of intent recognition, determine the user's operation type, including starting and stopping the device, adjusting parameters, querying device status, switching modes, and confirming operation details;

[0053] Generate corresponding control instructions according to the device type and operation intention, including startup commands, parameter adjustment, status query and protocol conversion, generate corresponding control commands and send them to the device;

[0054] Before sending operating instructions, a security check is performed and a communication channel is established based on the communication protocol and network architecture of the target device, including local control, remote control and instruction sending. The generated and verified operating instructions are sent to the target power equipment via the network, serial port and wirelessly.

[0055] Preferably, sending the operation instruction to the target power device to perform the corresponding operation specifically includes:

[0056] The user inputs the operation instruction through the device control interface, and the system analyzes the operation instruction input by the user and extracts the relevant target device and the specific operation to be performed;

[0057] Before sending an operation instruction, check the current status of the target power equipment, generate a control command based on the type of equipment and the communication protocol, and convert it into a digital signal or protocol packet;

[0058] According to the location and communication mode of the target device, a connection with the power device is established, and the generated control command is sent to the target power device through the communication protocol. After receiving the command, the target power device parses the command through the control system in the device, and the device performs the corresponding operation according to the command;

[0059] The device will provide real-time feedback on the operation results. After the operation is completed, the system will monitor the device status in real time and record the operation log.

[0060] Compared with the prior art, the present invention has the following beneficial effects:

[0061] The power equipment operation generation method based on semantic understanding and intention recognition in the present invention can effectively improve the management efficiency, operation convenience and safety of power equipment, and is particularly suitable for scenarios such as smart grids, smart power equipment, and home and industrial automation systems. Through the integration of multimodal input methods, natural language processing, speech recognition, gesture recognition and other technologies, this method provides a more intelligent, efficient and user-friendly solution for power equipment operation, and ensures the safety and reliability of operation through intelligent feedback and status monitoring. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] Figure 1 It is a step flow framework diagram of the present invention. DETAILED DESCRIPTION

[0063] The following description is used to disclose the present invention so that those skilled in the art can implement the present invention. The preferred embodiments described below are only examples, and those skilled in the art may think of other obvious variations.

[0064] Reference Figure 1 As shown, the method for generating power equipment operation based on semantic understanding and intention recognition includes:

[0065] Step 1:

[0066] Users input text commands through input boxes, mobile applications, smart devices and computer terminals;

[0067] The system receives and records the text input from the user, the text content including the control instructions of the power equipment;

[0068] The system performs preliminary preprocessing on the text, removing irrelevant characters, punctuation marks, and word segmentation;

[0069] The user issues a voice command through a voice input device, and voice recognition technology is used to convert the voice signal into text data;

[0070] The system evaluates the speech quality of the recognized text, handles noise and dialects, and performs unified subsequent processing of the converted text instructions and other types of input information, including gestures and touch information;

[0071] Users operate through gesture input devices. Gestures include pointing, waving, clicking and other touch action commands. The user's gestures are captured and recognized by a convolutional neural network algorithm.

[0072] The system converts the recognized gesture data into standard operation commands. When using a device with a touch screen, the system directly collects the user's touch operation information and maps it to specific operation commands;

[0073] The data from different input devices are uniformly formatted and integrated into standardized instruction data, and the collected multi-modal information is synchronized and matched in time;

[0074] The user confirms the input content. After the user confirms, the system will continue to generate the next operation instruction. If there is no confirmation or the input is wrong, the system prompts the user to re-enter.

[0075] Step 2:

[0076] Convert text into vector representation, using word vector context representation;

[0077] Use classification algorithms to classify the intent of the text and identify the user's intent type;

[0078] Split the text into small units, remove stop words, and convert word variants to their base form;

[0079] Extract specific named entities from the text and train the NER model using annotated datasets;

[0080] Among them, the NER model formula is:

[0081]

[0082] Where P is the conditional probability, y 1 ,y 2 ,…y n is the target sequence, x 1 ,x 2 ,…x n is the input sequence, f k (x i ,y i ) is the characteristic function, γ k is the weight of the feature, f' k (y i ,y i+1 ) is the transition characteristic function, γ” k is the weight associated with the transition feature, and Z(x) is the normalization factor;

