A scenic spot guide system and method based on real-time position dynamic updating

Abstract

This application relates to the technical field of scenic area positioning services, and discloses a scenic area navigation system and method based on real-time location dynamic updates, including: a terminal binding module, used to temporarily bind tourist identification information with corresponding scenic area navigation terminals, generate a binding relationship data packet and store it; a scenic area navigation terminal, used to respond to the scanning signal issued by the terminal identification unit, decrypt the binding relationship data packet and send the temporarily bound tourist identification information to the terminal identification unit; a terminal identification module, including multiple terminal identification units distributed in the scenic area, used to send scanning signals to a designated scanning area, receive tourist identification information and obtain the tourist's real-time location information in combination with its own deployment location, and dynamically adjust the acquisition accuracy and update frequency of the real-time location information according to the category of the designated scanning area; and a multimodal guidance module, used to provide multimodal guidance information to tourists based on real-time location information, tourist identification information and real-time data of the scenic area.

Description

Technical Field

[0001] This application relates to the technical field of scenic area positioning services, and in particular to a scenic area navigation system and method based on real-time location dynamic updates. Background Technology

[0002] With the rapid development of the cultural tourism industry, scenic area tour guide services have transformed from traditional manual guides to intelligent guides. Location-based guide systems have become the core support for improving the quality of scenic area services and achieving refined management. Existing scenic area guide systems mostly rely on single positioning technologies (such as GPS) to obtain tourists' locations and then push guidance content.

[0003] However, in practical applications, due to factors such as the complexity of scenic areas, the adaptability of positioning technology, and imperfect service logic, there are many technical bottlenecks that make it difficult to balance the needs of tourists for a precise experience with the efficient management of scenic areas. For example, the accuracy of location information acquisition lacks the ability to adapt to different scenarios and is not designed for indoor and outdoor scenarios. Fixed positioning modes either cause guidance deviations in core areas due to insufficient accuracy or lead to resource waste and increased terminal energy consumption in peripheral areas due to excessive accuracy.

[0004] Secondly, the over-reliance on users' handheld smart devices, such as mobile phones, for tourist location tracking makes it difficult to guarantee location stability, especially for groups like the elderly or children who require more accurate positioning. Finally, there is a lack of targeted guidance services and poor coordination between location updates and guidance push notifications. Existing guidance information is mostly pre-set and fixed, failing to accurately push notifications based on tourists' real-time locations, individual preferences, and dynamic data from the scenic area. This results in low interactivity and personalization, making it difficult to adapt to tourists of different ages and usage habits. Summary of the Invention

[0005] To address the aforementioned technical issues, this application provides a scenic area navigation system and method based on real-time location dynamic updates.

[0006] A scenic area navigation system based on real-time location updates includes: The terminal binding module is used to temporarily bind tourist identification information with the corresponding scenic area guide terminal, generate a binding relationship data packet and store it; the tourist identification information includes the location signal identifier of the tourist handheld smart terminal that is bound to the other tourist, tourist identity information, unique code and the start and end time of the binding effect, and the binding relationship data packet is encrypted using a symmetric encryption algorithm; The scenic area guide terminal is used to respond to the scanning signal sent by the terminal identification unit, decrypt the binding relationship data packet, and send the temporarily bound tourist identification information to the terminal identification unit; The terminal identification module includes multiple terminal identification units distributed throughout the scenic area, which are used to send the scanning signal to the designated scanning area, receive tourist identification information and obtain the real-time location information of tourists in combination with their own deployment location, and dynamically adjust the acquisition accuracy and update frequency of the real-time location information according to the category of the designated scanning area. The multimodal guidance module has a built-in AI recommendation model, which is used to provide tourists with multimodal guidance information based on the real-time location information, tourist identification information and real-time scenic area data. The multimodal guidance information includes interactive information, recommended route information and alarm information. The alarm information covers warnings of dangerous areas, warnings of violations and reminders of exceeding the tour time limit.

[0007] Optionally, the terminal identification module has a built-in location adaptive adjustment unit, used to dynamically adjust the acquisition accuracy and update frequency of the real-time location information. The specific process includes: S301: Pre-classify the designated scanning area into categories, dividing it into core scenic area, passageway area, danger warning area, and rest area according to scenic area function, and into outdoor area and indoor area according to indoor / outdoor conditions; preset the reference positioning accuracy for each type of area. Reference update frequency And the regional weighting coefficient K, where the K value of the core scenic area and the danger warning area is greater than the K value of the passage area and the rest area, and the K value of the indoor area is greater than the K value of the outdoor area of ​​the same functional level; S302: The terminal identification unit receives tourist identification information sent by the scenic area guide terminal, and calculates the real-time distance between the tourist and the terminal identification unit through the RSSI ranging algorithm based on its own deployment location to obtain the tourist's initial location coordinates. S303: The location adaptive adjustment unit matches the corresponding area category based on the tourist's initial location coordinates and retrieves the reference parameters for that area. , The K value is used to calculate the real-time positioning accuracy using an adaptive adjustment formula. With real-time update frequency : ; ; S304: The terminal identification module calculates... and The scanning signal power, sampling interval, and positioning calculation power allocation of the terminal identification unit are adjusted to achieve dynamic acquisition of real-time location information.

