Central control intention distribution method and system based on large model agent, and smart television

Abstract

The invention discloses a central control intention distribution method and system based on a large model agent and a smart television, and the method comprises the steps: collecting the voice audio of a user after awakening, and carrying out the recognition to obtain a voice recognition result; the voice recognition result is input into a large model agent center controller, text intention information of the user is recognized, the text intention information is sent to the agent which the user belongs to, and the agent answers the intention of the user and stores the dialogue information of the round in combination with the text intention information and the stored context; and when the user initiates a second round of dialogue, obtaining a second intention of the user, if the second intention does not belong to the intelligent agent, inputting the second intention into the large-model intelligent agent center controller again, identifying the second intention again, and distributing the second intention to the corresponding intelligent agent. According to the method, the functions of understanding complex instructions of the user and correctly distributing intentions to the corresponding intelligent agents are realized, the intelligent television dynamically switches different intelligent agents, and diversified requirements of the user are ensured to be timely and accurately responded.

Description

Technical Field

[0001] The present invention relates to the field of information processing technology, and in particular to a control intention distribution method and system based on a large model intelligent body, a smart television and a computer-readable storage medium. Background Art

[0002] Traditional large-model agents, which use artificial intelligence technology to generate natural language text or images in a specific field, need to enter a specific agent for independent use. After entering the agent, all intentions are defaulted to be related to the agent in this field. The intention is the task that the user wants the system to understand and execute. However, when the user wants to obtain information outside the field of the agent, he needs to exit the current agent and then enter other agents to achieve it.

[0003] Therefore, the existing technology cannot switch to other intelligent agents at will according to global needs, and the experience is not flexible enough. For example, if the traditional large model is already in the film and television intelligent agent, it cannot recognize and execute the operation of jumping from the film and television intelligent agent to the music intelligent agent. After switching the intelligent agent, the historical conversation is cleared, and the new intelligent agent cannot generate natural language text based on the conversation content of the previous intelligent agent. The user is required to express his intention again, which reduces the user experience.

[0004] Therefore, the prior art still needs to be improved and developed. Summary of the invention

[0005] The main purpose of the present invention is to provide a method, system, smart TV and computer-readable storage medium for distributing control intentions based on a large model intelligent agent, aiming to solve the problem that the existing technology cannot switch to other intelligent agents at will according to global needs, and after switching the intelligent agent, the historical conversation is cleared, and the user needs to express the intention again, which reduces the user experience.

[0006] To achieve the above object, the present invention provides a method for distributing control intentions in a large model intelligent agent, and the method for distributing control intentions in a large model intelligent agent comprises the following steps:

[0007] After the user is awakened by the wake-up word, the user's voice audio is collected, and the voice audio is recognized to obtain a voice recognition result;

[0008] The speech recognition result is input into the large model agent central control, the user's text intention information is identified, and the text intention information is sent to the agent to which it belongs. The agent combines the text intention information with the context stored in the agent, answers the user's intention, and saves the information of this round of dialogue;

[0009] When the user initiates a second round of conversation, the second intention of the user is obtained. If the second intention does not belong to the agent, the second intention is re-input into the large model agent central control, and the second intention is recognized again and distributed to the corresponding agent.

[0010] Optionally, the control intention distribution method based on a large model intelligent agent, wherein, after being awakened by the user's wake-up word, collecting the user's voice audio, recognizing the voice audio, and obtaining a voice recognition result, specifically includes:

[0011] Collect the wake-up word issued by the user, and when the wake-up word is verified, enter the device wake-up state and wait for the user's voice input;

[0012] The user's voice audio is collected, the voice audio is transmitted to a voice recognition service, and the voice audio is recognized by the voice recognition service to obtain a voice recognition result.

[0013] Optionally, the method for distributing intentions based on the large model agent central control, wherein the speech recognition result is input into the large model agent central control, the user's text intention information is recognized, the text intention information is sent to the agent to which it belongs, the agent combines the text intention information and the context stored in the agent, answers the user's intention, and saves the current round of dialogue information, specifically including:

[0014] The speech recognition result is input into the large model agent central control, and the large model agent central control recognizes the user's first text intention information according to the speech recognition result, and recognizes the first agent to which the first text intention information belongs, and sends the first text intention information to the first agent;

[0015] The first agent combines the first text intention information and the context stored in the first agent to answer the first intention of the user, and saves the current round of dialogue information in the first agent.

