Travel planning method and system based on multi-modal interaction

Abstract

The invention relates to the technical field of artificial intelligence, and discloses a travel planning method and system based on multi-modal interaction, and the method comprises the steps: setting a data transmission interface which is connected with a product database of a travel product platform and an information database of a social media platform, and accessing a semantic analysis model, a corresponding standardized data query service is established; semantic information of the standardized data query service is defined in a semantic analysis model, and the standardized data query service is packaged into a data query MCP service; the method comprises the following steps: constructing an information retrieval agent, an information screening agent and an itinerary planning agent based on an AI large model, and further creating a travel planning agent comprising the information retrieval agent, the information screening agent and the itinerary planning agent; connecting the travel planning agent with a data query MCP service and a map platform MCP service, and setting an output rendering component for the travel planning agent; the effect of improving the travel plan making efficiency is achieved.

Description

Technical Field

[0001] This application relates to the technical field of artificial intelligence, and in particular to a travel planning method and system based on multimodal interaction. Background Technology

[0002] Currently, travelers with travel or tourism needs mainly make travel plans by actively checking travel information, travel guides, and user reviews on social media and travel service platforms. However, the information that travelers can access on each platform is quite limited due to the platform's main business, making it difficult to collect multiple types of information on one platform. Furthermore, travelers' individual information gathering capabilities are limited, making it time-consuming and laborious to make travel plans.

[0003] Therefore, the aforementioned technologies suffer from low efficiency in helping passengers make travel plans. Summary of the Invention

[0004] To improve the efficiency of travel planning, this application provides a travel planning method and system based on multimodal interaction.

[0005] The first objective of this invention is achieved by the following technical solution: Multimodal interaction-based travel planning methods include: S1: Set up a data transmission interface connecting the product database of the travel product platform and the information database of the social media platform, and integrate a semantic parsing model to establish a corresponding standardized data query service; S2: Define the semantic information of the standardized data query service in the semantic parsing model, and encapsulate the standardized data query service into a data query MCP service; S3: Based on the AI ​​big model, construct an information retrieval intelligent agent, an information filtering intelligent agent, and a trip planning intelligent agent, and then create a trip planning intelligent agent that includes the information retrieval intelligent agent, the information filtering intelligent agent, and the trip planning intelligent agent; S4: Connect the travel planning agent to the data query MCP service and the map platform MCP service, and set up an output rendering component for the travel planning agent.

[0006] By adopting the above technical solution, data transmission interfaces are set up to connect the product database of the travel product platform and the information database of the social media platform. This allows for real-time acquisition of product and information data from these platforms. A semantic parsing model is then integrated to perform semantic analysis on the acquired data, providing data support for subsequent interactive travel planning and establishing a standardized data query service. Semantic information for the standardized data query service is defined within the semantic parsing model to enhance its functionality. This standardized data query service is encapsulated as a data query MCP service, facilitating subsequent integration with intelligent agents built on AI large-scale models. Information retrieval is then constructed based on the AI ​​large-scale model. The system comprises three intelligent agents: an information retrieval agent, an information filtering agent, and a trip planning agent. Further, a trip planning agent is created, incorporating these three agents and combining their functions, to facilitate efficient travel planning for passengers. This trip planning agent is then integrated with data query MCP services and map platform MCP services. By introducing MCP services, dynamic retrieval of external multi-source heterogeneous data is achieved. Compared to traditional retrieval enhancement frameworks that rely on static text vector libraries and semantic retrieval, this approach eliminates the need for local knowledge bases, enabling real-time synchronization of the latest data from existing systems and maintaining data freshness. Output rendering components are also included to enable rich modal information output.

[0007] In a preferred embodiment of this application, the step of setting up a data transmission interface connecting the product database of a travel product platform and the information database of a social media platform, and accessing a semantic parsing model to establish a corresponding standardized data query service, further includes: S11: Extract travel product data from the travel agency enterprise system based on the data collection cycle, extract key information fields from the travel product data and store them in the product database of the travel product platform; S12: Obtain travel information data from social media platforms based on data collection cycles and keyword groups, extract key information fields and feedback tendency information from the travel information data, and store them in the information database of the social media platform; S13: Set up a data transmission interface and semantic parsing model for the product database and information database to access the travel planning intelligent agent, and establish a standardized data query service; The key information fields include index information, travel destination, spending amount, personnel composition, time, and weather; the keyword group includes several keywords for identifying travel information data and several keywords for identifying key information fields and feedback tendency information from unstructured data; the feedback tendency information refers to the information on users' likes and dislikes towards travel projects on social media platforms.

