Security door state monitoring and early warning system and method based on internet of things

Abstract

This invention discloses an IoT-based security door status monitoring and early warning system and method, belonging to the field of security door monitoring technology. When a dangerous event occurs within a building, each security door node synchronously collects environmental and personnel perception data; analyzes the generated environmental feature vectors and anonymized access vectors, and associates and encapsulates them to form a feature data package; constructs a dynamic map of the building space; analyzes the comprehensive risk coefficient and personnel density distribution of each security door node in real time, and generates a spatial situation diagnosis report; based on the comprehensive risk coefficient and combined with the building's HVAC system operating data, simulates and generates a time-dimension-based risk diffusion heat map; generates a set of candidate collaborative control strategies, constructs an objective optimization function, executes the optimal collaborative control strategy, and performs coefficient correction based on the new data fed back by the security door nodes after strategy execution. This improves the accuracy and expected results of strategy implementation.

Description

Technical Field

[0001] This invention relates to the field of security door monitoring technology, specifically to a security door status monitoring and early warning system and method based on the Internet of Things. Background Technology

[0002] The intelligent security monitoring market, as an important part of the security industry, refers to comprehensive systems that utilize advanced technologies such as artificial intelligence, big data, the Internet of Things, and cloud computing to conduct real-time monitoring, early warning, analysis, and management of public places, corporate parks, residential areas, and other locations. In traditional building security and emergency management, its technical architecture typically follows a basic paradigm of passive response and system fragmentation.

[0003] Existing emergency response mechanisms rely heavily on pre-set, fixed plans and manual decision-making, exhibiting significant rigidity and lag. Whether it's closing fire doors, activating evacuation instructions, or forcing elevator descents, these actions are often triggered by simple thresholds or remote manual operation, making precise timing and logical coordination between systems difficult. This model is completely incapable of handling the dynamic spread of risks; traditional systems cannot anticipate these evolutionary paths, nor can they assess the potential global consequences of different control strategies in real time. It is difficult to achieve a fundamental improvement from passive alarm to proactive, forward-looking, and collaborative risk prevention and control.

[0004] Therefore, this invention discloses a security door status monitoring and early warning system and method based on the Internet of Things to solve the above problems. Summary of the Invention

[0005] The purpose of this invention is to provide a security door status monitoring and early warning system and method based on the Internet of Things to solve the problems raised in the prior art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a security door status monitoring and early warning method based on the Internet of Things, the method comprising the following steps: When a dangerous incident occurs inside the building, each security door node synchronously collects environmental perception data and personnel perception data; analyzes and generates environmental feature vectors and anonymized access vectors, and then associates and encapsulates them to form feature data packets; A dynamic map of building space is constructed based on the spatial relationships of security door nodes; the comprehensive risk coefficient and personnel density distribution of each security door node are analyzed in real time to generate a spatial situation diagnosis report; Based on the comprehensive risk coefficient and combined with the operating data of the building's HVAC system, a risk diffusion heat map with a time dimension is generated after simulation. Based on the status of controllable equipment within the building, a set of candidate collaborative control strategies is generated, an objective optimization function is constructed, the optimal collaborative control strategy is executed, and coefficient correction is performed based on the new data fed back by the security door nodes after the strategy is executed.

[0007] According to the above scheme, the feature data packet generation includes the following: Environmental and personnel perception data are collected synchronously through security door nodes deployed within the building; the environmental perception data includes temperature data from thermal imaging sensors and particulate matter concentration data from smoke detectors; the personnel perception data is anonymized video streams. The temperature data is analyzed to extract abnormally high temperature areas that exceed a preset temperature threshold. The area of ​​the abnormally high temperature area and its ratio to the total area are recorded as the abnormal proportion. The abnormal proportion and particulate matter concentration data are then used to generate an environmental feature vector. The anonymized video stream is processed to extract human body contours, and an anonymized passage vector with time sequence is generated through tracking and counting algorithms. The anonymized passage vector includes the number of people entering the monitoring area per unit time, the number of people leaving the monitoring area per unit time, and the average passage speed. The generated environmental feature vectors are associated and encapsulated with the anonymized travel vectors, and timestamps and node labels are added to the encapsulated data to form a feature data package.

[0008] This invention preprocesses the raw video stream and temperature data at the node end, extracting highly condensed feature values ​​such as anomaly ratio and passage vector, instead of uploading the raw massive data. This reduces network bandwidth pressure and the raw data processing burden on the central system, enabling the system to efficiently support large-scale node deployment. By anonymizing and extracting contours and counting data from the video stream, key pedestrian flow information is obtained while fundamentally avoiding the collection and transmission of personally identifiable information, complying with data privacy protection regulations and enhancing the feasibility of the solution.

