Intelligent park knowledge question and answer method and device based on large language model

Abstract

The present disclosure relates to a smart park knowledge question and answer method and device based on a large language model, an electronic device and a storage medium, and relates to the technical field of large language models, which can automatically diagnose the abnormality of park equipment based on multidimensional data and improve the intelligence level of the park. The method comprises: in response to a triggering operation of an abnormal function of equipment in a smart park knowledge question and answer page, displaying a first subpage, and in response to a question about the abnormal reason of a target equipment in the smart park input by a user in the first subpage, retrieving real-time equipment parameters and historical maintenance records of the target equipment and obtaining data of environmental sensors around the target equipment in a knowledge base; constructing a prompt word based on the question, the equipment parameters, the historical maintenance records and the data of the environmental sensors; inputting the prompt word into a large language model to obtain an abnormal diagnosis report for the target equipment, and displaying the abnormal diagnosis report on the first subpage.

Description

Technical Field

[0001] This disclosure relates to the field of natural language processing technology, specifically to a smart park knowledge question answering method, device, electronic device, and storage medium based on a large language model. Background Technology

[0002] A park is a comprehensive area formed by the aggregation of spaces around a specific function. It typically encompasses facilities such as offices, production, research and development, warehousing, and living services, serving the collaborative operation of enterprises, institutions, or individuals. Common types include industrial parks (such as manufacturing bases), science and technology parks (such as innovation business incubators), logistics parks (such as warehousing and transportation hubs), and campus parks (such as universities or corporate headquarters). Their core value lies in reducing operating costs, improving collaborative efficiency, and providing a platform for regional economic development through resource concentration and functional complementarity.

[0003] However, traditional industrial parks are centered on physical space and mainly rely on manual management and basic information technology. Their operation model is relatively rigid, the data interoperability between systems is poor, and there are problems such as delayed service response, which makes it difficult to meet the needs of modern enterprises for efficiency and intelligence. Summary of the Invention

[0004] To overcome the problems existing in related technologies, this disclosure provides a smart park knowledge question answering method, device, electronic device, storage medium and program product based on a large language model, in order to solve the defects in related technologies.

[0005] According to a first aspect of the present disclosure, a smart park knowledge question-answering method based on a large language model is provided, comprising:

[0006] A smart park knowledge Q&A page based on a large language model is displayed on the terminal device, wherein the smart park knowledge Q&A page is used to allow users to ask questions related to smart parks in at least one of the following ways: images, voice and text.

[0007] In response to the triggering operation of the device malfunction function on the smart park knowledge Q&A page, a first subpage for the user to ask questions about the device malfunction is displayed. In response to the question entered by the user on the first subpage asking about the cause of the malfunction of the target device in the smart park, the real-time device parameters and historical maintenance records of the target device are retrieved from the knowledge base, and data from the environmental sensors around the target device are obtained.

[0008] Based on the problem of the target equipment's abnormality, the equipment parameters, the historical maintenance records, and the data from the environmental sensors, prompt words are constructed. These prompt words are used to prompt the large language model to generate an abnormality diagnosis report for the target equipment based on the problem of the target equipment's abnormality, the equipment parameters, the historical maintenance records, and the data from the environmental sensors.

[0009] The prompt words are input into the large language model to obtain an anomaly diagnosis report for the target device, and the anomaly diagnosis report is displayed on the first subpage.

[0010] In one embodiment, the method further includes:

[0011] In response to the triggering of the maintenance operation function on the smart park knowledge Q&A page, a second subpage is displayed for the user to ask maintenance operation questions.

[0012] In response to the user's input of a maintenance operation question regarding the abnormal operation of the target qua device in the second sub-page, the user's physiological information is obtained from the user's wearable device, and the environmental information of the user's environment is obtained through the sensors on the terminal device;

[0013] Based on the physiological information and the environmental information, the user's stress state is determined. When the stress state indicates that the user is in a high-pressure state, the large language model generates operation steps information to instruct the user to perform abnormal maintenance on the target equipment based on the abnormal diagnosis report, the maintenance operation problem, the physiological information, and the environmental information.

[0014] The operation steps information is displayed on the second subpage and played in the form of voice.

[0015] In one embodiment, the method further includes:

[0016] In response to the triggering operation of the visitor function on the smart park knowledge Q&A page, a third subpage for the visitor is displayed, and in response to the user's input operation on the third subpage, emotion recognition is performed on the content corresponding to the input operation to obtain the user's first emotion information, wherein the content includes voice content and / or text content.

[0017] The user's facial expressions and body language are acquired through the camera on the terminal device, and the user's second emotional information is identified based on the facial expressions and body language.

[0018] Based on the first emotion information and the second emotion information, the target emotion information of the user is determined, and based on the target emotion information, the visitor type of the user is determined;

[0019] Based on the visitor type, the large language model generates introductory information about the smart park and guidance information for the user, and displays the introductory information and guidance information on the third subpage.

[0020] In one embodiment, generating introductory information about the smart park and smart park guidance information for the user based on the visitor type using the large language model includes:

[0021] When the visitor type indicates that the user is a hurried visitor, the first introduction information of the smart park and the first smart park guidance information for the user are generated by the big language model. The first smart park guidance information is navigation information generated by the big language model to guide the user to a first location. The first location is the park location that the user expects to reach, which is determined by the big language model based on the content corresponding to the input operation.

[0022] When the visitor type indicates that the user is a leisure visitor, the big language model generates a second introductory information for the smart park and a second smart park guidance information for the user. The second introductory information contains more content than the first introductory information. The second smart park guidance information is navigation information generated by the big language model to guide the user to a second location, which is a park location that the big language model predicts the user is interested in.

