An internet-of-things-based intelligent park regional management platform

Abstract

The application relates to the technical field of industrial automation control, and discloses a wisdom park regional management platform based on Internet of Things. The application comprises a regional feature construction module, which is used for fusing multi-source Internet of Things data to generate regional situation features; a dynamic influence calculation module between regions, which adopts a correlation model based on multi-dimensional feature similarity and space-time attenuation to calculate and construct a dynamic regional influence network in real time; an instruction optimization module, which is used for performing conflict detection and global collaborative optimization on multi-strategy instructions based on the influence network to generate a conflict-free optimal instruction set; and a control instruction execution module, which is used for realizing reliable instruction issuing and closed-loop feedback; the application realizes the leap from static regional monitoring to dynamic collaborative defense and the transformation from multi-instruction conflict to global collaborative control, and significantly improves the intelligent level and operation efficiency of park safety management and control.

Description

Technical Field

[0001] This invention relates to the field of industrial automation control technology, and in particular to a regionalized management platform for smart parks based on the Internet of Things. Background Technology

[0002] In recent years, the maturity of technologies such as the Internet of Things (IoT) and big data has driven the rapid development of smart industrial parks. In high-risk or large-scale integrated industrial parks such as those in the chemical and manufacturing industries, a large number of heterogeneous IoT systems have been deployed, including video surveillance, environmental sensors (such as gas and flame detectors), access control, fire alarm systems, and energy metering. These systems are typically designed to achieve independent monitoring and automated response to specific areas or types of risks within the park.

[0003] However, existing IoT-based park management platforms suffer from two fundamental technical flaws: First, existing platforms typically manage parks as static, isolated logical areas, with their monitoring and response logic often confined to a single area. Real security risks exhibit strong spatiotemporal propagation characteristics; existing technologies lack effective models to calculate, in real-time and quantitatively, the dynamic impact of a risk state in one area on another. Second, conflicts and lack of coordination among multiple system commands lead to control friction and inefficient execution. Different management subsystems within the park, such as security, energy efficiency, and access control, often issue automated control commands in parallel, easily resulting in conflicts. Existing platforms either employ simple static prioritization or rely on manual arbitration, often leading to internal friction within the control system itself. Summary of the Invention

[0004] The technical problem to be solved by this invention is that existing smart park management platforms have simple response logic and lack collaborative management. To address this, we propose a regionalized smart park management platform based on the Internet of Things.

[0005] To achieve the above objectives, this application adopts the following technical solution: a regionalized management platform for smart parks based on the Internet of Things, comprising:

[0006] The regional feature construction module is used to acquire environmental, equipment, personnel and vehicle data collected by various IoT sensing devices deployed in the park, clean, integrate and standardize the environmental, equipment and personnel and vehicle data, and generate regional state feature vectors that represent the comprehensive situation of each region based on predefined or dynamically generated logical management regions.

[0007] The inter-regional dynamic impact calculation module is connected to the regional feature construction module. It is used to calculate and quantify the dynamic impact intensity between any two logical management regions in real time based on the regional state feature vector, so as to construct and update the dynamic impact network that reflects the cross-regional propagation relationship of risks and situations within the park.

[0008] The instruction optimization module is connected to the inter-regional dynamic influence calculation module. It is used to receive control instruction requests for IoT execution devices from different management strategies, and based on the dynamic influence network, it performs conflict detection and global collaborative optimization on instruction requests that have spatial correlation or device resource competition relationship, and generates a set of conflict-free collaborative control instructions that meet multi-objective constraints.

[0009] The control command execution module, connected to the command optimization module, is used to convert the collaborative control command set into control commands adapted to the specific device communication protocol and reliably issue them. At the same time, it monitors the deviation between the command execution status and the expected effect, forming a closed-loop control.

[0010] Preferably, the region feature construction module is used for:

[0011] Taking the logical management area as a unit, all IoT sensing data within it are aggregated, and feature indicators under preset feature dimensions are extracted and calculated. The preset feature dimensions include environmental risk level, personnel density, key equipment operating load and management activity intensity.

