Cloud central control platform for intelligent digital exhibition hall

Abstract

The invention relates to the technical field of intelligent exhibition hall control systems, and discloses a cloud central control platform for an intelligent digital exhibition hall, and the platform comprises a cloud edge cooperative control module, a unified instruction analysis module, a control state feedback module, a linkage instruction generation module and a scene parameter synchronization module. Compared with a cloud exhibition hall which is single in exhibition hall control instruction mode, poor in linkage capability and high in instruction mistaken touch rate in the prior art, the technical problem that the control requirement of intelligent response is difficult to achieve especially under the conditions of dense audiences and complex scene switching is solved. According to the invention, by constructing a multi-modal input fusion mechanism, a semantic analysis and strategy model-based control instruction generation logic, an equipment state feedback closed-loop control mechanism and a cloud edge collaborative strategy optimization structure, multi-channel and scene adaptive combined control of an exhibition hall screen, lamplight, sound equipment and a guide system is realized; and the intelligent level of exhibition hall operation is improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of smart exhibition hall control systems, and in particular relates to a cloud-based central control platform for smart digital exhibition halls. Background Art

[0002] Digital exhibition halls are currently widely used in museums, science and technology museums, corporate pavilions, and other settings. They typically leverage cloud-based control platforms for centralized control of devices such as screens, lighting, audio, and guide systems. However, existing cloud-based control platforms often suffer from a single command input method, rigid control logic, high response latency, and weak system interoperability. These platforms struggle to meet the real-time, stable control requirements of exhibition halls, which face frequent interactions and complex scene switching. For example, in exhibition areas with dense crowds or frequent exhibit interaction, staff often need to frequently switch between different display modes and dynamically adjust device status (e.g., dimming, switching explanations, adjusting volume, etc.). However, existing control methods primarily rely on single-touch or preset timed scripts, lack support for multimodal input (such as voice, gestures, and sensors), and are unable to automatically adjust control strategies based on the exhibition hall's real-time status. This control approach is slow to respond to special scenarios such as sudden surges in visitor traffic, increased demand for exhibit interaction, or VIP receptions, and is prone to accidental operations, resulting in a poor visitor experience. Existing technologies cannot fully meet the comprehensive requirements of digital exhibition halls for low-latency response, highly intelligent linkage, and safe and fault-tolerant control under the conditions of "high interaction, high traffic, and complex scenarios." Therefore, there is an urgent need for an intelligent central control method with multimodal input fusion capabilities, a policy adaptive update mechanism, and a cloud-edge collaborative architecture. This method can achieve precise control and efficient linkage of exhibition hall equipment in complex and dynamic environments, thereby improving the intelligence level, stability, and interactive experience of digital exhibition hall operations. Summary of the Invention

[0003] In response to the above-mentioned technical deficiencies, the purpose of the present invention is to propose a cloud central control platform and method for smart digital exhibition halls, aiming to solve the technical problems of cloud exhibition halls in the existing technology with a single exhibition hall control command method, poor linkage capability and high command false trigger rate, especially under conditions of dense audiences and complex scene switching, which make it difficult to achieve intelligent response control requirements.

[0004] In order to solve the above technical problems, the present invention adopts the following technical solutions: The present invention provides a cloud-based control platform for a smart digital exhibition hall.

[0005] The cloud-based central control platform for the smart digital exhibition hall includes:

[0006] The cloud-edge collaborative control module is used to deploy the edge central control server locally in the exhibition hall and the policy optimization large model server in the cloud. It also uses the edge-cloud task division mechanism to establish a two-way communication channel between the edge central control server and the policy optimization large model server.

[0007] The unified instruction parsing module is used to collect multimodal input data, standardize the multimodal input data through the edge central control server, and obtain standard multimodal data; the standard multimodal data is sent to the policy optimization large model server through a two-way communication channel, and the policy optimization large model server processes the standard multimodal data and generates output structured control instructions;

[0008] The control state feedback module is used to execute structured control instructions, obtain the device status before and after the instruction execution in real time, calculate the instruction feedback deviation, and determine whether to trigger the fault-tolerant rollback mechanism;

[0009] The linkage command generation module is used to obtain the priority data and linkage dependency data of different control devices in the exhibition hall, and match the target scene mode based on the command feedback deviation , generate linkage instruction control sequence and execute it;

[0010] Scene parameter synchronization module, used to synchronize the target scene mode in the cloud The linkage command control sequence under is synchronized to the edge central control server through the cloud scheduling API interface.