[0016] Optionally, the method for distributing intentions based on the large model intelligent agent central control, wherein when the user initiates a second round of dialogue, the second intention of the user is obtained, and if the second intention does not belong to the intelligent agent, the second intention is re-input into the large model intelligent agent central control, and the second intention is recognized again and distributed to the corresponding intelligent agent, specifically including:

[0017] When the user initiates a second round of dialogue, second text intention information of the user is obtained, and the first agent determines whether the second text intention information belongs to the first agent;

[0018] If the second text intention information does not belong to the first agent, the second text intention information of the second round of dialogue is re-input into the large model agent central control, and the large model agent central control recognizes the second intention again and distributes it to the corresponding second agent.

[0019] Optionally, the control intention distribution method based on a large model agent, wherein after the first agent determines whether the second text intention information belongs to the first agent, the method further comprises:

[0020] If the second textual intention information belongs to the first agent, the first agent again answers the second intention of the user and saves the second round of dialogue information.

[0021] Optionally, in the method for distributing control intentions based on a large model intelligent agent, the textual intention information includes: the name of the intention and user voice description information.

[0022] Optionally, in the method for distributing control intentions based on a large model intelligent agent, the user voice description information includes: people, objects, colors, objects and locations.

[0023] In addition, to achieve the above-mentioned purpose, the present invention also provides a control intention distribution system based on a large model intelligent agent, wherein the control intention distribution system based on a large model intelligent agent includes:

[0024] A speech recognition module is used to collect the user's voice audio after being awakened by the user's wake-up word, recognize the voice audio, and obtain a speech recognition result;

[0025] The intention control module is used to input the speech recognition result into the large model agent central control, identify the user's text intention information, and send the text intention information to the agent to which it belongs. The agent combines the text intention information with the context stored in the agent to answer the user's intention and save the information of this round of dialogue;

[0026] The intention distribution module is used to obtain the second intention of the user when the user initiates the second round of dialogue. If the second intention does not belong to the agent, the second intention is re-input into the large model agent central control, and the second intention is recognized again and distributed to the corresponding agent.

[0027] In addition, to achieve the above-mentioned purpose, the present invention also provides a smart TV, wherein the smart TV includes: a memory, a processor, and a large model intelligent agent-based control intention distribution program stored in the memory and executable on the processor, and when the large model intelligent agent-based control intention distribution program is executed by the processor, the steps of the large model intelligent agent-based control intention distribution method as described above are implemented.

[0028] In addition, to achieve the above-mentioned purpose, the present invention also provides a computer-readable storage medium, wherein the computer-readable storage medium stores a control intention distribution program based on a large model intelligent agent, and when the control intention distribution program based on a large model intelligent agent is executed by a processor, the steps of the control intention distribution method based on a large model intelligent agent as described above are implemented.

[0029] In the present invention, after being awakened by the user's wake-up word, the user's voice audio is collected, the voice audio is recognized, and a voice recognition result is obtained; the voice recognition result is input into the large model intelligent agent central control, the user's text intention information is recognized, and the text intention information is sent to the corresponding intelligent agent, the intelligent agent combines the text intention information and the context stored in the intelligent agent, answers the user's intention, and saves the information of this round of dialogue; when the user initiates a second round of dialogue, the user's second intention is obtained. If the second intention does not belong to the intelligent agent, the second intention is re-input into the large model intelligent agent central control, and the second intention is recognized again and distributed to the corresponding intelligent agent. According to the development needs of smart TVs, the present invention realizes the function of understanding user complex instructions and correctly distributing intentions to corresponding intelligent agents, realizing efficient and intelligent user experience, and smart TVs can dynamically switch between different intelligent agents to ensure that users' diverse needs are responded to in a timely and accurate manner, further improving the convenience and intelligence level of smart TVs. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 It is a flow chart of a preferred embodiment of the control intention distribution method based on a large model intelligent body of the present invention;

[0031] Figure 2 It is a flow chart of the distribution process when there are multiple dialogues in a preferred embodiment of the control intention distribution method based on a large model intelligent agent of the present invention;

[0032] Figure 3 It is a structural diagram of a preferred embodiment of the control intention distribution system based on a large model intelligent body of the present invention;

[0033] Figure 4 It is a structural diagram of a preferred embodiment of the smart TV of the present invention. DETAILED DESCRIPTION

[0034] In order to make the purpose, technical solution and advantages of the present invention clearer and more specific, the present invention is further described in detail below with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0035] The control intention distribution method based on the large model intelligent agent described in the preferred embodiment of the present invention is as follows: Figure 1 and Figure 2 As shown, the control intention distribution method based on the large model intelligent agent includes the following steps:

[0036] Step S10: After being awakened by the user's wake-up word, the user's voice audio is collected, and the voice audio is recognized to obtain a voice recognition result.