[0012] The feature indicators of each dimension are combined to form a regional state feature vector, and then normalized so that the dynamic influence calculation module between regions can perform vectorized analysis and calculation.

[0013] Preferably, the inter-regional dynamic impact calculation module is used to: calculate the similarity between the regional state feature vectors of any two regions, as the basis for assessing their situational correlation;

[0014] Obtain the physical distance or logical topological distance between two regions, and calculate the spatial attenuation factor affecting the intensity based on the preset spatial attenuation coefficient;

[0015] Identify whether a significant event has occurred in the source region, and when it occurs, calculate the decay factor of the influence intensity over time based on the event intensity, the time of the event, and a preset time decay constant;

[0016] By combining similarity, spatial decay factor and temporal decay factor, the real-time unidirectional influence intensity between the two regions is dynamically synthesized.

[0017] Preferably, the instruction optimization module is used for:

[0018] Based on the dynamic influence network, the correlation strength between the regions involved in the current instruction request is identified, and instruction requests with correlation strength exceeding a preset threshold are judged to have potential cooperative or conflicting relationships.

[0019] For instruction requests with potential cooperative or conflicting relationships, as well as mutually exclusive instructions issued to the same execution device, a unified model is used to form an optimization problem with the goal of maximizing global cooperative utility.

[0020] By solving the optimization problem, the optimal instruction execution scheme under specified constraints is output, which is the cooperative control instruction set.

[0021] Preferably, when solving the optimization problem, the instruction optimization module introduces the historical switching frequency of the device as a constraint condition to penalize unnecessary or frequent state switching of key devices with stable historical states, thereby ensuring the smooth execution of control instructions.

[0022] Preferably, the control instruction execution module is used for:

[0023] The collaborative control instruction set is time-scheduled and protocol-encapsulated to generate an instruction sequence that can directly drive IoT execution devices.

[0024] Establish two-way communication with the execution device. After sending control commands, verify whether the device has reached the target state expected by the command through status readback or active monitoring, and feed back the verification results to the command optimization module.

[0025] Preferably, the control instruction execution module is also used for:

[0026] The verification results of instruction execution, especially cases of execution failure or deviation, are recorded as feedback data;

[0027] Feedback data is periodically sent to the instruction optimization module for adaptive adjustment and learning of its internal optimization model parameters or strategy weights, in order to continuously improve the accuracy of collaborative decision-making.

[0028] Preferably, the inter-regional dynamic impact calculation module and the instruction optimization module work together dynamically: when the inter-regional dynamic impact calculation module detects that the impact intensity of a certain region exceeds the alarm threshold due to a sudden event, it immediately triggers the instruction optimization module to enter the emergency optimization mode. In this mode, the instruction optimization module will prioritize the execution of safety-related instructions in other regions that are strongly associated with the region of the sudden event, and suspend non-critical management instructions that may cause conflicts.

[0029] Preferably, in the inter-regional dynamic influence calculation module, the dynamic influence intensity of region i on region j at time t is calculated. The formula used is:

[0030] ;

[0031] in, The intensity of the event source in region i at time t is quantified based on the alarm level or activity status of the sensors in that region. and These are the region state feature vectors of region i and region j at time t, respectively. This is a function for calculating the similarity between two feature vectors. The physical or logical distance between region i and region j; To control the spatial attenuation coefficient that affects the rate of attenuation with distance; This is the initial enhancement coefficient used to adjust the magnitude of the initial impact of the event. Let i be the start time of the most recent significant event in region i; To control the time decay constant of the event's impact decay rate over time, Set parameters based on the typical duration of the risk events of interest; , and The calibration is performed by analyzing the impact propagation data in historical records of emergencies and using optimization algorithms such as the least squares method.

