Intelligent exhibition item module clustering management method and system supporting multi-scene switching

Abstract

The invention discloses an intelligent exhibition item module clustering management method and system supporting multi-scene switching. The intelligent exhibition item module clustering management method comprises the following steps: step 1, real-time state acquisition; 2, generating a dynamic strategy; step 3, content updating and issuing; step 4, operation monitoring and fault processing; 5, multi-scene adaptation is carried out; and step 6, audience preference learning. According to the method, intelligent collaborative scheduling of exhibition item clusters is realized through a real-time state acquisition and dynamic strategy generation mechanism, the system can automatically optimize resource allocation based on multi-dimensional load evaluation and an audience interest model, the problem of resource scrambling caused by independent operation of traditional exhibition items is solved, content hot update and cluster-level synchronization are supported, and the system is suitable for popularization and application. The bottleneck of traditional display content homogenization is broken through, meanwhile, a predictive operation and maintenance closed loop is constructed by a dual-channel redundancy monitoring and three-level fault response system, the equipment downtime risk is remarkably reduced, and the system stability is improved.

Description

Technical Field

[0001] The present invention relates to the field of intelligent display, and in particular to a clustered management method and system for intelligent display modules supporting multi-scene switching. Background Art

[0002] In intelligent display scenarios such as museums, theme pavilions and commercial exhibitions, traditional exhibit management systems generally adopt a "single-point control-independent operation" technical architecture. Each exhibit (such as interactive screens, holographic projection devices, physical models, etc.) is equipped with an independent control system, and content playback and equipment start and stop are achieved through preset fixed programs or manual operations. Although this technical architecture can maintain basic operation in simple exhibition scenarios, as display needs develop in a diversified and dynamic direction, its technical defects are becoming increasingly prominent.

[0003] The existing technical system has three core defects: First, content scheduling is inefficient, and there is a lack of coordination mechanism between exhibits. Content switching requires manual operation on each display or relies on timed scripts. It is impossible to adjust the display rhythm in real time according to fluctuations in audience traffic, resulting in long waiting times for audiences during peak periods or idle equipment during off-peak periods; second, the resource allocation mechanism is rigid, and the traditional system can only roughly manage equipment load through preset thresholds. When multiple exhibits require high computing power support at the same time (such as holographic projection and AR interactive superposition scenes), the system often freezes or crashes due to resource competition; third, the interactive experience is highly homogenized, and the collection of audience behavior data is limited to simple counting. It is impossible to build individual interest portraits, nor can it dynamically optimize content combinations according to group characteristics, making it difficult for one-size-fits-all display content to meet differentiated needs.

[0004] Deeper technical bottlenecks are reflected in: the lack of a unified scheduling perspective for exhibit clusters, and the existing solutions can only achieve device-level monitoring rather than system-level optimization; the lack of environmental perception capabilities, and the failure to establish a multi-dimensional correlation model of audience behavior, exhibit status, and spatial parameters; the lagging content adaptation mechanism still uses a passive "preset-play" mode rather than active optimization based on real-time data. These defects directly lead to high exhibition operating costs, difficulty in evaluating display effects, and limited audience stay time. These have become the core pain points restricting the construction of smart exhibition halls. To this end, a clustered management method and system for intelligent exhibit modules that support multi-scene switching are proposed. Summary of the Invention

[0005] In view of the deficiencies in the prior art, the present invention provides a clustered management method and system for smart exhibition item modules that support multi-scene switching, so as to solve the problems raised in the above-mentioned background technology.

[0006] To achieve the above object, the present invention provides the following technical solution: a cluster management method for smart exhibition module supporting multi-scene switching, comprising the following steps: Step 1: Real-time status collection: Through the data collection module deployed in each exhibition module, the exhibition operation status, audience interaction data and environmental parameters are obtained at a preset sampling period; Step 2: Dynamic strategy generation: Based on the data collected in step 1, the data analysis module uses a multi-dimensional evaluation algorithm to calculate the load rate of exhibits and the audience interest index. The strategy generation module combines the scenario rule library with the audience interest model to generate an optimized playback strategy through a weighted decision tree. Step 3: Update and distribute content: The instruction issuing module uses the incremental update protocol to push the difference content data package to the designated exhibition item module, and triggers the hot update mechanism of the content rendering unit to perform content synchronous switching; Step 4: Operation monitoring and troubleshooting: The monitoring and maintenance module receives the status data of the display item through a dual-channel redundant link. When abnormal parameters are detected, it automatically matches the three-level response mechanism and simultaneously records the fault log; Step 5: Multi-scenario adaptation: Dynamically match the combination and playback order of exhibit content based on the preset scenario rule library, and achieve cluster-level content synchronization through inter-exhibit communication protocols, supporting smooth switching of exhibition themes within preset thresholds; Step 6: Audience Preference Learning: A deep learning algorithm is used to construct a spatiotemporal distribution model of audience interests. The model output weight parameters are fed back to the strategy generation module in real time, forming a closed-loop control system of "data collection-strategy optimization-content iteration"; The data acquisition module uses embedded sensors and wireless communication technology to collect data such as the CPU temperature, memory usage, visitor stay time (calculated by UWB base station positioning), and ambient light intensity of the exhibit at a preset period (e.g., every minute), and transmits it to the central control layer through an encrypted protocol. The dynamic strategy generation mechanism combines a multi-dimensional evaluation algorithm with a weighted decision tree, which can flexibly respond to changes in audience interests in different scenarios, optimize playback strategies, and enhance the attractiveness and interactivity of exhibitions. The content update is distributed using an incremental update protocol, which effectively reduces network load and improves the efficiency and accuracy of content updates. The operation monitoring and fault handling mechanism ensures the stable operation and rapid fault recovery of the system through dual-channel redundant links and a three-level response mechanism. The multi-scenario adaptation function supports smooth switching of exhibition themes and meets the diverse needs of exhibitions. The audience preference learning mechanism uses a deep learning algorithm to construct an audience interest model to form a closed-loop control system, continuously optimize exhibition content, and enhance the audience experience.

