Intelligent safety helmet management method and equipment based on Internet of Things

Abstract

The invention discloses an intelligent safety helmet management method and device based on the Internet of Things, and belongs to the technical field of wireless communication, and the method comprises the steps: obtaining head motion data collected by an inertial measurement unit in an intelligent safety helmet, and environment state data broadcasted by a wireless beacon deployed in an operation environment; analyzing the head movement data based on a preset dynamic model, generating physiological state features of the person, analyzing the environment state data, and identifying environment risk features; fusing the human physiological state features and the environmental risk features to generate a dynamic coupling risk level; and triggering a corresponding closed-loop safety intervention action based on the dynamic coupling risk level. According to the method, personnel physiological state features and position-based environment risk features are obtained, fusion operation is performed on the two features to generate a dynamic coupling risk level, and then a closed-loop safety intervention action is triggered, so that accurate identification and active intervention of personnel and environment coupling risks can be realized.

Description

Technical Field

[0001] This invention relates to the field of wireless communication technology, and in particular to a smart safety helmet management method and device based on the Internet of Things. Background Technology

[0002] As a product of the deep integration of personal protective equipment and Internet of Things (IoT) technology, smart safety helmets represent an important development direction in modern industrial safety management. By integrating various sensors and wireless communication modules, they expand the function of traditional safety helmets from simple physical protection to real-time perception and interaction of worker status and environmental information, providing a key technological carrier for achieving proactive and intelligent safety management.

[0003] In existing technologies, some smart safety helmet management solutions utilize GPS or UWB technology to acquire the location information of workers and upload the location data to a management platform via wireless network to achieve personnel positioning and trajectory tracking. Other solutions focus on monitoring the physiological state of personnel, such as detecting impact events through built-in sensors or using integrated cameras for facial recognition to determine fatigue levels. Regarding environmental perception, fixed sensors, such as gas detectors or infrared detectors, are typically installed in specific areas, and their alarm information is independently sent to the monitoring center.

[0004] However, existing technological solutions have significant shortcomings in practical applications. They typically collect and analyze personnel location information, physiological state information, and environmental risk information as independent dimensions, lacking in-depth analysis of the inherent correlations between these multi-source heterogeneous information. For example, the system may only know that an employee is fatigued and that a piece of equipment is operating at high speed, but it cannot automatically assess the dramatically amplified compound risk that would arise if the fatigued employee approached the high-speed operating equipment. This data-level fragmentation leads to incomplete and inaccurate risk assessments, and its safety intervention measures are often delayed, non-contextualized general alerts, making it difficult to cope with sudden risks in complex work scenarios. Summary of the Invention

[0005] To address the aforementioned issues, this invention provides an IoT-based smart safety helmet management method. This method acquires physiological characteristics of personnel and location-based environmental risk characteristics, and performs fusion calculations on these two characteristics to generate a dynamic coupled risk level, thereby triggering closed-loop safety intervention actions. This method enables accurate identification and proactive intervention of coupled risks between personnel and the environment.

[0006] The above objectives can be achieved through the following approach: The IoT-based smart safety helmet management method includes acquiring head motion data collected by an inertial measurement unit inside the smart safety helmet, and environmental status data broadcast by wireless beacons deployed in the work environment; analyzing the head motion data based on a preset dynamic model to generate personnel physiological state characteristics, and analyzing the environmental status data to identify environmental risk characteristics; fusing the personnel physiological state characteristics and the environmental risk characteristics to generate a dynamically coupled risk level; and triggering corresponding closed-loop safety intervention actions based on the dynamically coupled risk level.

[0007] Optionally, the step of acquiring environmental status data specifically includes: receiving a beacon signal broadcast by the wireless beacon that contains its own identity and real-time operating status; decoding the beacon signal and extracting the real-time operating status to form environmental status data; and matching the environmental status data with a preset environmental risk database to generate environmental risk features, wherein the environmental risk database is obtained from a preset historical security database.

[0008] Optionally, the dynamic model includes: analyzing the temporal pattern of head motion data within a preset time window to obtain head posture temporal parameters; comparing the head posture temporal parameters with a preset fatigue feature parameter set and a preset attention dissipation feature parameter set; and generating corresponding fatigue features or attention dissipation features as physiological state features of the person when the head posture temporal parameters satisfy the fatigue feature parameter set or the preset attention dissipation feature parameter set.

[0009] Optionally, generating the dynamic coupling risk level includes: constructing a risk coupling matrix with personnel physiological state characteristics and environmental risk characteristics as input dimensions; inputting the real-time generated personnel physiological state characteristics and environmental risk characteristics into the risk coupling matrix for calculation; and outputting the quantified dynamic coupling risk level based on the calculation result of the risk coupling matrix.