[0032] Preferably, the instruction optimization module is used to evaluate a candidate cooperative control instruction set when solving the optimization problem. Global collaborative utility function The formula is:

[0033] ;

[0034] in, This represents the total number of strategies currently participating in the optimization. For the first The intent vector of each strategy; This is the mean vector of all strategy intent vectors; Let be the Euclidean norm of the vector, which is used to measure the overall dispersion of all strategic intentions; For candidate instruction sets Total number of devices affected; For indicating functions, when the device exist The value is 1 if there is a logical conflict in the instruction; otherwise, it is 0. For equipment Frequency of state switching during the historical observation period; This is the conflict penalty intensity coefficient used to control the severity of command conflict penalties; This refers to the frequency adjustment coefficient used to adjust the weight of the penalty term on the effect of equipment switching frequency; the conflict penalty intensity coefficient. and frequency adjustment coefficient When deploying the platform, the administrator sets it in the configuration interface according to the park management strategy, and then dynamically adjusts it through iterative learning algorithms based on feedback data of the collaborative optimization results based on historical instructions.

[0035] The technical effects and advantages of this invention are as follows:

[0036] In this invention, by dynamically quantifying the intensity of influence between regions, the platform has achieved a leap from static isolated monitoring to dynamic proactive joint defense. It can accurately predict the risk propagation path and initiate minute-level preventive control of related areas, minimizing the blind spots and delays in the traditional passive response mode, and greatly improving the park's ability to suppress and coordinate sudden risks in the early stage.

[0037] By performing global collaborative optimization of multi-source instructions, the platform has achieved a transformation from multi-system instruction conflicts to global strategy collaboration. It can automatically resolve equipment control conflicts, balance multiple objectives such as safety and energy efficiency, and output a stable and deterministic optimal execution plan, completely avoiding system internal friction, equipment wear and tear, and safety blind spots. Attached Figure Description

[0038] The disclosure of this invention is illustrated with reference to the accompanying drawings. It should be understood that the drawings are for illustrative purposes only and are not intended to limit the scope of protection of this invention. In the drawings, the same reference numerals are used to refer to the same parts:

[0039] Figure 1 This is a schematic diagram of the module topology of the present invention. Detailed Implementation

[0040] It is readily understood that, based on the technical solution of this invention, those skilled in the art can propose various interchangeable structural methods and implementations without altering the essential spirit of the invention. Therefore, the following detailed embodiments and accompanying drawings are merely illustrative examples of the technical solution of this invention and should not be considered as the entirety of the invention or as limitations or restrictions on the technical solution of this invention.

[0041] Example 1

[0042] Reference Figure 1 As shown, the present invention provides a technical solution: a regional management platform for smart parks based on the Internet of Things, which mainly includes the following core modules: a regional feature construction module, a regional dynamic influence calculation module, an instruction optimization module, and a control instruction execution module.

[0043] The regional feature construction module is the platform's data entry point and basic perception layer. It is responsible for acquiring environmental, equipment, personnel, and vehicle data collected by various IoT sensing devices deployed in the park, including environmental sensors, video surveillance, access control systems, vehicle recognition systems, and equipment operation status sensors. This data includes temperature, humidity, PM2.5, CO2 concentration, personnel entry and exit records, vehicle trajectories, equipment energy consumption, and equipment fault alarms.

[0044] Specifically, the regional feature construction module first cleans, fuses, and standardizes these raw data. Data cleaning includes removing noise, filling in missing values, and correcting erroneous data. Data fusion integrates data from different sensors and protocols into a unified data model. Standardization normalizes or makes dimensionless data of different dimensions for subsequent unified analysis and calculation.

[0045] After the data processing is completed, the module aggregates all IoT sensing data within a predefined or dynamically generated logical management area, such as an area divided according to buildings, functional zones, or geographical locations. For example, a logical management area corresponds to an office building, a parking lot, or a production workshop.