[0037] Specifically, the smart TV collects the wake-up word (for example, "Xiao Wei Xiao Wei") issued by the user, and when the wake-up word is verified, the device enters the wake-up state (that is, the smart TV enters the wake-up state), and the smart TV waits for the user's voice input (that is, the user then initiates a voice conversation); the smart TV collects the user's voice audio, and the smart TV transmits the voice audio to a speech recognition service (ASR, Automatic Speech Recognition), and recognizes the voice audio through the speech recognition service ASR to obtain a speech recognition result (that is, the text corresponding to the voice).

[0038] Step S20: input the speech recognition result into the large model agent central control, identify the user's text intention information, send the text intention information to the agent to which it belongs, and the agent combines the text intention information and the context stored in the agent to answer the user's intention and save the information of this round of dialogue.

[0039] Specifically, the smart TV inputs the voice recognition result into the large model intelligent agent central control, and the large model intelligent agent central control identifies the user's first text intention information according to the voice recognition result, the first text intention information includes: the name of the intention and the user voice description information, the user voice description information includes: people, objects, colors, objects and positions and other information, and identifies the first intelligent agent to which the first text intention information belongs, and sends the first text intention information to the first intelligent agent; the first intelligent agent combines the first text intention information and the context saved in the first intelligent agent, answers the user's first intention, and saves the current round of dialogue information in the first intelligent agent.

[0040] Step S30: When the user initiates a second round of conversation, the second intention of the user is obtained. If the second intention does not belong to the agent, the second intention is re-input into the large model agent central control, and the second intention is recognized again and distributed to the corresponding agent.

[0041] Specifically, at this time, the user is in the first agent, and when the user initiates a second round of dialogue (or multiple rounds of dialogue), the second text intention information of the user is obtained, and the second text intention information includes: the name of the intention and the user voice description information, and the user voice description information includes: information such as people, objects, colors, objects and locations; the first agent determines whether the second text intention information belongs to the first agent; if the second text intention information belongs to the first agent, the first agent again answers the second intention of the user and saves the second round of dialogue information; if the second text intention information does not belong to the first agent, the second text intention information of the second round of dialogue is re-input into the large model agent central control, and the large model agent central control recognizes the second intention again and distributes it to the corresponding second agent.

[0042] According to the development needs of existing smart TVs, the present invention realizes the function of understanding complex user instructions and correctly distributing intentions to corresponding intelligent agents, thus achieving efficient and intelligent user experience. Through this technology, smart TVs can dynamically switch between different intelligent agents, ensuring that users' diverse needs are responded to in a timely and accurate manner, further improving the convenience and intelligence level of smart TVs. The key technology is to solve the problem of random switching of intelligent agents, and when switching between intelligent agents, the large model can track the context and realize the continued dialogue of a certain intelligent agent.

[0043] For example, taking the existing smart TV as an example, there is the following scenario: when the user says "I want to watch a movie", the traditional big model will recommend movies to the user, and the user says "I want to order a takeaway". At this time, the traditional intelligent body stays in the film and television intelligent body and cannot jump to the takeaway intelligent body. The user needs to exit the current intelligent body and then say the need to order a takeaway, and the previous conversation record will be cleared. The traditional big model cannot track the context to achieve the continuity of the conversation. At this time, the intention distribution method based on the big model intelligent body central control of the present invention can realize that after the user says to watch a movie and then says the command to order a takeaway, the big model will directly ask the user for the takeaway demand, and jump from the film and television intelligent body to the takeaway intelligent body. After the user orders the takeaway, when he says the command "I want to watch a movie in 2024", the intelligent body central control will recognize the user's intention to watch a movie, and will switch back to the film and television intelligent body at this time, and track the previous conversation, and recognize that the complete command is "a movie in 2024". The arbitrary switching of intelligent bodies and tracking of contexts are realized.