[0083] Entity classification based on the trained NER model;

[0084] Identify and analyze the relationship between entities, and distinguish entities with the same name but different meanings by analyzing the text that appears before and after the entity;

[0085] Analyze sentence structure through syntax tree, identify grammatical relations between entities, and store the identified entities and their relations in the knowledge base;

[0086] Organize the extracted entities and their categories into structured information;

[0087] Maintain the state of the conversation, including the user's preferences, known information, and historical conversations, and determine the most relevant information in the current conversation through context weighting technology;

[0088] Adjust the system’s response based on the user’s tone and sentiment, select response templates based on the identified intents and entities, and use generative models to generate natural language responses;

[0089] Adjust the response based on the user's historical behavior preferences. After generating the response, the system verifies it.

[0090] Based on the recognized intents and entities, they are mapped to system operations. After the actions are performed, the results are fed back to the user. The context state is updated based on the current conversation and execution results, and the conversation history is retained.

[0091] Step 3:

[0092] Identify the user's operation intention, identify the device information parameters related to the intention, and map it to specific power equipment operations according to the identified intention, wherein the power equipment operations include switch operations, parameter adjustment, and status query;

[0093] Among them, the intent recognition formula is:

[0094]

[0095] In the formula, P(I k |T) is the event I when the observation data T is known. k The conditional probability of occurrence, P(T|I k ) is the assumption that intention I k If it has already happened, the probability of the observed data T appearing, P(I K ) is intention I K The prior probability of itself, P(T) is the total probability of observing data T;

[0096] According to the entities in the user input, the type of power equipment involved is identified, and corresponding operation instructions are generated based on the intent;

[0097] Based on the results of intent recognition, determine the user's operation type, including starting and stopping the device, adjusting parameters, querying device status, switching modes, and confirming operation details;

[0098] Generate corresponding control instructions according to the device type and operation intention, including startup commands, parameter adjustment, status query and protocol conversion, generate corresponding control commands and send them to the device;

[0099] Before sending an operation instruction, a security check is performed, and a communication channel is established according to the communication protocol and network architecture of the target device, including local control, remote control and instruction sending, and the generated and verified operation instruction is sent to the target power device through the network, serial port and wireless mode;

[0100] According to the identified equipment and operation type, the corresponding command template is selected, the specific parameters in the command are filled in, and the generated instructions are converted into actual operations for controlling the power equipment;

[0101] Through the control system interface with the power equipment, the actual control instructions are sent, the equipment operation is executed, the system returns the operation result and confirms the operation status to the user, and performs status updates and error handling based on the response of the equipment.

[0102] Step 4:

[0103] The user inputs the operation instruction through the device control interface, and the system analyzes the operation instruction input by the user and extracts the relevant target device and the specific operation to be performed;

[0104] Before sending an operation instruction, check the current status of the target power equipment, generate a control command based on the type of equipment and the communication protocol, and convert it into a digital signal or protocol packet;

[0105] According to the location and communication mode of the target device, a connection with the power device is established, and the generated control command is sent to the target power device through the communication protocol. After receiving the command, the target power device parses the command through the control system in the device, and the device performs the corresponding operation according to the command;

[0106] The device will provide real-time feedback on the operation results. After the operation is completed, the system will monitor the device status in real time and record the operation log.

[0107] In summary, the advantages of the present invention are:

[0108] The power equipment operation generation method based on semantic understanding and intent recognition can effectively integrate multiple input methods to provide users with a more natural and convenient interaction method. This multimodal interaction method not only improves the user experience, but also can flexibly select the appropriate input method according to different scenarios and needs, thereby enhancing the intelligent level of operation.

[0109] The system uses natural language processing technology to perform semantic analysis on user input, accurately identify user intent, and then generate operation instructions that meet actual needs. Compared with traditional manual input and remote control operation, this automated instruction generation method can greatly improve operation efficiency and accuracy and reduce human errors.

[0110] By integrating advanced technologies such as voice recognition and gesture recognition, the system can understand the user's natural language or gesture commands. Users can directly issue commands without being familiar with complex control interfaces or operating procedures. This simplified operation method reduces the complexity of system operation, which is especially convenient for non-professionals and special scenarios.