[0046] Next, the module extracts and calculates feature metrics under preset feature dimensions, which include, but are not limited to:

[0047] Environmental risk levels, such as air quality index and fire risk level;

[0048] Population density, such as the real-time number of people in the area and the rate of population movement;

[0049] Key equipment operating loads, such as air conditioning energy consumption, server CPU utilization, and production line utilization rate;

[0050] And the intensity of management activities, such as the frequency of security patrols and the progress of cleaning operations.

[0051] These characteristic indicators are calculated from aggregated data through statistical analysis or machine learning models.

[0052] Finally, this module combines the feature indicators of each dimension to form a regional state feature vector and performs normalization processing. For example, the state feature vector of a region may be a multi-dimensional vector, where each component represents a feature dimension, such as [environmental risk index, personnel density index, equipment load index, management activity intensity index]; the normalization processing ensures that the weights of different feature dimensions are balanced in subsequent calculations, so that the dynamic impact calculation module between regions can perform vectorized analysis and calculation.

[0053] The inter-regional dynamic impact calculation module is connected to the regional feature construction module. It receives the regional state feature vectors of each region and is responsible for calculating and quantifying the dynamic impact intensity between any two logical management regions in real time, so as to construct and update a dynamic impact network that reflects the cross-regional propagation relationship of risks and situations within the park.

[0054] Specifically, the workflow of this module is as follows:

[0055] Similarity calculation: Calculate the similarity between the regional state feature vectors of any two regions, using cosine similarity, the reciprocal of Euclidean distance, etc., as the basis for evaluating their situational correlation; the higher the similarity, the more likely the two regions are to influence each other in the current state.

[0056] Obtain the spatial attenuation factor: Obtain the physical distance between two regions, such as straight-line distance, road distance, or logical topological distance, such as network hop count or functional correlation, and base it on a preset spatial attenuation coefficient. Calculate the spatial attenuation factor that affects the intensity; generally, the greater the distance, the smaller the influence and the smaller the attenuation factor.

[0057] Event identification and time decay factor calculation: Identify whether a significant event has occurred in the source area, such as a fire alarm, equipment failure, or abnormal gathering of people, and calculate the time decay factor based on the event intensity, the time of occurrence, and a preset time decay constant. The decay factor of the impact intensity over time is calculated; the impact is greatest in the early stage of the event, and then gradually decays over time.

[0058] Comprehensive calculation of unidirectional influence intensity: The real-time unidirectional influence intensity between two regions is dynamically synthesized by combining similarity, spatial decay factor, and temporal decay factor; the dynamic influence intensity of region i on region j at time t. The following formula is used for calculation:

[0059] ;

[0060] in, The intensity of the event source in region i at time t is quantified based on the alarm level or activity status of the sensors in that region.

[0061] and These are the region state feature vectors of region i and region j at time t, respectively.

[0062] This is a function for calculating the similarity between two feature vectors.

[0063] The physical or logical distance between region i and region j;

[0064] To control the spatial attenuation coefficient that affects the rate of attenuation with distance, the spatial attenuation coefficient is... This is used to characterize the rate at which risk intensity decreases with increasing physical distance. The larger the value, the faster the decay. For general risk diffusion, the preset value is 1.5; for high-risk chemical leaks, it is obtained by fitting historical leak data.

[0065] This is the initial enhancement coefficient used to adjust the magnitude of the initial impact of the event.

[0066] Let i be the start time of the most recent significant event in region i;

[0067] To control the time decay constant of the event's impact decay rate over time, Set according to the typical duration of the risk events of concern;

[0068] parameter , and By analyzing impact propagation data from historical records of emergencies, and using optimization algorithms such as the least squares method for calibration, the model can accurately reflect the actual patterns of impact propagation.

[0069] By continuously calculating and updating the intensity of these influences, this module constructs a dynamic influence network, where nodes represent logical management regions and edges represent the intensity and direction of dynamic influences between regions.

[0070] The instruction optimization module is connected to the inter-regional dynamic impact calculation module. It receives control instruction requests from different management strategies, such as security strategies, energy-saving strategies, and environmental control strategies, for IoT execution devices, such as lighting equipment, air conditioning, access control, fire sprinklers, and ventilation systems.