[0010] Optionally, constructing a risk coupling matrix with personnel physiological state characteristics and environmental risk characteristics as input dimensions further includes: setting initial weight values ​​based on the probability of safety accidents caused by different personnel physiological state characteristics and environmental risk characteristics in a preset historical safety database; monitoring the matching degree between dynamic coupling risk level and actual safety events in real time; and when a safety event actually occurs but the dynamic coupling risk level does not reach the corresponding warning threshold, using a gradient descent algorithm to iteratively optimize the weights of the two input dimensions in the risk coupling matrix according to the severity and feature correlation of the safety event.

[0011] Optionally, the closed-loop safety intervention actions include: determining whether the dynamic coupling risk level reaches a preset level 1 intervention threshold, and issuing a reminder to the wearer through a preset loudspeaker on the smart safety helmet; determining whether the dynamic coupling risk level reaches a preset level 2 intervention threshold, and if so, executing a coordinated warning intervention; determining whether the dynamic coupling risk level reaches a preset level 3 intervention threshold, and if so, executing an active environmental adaptation intervention.

[0012] Optionally, the collaborative warning intervention includes: generating a collaborative warning instruction, the collaborative warning instruction containing source information of the environmental risk characteristics; and broadcasting the collaborative warning instruction via a smart safety helmet.

[0013] Optionally, the proactive environmental adaptation intervention step includes: confirming that the work equipment associated with the environmental risk characteristics has an Internet of Things interface; obtaining the device control system address of the smart safety helmet; sending a preset control command to the device control system address; and executing the preset control command to achieve safe shutdown or reduced speed operation of the work equipment associated with the environmental risk characteristics.

[0014] Optionally, the method further includes: recording the physiological state characteristics of the personnel, the environmental risk characteristics, and the corresponding dynamic coupling risk levels in real time to generate a safety status log; aggregating the safety status logs to form a historical safety database for multiple wearers; and performing statistical analysis on the historical safety database to generate a risk trend report for optimizing management strategies.

[0015] Based on the same inventive concept, the present invention also provides an Internet of Things-based smart helmet management device, comprising: The multi-source data acquisition module is used to acquire head motion data collected by the inertial measurement unit inside the smart safety helmet, as well as environmental status data broadcast by wireless beacons deployed in the work environment; The feature extraction module is used to analyze the head movement data based on a preset dynamic model, generate physiological state features of the person, and analyze the environmental state data to identify environmental risk features. The dynamic coupling assessment module is used to integrate the physiological characteristics of the personnel with the environmental risk characteristics to generate a dynamic coupling risk level. The closed-loop intervention triggering module is used to trigger corresponding closed-loop safety intervention actions based on the dynamic coupling risk level.

[0016] Compared with the prior art, the present invention has the following advantages: 1. By acquiring head movement data collected by smart safety helmets and environmental status data broadcast by wireless beacons in the work environment, real-time dual-dimensional perception of personnel's physiological state and the risks of their environment is achieved. This overcomes the limitations of traditional safety monitoring systems that only focus on personnel location or the status of a single device. It can dynamically correlate personnel's inherent unsafe conditions, such as fatigue or inattention, with specific hazards in the external environment, thereby enabling precise and quantitative assessment of the coupled risks arising from personnel-environment interactions, and improving the comprehensiveness and depth of risk identification.

[0017] 2. A closed-loop, tiered intervention mechanism directly linked to dynamic risk levels has been established. Based on the risk level, equipment can automatically execute differentiated safety actions, ranging from individual alerts and collaborative warnings to proactive environmental control. This intelligent intervention strategy not only ensures the timeliness and effectiveness of safety responses but also, through collaborative warnings via wireless networks and direct control of IoT devices, constructs a collaborative safety protection system between people and between people and machines. This achieves a fundamental shift from passive monitoring to proactive intervention, enhancing the execution capability of on-site safety management.

[0018] 3. A data-driven self-optimization and continuous improvement mechanism has been introduced. By recording and aggregating the safety status logs of all personnel in real time, a historical safety database is formed, and in-depth statistical analysis is performed on it, enabling the equipment to reveal potential safety patterns and risk trends. More importantly, the risk assessment model itself has learning capabilities. Based on feedback from actual safety incidents, it uses a gradient descent algorithm to iteratively optimize the weights of the risk coupling matrix, allowing the accuracy of risk assessment to continuously improve with time and data accumulation. This gives the entire safety management system the ability to dynamically adapt and continuously evolve.

[0019] Other features and advantages of the invention will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures pointed out in the description, claims and drawings. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0021] Figure 1 This is a schematic diagram of the structure of the IoT-based smart safety helmet management method according to an embodiment of the present invention.

[0022] Figure 2 This is a schematic diagram of the risk level curve coupling personnel physiological state characteristics and environmental risk characteristics in an embodiment of the present invention.

[0023] Figure 3 This is a schematic diagram comparing the effects of risk coupling matrix weight optimization in an embodiment of the present invention.