[0007] Preferably, the data analysis module in step 2 adopts a weighted scoring mechanism to generate a priority index based on the load rate of the exhibits, the duration of visitor stay and the ambient light intensity; The weighted coefficients of the weighted scoring mechanism are adjusted in real time through a dynamic calibration algorithm, based on historical operating data of the exhibit, current visitor traffic peaks, and exhibition theme type. The exhibit load rate calculation incorporates device temperature parameters as implicit constraints. When the temperature of an exhibit module exceeds a threshold, its content scheduling priority is automatically lowered. Visitor dwell time is calculated using a spatial positioning algorithm, filtering invalid dwell records through UWB base station data and only counting dwell time within the effective viewing area. First, real-time data on exhibit load rate, visitor dwell time, and ambient light intensity is collected. Then, a dynamic calibration algorithm is used to calculate the weighted coefficients of each parameter based on the exhibit's historical operating data, current visitor traffic peaks, and exhibition theme type. Equipment temperature parameters are incorporated into the calculation of exhibit load rate. If the temperature exceeds a preset threshold, the content scheduling priority of that exhibit is lowered. Visitor dwell time is spatially located using UWB base stations, filtering out invalid stays to ensure that only dwell time within the effective viewing area is counted. Ultimately, a comprehensive priority index is generated. By comprehensively considering the exhibit load rate, visitor stay duration and ambient light intensity, dynamic optimization allocation of exhibit resources is achieved. The dynamic calibration algorithm of the weighted coefficient can be adjusted in real time according to the historical operation data of the exhibit, the current visitor flow peak and the exhibition theme type, ensuring the accuracy and adaptability of the scoring mechanism. The exhibit load rate calculation incorporates the equipment temperature parameters, effectively preventing performance degradation caused by overheating and improving the stability of the system. The precise calculation of the visitor stay duration is based on the spatial positioning technology of the UWB base station, filtering out invalid stays, ensuring the validity of the data, and providing a reliable basis for content scheduling. Overall, this mechanism helps to enhance the audience experience, optimize the utilization of exhibit resources, and realize intelligent management.

[0008] Preferably, the scene rule library in step 5 supports two modes: manual configuration and automatic learning. The manual configuration includes a visual rule editor. The visual rule editor supports drag-and-drop setting of scene switching conditions and linkage relationships between exhibits. The visual rule editor has a built-in exhibition industry template library. The automatic learning is based on a reinforcement learning algorithm, and the scene switching strategy network is trained through historical display data. The network input parameters include the density of audience flow trajectory, the frequency distribution of exhibit usage, and holiday feature labels. The scene switching condition setting includes the exhibit cluster-level timing constraints to ensure that the timing deviation of the synchronous switching of multi-module content is ≤50ms. In the manual configuration of the scenario rule library, the visual rule editor displays the layout and linkage relationships of exhibit items through a graphical interface. Users can drag and drop exhibit item icons to set switching conditions, such as time thresholds and visitor flow thresholds, and quickly build rules by selecting common templates for the exhibition industry from the built-in template library. In automatic learning mode, the reinforcement learning algorithm analyzes historical exhibition data, combines visitor flow trajectories, exhibit usage frequency, and holiday characteristics, and dynamically optimizes the scenario switching strategy to ensure that the timing deviation of synchronous switching of multi-module content is strictly controlled within 50ms. High-precision synchronization is maintained through a real-time calibration mechanism. The manual configuration mode simplifies the rule setting process through a visual editor. Users can quickly build scene switching logic that meets exhibition needs without any programming knowledge. The built-in industry template library further improves configuration efficiency and professionalism. The automatic learning mode uses reinforcement learning algorithms to continuously optimize switching strategies based on historical data to ensure that exhibit content can be dynamically adjusted according to audience behavior and exhibition rhythm to achieve personalized display. At the same time, strict timing constraints ensure the accuracy of synchronous switching of multi-module content, avoid display incoherence caused by timing deviations, and provide audiences with a smooth and immersive exhibition experience.

[0009] Preferably, the deep learning algorithm in step 6 adopts an LSTM network structure, and the input parameters include audience age distribution, exhibition item click heat map and social media keywords; The audience age distribution data is obtained through the face recognition system at the exhibition area entrance and is fuzzified to protect privacy. The display item click heat map is generated using a multi-level regional division strategy, dividing the display item screen into a functional operation area and a content display area, and calculating the click density of each area. The social media keyword processing includes a sentiment analysis module, which uses the BERT model to extract the sentiment polarity in audience comments and feeds it back into the audience interest model in step 2 as a content preference correction coefficient. During the LSTM network training phase, gradient clipping is used to prevent gradient explosion, and an early stopping mechanism is implemented to avoid overfitting. Furthermore, a sentiment dictionary is established to analyze the sentiment of social media keywords as auxiliary labels for the BERT model, improving the accuracy of sentiment polarity judgment. The LSTM network structure is used to process multiple data such as audience age distribution, exhibition item click heat map and social media keywords, which can accurately capture the temporal changes in audience interests. The age distribution is obtained and blurred through the face recognition system, which not only protects privacy but also provides valid data; the multi-level regional division strategy generates click heat map, which accurately reflects the audience's interactive behavior; combined with the sentiment tendency analysis of the BERT model, it can adjust content preferences in real time to form a closed-loop optimization.

