An intelligent guide content generation method and system, and a storage medium

CN122817569APending Publication Date: 2026-09-25CHENGDU JINGWEI TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202610979775.0
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-07-02
Publication Date
2026-09-25

AI Technical Summary

Technical Problem

[0013]为克服现有技术的不足,本发明提供了一种智能导览内容生成方法、系统及存储介质,解决了现有智慧展厅中仅依靠位置围栏触发讲解时容易误触发、错讲解、内容不够个性化的问题

Benefits of technology

[0053](1)本发明通过采用UWB 定位质量指标计算定位可信度,并将定位可信度作为导览内容触发和目标展项选择条件,降低相邻展项、边界区域和人员遮挡场景下的错讲概率,极大地提高了导览触发的准确性。

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122817569A_ABST
    Figure CN122817569A_ABST
Patent Text Reader

Abstract

The application discloses an intelligent guide content generation method and system and a storage medium. The method comprises the following steps: collecting data of a UWB positioning tag carried by an audience; calculating a positioning credibility score of a current position according to the positioning data; determining a candidate exhibit set containing at least one candidate exhibit based on the current position, an exhibition hall map and an exhibit space range; performing semantic matching on the candidate exhibits and calculating a comprehensive score of each candidate exhibit; determining a target exhibit according to the comprehensive score of each candidate exhibit and a preset false trigger prevention rule; and generating guide content suitable for the audience and outputting the guide content. The application can improve guide trigger accuracy, reduce the probability of mis-speaking in a boundary area and an adjacent exhibit scene, reduce false triggers when the audience passes by, and change the explanation content from a fixed template to controllable personalized generation.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention belongs to the field of smart exhibition hall and indoor intelligent tour guide technology, specifically relating to an intelligent tour guide content generation method, system and storage medium. Background Technology

[0002] Existing exhibition hall guided tour methods generally include fixed QR code explanations, Bluetooth beacon-triggered explanations, human explanations, manual selection of exhibits via mobile devices, and robot-guided tours. These methods can solve basic guided tour problems, but they still have significant shortcomings in actual use.

[0003] First, relying on QR codes or numbers for selection requires active operation from visitors. When visitors move around the exhibition hall, the system cannot accurately understand their current location and focus, making the explanation process easily interrupted.

[0004] Secondly, relying on Bluetooth RSSI (Received Signal Strength Indicator) or simple electronic fences typically triggers content only based on whether an exhibit has entered a certain area, lacking consideration for positioning quality and interference from adjacent exhibits. When two exhibits are close together, or when visitors are near the boundaries of exhibits, the system is prone to mis-talking, interrupting, or frequent switching.

[0005] Third, although some intelligent tour guide systems have introduced AI (Artificial Intelligence) content generation capabilities, the content they generate is often based solely on exhibit numbers or preset scripts, failing to fully incorporate on-site factors such as "whether the visitor is actually standing in front of the exhibit," "whether the current position is stable," and "what the visitor's previous visit was."

[0006] The smart exhibition hall market is shifting from simple multimedia displays to a closed-loop ecosystem based on artificial intelligence and precise positioning. Precise positioning is a crucial foundation for automated guided tours, and AI-powered guided tours are gradually moving from fixed scripts to dynamically generated content based on real-time location, feedback, and interest profiles. This trend is already reflected in smart exhibition hall market analysis data: the data points out that precise positioning technology is the physical foundation for achieving automated guided tours that allow for "changing scenery as you move," and generative artificial intelligence and large language models can dynamically generate personalized explanations based on visitor real-time feedback, location information, and interest profiles.

[0007] Currently, existing technologies still have the following problems:

[0008] 1. Location triggering is too simplistic. Existing solutions typically use fixed areas, fixed distances, or single signal strength thresholds for triggering, making it impossible to determine the reliability of the current location result.

[0009] 2. Adjacent exhibits are easily confused. In exhibition areas with densely packed display cases, close-knit exhibits, and narrow corridors, the system has difficulty determining the true focus of visitors when they are standing between two exhibits.

[0010] 3. Insufficient integration of content with audience engagement. Audiences of different ages, professional backgrounds, and purposes of visit require explanations of the same exhibit in varying depths and with different modes of expression, which traditional fixed scripts cannot adequately address.

[0011] 4. Lack of a mechanism to prevent accidental triggering. When a visitor briefly stops, passes by, turns back, or is obscured by the crowd, the system may immediately play the narration, causing disruption.

