Station information question and answer method, device and equipment for rail transit and storage medium

Abstract

The invention provides a rail transit station information question and answer method and device, equipment and a storage medium. The method comprises the steps that questions proposed by passengers are acquired; semantic information vectors of the questions are determined through an RAG model, and content blocks corresponding to the semantic information vectors are retrieved in a vector database; wherein the content blocks stored in the vector database are obtained based on an equipment meta-information document describing in-station position information and a station information document describing station comprehensive service information; each site corresponds to one equipment meta-information document and one site information document; integrating the corresponding content blocks to generate answers; and displaying the answer to the passenger. According to the method provided by the invention, real-time answer retrieval and generation can be realized through the RAG model, the provided answers are ensured to be accurate and have timeliness, in addition, downloading and installation are not needed when the method provided by the invention is implemented, the convenience of question answering is improved, and diversified query requirements of passengers can be met.

Description

Technical Field

[0001] The present application relates to the field of rail transit technology, and in particular to a method, device, equipment, and storage medium for questioning and answering station information of rail transit. Background Art

[0002] With the acceleration of urbanization, rail transit has become an indispensable part of urban residents' daily travel, and its service quality directly affects the quality of life of citizens. However, in high-traffic areas such as subway stations, passengers often face problems such as finding exits, transfer routes, and nearby service facilities.

[0003] In the prior art, rail transit subway station information query systems usually adopt the following methods:

[0004] 1. Traditional information query system: provides information through manually compiled paper or electronic maps, signboards, etc., but the information is not updated in a timely manner.

[0005] 2. Voice assistant: Some subway stations are equipped with voice assistants, but their functions are limited and cannot meet the diverse query needs of passengers.

[0006] 3. Mobile applications: Some subway companies have developed mobile applications to provide real-time information query services, but passengers are required to download and install them, which is inconvenient to use. Summary of the invention

[0007] In order to solve one of the above-mentioned technical defects, the present application provides a method, device, equipment and storage medium for rail transit station information question and answer.

[0008] In a first aspect of the present application, a method for questioning and answering station information of rail transit is provided, the method comprising:

[0009] Get questions from passengers;

[0010] Determine the semantic information vector of the problem through the RAG model, and retrieve the content block corresponding to the semantic information vector in the vector database; wherein the content block stored in the vector database is obtained based on the device meta information document describing the location information within the station and the site information document describing the comprehensive service information of the site; each site corresponds to a device meta information document and a site information document;

[0011] Integrate the corresponding content blocks to generate answers;

[0012] Show the answer to the passenger.

[0013] Optionally, before integrating the corresponding content blocks and generating the answer, the following steps may also be included:

[0014] Obtain map image information and / or navigation information corresponding to the question through the map module;

[0015] Integrate the corresponding content blocks and generate answers, including:

[0016] Integrate the corresponding content blocks, corresponding map image information and / or navigation information to generate answers.

[0017] Optionally, before determining the semantic information vector of the question by using the RAG model and searching the vector database for the content block corresponding to the semantic information vector, the method further includes:

[0018] Splitting the device meta information document and the site information document into information blocks;

[0019] Encode information blocks into high-dimensional vectors;

[0020] Store high-dimensional vectors in a vector database.

[0021] Optionally, the in-station location information includes one or more of the following: station information, in-station location information, and device information for executing the method;

[0022] The site comprehensive service information includes one or more of the following: site surrounding information, convenience information, navigation information, and lost and found information.

[0023] Optionally, display answers to the passenger, including:

[0024] presenting the answer to the passenger via one or more presentation methods;

[0025] The display modes include voice display mode, text display mode and image display mode.

[0026] Optionally, capture questions asked by passengers, including:

[0027] Obtaining questions from passengers through one or more acquisition methods;

[0028] Among them, the acquisition methods are: touch screen acquisition, sound acquisition device acquisition, image acquisition device acquisition, and virtual assistant acquisition.

[0029] In a second aspect of the present application, a rail transit station information question-and-answer device is provided, the device comprising:

[0030] An acquisition module, used to acquire questions raised by passengers;

[0031] A retrieval module, used to determine the semantic information vector of the question obtained by the acquisition module through the RAG model, and retrieve the content block corresponding to the semantic information vector in the vector database; wherein the content block stored in the vector database is obtained based on the device meta information document describing the location information within the station and the site information document describing the comprehensive service information of the site; each site corresponds to a device meta information document and a site information document;

[0032] A generation module is used to integrate the corresponding content blocks obtained by the retrieval module to generate answers;

[0033] The display module is used to show the answers generated by the generation module to passengers.

[0034] Optionally, the device is located inside the physical robot, or the device is located on the physical robot, or the device is located outside the physical robot and is connected to the physical robot by wire or wirelessly;

[0035] Physical robots are deployed at key locations of rail transit stations.

[0036] In a third aspect of the present application, an electronic device is provided, the electronic device being a physical robot, wherein the physical robot is deployed at a key position of a rail transit station;

[0037] The electronic device comprises:

[0038] Memory;

[0039] Processor; and

[0040] Computer programs;

[0041] The computer program is stored in the memory and is configured to be executed by the processor to implement the method as described in the first aspect above.

[0042] In a fourth aspect of the present application, a computer-readable storage medium is provided, wherein the device where the computer-readable storage medium is located is a physical robot, wherein the physical robot is deployed at a key position of a rail transit station;

[0043] The computer-readable storage medium stores a computer program; the computer program is executed by the processor to implement the method as described in the first aspect above.