[0071] The core function of this module is to perform conflict detection and global collaborative optimization on instruction requests that have spatial correlation or device resource competition based on dynamic influence networks, and generate a set of conflict-free collaborative control instructions that meet multi-objective constraints.

[0072] Specifically, the workflow of the instruction optimization module is as follows:

[0073] Identify potential collaborative or conflicting relationships: Based on the dynamic influence network, identify the correlation strength between the areas involved in the current instruction request; if the correlation strength exceeds a preset threshold, these instruction requests are determined to have potential collaborative or conflicting relationships; for example, when a fire occurs in one area and its influence on adjacent areas is very strong, fire-fighting instructions and evacuation instructions for these two areas need to be coordinated.

[0074] Unified modeling optimization problem: For instruction requests with potential cooperative or conflicting relationships, as well as mutually exclusive instructions issued to the same execution device, such as one instruction requiring the air conditioner to be turned on and another instruction requiring the air conditioner to be turned off, unified modeling is performed; the optimization problem aims to maximize the global cooperative utility; the global cooperative utility can comprehensively consider multiple dimensions such as safety, energy consumption, comfort, and efficiency.

[0075] Introducing historical switching frequency constraints for devices: When solving the optimization problem, the historical switching frequency of devices is introduced as a constraint. This constraint aims to penalize unnecessary or frequent state switching of critical devices with stable historical states, ensuring the smooth execution of control commands. For example, if a lighting device is frequently switched on and off in a short period of time, it will not only shorten its lifespan but may also affect the user experience. Through this constraint, the optimization algorithm will tend to choose the solution that has the least impact on device state switching.

[0076] Solving optimization problems: By solving optimization problems, using linear programming, integer programming, multi-objective optimization algorithms, reinforcement learning, etc., the optimal instruction execution scheme under specified constraints is output, i.e., the cooperative control instruction set.

[0077] For example, evaluating a candidate set of cooperative control instructions. Global collaborative utility function The following formula can be used:

[0078] ;

[0079] in, This represents the total number of strategies currently participating in the optimization.

[0080] For the first The intent vector of each strategy;

[0081] This is the mean vector of all strategy intent vectors;

[0082] Let be the Euclidean norm of the vector. This term measures the overall dispersion of all policy intentions; the smaller the value, the more cooperative the policies are.

[0083] For candidate instruction sets Total number of devices affected;

[0084] For indicating functions, when the device exist The value is 1 if there is a logical conflict in the instruction; otherwise, it is 0.

[0085] For equipment Frequency of state switching during the historical observation period;

[0086] This is the conflict penalty intensity coefficient used to control the severity of command conflict penalties;

[0087] This is a frequency adjustment coefficient used to adjust the weight of the penalty term on the effect of equipment switching frequency;

[0088] Conflict penalty intensity coefficient and frequency adjustment coefficient When deploying the platform, the administrator sets it in the configuration interface according to the park management strategy, and then dynamically adjusts it through iterative learning algorithms based on feedback data of the collaborative optimization results based on historical instructions.

[0089] The control command execution module is connected to the command optimization module. It receives the collaborative control command set, converts it into control commands adapted to the specific device communication protocol, and reliably issues them.

[0090] Specifically, the workflow of this module is as follows:

[0091] Timing and Protocol Encapsulation: The coordinated control instruction set is time-scheduled to determine the execution order and timing of each instruction, ensuring the effectiveness and coordination of the instructions. Subsequently, the scheduled instructions are protocol-encapsulated, converting them into instruction sequences that can directly drive IoT devices, such as those using protocols like Modbus, BACnet, MQTT, and Zigbee.

[0092] Two-way communication and status verification: Establish a two-way communication channel with the execution device; after sending a control command, verify whether the device has reached the target state expected by the command through status readback or active monitoring, such as through sensor feedback and device status reports; for example, if the command is to turn on the lighting, verify whether the lighting device has been turned on and the brightness meets the expectation.