[0012] 5. Lack of feedback loop. After the system plays content, it often fails to adjust subsequent navigation strategies based on viewer behavior such as dwell time, likes, questions, and skips. Summary of the Invention

[0013] To overcome the shortcomings of existing technologies, this invention provides an intelligent guide content generation method, system, and storage medium, which solves the problems of accidental triggering, incorrect explanations, and insufficient personalization of content when relying solely on location fences to trigger explanations in existing smart exhibition halls.

[0014] The technical solution adopted by the present invention to solve the above problems is:

[0015] A method for generating intelligent navigation content includes the following steps:

[0016] Data from the UWB positioning tags carried by the audience is collected to obtain information such as the audience's current location, timestamp, ranging results, number of base stations involved in positioning, positioning residual, and signal quality;

[0017] Calculate the location reliability score for the current location based on the location data;

[0018] Based on the current location, exhibition hall map, and exhibition space range, a set of candidate exhibition items is determined, which includes at least one candidate exhibition item.

[0019] Semantic matching is performed on the candidate exhibits, the relevance of each candidate exhibit to the audience status is calculated, and a comprehensive score is obtained for each candidate exhibit.

[0020] The target exhibit is determined based on the overall score of each candidate exhibit and the preset rules for preventing accidental triggering.

[0021] Based on the knowledge base content of the target exhibits, the explanation templates, and the generation constraints, guide content adapted to the audience is generated and output.

[0022] Furthermore, as a preferred technical solution, the calculation of the positioning reliability score of the current location adopts the following formula:

[0023] C_loc = w1·Q_anchor + w2·Q_residual + w3·Q_geometry + w4·Q_stability + w5·Q_zone;

[0024] Where C_loc represents the location reliability score, which ranges from 0 to 1; Q_anchor represents the effective base station number score; Q_residual represents the ranging residual score; Q_geometry represents the base station geometric distribution score; Q_stability represents the location continuity score; Q_zone represents the regional consistency score; and w1 to w5 are weighting coefficients.

[0025] Furthermore, as a preferred technical solution, the step of determining the candidate set includes:

[0026] Add the exhibits within the area where the audience's location coordinates are located to the candidate exhibit set;

[0027] Adjacent exhibits whose straight-line distance from the audience's location coordinates is less than a preset buffer distance are added to the candidate exhibit set.

[0028] Furthermore, as a preferred technical solution, the calculation of the comprehensive score of each candidate exhibit adopts the following formula:

[0029] ;

[0030] Where Score_i is the score of the i-th candidate exhibit, C_loc_i is the degree of support for the exhibit by the location credibility; D_i represents the spatial distance and viewing direction score; S_sem_i represents the semantic matching score; S_user_i represents the audience interest matching score; T_dwell_i represents the dwell time score; and P_i represents the penalty item.

[0031] Furthermore, as a preferred technical solution, the anti-false triggering rules include the following rules:

[0032] Minimum stay time rule: A guided tour is only allowed after a visitor has stayed within the service area of ​​a certain exhibit for a preset time threshold.

[0033] Stable threshold rule: It is not triggered when the location reliability is lower than the first threshold, and it is only triggered after it is higher than the second threshold and continues for a number of location cycles;

[0034] Hysteresis rule: Once a certain exhibit has been selected, the target exhibit will only be switched if the combined score of adjacent exhibits is consistently higher than that of the current exhibit by a certain margin.

[0035] Cooldown rule: Once the same exhibit has been explained, a complete explanation will not be triggered again within a set time.

[0036] Conflict confirmation rule: When multiple candidate exhibits have similar scores, a brief prompt will be displayed or the audience will be asked to make their own choice.

[0037] Furthermore, as a preferred technical solution, the step of generating the guide content includes:

[0038] Based on the audience status, retrieve the corresponding explanatory materials from the exhibit knowledge base;

[0039] Based on preset generation constraints, the generated content is limited to not exceeding the scope of the review knowledge base;

[0040] Based on the age group or professional level of the audience, differentiated explanation content is generated at the basic, extended, or professional levels.

[0041] Furthermore, as a preferred technical solution, a feedback optimization step is also included:

[0042] Record audience feedback on the guided tour content, including dwell time, likes, skips, or questions;

[0043] Adjust subsequent tour guidance strategies based on the feedback behavior and update the recommendation weights of the corresponding exhibits.