[0044] The present application provides a method, device, equipment, and storage medium for questioning and answering station information of rail transit, the method comprising: obtaining questions raised by passengers; determining the semantic information vector of the question through the RAG model, and retrieving the content block corresponding to the semantic information vector in the vector database; wherein the content block stored in the vector database is obtained based on the device meta information document describing the location information within the station and the site information document describing the comprehensive service information of the site; each site corresponds to a device meta information document and a site information document; integrating the corresponding content blocks to generate answers; and displaying the answers to passengers. The method provided in the present application can retrieve and generate answers in real time through the RAG model, ensuring that the answers provided are both accurate and timely. In addition, the method provided in the present application does not require downloading and installation when implemented, which improves the convenience of questioning and answering, and can meet the diverse query needs of passengers. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0046] Figure 1 A schematic diagram of a flow chart of a rail transit station information question and answer method provided in an embodiment of the present application;

[0047] Figure 2 A schematic diagram of an implementation architecture of a rail transit station information question and answer method provided in an embodiment of the present application;

[0048] Figure 3 A schematic diagram of the structure of a rail transit station information question and answer device provided in an embodiment of the present application;

[0049] Figure 4 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0050] In order to make the technical solutions and advantages in the embodiments of the present application more clearly understood, the exemplary embodiments of the present application are further described in detail below in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than an exhaustive list of all the embodiments. It should be noted that the embodiments in the present application and the features in the embodiments can be combined with each other without conflict.

[0051] In the process of realizing this application, the inventor found that with the acceleration of urbanization, rail transit has become an indispensable part of urban residents' daily travel, and its service quality directly affects the quality of life of citizens. However, in high-traffic areas such as subway stations, passengers often face problems such as finding exits, transfer routes, and nearby service facilities.

[0052] In the prior art, rail transit subway station information query systems usually adopt the following methods:

[0053] 1. Traditional information query system: provides information through manually compiled paper or electronic maps, signboards, etc., but the information is not updated in a timely manner and the interactivity is poor.

[0054] 2. Voice assistant: Some subway stations are equipped with voice assistants, but their functions are limited and cannot meet the diverse query needs of passengers.

[0055] 3. Mobile applications: Some subway companies have developed mobile applications to provide real-time information query services, but passengers are required to download and install them, which is inconvenient to use.

[0056] In response to the above problems, the present application provides a method, device, equipment, and storage medium for rail transit station information question and answer, the method comprising: obtaining questions raised by passengers; determining the semantic information vector of the question through the RAG model, and retrieving the content block corresponding to the semantic information vector in the vector database; wherein the content block stored in the vector database is obtained based on the device meta information document describing the location information within the station and the site information document describing the comprehensive service information of the site; each site corresponds to a device meta information document and a site information document; integrating the corresponding content blocks to generate answers; and displaying the answers to passengers. The method provided in the present application can retrieve and generate answers in real time through the RAG model, ensuring that the answers provided are both accurate and timely. In addition, the method provided in the present application does not require downloading and installation when implemented, which improves the convenience of question and answer and can meet the diverse query needs of passengers.

[0057] like Figure 1 As shown, this embodiment provides a rail transit station information question and answer method, and the implementation process of the method is as follows:

[0058] 101, get questions asked by passengers.

[0059] In specific implementation, the questions raised by the passengers can be obtained through one or more acquisition methods.

[0060] Among them, the acquisition methods are: touch screen acquisition, sound acquisition device acquisition, image acquisition device acquisition, and virtual assistant acquisition.

[0061] If the passenger asks a question by voice, then in step 101 the question provided by the passenger is collected by a sound collection device, and then the collected question is converted into text by a speech recognition module, and the text is used as the question raised by the passenger obtained in step 101.

[0062] If the passenger asks a question through an image, the question provided by the passenger is collected by an image acquisition device in step 101, and then the collected question is converted into text by an image recognition module. The text is used as the question raised by the passenger obtained in step 101.

[0063] 102. Determine the semantic information vector of the question through a RAG (Retrieval-Augmented Generation) model, and retrieve the content block corresponding to the semantic information vector in a vector database.

[0064] Among them, the content blocks stored in the vector database are obtained based on the device meta-information document describing the in-station location information and the site information document describing the site comprehensive service information. That is to say, before executing step 102, a plurality of content blocks will be stored in the vector database, some of which are obtained based on the in-station location information, and some of which are obtained based on the site comprehensive service information. The storage process is: 1) Obtain the device meta-information document describing the in-station location information and the site information document describing the site comprehensive service information. 2) Divide the device meta-information document and the site information document into information blocks (wherein an information block can be a paragraph, a sentence, or a whole representing an information), encode the information blocks into high-dimensional vectors, each high-order vector is an information block, and store the information blocks (i.e., high-dimensional vectors) in the vector database.

[0065] Among them, the in-station location information includes but is not limited to one or more of the following: station information (such as site facility descriptions, etc.), in-station location information (such as what channels, exits or service desks exist in each direction of the device itself, etc.) and equipment information for executing the rail transit site information Q&A method provided in this embodiment (such as equipment usage guidance, etc.).

[0066] The comprehensive service information of the station includes one or more of the following: information about the station's surroundings (such as buildings near the station (such as a subway entrance)), convenience information (such as convenience facilities and services in the station), navigation information (such as the nearest exit, station bus information, and transfer routes at the station), lost and found information, etc.