[0008] Optionally, the preset reference parameters corresponding to different categories of regions in the specified scanning area are: Core tourist area =2m、 =2 times / second, K=1.2; Danger warning area =1m、 =3 times / second, K=1.5; Passage Area =3m、 = 1 time / second, K = 1.0; rest area =5m、 =0.5 times / second, K=0.8; The indoor area is adjusted by adding a 1.2 times indoor correction factor to the K value of the outdoor area of ​​the same functional level.

[0009] Optionally, the outdoor area corresponds to RFID location positioning technology, and the indoor area corresponds to Bluetooth 5.0 AOA high-precision positioning technology; the deployment density of the terminal identification unit is adapted to the area category, with the deployment density of core scenic spots and danger warning areas being higher than that of passage areas and rest areas, and the deployment density of indoor areas being higher than that of outdoor areas of the same functional level.

[0010] Optionally, the terminal binding module is also used to push a binding renewal reminder to the scenic area guide terminal for a preset period of time before the binding effective start and end time expires. If no renewal confirmation instruction is received, the binding relationship is automatically terminated and the binding relationship data packet is deleted. The symmetric encryption algorithm adopts the AES-128 symmetric encryption algorithm, and the key is generated by the terminal binding module and the scenic area guide terminal through a two-way verification mechanism.

[0011] Optionally, in the RSSI ranging algorithm of the terminal identification module, the collected RSSI values ​​are smoothed by Gaussian filtering to filter out jump values ​​caused by environmental interference, and then the real-time distance between the tourist and the terminal identification unit is calculated by a logarithmic-distance path loss model. The logarithmic-distance path loss model is as follows: ; in, To receive signal strength, For reference distance The received signal strength at 1m, where n is the path loss exponent and d is the real-time distance to be determined. The variable is Gaussian noise.

[0012] Secondly, this application provides a scenic area navigation method based on real-time location dynamic updates, the method comprising the following steps: S1: The terminal binding module obtains tourist identification information and corresponding scenic area guide terminal device information, establishes a temporary binding relationship, generates an encrypted binding relationship data packet using the AES-128 symmetric encryption algorithm, sets the start and end time of binding effectiveness, and stores it in the cloud server and scenic area guide terminal. S2: The distributed terminal identification units continuously send scanning signals to their respective designated scanning areas. After the scenic area guide terminal enters the scanning area, it automatically responds to the scanning signal, decrypts the binding relationship data packet, and sends the decrypted tourist identification information to the corresponding terminal identification unit. S3: The terminal identification module calculates the initial location coordinates of the tourist based on the received tourist identification information and the deployment location of the terminal identification unit itself using the RSSI ranging algorithm. Combined with the category of the specified scanning area, it calculates the real-time positioning accuracy and real-time update frequency through an adaptive adjustment formula, adjusts the scanning parameters and computing power allocation of the terminal identification unit, and completes the dynamic acquisition of the tourist's real-time location information. S4: The multimodal guidance module inputs real-time location information, tourist identification information, and real-time scenic area data into the built-in AI recommendation model. Through model reasoning, it generates personalized recommended route information, adapted interactive information, and alarm information, and pushes them to the scenic area guide terminal to realize dynamic guide service. S5: The terminal identification module monitors in real time whether the tourist's real-time location information exceeds the permitted tour area of ​​the scenic area or enters a dangerous warning area. If it exceeds or enters such an area, the multimodal guidance module will be triggered to push alarm information of the corresponding level. At the same time, the tourist's location, identity information and alarm information will be synchronized to the scenic area management terminal.

[0013] Optionally, the steps also include: In step S3, the adaptive adjustment formula is: ; ; in, For real-time positioning accuracy, To update the frequency in real time, For regional benchmark positioning accuracy, K represents the regional benchmark update frequency, and K is the regional weight coefficient.