[0044] An intelligent navigation content generation system includes:

[0045] The positioning module includes a UWB positioning tag, a UWB base station, and a positioning calculation module, which is used to collect UWB signals and calculate the audience's position coordinates and positioning quality parameters.

[0046] The reliability assessment module is used to calculate the location reliability score based on the location quality parameters.

[0047] The candidate exhibit selection module is used to generate a set of candidate exhibits based on the audience's location, distance to nearby exhibits, exhibit orientation, and visitor flow.

[0048] The semantic matching and decision-making module is used to calculate the comprehensive score of each candidate exhibit by combining information such as exhibit semantic tags, audience profiles, and historical visit records, and to determine the target exhibit based on the rules for preventing accidental triggering.

[0049] The content generation module is used to generate guided tour content that is adapted to the audience based on the knowledge base content, explanation templates and generation constraints of the target exhibits;

[0050] The terminal presentation module is used to output the guided tour content to the viewer's terminal.

[0051] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method as claimed in any one of claims 1 to 7.

[0052] Compared with the prior art, the present invention has the following beneficial effects:

[0053] (1) This invention calculates the location reliability by using UWB location quality index and uses the location reliability as a condition for triggering guide content and selecting target exhibits, thereby reducing the probability of misleading in adjacent exhibits, boundary areas and human occlusion scenarios, and greatly improving the accuracy of guide triggering.

[0054] (2) This invention uses the audience’s current location, the spatial relationship of candidate exhibits, the semantic tags of exhibits and the audience’s visit profile to comprehensively score the target exhibits, and combines them with an anti-accidental touch mechanism to avoid the audience being frequently disturbed when passing by, thereby reducing the occurrence of accidental triggering, incorrect explanations and frequent switching.

[0055] (3) This invention improves the relevance between content by combining the guided tour content with information such as the visitor's location, interests, visit history and current scene.

[0056] (4) By recording feedback behaviors such as stopping, skipping, liking and asking questions, this invention can better optimize recommendation strategies and explanation levels, and provide higher quality tour guide content.

[0057] (5) This invention can generate different levels of guided tour content based on the target exhibit knowledge base and the audience status, and restricts the generated content from exceeding the scope of the approved knowledge base, thereby enabling the explanation content to be transformed from a fixed template to a controllable personalized generation, which greatly satisfies the personalized needs of different audiences.

[0058] (6) This invention can be implemented based on the existing UWB positioning system and exhibit knowledge base, without requiring each exhibit to be configured with a complex interactive device, thus facilitating engineering implementation. Attached Figure Description

[0059] Figure 1 This is a schematic diagram of the method flow of the present invention;

[0060] Figure 2 This is a schematic diagram of the system composition structure of the present invention;

[0061] Figure 3 This is a schematic diagram illustrating the location credibility and semantic matching process of the present invention;

[0062] Figure 4 This is a schematic diagram illustrating the prevention of false triggering and hysteresis judgment of the present invention. Detailed Implementation

[0063] The present invention will be further described in detail below with reference to the embodiments and accompanying drawings, but the embodiments of the present invention are not limited thereto.

[0064] Example

[0065] like Figure 1 As shown in this embodiment, a method for generating intelligent navigation content includes the following steps:

[0066] Data from the UWB positioning tags carried by the audience is collected to obtain information such as the audience's current location, timestamp, ranging results, number of base stations involved in positioning, positioning residual, and signal quality;

[0067] The location credibility score is calculated based on the location data. The location credibility score can be a value between 0 and 1. The higher the value, the more reliable the current location is.

[0068] Based on the current location, the exhibition hall map, and the spatial range of the exhibits, a set of candidate exhibits containing at least one candidate exhibit is determined; the candidate exhibits include not only exhibits within the area where the visitor is located, but also adjacent exhibits that are relatively close to each other;

[0069] Semantic matching is performed on the candidate exhibits, the relevance of each candidate exhibit to the audience status is calculated, and a comprehensive score is obtained for each candidate exhibit.

[0070] The target exhibit is determined based on the overall score of each candidate exhibit and the preset rules for preventing accidental triggering.

[0071] Based on the knowledge base content of the target exhibits, the explanation templates, and the generation constraints, guide content adapted to the audience is generated and output.