[0067] In addition, since the in-station location information and the comprehensive service information of the site will be different for each station, the in-station location information and the comprehensive service information of the site will be stored separately for each station, that is, each station corresponds to a device meta information document and a site information document. In the specific implementation, the content block corresponding to the semantic information vector can be retrieved from the content block corresponding to the site where the device that executes the rail transit site information question and answer method provided by this embodiment is located, thereby ensuring the accuracy of the answer. Among them, the content block corresponding to the site where the device that executes the rail transit site information question and answer method provided by this embodiment is located is a content block obtained based on the device meta information document and the site information document of the site where the device that executes the rail transit site information question and answer method provided by the embodiment is located.

[0068] In addition, in order to ensure the timeliness of the information stored in the vector database, all content blocks will be located and maintained. For example, the device meta information documents of each site (such as updating site facility descriptions, equipment information, etc.) and site information documents (such as updating nearby location information, lost and found information, etc.) are regularly maintained, and then based on the updated device meta information documents and site information documents, updated content blocks are obtained, and then the content blocks stored in the vector database are updated.

[0069] The information stored in the vector database is regularly updated according to the station's update mechanism (such as regularly maintaining location information near each station, lost and found information, etc.). Through this update mechanism, the continuous updating of the vector database is guaranteed, further improving the reliability of the information.

[0070] When executing step 102, the question obtained in step 101 can be first semantically understood through the natural language processing module to obtain the semantic information of the question, and then the semantic information of the question can be encoded into a vector through the RAG model to obtain a semantic information vector. Among the content blocks stored in the vector database, the distance between the vector of each content block and the semantic information vector is calculated, and the N content blocks with the smaller distance are determined as the content blocks corresponding to the semantic information vector. Therefore, the content blocks corresponding to the semantic information vector can all be content blocks obtained based on the device meta information document describing the location information within the site; they can also all be content blocks obtained based on the site information document describing the comprehensive service information of the site; they can also be partly content blocks obtained based on the device meta information document describing the location information within the site, and partly content blocks obtained based on the site information document describing the comprehensive service information of the site.

[0071] In addition, the rail transit station information question and answer method provided in this embodiment also provides map module access capability. When executed, the map module can provide map image information of the searched content near the current station based on the question, which can be used as an illustration of the answer. The map module can obtain the map image information and / or navigation information corresponding to the question by itself, or it can link to a third-party navigation service product to obtain the map image information and / or navigation information corresponding to the question output by the third-party navigation service product.

[0072] That is, if the question raised by the passenger involves map information (such as the question is related to the geographic location), the map image information and / or navigation information corresponding to the question is obtained through the map module. For example, if the question raised by the passenger obtained in step 101 involves map information, the question is input into the map module, and the map module analyzes the question. The corresponding map image information (i.e., map image information) searched near the station can be used as an illustration of the answer, or the navigation map information (i.e., navigation information) outside the station obtained by the map module can be used as the answer to the question.

[0073] 103. Integrate the corresponding content blocks to generate answers.

[0074] If the map image information and / or navigation information corresponding to the question is also obtained, then in step 103, the corresponding content block, the corresponding map image information and / or navigation information are integrated to generate an answer.

[0075] For example, the corresponding content block obtained in step 102 and the question raised by the passenger obtained in step 101 are input into the pre-trained integrated large model together (if there is map image information and / or navigation information corresponding to the question, the map image information and / or navigation information corresponding to the question will also be input into the pre-trained integrated large model), and the integrated large model will generate a reasonable answer based on the question and related content.

[0076] 104, show the answer to the passenger.

[0077] In specific implementation, the answers can be displayed to passengers through one or more display methods.

[0078] Among them, the display methods include voice display method, text display method, and image display method.

[0079] For example, the answer is converted into speech and played to the passenger.

[0080] The rail transit station information question and answer method provided in this embodiment can accurately provide users with the required information services, including but not limited to travel guidance, station navigation, surrounding services, etc., based on the user's voice inquiry, using the station location information, station comprehensive service information and data provided by the map module, thereby improving the information query experience of users in the subway station. The rail transit station information question and answer method provided in this embodiment significantly improves the understanding and response quality of questions and answers by introducing RAG technology, providing users with a more convenient and personalized service experience.

[0081] Although there are some intelligent question-and-answer systems for specific scenarios in the prior art, they often lack a deep understanding of and response capabilities for specific physical environments (such as subway stations). The rail transit station information question-and-answer method provided in this embodiment focuses more on specific applications in the rail transit environment. It not only provides more accurate and personalized services, but also enhances the method's understanding ability and response quality by introducing the RAG model.

[0082] The RAG model question-and-answer system in the prior art is usually aimed at general fields, while the rail transit station information question-and-answer method provided in this embodiment is targeted at the rail transit field and has stronger professionalism and pertinence. In addition, the rail transit station information question-and-answer method provided in this embodiment combines the document records regularly maintained by the rail transit company to ensure the timeliness and accuracy of the question-and-answer content. The rail transit station information question-and-answer method provided in this embodiment clearly uses a combination of text and graphics to display the answers, which can improve the user experience.

[0083] The rail transit station information question and answer method provided in this embodiment can solve the problems of untimely information update, limited service time and insufficient complex query processing capabilities existing in the prior art. The rail transit station information question and answer method provided in this embodiment can realize real-time data retrieval and generation by introducing the RAG model, ensuring that the information provided is both accurate and timely. At the same time, the rail transit station information question and answer method provided in this embodiment ensures the continuous update of the vector database through the update mechanism, further improving the reliability of the information.