[0011] Preferably, in the cloud-edge collaborative control module, the edge central control server is used to receive local multimodal instructions and structured control instructions fed back by the cloud, as well as to control the screen subsystem, lighting subsystem, audio subsystem and guide subsystem of the exhibition hall; the strategy optimization large model server is used to perform control strategy adaptive update tasks, AI command recognition tasks and crowd flow prediction tasks.

[0012] Preferably, in the unified instruction parsing module, the multimodal input data includes voice input data, exhibition hall touch terminal input data, exhibition hall mobile app input data and crowd density sensor input data; the output structured control instructions include screen subsystem control instructions, lighting subsystem control instructions, audio subsystem control instructions and guide subsystem control instructions.

[0013] Preferably, in the unified instruction parsing module, the step of standardizing the multimodal input data by the edge central control server specifically includes:

[0014] Receive multimodal input data through the edge control server, including voice input data, exhibition hall touch terminal input data, exhibition hall mobile app input data, and crowd density sensor input data;

[0015] Based on the multimodal input data, the input is classified into four basic types by data format, source port and context label, including voice input data , touch input data 、App input data , sensor input data , and generate a type identification code;

[0016] Constructing a unified input representation vector , including: For voice input data Use the lightweight speech recognition models BERT and CTC deployed locally on the edge control server to convert it into text instructions, and then extract the semantic vector of the speech input data; for touch input data and App input data Use predefined controls for one-hot coding and construct touch input data and App input data Semantic vector of sensor input data Directly build sensor input data for real-valued perception semantic vector of ; comprehensive speech input data , touch input data 、App input data , sensor input data The semantic vector of the unified input representation vector is constructed ;

[0017] Representing vectors for uniform input The value range differences of different dimensions in are first processed by linear normalization to obtain the optimized input representation vector;

[0018] Abnormal input detection and filtering are performed on the optimized input representation vector to obtain standardized multimodal input data.

[0019] Preferably, in the unified instruction parsing module, the step of processing the standard multimodal data by the strategy optimization large model server to generate and output structured control instructions specifically includes:

[0020] In the policy optimization large model server, the reinforcement learning structure DDPG is used to build the input control mapping model, and the standardized multimodal input data is mapped to the state space S of the input control mapping model. The input control mapping model policy function is expressed as , used to indicate that the Take action The probability distribution of action Used to represent a group of control instruction combinations;

[0021] Acquire historical multimodal input data and historical control instruction combinations, and train an input control mapping model based on the historical multimodal input data and historical control instruction combinations;

[0022] Will Input the trained input control mapping model, perform the inference operation, output the control instruction combination, and construct the structured control instruction according to the control instruction combination. The structured control instruction includes .

[0023] Preferably, in the control state feedback module, the steps of executing the structured control instruction, obtaining the device state before and after the instruction execution in real time, calculating the instruction feedback deviation, and determining whether to trigger the fault-tolerant rollback mechanism specifically include:

[0024] The output structured control instructions in the unified instruction parsing module are sent to the corresponding devices, and the state of each device at time t is Abstract in vector form, ,in, For the The current status of the device, including screen playback status, light brightness status, audio volume status, and crowd density sensor input status;

[0025] Real-time acquisition and recording of device status before instruction execution The device status after the instruction is executed , according to the device status before the instruction is executed The device status after the instruction is executed Calculate command feedback deviation ;

[0026] Preset command feedback deviation threshold , when the instruction feedback deviation Exceeding the command feedback deviation threshold When the error occurs, the fault-tolerant rollback mechanism is triggered; wherein, the fault-tolerant rollback mechanism is triggered: when the instruction feedback deviation Exceeding the command feedback deviation threshold When the resend instruction operation is taken, the resend instruction operation times When the rollback operation is initiated.