[0111] By effectively processing voice quality, noise, dialects, etc., the system can adapt to different environments and user habits to ensure the accuracy of command recognition. No matter what voice input device or gesture device the user uses, the system can respond in real time to ensure the accurate communication and execution of power equipment operation instructions;

[0112] Before generating an operation instruction, the system checks and verifies the status of the device to ensure the safety of the operation. For example, before performing a critical operation such as shutting down a device, the system will check whether the device is currently in normal operation and generate and transmit the operation instruction according to the device's communication protocol. Such a safety check mechanism can effectively avoid misoperation or unsafe operation.

[0113] After the operation instruction is sent, the system can monitor the execution status of the device in real time and feedback the results to the user after the operation is completed. If the device responds abnormally or the operation fails, the system will promptly prompt the user and handle the error to ensure the smooth operation of the device. Through this dynamic feedback mechanism, the user can clearly understand the real-time status of the device, further improving the reliability of the operation;

[0114] This method can adapt to different types of power equipment and generate corresponding control instructions according to the equipment type. Whether it is switch operation, parameter adjustment, or equipment status query, the system can intelligently identify the needs of the equipment and generate corresponding operation commands to ensure the accuracy of the instructions and the compatibility of the equipment.

[0115] Through the application of semantic understanding and intent recognition technology, the system can realize intelligent management of power equipment, which not only improves the efficiency of operation, but also enhances the automation level of the power system. Users can manage equipment in a simple and intelligent way, reducing dependence on manual operation and improving the overall intelligence and automation level of the power system.

[0116] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The above embodiments and descriptions only describe the principles of the present invention. The present invention may be subject to various changes and improvements without departing from the spirit and scope of the present invention. These changes and improvements fall within the scope of the present invention. The scope of protection claimed by the present invention is defined by the attached claims and their equivalents.

Claims

1. A method for generating power equipment operations based on semantic understanding and intention recognition, characterized in that: include: Collect user input information, including text instructions, voice commands, and gesture operation information; Use natural language processing technology to perform semantic analysis on the information and identify the user's intention; Generate corresponding power equipment operation instructions according to the recognized intention; The operation instruction is sent to the target power device to execute the corresponding operation.

2. The method for generating power equipment operation based on semantic understanding and intention recognition according to claim 1, characterized in that: The information collected from the user input, including text instructions, voice commands and gesture operation information, specifically includes: Users input text commands through input boxes, mobile applications, smart devices and computer terminals; The system receives and records the text input from the user, the text content including the control instructions of the power equipment; The system performs preliminary preprocessing on the text, removing irrelevant characters, punctuation marks, and word segmentation; The user issues a voice command through a voice input device, and voice recognition technology is used to convert the voice signal into text data; The system evaluates the speech quality of the recognized text, handles noise and dialects, and performs unified subsequent processing of the converted text instructions and other types of input information, including gestures and touch information; The user operates through a gesture input device; The system converts the recognized gesture data into standard operation commands. When using a device with a touch screen, the system directly collects the user's touch operation information and maps it to specific operation commands; Format the data from different input devices uniformly, integrate them into standardized command data, and synchronize and match the collected multimodal information in time; The user confirms the input content. After the user confirms, the system will continue to generate the next operation instruction. If there is no confirmation or the input is wrong, the system prompts the user to re-enter.

3. The method for generating power equipment operation based on semantic understanding and intention recognition according to claim 2, characterized in that: The using of natural language processing technology to perform semantic analysis on the information and identify the user's intention specifically includes: Convert text into vector representation, using word vector context representation; Use classification algorithms to classify the text and identify the user's intent type; Identify important entities in the text, including time, place, and person, to help the system understand the context; Integrate context information, generate corresponding system responses based on the recognized intent, and execute corresponding actions.

4. The method for generating power equipment operation based on semantic understanding and intention recognition according to claim 3 is characterized in that: The identification of important entities in the text, including time, place and person, helps the system understand the context, including: Split the text into small units, remove stop words, and convert word variants to their base form; Extract specific named entities from the text and train the NER model using annotated datasets; Entity classification based on the trained NER model; Identify and analyze the relationship between entities, and distinguish entities with the same name but different meanings by analyzing the text that appears before and after the entity; Analyze sentence structure through syntax tree, identify grammatical relationships between entities, and store the identified entities and their relationships in the knowledge base; Organize the extracted entities and their categories into structured information.