[0008] By adopting the above technical solution, travel product data is extracted from the target travel agency's enterprise system based on a preset data collection cycle. Since key information in travel product data is usually recorded in a specific format, key information fields are extracted from the travel product data and stored in the product database of the travel product platform to reduce the computer performance requirements for data storage. The key information fields contain index information for locating the data source, so that new key information fields can be checked or supplemented when necessary. Based on the data collection cycle and preset keyword groups, travel-related content such as travel guides, travel sharing, and travel-related news are obtained from the target social media platform as travel information data. Furthermore, based on keyword groups, key information fields and feedback tendency information are identified and extracted from the unstructured travel information data and stored in the information database of the social media platform to reduce storage space requirements. Data transmission interfaces and semantic parsing models are set up for the product database and information database to access the travel planning intelligent agent, so as to perform semantic parsing on the key information fields and feedback tendency information extracted from the acquired travel product data and travel information data, and participate in the subsequent travel planning process.

[0009] In a preferred embodiment of this application: the construction of an information retrieval agent, an information filtering agent, and a trip planning agent based on an AI large model, and the creation of a trip planning agent comprising the information retrieval agent, the information filtering agent, and the trip planning agent, includes: S31: Construct an information retrieval intelligent agent with natural language intelligent question answering function based on AI large model, so as to extract the first constraint information from the user's dialogue content, call the data query MCP service, retrieve and structure the first related data; S32: Construct an information filtering intelligent agent based on a multimodal visual big model to understand the content of the first associated data, filter content posts that meet the preset quality standards, and generate labeled text; S33: Based on a large AI model, a trip planning intelligent agent is built to extract second constraint information from the user's dialogue content to perform product screening and trip planning, and generate a travel plan form; The first constraint information includes travel destinations; the first associated data refers to several key information fields and feedback tendency information that conform to the first constraint information retrieved from the product database and information database, as well as the original content retrieved based on the index information; the second constraint information includes rigid constraints and flexible constraints, the rigid constraints including budget, number of people, time, and weather; the flexible constraints include travel preferences.

[0010] By adopting the above technical solutions, information retrieval, travel planning, and product recommendation are completed through multi-agent collaboration. Compared with single-agent frameworks or retrieval enhancement frameworks, this mechanism constructs a collaborative system composed of multiple specialized agents. Each agent is equipped with a few tools and focuses on limited functions, thus avoiding the problem of agents being unable to make reasonable choices due to too many tools, and improving the stability and interpretability of agent content output. By using a multimodal large model to perform semantic annotation on visual materials such as text, images, and videos, the problem of lack of semantic description of images on the Internet and social media, making them difficult for language large models to understand and difficult to associate with context is solved. This makes the final generated answers more visually appealing, helping users form an intuitive and three-dimensional understanding before traveling, improving the intuitiveness of information transmission and the accuracy of user decision-making.

[0011] In a preferred embodiment of this application: the creation of a travel planning agent comprising the information retrieval agent, the information filtering agent, and the trip planning agent includes: S34: Create a travel planning intelligent agent based on a large AI model and connect various professional intelligent agents; S35: The travel planning intelligent agent receives the user's dialogue information and engages in multi-round question-and-answer sessions with the user through natural language, driving subsequent information retrieval and itinerary planning tasks based on the cycle of "thinking-action-observation"; S36: The travel planning intelligent agent generates sub-tasks based on the identified user demand information, forwards the sub-tasks to the corresponding professional intelligent agents, collects and re-integrates the output content of each professional intelligent agent, and generates a final decision report. The specialized intelligent agents include information retrieval intelligent agents, information filtering intelligent agents, and itinerary planning intelligent agents.

[0012] By adopting the above technical solution, a travel planning intelligent agent is created based on an AI large-scale model and connected to an information retrieval intelligent agent, an information filtering intelligent agent, and a trip planning intelligent agent. This allows travel planning to be executed with the assistance of these specialized intelligent agents. The travel planning intelligent agent receives user dialogue information and engages in multi-round question-and-answer sessions with the user through natural language to obtain and supplement the user's constraints on the travel plan. This enables subsequent information retrieval and trip planning tasks to be driven based on a "thinking-action-observation" cycle. The travel planning intelligent agent generates sub-tasks based on the identified user needs information and forwards them to the corresponding specialized intelligent agents. This facilitates targeted problem-solving and planning by each specialized intelligent agent, collects and re-integrates the output content of each specialized intelligent agent, generates a final decision report, and improves the quality of the intelligent agent's content output.

[0013] In a preferred embodiment of this application: the step of connecting the travel planning agent to the data query MCP service and the map platform MCP service, and setting up an output rendering component for the travel planning agent, includes: S41: Connect the travel planning intelligent agent to the map platform MCP service, which integrates trip planning, route planning, and weather query tools; S42: The rendering component dynamically parses the content output of the travel planning intelligent agent, extracts multimedia links and HTML structure from the output content, and defines different style rendering templates accordingly.