[0027] Preferably, in the linkage instruction generation module, the priority data and linkage dependency data of different control devices in the exhibition hall are obtained, and the target scene mode is matched in combination with the instruction feedback deviation. , the steps of generating and executing the linkage instruction control sequence include:

[0028] Matching target scene mode based on instruction feedback deviation , is the target scene mode Build a device target state collection , where k is the number of devices;

[0029] Obtain the priority data and linkage dependency data of different control devices in the exhibition hall, and convert the device target state set into a linkage instruction control sequence executed in stages based on the priority data and linkage dependency data ;

[0030] Control commands are executed according to the linkage instruction control sequence executed in stages. After the control commands of each stage are executed, the status consistency check is performed. If any stage fails, the subsequent steps are interrupted, and the cause of the linkage failure and the affected equipment are automatically recorded.

[0031] The present invention also provides a cloud-based control method for a smart digital exhibition hall, comprising:

[0032] Step S10: deploy an edge central control server locally in the exhibition hall, deploy a policy optimization large model server in the cloud, and use the edge-cloud task division mechanism to establish a two-way communication channel between the edge central control server and the policy optimization large model server;

[0033] Step S20: Collect multimodal input data, standardize the multimodal input data through the edge central control server to obtain standard multimodal data; send the standard multimodal data to the policy optimization large model server through a two-way communication channel, and process the standard multimodal data through the policy optimization large model server to generate output structured control instructions;

[0034] Step S30: Execute the structured control instruction, obtain the device status before and after the instruction execution in real time, calculate the instruction feedback deviation, and determine whether to trigger the fault-tolerant rollback mechanism;

[0035] Step S40: Obtain the priority data and linkage dependency data of different control devices in the exhibition hall, and match the target scene mode in combination with the instruction feedback deviation , generate linkage instruction control sequence and execute it;

[0036] Step S50: Set the target scene mode in the cloud The linkage command control sequence under is synchronized to the edge central control server through the cloud scheduling API interface.

[0037] The present invention also provides a computer program product, including a cloud-based central control program for a smart digital exhibition hall. When the cloud-based central control program for a smart digital exhibition hall is executed by a processor, the cloud-based central control method for a smart digital exhibition hall is implemented.

[0038] The beneficial effects of the present invention are: compared with the cloud exhibition hall in the prior art, which has a single exhibition hall control command method, poor linkage capability and high command false touch rate, especially under conditions of dense audiences and complex scene switching, it is difficult to achieve the technical problem of intelligent response control requirements. This application realizes multi-channel and scene-adaptive joint control of exhibition hall screens, lights, audio and guide systems by constructing a multimodal input fusion mechanism, control command generation logic based on semantic analysis and strategy models, device status feedback closed-loop control mechanism and cloud-edge collaborative strategy optimization structure, thereby improving the intelligence level of exhibition hall operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0040] Figure 1 This is a platform schematic diagram of the first embodiment of a cloud-based central control platform for a smart digital exhibition hall according to the present invention.

[0041] Figure 2 This is a schematic diagram of the equipment of a cloud-based central control platform for a smart digital exhibition hall according to the present invention. DETAILED DESCRIPTION

[0042] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0043] Example 1: Figure 1 2 is a flow chart of the first embodiment of the cloud-based central control platform for a smart digital exhibition hall according to the present invention, which provides the first embodiment of the cloud-based central control platform for a smart digital exhibition hall according to the present invention.

[0044] In the first embodiment, the cloud-based control platform for the smart digital exhibition hall includes:

[0045] The cloud-edge collaborative control module is used to deploy the edge central control server locally in the exhibition hall and the policy optimization large model server in the cloud. It also uses the edge-cloud task division mechanism to establish a two-way communication channel between the edge central control server and the policy optimization large model server.

[0046] It should be noted that in the cloud-edge collaborative control module, the edge central control server is used to receive local multimodal instructions and structured control instructions fed back by the cloud, as well as to control the exhibition hall's screen subsystem, lighting subsystem, audio subsystem and guide subsystem; the strategy optimization large model server is used to perform control strategy adaptive update tasks, AI command recognition tasks and crowd flow prediction tasks.

[0047] It is understandable that the edge central control server, as the fastest-responding control node in the system, prioritizes the real-time execution of local instructions and status feedback updates, while handing over non-real-time tasks requiring policy optimization or semantic analysis to the cloud for processing, thereby improving control accuracy while ensuring low-latency operation of the system.