[0072] The location reliability in this invention does not simply treat UWB location results as absolutely reliable coordinates, but rather evaluates the quality of the current location result. In practical engineering, although UWB positioning has high accuracy, it is still affected by factors such as occlusion, multipath propagation, base station geometric distribution, population density, and tag pose changes. Therefore, determining whether the current location is reliable before triggering navigation is of practical significance.

[0073] The location confidence level C_loc can be calculated from at least one of the following factors:

[0074] 1. Number of participating base stations: The more effective base stations participating in the positioning, the higher the reliability.

[0075] 2. Ranging residual: The smaller the positioning solution residual, the more consistent the ranging results are, and the higher the reliability.

[0076] 3. Base station geometric distribution: The more uniform the base station distribution, the better the geometric accuracy factor and the higher the reliability.

[0077] 4. Positional continuity: Whether the change in position between the current position and the previous position conforms to the walking speed and movement path of a person;

[0078] 5. Regional Consistency: Whether the current location remains within the service area of ​​the same exhibition area or the same exhibit;

[0079] 6. Signal quality: including received signal strength, first-path quality, and non-line-of-sight (NLOS) assessment results.

[0080] Based on this, the present invention can use the following formula to calculate the location reliability score of the current location:

[0081] C_loc = w1·Q_anchor + w2·Q_residual + w3·Q_geometry + w4·Q_stability + w5·Q_zone;

[0082] Where C_loc represents the location reliability score, which ranges from 0 to 1; Q_anchor represents the effective base station number score; Q_residual represents the ranging residual score; Q_geometry represents the base station geometric distribution score; Q_stability represents the location continuity score; Q_zone represents the regional consistency score; and w1 to w5 are weighting coefficients.

[0083] Specifically, the steps for determining the candidate set include:

[0084] Add the exhibits within the area where the audience's location coordinates are located to the candidate exhibit set;

[0085] Adjacent exhibits whose straight-line distance from the audience's location coordinates is less than a preset buffer distance are added to the candidate exhibit set.

[0086] The purpose of candidate exhibit selection is to prevent the system from initially identifying the closest exhibit as the sole target. In an exhibition hall, visitors might be standing between two exhibits, facing an exhibit to their right, or passing by an exhibit without stopping. Using only the closest distance could easily lead to misjudgment.

[0087] The system can pre-define the spatial service area for each exhibit, including the exhibit's center coordinates, effective explanation area, best viewing area, adjacent exhibits, exhibit orientation, and recommended viewing direction. Once a visitor's location enters a certain area, the system adds exhibits within that area and a certain buffer distance to the candidate set.

[0088] like Figure 3 As shown, for example, in an exhibition area, two exhibits, a "Lunar Rover Model" and a "Chang'e 5 Return Capsule," are arranged adjacent to each other. When a visitor stands between the two exhibits, the candidate set contains both exhibits. The system then makes a decision based on factors such as dwell time, direction of movement, historical visit order, and semantic matching results, rather than directly selecting the closest one.

[0089] Exhibit semantic matching refers to the system not only determining whether a visitor is near a particular exhibit, but also whether that exhibit is relevant to the visitor's current state. The semantic matching here does not require a complex model; easily implemented engineering tools such as tags, keywords, topic categories, and knowledge base fields can be used.

[0090] Each exhibit can be pre-configured with the following information: exhibit name, exhibit number, exhibit category, theme tag, keywords, synonyms, suitable age group, explanation depth, related exhibits, standard explanation text, expert review content, and prohibited content range.

[0091] On the visitor side, a lightweight visitor profile can be generated, including the language chosen upon entry, age group, purpose of visit, exhibits visited, dwell time, likes or skips, and current question. The system matches exhibit tags with the visitor profile to obtain a semantic matching score S_sem. Based on this, the comprehensive score of each candidate exhibit can be calculated using the following formula:

[0092] ;

[0093] Where Score_i is the score of the i-th candidate exhibit, C_loc_i is the degree of support for the exhibit by the location credibility; D_i represents the spatial distance and viewing direction score; S_sem_i represents the semantic matching score; S_user_i represents the audience interest matching score; T_dwell_i represents the dwell time score; P_i represents the penalty item, such as similar content just presented, audience passing by quickly, score conflict between adjacent exhibits, etc.

[0094] When the difference between the highest score and the second highest score is less than a set threshold, the system may temporarily not play the full explanation, but wait for the next positioning update, or pop up a brief confirmation on the viewer's terminal, such as "Would you like to learn about the lunar rover model or the return capsule?" The content will be generated after the viewer makes a choice.