[0093] Feedback mechanism: The verification results, including whether the instruction was successfully executed and whether the execution effect deviates from the expectation, are fed back to the instruction optimization module. This feedback mechanism is the key to achieving closed-loop control.

[0094] Example 2

[0095] This invention further provides a dynamic collaborative mechanism and adaptive learning capability in an emergency mode.

[0096] In emergency mode, dynamic collaboration is enabled when the inter-regional dynamic impact calculation module detects that the impact intensity of a certain region exceeds the preset alarm threshold due to a sudden event, such as a fire, a dangerous gas leak, or a large-scale gathering of people. This will immediately trigger the instruction optimization module to enter emergency optimization mode.

[0097] In this emergency optimization mode, the instruction optimization module will adjust its optimization objectives and constraints, prioritizing the execution of safety-related instructions in other areas strongly associated with the emergency area, such as activating the fire protection system, opening emergency evacuation routes, and maximizing the operation of the ventilation system. It will also postpone non-critical management instructions that may conflict, such as shutting down air conditioning in energy-saving mode and adjusting lighting in non-emergency areas. This ensures that in emergency situations, the park's safety response can be guaranteed with the highest priority, minimizing risks and losses to the greatest extent possible.

[0098] The instruction execution feedback and adaptive learning control module not only verifies the instruction execution results, but also records the verification results, especially cases of execution failure or deviation, as feedback data. This feedback data includes the difference between the actual execution effect and the expectation, as well as the possible reasons for failure or deviation.

[0099] Subsequently, the control command execution module periodically sends feedback data to the command optimization module. The command optimization module uses this feedback data to adaptively adjust and learn the parameters or strategy weights of its internal optimization model. For example, if a device frequently fails to execute commands in a specific scenario, the optimization module may adjust the device's priority, command issuance strategy, or reassess its role in collaborative optimization. Through machine learning algorithms, the module continuously optimizes its decision-making model based on historical execution experience, thereby continuously improving the accuracy, robustness, and efficiency of collaborative decision-making and achieving true intelligent management.

[0100] Example 3

[0101] This embodiment uses a fire occurring in a certain area of ​​the park as an example to illustrate the workflow of the platform of this invention:

[0102] Regional feature construction: Smoke sensors, temperature sensors, video surveillance and other equipment in the fire area collect abnormal data; the regional feature construction module cleans and integrates this data and calculates that the environmental risk level of the area has increased sharply, forming a new regional state feature vector.

[0103] Inter-regional dynamic impact calculation: The inter-regional dynamic impact calculation module receives the updated fire area state feature vector and identifies the significant event: fire; it calculates the dynamic impact intensity of the fire area on surrounding adjacent areas, such as evacuation routes, adjacent office buildings, fire control rooms, etc.; for example, the impact intensity on evacuation routes will increase with the increase of personnel density, and the impact intensity on fire control rooms will increase with the need for emergency response; at the same time, it constructs and updates the dynamic impact network to show the propagation path and intensity of fire risk.

[0104] Command optimization: At this time, the security system may issue commands such as "turn on all fire sprinklers, start smoke exhaust fans, turn on emergency lighting, and unlock all access control systems." The energy-saving system may issue commands such as "turn off air conditioning in some areas and reduce lighting brightness" in non-emergency situations.

[0105] Upon receiving these instruction requests, the instruction optimization module, based on the dynamic influence network, identifies a high degree of correlation and potential conflict between instruction requests within the fire area and its strongly correlated areas.

[0106] Entering emergency optimization mode prioritizes safety-related commands. If a conflict is detected between "turn on fire sprinklers" and "turn off air conditioning" (e.g., sprinkler water may affect air conditioning equipment), or between "unlock access control" and "reduce lighting brightness" (e.g., sufficient lighting is needed during evacuation), then the system will be activated.