[0024] Figure 4 This is a schematic diagram of the structure of an IoT-based smart safety helmet management device according to an embodiment of the present invention. Detailed Implementation

[0025] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, 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, 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.

[0026] Reference Figure 1 One embodiment of the present invention proposes an Internet of Things-based smart safety helmet management method. This method acquires physiological state characteristics of personnel and location-based environmental risk characteristics, and performs fusion calculations on the two characteristics to generate a dynamic coupled risk level, thereby triggering closed-loop safety intervention actions. This method can achieve accurate identification and proactive intervention of coupled risks between personnel and the environment.

[0027] The method described in this embodiment specifically includes: It acquires head motion data collected by the inertial measurement unit inside the smart helmet, as well as environmental status data broadcast by wireless beacons deployed in the work environment; Based on a preset dynamic model, the head movement data is analyzed to generate physiological state characteristics of the person, and the environmental state data is analyzed to identify environmental risk characteristics. By integrating the physiological characteristics of the personnel with the environmental risk characteristics, a dynamically coupled risk level is generated; Based on the dynamic coupling risk level, a corresponding closed-loop safety intervention action is triggered.

[0028] Specifically, a dual-dimensional perception system integrating individual personnel status and work environment status is constructed. Inertial measurement units within smart safety helmets capture real-time head movement data reflecting the wearer's physiological state, while wireless beacons deployed on-site acquire real-time status data representing environmental hazards. Its core innovation lies in not evaluating these two data sources in isolation, but rather deeply integrating the analyzed personnel physiological characteristics with environmental risk characteristics through a comprehensive analysis model, generating a comprehensive risk level that dynamically reflects the coupling effect. Based on this dynamic coupled risk level, corresponding closed-loop safety interventions are automatically triggered, forming a complete technological closed loop from risk perception and intelligent assessment to proactive intervention. This enhances the intelligence, proactivity, and accuracy of operational safety management. By integrating physiological characteristics directly reflecting personnel safety status with dynamically changing environmental risk characteristics in real time, it overcomes the limitations of traditional safety monitoring that focuses only on a single risk source, enabling the identification of more hidden and dangerous coupled risks arising from the superposition of unsafe personnel states and unsafe environmental factors. This shifts the safety protection checkpoint forward, achieving a fundamental transformation from passive response after an accident to proactive early warning and intervention before accidents occur. A closed-loop intervention mechanism has been established to ensure that risks can be effectively addressed once identified, thereby constructing an intelligent safety protection system that can dynamically adapt to complex operating scenarios and fundamentally reducing the possibility of safety accidents caused by human and environmental factors.

[0029] Optionally, the step of acquiring environmental status data specifically includes: Receive the beacon signal broadcast by the wireless beacon, which includes its own identification and real-time operating status; Decode beacon signals and extract real-time operating status to form environmental status data; Environmental status data is matched with a preset environmental risk database to generate environmental risk characteristics. The environmental risk database is obtained from a preset historical safety database.

[0030] Specifically, multiple wireless beacons are pre-deployed in the work environment. These beacons, such as those using Bluetooth Low Energy (BLE) technology, are fixed to critical work equipment or used to delineate specific high-risk areas. Each beacon is configured with a unique identifier and can obtain the real-time operating status of its associated equipment, such as the operating status of a large crane, the real-time pressure value of a pressure vessel, or gas concentration information in a hazardous chemical storage area, through its built-in sensors or communication interfaces with external devices. These wireless beacons actively broadcast beacon signals containing their own identifier and real-time operating status at a preset broadcast period. The wireless communication module integrated in the smart safety helmet continuously scans and receives these beacon signals. Upon receiving a signal, the microprocessor inside the smart safety helmet immediately performs a decoding operation on the beacon signal, parsing the data frame according to a preset data protocol to accurately extract the wireless beacon's identifier and the real-time operating status representing the equipment's operating condition. The wireless beacon's identifier and the real-time operating status representing the equipment's operating condition constitute environmental status data. This environmental status data is then matched and queried against a preset environmental risk database. This environmental risk database was pre-generated through data mining and statistical analysis of a large-scale, pre-set historical safety database. The pre-set historical safety database stores a large number of historical safety event records, detailing the correlation between equipment status and accident occurrence under different operating conditions. Essentially, the environmental risk database is a knowledge base that directly maps specific wireless beacon identifiers and their different real-time operating states to specific environmental risk characteristics. It transforms raw equipment status data into standardized environmental risk characteristics that can be used for subsequent risk fusion calculations, such as the risk of mechanical rotation injury, the risk of falling objects from heights, or the risk of toxic gas leaks.