[0095] To address the issue of frequent system switching when visitors pass by exhibits, linger in boundary areas, or move back and forth between adjacent exhibits, this invention designs anti-accidental triggering rules, as follows:

[0096] Minimum stay time rule: A guided tour is only allowed after a visitor has stayed within the service area of ​​a certain exhibit for a preset time threshold.

[0097] Stable threshold rule: It is not triggered when the location reliability is lower than the first threshold, and it is only triggered after it is higher than the second threshold and continues for a number of location cycles;

[0098] Hysteresis rule: Once a certain exhibit has been selected, the target exhibit will only be switched if the combined score of adjacent exhibits is consistently higher than that of the current exhibit by a certain margin. For example... Figure 4 As shown;

[0099] Cooldown rule: Once the same exhibit has been explained, a complete explanation will not be triggered again within a set time.

[0100] Conflict confirmation rule: When multiple candidate exhibits have similar scores, a brief prompt will be displayed or the audience will be asked to make their own choice.

[0101] The steps for generating the navigation content in this embodiment include:

[0102] Based on the audience status, retrieve the corresponding explanatory materials from the exhibit knowledge base;

[0103] Based on preset generation constraints, the generated content is limited to not exceeding the scope of the review knowledge base;

[0104] Based on the age group or professional level of the audience, differentiated explanation content is generated at the basic, extended, or professional levels.

[0105] Specifically, once the system identifies the target exhibit, the guide content generation module retrieves corresponding information from the exhibit knowledge base and generates explanation content according to the audience's status. To ensure content reliability, this invention prioritizes using verified exhibit knowledge, explanation texts, image descriptions, and expert materials as the basis for generation, rather than allowing the model to generate content freely.

[0106] The generated content can be divided into three levels:

[0107] The first floor provides basic explanations, suitable for first-time visitors or children. The language is concise, focusing on what the exhibits are, what their purpose is, and why they are important.

[0108] The second floor provides extended explanations, suitable for general adult audiences, and adds background stories, technical principles, historical developments, or related exhibits.

[0109] The third layer provides professional explanations, suitable for industry professionals or in-depth visitors, and includes content on parameters, structure, technology, research process, etc.

[0110] Constraints can be set during content generation, such as prohibiting fabricated exhibit dates, exceeding the scope of the review knowledge base, using expressions inappropriate for the audience's age, and mixing information from adjacent exhibits into the current exhibit's explanation. After the system generates content, it can be output through text, speech synthesis, headphones, mini-programs, screens, or robots.

[0111] To continuously optimize the guided tour strategy, this embodiment also includes a feedback optimization step. Specifically, the system can record visitor feedback on the guided tour content, including whether they listened to the entire presentation, liked it, asked further questions, skipped parts, or stayed after the explanation. This feedback does not directly change the actual content of the exhibits, but is used to adjust subsequent guided tour strategies, such as increasing the recommendation weight of certain themes or reducing repetitive content that visitors have already heard.

[0112] For example, if a visitor lingers at the "Space Engine" exhibit and continues to ask questions about propulsion principles, the system can determine that they have a high level of interest in the topic of aerospace technology. When they subsequently enter the "Return Capsule" exhibit, the system can appropriately add technical content such as reentry and return, and thermal protection materials, in addition to the basic explanation.

[0113] like Figure 2 As shown in the figure, this embodiment provides an intelligent navigation content generation system, including:

[0114] The positioning module includes a UWB positioning tag, a UWB base station, and a positioning calculation module, which is used to collect UWB signals and calculate the audience's position coordinates and positioning quality parameters.

[0115] The reliability assessment module is used to calculate the location reliability score based on the location quality parameters.

[0116] The candidate exhibit selection module is used to generate a set of candidate exhibits based on the audience's location, distance to nearby exhibits, exhibit orientation, and visitor flow.

[0117] The semantic matching and decision-making module is used to calculate the comprehensive score of each candidate exhibit by combining information such as exhibit semantic tags, audience profiles, and historical visit records, and to determine the target exhibit based on the rules for preventing accidental triggering.

[0118] The content generation module is used to generate guided tour content that is adapted to the audience based on the knowledge base content, explanation templates and generation constraints of the target exhibits;

[0119] The terminal presentation module is used to output the guided tour content to the viewer's terminal.