[0010] Preferably, the three-level response mechanism in step 4 includes: In the event of a Level 1 fault, a data snapshot backup is performed before the system automatically restarts, and the integrity of key configuration files is automatically verified after the system restarts. When switching to the backup module for a Level 2 fault, dual-machine hot standby synchronization technology is used to achieve frame-level synchronization of playback content. When a Level 3 fault triggers manual intervention, the system automatically generates a fault diagnosis report containing hardware logs, network packet capture data, and a flowchart of recommended operations. The diagnosis report is synchronized to the operation and maintenance terminal via the inter-exhibit communication protocol. Before automatically restarting in the event of a Level 1 fault, the system automatically performs a data snapshot backup using a preset script, covering critical business data and configuration files. In the event of a Level 2 fault, the backup module uses dual-machine hot standby technology to achieve millisecond-level switching, ensuring seamless playback of content. After a Level 3 fault is triggered, the system automatically collects hardware logs and network packet capture data, and generates an action suggestion flowchart based on preset rules. This flowchart is pushed to the operation and maintenance terminal in real time via the inter-exhibit communication protocol to guide manual intervention. The three-level response mechanism provides a comprehensive and efficient troubleshooting solution for the clustered management of smart exhibition modules. Data snapshot backup and key configuration file verification for level one faults ensure data consistency and integrity after system restart, reducing the risk of data loss. For level two faults, dual-machine hot standby synchronization technology ensures the continuity of playback content and seamless audience experience through millisecond-level switching. In the event of a level three fault, the system automatically generates a fault diagnosis report that provides detailed fault information and operational suggestions to operation and maintenance personnel, greatly shortening the troubleshooting and repair time.

[0011] Preferably, the incremental update protocol in step 3 only transmits differential content data packets to reduce network load; the incremental update protocol includes differential content identification based on a binary hash comparison algorithm, triggering an update only when the file MD5 value changes; data packet fragmentation transmission uses variable length encoding technology, and the fragment size is dynamically adjusted according to the network bandwidth; the receiving end sets a cache preloading mechanism to pre-download frequently switched content segments during non-peak hours, and the preloading priority is dynamically adjusted by the priority index in step 2; In the incremental update protocol of step three, in addition to identifying content differences based on a binary hash comparison algorithm (MD5 value changes trigger updates), file size comparison can also be added as an auxiliary verification. If the MD5 values ​​are consistent but the file size is abnormal, re-verification or a full update process is triggered. Variable-length encoding technology for fragmented data packet transmission can optimize the fragmentation threshold based on historical bandwidth data. The receiving-end cache preloading mechanism dynamically adjusts the preloading time window for high-frequency content segments by analyzing the audience's traffic cycle, ensuring efficient resource utilization during non-peak hours. This incremental update protocol significantly reduces network load by transmitting only differential content data packets, avoiding bandwidth waste caused by complete content updates. Differential content identification is based on a binary hash comparison algorithm to ensure the accuracy and efficiency of updates. Updates are triggered only when the file MD5 value changes, avoiding unnecessary update operations. Data packet fragmentation transmission uses variable length coding technology to dynamically adjust the fragment size according to network bandwidth, further optimizing transmission efficiency. The cache preloading mechanism set at the receiving end pre-downloads high-frequency switching content fragments during non-peak hours, improving the response speed of content switching. The preloading priority is dynamically adjusted by the priority index to ensure the rational allocation of resources.

[0012] The system for cluster management of intelligent exhibition modules supporting multi-scene switching adopts the above-mentioned cluster management method for intelligent exhibition modules supporting multi-scene switching, including: A data acquisition layer, a central control layer, an execution terminal layer, and a monitoring and maintenance layer, wherein the output end of the data acquisition layer is electrically connected to the input end of the central control layer, the central control layer is electrically connected to the execution terminal layer, and the central control layer is electrically connected to the monitoring and maintenance layer; The data acquisition layer is composed of a data acquisition module composed of a distributed sensor array, which is embedded in each exhibit body and the exhibition area environment; The central control layer integrates an industrial-grade control host with a data analysis module, a strategy generation module, and an instruction issuing module; The execution terminal layer configures a display module cluster of the intelligent playback terminal, each terminal including a content rendering unit and a state feedback unit; The monitoring and maintenance layer is managed by a visual operation and maintenance platform managed by the monitoring and maintenance module and deployed on the control center large screen and mobile terminals; The data collection layer can achieve multi-dimensional data collection by embedding temperature, humidity, and light sensors and audience positioning UWB base stations in the exhibits themselves, and deploying air quality and crowd counters in the exhibition environment. The central control layer uses an industrial-grade host and integrates data analysis (Python / TensorFlow), strategy generation (decision tree / reinforcement learning), and command issuance (MQTT protocol) modules. The execution terminal layer uses a smart playback terminal cluster with a built-in embedded Linux system and GPU rendering unit to support 4K video decoding. The monitoring and maintenance layer deploys a visualization platform and integrates AR remote collaboration and fault prediction algorithms to achieve full-link monitoring. The data collection layer uses a distributed sensor array to accurately capture the operating status and environmental parameters of exhibits, providing data support for strategy generation. The central control layer integrates multiple modules and dynamically generates optimized playback strategies through multi-dimensional evaluation algorithms and weighted decision trees to ensure that the content is highly matched with the audience's interests. The execution terminal layer configures an intelligent playback cluster, supports a hot update mechanism, and realizes rapid and synchronous switching of content.