[0031] Optionally, the dynamic model includes: Analyze the temporal patterns of head motion data within a preset time window to obtain head posture temporal parameters; The head posture time sequence parameters are compared with the preset fatigue feature parameter set and the preset attention dissipation feature parameter set; When the head posture timing parameters satisfy the fatigue feature parameter set or the preset attention dissipation feature parameter set, the corresponding fatigue feature or attention dissipation feature is generated as the physiological state feature of the person.

[0032] Specifically, the head motion data is collected by the inertial measurement unit (IMU) inside the smart helmet, and features characterizing the person's physiological state are extracted from it. The IMU, typically composed of a three-axis accelerometer and a three-axis gyroscope, can output raw three-dimensional acceleration and angular velocity data of the head in real time at high frequency, forming head motion data. Instead of analyzing the data within a preset time window, a Kalman filter algorithm is used to process the raw acceleration and angular velocity data, calculating more stable and accurate head posture information, such as pitch, roll, and yaw angles. Based on the time series of head posture information, quantified head posture temporal parameters are calculated. These parameters include, but are not limited to, the frequency and amplitude of head swaying in the vertical direction to identify nodding or drowsy movements; the peak value and rate of change of head angular velocity to identify distracting movements such as rapid head turns; and the duration and dispersion of the head maintaining a specific tilt angle. The real-time calculated head posture time-series parameters are matched and compared with two pre-established databases: a preset fatigue feature parameter set and a preset inattention disorientation feature parameter set. These two parameter sets are knowledge bases derived from physiological calibration and data analysis of a large number of experimental samples, defining the quantitative boundaries of typical head movement patterns constituting fatigue or inattention disorientation. The preset inattention disorientation feature parameter set includes definitions for parameters such as excessively rapid head rotation angular velocity and excessively frequent switching of gaze direction. When the combination of head posture time-series parameters calculated within a time window satisfies the rules or model criteria defined in the fatigue feature parameter set, fatigue features are generated. If the parameters satisfy the conditions of the preset inattention disorientation feature parameter set, inattention disorientation features are generated. The generated fatigue or inattention disorientation features are output as physiological state characteristics of the individual.

[0033] Optionally, the generation of dynamic coupling risk levels includes: Construct a risk coupling matrix with personnel physiological state characteristics and environmental risk characteristics as input dimensions; The real-time generated physiological characteristics of personnel and environmental risk characteristics are input into the risk coupling matrix for calculation; Based on the calculation results of the risk coupling matrix, the quantified dynamic coupling risk level is output.

[0034] Specifically, a risk coupling matrix is ​​pre-constructed, using personnel physiological state characteristics and environmental risk characteristics as input dimensions. Logically, the risk coupling matrix is ​​a two-dimensional decision model, with its row vectors corresponding to discretized personnel physiological state characteristics and its column vectors corresponding to discretized environmental risk characteristics. The value of each intersection point in the matrix is ​​a pre-defined quantitative risk score based on historical security big data analysis and a security expert knowledge base. In actual operation, the real-time generated specific personnel physiological state characteristics and environmental risk characteristics are mapped to specific input coordinates of this matrix. Subsequently, these two features are input into the risk coupling matrix for calculation. For dynamically coupled risks, we have: , in, The dynamic coupling risk level, representing the final output, is a quantified numerical value. It is a risk factor score obtained by quantifying the physiological characteristics of the input personnel. The score is directly output by the dynamic model or obtained according to the preset mapping rules. For example, the score is 0 for normal state, 0.4 for mild fatigue, and 0.7 for inattention. It is a risk factor score obtained by quantifying the input environmental risk characteristics. This score is also based on the matching results of the environmental risk database and preset rules. For example, the safe area is 0, the equipment operation area is 0.5, and the high-risk operation area is 0.9. and These are the initial weight values ​​for personnel physiological characteristics and environmental risk characteristics, respectively. They are dimensionless coefficients set based on the statistical analysis of the contribution of these two types of factors to the occurrence of accidents in the historical safety database. They are used to adjust the relative importance of the two types of risks in the comprehensive assessment and usually meet the normalization condition that they add up to 1.

[0035] Optionally, constructing a risk coupling matrix with personnel physiological state characteristics and environmental risk characteristics as input dimensions further includes: Based on the probability of safety accidents caused by different physiological state characteristics of the personnel and environmental risk characteristics in the preset historical safety database, initial weight values ​​are set respectively; The system monitors the matching degree between the dynamic coupling risk level and the actual security event in real time. When a security event actually occurs but the dynamic coupling risk level does not reach the corresponding warning threshold, the gradient descent algorithm is used to iteratively optimize the weights of the two input dimensions in the risk coupling matrix based on the severity and feature correlation of the security event.