[0044] First, the user wakes up the device through the wake-up word. Take the smart TV as an example: Xiaowei Xiaowei. The device receives the wake-up signal and waits for the user's voice input. Then the user initiates a voice conversation. The terminal collects the user's audio and transmits it to the voice recognition service ASR, recognizes the text corresponding to the voice, and then inputs the text into the large model intelligent agent central control to identify which intelligent agent the intent belongs to and distribute the intent to the module of the intelligent agent. Each intelligent agent is responsible for processing a certain type of intent, so after the central control recognizes the text intent, it distributes the intent to the corresponding intelligent agent. The intent information includes the name of the intent and the person, object, color, object, location and other information described by the user's voice. The intelligent agent will combine the above information and the context saved in the intelligent agent to answer the intent and save the information of this round of conversation. At this time, the user is in a certain intelligent agent. When the user initiates the second round of conversation, the intelligent agent will first determine whether the intent belongs to the intelligent agent. If it does, it will answer the intent again and save the information of this round of conversation. If it does not belong, the text of this round of conversation will be re-entered into the large model intelligent agent central control, and the central control will recognize the intent again and distribute it to the corresponding intelligent agent. Therefore, the present invention realizes the integration of multiple intelligent agents and enables arbitrary switching, making the user experience more flexible and intelligent.

[0045] Through the intention understanding ability of the big model, compared with the traditional big model intelligent agent, the present invention enables the intelligent agent to have the logic of judging exit, and the intelligent agent can be switched at will without human intervention. The context recorded independently by each intelligent agent enables the big model to track the conversation records, making the conversation management clearer and the product experience more intelligent.

[0046] Furthermore, if Figure 3 As shown, based on the above-mentioned control intention distribution method based on a large model intelligent agent, the present invention also provides a control intention distribution system based on a large model intelligent agent, wherein the control intention distribution system based on a large model intelligent agent includes:

[0047] A speech recognition module 51 is used to collect the user's voice audio after being awakened by the user's wake-up word, recognize the voice audio, and obtain a speech recognition result;

[0048] The intention control module 52 is used to input the speech recognition result into the large model agent central control, identify the user's text intention information, and send the text intention information to the agent to which it belongs. The agent combines the text intention information and the context stored in the agent to answer the user's intention and save the information of this round of dialogue;

[0049] The intention distribution module 53 is used to obtain the second intention of the user when the user initiates the second round of dialogue. If the second intention does not belong to the agent, the second intention is re-input into the large model agent central control, and the second intention is recognized again and distributed to the corresponding agent.

[0050] Furthermore, if Figure 4 As shown, based on the above-mentioned large model intelligent agent-based control intention distribution method and system, the present invention also provides a smart TV accordingly, and the smart TV includes a processor 10, a memory 20 and a display 30. Figure 4 Only some components of the smart TV are shown, but it should be understood that it is not required to implement all the components shown, and more or fewer components may be implemented instead.

[0051] In some embodiments, the memory 20 may be an internal storage unit of the smart TV, such as a hard disk or memory of the smart TV. In other embodiments, the memory 20 may also be an external storage device of the smart TV, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), etc. equipped on the smart TV. Further, the memory 20 may also include both an internal storage unit of the smart TV and an external storage device. The memory 20 is used to store application software and various types of data installed on the smart TV, such as the program code of the smart TV installed. The memory 20 may also be used to temporarily store data that has been output or is to be output. In one embodiment, a control intention distribution program 40 based on a large model intelligent agent is stored on the memory 20, and the control intention distribution program 40 based on a large model intelligent agent can be executed by the processor 10, thereby realizing the control intention distribution method based on a large model intelligent agent in the present application.

[0052] In some embodiments, the processor 10 can be a central processing unit (CPU), a microprocessor or other data processing chip, used to run the program code or process data stored in the memory 20, such as executing the control intention distribution method based on the large model intelligent agent.

[0053] In some embodiments, the display 30 may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, an OLED (Organic Light-Emitting Diode) touch device, etc. The display 30 is used to display information on the smart TV and to display a visual user interface. The processor 10, the memory 20, and the display 30 of the smart TV communicate with each other through a system bus.

[0054] In one embodiment, when the processor 10 executes the control intention distribution program 40 based on the large model agent in the memory 20, the following steps are implemented:

[0055] After the user is awakened by the wake-up word, the user's voice audio is collected, and the voice audio is recognized to obtain a voice recognition result;

[0056] The speech recognition result is input into the large model agent central control, the user's text intention information is identified, and the text intention information is sent to the agent to which it belongs. The agent combines the text intention information with the context stored in the agent, answers the user's intention, and saves the information of this round of dialogue;

[0057] When the user initiates a second round of conversation, the second intention of the user is obtained. If the second intention does not belong to the agent, the second intention is re-input into the large model agent central control, and the second intention is recognized again and distributed to the corresponding agent.