[0013] Preferably, the data acquisition layer is equipped with a self-calibration module, and each node of the data acquisition layer is equipped with a redundant sensor group to support automatic switching of the backup acquisition channel when data is abnormal; The self-calibration module regularly calibrates sensor outputs using a built-in standard signal source to ensure data accuracy. When the primary sensor data is abnormal, the system automatically detects and switches to the backup sensor in the redundant sensor group, while recording the switching event to ensure the continuity and reliability of data collection. The self-calibration module can regularly calibrate sensors to reduce measurement errors caused by long-term use or environmental changes, ensuring data accuracy. The setting of the redundant sensor group provides a backup channel for data collection. When the main sensor fails, the system can quickly switch to the backup sensor to ensure the continuity of data collection and avoid affecting the normal operation of the entire smart exhibition module cluster management system due to data interruption.

[0014] Preferably, the execution terminal layer has a built-in liquid cooling and heat dissipation system, which is connected to the central control layer through a thermal management module and supports dynamic speed control; The liquid cooling system consists of a high-efficiency liquid cooling plate, a circulation pump, a radiator, and a thermal management module. The liquid cooling plate is attached to the heat-generating components of the exhibit module. The circulation pump drives the coolant circulation, which dissipates heat through the radiator. The thermal management module adjusts the circulation pump speed in real time according to the exhibit temperature, achieving dynamic speed control to ensure stable operation of the exhibit. Compared with traditional air cooling, the liquid cooling system can remove heat more efficiently, lower the operating temperature of the exhibition module, and reduce the risk of performance degradation or damage due to overheating. The dynamic speed control automatically adjusts the cooling efficiency according to the actual workload of the exhibit, which not only ensures the cooling effect, but also avoids energy waste and improves overall energy efficiency. In addition, the design also enhances the maintainability and scalability of the system, facilitating future upgrades or optimizations of the cooling system, and providing a solid hardware foundation for the clustered management of intelligent exhibition modules.

[0015] Preferably, the monitoring and maintenance layer includes an AR module, the AR module integrates a six-degree-of-freedom positioning sensor, the console is provided with a mechanical emergency switch, and the protective shell meets the IP65 protection grade; The AR module uses a built-in six-degree-of-freedom positioning sensor to capture the operator's spatial position and posture information in real time. Combined with the three-dimensional model of the exhibit, it presents a virtual and real-world operation and maintenance guidance interface on mobile terminals or the large screen of the control center. Through this interface, operators can intuitively view the internal structure of the exhibit, locate fault points, and receive repair steps, improving troubleshooting efficiency. The monitoring and maintenance layer integrates AR modules, which significantly improves the intelligence and efficiency of operation and maintenance work. The AR module provides operation and maintenance personnel with accurate spatial positioning and posture perception through six-degree-of-freedom positioning sensors. Combined with the three-dimensional model of the exhibits, it realizes the intuitive integration of virtual and reality. This intuitive operation and maintenance guidance interface enables operation and maintenance personnel to quickly locate the fault point and reduce the troubleshooting time. At the same time, the maintenance step prompts provided by the AR module reduce the difficulty of operation and improve the accuracy of maintenance. In addition, the mechanical emergency switch set on the console can quickly cut off the power supply in an emergency to ensure the safety of personnel; the protective casing meets the IP65 protection level, ensuring the stable operation of the equipment in harsh environments and extending the service life of the equipment.

[0016] In summary, compared with the prior art, the present invention provides a clustered management method and system for smart exhibition modules that supports multi-scene switching, which has the following beneficial effects: This invention realizes intelligent resource scheduling and dynamic content adaptation of exhibition clusters through a collaborative mechanism of real-time status collection and dynamic strategy generation. The rigid resource allocation problem caused by the independent operation of traditional exhibition items is solved through the multi-dimensional load evaluation algorithm of the data analysis module in the present invention. The system can dynamically adjust the content playback strategy based on the real-time fusion calculation of the exhibition item load rate, audience flow and environmental parameters, effectively avoiding system jams caused by resource competition in scenarios with high loads of multiple exhibition items at the same time, shortening the content switching response cycle, and improving cluster operation efficiency. Through the deep integration of the audience interest spatiotemporal distribution model and the incremental update protocol, a differentiated interactive display system has been constructed. Traditional solutions can only collect extensive audience data, while the audience behavior portrait constructed by the present invention using a deep learning algorithm can capture the migration trajectory of group interests in real time. The hot update mechanism of the content rendering unit is used to achieve local content replacement, avoiding the transmission pressure caused by the full update. In combination with the cluster-level content synchronization achieved by the inter-exhibit communication protocol, the logical connection between adjacent exhibits is smooth, forming an immersive dynamic narrative flow, breaking through the bottleneck of traditional display content homogeneity. Through dual-channel redundant monitoring and a three-level fault response system, an intelligent operation and maintenance closed loop of the exhibition cluster has been established. The traditional system relies on a passive maintenance mode of manual inspections, which has been upgraded to predictive active maintenance in this invention: the monitoring module ensures the integrity of status data through a dual-link redundant design. When equipment abnormalities are detected, the three-level response mechanism automatically executes a gradient processing process of frequency reduction, standby module switching, and engineer dispatch. Combined with the fault tree analysis function of the visual operation and maintenance platform, it can achieve rapid root cause location and significantly reduce equipment downtime rate, comprehensively improve system stability and reduce operation and maintenance labor costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1This is a step diagram of the invention's method for clustering management of intelligent exhibition modules that supports multi-scene switching.