[0084] In the specific implementation, questions raised by passengers can be obtained through a virtual assistant, thereby realizing the combination of the rail transit station information question and answer method provided in this embodiment and a virtual assistant, which can better adapt to service needs in different scenarios and greatly improve the user experience.

[0085] The rail transit station information question and answer method provided in this embodiment can improve the professionalism and accuracy of the rail transit field question and answer system, adapt to field updates, and provide a convenient interaction method. At the same time, the rail transit station information question and answer method provided in this embodiment supports voice interaction, making the operation easier and suitable for users of different ages.

[0086] Taking a passenger asking a question by voice as an example, the implementation process of the method provided in this embodiment is described again. Figure 2 As shown, in step 101, the passenger's question will be converted into text through the voice recognition module. In step 102, the question is semantically understood through the natural language processing module to obtain the semantic information of the question, the semantic information vector of the question is determined through the RAG model, and the content block corresponding to the semantic information vector is retrieved in the vector database, and the image and text description information of the location are obtained through the map module. In step 103, the corresponding content blocks are integrated to generate answers. In step 104, the answer is converted into voice and fed back to the passenger. In addition, the station location information and station comprehensive service information of each station are regularly maintained according to the station's update mechanism.

[0087] In specific implementation, the device (such as a physical robot) that executes the site information question and answer method for rail transit provided in this embodiment can be deployed at key locations (such as entrances, exits and transfer areas) of the site (such as subway entrances). The device has a built-in RAG model and a vector database, and the vector database pre-stores content blocks obtained based on the device meta-information document of the site and content blocks obtained based on the site information document of the site. The device can interact with passengers through voice recognition. When the user asks for relevant information through voice, the device can quickly call the RAG model through the method provided in this embodiment, combined with the data of the map module, to provide the user with detailed answers, including map image information and / or navigation information corresponding to the question. Answer common questions about route guidance, transfer information, etc. within the station. At the same time, the device is also equipped with a touch screen for passengers to enter text inquiries, which increases the flexibility of use.

[0088] In addition, the device can also have a remote update function, allowing rail transit companies to regularly update the site's device meta information documents and site information documents through the network, so that the vector database can update its stored content blocks according to the updated device meta information documents and site information documents, ensuring the timeliness of the content blocks stored in the vector database. In addition, the device can also automatically optimize service strategies and improve service quality based on user usage habits and feedback.

[0089] In addition, when the device fails to work for some reason, the virtual assistant can automatically take over the service to ensure the integrity of the service. The virtual assistant obtains questions raised by passengers through the QR code on the device and sends answers to the passengers' mobile devices. The virtual assistant is also based on the RAG model and can provide services to passengers through mobile applications without physical equipment. This method is particularly suitable for periods of high passenger flow and can effectively alleviate the pressure on equipment (such as physical robots) that execute the rail transit site information question and answer method provided in this embodiment. At the same time, the virtual assistant can also collect feedback information from passengers to optimize the service strategy of equipment (such as physical robots) that executes the rail transit site information question and answer method provided in this embodiment. The document can be jointly maintained by passengers and there is an audit mechanism.

[0090] In order to adapt to different usage scenarios, the device can be equipped with multiple language support capabilities to meet the needs of international tourists. At the same time, the device can also provide personalized services to frequent customers through facial recognition technology, such as remembering the frequent customers' common routes and preferences and providing more considerate navigation suggestions. At the same time, the device can also integrate more sensors, such as cameras and infrared sensors, to enhance its environmental perception capabilities and provide richer services.

[0091] The rail transit station information question and answer method provided in this embodiment can realize question and answer of various entity direction information based on the station location based on the RAG enhancement mechanism of the station location, such as answering the station direction information of the current passenger's location, such as what channels are on the right or left, or in front of the left of the device (such as a physical robot) executing the rail transit station information question and answer method provided in this embodiment, and where to go, all of which can provide customers with more humane answer results.

[0092] The multi-language support capability in the rail transit station information question and answer method provided in this embodiment can support multiple languages ​​to meet the needs of different passengers.

[0093] The real-time update mechanism of the rail transit station information question and answer method provided in this embodiment can regularly update the information stored in the vector database according to the station's update mechanism.

[0094] The intelligent navigation function of the rail transit station information question and answer method provided in this embodiment can be combined with a map module to provide intelligent navigation services for passengers.

[0095] The rail transit station information question-and-answer method provided in this embodiment continuously optimizes service strategies and improves user experience by collecting and analyzing user behavior data.

[0096] This embodiment provides a method for questioning and answering station information of rail transit, which obtains questions raised by passengers; determines the semantic information vector of the question through the RAG model, and retrieves the content block corresponding to the semantic information vector in the vector database; wherein the content block stored in the vector database is obtained based on the device meta information document describing the location information within the station and the site information document describing the comprehensive service information of the site; each site corresponds to a device meta information document and a site information document; integrates the corresponding content blocks to generate answers; and displays the answers to passengers. The method provided in this embodiment can retrieve and generate answers in real time through the RAG model, ensuring that the answers provided are both accurate and timely. In addition, this method does not require downloading and installation when implemented, which improves the convenience of questioning and answering, and can meet the diverse query needs of passengers.