[0058] Among them, when waking up according to the user's wake-up word, collecting the user's voice audio, recognizing the voice audio, and obtaining a voice recognition result specifically includes:

[0059] Collect the wake-up word issued by the user, and when the wake-up word is verified, enter the device wake-up state and wait for the user's voice input;

[0060] The user's voice audio is collected, the voice audio is transmitted to a voice recognition service, and the voice audio is recognized by the voice recognition service to obtain a voice recognition result.

[0061] The process of inputting the speech recognition result into the large model agent central control, identifying the user's text intention information, sending the text intention information to the agent to which it belongs, and the agent combining the text intention information with the context stored in the agent to answer the user's intention and save the current round of dialogue information specifically includes:

[0062] The speech recognition result is input into the large model agent central control, and the large model agent central control recognizes the user's first text intention information according to the speech recognition result, and recognizes the first agent to which the first text intention information belongs, and sends the first text intention information to the first agent;

[0063] The first agent combines the first text intention information and the context stored in the first agent to answer the first intention of the user, and saves the current round of dialogue information in the first agent.

[0064] When the user initiates a second round of dialogue, the second intention of the user is obtained. If the second intention does not belong to the agent, the second intention is re-input into the large model agent central control, and the second intention is recognized again and distributed to the corresponding agent, which specifically includes:

[0065] When the user initiates a second round of dialogue, second text intention information of the user is obtained, and the first agent determines whether the second text intention information belongs to the first agent;

[0066] If the second text intention information does not belong to the first agent, the second text intention information of the second round of dialogue is re-input into the large model agent central control, and the large model agent central control recognizes the second intention again and distributes it to the corresponding second agent.

[0067] After the first agent determines whether the second text intention information belongs to the first agent, the following further comprises:

[0068] If the second textual intention information belongs to the first agent, the first agent again answers the second intention of the user and saves the second round of dialogue information.

[0069] The text intent information includes: the name of the intent and user voice description information.

[0070] Optionally, in the method for distributing control intentions based on a large model intelligent agent, the user voice description information includes: people, objects, colors, objects and locations.

[0071] The present invention also provides a computer-readable storage medium, wherein the computer-readable storage medium stores a program for distributing control intentions based on a large model intelligent agent, and when the program for distributing control intentions based on a large model intelligent agent is executed by a processor, the steps of the method for distributing control intentions based on a large model intelligent agent as described above are implemented.

[0072] In summary, the present invention provides a method, system, smart TV and computer-readable storage medium for distributing intentions based on a large model intelligent agent central control. The method includes: after being awakened according to the user's wake-up word, collecting the user's voice audio, recognizing the voice audio, and obtaining a voice recognition result; inputting the voice recognition result into the large model intelligent agent central control, recognizing the user's text intention information, and sending the text intention information to the intelligent agent to which it belongs. The intelligent agent combines the text intention information and the context stored in the intelligent agent to answer the user's intention and save the current round of dialogue information; when the user initiates a second round of dialogue, the user's second intention is obtained. If the second intention does not belong to the intelligent agent, the second intention is re-input into the large model intelligent agent central control, and the second intention is recognized again and distributed to the corresponding intelligent agent. According to the development needs of smart TVs, the present invention realizes the function of understanding user complex instructions and correctly distributing intentions to corresponding intelligent agents, realizing efficient and intelligent user experience, and the smart TV can dynamically switch between different intelligent agents to ensure that the user's diverse needs are responded to in a timely and accurate manner, further improving the convenience and intelligence level of smart TVs.

[0073] It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or smart TV including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or smart TV. In the absence of further restrictions, an element defined by the sentence "includes a ..." does not exclude the existence of other identical elements in the process, method, article or smart TV including the element.

[0074] Of course, those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing related hardware (such as a processor, a controller, etc.) through a computer program, and the program can be stored in a computer-readable storage medium that can be read by a computer, and the program can include the processes of the above-mentioned method embodiments when executed. The computer-readable storage medium can be a memory, a disk, an optical disk, etc.