[0018] Figure 2 This is a system diagram of the invention's clustered management of intelligent exhibition modules that supports multi-scene switching. DETAILED DESCRIPTION

[0019] The present invention provides a technical solution to support the cluster management method of intelligent exhibition module with multi-scene switching. Figure 1 , including the following steps: Step 1: Real-time status collection: Through the data collection module deployed in each exhibition module, the exhibition operation status, audience interaction data and environmental parameters are obtained at a preset sampling period; Step 2: Dynamic strategy generation: Based on the data collected in step 1, the data analysis module uses a multi-dimensional evaluation algorithm to calculate the load rate of exhibits and the audience interest index. The strategy generation module combines the scenario rule library with the audience interest model to generate an optimized playback strategy through a weighted decision tree. Step 3: Update and distribute content: The instruction issuing module uses the incremental update protocol to push the difference content data package to the designated exhibition item module, and triggers the hot update mechanism of the content rendering unit to perform content synchronous switching; Step 4: Operation monitoring and troubleshooting: The monitoring and maintenance module receives the status data of the display item through a dual-channel redundant link. When abnormal parameters are detected, it automatically matches the three-level response mechanism and simultaneously records the fault log; Step 5: Multi-scenario adaptation: Dynamically match the combination and playback order of exhibit content based on the preset scenario rule library, and achieve cluster-level content synchronization through inter-exhibit communication protocols, supporting smooth switching of exhibition themes within preset thresholds; Step 6: Audience Preference Learning: A deep learning algorithm is used to construct a spatiotemporal distribution model of audience interests. The model output weight parameters are fed back to the strategy generation module in real time, forming a closed-loop control system of "data collection-strategy optimization-content iteration"; The data acquisition module uses embedded sensors and wireless communication technology to collect data such as the CPU temperature, memory usage, visitor stay time (calculated by UWB base station positioning), and ambient light intensity of the exhibit at a preset period (e.g., every minute), and transmits it to the central control layer through an encrypted protocol. The dynamic strategy generation mechanism combines a multi-dimensional evaluation algorithm with a weighted decision tree, which can flexibly respond to changes in audience interests in different scenarios, optimize playback strategies, and enhance the attractiveness and interactivity of exhibitions. The content update is distributed using an incremental update protocol, which effectively reduces network load and improves the efficiency and accuracy of content updates. The operation monitoring and fault handling mechanism ensures the stable operation and rapid fault recovery of the system through dual-channel redundant links and a three-level response mechanism. The multi-scenario adaptation function supports smooth switching of exhibition themes and meets the diverse needs of exhibitions. The audience preference learning mechanism uses a deep learning algorithm to construct an audience interest model to form a closed-loop control system, continuously optimize exhibition content, and enhance the audience experience.

[0020] See also Figure 1 ,The data analysis module in step 2 adopts a weighted scoring mechanism, ,which comprehensively considers the load rate of exhibits, the length of ,visitors’ stay and the ambient light intensity to generate a priority ,index; The weighted coefficients of the weighted scoring mechanism are adjusted in real time through a dynamic calibration algorithm, based on historical operating data of the exhibit, current visitor traffic peaks, and exhibition theme type. The exhibit load rate calculation incorporates device temperature parameters as implicit constraints. When the temperature of an exhibit module exceeds the threshold, its content scheduling priority is automatically lowered. Visitor dwell time is calculated using a spatial positioning algorithm, filtering invalid dwell records through UWB base station data and only counting dwell time within the effective viewing area. First, real-time data on exhibit load rate, visitor dwell time, and ambient light intensity is collected. Then, a dynamic calibration algorithm is used to calculate the weighted coefficients of each parameter based on the exhibit's historical operating data, current visitor traffic peaks, and exhibition theme type. Equipment temperature parameters are incorporated into the calculation of exhibit load rate. If the temperature exceeds a preset threshold, the content scheduling priority of that exhibit is lowered. Visitor dwell time is spatially located using UWB base stations, filtering out invalid stays to ensure that only dwell time within the effective viewing area is counted. Ultimately, a comprehensive priority index is generated. By comprehensively considering the exhibit load rate, visitor stay duration and ambient light intensity, dynamic optimization allocation of exhibit resources is achieved. The dynamic calibration algorithm of the weighted coefficient can be adjusted in real time according to the historical operation data of the exhibit, the current visitor flow peak and the exhibition theme type, ensuring the accuracy and adaptability of the scoring mechanism. The exhibit load rate calculation incorporates the equipment temperature parameters, effectively preventing performance degradation caused by overheating and improving the stability of the system. The precise calculation of the visitor stay duration is based on the spatial positioning technology of the UWB base station, filtering out invalid stays, ensuring the validity of the data, and providing a reliable basis for content scheduling. Overall, this mechanism helps to enhance the audience experience, optimize the utilization of exhibit resources, and realize intelligent management.

[0021] See also Figure 1The scene rule library in step 5 supports two modes: manual configuration and automatic learning. Manual configuration includes a visual rule editor. The visual rule editor supports drag-and-drop setting of scene switching conditions and exhibition item linkage relationships. The visual rule editor has a built-in exhibition industry template library. Automatic learning is based on a reinforcement learning algorithm, using historical display data to train a scene switching strategy network. Network input parameters include audience flow density, exhibit usage frequency distribution, and holiday feature labels. The scene switching condition settings include exhibit cluster-level timing constraints to ensure that the timing deviation of synchronous switching of multi-module content is ≤50ms. In the manual configuration of the scenario rule library, the visual rule editor displays the layout and linkage relationships of exhibit items through a graphical interface. Users can drag and drop exhibit item icons to set switching conditions, such as time thresholds and visitor flow thresholds, and quickly build rules by selecting common templates for the exhibition industry from the built-in template library. In automatic learning mode, the reinforcement learning algorithm analyzes historical exhibition data, combines visitor flow trajectories, exhibit usage frequency, and holiday characteristics, and dynamically optimizes the scenario switching strategy to ensure that the timing deviation of synchronous switching of multi-module content is strictly controlled within 50ms. High-precision synchronization is maintained through a real-time calibration mechanism. The manual configuration mode simplifies the rule setting process through a visual editor. Users can quickly build scene switching logic that meets exhibition needs without any programming knowledge. The built-in industry template library further improves configuration efficiency and professionalism. The automatic learning mode uses reinforcement learning algorithms to continuously optimize switching strategies based on historical data to ensure that exhibit content can be dynamically adjusted according to audience behavior and exhibition rhythm to achieve personalized display. At the same time, strict timing constraints ensure the accuracy of synchronous switching of multi-module content, avoid display incoherence caused by timing deviations, and provide audiences with a smooth and immersive exhibition experience.