[0014] In summary, this application includes at least one of the following beneficial technical effects: This invention employs a dual-region classification system to adapt differentiated positioning technologies and accuracy standards. Outdoors, RFID positioning balances coverage and cost, while indoor positioning uses Bluetooth to ensure accuracy. Furthermore, it adjusts update frequency based on functional areas, effectively addressing the limitations of existing systems where single positioning technologies cannot adapt to all scenic areas and where accuracy and energy consumption are unbalanced. This achieves optimal matching of positioning performance and resource consumption across different scenarios. The temporary binding mechanism, combined with symmetric encryption, validity period limits, and renewal reminders, significantly enhances binding security and flexibility. It avoids issues such as identity confusion and information leakage, and adapts to diverse scenarios including individual customer self-binding and group bulk binding, reducing the resource consumption of invalid bindings.

[0015] The multimodal guidance module relies on an AI recommendation model to push personalized content based on tourists' real-time location, individual preferences, and dynamic data of the scenic area. At the same time, it adapts to different terminal types and optimizes the push method, significantly improving the targeting and interactivity of the tour guide service and meeting the usage needs of different tourist groups.

[0016] The entire system and methodology, through dynamic location updates and coordinated linkage with tour guide services and scenic area management, can not only provide tourists with a coherent, accurate, and personalized tour experience, but also help scenic areas to monitor tourist distribution in real time, quickly handle safety hazards such as crossing boundaries, improve the level of refined management, adapt to the application needs of various scenic areas, and have strong practicality and promotional value. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of the architecture of the scenic area guide system based on real-time location dynamic updates in this application. Detailed Implementation

[0018] The embodiments of this application are described in detail below, and examples of the embodiments are shown in the accompanying drawings.

[0019] In the description of this specification, the references to "certain embodiments," "one embodiment," "some embodiments," "illustrative embodiment," "example," "specific example," or "some examples" refer to specific features, structures, materials, or characteristics described in connection with the described embodiment or example, which are included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0020] This application discloses a scenic area navigation system based on real-time location dynamic updates, referring to... Figure 1 ,include: The terminal binding module is used to temporarily bind tourist identification information with the corresponding scenic area guide terminal, generate a binding relationship data packet and store it; the tourist identification information includes the location signal identifier of the tourist handheld smart terminal that is bound to the other tourist, tourist identity information, unique code and the start and end time of the binding effect, and the binding relationship data packet is encrypted using a symmetric encryption algorithm; The scenic area guide terminal is used to respond to the scanning signal sent by the terminal identification unit, decrypt the binding relationship data packet, and send the temporarily bound tourist identification information to the terminal identification unit; The terminal identification module includes multiple terminal identification units distributed throughout the scenic area, which are used to send the scanning signal to the designated scanning area, receive tourist identification information and obtain the real-time location information of tourists in combination with their own deployment location, and dynamically adjust the acquisition accuracy and update frequency of the real-time location information according to the category of the designated scanning area. The multimodal guidance module has a built-in AI recommendation model, which is used to provide tourists with multimodal guidance information based on the real-time location information, tourist identification information and real-time scenic area data. The multimodal guidance information includes interactive information, recommended route information and alarm information. The alarm information covers warnings of dangerous areas, warnings of violations and reminders of exceeding the tour time limit. The terminal identification module has a built-in location adaptive adjustment unit, which is used to dynamically adjust the acquisition accuracy and update frequency of the real-time location information. The specific process includes: S301: Pre-classify the designated scanning area into categories, dividing it into core scenic area, passageway area, danger warning area, and rest area according to scenic area function, and into outdoor area and indoor area according to indoor / outdoor conditions; preset the reference positioning accuracy for each type of area. Reference update frequency And the regional weighting coefficient K, where the K value of the core scenic area and the danger warning area is greater than the K value of the passage area and the rest area, and the K value of the indoor area is greater than the K value of the outdoor area of ​​the same functional level; S302: The terminal identification unit receives tourist identification information sent by the scenic area guide terminal, and calculates the real-time distance between the tourist and the terminal identification unit through the RSSI ranging algorithm based on its own deployment location to obtain the tourist's initial location coordinates. S303: The location adaptive adjustment unit matches the corresponding area category based on the tourist's initial location coordinates and retrieves the reference parameters for that area. , The K value is used to calculate the real-time positioning accuracy using an adaptive adjustment formula. With real-time update frequency : ; ; S304: The terminal identification module calculates... and The scanning signal power, sampling interval, and positioning calculation power allocation of the terminal identification unit are adjusted to achieve dynamic acquisition of real-time location information.

[0021] In this embodiment, the invention is further described in detail in the context of a 4A-level comprehensive scenic area (including an indoor museum and outdoor natural landscapes). The scenic area covers a total area of ​​6 square kilometers and includes two indoor areas (folk museum and karst cave exhibition hall) and several outdoor areas (ancient building complex, natural trails, dangerous water areas, and rest service areas). The system described in this invention is deployed therein and is suitable for both individual and group tourists.