[0009] According to the above scheme, the generation of the space situation diagnosis report includes the following: All security door nodes are used as a vertex set; if the areas monitored by two security door nodes are not physically isolated in the building structure, an edge is established between the corresponding security door nodes to form an edge set; a building space graph is constructed based on the vertex set and edge set; dynamic weights are analyzed based on the anonymized access vectors of the two security door nodes and the distance between the two security door nodes. The dynamic weight is equal to the sum of the passage volume of the two security gate nodes divided by the sum of the average passage speeds of the two security gate nodes, and then divided by the distance between the two security gate nodes; the sum of the passage volume of the two security gate nodes is equal to the sum of the number of people entering the monitored area per unit time and the number of people leaving the monitored area per unit time at the two security gate nodes; Based on the environmental feature vectors of each security door node, the normalized anomaly ratio and particulate matter concentration data are weighted and fused to generate local risk assessment indicators; based on the topology and dynamic weights of the edges of the building space map, the comprehensive risk coefficient of each node is calculated through an iterative algorithm; the iteration continues until the comprehensive risk coefficients of all security door nodes converge, resulting in a global comprehensive risk coefficient matrix. Based on the counting information in the passage vectors of all security door nodes, combined with the geographical coordinates of the security door nodes, a real-time personnel density distribution map is generated on the continuous coordinate system of the entire building floor plan using a spatial interpolation algorithm. All security door nodes with a comprehensive risk coefficient exceeding a preset safety threshold are selected and sorted in descending order of their comprehensive risk coefficients. A list of high-risk nodes is generated by attaching security door node identifiers and location descriptions. From the personnel density distribution map, continuous areas with density values ​​exceeding a preset congestion threshold are identified and marked on the building floor plan. The edges with the highest edge weights are identified as the critical paths for personnel flow. Generate a spatial situation diagnostic report, which includes at least: a list of high-risk nodes, identification of areas where people gather, and critical paths for the movement of people.

[0010] This invention incorporates building spatial structure and real-time pedestrian flow into risk value analysis. The risk value of each node depends not only on local sensors but also on the coupling effect of risks from other nodes within its connected area. This simulates the actual propagation characteristics of risks (such as smoke and panic), resulting in a significant improvement in accuracy compared to judging each node in isolation. The definition of dynamic weights (positively correlated with traffic volume and inversely correlated with average speed and distance) has physical meaning. It quantifies the mediating role of pedestrian flow in risk propagation, enabling the model to more accurately identify the channels through which risks spread most rapidly.

[0011] According to the above scheme, the prediction of risk field and population density distribution map includes the following: The comprehensive risk coefficient is used as the initial risk status of each security door node, and the personnel density distribution map is used as the initial personnel exposure basis. The operation data of the building's HVAC system is acquired in real time, including the air supply direction and intensity of the ventilation openings. The risk increment of any adjacent security door node is analyzed to simulate the risk field in several future time steps. Based on the fact that people tend to move from relatively high-risk, high-density areas to relatively low-risk, low-density areas; the speed and direction of movement are jointly influenced by real-time risk gradients, current population density, and known evacuation routes; the predicted population density distribution map is updated at each time step using a pre-trained crowd flow model. After the simulation is completed, the risk field and personnel density distribution maps of each time step are integrated to generate a series of risk diffusion heat maps for several future time periods, and potential bottleneck areas where personnel density exceeds the threshold and the corresponding time nodes are identified.

[0012] According to the above scheme, the incremental transmission risk is equal to the product of the dynamic weight, the comprehensive risk coefficient, and the airflow direction influence factor, divided by the square of the distance between the two security door nodes. The airflow direction influence factor is determined based on the air supply direction of the vent and the direction between the security door nodes. If the air supply direction of the vent is consistent with the direction of security door nodes i to j, then the airflow direction influence factor θ is... （i，j） The first airflow direction influence factor is θ; if the air supply direction of the vent is inconsistent with the direction of security door nodes i to j, then the airflow direction influence factor θ is... （i，j） The first airflow direction influence factor and the second airflow direction influence factor are preset by the system. The first airflow direction influence factor is greater than one, and the second airflow direction influence factor is less than one.

[0013] This invention, through spatiotemporal simulation, enables the system to anticipate the evolution trend and potential bottleneck areas of risks before they fully spread or crowd congestion occurs. This predictability is crucial for improving emergency response efficiency, providing a valuable time window for formulating and issuing control strategies. It explicitly introduces a rule that people tend to move from high-risk, high-density areas to low-risk, low-density areas, and updates the density map through a pre-trained model. This ensures that risk simulation and personnel evacuation dynamics are bidirectionally coupled and mutually influential, making the predicted personnel exposure risk indicators more realistic and reliable.