[0036] Specifically, initial weight values ​​are set based on a pre-set historical safety database. This database records a large number of past safety incidents or near misses, and associates them with the physiological characteristics of personnel and environmental risk characteristics at the time of the incident. Through data mining and statistical analysis, the conditional probability of a safety accident occurring when different physiological characteristics and environmental risk characteristics occur individually or in combination is calculated. Initial weight values ​​are then set for the physiological characteristics of personnel and environmental risk characteristics in the risk coupling matrix. The matching degree between the dynamic coupling risk level and the actual safety event on site is monitored in real time. When a specific negative feedback event occurs, i.e., a safety event actually occurs on site, but the dynamic coupling risk level calculated by the equipment fails to reach the pre-set warning threshold that should trigger an alarm, the optimization mechanism is triggered. The gradient descent algorithm is used to iteratively optimize the weights of the two input dimensions in the risk coupling matrix. The gradient descent algorithm is an optimization algorithm that finds the minimum value of a function, and the optimization goal is to minimize the error between the predicted risk and the actual risk. The severity of the safety event is assessed, and the specific physiological characteristics of personnel and environmental risk characteristics associated with the event are identified to determine the feature correlation degree. Based on the severity of the event and the correlation between features, the gradient descent algorithm calculates the gradient direction and step size for weight adjustment, and specifically increases the weights corresponding to the features that caused the misjudgment. For the calculation logic of personnel-environment coupling risk, please refer to [link to relevant documentation]. Figure 2 By initially weighting and integrating physiological and environmental risk factors, the risk levels of different combinations clearly correspond to the three-level intervention thresholds, providing a quantitative basis for closed-loop intervention. The weight optimization effect is as follows: Figure 3 As shown, the risk level of the "distracted attention + slippery ground" scenario improved from 0.68 to 0.72 after optimization, exceeding the secondary intervention threshold and avoiding underreporting, thus confirming the effectiveness of the feedback optimization mechanism. Optionally, the closed-loop safety intervention actions include: Determine whether the dynamic coupling risk level has reached the preset first-level intervention threshold, and issue a reminder to the wearer through a preset loudspeaker on the smart safety helmet; Determine whether the dynamic coupling risk level has reached the preset level 2 intervention threshold; if so, execute a collaborative warning intervention. Determine whether the dynamic coupling risk level has reached the preset level three intervention threshold. If so, perform proactive environmental adaptation intervention.

[0037] Specifically, the system determines whether the dynamic coupling risk level reaches or exceeds a preset Level 1 intervention threshold, which corresponds to a relatively low risk level. If the condition is met, the microprocessor inside the smart helmet immediately triggers a reminder action targeting the wearer. This involves driving a preset miniature speaker on the helmet to play a pre-set voice prompt and emit a specific warning sound effect, providing a direct and low-interference reminder to the individual. If the dynamic coupling risk level continues to rise and reaches or exceeds the preset Level 2 intervention threshold, the device determines that the risk is beyond the individual's awareness and requires the cooperation of those around them for prevention. At this point, it will automatically execute a collaborative warning intervention. When the risk level further escalates, reaching or exceeding the preset Level 3 intervention threshold (the highest level), it indicates that the current state constitutes a serious or immediate danger. A simple warning is insufficient to ensure safety, and the device will immediately initiate the highest level of proactive environmental adaptation intervention, actively intervening and changing the risk state of the environment.

[0038] Optionally, the collaborative alert intervention includes: Generate a collaborative warning instruction, wherein the collaborative warning instruction contains source information of the environmental risk characteristics; Alignment instructions are broadcast via smart safety helmets.

[0039] Specifically, when the dynamic coupling risk level reaches a preset secondary intervention threshold, a collaborative warning instruction is generated. This instruction is a structured data set containing not only the warning message itself but also information about the source of the environmental risk that triggered the warning. For example, if the risk is caused by the coupling of a nearby high-speed crane and the wearer's inattentive state, the collaborative warning instruction will explicitly include the identification of the crane's wireless beacon. After the instruction is generated, the smart helmet uses its integrated wireless communication module to broadcast the collaborative warning instruction. The broadcast signal covers a preset collaborative work area, ensuring that workers wearing other smart helmets or deployed smart warning devices within that area can receive it. Other smart terminals that receive the instruction immediately parse its content and trigger their respective warning devices, such as through voice broadcasts or sound and light alarms, thereby achieving targeted and precise notification to those in the vicinity.

[0040] Optionally, the proactive environmental adaptation intervention steps include: Confirm that the operating equipment associated with the aforementioned environmental risk characteristics has an Internet of Things (IoT) interface; Obtain the device control system address of the smart safety helmet; Send preset control commands to the device control system address; Execute preset control commands to achieve safe shutdown or reduced speed operation of the work equipment associated with environmental risk characteristics.