[0022] The terminal binding module employs a turnstile and an encrypted server, with built-in AES encryption algorithm, providing both online and offline binding channels. It handles the core function of temporarily binding tourist identification information to the scenic area's navigation terminals. Individual tourists bind their devices via the scenic area's official WeChat account or offline service desk, entering their name, contact information, and other identity information. The system automatically generates a unique code, obtains the tourist's mobile phone's Bluetooth and RFID positioning signals, sets the binding activation time based on the ticket validity period (e.g., 9:00-18:00 on the same day), generates an encrypted binding relationship data packet, and simultaneously stores it on the server and the scenic area's management backend, while also pushing it to the scenic area's dedicated APP. For group tourists, the tour leader submits the group number and member identity information. The terminal binding module batch-binds member information to the group's navigation devices, sets the binding activation time to the group's tour duration (e.g., 4 hours), and assigns personalized supplementary information permissions to each navigation device. Thirty minutes before the binding takes effect, a pop-up window and voice notification on the tourist's mobile APP or the group's navigation device remind them to renew. Individual tourists can extend their binding by 2 hours, while groups can have their binding extended by the tour leader. If no renewal is initiated, the binding relationship is automatically released after the effective time expires, and the data packet is deleted simultaneously.

[0023] The location signal identifier in the tourist identification information corresponds to the tourist's mobile phone Bluetooth and RFID location signal identifiers. The unique code is a randomly generated 64-bit string, ensuring that each binding relationship is unique and traceable, avoiding identity confusion. The effective start and end time of the binding can be set by the tourist or automatically generated by the scenic area based on the ticket validity period and group tour duration. The binding relationship data packet is encrypted using the AES symmetric encryption algorithm, and the key is verified and confirmed bidirectionally by the terminal binding module and the scenic area's guide terminal to prevent information leakage or tampering.

[0024] The scenic area guide terminal is used to respond to the scanning signal emitted by the terminal identification unit, decrypt the binding relationship data packet, and send the temporarily bound tourist identification information to the terminal identification unit. The scenic area guide terminal can include tourist handheld mobile terminals (adapted to the scenic area's dedicated APP and supporting iOS and Android systems), wearable guide devices (wristbands, headphones), and group-dedicated guide devices. It has the functions of scanning signal response, encrypted data packet decryption, tourist identification information transmission, and multimodal information reception and display. It can adapt the guidance information push method according to its own device type to ensure that the guidance content is effectively delivered.

[0025] The terminal identification module includes multiple terminal identification units distributed throughout the scenic area, which are used to send the scanning signal to the designated scanning area, receive tourist identification information and obtain the real-time location information of tourists in combination with their own deployment location, and dynamically adjust the acquisition accuracy and update frequency of the real-time location information according to the category of the designated scanning area. The terminal identification module uses RFID and Bluetooth dual-mode positioning technology and is deployed differently according to regional categories: Based on function, one warning sign is deployed every 50 meters in core scenic areas (ancient building complexes, exhibition hall core areas), one warning sign is deployed every 30 meters in danger warning areas (cliff edges, waterside areas), one warning sign is deployed every 100 meters in passage areas (main roads), and one warning sign is deployed every 150 meters in rest areas (viewing platforms, service areas). Based on indoor and outdoor classification, RFID positioning units are deployed in outdoor areas, while Bluetooth positioning units are deployed in indoor areas (museum exhibition halls, caves, visitor centers). The deployment density of indoor Bluetooth units is higher than that of outdoor RFID units with the same functional level (e.g., one unit every 30 meters in core indoor scenic spots) to avoid indoor signal obstruction affecting the positioning effect.

[0026] The terminal identification unit continuously sends scanning signals at a frequency of 1 second, with a scanning radius corresponding to half the deployment spacing to avoid signal overlap and interference. After tourists enter the scenic area, the scenic area guide terminal automatically responds to the scanning signal, decrypts the binding relationship data packet, and sends tourist identification information. The terminal identification module determines real-time location information by combining the unit deployment location and area category, while adjusting accuracy and update frequency: when tourists are in the indoor folk museum, real-time location is obtained through Bluetooth positioning, with accuracy controlled within 0.8 meters and an update frequency of 2 times / second as the baseline value; when tourists walk to the outdoor ancient building complex, it switches to RFID positioning, with accuracy controlled within 2 meters and an update frequency maintained at 2 times / second; when tourists enter the outdoor main road (passage area), the RFID positioning accuracy is adjusted to 3 meters, and the update frequency is reduced to 1 time / second; when tourists stay in the outdoor rest area, the update frequency is further reduced to 0.5 times / second to reduce terminal power consumption and achieve a balance between accuracy and power consumption.