[0014] According to the above scheme, the coefficient correction includes the following: Based on the status of controllable equipment within the building, a set of candidate collaborative control strategies is generated; each strategy contains combined control instructions for at least two of the following types of equipment: opening and closing status and direction control instructions for each security door node, forced landing and operation control instructions for the elevator system, and supply and exhaust air volume and direction control instructions for the HVAC system. Based on each strategy in the candidate collaborative control strategy set, taking the risk field and personnel density distribution map before the strategy is executed as the starting point and the time of strategy execution as the initial simulation time, a spatiotemporal simulation covering several future time periods is run once according to the iterative method of predicting the risk field and personnel density distribution map; the average value of the risk field and personnel density distribution map output by the spatiotemporal simulation is analyzed as the prediction result of the corresponding strategy. Based on the prediction results after the execution strategy, the target optimization function value is calculated; the target optimization function value is a weighted fusion index of the sum of the predicted comprehensive risk coefficients of all security door nodes and the personnel exposure risk score; the strategy that maximizes the target optimization function value is selected as the optimal collaborative control strategy and sent to the corresponding device; The personnel exposure risk integral is equal to the integral over time of the product of the predicted personnel density and the predicted comprehensive risk coefficient. After executing the optimal collaborative control strategy, new security door node sensing data is received; the predicted data before strategy execution is compared with the actual observation data after strategy execution. If the prediction error continues to exceed the preset threshold, the local risk weight coefficient and the airflow direction influence factor in the comprehensive risk coefficient calculation are adjusted to correct the model parameters.

[0015] The evaluation of the strategy in this invention is not based on simple rules, but rather on sandbox simulation using a pre-calculation model. The system quickly predicts the situational changes over a period of time after each candidate strategy is implemented, and quantitatively calculates the objective function value based on the simulation results. This ensures that the final optimal collaborative control strategy is scientifically predicted and verified, significantly improving the accuracy and expected results of strategy implementation. By comparing the error between prediction and reality and adjusting the model parameters in reverse, the system achieves online learning. This allows the entire system to adapt to changes in building layout, equipment aging, and changes in personnel behavior, continuously optimizing its diagnostic and predictive accuracy, and avoiding the problem of traditional systems gradually diminishing in effectiveness due to model stagnation.

[0016] Another aspect of this application provides an IoT-based security door status monitoring and early warning system, which is applied to the above-mentioned IoT-based security door status monitoring and early warning method. The system includes the data encapsulation module, spatial situation diagnosis module, risk prediction module, and strategy execution feedback module. When a dangerous event occurs within the building, the data encapsulation module synchronously collects environmental perception data and personnel perception data from each security door node; analyzes and generates environmental feature vectors and anonymized access vectors, and then associates and encapsulates them to form feature data packets. The spatial situation diagnosis module constructs a dynamic map of the building space based on the spatial relationships of security door nodes; it analyzes the comprehensive risk coefficient and personnel density distribution of each security door node in real time and generates a spatial situation diagnosis report. The risk prediction module generates a time-dimensional risk diffusion heat map based on a comprehensive risk coefficient and combined with the operating data of the building's HVAC system. The strategy execution feedback module generates a set of candidate collaborative control strategies based on the status of controllable equipment within the building, constructs a target optimization function, executes the optimal collaborative control strategy, and performs coefficient correction based on the new data fed back by the security door nodes after strategy execution.

[0017] According to the above scheme, the data encapsulation module includes a sensing data acquisition unit and a data processing and packaging unit; The sensing data acquisition unit is used to simultaneously collect environmental sensing data and personnel sensing data through security door nodes deployed in the building; wherein, the environmental sensing data includes temperature data from thermal imaging sensors and particulate matter concentration data from smoke detectors; and the personnel sensing data is anonymized video stream. The data processing and packaging unit is used to analyze the temperature data and particulate matter concentration data to generate an environmental feature vector; process the anonymized video stream to generate an anonymized access vector; associate and encapsulate the generated environmental feature vector with the anonymized access vector; and add timestamps and node tags to the encapsulated data to form a feature data package.

[0018] According to the above scheme, the spatial situation diagnosis module includes a risk analysis unit, a personnel gathering analysis unit, and a spatial situation diagnosis report analysis unit; The risk analysis unit is used to construct a building space map. Based on the environmental feature vectors of each security door node, it performs weighted fusion of normalized anomaly ratio and particulate matter concentration data to generate local risk assessment indicators. Based on the topology and dynamic weights of the edges of the building space map, it calculates the comprehensive risk coefficient of each node through an iterative algorithm. The personnel aggregation analysis unit is used to generate a real-time personnel density distribution map on the continuous coordinate system of the entire building plan based on the counting information in the passage vectors of all security door nodes and the geographical coordinates of the security door nodes, using spatial interpolation algorithms (such as inverse distance weighting or Kriging interpolation). The spatial situation diagnosis report analysis unit is used to filter out all security door nodes whose comprehensive risk coefficient exceeds the preset safety threshold, and arrange them in descending order of their comprehensive risk coefficient, attaching security door node identification and location description to generate a high-risk node list; from the personnel density distribution map, it identifies continuous areas whose density value exceeds the preset congestion threshold and marks them on the building floor plan; it identifies the several edges with the highest edge weight as the critical paths of personnel flow; and generates a spatial situation diagnosis report.