[0023] In one embodiment, displaying the smart park knowledge Q&A page based on a large language model on the terminal device includes:

[0024] In response to a vehicle user scanning a QR code at the target entrance of the smart park, a smart park knowledge Q&A page based on a large language model is displayed on the terminal device held by the vehicle user.

[0025] The method further includes:

[0026] In response to the triggering operation of the parking function on the smart park knowledge Q&A page, a fourth subpage is displayed for the vehicle user to ask questions about parking in the park.

[0027] In response to the parking navigation question entered by the vehicle user on the fourth subpage, the system uses the large language model to query the underground parking information of the smart park for the nearest available parking space based on the target park location in the parking navigation question, and then uses the large language model to generate parking navigation information based on the available parking space and the location of the target entrance.

[0028] The parking navigation information is displayed on the fourth subpage and played in voice format.

[0029] In one embodiment, the method further includes:

[0030] In response to the triggering of the emergency drill function on the smart park knowledge Q&A page, a fifth subpage is displayed for the user to ask questions about the emergency drill.

[0031] In response to the user's inquiry about the emergency drill plan entered on the fifth subpage, the system determines the real-time environmental information and real-time operating information of the equipment in the smart park through sensors in the smart park, determines the real-time weather information of the smart park through a meteorological data interface, and acquires real-time video of the smart park from cameras in the smart park. Based on the real-time video, the system determines the real-time pedestrian flow information and real-time traffic information of the smart park. The real-time environmental information includes the temperature and / or humidity of the smart park.

[0032] Based on the real-time environmental information, real-time operational information, real-time weather information, real-time pedestrian flow information, and real-time traffic information, the large language model generates an emergency drill plan for the smart park.

[0033] The emergency drill plan is displayed on the fifth subpage.

[0034] In one embodiment, the method further includes:

[0035] In response to the triggering operation of the work status function on the smart park knowledge Q&A page, a sixth subpage is displayed for the user to inquire about the work status of the staff in the smart park;

[0036] In response to a question entered on the sixth subpage inquiring about the work status of a specific staff member, the system acquires the facial expressions and body language of the specific staff member through the camera at the specific staff member's location, and identifies the third emotional information of the specific staff member based on the facial expressions and body language using the large language model.

[0037] When the third emotional information indicates that the specific staff member is in a low mood, the large language model generates emotion regulation suggestions, wherein the emotion regulation suggestions include recommended relaxation places within a preset range and / or music for soothing emotions;

[0038] The emotion regulation suggestions are displayed on the sixth subpage and sent to the terminal device held by the specific staff member.

[0039] According to a second aspect of the present disclosure, a smart park knowledge question-answering device based on a large language model is provided, comprising:

[0040] The first display module is used to display a smart park knowledge Q&A page based on a large language model on a terminal device, wherein the smart park knowledge Q&A page is used for users to ask questions related to smart parks in at least one of the following ways: images, voice and text.

[0041] The second display module is used to respond to the triggering operation of the device malfunction function on the smart park knowledge Q&A page, display a first subpage for the user to ask about the device malfunction, and respond to the question entered by the user on the first subpage asking about the cause of the target device malfunction in the smart park, retrieve the real-time device parameters and historical maintenance records of the target device in the knowledge base, and obtain the data of the environmental sensors around the target device.

[0042] A construction module is used to construct prompt words based on the problem of the target device's abnormality, the device parameters, the historical maintenance records, and the data from the environmental sensors. The prompt words are used to prompt the large language model to generate an abnormality diagnosis report for the target device based on the problem of the target device's abnormality, the device parameters, the historical maintenance records, and the data from the environmental sensors.

[0043] The third display module is used to input the prompt words into the large language model to obtain an anomaly diagnosis report for the target device, and to display the anomaly diagnosis report on the first sub-page.

[0044] According to a third aspect of the present disclosure, an electronic device is provided, the electronic device including a memory and a processor, the memory being configured to store computer instructions executable on the processor, and the processor being configured to implement the method described in any one of the first aspects when executing the computer instructions.

[0045] According to a fourth aspect of the present disclosure, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the method described in any of the first aspects.

[0046] The technical solutions provided by the embodiments of this disclosure may include the following beneficial effects:

[0047] The smart park knowledge-based question-and-answer method based on a large language model provided in this disclosure supports users in asking questions related to the smart park using at least one of images, voice, and text, i.e., it supports multimodal input, thereby improving input accuracy and consequently the accuracy of the knowledge-and-answer results. Secondly, the smart park knowledge-and-answer page provides a device anomaly function. Responding to a user's question about the cause of a target device malfunction entered on the first subpage corresponding to this function, the method retrieves real-time device parameters and historical maintenance records of the target device from the knowledge base, as well as data from environmental sensors surrounding the target device. The large language model can then combine this multi-dimensional data (i.e., real-time device parameters, historical maintenance records, and environmental sensor data) to generate an anomaly diagnosis report for the target device, thereby improving the accuracy of anomaly diagnosis and the accuracy of the park knowledge-and-answer results. Furthermore, the large language model enables automatic anomaly diagnosis of equipment in the park, eliminating the need for manual on-site anomaly diagnosis, thus improving the efficiency of park anomaly diagnosis and enhancing the intelligence level of the smart park. Attached Figure Description

[0048] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with the invention and, together with the description, serve to explain the principles of the invention.

[0049] Figure 1 This is a flowchart illustrating an exemplary embodiment of the present disclosure of a smart park knowledge question answering method based on a large language model;

[0050] Figure 2 This is a schematic diagram illustrating the structure of a smart park knowledge question-answering device based on a large language model, as shown in an exemplary embodiment of this disclosure.

[0051] Figure 3 This is a structural block diagram of an electronic device illustrated in an exemplary embodiment of the present disclosure. Detailed Implementation

[0052] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numerals in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this disclosure. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this disclosure as detailed in the appended claims.