[0107] By solving the optimization problem, a set of collaborative control instructions is generated: safety instructions such as "turn on fire sprinklers", "start smoke exhaust fans", "turn on emergency lighting", and "unlock all access control" are executed first; non-critical energy-saving instructions, such as "turn off air conditioning in some areas", may be postponed or adjusted; at the same time, considering the historical switching frequency of the equipment, unnecessary frequent operation of critical fire protection equipment is avoided.

[0108] Control command execution: The control command execution module performs timing scheduling and protocol encapsulation of the collaborative control command set, and reliably sends it to fire sprinkler systems, smoke exhaust fans, emergency lighting equipment, access control systems, etc.

[0109] After the command is issued, the module will read back through sensors such as spray pressure, fan speed, lighting brightness, and access control status to verify whether the command has been successfully executed and monitor the execution effect.

[0110] If a sprinkler head is found to be not activating properly, or a door access control is not unlocked, this information will be recorded as feedback data.

[0111] Feedback and Adaptation: The control command execution module sends the above feedback data to the command optimization module. The command optimization module uses this data to analyze the reasons for command execution failure and adjusts its optimization model parameters or strategy weights accordingly. For example, if the access control system in a certain area frequently fails to unlock in an emergency, the optimization module may adjust the priority of the access control system or suggest that it be repaired, and consider its reliability factors in subsequent command optimization, thereby continuously improving the platform's decision-making accuracy and emergency response capabilities.

[0112] The technical scope of this invention is not limited to the content described above. Those skilled in the art can make various modifications and variations to the above embodiments without departing from the technical concept of this invention, and all such modifications and variations should fall within the protection scope of this invention.

Claims

1. A regionalized management platform for smart parks based on the Internet of Things, characterized in that, include: The regional feature construction module is used to acquire environmental, equipment, and personnel and vehicle data collected by various IoT sensing devices deployed in the park, clean, fuse and standardize the environmental, equipment and personnel and vehicle data, and generate regional state feature vectors representing the comprehensive situation of each region based on predefined or dynamically generated logical management regions. The inter-regional dynamic impact calculation module, connected to the region feature construction module, is used to calculate the similarity between the region state feature vectors of any two logical management regions based on the region state feature vector, obtain the physical distance or logical topological distance between the two logical management regions, calculate the spatial decay factor of the impact intensity based on the preset spatial decay coefficient, identify whether a significant event has occurred in the logical management region as the source region, and calculate the time decay factor of the impact intensity as time evolves based on the event intensity, the event occurrence time and the preset time decay constant when a significant event occurs. The module combines the similarity, spatial decay factor and time decay factor to calculate the dynamic impact intensity of region i on region j at time t. ； in, The intensity of the event source in region i at time t is quantified based on the alarm level or activity status of the sensors in that region. and These are the region state feature vectors of region i and region j at time t, respectively. This is a function for calculating the similarity between two feature vectors. The physical or logical distance between region i and region j; To control the spatial attenuation coefficient that affects the rate of attenuation with distance; This is the initial enhancement coefficient used to adjust the magnitude of the initial impact of the event. Let i be the start time of the most recent significant event in region i; The time decay constant is used to control the rate at which the impact of an event decays over time; Based on the dynamic impact intensity, a dynamic impact network reflecting the cross-regional propagation relationship of risks and situations within the park is constructed and updated; The instruction optimization module, connected to the inter-regional dynamic influence calculation module, is used to receive control instruction requests for IoT execution devices from different management strategies, and based on the dynamic influence network, to perform conflict detection and global collaborative optimization on instruction requests that have spatial correlation or device resource competition relationships, and generate a set of conflict-free collaborative control instructions that meet multi-objective constraints. The control command execution module, connected to the command optimization module, is used to convert the collaborative control command set into control commands adapted to the specific device communication protocol and reliably issue them, while monitoring the deviation between the command execution status and the expected effect to form a closed-loop control.