[0019] According to the above scheme, the risk prediction module includes a risk field analysis unit and a personnel density distribution analysis unit; The method uses the comprehensive risk coefficient as the initial risk state of each security door node, the personnel density distribution map as the initial personnel exposure basis, and the real-time acquisition of the building's HVAC system operation data, including the air supply direction and intensity of the ventilation openings; analyzes the propagation risk increment for any adjacent security door node, and simulates the risk field within several future time steps. The personnel density distribution analysis unit is used to analyze the tendency of people to move from relatively high-risk, high-density areas to relatively low-risk, low-density areas. The speed and direction of movement are jointly affected by the real-time risk gradient, the current personnel density, and known evacuation routes. The predicted personnel density distribution map is updated at each time step using a pre-trained crowd flow model.

[0020] According to the above scheme, the strategy execution feedback module includes a prediction execution unit and a feedback comparison and correction unit; The execution prediction unit is used to generate a set of candidate collaborative control strategies based on the status of controllable equipment in the building; based on each strategy in the set of candidate collaborative control strategies, it runs a spatiotemporal simulation covering several future time periods according to the iterative method of risk field and personnel density distribution map prediction; based on the prediction results after executing the strategy, it calculates the target optimization function value; selects the strategy that maximizes the target optimization function value as the optimal collaborative control strategy, and sends it to the corresponding equipment; The feedback comparison and correction unit is used to receive new security door node sensing data after executing the optimal collaborative control strategy; compare the predicted data before strategy execution with the actual observation data after strategy execution; if the prediction error continues to exceed the preset threshold, adjust the local risk weight coefficient and the air flow direction influence factor in the comprehensive risk coefficient calculation to correct the model parameters.

[0021] Compared with existing technologies, the beneficial effects of this invention are as follows: This invention preprocesses the raw video stream and temperature data at the node end, extracting highly condensed feature values ​​such as anomaly ratio and traffic vector, instead of uploading the raw massive data. This greatly reduces network bandwidth pressure and the raw data processing burden of the central system, enabling the system to efficiently support large-scale node deployment. By anonymizing the contour extraction and counting of the video stream, key pedestrian flow information is obtained while fundamentally avoiding the collection and transmission of personal identity information, complying with data privacy protection regulations and enhancing the feasibility of the solution. This invention incorporates building space structure and real-time pedestrian flow into risk value analysis. The risk value of each node depends not only on local sensors but also on the coupling effect of risks from other nodes in its connected area. This simulates the actual propagation characteristics of risk, resulting in a qualitative improvement in accuracy compared to judging each node in isolation. The definition of dynamic weights has physical meaning. It quantifies the mediating role of pedestrian flow in risk propagation, enabling the model to more accurately identify the channels through which risks spread most rapidly. Through spatiotemporal extrapolation, this invention allows the system to predict the evolution trend and potential bottleneck areas before the risk fully spreads or crowd congestion forms. This foresight is key to improving emergency response efficiency, providing a valuable time window for formulating and issuing control strategies. It explicitly introduces rules indicating a tendency for personnel to move from high-risk, high-density areas to low-risk, low-density areas, and updates the density map through a pre-trained model. This ensures that risk simulation and personnel evacuation dynamics are bidirectionally coupled and mutually influential, making the predicted personnel exposure risk indicators more realistic and reliable. The evaluation of the strategy in this invention is not based on simple rules, but rather on sand table simulation using a pre-calculation model. The system quickly pre-simulates the situational changes over a period of time after each candidate strategy is implemented, and quantitatively calculates the objective function value based on the simulation results. This ensures that the final optimal collaborative control strategy is scientifically predicted and verified, significantly improving the accuracy and expected effects of strategy implementation. By comparing the error between prediction and reality and adjusting the model parameters in reverse, the system achieves online learning. This allows the entire system to adapt to changes in building layout, equipment aging, and changes in personnel behavior, continuously optimizing its diagnostic and predictive accuracy, and avoiding the problem of traditional systems gradually diminishing in effectiveness due to model stagnation. Attached Figure Description

[0022] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings: Figure 1 This is a flowchart illustrating the IoT-based security door status monitoring and early warning method of the present invention. Figure 2 This is a schematic diagram of the structure of the Internet of Things-based security door status monitoring and early warning system of the present invention. Detailed Implementation

[0023] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0024] Please see Figure 1 The present invention provides a technical solution: a security door status monitoring and early warning method based on the Internet of Things, the method comprising the following steps: When a dangerous incident occurs inside the building, each security door node synchronously collects environmental perception data and personnel perception data; analyzes and generates environmental feature vectors and anonymized access vectors, and then associates and encapsulates them to form feature data packets; The feature data packet generation includes the following: Environmental and personnel perception data are collected synchronously through security door nodes deployed within the building; the environmental perception data includes temperature data from thermal imaging sensors and particulate matter concentration data from smoke detectors; the personnel perception data is anonymized video streams. Analyze the temperature data, extract abnormally high temperature areas where the temperature exceeds the preset temperature threshold, and record the area of ​​the abnormally high temperature area and its ratio to the total area as the abnormal proportion; then combine the abnormal proportion with the particulate matter concentration data to generate an environmental feature vector. The anonymized video stream is processed to extract human body contours, and an anonymized passage vector with time sequence is generated through tracking and counting algorithms. The anonymized passage vector includes the number of people entering the monitoring area per unit time, the number of people leaving the monitoring area per unit time, and the average passage speed. The generated environmental feature vectors are associated and encapsulated with the anonymized travel vectors, and timestamps and node labels are added to the encapsulated data to form a feature data package.