[0053] The terminology used in this disclosure is for the purpose of describing particular embodiments only and is not intended to be limiting of the disclosure. The singular forms “a,” “the,” and “the” as used in this disclosure and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used herein refers to and includes any and all possible combinations of one or more of the associated listed items.

[0054] It should be understood that although the terms first, second, third, etc., may be used in this disclosure to describe various information, such information should not be limited to these terms. These terms are used only to distinguish information of the same type from one another. For example, without departing from the scope of this disclosure, first information may also be referred to as second information, and similarly, second information may also be referred to as first information. Depending on the context, the word "if" as used herein may be interpreted as "when," "when," or "in response to determination."

[0055] As the background technology states, traditional industrial parks are centered on physical space and mainly rely on manual management and basic information technology. Their operation model is relatively rigid, the data interoperability between systems is poor, and there are problems such as delayed service response, making it difficult to meet the needs of modern enterprises for efficiency and intelligence.

[0056] For example, the inventors discovered that in equipment fault diagnosis scenarios, traditional industrial parks typically require manual on-site fault diagnosis. In large parks, reaching the location of the fault takes considerable time, and manual fault diagnosis usually relies on human experience for on-site inspection. Therefore, equipment fault diagnosis in traditional industrial parks is inefficient and cannot meet the demands of modern enterprises for high efficiency and intelligent systems.

[0057] Based on this, in a first aspect, at least one embodiment of this disclosure provides a smart park knowledge question answering method based on a large language model. Please refer to... Figure 1 The diagram illustrates the process of the method, including steps S101 to S104.

[0058] In step S101, a smart park knowledge Q&A page based on a large language model is displayed on the terminal device. This smart park knowledge Q&A page allows users to ask questions related to the smart park using at least one of the following methods: images, voice, and text.

[0059] In step S102, in response to the triggering operation of the device malfunction function on the smart park knowledge Q&A page, a first subpage for the user to ask questions about device malfunctions is displayed. In response to the question entered by the user on the first subpage asking about the cause of the target device malfunction in the smart park, the real-time device parameters and historical maintenance records of the target device are retrieved from the knowledge base, and data from the environmental sensors around the target device are obtained.

[0060] For example, the knowledge base is used to store real-time equipment parameters and historical maintenance records of various devices in the smart park. Environmental sensors around the target equipment may include, for example, wind pressure sensors, humidity sensors, temperature sensors, etc., but this disclosure does not limit the scope of these sensors.

[0061] For example, in this embodiment of the disclosure, a user can upload images of device malfunctions, then extract image features using a ResNet-50 structure, search in a vector library, associate them with device maintenance video clips, and obtain the device's historical maintenance records.

[0062] In step S103, a prompt word is constructed based on the problem causing the target device's malfunction, the device parameters, the historical maintenance records, and the data from the environmental sensors. This prompt word prompts the large language model to generate an anomaly diagnosis report for the target device based on the problem causing the malfunction, the device parameters, the historical maintenance records, and the data from the environmental sensors.

[0063] In step S104, the prompt word is input into the large language model to obtain an anomaly diagnosis report for the target device, and the anomaly diagnosis report is displayed on the first sub-page.

[0064] It should be understood that the smart park knowledge Q&A page in this embodiment provides multiple functions for smart parks, such as equipment malfunction function, maintenance operation function, visitor function, parking function, emergency drill function, and working status function. This embodiment does not limit these functions, and each of the aforementioned functions will be described in detail below.

[0065] For example, in the case of equipment malfunction, in response to the user's voice question "Possible reasons for insufficient airflow from the air conditioner in Building 3" entered on the first subpage, the system can retrieve real-time equipment parameters and historical maintenance records of the air conditioner fan from the knowledge base, as well as data from environmental sensors around the air conditioner fan. Then, based on the question "Possible reasons for insufficient airflow from the air conditioner in Building 3," the equipment parameters, historical maintenance records, and environmental sensor data, a prompt is constructed. Finally, the prompt is input into a large language model, resulting in an anomaly diagnosis report generated by the large language model for the target equipment: "Filter blockage probability 72%, immediate cleaning recommended."

[0066] Therefore, the large language model can combine multi-dimensional data (i.e., real-time equipment parameters, historical maintenance records, and environmental sensor data) to generate anomaly diagnostic reports for target equipment, thereby improving the accuracy of anomaly diagnosis and the accuracy of knowledge-based question-and-answer results within the park. Furthermore, the large language model can automatically diagnose anomalies in equipment within the park, eliminating the need for manual on-site diagnosis, thus improving the efficiency of anomaly diagnosis and enhancing the overall intelligence of the smart park. Moreover, it supports multimodal user input, thereby improving input accuracy and ultimately enhancing the accuracy of knowledge-based question-and-answer results.

[0067] In one embodiment, in response to a triggering operation of the maintenance operation function on the smart park knowledge Q&A page, a second subpage for the user to ask maintenance operation questions can be displayed; in response to the user's input of a maintenance operation question regarding the abnormal operation of the target device on the second subpage, the user's physiological information is obtained from the user's wearable device, and the environmental information of the user's environment is obtained through sensors on the terminal device; based on the physiological information and the environmental information, the user's stress state is determined, and if the stress state indicates that the user is in a high-pressure state, the large language model generates operation step information to instruct the user to perform abnormal maintenance on the target device based on the abnormal diagnosis report, the maintenance operation question, the physiological information, and the environmental information; the operation step information is displayed on the second subpage and played in the form of voice.

[0068] The user's stress state can be a numerical value representing the degree of stress. Correspondingly, a high-pressure state can be a value that is higher than a preset threshold.

[0069] It should be understood that fault repair in traditional industrial parks typically relies on the experience of maintenance personnel. If these personnel are under stress, the repair results may be affected. However, in this embodiment, the stress level of the maintenance personnel can be identified. If the personnel are under stress, a large language model is used to assist them in fault repair, thereby ensuring the repair effect.