[0041] Specifically, when an environmental adaptive intervention is triggered, the system queries a pre-defined equipment asset database to confirm whether the operational equipment directly associated with the highest level of environmental risk has an IoT interface. This database stores the IoT connectivity and communication protocol information of all critical equipment in the operational environment. After confirming that the equipment has an IoT interface, the system obtains the equipment's control system address and sends a pre-defined control command, based on the equipment type and risk scenario, to that address. The command explicitly requires the equipment to perform risk-avoidance actions such as safe shutdown or reduced speed. Upon receiving and verifying the command's validity, the equipment's control system immediately executes it.

[0042] Optionally, the method further includes: The system records the physiological state characteristics of the personnel, the environmental risk characteristics, and the corresponding dynamic coupling risk levels in real time, and generates a safety status log. Aggregate security status logs to form a historical security database for multiple wearers; Perform statistical analysis on historical security databases to generate risk trend reports for optimizing management strategies.

[0043] Specifically, a data tuple is generated for each wearer, containing personal physiological characteristics, environmental risk characteristics, and the final dynamic coupling risk level, which together constitute the original safety status log. All safety status logs are aggregated to form a centralized historical safety database. This historical safety database is analyzed to generate a risk trend report for optimizing management strategies. This analysis process includes introducing a risk contribution rate indicator to assess the contribution of specific risk combinations to overall high-risk events; the calculation formula is as follows: , in, This represents the risk contribution rate resulting from the combination of specific personnel physiological state characteristics i and environmental risk characteristics j. In the historical safety database, this represents the total number of events where the specific combination (i,j) occurs and its dynamic coupling risk level exceeds a preset high-risk threshold. This value is obtained by filtering and counting the database under multiple conditions. N_{t} represents the total number of events in the same database where the dynamic coupling risk level exceeds this high-risk threshold. This value is obtained by filtering and counting the risk level under a single condition. To measure the average risk level of a specific risk combination, an average risk intensity index is introduced, and its calculation formula is as follows: , in, This represents the average risk intensity of a specific combination (i,j). It is the actual value of the dynamic coupling risk level recorded in the security status log when the combination occurs for the kth time. This represents the total number of times the specific combination (i,j) appears in the historical safety database, regardless of its risk level. This indicator reflects the inherent, averaged level of danger in a given scenario. To assess individual safety performance, a personal risk index is introduced, calculated as follows: , in, A personal risk index representing a specific wearer u. It is the dynamic coupling risk level recorded in the m-th security status log of the wearer. This refers to the total number of safety status log entries generated by the wearer during the analysis period. This index comprehensively reflects the individual's average risk exposure level during work. These calculated quantitative indicators collectively constitute the core content of the risk trend report.

[0044] Based on the same inventive concept, such as Figure 4 As shown, the present invention also provides an IoT-based smart helmet management device, the device comprising: The multi-source data acquisition module is used to acquire head motion data collected by the inertial measurement unit inside the smart safety helmet, as well as environmental status data broadcast by wireless beacons deployed in the work environment; The feature extraction module is used to analyze the head movement data based on a preset dynamic model, generate physiological state features of the person, and analyze the environmental state data to identify environmental risk features. The dynamic coupling assessment module is used to integrate the physiological characteristics of the personnel with the environmental risk characteristics to generate a dynamic coupling risk level. The closed-loop intervention triggering module is used to trigger corresponding closed-loop safety intervention actions based on the dynamic coupling risk level.

[0045] Example 1: To verify the feasibility of this invention in practice, it was applied to the construction site of a large cross-sea bridge project. The project site environment was complex, including multiple high-risk scenarios such as high-altitude work areas, large machinery work areas, and material transportation channels. The workers faced a high level of complex safety risks, and traditional safety management methods were insufficient to achieve real-time, dynamic monitoring and intervention of personnel status and environmental risks.

[0046] This project selected 100 frontline workers for a 6-month application test. Fifty workers wore the smart safety helmet described in this invention as the experimental group, while the other 50 wore traditional safety helmets as the control group. Low-power Bluetooth wireless beacons were pre-deployed at the project site in critical operational equipment such as tower cranes and concrete mixer trucks, and in high-risk areas such as bridge pier edges and precast component hoisting areas. These beacons can broadcast the status of their attached equipment or the risk attributes of the area.

[0047] In practice, the smart safety helmets worn by the experimental group personnel collected head movement data in real time through their built-in inertial measurement unit and received environmental status data broadcast from surrounding wireless beacons via a wireless communication module. For example, at 2:30 PM one afternoon, worker A028 in the experimental group was organizing materials in a material storage area. The dynamic model of his smart safety helmet analyzed his head movement data within a 30-second time window and calculated that his head posture time series parameters showed a low-frequency, large-amplitude nodding pattern. These parameters met the preset fatigue characteristic parameter set, generating the physiological state characteristics of a person with mild fatigue. At this time, no large equipment was operating in the area, and the environmental risk characteristic was a safe zone. The risk coupling matrix fused the mild fatigue characteristics with the safe zone characteristics and their corresponding quantitative risk factors of 0.4 and 0.1, with initial weights set to 0.6 and 0.4, resulting in a dynamic coupling risk level of 0.28. This level exceeded the preset first-level intervention threshold of 0.25, and the smart safety helmet immediately played a voice reminder to worker A028 through a preset loudspeaker: "Fatigue detected. Please rest and ensure safety."