[0025] A dynamic map of building space is constructed based on the spatial relationships of security door nodes; the comprehensive risk coefficient and personnel density distribution of each security door node are analyzed in real time to generate a spatial situation diagnosis report; The generation of the space situation diagnostic report includes the following: All security door nodes are used as the vertex set; if the areas monitored by two security door nodes are not physically isolated in the building structure, an edge is established between the corresponding security door nodes to form an edge set; a building space graph is constructed based on the vertex set and the edge set; dynamic weights are analyzed based on the anonymized access vectors of the two security door nodes and the distance between the two security door nodes. The dynamic weight equals the sum of the passage volume of the two security gate nodes divided by the sum of the average passage speeds of the two security gate nodes, and then divided by the distance between the two security gate nodes; the sum of the passage volume of the two security gate nodes equals the sum of the number of people entering the monitored area per unit time and the number of people leaving the monitored area per unit time at the two security gate nodes; Based on the environmental feature vectors of each security door node, the normalized anomaly ratio and particulate matter concentration data are weighted and fused to generate local risk assessment indicators; based on the topology of the building space graph and the dynamic weights of the edges, the comprehensive risk coefficient of each node is calculated through an iterative algorithm; the iteration continues until the comprehensive risk coefficients of all security door nodes converge, resulting in a global comprehensive risk coefficient matrix. Example 1: In this example: ; Among them, R i (n) represents the comprehensive risk coefficient of the i-th security door node after the nth iteration; Rlocal i Let γ represent the local risk assessment index of the i-th security door node; γ is the local risk weight coefficient; N(i) is the set of neighboring nodes of security door node i; w （i，j） R represents the dynamic weights of the i-th security door node and the j-th security door node; j (n-1) represents the comprehensive risk coefficient of the j-th security door node after the nth iteration; Based on the counting information in the passage vectors of all security door nodes, combined with the geographical coordinates of the security door nodes, a real-time personnel density distribution map is generated on the continuous coordinate system of the entire building floor plan using the inverse distance weighting method or the Kriging interpolation method. All security door nodes with a comprehensive risk coefficient exceeding a preset safety threshold are selected and sorted in descending order of their comprehensive risk coefficients. A list of high-risk nodes is generated by attaching security door node identifiers and location descriptions. From the personnel density distribution map, continuous areas with density values ​​exceeding a preset congestion threshold are identified and marked on the building floor plan. The edges with the highest edge weights are identified as the critical paths for personnel flow. Generate a spatial situation diagnostic report, which should include at least: a list of high-risk nodes, identification of areas where people congregate, and critical paths for the movement of people.

[0026] Based on the comprehensive risk coefficient and combined with the operating data of the building's HVAC system, a risk diffusion heat map with a time dimension is generated after simulation. The prediction of risk field and population density distribution maps includes the following: The comprehensive risk coefficient is used as the initial risk status of each security door node, and the personnel density distribution map is used as the initial personnel exposure basis. The operation data of the building's HVAC system is acquired in real time, including the air supply direction and intensity of the ventilation openings. The risk increment of any adjacent security door node is analyzed to simulate the risk field in several future time steps. Based on the fact that people tend to move from relatively high-risk, high-density areas to relatively low-risk, low-density areas; the speed and direction of movement are jointly influenced by real-time risk gradients, current population density, and known evacuation routes; the predicted population density distribution map is updated at each time step using a pre-trained crowd flow model. After the simulation is completed, the risk field and personnel density distribution maps of each time step are integrated to generate a series of risk diffusion heat maps for several future time periods, and potential bottleneck areas where personnel density exceeds the threshold and the corresponding time nodes are identified.

[0027] The incremental risk of transmission is equal to the product of the dynamic weight, the comprehensive risk coefficient, and the airflow direction influence factor, divided by the square of the distance between the two security door nodes. The airflow direction influence factor is determined based on the air supply direction of the vent and the direction between the security door nodes. If the air supply direction of the vent is consistent with the direction of security door nodes i to j, then the airflow direction influence factor θ is... （i，j） The first airflow direction influence factor is θ; if the air supply direction of the vent is inconsistent with the direction of security door nodes i to j, then the airflow direction influence factor θ is... （i，j） The first airflow direction influence factor is the second airflow direction influence factor; the first airflow direction influence factor and the second airflow direction influence factor are preset by the system, the first airflow direction influence factor is greater than one, and the second airflow direction influence factor is less than one.

[0028] Based on the status of controllable equipment within the building, a set of candidate collaborative control strategies is generated, an objective optimization function is constructed, the optimal collaborative control strategy is executed, and coefficient correction is performed based on the new data fed back by the security door nodes after the strategy is executed.