[0070] For example, in response to a repair technician's input of a repair operation question regarding an air conditioner malfunction on the second subpage, physiological information (such as heart rate) is obtained from the technician's wristband, and environmental information (such as noise level, workspace illumination, and temperature) is obtained from sensors on the terminal device. Then, based on the physiological and environmental information, the technician's stress state is determined. If the stress state indicates the technician is under high pressure, a large language model is used to generate operational steps instructing the technician to perform repairs on the air conditioner based on the air conditioner's malfunction diagnosis report, the repair operation question, the physiological information, and the environmental information.

[0071] For example, if the anomaly diagnosis report indicates a fault in the air conditioner's fan bearing, the generated operation steps could be: "You now need: ① a blue-handled Allen wrench (in the second layer of the tool kit) ② a red torque gauge (with a yellow label). Note! The current screw is secured with anti-loosening adhesive; you need to first gently turn it clockwise 15° to remove the adhesive layer. Now loosen the first screw (located at the 3 o'clock position on the north side). Insert a feeler gauge into the bearing clearance; when a 0.3mm thin gauge cannot be inserted, the clearance is within the acceptable range. Now begin the inspection, and say 'Inspection complete' upon completion." The operation steps generated by the large language model can include spatial positioning descriptions and feature identifiers, decompose complex actions, provide spatial orientation guidance to maintenance personnel, quantify operation standards, and add voice confirmation nodes. Combined with voice playback, this better assists maintenance personnel in troubleshooting and improves repair efficiency.

[0072] Using the above example, if there is no operation within a preset time (e.g., 30 seconds) after a certain step is guided by voice, it can automatically switch to a flashing graphic reminder (e.g., a red arrow pointing to the part to be operated is displayed through the maintenance glasses worn by the maintenance personnel). Alternatively, when an incorrect tool is detected (e.g., using an 8mm wrench to tighten a 10mm screw), a prompt can be generated through the large language model: Stop operation! The current screw requires a 10mm wrench. The correct tool is on the third layer of the tool cart to your left rear, and this prompt can be played by voice.

[0073] In one embodiment, in response to a triggering operation of the visitor function on the smart park knowledge Q&A page, a third subpage for visitors can be displayed. In response to the user's input on the third subpage, emotion recognition is performed on the content corresponding to the input to obtain the user's first emotion information, wherein the content includes voice content and / or text content. The user's facial expressions and body language are acquired through a camera on the terminal device, and the user's second emotion information is identified based on the expressions and body language. Based on the first and second emotion information, the user's target emotion information is determined, and based on the target emotion information, the user's visitor type is determined. Based on the visitor type using the large language model, introductory information about the smart park and smart park guidance information for the user are generated, and the introductory information and smart park guidance information are displayed on the third subpage.

[0074] In other words, when visitors enter the park, the reception method can be quickly adjusted based on the visitors' input, facial expressions, and body language.

[0075] In one embodiment, generating introductory information for the smart park and smart park guidance information for the user based on the visitor type using the large language model includes: when the visitor type indicates that the user is a hurried visitor, generating first introductory information for the smart park and first smart park guidance information for the user using the large language model, wherein the first smart park guidance information is navigation information generated by the large language model to guide the user to a first location, the first location being the park location that the user expects to reach, determined by the large language model based on the content corresponding to the input operation; when the visitor type indicates that the user is a leisure visitor, generating second introductory information for the smart park and second smart park guidance information for the user using the large language model, wherein the second introductory information has more content than the first introductory information, and the second smart park guidance information is navigation information generated by the large language model to guide the user to a second location, the second location being the park location that the user is interested in, predicted by the large language model.

[0076] In other words, for visitors who appear to be in a hurry, the large language model can concisely provide key information and quick guidance; for visitors with more time, the large language model can provide more detailed park introductions and interactive experience recommendations. Thus, different reception methods can be automatically provided for different types of visitors, improving visitor service effectiveness.

[0077] For example, the initial introduction information may include the name of the park, the approximate location and purpose of the main functional areas (such as office areas, leisure areas, exhibition areas, etc.), and the direction and distance of important facilities around the visitor's current location, such as restrooms, elevators, and exits. For instance, the initial introduction information could be generated using a large language model as follows: "You are currently at the main entrance of the smart park. The office building reception is 50 meters ahead. There is a coffee shop on the left side of the building, a restroom on the right side, and the elevator is located in the center of the office building lobby."

[0078] For example, the first smart park guidance information can enable visitors to reach their destination most conveniently, and can also provide some key node markers or reference points to help visitors quickly locate themselves. For example, "Please walk straight along this main road, turn left at the second intersection, and the meeting room you are going to is 100 meters ahead. You will see a large fountain as a marker along the way." In addition, the first smart park guidance information can be displayed in a concise and clear way, such as displaying a route map on the terminal device screen, or providing direction and distance prompts through voice prompts. This disclosure embodiment does not limit this approach.

[0079] For example, in addition to the usual introduction to the park layout and functional areas, the second set of introductory information can also include background information such as the park's development history, design concept, advanced technologies used, and awards received. For instance, it can explain how the park utilizes an intelligent energy management system to achieve energy conservation and emission reduction, or how the buildings within the park are designed and constructed in accordance with green and environmentally friendly principles. It can also showcase the park's corporate culture, employee activities, and other content, allowing visitors to gain a more comprehensive understanding of the park as a whole.