[0014] By adopting the above technical solutions, the travel planning intelligent agent is connected to the data query MCP service to obtain relevant products and content posts from travel product platforms and social media platforms for travel planning; it is also connected to the map platform MCP service and integrates multiple tool calling methods such as itinerary planning, route planning, and weather query to serve as reference information for travel planning; the rendering component dynamically parses the content output of the travel planning intelligent agent on the front end, extracts multimedia links and HTML structure from the output content through custom tag pairs, and defines different style rendering templates to achieve rich modal information output.

[0015] The second objective of this application is achieved by the following technical solution: A multimodal interaction-based travel planning system includes: The data query service establishment module is used to set up the data transmission interface connecting the product database of the travel product platform and the information database of the social media platform, and to connect to the semantic parsing model to establish a corresponding standardized data query service. The data query service encapsulation module is used to define the semantic information of the standardized data query service in the semantic parsing model and encapsulate the standardized data query service into a data query MCP service. The travel planning intelligent agent creation module is used to construct an information retrieval intelligent agent, an information filtering intelligent agent, and a trip planning intelligent agent based on an AI big model, and then create a travel planning intelligent agent that includes the information retrieval intelligent agent, the information filtering intelligent agent, and the trip planning intelligent agent. The output rendering settings module is used to connect the travel planning intelligent agent to the data query MCP service and the map platform MCP service, and to set up the output rendering components for the travel planning intelligent agent.

[0016] In a preferred embodiment of this application, the data query service establishment module includes: The travel product data extraction submodule is used to extract travel product data from the travel agency enterprise system based on the data collection cycle, extract key information fields from the travel product data and store them in the product database of the travel product platform. The travel information data extraction submodule is used to obtain travel information data from social media platforms based on data collection cycles and keyword groups, extract key information fields and feedback tendency information from the travel information data, and store them in the information database of the social media platform. The data query service submodule is used to set up data transmission interfaces and semantic parsing models for accessing the travel planning intelligent agent in the product database and information database, and to establish standardized data query services. The key information fields include index information, travel destination, spending amount, personnel composition, time, and weather; the keyword group includes several keywords for identifying travel information data and several keywords for identifying key information fields and feedback tendency information from unstructured data; the feedback tendency information refers to the information on users' likes and dislikes towards travel projects on social media platforms.

[0017] In a preferred embodiment of this application, the travel planning agent creation module includes: The first associated data acquisition submodule is used to build an information retrieval intelligent agent with natural language intelligent question answering function based on the AI ​​big model, so as to extract the first constraint information from the user's dialogue content, call the data query MCP service, retrieve and structure the first associated data; The annotation text generation submodule is used to build an information filtering agent based on a multimodal visual big model to understand the content of the first associated data, filter content that meets the preset quality standards, and generate annotation text. The travel plan form generation submodule is used to build an intelligent agent for trip planning based on a large AI model. It extracts second constraint information from the user's dialogue content to perform product filtering and trip planning, and generates a travel plan form. The first constraint information includes travel destinations; the first associated data refers to several key information fields and feedback tendency information that conform to the first constraint information retrieved from the product database and information database, as well as the original content retrieved based on the index information; the second constraint information includes rigid constraints and flexible constraints, the rigid constraints including budget, number of people, time, and weather; the flexible constraints include travel preferences.

[0018] The third objective of this invention is achieved by the following technical solution: A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the computer program, implements the steps of the multimodal interaction-based travel planning method as described in any of the preceding claims.

[0019] The fourth objective of this invention is achieved by the following technical solution: A computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the multimodal interaction-based travel planning method described in any of the preceding claims.

[0020] In summary, this application includes at least one of the following beneficial technical effects: 1. Set up data transmission interfaces to connect the product database of the travel product platform and the information database of the social media platform, so as to obtain product and information data from the travel product platform and social media platform in real time. Integrate with the semantic parsing model to perform semantic analysis on the acquired data for subsequent interactive travel planning data support, thereby establishing a standardized data query service; define the semantic information of the standardized data query service in the semantic parsing model to improve the semantic parsing function, and encapsulate the standardized data query service into a data query MCP service to facilitate subsequent integration with intelligent agents built based on AI big data models; build an information retrieval intelligent agent based on AI big data models, The system first develops an information filtering agent and a trip planning agent. Then, it further creates a comprehensive travel planning agent that combines these three agents, facilitating efficient travel planning for passengers. This travel planning agent is then integrated with data query MCP services and map platform MCP services to enable dynamic retrieval of external multi-source heterogeneous data. Compared to traditional retrieval enhancement frameworks that rely on static text vector libraries and semantic retrieval, this approach eliminates the need for local knowledge bases, allowing for real-time synchronization of the latest data from existing systems and maintaining data freshness. Output rendering components are also included to enable rich modal information output.

