Fall detection early warning method and device, and electronic device

Abstract

The application provides a falling detection early warning method and device and electronic equipment, and belongs to the technical field of personnel safety monitoring. The method comprises the following steps: obtaining historical falling accident experimental data, determining a falling accident prior probability based on the historical falling accident experimental data; determining a falling accident posterior probability of a target monitoring object based on a first posterior probability, a second posterior probability and a third posterior probability; determining a falling accident level of the target monitoring object based on a probability value of the falling accident posterior probability, and executing an accident intervention strategy corresponding to the falling accident level. The above scheme can effectively improve the timeliness and accuracy of falling accident monitoring.

Description

Technical Field

[0001] This invention relates to the field of personnel safety monitoring technology, specifically to fall detection and early warning methods, devices, and electronic equipment. Background Technology

[0002] Currently, wearable devices are commonly used for fall detection in hazardous scenarios such as working at heights. Existing technologies mainly rely on a single acceleration threshold method to determine a fall. This involves using a built-in accelerometer to collect acceleration signals of human movement, and when the peak acceleration exceeds a preset threshold, a fall is detected and an alarm is triggered.

[0003] However, the peak acceleration generated by normal human activities (such as running and jumping) can sometimes exceed the acceleration during a fall, while the peak acceleration during a fall can sometimes be lower than a set threshold. Relying solely on acceleration as evidence is insufficient to distinguish between normal movement and a real fall, leading to high false alarm and false negative rates. Secondly, current technologies lack a precise probabilistic assessment of fall events and cannot comprehensively judge based on multi-dimensional information such as human physiological responses and changes in environmental altitude, making it difficult to guarantee detection accuracy in complex construction scenarios. Finally, current technologies typically only provide a single alarm output, lacking a tiered intervention mechanism for different probability levels. This results in abnormal states, potential hazards, and confirmed falls not being handled differently, leading to insufficient timeliness and accuracy in rescue responses.

[0004] Therefore, there is an urgent need for a fall detection and early warning method that can integrate multidimensional evidence for probabilistic assessment, reduce false alarm and false negative rates, and implement differentiated intervention strategies based on different probability levels, in order to solve the technical problems of inaccurate judgment based on single evidence and lack of graded intervention in existing technologies. Summary of the Invention

[0005] To address the aforementioned issues, this application provides a fall detection and early warning method to improve the accuracy of fall accident risk monitoring. The technical solution is as follows: In a first aspect, the present invention provides a fall detection and early warning method, comprising: Acquire historical fall accident experimental data and determine the prior probability of fall accidents based on the historical fall accident experimental data; Acquire motion characteristic data, physiological characteristic data, and environmental characteristic data of the target monitoring object; The first posterior probability is obtained based on the first data likelihood of the prior probability of the fall accident and the motion feature data; the second posterior probability is obtained based on the second data likelihood of the prior probability of the fall accident and the physiological feature data; and the third posterior probability is obtained based on the third data likelihood of the prior probability of the fall accident and the environmental feature data. Based on the first posterior probability, the second posterior probability, and the third posterior probability, the posterior probability of a fall accident of the target monitoring object is determined. The fall accident level of the target monitoring object is determined based on the probability value of the posterior probability of the fall accident, and the corresponding accident intervention strategy is executed.

[0006] Combining the first aspect and the above implementation methods, in some possible implementation methods, the first posterior probability is obtained based on the first data likelihood of the prior probability of the fall accident and the motion feature data, including: Extract acceleration characteristic data, peak acceleration, and angular velocity change data of the target monitoring object from motion characteristic data; The data likelihoods corresponding to acceleration feature data, peak acceleration, and angular velocity change data are integrated to obtain the first data likelihood. The first posterior probability is obtained by solving the probability problem based on the first data likelihood and the prior probability of the fall accident.