[0029] The coefficient correction includes the following: Based on the status of controllable equipment within the building, a set of candidate collaborative control strategies is generated; each strategy contains combined control instructions for at least two of the following types of equipment: opening and closing status and direction control instructions for each security door node, forced landing and operation control instructions for the elevator system, and supply and exhaust air volume and direction control instructions for the HVAC system. Based on each strategy in the candidate collaborative control strategy set, taking the risk field and personnel density distribution map before the strategy is executed as the starting point and the time of strategy execution as the initial simulation time, a spatiotemporal simulation covering several future time periods is run once according to the iterative method of predicting the risk field and personnel density distribution map; the average value of the risk field and personnel density distribution map output by the spatiotemporal simulation is analyzed as the prediction result of the corresponding strategy. Based on the prediction results after the execution strategy, the target optimization function value is calculated; the target optimization function value is a weighted fusion index of the sum of the predicted comprehensive risk coefficients of all security door nodes and the personnel exposure risk score; the strategy that maximizes the target optimization function value is selected as the optimal collaborative control strategy and distributed to the corresponding devices. The personnel exposure risk score is equal to the integral over time of the product of the predicted personnel density and the predicted comprehensive risk coefficient. After executing the optimal collaborative control strategy, new security door node sensing data is received; the predicted data before the strategy is executed is compared with the actual observation data after the strategy is executed. If the prediction error continues to exceed the preset threshold, the local risk weight coefficient in the comprehensive risk coefficient calculation and the air flow direction influence factor are adjusted to correct the model parameters.

[0030] Please see Figure 2 The present invention provides a technical solution: a security door status monitoring and early warning system based on the Internet of Things, which includes a data encapsulation module, a spatial situation diagnosis module, a risk prediction module, and a strategy execution feedback module; When a dangerous event occurs within the building, the data encapsulation module synchronously collects environmental perception data and personnel perception data from each security door node; analyzes and generates environmental feature vectors and anonymized access vectors, and then associates and encapsulates them to form feature data packets. The spatial situation diagnosis module constructs a dynamic map of the building space based on the spatial relationships of security door nodes; it analyzes the comprehensive risk coefficient and personnel density distribution of each security door node in real time and generates a spatial situation diagnosis report. The risk prediction module generates a time-dimensional risk diffusion heat map based on a comprehensive risk coefficient and combined with the operating data of the building's HVAC system. The strategy execution feedback module generates a set of candidate collaborative control strategies based on the status of controllable equipment within the building, constructs an objective optimization function, executes the optimal collaborative control strategy, and performs coefficient correction based on the new data fed back by the security door nodes after strategy execution.

[0031] The data encapsulation module includes a sensing data acquisition unit and a data processing and packaging unit; The sensing data acquisition unit is used to simultaneously collect environmental sensing data and personnel sensing data through security door nodes deployed within the building; the environmental sensing data includes temperature data from thermal imaging sensors and particulate matter concentration data from smoke detectors; the personnel sensing data is anonymized video streams. The data processing and packaging unit is used to analyze temperature data and particulate matter concentration data to generate environmental feature vectors; it processes the anonymized video stream to generate anonymized pass vectors, associates and encapsulates the generated environmental feature vectors with the anonymized pass vectors, and adds timestamps and node tags to the encapsulated data to form feature data packages.

[0032] The space situation diagnosis module includes a risk analysis unit, a personnel gathering analysis unit, and a space situation diagnosis report analysis unit; The risk analysis unit is used to construct the building space map. Based on the environmental feature vectors of each security door node, it performs weighted fusion of normalized anomaly ratio and particulate matter concentration data to generate local risk assessment indicators. Based on the topology of the building space map and the dynamic weights of the edges, it calculates the comprehensive risk coefficient of each node through an iterative algorithm. The personnel aggregation analysis unit is used to generate a real-time personnel density distribution map on the continuous coordinate system of the entire building plan based on the counting information in the passage vectors of all security door nodes and the geographical coordinates of the security door nodes, using spatial interpolation algorithms (such as inverse distance weighting or kriging interpolation). The spatial situation diagnosis report analysis unit is used to filter out all security door nodes whose comprehensive risk coefficient exceeds the preset safety threshold, and sort them in descending order of their comprehensive risk coefficient, attaching security door node identification and location description to generate a list of high-risk nodes; from the personnel density distribution map, it identifies continuous areas whose density value exceeds the preset congestion threshold and marks them on the building floor plan; it identifies the edges with the highest edge weights as the critical paths for personnel flow; and generates a spatial situation diagnosis report.