[0048] It should be understood that the policy optimization large model server is not only used to update the policy logic, but also undertakes the task of semantic nesting and parsing of voice commands and perception events, ensuring that the fusion control of multimodal input has the ability to understand the context, further reducing the risk of misoperation and improving the intelligence level of control decisions.

[0049] For example, when a sudden increase in crowds is detected in an area with dense crowds, the edge central control server will promptly report the status changes to the strategy optimization large model server based on the information collected by sensors; the latter will determine that the area will enter a peak period through the crowd prediction model, and then automatically adjust the lighting brightness, guide audio playback, screen content push and other control strategies in the area, and synchronously return them to the edge central control server for execution through structured control instructions, realizing full-link high-response intelligent scene linkage control.

[0050] The unified instruction parsing module is used to collect multimodal input data, standardize the multimodal input data through the edge central control server, and obtain standard multimodal data; the standard multimodal data is sent to the policy optimization large model server through a two-way communication channel, and the policy optimization large model server processes the standard multimodal data and generates output structured control instructions;

[0051] It should be noted that in the unified instruction parsing module, the steps of standardizing the multimodal input data through the edge central control server specifically include: receiving multimodal input data through the edge central control server, the multimodal input data including voice input data, exhibition hall touch terminal input data, exhibition hall mobile app input data and crowd density sensor input data; classifying the input into four basic types according to the multimodal input data by data format, source port and context label, including voice input data , touch input data 、App input data , sensor input data , and generate a type identification code; construct a unified input representation vector , including: For voice input data Use the lightweight speech recognition models BERT and CTC deployed locally on the edge control server to convert it into text instructions, and then extract the semantic vector of the speech input data; for touch input data and App input data Use predefined controls for one-hot coding and construct touch input data and App input data Semantic vector of sensor input data Directly build sensor input data for real-valued perception semantic vector of ; comprehensive speech input data , touch input data 、App input data , sensor input data The semantic vector of the unified input representation vector is constructed ; For uniform input representation vector The value range differences of different dimensions in the dataset are first processed by linear normalization to obtain the optimized input representation vector; abnormal input detection and filtering are performed on the optimized input representation vector to obtain standardized multimodal input data.

[0052] In the unified instruction parsing module, the standard multimodal data is processed by the policy optimization large model server to generate output structured control instructions. Specifically, in the policy optimization large model server, the reinforcement learning structure DDPG is used to build the input control mapping model, and the standardized multimodal input data is mapped to the state space S of the input control mapping model. The input control mapping model policy function is expressed as , used to indicate that the Take action The probability distribution of action Used to represent a set of control instruction combinations; obtain historical multimodal input data and historical control instruction combinations, and train an input control mapping model based on the historical multimodal input data and historical control instruction combinations; Input the trained input control mapping model, perform the inference operation, output the control instruction combination, and construct the structured control instruction according to the control instruction combination. The structured control instruction includes .

[0053] It can be understood that the standardized input data not only eliminates the heterogeneity between different input sources, but also enhances the accurate matching of instructions through context encoding, providing more stable and consistent state input for subsequent policy model reasoning, and improving the robustness and generalization ability of control decisions.

[0054] It should be understood that compared to traditional fixed-rule-based control methods, the DDPG strategy optimization model can achieve "multi-condition-multi-objective-adaptive" reasoning of control actions by continuously iteratively learning the state-behavior feedback relationship of the exhibition hall, significantly improving the system's control decision-making capabilities in high-interaction scenarios.

[0055] For example, when a visitor says "Please talk about this exhibit" and the crowd sensor detects a surge in the number of people in the area, the system converts the speech into text and extracts the intention of "starting the explanation". Combined with the crowd flow status, it is judged as "peak tour mode". The strategy model outputs an action combination including: "playing the explanation video on the screen, focusing the light on the exhibit, and playing the commentary audio on the speakers", and sends it to the corresponding device through structured instructions to achieve intelligent cross-system collaboration.