[0080] For example, the guidance information in a second smart park could be personalized interactive experience recommendations based on a large language model, tailored to the visitor's interests and available time. For instance, if a visitor is interested in technology, they could be recommended to visit the park's intelligent robot R&D center, watch robot demonstrations and demonstrations, and even experience simple interactive communication with the robots. Alternatively, they could be recommended to participate in the park's technology lectures, innovation workshops, and other activities, allowing them to deeply engage with and experience the park's technological atmosphere and innovative capabilities. For visitors who enjoy culture and art, they could be recommended to visit art exhibitions and creative studios within the park, and interactive art creation experiences such as pottery making and painting could be offered. It should be understood that if a visitor is entering the park for the first time, their interests can be predicted based on the interests of most visitors; if a visitor is not a first-timer, their interests can be predicted based on their past visit information.

[0081] In one embodiment, displaying a smart park knowledge Q&A page based on a large language model on a terminal device includes: in response to a vehicle user scanning a QR code at the target entrance of the smart park, displaying the smart park knowledge Q&A page based on a large language model on the terminal device held by the vehicle user. Correspondingly, in response to a triggering operation of the parking function on the smart park knowledge Q&A page, a fourth subpage for the vehicle user to inquire about parking issues in the park can also be displayed; in response to a parking navigation question entered by the vehicle user on the fourth subpage, the large language model queries the underground parking information of the smart park for the nearest available parking space based on the target park location in the parking navigation question, and generates parking navigation information based on the available parking space and the location of the target entrance using the large language model; the parking navigation information is displayed on the fourth subpage and played in voice format.

[0082] It should be understood that first-time visitors often face parking difficulties, mainly due to the park's large size, making it difficult to locate parking lots correctly; even if a parking lot is found, finding an available space can be challenging. Additionally, park parking is typically underground, where GPS signals are weak, rendering most navigation software unusable. Therefore, even if a visitor's companion finds an available space first, the visitor may still be unable to drive to that space correctly.

[0083] In this embodiment, users can scan a QR code at the entrance of the smart park to access a smart park knowledge Q&A page. On this page, users can further activate the parking function to search for available parking spaces on a fourth sub-page. After finding an available parking space, a large language model can be used to generate parking navigation information based on the location of the available parking space and the target entrance. This navigation information can be displayed as an image on the fourth sub-page, showing the navigation route, or played as audio, including landmarks along the way.

[0084] For example, the above parking navigation information could be: "Hello! Welcome to the underground parking lot. Please follow the route to the available parking space: Go straight for 50 meters from the entrance: Turn left at the first intersection. You will see a yellow sign for 'Area A Elevator Entrance' on the left wall; After passing the elevator entrance: Continue straight for about 100 meters. Look for the green fire hydrant (number F-03) on the right pillar; Turn right at the green fire hydrant: Enter the B area passage. There is a blue arrow marked 'B2-08' on the ground. Follow the direction of the arrow; After passing two rows of pillars: A fluorescent sign for 'Charging Pile Area' hangs above. Turn left to reach the parking space (number B2-12)."

[0085] In one embodiment, in response to the triggering operation of the emergency drill function on the smart park knowledge Q&A page, a fifth subpage for the user to ask questions about the emergency drill can be displayed; in response to the user's question about the emergency drill plan entered on the fifth subpage, the real-time environmental information and real-time operating information of the equipment in the smart park are determined by sensors in the smart park, the real-time weather information of the smart park is determined by a meteorological data interface, and the real-time video of the smart park is acquired from cameras in the smart park. Based on the real-time video, the real-time pedestrian flow information and real-time traffic information of the smart park are determined, wherein the real-time environmental information includes the temperature and / or humidity of the smart park; an emergency drill plan for the smart park is generated by the large language model based on the real-time environmental information, the real-time operating information, the real-time weather information, the real-time pedestrian flow information, and the real-time traffic information; and the emergency drill plan is displayed on the fifth subpage.

[0086] In other words, by combining real-time environmental, meteorological, pedestrian, and traffic information, as well as the real-time operational information of equipment within the smart park, a large language model can be used to automatically generate emergency drill plans for the park, reducing the manpower and time spent in the process. Furthermore, because the plans are generated using real-time park information, they can be dynamically adjusted, thereby improving their practicality.

[0087] In one embodiment, in response to a triggering operation of the work status function on the smart park knowledge Q&A page, a sixth subpage can be displayed for the user to inquire about the work status of staff in the smart park; in response to a question about the work status of a specific staff member entered on the sixth subpage, the facial expressions and body language of the specific staff member can be acquired through a camera at the specific staff member's location, and the third emotion information of the specific staff member can be identified based on the facial expressions and body language using the large language model; when the third emotion information indicates that the specific staff member is in a low mood, an emotion regulation suggestion can be generated through the large language model, wherein the emotion regulation suggestion includes recommended relaxation places within a preset range and / or music for soothing emotions; the emotion regulation suggestion is displayed on the sixth subpage and sent to the terminal device held by the specific staff member.

[0088] It should be understood that facial expressions and body language can be converted into text and input into a large language model, which can then identify the emotional information of specific staff members. When this emotional information indicates that a specific staff member is feeling down, the large language model can further generate emotion regulation suggestions, such as recommending nearby relaxation venues or playing soothing music. In this way, deep emotional interaction can be achieved with staff in the park through multiple modalities such as voice and facial expressions, enhancing the park's intelligence.

[0089] According to a second aspect of the embodiments of this disclosure, a smart park knowledge question-answering device 200 based on a large language model is provided. Please refer to... Figure 2 The device 200 includes:

[0090] The first display module 201 is used to display a smart park knowledge Q&A page based on a large language model on a terminal device, wherein the smart park knowledge Q&A page is used for users to ask questions related to smart parks in at least one of the following ways: pictures, voice and text.

[0091] The second display module 202 is used to respond to the triggering operation of the device abnormal function in the smart park knowledge Q&A page, display a first subpage for the user to ask about the device abnormality problem, and respond to the question entered by the user on the first subpage asking about the cause of the target device abnormality in the smart park, retrieve the real-time device parameters and historical maintenance records of the target device in the knowledge base, and obtain the data of the environmental sensors around the target device.