[0022] See also Figure 1 ,The deep learning algorithm in step six adopts the LSTM network structure, and the input ,parameters include audience age distribution, exhibition item click heat map, and social media keywords; Visitor age distribution data was obtained through the facial recognition system at the exhibition entrance and fuzzified to protect privacy. A multi-level zone partitioning strategy was used to generate the heat map for clicks on exhibit items, dividing the display screen into a functional operation area and a content display area, and calculating click density for each. Social media keyword processing included a sentiment analysis module, which used the BERT model to extract the sentiment polarity from visitor comments and used it as a content preference correction coefficient to feed back into the visitor interest model in step two. During the LSTM network training phase, gradient clipping is used to prevent gradient explosion, and an early stopping mechanism is implemented to avoid overfitting. Furthermore, a sentiment dictionary is established to analyze the sentiment of social media keywords as auxiliary labels for the BERT model, improving the accuracy of sentiment polarity judgment. The LSTM network structure is used to process multiple data such as audience age distribution, exhibition item click heat map and social media keywords, which can accurately capture the temporal changes in audience interests. The age distribution is obtained and blurred through the face recognition system, which not only protects privacy but also provides valid data; the multi-level regional division strategy generates click heat map, which accurately reflects the audience's interactive behavior; combined with the sentiment tendency analysis of the BERT model, it can adjust content preferences in real time to form a closed-loop optimization.

[0023] See also Figure 1 , the three-level response mechanism in step 4 includes: In the event of a Level 1 fault, a data snapshot backup is performed before the system automatically restarts, and the integrity of key configuration files is automatically verified after the system restarts. When switching to the backup module for a Level 2 fault, dual-machine hot standby synchronization technology is used to achieve frame-level synchronization of playback content. When a Level 3 fault triggers manual intervention, the system automatically generates a fault diagnosis report containing hardware logs, network packet capture data, and a flowchart of recommended operations. The diagnosis report is synchronized to the operation and maintenance terminal via the inter-exhibit communication protocol. Before automatically restarting in the event of a Level 1 fault, the system automatically performs a data snapshot backup using a preset script, covering critical business data and configuration files. In the event of a Level 2 fault, the backup module uses dual-machine hot standby technology to achieve millisecond-level switching, ensuring seamless playback of content. After a Level 3 fault is triggered, the system automatically collects hardware logs and network packet capture data, and generates an action suggestion flowchart based on preset rules. This flowchart is pushed to the operation and maintenance terminal in real time via the inter-exhibit communication protocol to guide manual intervention. The three-level response mechanism provides a comprehensive and efficient troubleshooting solution for the clustered management of smart exhibition modules. Data snapshot backup and key configuration file verification for level one faults ensure data consistency and integrity after system restart, reducing the risk of data loss. For level two faults, dual-machine hot standby synchronization technology ensures the continuity of playback content and seamless audience experience through millisecond-level switching. In the event of a level three fault, the system automatically generates a fault diagnosis report that provides detailed fault information and operational suggestions to operation and maintenance personnel, greatly shortening the troubleshooting and repair time.

[0024] See also Figure 1 The incremental update protocol in step 3 only transmits differential content data packets to reduce network load; the incremental update protocol includes differential content identification, which is based on a binary hash comparison algorithm and triggers an update only when the file MD5 value changes; the data packet fragment transmission adopts variable length coding technology and dynamically adjusts the fragment size according to the network bandwidth; the receiving end sets a cache preloading mechanism to pre-download high-frequency switching content fragments during non-peak hours, and the preloading priority is dynamically adjusted by the priority index in step 2; In the incremental update protocol of step three, in addition to identifying content differences based on a binary hash comparison algorithm (MD5 value changes trigger updates), file size comparison can also be added as an auxiliary verification. If the MD5 values ​​are consistent but the file size is abnormal, re-verification or a full update process is triggered. Variable-length encoding technology for fragmented data packet transmission can optimize the fragmentation threshold based on historical bandwidth data. The receiving-end cache preloading mechanism dynamically adjusts the preloading time window for high-frequency content segments by analyzing the audience's traffic cycle, ensuring efficient resource utilization during non-peak hours. This incremental update protocol significantly reduces network load by transmitting only differential content data packets, avoiding bandwidth waste caused by complete content updates. Differential content identification is based on a binary hash comparison algorithm to ensure the accuracy and efficiency of updates. Updates are triggered only when the file MD5 value changes, avoiding unnecessary update operations. Data packet fragmentation transmission uses variable length coding technology to dynamically adjust the fragment size according to network bandwidth, further optimizing transmission efficiency. The cache preloading mechanism set at the receiving end pre-downloads high-frequency switching content fragments during non-peak hours, improving the response speed of content switching. The preloading priority is dynamically adjusted by the priority index to ensure the rational allocation of resources.