[0097] Based on the same inventive concept of the rail transit station information question and answer method, this embodiment provides a rail transit station information question and answer device, see Figure 3 , the device comprises:

[0098] The acquisition module 301 is used to acquire questions raised by passengers.

[0099] The retrieval module 302 is used to determine the semantic information vector of the question obtained by the acquisition module 301 through the RAG model, and retrieve the content block corresponding to the semantic information vector in the vector database. The content block stored in the vector database is obtained based on the device meta information document describing the location information in the station and the site information document describing the comprehensive service information of the site; each site corresponds to a device meta information document and a site information document.

[0100] The generation module 303 is used to integrate the corresponding content blocks obtained by the retrieval module 302 to generate an answer.

[0101] The display module 304 is used to display the answers generated by the generation module 303 to the passengers.

[0102] The device further includes: a map answer acquisition module, which is used to acquire map image information and / or navigation information corresponding to the question through the map module.

[0103] The generation module 303 is used to integrate the corresponding content blocks, the corresponding map image information and / or navigation information to generate an answer.

[0104] The device further includes: an information storage module, which is used to divide the device meta information document and the site information document into information blocks, encode the information blocks into high-dimensional vectors, and store the high-dimensional vectors in a vector database.

[0105] The in-station location information includes one or more of the following: station information, in-station location information, and equipment information of the execution method.

[0106] The site comprehensive service information includes one or more of the following: site surrounding information, convenience information, navigation information, and lost and found information.

[0107] The display module 304 is used to display the answers to the passengers through one or more display methods.

[0108] The display modes include voice display mode, text display mode and image display mode.

[0109] The acquisition module 301 is used to acquire questions raised by passengers through one or more acquisition methods.

[0110] Among them, the acquisition methods are: touch screen acquisition, sound acquisition device acquisition, image acquisition device acquisition, and virtual assistant acquisition.

[0111] The station information question-and-answer device for rail transit provided in this embodiment can accurately provide users with the required information services, including but not limited to travel guidance, station navigation, surrounding services, etc., based on the user's voice inquiry, using the station location information, station comprehensive service information, and data provided by the map module. The station information question-and-answer device for rail transit provided in this embodiment significantly improves the understanding and response quality of the question-and-answer robot by introducing the RAG model, providing users with a more convenient and personalized service experience.

[0112] In a specific implementation, the device provided in this embodiment may be located inside a physical robot (e.g., the physical robot is the device provided in this embodiment), or, the device is located on the physical robot (e.g., the device provided in this embodiment is installed on the body of the physical robot), or, the device is located outside the physical robot and is connected to the physical robot by wire or wirelessly (e.g., the physical robot is connected to the device provided in this embodiment via a wireless network connection).

[0113] The physical robot is deployed at key locations of rail transit stations. For example, the physical robot can be deployed at key locations (such as entrances, exits, and transfer areas) of rail transit stations (such as subway entrances).

[0114] The device provided in this embodiment has a built-in RAG model and a vector database, and the vector database pre-stores content blocks obtained based on the device meta information document of the site and content blocks obtained based on the site information document of the site. The physical robot can interact with passengers through voice recognition. When the user asks for relevant information through voice, the physical robot can quickly call the RAG model through the device provided in this embodiment, and provide the user with detailed answers in combination with the data of the map module, including map image information and / or navigation information corresponding to the question. Answer common questions about route guidance within the station, transfer information, etc. At the same time, the physical robot is also equipped with a touch screen for passengers to enter text inquiries, which increases the flexibility of use.

[0115] In addition, the physical robot can also have a remote update function, allowing the rail transit company to regularly update the equipment meta information documents and site information documents of the site through the network, so that the vector database can update its stored content blocks according to the updated equipment meta information documents and site information documents, ensuring the timeliness of the content blocks stored in the vector database. In addition, the physical robot can also automatically optimize service strategies and improve service quality based on user usage habits and feedback.

[0116] In addition, when the physical robot is unable to work for some reason, the virtual assistant can automatically take over the service to ensure the integrity of the service. The virtual assistant obtains questions raised by passengers through the QR code on the physical robot, and sends answers to the passenger's mobile physical robot. The virtual assistant also uses the device provided in this embodiment, based on the RAG model, to provide services to passengers through mobile applications in the absence of a physical robot. This method is particularly suitable for periods of high passenger flow and can effectively relieve the pressure on physical robots. At the same time, the virtual assistant can also collect feedback information from passengers to optimize the service strategy of the physical robot. The documents can be jointly maintained by passengers, and there is an audit mechanism.

[0117] In order to adapt to different usage scenarios, the physical robot can be equipped with multi-language support capabilities to meet the needs of international tourists. At the same time, the physical robot can also provide personalized services to frequent customers through facial recognition technology, such as remembering the frequent customers' common routes and preferences and providing more considerate navigation suggestions. At the same time, the physical robot can also integrate more sensors, such as cameras and infrared sensors, to enhance its environmental perception capabilities and provide richer services.

[0118] The physical robot can use the device provided in this embodiment to realize questions and answers based on the direction information of each entity in the station based on the RAG enhancement mechanism of the station location. For example, it can answer the station direction information of the current passenger's location, such as what channels are on the right or left, or in front of the left of the physical robot, and where to go. The results can provide customers with more humane answers.

[0119] The multilingual support capability of this physical robot can support multiple languages ​​to meet the needs of different passengers.

[0120] The real-time update mechanism of the physical robot can periodically update the information stored in the vector database according to the update mechanism of the station through the device provided in this embodiment.