[0021] 2. By using multi-agent collaboration to complete information retrieval, travel planning, and product recommendation, compared to a single-agent framework or retrieval enhancement framework, this mechanism constructs a collaborative system composed of multiple specialized agents. Each agent is equipped with a few tools and focuses on limited functions. Therefore, it avoids the problem of agents being unable to make reasonable choices due to too many tools, and improves the stability and interpretability of agent content output.

[0022] 3. By using a multimodal large model to semantically annotate visual materials such as text, images, and videos, we solve the problem that images on the Internet and social media lack semantic descriptions, are difficult for language large models to understand, and are difficult to associate with context. This makes the generated answers more visually appealing and helps users form an intuitive and three-dimensional understanding before traveling, improving the intuitiveness of information delivery and the accuracy of user decision-making. Attached Figure Description

[0023] Figure 1 This is a flowchart of the multimodal interaction-based travel planning method in Embodiment 1 of this application.

[0024] Figure 2 This is a principle block diagram of the multimodal interaction-based travel planning system in Embodiment 2 of this application.

[0025] Figure 3 This is a schematic diagram of the device in Embodiment 3 of this application. Detailed Implementation

[0026] The following is in conjunction with the appendix Figures 1 to 3 This application will be described in further detail.

[0027] Example 1 Reference Figure 1 This application discloses a travel planning method based on multimodal interaction, which specifically includes the following steps: S1: Set up a data transmission interface connecting the product database of the travel product platform and the information database of the social media platform, and integrate a semantic parsing model to establish a corresponding standardized data query service.

[0028] In this embodiment, there can be one or more travel product platforms, which can be set according to actual needs.

[0029] Specifically, data transmission interfaces are set up to connect the product database of the travel product platform and the information database of the social media platform, so as to obtain product and information data from the travel product platform and social media platform in real time. The data is then connected to a semantic parsing model to perform semantic parsing on the acquired data, which is used to support the data of subsequent interactive travel planning, thereby establishing a standardized data query service.

[0030] S11: Extract travel product data from the travel agency's enterprise system based on the data collection cycle, extract key information fields from the travel product data, and store them in the product database of the travel product platform.

[0031] In this embodiment, the data collection period refers to the preset period for extracting the latest travel product data. Preferably, the data collection period can be set to one hour. The travel agency enterprise system can be an enterprise management system or a data platform system. Key information fields include index information and descriptive information on travel destinations, consumption amounts, personnel composition, time, and weather. The index information refers to the information used to locate the original content posting location.

[0032] Specifically, travel product data is extracted from the target travel agency's enterprise system based on a preset data collection cycle. Since key information in travel product data is usually recorded in a specific format, key information fields are extracted from the travel product data and stored in the product database of the travel product platform to reduce the computer performance requirements for data storage. Furthermore, the key information fields contain index information for locating the source of the data, so that new key information fields can be checked or supplemented when necessary.

[0033] S12: Obtain travel information data from social media platforms based on data collection cycles and keyword groups, extract key information fields and feedback tendency information from the travel information data, and store them in the information database of the social media platform.

[0034] In this embodiment, the keyword group includes several keywords for identifying travel information data and several keywords for identifying key information fields and feedback tendency information from unstructured data, so as to find content posts related to travel / trip; the content posts can be presented in the form of text, images, videos or any combination of the three, including content posts related to relevant product information and social media posts; feedback tendency information refers to the information on the likes and dislikes of social media platform users for travel projects.

[0035] Specifically, based on the data collection cycle and preset keyword groups, travel-related content such as travel guides, travel sharing, and travel-related news is obtained from the target social media platform as travel information data. Furthermore, based on the keyword groups, key information fields and feedback tendency information are identified and extracted from the unstructured travel information data and stored in the information database of the social media platform to reduce storage space requirements. The target social media platform refers to the social media platform from which the travel information data in this application method originates.

[0036] S13: Set up data transmission interfaces and semantic parsing models for the product database and information database to access the travel planning intelligent agent, and establish standardized data query services.

[0037] Specifically, data transmission interfaces and semantic parsing models are set up for the product database and information database to access the travel planning intelligent agent, so as to perform semantic parsing on key information fields and feedback tendency information extracted from the acquired travel product data and travel information data, and participate in the subsequent travel planning process.

[0038] S2: Define the semantic information of the standardized data query service in the semantic parsing model, and encapsulate the standardized data query service into a data query MCP service.

[0039] In this embodiment, semantic information includes method name and function, meaning of input parameters, and format of output parameters.

[0040] Specifically, the semantic information of the standardized data query service is defined in the semantic parsing model to improve the semantic parsing function. The standardized data query service is encapsulated as a data query MCP service, which facilitates the subsequent access of intelligent agents built on the AI ​​big model.