2. The IoT-based smart park regional management platform according to claim 1, characterized in that: The region feature construction module is used for: Taking a logical management area as a unit, all IoT sensing data within it are aggregated, and feature indicators under preset feature dimensions are extracted and calculated. The preset feature dimensions include environmental risk level, personnel density, key equipment operating load, and management activity intensity. The feature indicators of each dimension are combined to form the regional state feature vector, and then normalized so that the dynamic influence calculation module between regions can perform vectorized analysis and calculation.

3. The IoT-based smart park regional management platform according to claim 1, characterized in that: The instruction optimization module is used for: Based on the dynamic influence network, the correlation strength between the regions involved in the current instruction request is identified, and instruction requests with correlation strength exceeding a preset threshold are judged to have potential cooperative or conflicting relationships. The instruction requests with potential cooperative or conflicting relationships, as well as the mutually exclusive instructions issued to the same execution device, are uniformly modeled to form an optimization problem with the goal of maximizing global cooperative utility. By solving the optimization problem, the optimal instruction execution scheme under specified constraints is output, which is the cooperative control instruction set.

4. The IoT-based smart park regional management platform according to claim 3, characterized in that: When solving the optimization problem, the instruction optimization module introduces the historical switching frequency of the device as a constraint to penalize unnecessary or frequent state switching of key devices with stable historical states, thereby ensuring the smooth execution of control instructions.

5. The IoT-based smart park regional management platform according to claim 1, characterized in that: The control command execution module is used for: The collaborative control instruction set is time-scheduled and protocol-encapsulated to generate an instruction sequence that can directly drive IoT execution devices. Establish two-way communication with the execution device. After sending control commands, verify whether the device has reached the target state expected by the command through status readback or active monitoring, and feed back the verification results to the command optimization module.

6. The IoT-based smart park regional management platform according to claim 5, characterized in that: The control command execution module is also used for: The verification results of instruction execution, especially cases of execution failure or deviation, are recorded as feedback data; The feedback data is periodically sent to the instruction optimization module for adaptive adjustment and learning of its internal optimization model parameters or strategy weights, so as to continuously improve the accuracy of collaborative decision-making.

7. The IoT-based smart park regional management platform according to claim 1, characterized in that: The inter-regional dynamic impact calculation module and the instruction optimization module work together dynamically: when the inter-regional dynamic impact calculation module detects that the impact intensity of a certain region exceeds the alarm threshold due to a sudden event, it immediately triggers the instruction optimization module to enter the emergency optimization mode. In this mode, the instruction optimization module will prioritize the execution of security-related instructions in other regions that are strongly associated with the region of the sudden event, and suspend non-critical management instructions that may cause conflicts.

8. The IoT-based smart park regional management platform according to claim 1, characterized in that: The The parameter is set according to the typical duration of the risk event of concern; , and The calibration is performed by analyzing the impact propagation data in historical records of emergencies and using optimization algorithms such as the least squares method.

9. A regionalized management platform for smart parks based on the Internet of Things as described in claim 4, characterized in that: The instruction optimization module is used to evaluate a candidate cooperative control instruction set when solving the optimization problem. Global collaborative utility function The formula is: ； in, This represents the total number of strategies currently participating in the optimization. For the first The intent vector of each strategy; This is the mean vector of all strategy intent vectors; Let be the Euclidean norm of the vector, which is used to measure the overall dispersion of all strategic intentions; For candidate instruction sets Total number of devices affected; For indicating functions, when the device exist The value is 1 if there is a logical conflict in the instruction; otherwise, it is 0. For equipment Frequency of state switching during the historical observation period; This is the conflict penalty intensity coefficient used to control the severity of command conflict penalties; This refers to the frequency adjustment coefficient used to adjust the weight of the penalty term on the effect of equipment switching frequency; the conflict penalty intensity coefficient. and frequency adjustment coefficient When deploying the platform, the administrator sets it in the configuration interface according to the park management strategy, and then dynamically adjusts it through iterative learning algorithms based on feedback data of the collaborative optimization results based on historical instructions.

Images