[0027] After receiving the signal from the scenic area guide terminal, the terminal identification unit calculates the tourist's initial location coordinates through the following steps: RSSI value filtering: Gaussian filtering is performed on 20 continuously acquired RSSI values ​​to remove the maximum and minimum values ​​and then take the average value to filter out signal jumps caused by environmental occlusion and multipath effects, resulting in smoothed RSSI values. Distance Calculation: The real-time distance between the tourist and the terminal identification unit is calculated using a logarithmic-distance path loss model. The model formula is as follows: ; in, To receive signal strength, For reference distance Received signal strength at 1m This is a Gaussian noise variable. In this embodiment, the RFID unit... =-45dBm, Bluetooth unit =-58dBm; n is the path loss index, n=2.2 in open outdoor environments and n=3.0 in occluded indoor environments; d is the real-time distance between the tourist and the terminal recognition unit. The variable is a Gaussian noise with a mean of 0. In this embodiment, we take... =3dB; Using the ranging results from at least three terminal identification units, the initial planar location coordinates (x, y) of the tourist are calculated using the trilateration method, thus completing the initial position acquisition.

[0028] Visitors enter the core indoor attraction area. =1.2×1.2=1.44, calculated as follows =2 / 1.44≈1.39m =2×1.44≈2.88 times / second, the terminal identification module adjusts the scanning interval of the Bluetooth unit to 347ms, and sets the number of positioning calculation iterations to 8 times, ensuring positioning accuracy within 1.5m and an update frequency of 3 times / second; when tourists enter the outdoor rest area, =0.8, calculated as follows =5 / 0.8=6.25m, =0.5×0.8=0.4 times / second. The terminal identification module adjusts the scanning interval of the RFID unit to 2.5s, reduces the transmission power and the number of calculation iterations, and reduces the standby power consumption of the terminal.

[0029] In addition, if the scenic area's navigation terminal adopts the lowest-cost passive RFID solution (an RFID card), it can be associated with a mobile phone with a dedicated scenic area app installed during the binding phase. The mobile phone can transmit its Bluetooth name to the terminal binding module via the internet. The terminal binding module can be a turnstile with facial recognition functionality. When tourists enter the scenic area and pass through the turnstile for verification, they can place the scenic area navigation terminal on the turnstile's recognition area to write their identification information. When tourists enter the indoor area from outdoors, the terminal recognition module can additionally activate the Bluetooth positioning function, establishing a Bluetooth connection with the tourist's mobile phone by reading the mobile phone's Bluetooth name stored on the scenic area navigation terminal to achieve positioning. This solution is suitable for low-cost scenarios. On the one hand, the scenic area navigation terminal does not require battery power or a Bluetooth module; on the other hand, it is easy to recycle and reuse. Tourist identification information can be deleted and overwritten when leaving or entering the park, ensuring data security.

[0030] The multimodal guidance module is used to provide tourists with multimodal guidance information based on the real-time location information, tourist identification information, and real-time scenic area data. The multimodal guidance information includes interactive information, recommended route information, and alarm information. The alarm information covers warnings of dangerous areas, warnings of violations, and reminders of exceeding the visit time limit.

[0031] The AI ​​recommendation model built into the multimodal guidance module has been optimized and adapted using two years of historical visitor data from the scenic area. Input parameters include real-time location information, visitor age and preference tags (selected by the visitor during the binding phase), real-time visitor flow data for each attraction (updated every 5 minutes), and opening status. When a visitor (a young visitor with a preference for "humanities and history") enters the indoor museum, the model outputs a detailed audio explanation of the artifacts in the exhibition hall (interactive information) and recommends a route connecting less popular exhibits (avoiding crowd congestion). When a visitor approaches an unopened area (danger warning area) in the indoor cave exhibition hall, the terminal recognition module detects the abnormal location via Bluetooth positioning, triggering step S5. The multimodal guidance module pushes a level one alarm (wristband vibration + voice reminder + APP pop-up), and simultaneously synchronizes the visitor's real-time location and identity information to the scenic area management terminal. Staff then push return instructions to the visitor through the management terminal and simultaneously go to the site to guide the flow of visitors. When tourists (especially elderly tourists) switch to the outdoor ancient building complex, the model automatically adjusts the narration speed, recommends gentle routes, avoids steep slope areas, and simultaneously pushes the location information of nearby rest areas; when the tourist's visit time is close to the binding expiration time, a level 3 alarm (text + voice reminder) is pushed to inform them of the remaining visit time.