[0041] S3: Based on the AI ​​big model, construct an information retrieval intelligent agent, an information filtering intelligent agent, and a trip planning intelligent agent, and then create a trip planning intelligent agent that includes the information retrieval intelligent agent, the information filtering intelligent agent, and the trip planning intelligent agent.

[0042] Specifically, an information retrieval intelligent agent, an information filtering intelligent agent, and a trip planning intelligent agent are constructed based on a large AI model. Furthermore, a travel planning intelligent agent is created that includes the information retrieval intelligent agent, the information filtering intelligent agent, and the trip planning intelligent agent and realizes the functional combination, which facilitates the efficient formulation of travel plans for passengers in the future.

[0043] Step S3 also includes: S31: Construct an information retrieval intelligent agent with natural language intelligent question answering function based on AI large model, so as to extract the first constraint information from the user's dialogue content, call the data query MCP service, retrieve and structure the first related data.

[0044] In this embodiment, the first constraint information includes travel destinations; the first associated data refers to several key information fields and feedback tendency information that conform to the first constraint information retrieved from the product database and information database, as well as the original content retried based on the index information; Specifically, based on existing large-scale AI models such as intelligent question-answering systems / models, an information retrieval intelligent agent with natural language intelligent question-answering capabilities is constructed to learn about users' needs and constraints for travel or tourism through dialogue. Based on the user's first constraint information, the data query MCP service is invoked to retrieve first-related data that matches the first constraint information from product databases and information databases. The first-related data is then organized in a structured manner to obtain the retrieval results, which can be used as reference data for subsequent travel planning.

[0045] S32: Construct an information filtering agent based on a multimodal visual big data model to understand the content of the first associated data, filter content posts that meet the preset quality standards, and generate labeled text.

[0046] In this embodiment, the multimodal visual big model refers to an existing AI big model with functions such as image feature recognition, image feature understanding, and semantic recognition; the preset quality standard refers to the standard used to filter content posts, including requirements for multiple indicators such as content relevance, popularity, timeliness, content richness, and feedback tendency. The requirements for indicators such as content relevance, popularity, timeliness, and content richness can be set with reference to existing platform quality rules or according to actual needs; the labeled text refers to the text labeling of the content description and media quality of the relevant content posts.

[0047] Specifically, an information filtering agent is constructed based on the existing multimodal visual big data model. The agent performs content understanding on the content patches in the first associated data using semantic recognition and image recognition algorithms, thereby filtering out content patches that meet the preset quality standards and generating corresponding labeled text based on the content understanding.

[0048] S33: Based on a large AI model, a smart agent for trip planning is built. Secondary constraint information is extracted from the user's dialogue content to perform product screening and trip planning, and a travel plan form is generated.

[0049] In this embodiment, the second constraint information includes rigid constraints and flexible constraints. Rigid constraints include budget, number of people, time, and weather; flexible constraints include travel preferences.

[0050] Specifically, a trip planning intelligent agent is built based on the existing large AI model. It further extracts second constraint information from the user's dialogue content, including rigid constraints and flexible constraints, so as to perform product screening and trip planning based on the second constraint information, thereby generating a travel plan form, which facilitates the improvement of the efficiency of travel plan formulation.

[0051] Furthermore, by collaborating with multiple agents to complete information retrieval, travel planning, and product recommendations, this mechanism constructs a collaborative system composed of multiple specialized agents compared to single-agent frameworks or retrieval enhancement frameworks. Each agent is equipped with a few tools and focuses on limited functions, thus avoiding the problem of agents being unable to make reasonable choices due to too many tools, and improving the stability and interpretability of agent content output. By using a multimodal large model to semantically annotate visual materials such as text, images, and videos, it solves the problem that images on the Internet and social media lack semantic descriptions, are difficult to understand by language large models, and are difficult to associate with context. This makes the final generated answers more visually appealing, helping users form an intuitive and three-dimensional understanding before traveling, improving the intuitiveness of information delivery and the accuracy of user decision-making.

[0052] Step S3 also includes: S34: Create a travel planning intelligent agent based on a large AI model and connect various professional intelligent agents.

[0053] In this embodiment, the specialized intelligent agents include an information retrieval intelligent agent, an information filtering intelligent agent, and a trip planning intelligent agent.

[0054] Specifically, a travel planning intelligent agent is created based on a large AI model and connected to an information retrieval intelligent agent, an information filtering intelligent agent, and a trip planning intelligent agent, so that travel planning can be executed with the assistance of these specialized intelligent agents.

[0055] S35: Receives user dialogue information to conduct multi-round question-and-answer sessions with the user using natural language, driving subsequent information retrieval and itinerary planning tasks based on a cycle of "thinking-action-observation".