5. The method for generating power equipment operation based on semantic understanding and intention recognition according to claim 4, characterized in that: The specific named entities are extracted from the text and the NER model is trained using the labeled data set. include: Among them, the NER model formula is: Where P is the conditional probability, y1,y2,…y n is the target sequence, x1,x2,…x n is the input sequence, f k (x i ,y i ) is the characteristic function, γ k is the weight of the feature, f , k (y i ,y i+1 ) is the transition characteristic function, γ ,, k is the weight associated with the transition feature and Z(x) is the normalization factor.

6. The method for generating power equipment operation based on semantic understanding and intention recognition according to claim 5, characterized in that: The fusion of context information, generating a corresponding system response according to the recognized intent, and executing the corresponding action specifically include: Maintaining state in a session; Adjust the system’s response based on the user’s tone and sentiment, select response templates based on the identified intents and entities, and use generative models to generate natural language responses; Adjust the response based on the user's historical behavior preferences. After generating the response, the system verifies it. Based on the recognized intents and entities, they are mapped to system operations. After the actions are performed, the results are fed back to the user. The context state is updated based on the current conversation and execution results, and the conversation history is retained.

7. The method for generating power equipment operation based on semantic understanding and intention recognition according to claim 6, characterized in that: The generating of corresponding power equipment operation instructions according to the identified intention specifically includes: Identify the user's operation intention, identify the equipment information parameters related to the intention, and map it to specific power equipment operations based on the identified intention; According to the entities in the user input, the type of power equipment involved is identified, and corresponding operation instructions are generated based on the intent; According to the identified equipment and operation type, the corresponding command template is selected, the specific parameters in the command are filled in, and the generated instructions are converted into actual operations for controlling the power equipment; Through the control system interface with the power equipment, the actual control instructions are sent, the equipment operation is executed, the system returns the operation result and confirms the operation status to the user, and performs status updates and error handling based on the response of the equipment.

8. The method for generating electric power equipment operation based on semantic understanding and intention recognition according to claim 7, characterized in that: The identification of the user's operation intention and the identification of the device information parameters related to the intention are specific include: Among them, the intent recognition formula is: In the formula, P(I k |T) is the event I when the observation data T is known. k The conditional probability of occurrence, P(T|I k ) is the assumption that intention I k If it has already happened, the probability of the observed data T appearing, P(I K ) is intention I K The prior probability of itself, P(T) is the total probability of observing data T.

9. The method for generating power equipment operation based on semantic understanding and intention recognition according to claim 8, characterized in that: The step of identifying the type of electric power equipment involved according to the entity in the user input and generating the corresponding operation instruction based on the intention specifically includes: Based on the results of intent recognition, determine the user's operation type, including starting and stopping the device, adjusting parameters, querying device status, switching modes, and confirming operation details; Generate corresponding control instructions according to the device type and operation intention, including startup commands, parameter adjustment, status query and protocol conversion, generate corresponding control commands and send them to the device; Before sending an operation instruction, a security check is performed and a communication channel is established based on the communication protocol and network architecture of the target device.

10. The method for generating power equipment operation based on semantic understanding and intention recognition according to claim 9, characterized in that: The sending of the operation instruction to the target power device to perform the corresponding operation specifically includes: The user inputs the operation instruction through the device control interface, and the system analyzes the operation instruction input by the user and extracts the relevant target device and the specific operation to be performed; Before sending an operation instruction, check the current status of the target power equipment, generate a control command based on the type of equipment and the communication protocol, and convert it into a digital signal or protocol packet; According to the location and communication mode of the target device, a connection with the power device is established, and the generated control command is sent to the target power device through the communication protocol. After the target power device receives the command, the control system in the device parses the command, and the device performs the corresponding operation according to the command; The device will provide real-time feedback on the operation results. After the operation is completed, the system will monitor the device status in real time and record the operation log.

Images