[0048] In another scenario, at 10:10 AM one morning, worker B015 in the experimental group approached a tower crane undergoing hoisting operations. His safety helmet detected excessively rapid head rotation and frequent gaze shifts, generating physiological characteristics of inattentiveness. Simultaneously, the helmet received a signal from a wireless beacon on the tower crane, which, after decoding, identified environmental risk characteristics related to mechanical rotation injury. The calculated dynamic coupling risk level was 0.74. This level exceeded the secondary intervention threshold of 0.70, prompting immediate coordinated warning intervention. Worker B015's helmet generated a coordinated warning instruction containing the tower crane's identification and broadcast it. Other safety helmets within a 50-meter radius of the experimental group, upon receiving the instruction, immediately broadcast to their wearers: "Please note that worker B015 is in a dangerous situation near tower crane number 3. Please pay attention."

[0049] In a high-risk incident, at 16:45 one afternoon, worker C033 in the experimental group, due to continuous work, exhibited extremely low head posture dispersion, remaining leaning forward and stationary for an extended period, indicating severe fatigue. He mistakenly entered a precast component production area where automated rebar tying was underway. The area's beacons broadcast environmental risk characteristics of the high-risk automated operation. The dynamic coupling risk level spiked to 0.90, exceeding the Level 3 intervention threshold of 0.85. An immediate proactive environmental intervention was implemented. After confirming that the automated tying equipment possessed an IoT interface, a pre-set safety shutdown control command was immediately sent to its control system address. The tying equipment safely shut down within 1.5 seconds, averting a serious accident.

[0050] This invention also demonstrates the self-optimization capability of the risk coupling matrix. On a certain day, a worker slipped and fell slightly due to a combination of slippery ground and inattention. The calculated risk level at the time was 0.68, which did not reach the secondary warning threshold of 0.70. This event was recorded as negative feedback, and the gradient descent algorithm was used to iteratively optimize the weights related to inattention and slippery ground risks in the risk coupling matrix. The weights were adjusted to 0.65 and 0.35, respectively, improving the model's sensitivity to this type of combined risk.

[0051] Throughout the testing period, the physiological characteristics, environmental risk characteristics, and dynamically coupled risk levels of all personnel were recorded in real time, forming a historical safety database. Statistical analysis of this database generated a risk trend report, which showed that 2:00 PM to 4:00 PM was the peak period for fatigue, with the highest coupling degree with mechanical injury risk. Based on this report, the project management adjusted the work plan, adding mandatory rest time during this period.

[0052] Data comparison shows that the experimental group recorded approximately 65% ​​fewer near misses over six months compared to the control group, with no major safety incidents occurring, while the control group experienced two reportable safety incidents. This invention achieves precise coupled assessment and closed-loop intervention of risks to workers and the environment, transforming safety management from reactive response to proactive prevention, thereby improving the inherent safety level of construction sites.

[0053] Table 1. Records of Risk Monitoring and Tiered Intervention for Smart Helmets

[0054] Table 2 Self-optimization of risk coupling matrix weights

[0055] Table 3 Comparison of safety performance between the experimental group and the control group (6 months)

[0056] As can be seen from the data in Tables 1 to 3 above, the method of this invention demonstrates excellent safety management results in practical applications. Table 1 clearly shows that it can accurately calculate risk levels and execute corresponding, differentiated intervention actions based on different combinations of personnel and environmental risks, achieving a complete closed loop from individual reminders to environmental control. The data in Table 2 demonstrates its self-learning and optimization capabilities; through feedback from actual events, the model can continuously improve its predictive accuracy. The final comparative data in Table 3 intuitively proves the beneficial effects of this invention. Compared with traditional management methods, the experimental group applying this invention showed a significant reduction in the incidence of both near misses and actual accidents, which fully demonstrates the enormous advantages of this invention in achieving proactive and intelligent accident prevention through technological means.

[0057] It should be noted that the electrical connections between the various units described above do not necessarily represent direct or indirect connections. Any indirect connection method can be applied to the embodiments of the present invention as long as it achieves the purpose of the present invention. The above descriptions are merely exemplary embodiments of the present invention and should not be construed as limiting the scope of the present invention.

[0058] All equivalent changes and modifications made in accordance with the teachings of this invention are still within the scope of this invention. Those skilled in the art will readily conceive of other embodiments of this invention upon considering the specification and the disclosure of practical truth. This application is intended to cover any variations, uses, or adaptations of this invention that follow the general principles of this invention and include common knowledge or conventional techniques in the art not described herein.