[0025] The system for cluster management of intelligent exhibition modules that supports multi-scene switching adopts the above-mentioned cluster management method for intelligent exhibition modules that supports multi-scene switching. Figure 1 and Figure 2 ,include: Data acquisition layer, central control layer, execution terminal layer and monitoring and maintenance layer, the output end of the data acquisition layer is electrically connected to the input end of the central control layer, the central control layer is electrically connected to the execution terminal layer, and the central control layer is electrically connected to the monitoring and maintenance layer; The data acquisition layer consists of a data acquisition module composed of a distributed sensor array, which is embedded in each exhibit and the exhibition environment; The central control layer integrates an industrial-grade control host with data analysis modules, strategy generation modules, and instruction issuance modules; The terminal layer configures the display module cluster of the intelligent playback terminal, each terminal includes a content rendering unit and a status feedback unit; The monitoring and maintenance layer is managed by a visual operation and maintenance platform managed by the monitoring and maintenance module and deployed on the control center's large screen and mobile terminals; The data collection layer can achieve multi-dimensional data collection by embedding temperature, humidity, and light sensors and audience positioning UWB base stations in the exhibits themselves, and deploying air quality and crowd counters in the exhibition environment. The central control layer uses an industrial-grade host and integrates data analysis (Python / TensorFlow), strategy generation (decision tree / reinforcement learning), and command issuance (MQTT protocol) modules. The execution terminal layer uses a smart playback terminal cluster with a built-in embedded Linux system and GPU rendering unit to support 4K video decoding. The monitoring and maintenance layer deploys a visualization platform and integrates AR remote collaboration and fault prediction algorithms to achieve full-link monitoring. The data collection layer uses a distributed sensor array to accurately capture the operating status and environmental parameters of exhibits, providing data support for strategy generation. The central control layer integrates multiple modules and dynamically generates optimized playback strategies through multi-dimensional evaluation algorithms and weighted decision trees to ensure that the content is highly matched with the audience's interests. The execution terminal layer configures an intelligent playback cluster, supports a hot update mechanism, and realizes rapid and synchronous switching of content.

[0026] See also Figure 1 and Figure 2 ,The data acquisition layer is equipped with a self-calibration module, and each node of the data acquisition layer is equipped with a redundant sensor group, which supports automatic switching of backup acquisition channels when data is abnormal; The self-calibration module regularly calibrates sensor outputs using a built-in standard signal source to ensure data accuracy. When the primary sensor data is abnormal, the system automatically detects and switches to the backup sensor in the redundant sensor group, while recording the switching event to ensure the continuity and reliability of data collection. The self-calibration module can regularly calibrate sensors to reduce measurement errors caused by long-term use or environmental changes, ensuring data accuracy. The setting of the redundant sensor group provides a backup channel for data collection. When the main sensor fails, the system can quickly switch to the backup sensor to ensure the continuity of data collection and avoid affecting the normal operation of the entire smart exhibition module cluster management system due to data interruption.

[0027] See also Figure 1 and Figure 2 ,The execution terminal layer has a built-in liquid cooling system, which is connected to the central control layer through a thermal management module and supports dynamic speed control; The liquid cooling system consists of a high-efficiency liquid cooling plate, a circulation pump, a radiator, and a thermal management module. The liquid cooling plate is attached to the heat-generating components of the exhibit module. The circulation pump drives the coolant circulation, which dissipates heat through the radiator. The thermal management module adjusts the circulation pump speed in real time according to the exhibit temperature, achieving dynamic speed control to ensure stable operation of the exhibit. Compared with traditional air cooling, the liquid cooling system can remove heat more efficiently, lower the operating temperature of the exhibition module, and reduce the risk of performance degradation or damage due to overheating. The dynamic speed control automatically adjusts the cooling efficiency according to the actual workload of the exhibit, which not only ensures the cooling effect, but also avoids energy waste and improves overall energy efficiency. In addition, the design also enhances the maintainability and scalability of the system, facilitating future upgrades or optimizations of the cooling system, and providing a solid hardware foundation for the clustered management of intelligent exhibition modules.

[0028] See also Figure 1 and Figure 2 ,The monitoring and maintenance layer includes an AR module, which integrates a six-degree-of-freedom positioning sensor, a mechanical emergency switch is set on the console, and the protective shell meets the IP65 protection grade; The AR module uses a built-in six-degree-of-freedom positioning sensor to capture the operator's spatial position and posture information in real time. Combined with the three-dimensional model of the exhibit, it presents a virtual and real-world operation and maintenance guidance interface on mobile terminals or the large screen of the control center. Through this interface, operators can intuitively view the internal structure of the exhibit, locate fault points, and receive repair steps, improving troubleshooting efficiency. The monitoring and maintenance layer integrates AR modules, which significantly improves the intelligence and efficiency of operation and maintenance work. The AR module provides operation and maintenance personnel with accurate spatial positioning and posture perception through six-degree-of-freedom positioning sensors. Combined with the three-dimensional model of the exhibits, it realizes the intuitive integration of virtual and reality. This intuitive operation and maintenance guidance interface enables operation and maintenance personnel to quickly locate the fault point and reduce the troubleshooting time. At the same time, the maintenance step prompts provided by the AR module reduce the difficulty of operation and improve the accuracy of maintenance. In addition, the mechanical emergency switch set on the console can quickly cut off the power supply in an emergency to ensure the safety of personnel; the protective casing meets the IP65 protection level, ensuring the stable operation of the equipment in harsh environments and extending the service life of the equipment.

[0029] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that includes a list of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus.