[0033] The risk prediction module includes a risk field analysis unit and a personnel density distribution analysis unit; The comprehensive risk coefficient is used as the initial risk status of each security door node, and the personnel density distribution map is used as the initial personnel exposure basis. The operation data of the building's HVAC system is acquired in real time, including the air supply direction and intensity of the ventilation openings. The risk increment of any adjacent security door node is analyzed to simulate the risk field in several future time steps. The population density distribution analysis unit is used to analyze the tendency of people to move from relatively high-risk, high-density areas to relatively low-risk, low-density areas. The speed and direction of movement are jointly influenced by the real-time risk gradient, the current population density, and known evacuation routes. The predicted population density distribution map is updated at each time step using a pre-trained population flow model.

[0034] The strategy execution feedback module includes a prediction execution unit and a feedback comparison and correction unit; The execution prediction unit is used to generate a set of candidate collaborative control strategies based on the status of controllable equipment in the building; based on each strategy in the candidate collaborative control strategy set, it runs a spatiotemporal simulation covering several future time periods according to the iterative method of predicting risk field and personnel density distribution map; based on the prediction results after executing the strategy, it calculates the target optimization function value; selects the strategy that maximizes the target optimization function value as the optimal collaborative control strategy, and sends it to the corresponding equipment; The feedback comparison and correction unit is used to receive new security door node sensing data after the optimal collaborative control strategy is executed; it compares the predicted data before the strategy is executed with the actual observation data after the strategy is executed. If the prediction error continues to exceed the preset threshold, it adjusts the local risk weight coefficient in the comprehensive risk coefficient calculation and adjusts the air flow direction influence factor to correct the model parameters.

[0035] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "include," "contain," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus.

[0036] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims.

Claims

1. A security door status monitoring and early warning method based on the Internet of Things, characterized in that, The method includes the following steps: When a dangerous incident occurs inside the building, each security door node synchronously collects environmental perception data and personnel perception data; analyzes and generates environmental feature vectors and anonymized access vectors, and then associates and encapsulates them to form feature data packets; Constructing a dynamic architectural space diagram based on the spatial relationships of security door nodes; Real-time analysis of the comprehensive risk coefficient and personnel density distribution of each security door node, generating a spatial situation diagnosis report; Based on the comprehensive risk coefficient and combined with the operating data of the building's HVAC system, a risk diffusion heat map with a time dimension is generated after simulation. Based on the status of controllable equipment within the building, a set of candidate collaborative control strategies is generated, an objective optimization function is constructed, the optimal collaborative control strategy is executed, and coefficient correction is performed based on the new data fed back by the security door nodes after the strategy is executed.

2. The IoT-based security door status monitoring and early warning method according to claim 1, characterized in that: The feature data packet generation includes the following: Environmental and personnel perception data are collected synchronously through security door nodes deployed within the building; the environmental perception data includes temperature data from thermal imaging sensors and particulate matter concentration data from smoke detectors; the personnel perception data is anonymized video streams. The temperature data is analyzed to extract abnormally high temperature areas that exceed a preset temperature threshold. The area of ​​the abnormally high temperature area and its ratio to the total area are recorded as the abnormal proportion. The abnormal proportion and particulate matter concentration data are then used to generate an environmental feature vector. The anonymized video stream is processed to extract human body contours, and an anonymized passage vector with time sequence is generated through tracking and counting algorithms. The anonymized passage vector includes the number of people entering the monitoring area per unit time, the number of people leaving the monitoring area per unit time, and the average passage speed. The generated environmental feature vectors are associated and encapsulated with the anonymized travel vectors, and timestamps and node labels are added to the encapsulated data to form a feature data package.

3. The IoT-based security door status monitoring and early warning method according to claim 2, characterized in that: The generation of the space situation diagnostic report includes the following: All security door nodes are used as a vertex set; if the areas monitored by two security door nodes are not physically isolated in the building structure, an edge is established between the corresponding security door nodes to form an edge set; a building space graph is constructed based on the vertex set and edge set; dynamic weights are analyzed based on the anonymized access vectors of the two security door nodes and the distance between the two security door nodes. Based on the environmental feature vectors of each security door node, the normalized anomaly ratio and particulate matter concentration data are weighted and fused to generate local risk assessment indicators; based on the topology and dynamic weights of the edges of the building space map, the comprehensive risk coefficient of each node is calculated through an iterative algorithm; the iteration continues until the comprehensive risk coefficients of all security door nodes converge, resulting in a global comprehensive risk coefficient matrix. Based on the counting information in the passage vectors of all security door nodes, combined with the geographical coordinates of the security door nodes, a real-time personnel density distribution map is generated on the continuous coordinate system of the entire building floor plan using a spatial interpolation algorithm. All security door nodes with a comprehensive risk coefficient exceeding a preset safety threshold are selected and sorted in descending order of their comprehensive risk coefficients. A list of high-risk nodes is generated by attaching security door node identifiers and location descriptions. From the personnel density distribution map, continuous areas with density values ​​exceeding a preset congestion threshold are identified and marked on the building floor plan. The edges with the highest edge weights are identified as the critical paths for personnel flow. Generate a spatial situation diagnostic report, which includes at least: a list of high-risk nodes, identification of areas where people gather, and critical paths for the movement of people.