[0007] Combining the first aspect and the above implementation methods, in some possible implementation methods, a second posterior probability is obtained based on the second data likelihood of the prior probability of the fall accident and the physiological characteristic data, including: Extract the heart rate change rate, heart rate recovery curve, and instantaneous heart rate value of the target monitoring object from physiological characteristic data; The second data likelihood is obtained by integrating the heart rate change rate, heart rate recovery curve, and the data likelihood corresponding to the instantaneous heart rate value. The second posterior probability is obtained by solving the probability problem based on the second data likelihood and the prior probability of the fall accident.

[0008] Combining the first aspect and the above implementation methods, in some possible implementation methods, a third posterior probability is obtained based on the third data likelihood of the prior probability of the fall accident and the environmental feature data, including: Extract altitude descent rate, total altitude descent, and altitude change curvature from environmental feature data; The third data likelihood is obtained by integrating the data likelihoods corresponding to the rate of altitude descent, the total amount of altitude descent, and the curvature of altitude change. The probability is calculated by solving the probability problem based on the third data likelihood and the prior probability of the fall accident, and the third posterior probability is obtained.

[0009] Combining the first aspect and the above implementation methods, in some possible implementation methods, an accident intervention strategy corresponding to the fall accident level is executed, including: If the fall accident level is Level 1, then a preset risk warning message will be pushed to the target monitoring object based on the monitoring device worn by the target monitoring object; If the fall accident is classified as Level 2, then the system will push preset risk consultation information to the target monitoring object based on the monitoring device and obtain feedback information from the target monitoring object based on the monitoring device. If the fall accident level is Level 3, a preset fall warning sound will be broadcast based on the monitoring device, and the feedback operation signal of the target monitoring object will be obtained based on the monitoring device.

[0010] Combining the first aspect and the above implementation methods, in some possible implementation methods, if the fall accident level is level three, then after broadcasting a preset fall warning sound based on the monitoring device and obtaining the feedback operation signal of the target monitored object based on the monitoring device, the following steps are included: Based on the target monitoring object and the corresponding fall accident level, generate assistance and rescue information; Send assistance and rescue information to wearable devices within a preset range of the target monitoring object, and then send the assistance and rescue information to a preset rescue service platform.

[0011] Combining the first aspect and the above implementation methods, in some possible implementation methods, after acquiring the motion characteristic data, physiological characteristic data, and environmental characteristic data of the target monitoring object, the following are included: Based on motion characteristic data, determine the motion state of the target monitoring object; When the motion state indicates the working state, the fatigue state level of the target monitoring object is determined based on physiological characteristic data; Implement fatigue risk intervention strategies corresponding to the fatigue state level.

[0012] Combining the first aspect and the above implementation methods, in some possible implementation methods, after determining the motion state of the target monitoring object based on motion feature data, the following steps are included: When the activity status indicates the working status, the risk level of heatstroke is determined based on physiological and environmental characteristic data. Implement heatstroke risk intervention strategies corresponding to the fatigue level.

[0013] Secondly, the present invention also provides a fall detection and early warning device, comprising: Thirdly, the present invention also provides an electronic device, including a data acquisition unit, a memory, a processor, and a policy executor, wherein, The data acquisition device is used to acquire experimental data from historical fall accidents; and to acquire motion characteristic data, physiological characteristic data, and environmental characteristic data of the target monitoring object. Memory, used to store executable program code; The processor is used to determine the prior probability of a fall accident from experimental data based on historical fall accidents; generate a first posterior probability, a second posterior probability, and a third posterior probability; determine the posterior probability of a fall accident of the target monitoring object based on the first posterior probability, the second posterior probability, and the third posterior probability; and determine the fall accident level of the target monitoring object based on the probability value of the posterior probability of the fall accident. The strategy executor is used to execute the accident intervention strategy corresponding to the fall accident level.