[0056] The control state feedback module is used to execute structured control instructions, obtain the device status before and after the instruction execution in real time, calculate the instruction feedback deviation, and determine whether to trigger the fault-tolerant rollback mechanism;

[0057] It should be noted that in the control state feedback module, the steps of executing the structured control instruction, obtaining the device state before and after the instruction execution in real time, calculating the instruction feedback deviation, and judging whether to trigger the fault-tolerant rollback mechanism specifically include: issuing the output structured control instruction in the unified instruction parsing module to the corresponding device, and setting the state of each device at time t to the state of the device. Abstract in vector form, ,in, For the The current status of the device, including screen playback status, light brightness status, audio volume status and crowd density sensor input status; real-time acquisition and recording of the device status before the command is executed The device status after the instruction is executed , according to the device status before the instruction is executed The device status after the instruction is executed Calculate command feedback deviation ;Preset instruction feedback deviation threshold , when the instruction feedback deviation Exceeding the command feedback deviation threshold When the error occurs, the fault-tolerant rollback mechanism is triggered; wherein, the fault-tolerant rollback mechanism is triggered: when the instruction feedback deviation Exceeding the command feedback deviation threshold When the resend instruction operation is taken, the resend instruction operation times When the rollback operation is initiated.

[0058] It can be understood that the vectorized modeling of device status not only improves the uniformity of the feedback data structure, but also provides a calculation basis for the error evaluation of instruction feedback, enabling the system to achieve high-precision execution effect evaluation.

[0059] It should be understood that compared with the traditional fuzzy fault-tolerant strategy based on "whether to respond" judgment, this embodiment achieves more refined fault-tolerant trigger conditions by quantifying the difference and setting multi-dimensional deviation thresholds, while avoiding the problems of misjudgment or excessive fault tolerance, and ensuring the reliability and control consistency of the system.

[0060] For example, when the system issues an instruction to the lighting subsystem to "adjust the lighting brightness in exhibition area A to 80%", if the current state is 65%, but the actual feedback value is 68% (deviation 12%), which is greater than the set tolerance of 5%, the command resend operation is triggered; if abnormal feedback is still received after three consecutive executions, the light brightness will be restored to the original value of 65%, and the lighting controller will be marked as "abnormal state", prompting the operation and maintenance personnel to perform maintenance intervention.

[0061] The linkage command generation module is used to obtain the priority data and linkage dependency data of different control devices in the exhibition hall, and match the target scene mode based on the command feedback deviation , generate linkage instruction control sequence and execute it;

[0062] It should be noted that in the linkage command generation module, the priority data and linkage dependency data of different control devices in the exhibition hall are obtained, and the target scene mode is matched with the command feedback deviation. , generating and executing the linkage command control sequence, specifically including: matching the target scene mode based on the command feedback deviation , is the target scene mode Build a device target state collection , where k is the number of devices; obtain the priority data and linkage dependency data of different control devices in the exhibition hall, and convert the device target state set into a linkage instruction control sequence executed in stages according to the priority data and linkage dependency data ; Execute control commands according to the linkage instruction control sequence executed in stages. After the control commands of each stage are executed, the status consistency check is performed. If any stage fails, the subsequent steps are interrupted, and the cause of the linkage failure and the affected equipment are automatically recorded.

[0063] It is understandable that priority data and linkage dependency information jointly determine the execution order and dependency constraints in the control process. By integrating the topological sorting algorithm to consider the issues of "who to execute first" and "who depends on whom", it can effectively avoid the erroneous issuance of instructions when key equipment is not ready, thereby improving the logical integrity and timing correctness of the linkage.

[0064] It should be understood that through the phased execution and phased verification mechanism, the process-based and structured management of the control logic between complex equipment groups is achieved. When the control is interrupted or fails, the specific affected links can be quickly located, which facilitates system self-healing or manual intervention, greatly enhancing the stability and maintainability of the exhibition hall operation.

[0065] For example, if the system determines that the current crowded area should switch to "peak tour mode", the linkage command control sequence may be as follows: Stage 1: turn on the lights in exhibition area A (high priority, lighting must be turned on first); Stage 2: play the exhibit explanation audio and activate the corresponding audio; Stage 3: switch the screen content to the interactive tour interface; if the audio module in stage 2 fails to feedback the "playing" status due to communication interruption, the system will interrupt the execution, and at the same time record "audio equipment A01 command failure, screen S05 dependency delay" and issue an alarm prompt to prevent the out-of-order experience of playing content before the explanation.