[0121] The intelligent navigation function of the physical robot can provide intelligent navigation services for passengers through the device provided in this embodiment combined with a map module.

[0122] This physical robot continuously optimizes service strategies and improves user experience by collecting and analyzing user behavior data.

[0123] The device provided in this embodiment can retrieve and generate answers in real time through the RAG model, ensuring that the answers provided are both accurate and timely. In addition, the method provided in this application does not require downloading and installation during implementation, which improves the convenience of question and answering and can meet the diverse query needs of passengers.

[0124] Based on the same inventive concept of the method for questioning and answering station information of rail transit, this embodiment provides an electronic device, which is a physical robot, wherein the physical robot is deployed at a key position of a rail transit station.

[0125] The electronic equipment Figure 4 As shown, it includes: a memory 401, a processor 402, and a computer program.

[0126] The computer program is stored in the memory 401 and is configured to be executed by the processor 402 to implement the above-mentioned rail transit station information question and answer method.

[0127] Specifically,

[0128] Get questions from passengers.

[0129] The semantic information vector of the question is determined through the RAG model, and the content block corresponding to the semantic information vector is retrieved in the vector database. Among them, the content blocks stored in the vector database are obtained based on the device meta information document describing the location information within the station and the site information document describing the comprehensive service information of the site. Each site corresponds to a device meta information document and a site information document.

[0130] Integrate the corresponding content blocks to generate answers.

[0131] Show the answer to the passenger.

[0132] Optionally, before integrating the corresponding content blocks and generating the answer, the following steps may also be included:

[0133] Obtain map image information and / or navigation information corresponding to the question through the map module.

[0134] Integrate the corresponding content blocks and generate answers, including:

[0135] Integrate the corresponding content blocks, corresponding map image information and / or navigation information to generate answers.

[0136] Optionally, before determining the semantic information vector of the question by using the RAG model and searching the vector database for the content block corresponding to the semantic information vector, the method further includes:

[0137] The device meta information document and the site information document are divided into information blocks.

[0138] Encode information blocks into high-dimensional vectors.

[0139] Store high-dimensional vectors in a vector database.

[0140] Optionally, the in-station location information includes one or more of the following: station information, in-station location information, and equipment information of the execution method.

[0141] The site comprehensive service information includes one or more of the following: site surrounding information, convenience information, navigation information, and lost and found information.

[0142] Optionally, display answers to the passenger, including:

[0143] The answer is presented to the passenger through one or more presentation methods.

[0144] The display modes include voice display mode, text display mode and image display mode.

[0145] Optionally, capture questions asked by passengers, including:

[0146] The questions asked by the passengers are obtained through one or more obtaining methods.

[0147] Among them, the acquisition methods are: touch screen acquisition, sound acquisition device acquisition, image acquisition device acquisition, and virtual assistant acquisition.

[0148] The electronic device provided in this embodiment can accurately provide the user with the required information services, including but not limited to travel guidance, station navigation, surrounding services, etc., based on the user's voice inquiry, using the station location information, station comprehensive service information, and data provided by the map module. The electronic device provided in this embodiment significantly improves the understanding and response quality of the question-answering robot by introducing the RAG model, providing users with a more convenient and personalized service experience.

[0149] In a specific implementation, the electronic device provided in this embodiment may be a physical robot. The physical robot may be deployed at key locations of rail transit stations, such as key locations (such as entrances, exits, and transfer areas) of rail transit stations (such as subway entrances).

[0150] The physical robot has a built-in RAG model and a vector database, and the vector database pre-stores content blocks obtained based on the device meta information document of the site and content blocks obtained based on the site information document of the site. The physical robot can interact with passengers through voice recognition. When the user asks for relevant information through voice, the physical robot can quickly call the RAG model and provide users with detailed answers, including map image information and / or navigation information corresponding to the question, combined with the data of the map module. Answer common questions about route guidance within the station, transfer information, etc. At the same time, the physical robot is also equipped with a touch screen for passengers to enter text queries, which increases the flexibility of use.

[0151] In addition, the physical robot can also have a remote update function, allowing the rail transit company to regularly update the equipment meta information documents and site information documents of the site through the network, so that the vector database can update its stored content blocks according to the updated equipment meta information documents and site information documents, ensuring the timeliness of the content blocks stored in the vector database. In addition, the physical robot can also automatically optimize service strategies and improve service quality based on user usage habits and feedback.

[0152] In addition, if the physical robot cannot work for some reason, the virtual assistant can automatically take over the service to ensure the integrity of the service. The virtual assistant obtains the questions raised by the passengers through the QR code on the physical robot and sends the answers to the passenger's mobile physical robot. The virtual assistant is also based on the RAG model and can provide services to passengers through mobile applications in the absence of physical robots. This method is particularly suitable for periods of high passenger flow and can effectively relieve the pressure on physical robots. At the same time, the virtual assistant can also collect feedback from passengers to optimize the service strategy of the physical robot. The documents can be jointly maintained by passengers and there is an audit mechanism.

[0153] In order to adapt to different usage scenarios, the physical robot can be equipped with multi-language support capabilities to meet the needs of international tourists. At the same time, the physical robot can also provide personalized services to frequent customers through facial recognition technology, such as remembering the frequent customers' common routes and preferences and providing more considerate navigation suggestions. At the same time, the physical robot can also integrate more sensors, such as cameras and infrared sensors, to enhance its environmental perception capabilities and provide richer services.