[0056] Specifically, it receives user dialogue information and engages in multi-round question-and-answer sessions with users through natural language in order to obtain and supplement the user's constraints on the travel plan. This enables subsequent information retrieval and itinerary planning tasks to be driven based on a cycle of "thinking-action-observation".

[0057] S36: Generate sub-tasks based on the identified user needs information, forward the sub-tasks to the corresponding professional intelligent agents, collect and re-integrate the output content of each intelligent agent, and generate a final decision report.

[0058] Specifically, sub-tasks are generated based on the identified user needs and forwarded to the corresponding professional intelligent agents. This allows each professional intelligent agent to address problems in a targeted manner, develop planning solutions, collect and re-integrate the output content of each intelligent agent, generate a final decision report, and improve the quality of the intelligent agent's content output.

[0059] S4: Connect the travel planning agent to the data query MCP service and the map platform MCP service, and set up an output rendering component for the travel planning agent.

[0060] Specifically, the travel planning intelligent agent is connected to the data query MCP service and the map platform MCP service. By introducing the MCP service, dynamic retrieval of external multi-source heterogeneous data can be achieved. Compared with the traditional retrieval enhancement framework that relies on static text vector libraries and semantic retrieval, it does not need to rely on local knowledge bases. It can synchronize the latest data of the existing system in real time, keep the data fresh, and set up output rendering components to achieve rich modal information output.

[0061] Step S4 also includes: S41: Connect the travel planning intelligent agent to the map platform MCP service, which integrates trip planning, route planning, and weather query tools.

[0062] Specifically, the travel planning intelligent agent is connected to the data query MCP service to obtain relevant products and content posts from travel product platforms and social media platforms for travel planning; it is also connected to the map platform MCP service and integrates multiple tool calling methods such as itinerary planning, route planning, and weather query to serve as reference information for travel planning.

[0063] S42: The rendering component dynamically parses the content output of the travel planning intelligent agent, extracts multimedia links and HTML structure from the output content, and defines different style rendering templates accordingly.

[0064] Specifically, the rendering component dynamically parses the content output of the travel planning intelligent agent on the front end, extracts multimedia links and HTML structure from the output content through custom tag pairs, and defines different style rendering templates to achieve rich modal information output.

[0065] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0066] Example 2 This application discloses a travel planning system based on multimodal interaction, which corresponds to the travel planning method based on multimodal interaction in the above embodiments.

[0067] like Figure 2As shown, the multimodal interaction-based travel planning system includes a data query service creation module, a data query service encapsulation module, a travel planning intelligent agent creation module, and an output rendering settings module. Detailed descriptions of each functional module are as follows: The data query service establishment module is used to set up the data transmission interface connecting the product database of the travel product platform and the information database of the social media platform, and to connect to the semantic parsing model to establish a corresponding standardized data query service. The data query service encapsulation module is used to define the semantic information of the standardized data query service in the semantic parsing model and encapsulate the standardized data query service into a data query MCP service. The travel planning intelligent agent creation module is used to construct an information retrieval intelligent agent, an information filtering intelligent agent, and a trip planning intelligent agent based on an AI big model, and then create a travel planning intelligent agent that includes the information retrieval intelligent agent, the information filtering intelligent agent, and the trip planning intelligent agent. The output rendering settings module is used to connect the travel planning intelligent agent to the data query MCP service and the map platform MCP service, and to set up the output rendering components for the travel planning intelligent agent.

[0068] The data query service establishment module includes: The travel product data extraction submodule is used to extract travel product data from the travel agency enterprise system based on the data collection cycle, extract key information fields from the travel product data and store them in the product database of the travel product platform. The travel information data extraction submodule is used to obtain travel information data from social media platforms based on data collection cycles and keyword groups, extract key information fields and feedback tendency information from the travel information data, and store them in the information database of the social media platform. The data query service submodule is used to set up data transmission interfaces and semantic parsing models for the product database and information database to access the travel planning intelligent agent, and to establish standardized data query services.

[0069] The travel planning intelligent agent creation module includes: The first associated data acquisition submodule is used to build an information retrieval intelligent agent with natural language intelligent question answering function based on the AI ​​big model, so as to extract the first constraint information from the user's dialogue content, call the data query MCP service, retrieve and structure the first associated data; The annotation text generation submodule is used to build an information filtering agent based on a multimodal visual big model to understand the content of the first associated data, filter content that meets the preset quality standards, and generate annotation text. The travel plan form generation submodule is used to build an intelligent agent for trip planning based on a large AI model. It extracts secondary constraint information from the user's dialogue content to perform product filtering and trip planning, and generates a travel plan form.