[0066] Scene parameter synchronization module, used to synchronize the target scene mode in the cloud The linkage command control sequence under is synchronized to the edge central control server through the cloud scheduling API interface.

[0067] It should be understood that in order to ensure synchronization security and consistency, this module supports the API interface's two-way handshake verification mechanism, failure retry mechanism, and version number alignment mechanism to ensure that the parameter package is not tampered with, lost, or executed incorrectly during transmission. It also supports a timestamp-based sequence protection mechanism to ensure the edge side's "latest parameter first" execution strategy.

[0068] For example, after the cloud detects that the current state should switch to "Closed Inspection Mode," it generates the following control sequence: first, turn off all screens (L1), then dim the lights (L2), and finally activate the inspection robot (L3). This sequence is pushed to the edge control server via the scheduling interface. Upon receiving this sequence, the edge control server selects the execution time 10 minutes before closing time based on the current time, task priority, and device feedback, to avoid inadvertent actions such as turning off lights or stopping explanations before visitors have left.

[0069] Embodiment 2: In addition, the present invention provides a cloud-based control method for a smart digital exhibition hall, which uses a cloud-based control platform for a smart digital exhibition hall in the above embodiment to solve the technical problems of cloud-based control for a smart digital exhibition hall. Compared with the existing technology, the beneficial effects of the cloud-based control method for a smart digital exhibition hall provided by the present invention are the same as the beneficial effects of the cloud-based control platform for a smart digital exhibition hall provided by the above embodiment, and the other technical features of the cloud-based control method for a smart digital exhibition hall are the same as those disclosed in the above embodiment, and are not further described here.

[0070] Example 3: The present invention provides a cloud-based control device for a smart digital exhibition hall. Figure 2 A cloud control device for a smart digital exhibition hall includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute a cloud control method for a smart digital exhibition hall in the above-mentioned embodiment one. A cloud control device for a smart digital exhibition hall in an embodiment of the present invention may include but is not limited to mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Descriptions), PMPs (Portable Media Players), vehicle-mounted terminals (such as vehicle-mounted navigation terminals), etc., and fixed terminals such as digital TVs, desktop computers, etc. A cloud control device for a smart digital exhibition hall is merely an example and should not impose any limitations on the functions and scope of use of the embodiments of the present invention. A cloud-based central control device for a smart digital exhibition hall may include a processing device 1001 (e.g., a central processing unit, graphics processing unit, etc.), which can perform various appropriate actions and processes based on programs stored in a read-only memory (ROM) 1002 or programs loaded from a storage device 1003 into a random access memory (RAM) 1004. RAM 1004 also stores various programs and data required for the operation of the cloud-based central control device for a smart digital exhibition hall. Processing device 1001, ROM 1002, and RAM 1004 are interconnected via a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Typically, the following platforms can be connected to I / O interface 1006: input devices 1007 including, for example, a touchscreen, touchpad, keyboard, mouse, image sensor, microphone, accelerometer, gyroscope, etc.; output devices 1008 including, for example, a liquid crystal display (LCD), speaker, vibrator, etc.; storage devices 1003 including, for example, magnetic tape, hard disk, etc.; and communication devices 1009. Communication devices 1009 can allow a cloud-based control device for a smart digital exhibition hall to communicate wirelessly or wired with other devices to exchange data. Although the figure shows a cloud-based control device for a smart digital exhibition hall with various platforms, it should be understood that implementation or presence of all illustrated platforms is not required. More or fewer platforms may alternatively be implemented or present.

[0071] Example 4: The present invention also provides a computer program product, including a computer program. When executed by a processor, the computer program implements the steps of the cloud-based control method for a smart digital exhibition hall as described above. The computer program product provided by the present invention can solve the technical problems of cloud-based control for a smart digital exhibition hall. Compared with the prior art, the beneficial effects of the computer program product provided by the present invention are the same as those of the cloud-based control method for a smart digital exhibition hall provided in the above embodiment, and are not further described here.