[0030] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A cluster management method for intelligent exhibition modules supporting multi-scene switching, characterized in that: The steps include: Step 1: Real-time status collection: Through the data collection module deployed in each exhibition module, the exhibition operation status, audience interaction data and environmental parameters are obtained at a preset sampling period; Step 2: Dynamic strategy generation: Based on the data collected in step 1, the data analysis module uses a multi-dimensional evaluation algorithm to calculate the load rate of exhibits and the audience interest index. The strategy generation module combines the scenario rule library with the audience interest model to generate an optimized playback strategy through a weighted decision tree. Step 3: Update and distribute content: The instruction issuing module uses the incremental update protocol to push the difference content data package to the designated exhibition item module, and triggers the hot update mechanism of the content rendering unit to perform content synchronous switching; Step 4: Operation monitoring and troubleshooting: The monitoring and maintenance module receives the status data of the display item through a dual-channel redundant link. When abnormal parameters are detected, it automatically matches the three-level response mechanism and simultaneously records the fault log; Step 5: Multi-scenario adaptation: Dynamically match the combination and playback order of exhibit content based on the preset scenario rule library, and achieve cluster-level content synchronization through the inter-exhibit communication protocol; Step 6: Audience Preference Learning: The audience interest spatiotemporal distribution model is constructed through deep learning algorithms, and the model output weight parameters are fed back to the strategy generation module in real time.

2. The method for cluster management of smart display modules supporting multi-scenario switching according to claim 1, characterized in that: The data analysis module in step 2 adopts a weighted scoring mechanism to generate a priority index based on the load rate of exhibits, the duration of visitor stay and the ambient light intensity; The weighted coefficients of the weighted scoring mechanism are adjusted in real time through a dynamic calibration algorithm, and the calibration is based on historical operation data of the exhibits, current peak visitor traffic, and exhibition theme type.

3. The method for cluster management of smart display modules supporting multi-scenario switching according to claim 1, characterized in that: The scenario rule library in step 5 supports two modes: manual configuration and automatic learning. The manual configuration includes a visual rule editor. The visual rule editor supports drag-and-drop setting of scene switching conditions and linkage relationships between exhibits. The visual rule editor has a built-in exhibition industry template library. The automatic learning is based on a reinforcement learning algorithm, and the scene switching strategy network is trained through historical display data. The network input parameters include the density of audience flow trajectory, the frequency distribution of exhibit usage, and holiday feature labels. The scene switching condition settings include cluster-level timing constraints for exhibits.

4. The method for cluster management of smart display modules supporting multi-scenario switching according to claim 1, characterized in that: The deep learning algorithm in step 6 adopts an LSTM network structure, and the input parameters include the age distribution of the audience, the heat map of the exhibition items clicked, and social media keywords; The audience age distribution data is obtained through the face recognition system at the exhibition area entrance and fuzzy processing is used to protect privacy. The display item click heat map is generated using a multi-level area division strategy, dividing the display item screen into a functional operation area and a content display area, and calculating the click density of each area. The social media keyword processing includes a sentiment analysis module, which extracts the sentiment polarity in audience comments through the BERT model and feeds it back to the audience interest model in step 2 as a content preference correction coefficient.

5. The cluster management method for intelligent exhibition module supporting multi-scene switching according to claim 1, characterized in that: The three-level response mechanism in step 4 includes: In the event of a level 1 fault, a data snapshot backup is performed before automatic restart, and the integrity of key configuration files is automatically verified after restart. When switching to the backup module for a level 2 fault, dual-machine hot standby synchronization technology is used to achieve frame-level synchronization of playback content. When a level 3 fault triggers manual intervention, the system automatically generates a fault diagnosis report, which includes hardware logs, network packet capture data, and an operation suggestion flowchart. The diagnosis report is synchronized to the operation and maintenance terminal through the inter-exhibit communication protocol.

6. The method for cluster management of smart display modules supporting multi-scenario switching according to claim 1, characterized in that: The incremental update protocol in step three only transmits differential content data packets. The incremental update protocol includes differential content identification based on a binary hash comparison algorithm, and triggers an update only when the file MD5 value changes.

7. A system for cluster management of smart display modules supporting multi-scene switching, adopting a cluster management method for smart display modules supporting multi-scene switching according to any one of claims 1 to 6, characterized in that: include: A data acquisition layer, a central control layer, an execution terminal layer, and a monitoring and maintenance layer, wherein the output end of the data acquisition layer is electrically connected to the input end of the central control layer, the central control layer is electrically connected to the execution terminal layer, and the central control layer is electrically connected to the monitoring and maintenance layer; The data acquisition layer is a data acquisition module composed of a distributed sensor array; The central control layer integrates an industrial-grade control host with a data analysis module, a strategy generation module, and an instruction issuing module; The execution terminal layer configures a display module cluster of the intelligent playback terminal, each terminal including a content rendering unit and a state feedback unit; The monitoring and maintenance layer is managed by a visual operation and maintenance platform managed by the monitoring and maintenance module and deployed on the large screen of the control center and mobile terminals.

8. The system for clustered management of intelligent exhibition modules supporting multi-scenario switching according to claim 7, characterized in that: The data acquisition layer is equipped with a self-calibration module, and each node of the data acquisition layer is equipped with a redundant sensor group.

9. The system for clustered management of intelligent exhibition modules supporting multi-scene switching according to claim 7, characterized in that: The execution terminal layer has a built-in liquid cooling and heat dissipation system, and the liquid cooling and heat dissipation system is connected to the central control layer through a thermal management module.

10. The system for clustered management of intelligent exhibition modules supporting multi-scene switching according to claim 7, characterized in that: The monitoring and maintenance layer includes an AR module, and the AR module integrates a six-degree-of-freedom positioning sensor.

Images