[0154] The physical robot's RAG enhancement mechanism based on the station's location can realize questions and answers about the direction information of various entities based on the station's location. For example, it can answer the station's directions for the current passenger's location, such as what passages are to the right or left, or to the left front of the physical robot, and which way to go. The results can provide customers with more humane answers.

[0155] The multilingual support capability of this physical robot can support multiple languages ​​to meet the needs of different passengers.

[0156] The real-time update mechanism of the physical robot can regularly update the information stored in the vector database according to the station's update mechanism.

[0157] The intelligent navigation function of this physical robot can be combined with a map module to provide intelligent navigation services for passengers.

[0158] This physical robot continuously optimizes service strategies and improves user experience by collecting and analyzing user behavior data.

[0159] The electronic device provided in this embodiment has a computer program executed by a processor so that answers can be retrieved and generated in real time through a RAG model, ensuring that the answers provided are both accurate and timely. In addition, the method provided in this application does not require downloading and installation during implementation, which improves the convenience of question and answering and can meet the diverse query needs of passengers.

[0160] Based on the same inventive concept of the rail transit station information question and answer method, this embodiment provides a computer-readable storage medium, and the device where the computer-readable storage medium is located is a physical robot, wherein the physical robot is deployed at a key position of a rail transit station.

[0161] The computer readable storage medium stores a computer program, which is executed by a processor to implement the above-mentioned rail transit station information question and answer method.

[0162] Specifically,

[0163] Get questions from passengers.

[0164] The semantic information vector of the question is determined through the RAG model, and the content block corresponding to the semantic information vector is retrieved in the vector database. Among them, the content blocks stored in the vector database are obtained based on the device meta information document describing the location information within the station and the site information document describing the comprehensive service information of the site. Each site corresponds to a device meta information document and a site information document.

[0165] Integrate the corresponding content blocks to generate answers.

[0166] Show the answer to the passenger.

[0167] Optionally, before integrating the corresponding content blocks and generating the answer, the following steps may also be included:

[0168] Obtain map image information and / or navigation information corresponding to the question through the map module.

[0169] Integrate the corresponding content blocks and generate answers, including:

[0170] Integrate the corresponding content blocks, corresponding map image information and / or navigation information to generate answers.

[0171] Optionally, before determining the semantic information vector of the question by using the RAG model and searching the vector database for the content block corresponding to the semantic information vector, the method further includes:

[0172] The device meta information document and the site information document are divided into information blocks.

[0173] Encode information blocks into high-dimensional vectors.

[0174] Store high-dimensional vectors in a vector database.

[0175] Optionally, the in-station location information includes one or more of the following: station information, in-station location information, and equipment information of the execution method.

[0176] The site comprehensive service information includes one or more of the following: site surrounding information, convenience information, navigation information, and lost and found information.

[0177] Optionally, display answers to the passenger, including:

[0178] The answer is presented to the passenger through one or more presentation methods.

[0179] The display modes include voice display mode, text display mode and image display mode.

[0180] Optionally, capture questions asked by passengers, including:

[0181] The questions asked by the passengers are obtained through one or more obtaining methods.

[0182] Among them, the acquisition methods are: touch screen acquisition, sound acquisition device acquisition, image acquisition device acquisition, and virtual assistant acquisition.

[0183] The computer-readable storage medium provided in this embodiment can accurately provide users with the required information services, including but not limited to travel guidance, station navigation, surrounding services, etc., based on the user's voice inquiry, using the station location information, station comprehensive service information, and data provided by the map module. The station information question-and-answer device for rail transit provided in this embodiment significantly improves the understanding and response quality of the question-and-answer robot by introducing the RAG model, providing users with a more convenient and personalized service experience.

[0184] In a specific implementation, the device where the computer-readable storage medium provided in this embodiment is located may be a physical robot. The physical robot may be deployed at key locations of rail transit stations, such as key locations (such as entrances, exits, and transfer areas) of rail transit stations (such as subway entrances).

[0185] The physical robot has a built-in RAG model and a vector database, and the vector database pre-stores content blocks obtained based on the device meta information document of the site and content blocks obtained based on the site information document of the site. The physical robot can interact with passengers through voice recognition. When the user asks for relevant information through voice, the physical robot can quickly call the RAG model and provide users with detailed answers, including map image information and / or navigation information corresponding to the question, combined with the data of the map module. Answer common questions about route guidance within the station, transfer information, etc. At the same time, the physical robot is also equipped with a touch screen for passengers to enter text queries, which increases the flexibility of use.

[0186] In addition, the physical robot can also have a remote update function, allowing the rail transit company to regularly update the equipment meta information documents and site information documents of the site through the network, so that the vector database can update its stored content blocks according to the updated equipment meta information documents and site information documents, ensuring the timeliness of the content blocks stored in the vector database. In addition, the physical robot can also automatically optimize service strategies and improve service quality based on user usage habits and feedback.

[0187] In addition, if the physical robot cannot work for some reason, the virtual assistant can automatically take over the service to ensure the integrity of the service. The virtual assistant obtains the questions raised by the passengers through the QR code on the physical robot and sends the answers to the passenger's mobile physical robot. The virtual assistant is also based on the RAG model and can provide services to passengers through mobile applications in the absence of physical robots. This method is particularly suitable for periods of high passenger flow and can effectively relieve the pressure on physical robots. At the same time, the virtual assistant can also collect feedback from passengers to optimize the service strategy of the physical robot. The documents can be jointly maintained by passengers and there is an audit mechanism.