[0014] The beneficial effects of this invention are as follows: By acquiring motion characteristic data, physiological characteristic data, and environmental characteristic data of the target monitoring object, and combining this with the prior probability of a fall accident determined from historical fall accident experimental data, the first data likelihood of the motion characteristic data, the second data likelihood of the physiological characteristic data, and the third data likelihood of the environmental characteristic data are calculated respectively. This yields the first posterior probability, the second posterior probability, and the third posterior probability, achieving a probabilistic fusion evaluation of multi-dimensional evidence. This avoids the deficiency of relying solely on single evidence, which cannot accurately distinguish between normal motion and a real fall, reducing false alarm and false negative rates, and improving the accuracy of fall detection. Furthermore, by determining the posterior probability of a fall accident for the target monitoring object based on the first, second, and third posterior probabilities, and determining the fall accident level based on the probability value of the fall accident posterior probability, and executing the corresponding accident intervention strategy, this invention avoids the deficiency of single alarm output in existing technologies. It achieves differentiated responses and precise interventions for different probability levels, improving the timeliness of rescue responses. Attached Figure Description To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0015] Figure 1 This is a flowchart illustrating the fall detection and early warning method provided by the present invention; Figure 2 For the present invention Figure 1 A schematic diagram of an embodiment of S103; Figure 3 For the present invention Figure 1 A schematic diagram of an embodiment of S103; Figure 4 For the present invention Figure 1 A schematic diagram of an embodiment of S103; Figure 5 For the present invention Figure 1 A schematic diagram of an embodiment of S102; Figure 6 This is a schematic diagram of the structure of the fall detection and early warning device provided by the present invention; Figure 7 This is a schematic diagram of the structure of the electronic device provided by the present invention. Detailed Implementation

[0016] 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 a part of the embodiments of the present invention, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0017] In the description of the embodiments of the present invention, unless otherwise stated, "multiple" means two or more. "And / or" describes the relationship between related objects, indicating that there can be three relationships. For example, A and / or B can represent three situations: A exists alone, A and B exist simultaneously, and B exists alone.

[0018] The terms "first," "second," etc., used in the embodiments of this invention are for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of technical features indicated. Therefore, a technical feature defined with "first" or "second" may explicitly or implicitly include at least one of that feature.

[0019] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of the invention. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0020] It should be noted that the execution architecture of the fall detection and early warning method provided in this application embodiment is based on a three-tier architecture system consisting of a smart wearable device, an edge computing gateway, and a cloud-based health management platform. The smart wearable device is worn on the wrist of the construction worker and has a built-in main control chip and multi-dimensional physiological parameter sensors, including but not limited to a heart rate sensor, body temperature sensor, triaxial accelerometer, barometric altimeter, and ambient temperature and humidity sensor, as well as an alarm actuator, including but not limited to a linear vibration motor, RGB three-color LED, and speaker. It is responsible for the synchronous acquisition of multi-dimensional physiological parameters, local feature extraction, lightweight neural network model inference, real-time fall risk assessment, and local alarm triggering.

[0021] Edge computing gateways are deployed at the boundaries of work and living areas. Each gateway covers a preset radius and has a built-in computing module. It collects data from nearby smart wearable devices via wireless communication, acting as a relay between smart wearable devices and the cloud. It performs complex reasoning tasks such as HRV analysis, fatigue assessment, and anomaly detection, and is responsible for forwarding and broadcasting alarm information in multi-level coordinated rescue operations.

[0022] The cloud-based health management platform is deployed in the enterprise data center and connected to the edge computing gateway via a mobile network. It is responsible for aggregating the health data of all construction personnel in the project and provides value-added functions such as health record management, trend analysis, report export, and dynamic configuration and distribution of alarm thresholds. It also supports updating the better model trained in the cloud to the edge computing gateway and smart wearable devices via OTA. Furthermore, the platform is integrated with the BIM system and personnel management system to achieve data linkage, thus forming a complete method execution architecture from terminal data collection and simple inference, complex calculation and alarm relay at the edge, to cloud data aggregation and global management.

[0023] This application provides a fall detection and early warning method, wherein the subject executing the fall detection and early warning method is a fall detection and early warning device or an electronic device equipped with a fall detection and early warning device. The following is a detailed description; it should be noted that the order of description of the following embodiments is not intended to limit the preferred order of the embodiments.