[0092] The construction module 203 is used to construct prompt words based on the problem of the target device's abnormality, the device parameters, the historical maintenance records, and the data from the environmental sensors. The prompt words are used to prompt the large language model to generate an abnormality diagnosis report for the target device based on the problem of the target device's abnormality, the device parameters, the historical maintenance records, and the data from the environmental sensors.

[0093] The third display module 204 is used to input the prompt words into the large language model to obtain an anomaly diagnosis report for the target device, and to display the anomaly diagnosis report on the first sub-page.

[0094] In some embodiments of this disclosure, the device 200 further includes a maintenance operation module for:

[0095] In response to the triggering of the maintenance operation function on the smart park knowledge Q&A page, a second subpage is displayed for the user to ask maintenance operation questions.

[0096] In response to the user's input of a maintenance operation question regarding the abnormal operation of the target device on the second sub-page, the user's physiological information is obtained from the user's wearable device, and the environmental information of the user's environment is obtained through the sensors on the terminal device;

[0097] Based on the physiological information and the environmental information, the user's stress state is determined. When the stress state indicates that the user is in a high-pressure state, the large language model generates operation steps information to instruct the user to perform abnormal maintenance on the target equipment based on the abnormal diagnosis report, the maintenance operation problem, the physiological information, and the environmental information.

[0098] The operation steps information is displayed on the second subpage and played in the form of voice.

[0099] In some embodiments of this disclosure, the device 200 further includes a visitor module for:

[0100] In response to the triggering operation of the visitor function on the smart park knowledge Q&A page, a third subpage for the visitor is displayed, and in response to the user's input operation on the third subpage, emotion recognition is performed on the content corresponding to the input operation to obtain the user's first emotion information, wherein the content includes voice content and / or text content.

[0101] The user's facial expressions and body language are acquired through the camera on the terminal device, and the user's second emotional information is identified based on the facial expressions and body language.

[0102] Based on the first emotion information and the second emotion information, the target emotion information of the user is determined, and based on the target emotion information, the visitor type of the user is determined;

[0103] Based on the visitor type, the large language model generates introductory information about the smart park and guidance information for the user, and displays the introductory information and guidance information on the third subpage.

[0104] In some embodiments of this disclosure, the visitor module is specifically used for:

[0105] When the visitor type indicates that the user is a hurried visitor, the first introduction information of the smart park and the first smart park guidance information for the user are generated by the big language model. The first smart park guidance information is navigation information generated by the big language model to guide the user to a first location. The first location is the park location that the user expects to reach, which is determined by the big language model based on the content corresponding to the input operation.

[0106] When the visitor type indicates that the user is a leisure visitor, the big language model generates a second introductory information for the smart park and a second smart park guidance information for the user. The second introductory information contains more content than the first introductory information. The second smart park guidance information is navigation information generated by the big language model to guide the user to a second location, which is a park location that the big language model predicts the user is interested in.

[0107] In some embodiments of this disclosure, the first display module 201 is specifically used for:

[0108] In response to a vehicle user scanning a QR code at the target entrance of the smart park, a smart park knowledge Q&A page based on a large language model is displayed on the terminal device held by the vehicle user.

[0109] The device 200 further includes a parking module for:

[0110] In response to the triggering operation of the parking function on the smart park knowledge Q&A page, a fourth subpage is displayed for the vehicle user to ask questions about parking in the park.

[0111] In response to the parking navigation question entered by the vehicle user on the fourth subpage, the system uses the large language model to query the underground parking information of the smart park for the nearest available parking space based on the target park location in the parking navigation question, and then uses the large language model to generate parking navigation information based on the available parking space and the location of the target entrance.

[0112] The parking navigation information is displayed on the fourth subpage and played in voice format.

[0113] In some embodiments of this disclosure, the device 200 further includes an emergency drill module for:

[0114] In response to the triggering of the emergency drill function on the smart park knowledge Q&A page, a fifth subpage is displayed for the user to ask questions about the emergency drill.

[0115] In response to the user's inquiry about the emergency drill plan entered on the fifth subpage, the system determines the real-time environmental information and real-time operating information of the equipment in the smart park through sensors in the smart park, determines the real-time weather information of the smart park through a meteorological data interface, and acquires real-time video of the smart park from cameras in the smart park. Based on the real-time video, the system determines the real-time pedestrian flow information and real-time traffic information of the smart park. The real-time environmental information includes the temperature and / or humidity of the smart park.

[0116] Based on the real-time environmental information, real-time operational information, real-time weather information, real-time pedestrian flow information, and real-time traffic information, the large language model generates an emergency drill plan for the smart park.

[0117] The emergency drill plan is displayed on the fifth subpage.

[0118] In some embodiments of this disclosure, the device 200 further includes an operating status module for:

[0119] In response to the triggering operation of the work status function on the smart park knowledge Q&A page, a sixth subpage is displayed for the user to inquire about the work status of the staff in the smart park;

[0120] In response to a question entered on the sixth subpage inquiring about the work status of a specific staff member, the system acquires the facial expressions and body language of the specific staff member through the camera at the specific staff member's location, and identifies the third emotional information of the specific staff member based on the facial expressions and body language using the large language model.

[0121] When the third emotional information indicates that the specific staff member is in a low mood, the large language model generates emotion regulation suggestions, wherein the emotion regulation suggestions include recommended relaxation places within a preset range and / or music for soothing emotions;

[0122] The emotion regulation suggestions are displayed on the sixth subpage and sent to the terminal device held by the specific staff member.

[0123] Regarding the apparatus in the above embodiments, the specific manner in which each module performs its operation has been described in detail in the embodiments of the method in the first aspect, and will not be elaborated upon here.