[0070] For specific limitations regarding the multimodal interaction-based travel planning system, please refer to the limitations of the multimodal interaction-based travel planning method mentioned above, which will not be repeated here. Each module in the multimodal interaction-based travel planning system can be implemented entirely or partially through software, hardware, or a combination thereof. Each module can be embedded in the processor of the computer device in hardware form or independent of it, or it can be stored in the memory of the computer device in software form, so that the processor can call and execute the corresponding operations of each module.

[0071] In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 3 As shown, the computer device includes a processor, memory, network interface, and database connected via a system bus. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and database. The internal memory provides the environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The database stores data such as product databases, information databases, semantic parsing models, semantic information, information retrieval agents, information filtering agents, itinerary planning agents, travel planning agents, and rendering components. The network interface is used for communication with external terminals via a network connection. When the computer program is executed by the processor, it implements a multimodal interaction-based travel planning method.

[0072] In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to perform the following steps: S1: Set up a data transmission interface connecting the product database of the travel product platform and the information database of the social media platform, and integrate a semantic parsing model to establish a corresponding standardized data query service; S2: Define the semantic information of the standardized data query service in the semantic parsing model, and encapsulate the standardized data query service into a data query MCP service; S3: Based on the AI ​​big model, construct an information retrieval intelligent agent, an information filtering intelligent agent, and a trip planning intelligent agent, and then create a trip planning intelligent agent that includes the information retrieval intelligent agent, the information filtering intelligent agent, and the trip planning intelligent agent; S4: Connect the travel planning agent to the data query MCP service and the map platform MCP service, and set up an output rendering component for the travel planning agent.

[0073] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, the computer program performing the following steps when executed by a processor: S1: Set up a data transmission interface connecting the product database of the travel product platform and the information database of the social media platform, and integrate a semantic parsing model to establish a corresponding standardized data query service; S2: Define the semantic information of the standardized data query service in the semantic parsing model, and encapsulate the standardized data query service into a data query MCP service; S3: Based on the AI ​​big model, construct an information retrieval intelligent agent, an information filtering intelligent agent, and a trip planning intelligent agent, and then create a trip planning intelligent agent that includes the information retrieval intelligent agent, the information filtering intelligent agent, and the trip planning intelligent agent; S4: Connect the travel planning agent to the data query MCP service and the map platform MCP service, and set up an output rendering component for the travel planning agent.

[0074] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), Synchlink, DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and RAMbus dynamic RAM (RDRAM), etc.

[0075] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is used as an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0076] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. A travel planning method based on multimodal interaction, characterized in that, include: S1: Set up a data transmission interface connecting the product database of the travel product platform and the information database of the social media platform, and integrate a semantic parsing model to establish a corresponding standardized data query service; S2: Define the semantic information of the standardized data query service in the semantic parsing model, and encapsulate the standardized data query service into a data query MCP service; S3: Based on the AI ​​big model, construct an information retrieval intelligent agent, an information filtering intelligent agent, and a trip planning intelligent agent, and then create a trip planning intelligent agent that includes the information retrieval intelligent agent, the information filtering intelligent agent, and the trip planning intelligent agent; S4: Connect the travel planning agent to the data query MCP service and the map platform MCP service, and set up an output rendering component for the travel planning agent.

2. The travel planning method based on multimodal interaction according to claim 1, characterized in that: The step of setting up a data transmission interface connecting the product database of the travel product platform and the information database of the social media platform, and integrating a semantic parsing model to establish a corresponding standardized data query service, further includes: S11: Extract travel product data from the travel agency enterprise system based on the data collection cycle, extract key information fields from the travel product data and store them in the product database of the travel product platform; S12: Obtain travel information data from social media platforms based on data collection cycles and keyword groups, extract key information fields and feedback tendency information from the travel information data, and store them in the information database of the social media platform; S13: Set up a data transmission interface and semantic parsing model for the product database and information database to access the travel planning intelligent agent, and establish a standardized data query service; The key information fields include index information, travel destination, spending amount, personnel composition, time, and weather; the keyword group includes several keywords for identifying travel information data and several keywords for identifying key information fields and feedback tendency information from unstructured data; the feedback tendency information refers to the information on users' likes and dislikes towards travel projects on social media platforms.

3. The travel planning method based on multimodal interaction according to claim 1, characterized in that: The process involves constructing an information retrieval agent, an information filtering agent, and a trip planning agent based on a large AI model, and then creating a trip planning agent that includes these three agents: S31: Construct an information retrieval intelligent agent with natural language intelligent question answering function based on AI large model, so as to extract the first constraint information from the user's dialogue content, call the data query MCP service, retrieve and structure the first related data; S32: Construct an information filtering intelligent agent based on a multimodal visual big model to understand the content of the first associated data, filter content posts that meet the preset quality standards, and generate labeled text; S33: Based on a large AI model, a trip planning intelligent agent is built to extract second constraint information from the user's dialogue content to perform product screening and trip planning, and generate a travel plan form; The first constraint information includes travel destinations; the first associated data refers to several key information fields and feedback tendency information that conform to the first constraint information retrieved from the product database and information database, as well as the original content retrieved based on the index information; the second constraint information includes rigid constraints and flexible constraints, the rigid constraints including budget, number of people, time, and weather; the flexible constraints include travel preferences.