[0032] The AI ​​recommendation model built into the multimodal guidance module adopts a hybrid network architecture with an improved Wide&Deep + spatial attention mechanism. The complete application process and technical details are as follows: The model is divided into 5 core units, namely: feature preprocessing unit, wide linear layer, deep layer, spatial attention layer, and multi-task output layer. The functions and processing logic of each unit are as follows: Feature preprocessing unit: Responsible for quantizing and normalizing the input features. The input features are divided into 4 categories: Spatial characteristics: real-time location coordinates (x, y) of tourists, category of the area, coordinates and distances of surrounding attractions; User attribute characteristics: tourist age, preference tags (such as cultural history, natural landscapes, family entertainment, etc., using one-hot encoding), tour duration, and historical tour trajectory; Environmental characteristics: Real-time crowd density at each attraction (normalized to the 0-1 range), weather conditions, and current time period; Resource characteristics: the opening status of the attraction, the duration of the guided tour, and the physical fitness level required for visiting the attraction.

[0033] Wide Linear Layer: Employs a generalized linear model to perform cross-product transformations on discrete features, such as feature cross-features like "preference label - attraction type" and "age - fitness level," learning explicit association rules between features to ensure the interpretability of recommendation results.

[0034] Deep layer: A 3-layer fully connected neural network is used, with 256, 128 and 64 neurons in the hidden layer respectively. The ReLU function is used as the activation function to perform nonlinear transformation on continuous features and encoded discrete features, learn the implicit correlation between features and improve the generalization ability of the model.

[0035] Spatial Attention Layer: This is an attention mechanism customized for scenic area tour scenarios. Centered on the visitor's current location, it assigns dynamic attention weights to the features of attractions within a 1000m radius. The weight calculation formula is as follows: In the formula: Let be the attention weight for the i-th attraction; Let represent the matching degree (0-1 range) between the i-th attraction and the tourist preference label. Let be the straight-line distance between the i-th attraction and the tourist's current location; It is the sum of the matching degree-distance ratios of all surrounding attractions.

[0036] Through the spatial attention layer, the model can focus on attractions that are close to tourists and have a high degree of preference matching, which solves the pain point of traditional recommendation models that do not combine real-time spatial location and improves the synergy between location updates and guidance push.

[0037] Multi-task output layer: Employs a multi-task learning framework to simultaneously complete inference for two sub-tasks, sharing underlying features and setting independent output heads for each. The route planning subtask output header outputs the optimal tour route's sequence of attractions, path length, and estimated tour duration. It uses the A* path planning algorithm combined with the attraction weights output by the model to complete the final route generation. Interactive Information Adaptation Subtask Output Header: Outputs explanation content and interactive prompts that match the tourist's current location, age, and preferences. For elderly tourists, it automatically reduces the speaking speed and enlarges the font size; for children, it outputs fun science content; and for tourists with cultural preferences, it outputs in-depth historical explanations.

[0038] The offline training process of the model: Dataset construction: Collect 1.2 million visitor visit data from the scenic area over the past two years, including visitor trajectories, visitor attribute tags, scenic area visitor flow data, and basic information about attractions. After cleaning, deduplication, and missing value filling of the dataset, it is divided into training set, validation set, and test set in a ratio of 8:1:1. Loss function design: The total loss function is the weighted sum of the losses of the two sub-tasks, and the formula is: In the formula: =0.6, =0.4; The comprehensive loss for the route planning sub-task is composed of a weighted average of path length loss, pedestrian density loss, and preference matching degree loss. The cross-entropy loss function is adapted to the interactive information of the subtasks; Model training: The Adam optimizer was used, with an initial learning rate of 0.001, a batch size of 64, and 100 iterations. Training was stopped early when the validation set loss no longer decreased for 5 consecutive iterations. Performance validation: The test set validated that the model's route recommendation accuracy reached 92.3% and the user matching rate of the explanation content reached 89.7%, which are 11.2% and 9.5% higher than the traditional Wide&Deep model, respectively.

[0039] The model's online learning mechanism: After the model is deployed, real-time visitor flow data and attraction status data are collected every 5 minutes, and visitor behavior feedback (such as route clicks, listening time for guided tours, and duration of stay at attractions) is collected every hour. The model parameters are updated through incremental learning to ensure the real-time and adaptability of the recommended content and avoid the problem of fixed recommended content being out of touch with the real-time situation of the scenic area.