[0120] UWB positioning tags are worn or carried by visitors and can take the form of badges, wristbands, audio guides, or other portable terminals. UWB base stations are deployed on the ceiling, walls, or near display cases in the exhibition hall to receive tag signals and work with the positioning engine to calculate the visitor's location. The positioning calculation module outputs data such as the visitor's location coordinates, timestamp, ranging quality, and positioning mode.

[0121] The location reliability assessment module scores the reliability of the location results. The candidate exhibit selection module forms a set of candidate exhibits based on the visitor's location, distance to nearby exhibits, exhibit orientation, and visitor flow. The semantic matching and decision-making module further combines exhibit semantic tags, visitor profiles, and historical visit records to calculate a comprehensive score for each candidate exhibit. The guide content generation module generates text, audio, or multimedia guide content based on the target exhibit's knowledge content and the visitor's state.

[0122] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method as claimed in any one of claims 1 to 7.

[0123] Below are examples illustrating three specific guided tour scenarios:

[0124] I. Guided Tour of the Aerospace Exhibition Area for Children

[0125] A science and technology museum has set up a space exploration and lunar exploration exhibition area, with models of the Chang'e 5 return capsule, lunar rover, and rocket engine cross-section displayed adjacent to each other. When a visitor named Xiaoming entered the exhibition hall, he was identified as a child by a UWB location tag attached to his badge via a mini-program. The system recorded him as a child visitor, and the narration was set to Chinese with a basic level of detail.

[0126] Xiaoming walked near the "Chang'e 5 Return Capsule Model". The system obtained five consecutive positioning results with an average positioning reliability of 0.86, and Xiaoming stayed within the optimal viewing area of ​​the return capsule for more than 8 seconds. Among the candidate exhibits, the "Return Capsule Model" had the highest spatial score, and its semantic tags "lunar sampling, return, spacecraft" were related to the theme of Xiaoming's previous exhibit, "lunar sample display," resulting in a higher overall score than the "lunar rover model." Based on this, the system determined the target exhibit to be the "Chang'e 5 Return Capsule Model."

[0127] The system generates the following type of explanation: "What you are seeing now is a model of the Chang'e 5 return capsule. Its mission is to safely return the precious package brought back from the moon to Earth. The return capsule enters the atmosphere at high speed and its exterior will be subjected to very high temperatures, so a special heat-resistant structure is required."

[0128] If Xiaoming does not stop but merely passes in front of the return capsule, the system will not trigger a full explanation due to insufficient stopping time. Instead, it will only display the exhibit name and a brief prompt "Stop to listen to the explanation" on the mini-program interface.

[0129] II. Preventing Mistakes at the Boundary of Adjacent Exhibits

[0130] A company's exhibition hall displayed two exhibits, "UWB Positioning Base Station" and "UWB Positioning Tag," side by side, with the two exhibits quite close together. Visitors standing between them might easily switch back and forth between the two explanations based solely on distance.

[0131] With this invention, the system first determines the location reliability. When the reliability is 0.58 and the location falls within the buffer zone between two exhibits, the system does not immediately play the full explanation but observes the subsequent three location cycles. If the visitor then approaches the "Location Tag" exhibit and remains stationary in front of it, the system then triggers the explanation of the "Location Tag." If the visitor is still between the two, the system can display a selection prompt: "Would you like to learn about the location base station or the location tag?"

[0132] This process prevents the system from immediately playing incorrect content due to a single unstable positioning result.

[0133] III. In-depth explanation for adult professional audiences

[0134] An industrial exhibition hall received a customer from the power industry. Upon entering the hall, the customer selected the "professional visit mode" and repeatedly lingered on topics such as "indoor high-precision positioning" and "safety linkage" in the preceding exhibits. When the customer entered the "Smart Exhibition Hall Platform Architecture" exhibit, in addition to displaying a basic introduction, the system automatically supplemented the information with details on the system architecture, how UWB positioning data drives the generation of guided tour content, and how it links with the central control system.

[0135] When the same exhibit is presented to the general public, the system only explains that "the system can know your location in the exhibition hall and automatically explain the content in front of you"; when it is presented to professional customers, it further explains that "the system first evaluates the credibility of the UWB positioning results, then combines the semantic tags of the exhibit with the visitor history to calculate the target exhibit, and finally calls the knowledge base to generate the guide content."