[0072] In particular, according to the embodiments disclosed in the present invention, the processes described above with reference to the flowcharts can be implemented as computer platform programs. For example, the embodiments disclosed in the present invention include a computer program product comprising a computer program carried on a computer-readable medium, the computer program comprising program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network via a communication device, or installed from a storage device 1003, or installed from a ROM 1002. When the computer program is executed by the processing device 1001, the above-mentioned functions defined in the method of the embodiment disclosed in the present invention are performed.

[0073] It should be understood that the various parts disclosed in the present invention can be implemented using hardware, platform, firmware or a combination thereof. In the description of the above embodiments, specific features, structures, materials or characteristics can be combined in any one or more embodiments or examples in a suitable manner.

[0074] Obviously, those skilled in the art may make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if such changes and modifications fall within the scope of the claims and their equivalents, the present invention is intended to include such changes and modifications.

Claims

1. A cloud-based control platform for a smart digital exhibition hall, characterized in that the platform include: The cloud-edge collaborative control module is used to deploy the edge central control server locally in the exhibition hall and the policy optimization large model server in the cloud. It also uses the edge-cloud task division mechanism to establish a two-way communication channel between the edge central control server and the policy optimization large model server. The unified instruction parsing module is used to collect multimodal input data, standardize the multimodal input data through the edge central control server, and obtain standard multimodal data; the standard multimodal data is sent to the policy optimization large model server through a two-way communication channel, and the policy optimization large model server processes the standard multimodal data and generates output structured control instructions; The control state feedback module is used to execute structured control instructions, obtain the device status before and after the instruction execution in real time, calculate the instruction feedback deviation, and determine whether to trigger the fault-tolerant rollback mechanism; The linkage command generation module is used to obtain the priority data and linkage dependency data of different control devices in the exhibition hall, and match the target scene mode based on the command feedback deviation , generate linkage instruction control sequence and execute it; Scene parameter synchronization module, used to synchronize the target scene mode in the cloud The linkage command control sequence under is synchronized to the edge central control server through the cloud scheduling API interface.

2. The cloud-based control platform for a smart digital exhibition hall according to claim 1, characterized in that: In the cloud-edge collaborative control module, the edge central control server is used to receive local multimodal commands and structured control commands fed back by the cloud, as well as to control the exhibition hall's screen subsystem, lighting subsystem, audio subsystem and guide subsystem; the strategy optimization large model server is used to perform control strategy adaptive update tasks, AI command recognition tasks and crowd flow prediction tasks.

3. The cloud-based control platform for a smart digital exhibition hall according to claim 1, characterized in that: In the unified command parsing module, multimodal input data includes voice input data, exhibition hall touch terminal input data, exhibition hall mobile app input data and crowd density sensor input data; the output structured control instructions include screen subsystem control instructions, lighting subsystem control instructions, audio subsystem control instructions and guide subsystem control instructions.

4. The cloud-based control platform for a smart digital exhibition hall according to claim 1, characterized in that: In the unified command parsing module, the steps for standardizing multimodal input data through the edge central control server include: Receive multimodal input data through the edge control server, including voice input data, exhibition hall touch terminal input data, exhibition hall mobile app input data, and crowd density sensor input data; Based on the multimodal input data, the input is classified into four basic types by data format, source port and context label, including voice input data , touch input data 、App input data , sensor input data , and generate a type identification code; Constructing a unified input representation vector , including: For voice input data Use the lightweight speech recognition models BERT and CTC deployed locally on the edge control server to convert it into text instructions, and then extract the semantic vector of the speech input data; for touch input data and App input data Use predefined controls for one-hot coding and construct touch input data and App input data Semantic vector of sensor input data Directly build sensor input data for real-valued perception semantic vector of ; comprehensive speech input data , touch input data 、App input data , sensor input data The semantic vector of the unified input representation vector is constructed ; Representing vectors for uniform input The value range differences of different dimensions in are first processed by linear normalization to obtain the optimized input representation vector; Abnormal input detection and filtering are performed on the optimized input representation vector to obtain standardized multimodal input data.