[0124] According to a third aspect of the embodiments of this disclosure, please refer to the appendix. Figure 3The diagram illustrates an exemplary block diagram of an electronic device 700, which may include a processor 701 and a memory 702. The electronic device 700 may also include one or more of a multimedia component 703, an input / output (I / O) interface 704, and a communication component 705.

[0125] The processor 701 controls the overall operation of the electronic device 700 to complete all or part of the steps in any of the methods described above. The memory 702 stores various types of data to support the operation of the electronic device 700. This data may include, for example, instructions for any application or method operating on the electronic device 700, and application-related data, such as the aforementioned anomaly diagnostic reports, operation step information, smart park introduction information, and smart park guidance information. The memory 702 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read-Only Memory (EPROM), Programmable Read-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. Multimedia component 703 may include a screen and an audio component. The screen may be, for example, a touchscreen, and the audio component is used to output and / or input audio signals. For example, the audio component may include a microphone for receiving external audio signals. The received audio signals may be further stored in memory 702 or transmitted via communication component 705. The audio component also includes at least one speaker for outputting audio signals. I / O interface 704 provides an interface between processor 701 and other interface modules, such as a keyboard, mouse, buttons, etc. These buttons may be virtual or physical buttons. Communication component 705 is used for wired or wireless communication between the electronic device 700 and other devices. Wireless communication may include Wi-Fi, Bluetooth, Near Field Communication (NFC), 2G, 3G, or 4G, or a combination of these. Therefore, the corresponding communication component 705 may include a Wi-Fi module, a Bluetooth module, or an NFC module.

[0126] In an exemplary embodiment, the electronic device 700 may be implemented by one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field-programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components to perform any of the methods described above.

[0127] In another exemplary embodiment, a computer-readable storage medium including program instructions is also provided, which, when executed by a processor, implement the steps of any of the methods described above. For example, the computer-readable storage medium may be the memory 702 including the program instructions described above, which may be executed by the processor 701 of the electronic device 700 to perform any of the methods described above.

[0128] In another exemplary embodiment, a computer program product is also provided, which includes a computer program executable by a processor, wherein the computer program, when executed by the processor, implements the steps of any of the methods described above.

[0129] The preferred embodiments of this disclosure have been described in detail above with reference to the accompanying drawings. However, this disclosure is not limited to the specific details of the above embodiments. Within the scope of the technical concept of this disclosure, various simple modifications can be made to the technical solutions of this disclosure, and these simple modifications all fall within the protection scope of this disclosure.

[0130] It should also be noted that the various specific technical features described in the above specific embodiments can be combined in any suitable manner without contradiction. In order to avoid unnecessary repetition, this disclosure will not describe the various possible combinations separately.

[0131] Furthermore, various different embodiments of this disclosure can be combined in any way, as long as they do not violate the spirit of this disclosure, they should also be regarded as the content disclosed in this disclosure.

Claims

1. A knowledge-based question-answering method for smart parks based on a large language model, characterized in that, include: A smart park knowledge Q&A page based on a large language model is displayed on the terminal device, wherein the smart park knowledge Q&A page is used to allow users to ask questions related to smart parks in at least one of the following ways: images, voice and text. In response to the triggering operation of the device malfunction function on the smart park knowledge Q&A page, a first subpage for the user to ask questions about the device malfunction is displayed. In response to the question entered by the user on the first subpage asking about the cause of the malfunction of the target device in the smart park, the real-time device parameters and historical maintenance records of the target device are retrieved from the knowledge base, and data from the environmental sensors around the target device are obtained. Based on the problem of the target equipment's abnormality, the equipment parameters, the historical maintenance records, and the data from the environmental sensors, prompt words are constructed. These prompt words are used to prompt the large language model to generate an abnormality diagnosis report for the target equipment based on the problem of the target equipment's abnormality, the equipment parameters, the historical maintenance records, and the data from the environmental sensors. The prompt words are input into the large language model to obtain an anomaly diagnosis report for the target device, and the anomaly diagnosis report is displayed on the first subpage. In response to the triggering of the maintenance operation function on the smart park knowledge Q&A page, a second subpage is displayed for the user to ask maintenance operation questions. In response to the user's input of a maintenance operation question regarding the abnormal operation of the target device on the second sub-page, the user's physiological information is obtained from the user's wearable device, and the environmental information of the user's environment is obtained through the sensors on the terminal device; Based on the physiological information and the environmental information, the user's stress state is determined. When the stress state indicates that the user is in a high-pressure state, the large language model generates operation steps information to instruct the user to perform abnormal maintenance on the target equipment based on the abnormal diagnosis report, the maintenance operation problem, the physiological information, and the environmental information. The operation step information is displayed on the second subpage and played in the form of voice. In response to the triggering operation of the visitor function on the smart park knowledge Q&A page, a third subpage for the visitor is displayed, and in response to the user's input operation on the third subpage, emotion recognition is performed on the content corresponding to the input operation to obtain the user's first emotion information, wherein the content includes voice content and / or text content. The user's facial expressions and body language are acquired through the camera on the terminal device, and the user's second emotional information is identified based on the facial expressions and body language. Based on the first emotion information and the second emotion information, the target emotion information of the user is determined, and based on the target emotion information, the visitor type of the user is determined, wherein the visitor type is used to characterize whether the user is a busy visitor or a leisure visitor. Based on the visitor type, the large language model generates introductory information about the smart park and guidance information for the user, and displays the introductory information and guidance information on the third subpage.