[0040] Example of model inference implementation: When a 25-year-old tourist with a preference for "humanities and history" enters the core exhibition area of ​​the indoor folk museum, the model executes a complete reasoning process: The feature preprocessing unit receives the following inputs: real-time location coordinates of the visitor (matching the core indoor attraction area), age 25, preference tag "humanities and history", current visitor density in the exhibition area is 0.7, and the opening status and location information of 10 surrounding exhibits; The Wide linear layer learns the explicit relationship between "humanities and history preferences and cultural relics exhibits" and assigns basic weights to cultural relics exhibits. Deep learning layers learn the implicit correlation between "young tourists - long-duration in-depth explanations - low-traffic routes" and output appropriate explanation durations and route preferences. The spatial attention layer centers on the visitor's current location and assigns dynamic attention weights to cultural relics and exhibits within a 50-meter radius. The closer the exhibit is and the higher its historical value, the higher the weight. The multi-task output layer completes the reasoning: the route planning sub-task output connects 8 high-weight cultural relics exhibits and the tour route avoids crowds, with an estimated tour time of 40 minutes; the interactive information adaptation sub-task output provides in-depth historical explanations of the exhibits through audio and text interactive information, which are simultaneously pushed to the tourist's scenic area guide terminal.

[0041] Secondly, this application provides a scenic area navigation method based on real-time location dynamic updates, the method comprising the following steps: S1: The terminal binding module obtains tourist identification information and corresponding scenic area guide terminal device information, establishes a temporary binding relationship and generates an encrypted binding relationship data packet, sets the start and end time of the binding effect and then stores it; S2: The distributed terminal identification units continuously send scanning signals to their respective designated scanning areas, and the scenic area guide terminal responds to the scanning signals and sends the decrypted tourist identification information. S3: The terminal identification module determines the real-time location information of tourists based on the received tourist identification information and its own deployment location, and adjusts the acquisition accuracy and update frequency of the real-time location information in combination with the specified scanning area category; S4: The multimodal guidance module generates multimodal guidance information based on real-time location information, tourist identification information, and real-time data of the scenic area, and pushes it to the scenic area guide terminal to realize dynamic guide service; S5: The terminal identification module monitors in real time whether the tourist's real-time location information exceeds the permitted tour area of ​​the scenic area. If it does, the multimodal guidance module is triggered to push alarm information, and the tourist's location and alarm information are synchronized to the scenic area management terminal.

[0042] Although embodiments of this application have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting this application. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of this application.

Claims

1. A scenic area navigation system based on real-time location dynamic updates, characterized in that, include: The terminal binding module is used to temporarily bind tourist identification information with the corresponding scenic area guide terminal, generate a binding relationship data packet and store it; The tourist identification information includes the location signal identifiers of the tourist handheld smart terminals that are linked together, tourist identity information, unique codes, and the start and end times of the linkage. The data packets of the linkage relationship are encrypted using a symmetric encryption algorithm. The scenic area guide terminal is used to respond to the scanning signal sent by the terminal identification unit, decrypt the binding relationship data packet, and send the temporarily bound tourist identification information to the terminal identification unit; The terminal identification module includes multiple terminal identification units distributed throughout the scenic area, which are used to send the scanning signal to the designated scanning area, receive tourist identification information and obtain the real-time location information of tourists in combination with their own deployment location, and dynamically adjust the acquisition accuracy and update frequency of the real-time location information according to the category of the designated scanning area. The multimodal guidance module has a built-in AI recommendation model, which is used to provide tourists with multimodal guidance information based on the real-time location information, tourist identification information and real-time scenic area data. The multimodal guidance information includes interactive information, recommended route information and alarm information. The alarm information covers warnings of dangerous areas, warnings of violations and reminders of exceeding the tour time limit.

2. The scenic area navigation system based on real-time location dynamic updates according to claim 1, characterized in that, The terminal identification module has a built-in location adaptive adjustment unit, which is used to dynamically adjust the acquisition accuracy and update frequency of the real-time location information. The specific process includes: S301: Pre-classify the designated scanning area into categories, dividing it into core scenic area, passageway area, danger warning area, and rest area according to scenic area function, and into outdoor area and indoor area according to indoor / outdoor conditions; preset the reference positioning accuracy for each type of area. Reference update frequency And the regional weighting coefficient K, where the K value of the core scenic area and the danger warning area is greater than the K value of the passage area and the rest area, and the K value of the indoor area is greater than the K value of the outdoor area of ​​the same functional level; S302: The terminal identification unit receives tourist identification information sent by the scenic area guide terminal, and calculates the real-time distance between the tourist and the terminal identification unit through the RSSI ranging algorithm based on its own deployment location to obtain the tourist's initial location coordinates. S303: The location adaptive adjustment unit matches the corresponding area category based on the tourist's initial location coordinates and retrieves the reference parameters for that area. , The K value is used to calculate the real-time positioning accuracy using an adaptive adjustment formula. With real-time update frequency : ； ； S304: The terminal identification module calculates... and The scanning signal power, sampling interval, and positioning calculation power allocation of the terminal identification unit are adjusted to achieve dynamic acquisition of real-time location information.