Claims

1. A method for managing smart safety helmets based on the Internet of Things, characterized by, The method comprises: acquiring head movement data collected by an inertial measurement unit in a smart safety helmet, and environment state data broadcast by a wireless beacon deployed in a work environment; based on a preset dynamic model, analyzing the head movement data to generate personnel physiological state features, and analyzing the environment state data to identify environment risk features; fusing the personnel physiological state features and the environment risk features to generate a dynamic coupling risk level; based on the dynamic coupling risk level, triggering corresponding closed-loop safety intervention actions.

2. The IoT based smart safety helmet management method as claimed in claim 1 wherein, The step of acquiring environment state data specifically comprises: receiving a beacon signal broadcast by the wireless beacon, the beacon signal containing an identity of the wireless beacon and real-time working state; decoding the beacon signal to extract the real-time working state to form environment state data; matching the environment state data with a preset environment risk database to generate environment risk features, the environment risk database being obtained from a preset historical safety database.

3. The IoT based smart safety cap management method as claimed in claim 1 wherein, The dynamic model comprises: analyzing the time sequence mode of the head movement data within a preset time window to obtain head posture time sequence parameters; comparing the head posture time sequence parameters with a preset set of fatigue feature parameters and a preset set of attention distraction feature parameters; when the head posture time sequence parameters meet the set of fatigue feature parameters or the set of attention distraction feature parameters, generating corresponding fatigue features or attention distraction features as personnel physiological state features.

4. The IoT based smart safety helmet management method as claimed in claim 1 wherein, The generation of the dynamic coupling risk level comprises: constructing a risk coupling matrix with personnel physiological state features and environment risk features as input dimensions; inputting the real-time generated personnel physiological state features and environment risk features into the risk coupling matrix for operation; outputting a quantized dynamic coupling risk level according to the operation result of the risk coupling matrix.

5. The IoT based smart safety helmet management method as claimed in claim 1 wherein, The construction of the risk coupling matrix with personnel physiological state features and environment risk features as input dimensions further comprises: based on the probabilities of different personnel physiological state features and environment risk features leading to safety accidents in the preset historical safety database, setting initial weight values respectively; real-time monitoring the matching degree of the dynamic coupling risk level and actual safety events, when an actual safety event occurs but the dynamic coupling risk level does not reach a corresponding warning threshold, iteratively optimizing the weights of the two input dimensions in the risk coupling matrix according to the severity and feature correlation degree of the safety event using a gradient descent algorithm.

6. The IoT based smart safety helmet management method as claimed in claim 1 wherein, The closed-loop safety intervention actions comprise: judging whether the dynamic coupling risk level reaches a preset first-level intervention threshold, and executing a reminder for the wearer through a preset loudspeaker on the smart safety helmet; judging whether the dynamic coupling risk level reaches a preset second-level intervention threshold, and if so, executing a collaborative warning intervention; judging whether the dynamic coupling risk level reaches a preset third-level intervention threshold, and if so, executing an environment active adaptation intervention.

7. The IoT based smart safety cap management method as claimed in claim 1 wherein, The collaborative warning intervention comprises: generating a collaborative warning instruction, the collaborative warning instruction containing source information of the environment risk features; broadcasting the collaborative warning instruction through the smart safety helmet.

8. The IoT based smart safety cap management method as claimed in claim 1 wherein, The environment active adaptation intervention step comprises: confirming that a work device associated with the environment risk features has an Internet of Things interface; An address of a device control system of the intelligent safety helmet is acquired; A preset control instruction is sent to the address of the device control system; The preset control instruction is executed, and safe shutdown or reduced speed operation of the work equipment associated with the environmental risk feature is realized.

9. The IoT based smart safety cap management method as claimed in claim 1 wherein, The method further includes: Real-time recording of the personnel physiological state feature, the environmental risk feature, and the corresponding dynamic coupling risk level to generate a safety state log; Aggregation of the safety state log to form a historical safety database of multiple wearers; Statistical analysis of the historical safety database to generate a risk trend report for optimizing a management strategy.

10. The intelligent safety helmet management device based on the Internet of Things, applied to the intelligent safety helmet management method based on the Internet of Things as claimed in any one of claims 1-9, characterized in that, The device includes: A multi-source data acquisition module for acquiring head movement data collected by an inertial measurement unit in the intelligent safety helmet and environmental state data broadcast by a wireless beacon deployed in a work environment; A feature extraction module for analyzing the head movement data based on a preset dynamic model to generate a personnel physiological state feature and analyzing the environmental state data to identify an environmental risk feature; A dynamic coupling evaluation module for fusing the personnel physiological state feature and the environmental risk feature to generate a dynamic coupling risk level; A closed-loop intervention triggering module for triggering a corresponding closed-loop safety intervention action based on the dynamic coupling risk level.

Images