2. The smart park knowledge question-and-answer method according to claim 1, characterized in that, The step of generating introductory information about the smart park and guidance information for the user based on the visitor type using the large language model includes: When the visitor type indicates that the user is a hurried visitor, the first introduction information of the smart park and the first smart park guidance information for the user are generated by the big language model. The first smart park guidance information is navigation information generated by the big language model to guide the user to a first location. The first location is the park location that the user expects to reach, which is determined by the big language model based on the content corresponding to the input operation. When the visitor type indicates that the user is a leisure visitor, the big language model generates a second introductory information for the smart park and a second smart park guidance information for the user. The second introductory information contains more content than the first introductory information. The second smart park guidance information is navigation information generated by the big language model to guide the user to a second location, which is a park location that the big language model predicts the user is interested in.

3. The smart park knowledge question-and-answer method according to claim 1 or 2, characterized in that, The display of the smart park knowledge Q&A page based on a large language model on the terminal device includes: In response to a vehicle user scanning a QR code at the target entrance of the smart park, a smart park knowledge Q&A page based on a large language model is displayed on the terminal device held by the vehicle user. The method further includes: In response to the triggering operation of the parking function on the smart park knowledge Q&A page, a fourth subpage is displayed for the vehicle user to ask questions about parking in the park. In response to the parking navigation question entered by the vehicle user on the fourth subpage, the system uses the large language model to query the underground parking information of the smart park for the nearest available parking space based on the target park location in the parking navigation question, and then uses the large language model to generate parking navigation information based on the available parking space and the location of the target entrance. The parking navigation information is displayed on the fourth subpage and played in voice format.

4. The smart park knowledge question-and-answer method according to claim 1 or 2, characterized in that, The method further includes: In response to the triggering of the emergency drill function on the smart park knowledge Q&A page, a fifth subpage is displayed for the user to ask questions about the emergency drill. In response to the user's inquiry about the emergency drill plan entered on the fifth subpage, the system determines the real-time environmental information and real-time operating information of the equipment in the smart park through sensors in the smart park, determines the real-time weather information of the smart park through a meteorological data interface, and acquires real-time video of the smart park from cameras in the smart park. Based on the real-time video, the system determines the real-time pedestrian flow information and real-time traffic information of the smart park. The real-time environmental information includes the temperature and / or humidity of the smart park. Based on the real-time environmental information, real-time operational information, real-time weather information, real-time pedestrian flow information, and real-time traffic information, the large language model generates an emergency drill plan for the smart park. The emergency drill plan is displayed on the fifth subpage.

5. The smart park knowledge question-and-answer method according to claim 1 or 2, characterized in that, The method further includes: In response to the triggering operation of the work status function on the smart park knowledge Q&A page, a sixth subpage is displayed for the user to inquire about the work status of the staff in the smart park; In response to a question about the staff's work status entered on the sixth subpage, the system acquires the staff's facial expressions and body language through the camera at the staff's location, and identifies the staff's third emotional information based on the facial expressions and body language using the large language model. When the third emotional information indicates that the staff member is in a low mood, the large language model generates emotion regulation suggestions, which include recommended relaxation places and / or music for soothing emotions within a preset range. The emotion regulation suggestions are displayed on the sixth subpage and then sent to the terminal device held by the staff.

6. A smart park knowledge-based question-and-answer device based on a large language model, characterized in that, include: The first display module is used to display a smart park knowledge Q&A page based on a large language model on a terminal device, wherein the smart park knowledge Q&A page is used for users to ask questions related to smart parks in at least one of the following ways: images, voice and text. The second display module is used to respond to the triggering operation of the device malfunction function on the smart park knowledge Q&A page, display a first subpage for the user to ask about the device malfunction, and respond to the question entered by the user on the first subpage asking about the cause of the target device malfunction in the smart park, retrieve the real-time device parameters and historical maintenance records of the target device in the knowledge base, and obtain the data of the environmental sensors around the target device. A construction module is used to construct prompt words based on the problem of the target device's abnormality, the device parameters, the historical maintenance records, and the data from the environmental sensors. The prompt words are used to prompt the large language model to generate an abnormality diagnosis report for the target device based on the problem of the target device's abnormality, the device parameters, the historical maintenance records, and the data from the environmental sensors. The third display module is used to input the prompt words into the large language model to obtain an anomaly diagnosis report for the target device, and to display the anomaly diagnosis report on the first sub-page. The maintenance operation module is used to respond to the triggering operation of the maintenance operation function on the smart park knowledge Q&A page, displaying a second subpage for the user to ask maintenance operation questions; responding to the user's input of a maintenance operation question regarding the abnormal operation of the target equipment on the second subpage, acquiring the user's physiological information from the user's wearable device, and acquiring the environmental information of the user's environment through sensors on the terminal device; determining the user's stress state based on the physiological information and the environmental information, and if the stress state indicates that the user is in a high-pressure state, generating operation step information to instruct the user to perform abnormal maintenance on the target equipment based on the abnormal diagnosis report, the maintenance operation question, the physiological information, and the environmental information using the large language model; displaying the operation step information on the second subpage, and playing the operation step information in the form of voice; The visitor module is used to respond to the triggering operation of the visitor function on the smart park knowledge Q&A page, display a third subpage for visitors, and respond to the user's input operation on the third subpage, perform emotion recognition on the content corresponding to the input operation to obtain the user's first emotion information, wherein the content includes voice content and / or text content; acquire the user's facial expressions and body language through the camera on the terminal device, and identify the user's second emotion information based on the facial expressions and body language; determine the user's target emotion information based on the first emotion information and the second emotion information, and determine the user's visitor type based on the target emotion information, wherein the visitor type is used to characterize the user as a busy visitor or a leisure visitor; generate the smart park introduction information and smart park guidance information for the user based on the visitor type through the large language model, and display the introduction information and smart park guidance information on the third subpage.

7. An electronic device, characterized in that, The electronic device includes a memory and a processor, the memory being used to store computer instructions executable on the processor, and the processor being used to implement the steps of the method according to any one of claims 1-5 when executing the computer instructions.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by a processor, it implements the steps of the method according to any one of claims 1-5.

Images