5. The cloud-based control platform for a smart digital exhibition hall according to claim 1, characterized in that: In the unified instruction parsing module, the steps of processing standard multimodal data through the strategy optimization large model server and generating output structured control instructions include: In the policy optimization large model server, the reinforcement learning structure DDPG is used to build the input control mapping model, and the standardized multimodal input data is mapped to the state space S of the input control mapping model. The input control mapping model policy function is expressed as , used to indicate that the Take action The probability distribution of action Used to represent a group of control instruction combinations; Acquire historical multimodal input data and historical control instruction combinations, and train an input control mapping model based on the historical multimodal input data and historical control instruction combinations; Will Input the trained input control mapping model, perform the inference operation, output the control instruction combination, and construct the structured control instruction according to the control instruction combination. The structured control instruction includes 。 6. The cloud-based control platform for a smart digital exhibition hall according to claim 1, characterized in that: In the control state feedback module, the steps of executing structured control instructions, obtaining the device status before and after the instruction execution in real time, calculating the instruction feedback deviation, and determining whether to trigger the fault-tolerant rollback mechanism include: The output structured control instructions in the unified instruction parsing module are sent to the corresponding devices, and the state of each device at time t is Abstract in vector form, ,in, For the The current status of the device, including screen playback status, light brightness status, audio volume status, and crowd density sensor input status; Real-time acquisition and recording of device status before instruction execution The device status after the instruction is executed , according to the device status before the instruction is executed The device status after the instruction is executed Calculate command feedback deviation ; Preset command feedback deviation threshold , when the instruction feedback deviation Exceeding the command feedback deviation threshold When the error occurs, the fault-tolerant rollback mechanism is triggered; wherein, the fault-tolerant rollback mechanism is triggered: when the instruction feedback deviation Exceeding the command feedback deviation threshold When the resend instruction operation is taken, the resend instruction operation times When the rollback operation is initiated.

7. The cloud-based control platform for a smart digital exhibition hall according to claim 1, characterized in that: In the linkage command generation module, the priority data and linkage dependency data of different control devices in the exhibition hall are obtained, and the target scene mode is matched with the command feedback deviation. , the steps of generating and executing the linkage instruction control sequence include: Matching target scene mode based on instruction feedback deviation , is the target scene mode Build a device target state collection , where k is the number of devices; Obtain the priority data and linkage dependency data of different control devices in the exhibition hall, and convert the device target state set into a linkage instruction control sequence executed in stages based on the priority data and linkage dependency data ; Control commands are executed according to the linkage instruction control sequence executed in stages. After the control commands of each stage are executed, the status consistency check is performed. If any stage fails, the subsequent steps are interrupted, and the cause of the linkage failure and the affected equipment are automatically recorded.

8. A cloud-based control method for a smart digital exhibition hall, applied to a cloud-based control platform for a smart digital exhibition hall according to any one of claims 1 to 7, characterized in that: Methods include: Step S10: deploy an edge central control server locally in the exhibition hall, deploy a policy optimization large model server in the cloud, and use the edge-cloud task division mechanism to establish a two-way communication channel between the edge central control server and the policy optimization large model server; Step S20: Collect multimodal input data, standardize the multimodal input data through the edge central control server to obtain standard multimodal data; send the standard multimodal data to the policy optimization large model server through a two-way communication channel, and process the standard multimodal data through the policy optimization large model server to generate output structured control instructions; Step S30: Execute the structured control instruction, obtain the device status before and after the instruction execution in real time, calculate the instruction feedback deviation, and determine whether to trigger the fault-tolerant rollback mechanism; Step S40: Obtain the priority data and linkage dependency data of different control devices in the exhibition hall, and match the target scene mode in combination with the instruction feedback deviation , generate linkage instruction control sequence and execute it; Step S50: Set the target scene mode in the cloud The linkage command control sequence under is synchronized to the edge central control server through the cloud scheduling API interface.

9. A cloud-based control device for a smart digital exhibition hall, characterized in that: The cloud control device for the smart digital exhibition hall includes: a memory, a processor, and a cloud control program for the smart digital exhibition hall stored in the memory and runnable on the processor. When the cloud control program for the smart digital exhibition hall is executed by the processor, a cloud control platform for the smart digital exhibition hall according to any one of claims 1 to 7 is implemented.

10. A computer program product, characterized in that The computer program product includes a cloud central control program for a smart digital exhibition hall, and when the cloud central control program for a smart digital exhibition hall is executed by a processor, a cloud central control platform for a smart digital exhibition hall according to any one of claims 1 to 7 is implemented.

Images