[0075] It should be understood that the application of the present invention is not limited to the above examples. For ordinary technicians in this field, improvements or changes can be made based on the above description. All these improvements and changes should fall within the scope of protection of the claims attached to the present invention.

Claims

1. A control intention distribution method based on a large model intelligent agent, characterized in that: The control intention distribution method based on the large model intelligent agent includes: After the user is awakened by the wake-up word, the user's voice audio is collected, and the voice audio is recognized to obtain a voice recognition result; The speech recognition result is input into the large model agent central control, the user's text intention information is identified, and the text intention information is sent to the agent to which it belongs. The agent combines the text intention information with the context stored in the agent, answers the user's intention, and saves the information of this round of dialogue; When the user initiates a second round of conversation, the second intention of the user is obtained. If the second intention does not belong to the agent, the second intention is re-input into the large model agent central control, and the second intention is recognized again and distributed to the corresponding agent.

2. The control intention distribution method based on a large model agent according to claim 1 is characterized in that: After the user is awakened by the wake-up word, the user's voice audio is collected, and the voice audio is recognized to obtain a voice recognition result, specifically including: Collect the wake-up word issued by the user, and when the wake-up word is verified, enter the device wake-up state and wait for the user's voice input; The user's voice audio is collected, the voice audio is transmitted to a voice recognition service, and the voice audio is recognized by the voice recognition service to obtain a voice recognition result.

3. The control intention distribution method based on a large model agent according to claim 2 is characterized in that: The speech recognition result is input into the large model agent central control, the text intention information of the user is recognized, the text intention information is sent to the agent to which it belongs, the agent combines the text intention information with the context stored in the agent, answers the user's intention, and saves the information of this round of dialogue, specifically including: The speech recognition result is input into the large model agent central control, and the large model agent central control recognizes the user's first text intention information according to the speech recognition result, and recognizes the first agent to which the first text intention information belongs, and sends the first text intention information to the first agent; The first agent combines the first text intention information and the context stored in the first agent to answer the first intention of the user, and saves the current round of dialogue information in the first agent.

4. The control intention distribution method based on a large model agent according to claim 3 is characterized in that: When the user initiates a second round of dialogue, the second intention of the user is obtained. If the second intention does not belong to the agent, the second intention is re-input into the large model agent central control, and the second intention is recognized again and distributed to the corresponding agent, specifically including: When the user initiates a second round of dialogue, second text intention information of the user is obtained, and the first agent determines whether the second text intention information belongs to the first agent; If the second text intention information does not belong to the first agent, the second text intention information of the second round of dialogue is re-input into the large model agent central control, and the large model agent central control recognizes the second intention again and distributes it to the corresponding second agent.

5. The control intention distribution method based on a large model agent according to claim 4 is characterized in that: After the first agent determines whether the second text intention information belongs to the first agent, the following further includes: If the second textual intention information belongs to the first agent, the first agent again answers the second intention of the user and saves the second round of dialogue information.

6. The control intention distribution method based on a large model agent according to claim 1 is characterized in that: The text intent information includes: the name of the intent and user voice description information.

7. The control intention distribution method based on a large model agent according to claim 6 is characterized in that: The user voice description information includes: person, object, color, subject and location.

8. A control intention distribution system based on a large model intelligent agent, characterized in that: The control intention distribution system based on the large model intelligent agent includes: A speech recognition module is used to collect the user's voice audio after being awakened by the user's wake-up word, recognize the voice audio, and obtain a speech recognition result; The intention control module is used to input the speech recognition result into the large model agent central control, identify the user's text intention information, and send the text intention information to the agent to which it belongs. The agent combines the text intention information with the context stored in the agent to answer the user's intention and save the information of this round of dialogue; The intention distribution module is used to obtain the second intention of the user when the user initiates the second round of dialogue. If the second intention does not belong to the agent, the second intention is re-input into the large model agent central control, and the second intention is recognized again and distributed to the corresponding agent.

9. A smart TV, characterized in that: The smart TV includes: a memory, a processor, and a large model intelligent agent-based control intention distribution program stored in the memory and executable on the processor. When the large model intelligent agent-based control intention distribution program is executed by the processor, the steps of the large model intelligent agent-based control intention distribution method as described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a control intention distribution program based on a large model intelligent agent, and when the control intention distribution program based on a large model intelligent agent is executed by a processor, the steps of the control intention distribution method based on a large model intelligent agent are implemented as described in any one of claims 1-7.

Images