[0136] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Based on the technical essence of the present invention, any simple modifications, equivalent substitutions, and improvements made to the above embodiments within the spirit and principles of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A method for generating intelligent navigation content, characterized in that, Includes the following steps: Data from the UWB positioning tags carried by the audience is collected to obtain information such as the audience's current location, timestamp, ranging results, number of base stations involved in positioning, positioning residual, and signal quality; Calculate the location reliability score for the current location based on the location data; Based on the current location, exhibition hall map, and exhibition space range, a set of candidate exhibition items is determined, which includes at least one candidate exhibition item. Semantic matching is performed on the candidate exhibits, the relevance of each candidate exhibit to the audience status is calculated, and a comprehensive score is obtained for each candidate exhibit. The target exhibit is determined based on the overall score of each candidate exhibit and the preset rules for preventing accidental triggering. Based on the knowledge base content of the target exhibits, the explanation templates, and the generation constraints, guide content adapted to the audience is generated and output.

2. The method according to claim 1, characterized in that, The location reliability score for the current position is calculated using the following formula: C_loc = w1·Q_anchor + w2·Q_residual + w3·Q_geometry + w4·Q_stability+ w5·Q_zone; Where C_loc represents the location reliability score, which ranges from 0 to 1; Q_anchor represents the effective base station number score; Q_residual represents the ranging residual score; Q_geometry represents the base station geometric distribution score; Q_stability represents the location continuity score; Q_zone represents the regional consistency score; and w1 to w5 are weighting coefficients.

3. The method according to claim 1, characterized in that, The steps for determining the candidate set include: Add the exhibits within the area where the audience's location coordinates are located to the candidate exhibit set; Adjacent exhibits whose straight-line distance from the audience's location coordinates is less than a preset buffer distance are added to the candidate exhibit set.

4. The method according to claim 1 or 3, characterized in that, The overall score for each candidate exhibit is calculated using the following formula: ; Where Score_i is the score of the i-th candidate exhibit, C_loc_i is the degree of support for the exhibit by the location credibility; D_i represents the spatial distance and viewing direction score; S_sem_i represents the semantic matching score; S_user_i represents the audience interest matching score; T_dwell_i represents the dwell time score; and P_i represents the penalty item.

5. The method according to claim 1, characterized in that, The rules for preventing accidental triggering include the following: Minimum stay time rule: A guided tour is only allowed after a visitor has stayed within the service area of ​​a certain exhibit for a preset time threshold. Stable threshold rule: It is not triggered when the location reliability is lower than the first threshold, and it is only triggered after it is higher than the second threshold and continues for a number of location cycles; Hysteresis rule: Once a certain exhibit has been selected, the target exhibit will only be switched if the combined score of adjacent exhibits is consistently higher than that of the current exhibit by a certain margin. Cooldown rule: Once the same exhibit has been explained, a complete explanation will not be triggered again within a set time. Conflict confirmation rule: When multiple candidate exhibits have similar scores, a brief prompt will be displayed or the audience will be asked to make their own choice.

6. The method according to claim 1, characterized in that, The steps for generating the navigation content include: Based on the audience status, retrieve the corresponding explanatory materials from the exhibit knowledge base; Based on preset generation constraints, the generated content is limited to not exceeding the scope of the review knowledge base; Based on the age group or professional level of the audience, differentiated explanation content is generated at the basic, extended, or professional levels.

7. The method according to claim 1, characterized in that, It also includes feedback optimization steps: Record audience feedback on the guided tour content, including dwell time, likes, skips, or questions; Adjust subsequent tour guidance strategies based on the feedback behavior and update the recommendation weights of the corresponding exhibits.

8. An intelligent navigation content generation system, characterized in that, include: The positioning module includes a UWB positioning tag, a UWB base station, and a positioning calculation module, which is used to collect UWB signals and calculate the audience's position coordinates and positioning quality parameters. The reliability assessment module is used to calculate the location reliability score based on the location quality parameters. The candidate exhibit selection module is used to generate a set of candidate exhibits based on the audience's location, distance to nearby exhibits, exhibit orientation, and visitor flow. The semantic matching and decision-making module is used to calculate the comprehensive score of each candidate exhibit by combining information such as exhibit semantic tags, audience profiles, and historical visit records, and to determine the target exhibit based on the rules for preventing accidental triggering. The content generation module is used to generate guided tour content that is adapted to the audience based on the knowledge base content, explanation templates and generation constraints of the target exhibits; The terminal presentation module is used to output the guided tour content to the viewer's terminal.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method as described in any one of claims 1 to 7.