4. The travel planning method based on multimodal interaction according to claim 3, characterized in that: The creation of a travel planning agent that includes the information retrieval agent, the information filtering agent, and the trip planning agent includes: S34: Create a travel planning intelligent agent based on a large AI model and connect various professional intelligent agents; S35: The travel planning intelligent agent receives the user's dialogue information and conducts multi-round question and answer with the user through natural language, driving subsequent information retrieval and itinerary planning tasks based on the cycle of "thinking-action-observation"; S36: The travel planning intelligent agent generates sub-tasks based on the identified user demand information, forwards the sub-tasks to the corresponding professional intelligent agents, collects and re-integrates the output content of each professional intelligent agent, and generates a final decision report. The specialized intelligent agents include information retrieval intelligent agents, information filtering intelligent agents, and itinerary planning intelligent agents.

5. The travel planning method based on multimodal interaction according to claim 1, characterized in that: The step of connecting the travel planning intelligent agent to the data query MCP service and the map platform MCP service, and setting up an output rendering component for the travel planning intelligent agent, includes: S41: Connect the travel planning intelligent agent to the map platform MCP service, which integrates trip planning, route planning, and weather query tools; S42: The rendering component dynamically parses the content output of the travel planning intelligent agent, extracts multimedia links and HTML structure from the output content, and defines different style rendering templates accordingly.

6. A tourism travel planning system based on multimodal interaction, characterized in that: include: The data query service establishment module is used to set up the data transmission interface connecting the product database of the travel product platform and the information database of the social media platform, and to connect to the semantic parsing model to establish a corresponding standardized data query service. The data query service encapsulation module is used to define the semantic information of the standardized data query service in the semantic parsing model and encapsulate the standardized data query service into a data query MCP service. The travel planning intelligent agent creation module is used to construct an information retrieval intelligent agent, an information filtering intelligent agent, and a trip planning intelligent agent based on an AI big model, and then create a travel planning intelligent agent that includes the information retrieval intelligent agent, the information filtering intelligent agent, and the trip planning intelligent agent. The output rendering settings module is used to connect the travel planning intelligent agent to the data query MCP service and the map platform MCP service, and to set up the output rendering components for the travel planning intelligent agent.

7. The travel planning system based on multimodal interaction according to claim 6, characterized in that, The data query service establishment module includes: The travel product data extraction submodule is used to extract travel product data from the travel agency enterprise system based on the data collection cycle, extract key information fields from the travel product data and store them in the product database of the travel product platform. The travel information data extraction submodule is used to obtain travel information data from social media platforms based on data collection cycles and keyword groups, extract key information fields and feedback tendency information from the travel information data, and store them in the information database of the social media platform. The data query service submodule is used to set up data transmission interfaces and semantic parsing models for accessing the travel planning intelligent agent in the product database and information database, and to establish standardized data query services. The key information fields include index information, travel destination, spending amount, personnel composition, time, and weather; the keyword group includes several keywords for identifying travel information data and several keywords for identifying key information fields and feedback tendency information from unstructured data; the feedback tendency information refers to the information on users' likes and dislikes towards travel projects on social media platforms.

8. The travel planning system based on multimodal interaction according to claim 6, characterized in that, The travel planning intelligent agent creation module includes: The first associated data acquisition submodule is used to build an information retrieval intelligent agent with natural language intelligent question answering function based on the AI ​​big model, so as to extract the first constraint information from the user's dialogue content, call the data query MCP service, retrieve and structure the first associated data; The annotation text generation submodule is used to build an information filtering agent based on a multimodal visual big model to understand the content of the first associated data, filter content that meets the preset quality standards, and generate annotation text. The travel plan form generation submodule is used to build an intelligent agent for trip planning based on a large AI model. It extracts second constraint information from the user's dialogue content to perform product filtering and trip planning, and generates a travel plan form. The first constraint information includes travel destinations; the first associated data refers to several key information fields and feedback tendency information that conform to the first constraint information retrieved from the product database and information database, as well as the original content retrieved based on the index information; the second constraint information includes rigid constraints and flexible constraints, the rigid constraints including budget, number of people, time, and weather; the flexible constraints include travel preferences.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the multimodal interaction-based travel planning method as described in any one of claims 1 to 5.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the multimodal interaction-based travel planning method as described in any one of claims 1 to 5.

Images