4. The IoT-based security door status monitoring and early warning method according to claim 3, characterized in that: The dynamic weight is equal to the sum of the passage volume of the two security gate nodes divided by the sum of the average passage speed of the two security gate nodes, and then divided by the distance between the two security gate nodes; the sum of the passage volume of the two security gate nodes is equal to the sum of the number of people entering the monitoring area per unit time and the number of people leaving the monitoring area per unit time.

5. The IoT-based security door status monitoring and early warning method according to claim 4, characterized in that: The prediction of risk field and population density distribution maps includes the following: The comprehensive risk coefficient is used as the initial risk status of each security door node, and the personnel density distribution map is used as the initial personnel exposure basis. The operation data of the building's HVAC system is acquired in real time, including the air supply direction and intensity of the ventilation openings. The risk increment of any adjacent security door node is analyzed to simulate the risk field in several future time steps. Based on the fact that people tend to move from relatively high-risk, high-density areas to relatively low-risk, low-density areas; the speed and direction of movement are jointly influenced by real-time risk gradients, current population density, and known evacuation routes; the predicted population density distribution map is updated at each time step using a pre-trained crowd flow model. After the simulation is completed, the risk field and personnel density distribution maps of each time step are integrated to generate a series of risk diffusion heat maps for several future time periods, and potential bottleneck areas where personnel density exceeds the threshold and the corresponding time nodes are identified.

6. The IoT-based security door status monitoring and early warning method according to claim 5, characterized in that: The increment of transmission risk is equal to the product of dynamic weight, comprehensive risk coefficient, and airflow direction influence factor, divided by the square of the distance between the two security door nodes. The airflow direction influence factor is determined based on the air supply direction of the vent and the direction between the security door nodes. If the air supply direction of the vent is consistent with the direction of security door nodes i to j, then the airflow direction influence factor θ is... （i，j） The first airflow direction influence factor is θ; if the air supply direction of the vent is inconsistent with the direction of security door nodes i to j, then the airflow direction influence factor θ is... （i，j） The first airflow direction influence factor and the second airflow direction influence factor are preset by the system. The first airflow direction influence factor is greater than one, and the second airflow direction influence factor is less than one.

7. The IoT-based security door status monitoring and early warning method according to claim 6, characterized in that: The coefficient correction includes the following: Based on the status of controllable equipment within the building, a set of candidate collaborative control strategies is generated; each strategy contains combined control instructions for at least two of the following types of equipment: opening and closing status and direction control instructions for each security door node, forced landing and operation control instructions for the elevator system, and supply and exhaust air volume and direction control instructions for the HVAC system. Based on each strategy in the candidate collaborative control strategy set, taking the risk field and personnel density distribution map before the strategy is executed as the starting point and the time of strategy execution as the initial simulation time, a spatiotemporal simulation covering several future time periods is run once according to the iterative method of predicting the risk field and personnel density distribution map; the average value of the risk field and personnel density distribution map output by the spatiotemporal simulation is analyzed as the prediction result of the corresponding strategy. Based on the prediction results after the execution strategy, the target optimization function value is calculated; the target optimization function value is a weighted fusion index of the sum of the predicted comprehensive risk coefficients of all security door nodes and the personnel exposure risk score; the strategy that maximizes the target optimization function value is selected as the optimal collaborative control strategy and sent to the corresponding device; After executing the optimal collaborative control strategy, new security door node sensing data is received; the predicted data before strategy execution is compared with the actual observation data after strategy execution. If the prediction error continues to exceed the preset threshold, the local risk weight coefficient and the airflow direction influence factor in the comprehensive risk coefficient calculation are adjusted to correct the model parameters.

8. The IoT-based security door status monitoring and early warning method according to claim 7, characterized in that: The personnel exposure risk integral is equal to the integral over time of the product of the predicted personnel density and the predicted comprehensive risk coefficient.

9. A security door status monitoring and early warning system based on the Internet of Things (IoT), wherein the system is applied to the security door status monitoring and early warning method based on the IoT as described in any one of claims 1-8, characterized in that, The system includes the data encapsulation module, the spatial situation diagnosis module, the risk prediction module, and the strategy execution feedback module; When a dangerous event occurs within the building, the data encapsulation module synchronously collects environmental perception data and personnel perception data from each security door node; analyzes and generates environmental feature vectors and anonymized access vectors, and then associates and encapsulates them to form feature data packets. The spatial situation diagnosis module constructs a dynamic map of the building space based on the spatial relationships of security door nodes; Real-time analysis of the comprehensive risk coefficient and personnel density distribution of each security door node, generating a spatial situation diagnosis report; The risk prediction module generates a time-dimensional risk diffusion heat map based on a comprehensive risk coefficient and combined with the operating data of the building's HVAC system. The strategy execution feedback module generates a set of candidate collaborative control strategies based on the status of controllable equipment within the building, constructs a target optimization function, executes the optimal collaborative control strategy, and performs coefficient correction based on the new data fed back by the security door nodes after strategy execution.

Images