[0188] In order to adapt to different usage scenarios, the physical robot can be equipped with multi-language support capabilities to meet the needs of international tourists. At the same time, the physical robot can also provide personalized services to frequent customers through facial recognition technology, such as remembering the frequent customers' common routes and preferences and providing more considerate navigation suggestions. At the same time, the physical robot can also integrate more sensors, such as cameras and infrared sensors, to enhance its environmental perception capabilities and provide richer services.

[0189] The physical robot's RAG enhancement mechanism based on the station's location can realize questions and answers about the direction information of various entities based on the station's location. For example, it can answer the station's directions for the current passenger's location, such as what passages are to the right or left, or to the left front of the physical robot, and which way to go. The results can provide customers with more humane answers.

[0190] The multilingual support capability of this physical robot can support multiple languages ​​to meet the needs of different passengers.

[0191] The real-time update mechanism of the physical robot can regularly update the information stored in the vector database according to the station's update mechanism.

[0192] The intelligent navigation function of this physical robot can be combined with a map module to provide intelligent navigation services for passengers.

[0193] This physical robot continuously optimizes service strategies and improves user experience by collecting and analyzing user behavior data.

[0194] The computer-readable storage medium provided in this embodiment has a computer program thereon which is executed by a processor to retrieve and generate answers in real time through a RAG model, thereby ensuring that the answers provided are both accurate and timely. In addition, the method provided in this application does not require downloading and installation during implementation, thereby improving the convenience of question and answering and meeting the diverse query needs of passengers.

[0195] Those skilled in the art will appreciate that the embodiments of the present application can be provided as methods, systems, or computer program products. Therefore, the present application can adopt the form of complete hardware embodiments, complete software embodiments, or embodiments in combination with software and hardware. Moreover, the present application can adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code. The scheme in the embodiments of the present application can be implemented in various computer languages, for example, object-oriented programming language Java and literal scripting language JavaScript, etc.

[0196] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0197] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0198] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0199] Although the preferred embodiments of the present application have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications falling within the scope of the present application.

[0200] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is also intended to include these modifications and variations.

Claims

1. A method for questioning and answering station information of rail transit, characterized in that: The method comprises: Get questions from passengers; Determine the semantic information vector of the problem through the RAG model, and retrieve the content block corresponding to the semantic information vector in the vector database; wherein the content block stored in the vector database is obtained based on the device meta information document describing the location information within the station and the site information document describing the comprehensive service information of the site; each site corresponds to a device meta information document and a site information document; Integrate the corresponding content blocks to generate answers; The answer is displayed to the passenger.

2. The method according to claim 1, characterized in that The content blocks corresponding to the integration, before generating the answer, also include: Obtaining map image information and / or navigation information corresponding to the question through a map module; The integration of the corresponding content blocks to generate answers includes: Integrate the corresponding content blocks, corresponding map image information and / or navigation information to generate answers.

3. The method according to claim 1, characterized in that Before determining the semantic information vector of the question by using the RAG model and searching the content block corresponding to the semantic information vector in the vector database, the method further includes: Splitting the device meta information document and the site information document into information blocks; Encode information blocks into high-dimensional vectors; The high-dimensional vectors are stored in the vector database.

4. The method according to claim 1, characterized in that: The in-station location information includes one or more of the following: station information, in-station location information, and device information for executing the method; The site comprehensive service information includes one or more of the following: site surrounding information, convenience information, navigation information, and lost and found information.

5. The method according to claim 1, characterized in that The step of displaying the answer to the passenger comprises: presenting the answer to the passenger through one or more presentation methods; The display method includes voice display method, text display method and image display method.

6. The method according to claim 1, characterized in that The questions asked by passengers include: Obtaining questions from passengers through one or more acquisition methods; Among them, the acquisition methods are: touch screen acquisition, sound acquisition device acquisition, image acquisition device acquisition, and virtual assistant acquisition.

7. A station information question and answer device for rail transit, characterized in that: The device comprises: An acquisition module, used to acquire questions raised by passengers; A retrieval module, used to determine the semantic information vector of the question obtained by the acquisition module through the RAG model, and retrieve the content block corresponding to the semantic information vector in the vector database; wherein the content block stored in the vector database is obtained based on the device meta information document describing the location information within the station and the site information document describing the comprehensive service information of the site; each site corresponds to a device meta information document and a site information document; A generation module, used to integrate the corresponding content blocks obtained by the retrieval module to generate an answer; A display module is used to display the answer generated by the generation module to the passenger.

8. The device according to claim 7, characterized in that The device is located inside the physical robot, or the device is located on the physical robot, or the device is located outside the physical robot and is connected to the physical robot by wire or wirelessly; The physical robot is deployed at key locations of rail transit stations.

9. An electronic device, characterized in that: The electronic device is a physical robot, wherein the physical robot is deployed at a key position of a rail transit station; The electronic device comprises: Memory; Processor; and Computer programs; The computer program is stored in the memory and is configured to be executed by the processor to implement the method according to any one of claims 1 to 6.

10. A computer-readable storage medium, characterized in that: The device where the computer-readable storage medium is located is a physical robot, wherein the physical robot is deployed at a key position of a rail transit station; The computer-readable storage medium stores a computer program; the computer program is executed by a processor to implement the method according to any one of claims 1 to 6.

Images