3. The scenic area navigation system based on real-time location dynamic updates according to claim 2, characterized in that, The preset baseline parameters corresponding to the different categories of regions in the specified scanning area are: Core tourist area =2m、 =2 times / second, K=1.2; Danger warning area =1m、 =3 times / second, K=1.5; Passage Area =3m、 = 1 time / second, K = 1.0; rest area =5m、 =0.5 times / second, K=0.8; The indoor area is adjusted by adding a 1.2 times indoor correction factor to the K value of the outdoor area of ​​the same functional level.

4. The scenic area navigation system based on real-time location dynamic updates according to claim 1, characterized in that, The outdoor area uses RFID location positioning technology, while the indoor area uses Bluetooth 5.0 AOA high-precision positioning technology. The deployment density of the terminal identification unit is adapted to the area category. The deployment density of core scenic spots and danger warning areas is higher than that of passageways and rest areas, and the deployment density of indoor areas is higher than that of outdoor areas of the same functional level.

5. The scenic area navigation system based on real-time location dynamic updates according to claim 1, characterized in that, The terminal binding module is also used to push a binding renewal reminder to the scenic area guide terminal for a preset period of time before the binding effective start and end time expires. If no renewal confirmation instruction is received, the binding relationship is automatically terminated and the binding relationship data packet is deleted. The symmetric encryption algorithm adopts the AES-128 symmetric encryption algorithm, and the key is generated by the terminal binding module and the scenic area guide terminal through a two-way verification mechanism.

6. The scenic area navigation system based on real-time location dynamic updates according to claim 1, characterized in that, In the RSSI ranging algorithm of the terminal identification module, the collected RSSI values ​​are smoothed by Gaussian filtering to filter out jump values ​​caused by environmental interference. Then, the real-time distance between the tourist and the terminal identification unit is calculated by the logarithmic-distance path loss model. The logarithmic-distance path loss model is as follows: ； in, To receive signal strength, For reference distance The received signal strength at 1m, where n is the path loss exponent and d is the real-time distance to be determined. The variable is Gaussian noise.

7. The scenic area navigation system and method based on real-time location dynamic updates according to claim 1, characterized in that, The method, applied to the scenic area navigation system based on real-time location dynamic updates as described in any one of claims 1-6, comprises the following steps: S1: The terminal binding module obtains tourist identification information and corresponding scenic area guide terminal device information, establishes a temporary binding relationship, generates an encrypted binding relationship data packet using the AES-128 symmetric encryption algorithm, sets the start and end time of binding effectiveness, and stores it in the cloud server and scenic area guide terminal. S2: The distributed terminal identification units continuously send scanning signals to their respective designated scanning areas. After the scenic area guide terminal enters the scanning area, it automatically responds to the scanning signal, decrypts the binding relationship data packet, and sends the decrypted tourist identification information to the corresponding terminal identification unit. S3: The terminal identification module calculates the initial location coordinates of the tourist based on the received tourist identification information and the deployment location of the terminal identification unit itself using the RSSI ranging algorithm. Combined with the category of the specified scanning area, it calculates the real-time positioning accuracy and real-time update frequency through an adaptive adjustment formula, adjusts the scanning parameters and computing power allocation of the terminal identification unit, and completes the dynamic acquisition of the tourist's real-time location information. S4: The multimodal guidance module inputs real-time location information, tourist identification information, and real-time scenic area data into the built-in AI recommendation model. Through model reasoning, it generates personalized recommended route information, adapted interactive information, and alarm information, and pushes them to the scenic area guide terminal to realize dynamic guide service. S5: The terminal identification module monitors in real time whether the tourist's real-time location information exceeds the permitted tour area of ​​the scenic area or enters a dangerous warning area. If it exceeds or enters such an area, the multimodal guidance module will be triggered to push alarm information of the corresponding level. At the same time, the tourist's location, identity information and alarm information will be synchronized to the scenic area management terminal.

8. The scenic area navigation method based on real-time location dynamic updates according to claim 7, characterized in that, In step S3, the adaptive adjustment formula is: ； ； in, For real-time positioning accuracy, To update the frequency in real time, For regional benchmark positioning accuracy, K represents the regional benchmark update frequency, and K is the regional weight coefficient.

Images