[0024] Please see Figure 1 , Figure 1 This is a flowchart illustrating the fall detection and early warning method provided in an embodiment of this application. Figure 1 As shown, the method in this application embodiment may include the following steps S101-S105: S101, Obtain historical fall accident experimental data, and determine the prior probability of fall accidents based on the historical fall accident experimental data.

[0025] In this embodiment, historical fall accident experimental data refers to the number of fall events recorded in high-altitude work scenarios during historical observation periods, along with corresponding statistical sample data of normal operations. It also includes fall characteristic data generated through simulation platforms. The prior probability of a fall accident refers to the initial probability of a fall event occurring based on historical statistical patterns before obtaining real-time observational evidence of the current target monitoring object.

[0026] In one feasible implementation, the total number of observation days and the total number of fall events within the historical observation period are statistically analyzed, and the ratio of the number of fall events to the total number of observation days is calculated as the prior probability of the fall accident.

[0027] For example, based on historical data statistics, the average daily probability of a fall for workers working at height is approximately [missing information]. That is, one fall occurs for every 100,000 people-days, and this statistical ratio is taken as the prior probability value of a fall accident.

[0028] S102, acquire motion characteristic data, physiological characteristic data and environmental characteristic data of the target monitoring object.

[0029] In this embodiment, the target monitoring object refers to construction workers wearing wearable devices, such as smart wearable devices. Motion characteristic data refers to acceleration and angular velocity data reflecting human motion state and posture collected by the smart wearable device via a three-axis accelerometer and gyroscope. Physiological characteristic data refers to heart rate, blood oxygen, and body temperature data reflecting human physiological state collected by the heart rate sensor and body temperature sensor of the smart wearable device. Environmental characteristic data refers to data reflecting air pressure, altitude, temperature, and humidity of the working environment collected by the barometric altimeter and ambient temperature and humidity sensor of the smart wearable device.

[0030] In one feasible implementation, a multi-dimensional physiological parameter sensor built into a wearable device worn by the target being monitored is used to acquire motion characteristic data, physiological characteristic data, and environmental characteristic data of the target being monitored by sampling at a unified clock using a time synchronization mechanism. For example, triaxial acceleration and angular velocity are collected as motion feature data by the BMI160 chip built into the wearable device, heart rate and blood oxygen are collected as physiological feature data by the MAX30102 chip of the wearable device, and atmospheric pressure and relative altitude change are collected as environmental feature data by the 1BMP280 chip of the wearable device. All sensor data are accompanied by a timestamp with an accuracy of 1 millisecond to ensure the time consistency of multi-source data fusion calculation.

[0031] S103, based on the first data likelihood of the prior probability of the fall accident and the motion feature data, a first posterior probability is obtained; based on the second data likelihood of the prior probability of the fall accident and the physiological feature data, a second posterior probability is obtained; based on the third data likelihood of the prior probability of the fall accident and the environmental feature data, a third posterior probability is obtained.

[0032] In this embodiment, the first data likelihood refers to the likelihood of observing motion characteristic evidence at the time of the fall. The first posterior probability refers to the posterior probability of the fall event given motion characteristic evidence. The second data likelihood refers to the likelihood of observing physiological characteristic evidence at the time of the fall. The second posterior probability refers to the posterior probability of the fall event given physiological characteristic evidence. The third data likelihood refers to the likelihood of observing environmental characteristic evidence at the time of the fall. The third posterior probability refers to the posterior probability of the fall event given environmental characteristic evidence.

[0033] In one feasible implementation, each source of evidence independently calculates its own likelihood, and then combines this likelihood with the prior probability of the fall accident to calculate the posterior probability of each piece of evidence. For example, the Bayes' theorem formula in step S103 above takes the following form:

[0034] in, For given evidence The posterior probability of a fall event. Evidence observed at the time of the fall likelihood Let be the prior probability of the fall event. As evidence The marginal probability. Motion feature data, physiological feature data, and environmental feature data are used as evidence, respectively. By inputting the above formula, the first data likelihood, the second data likelihood, and the third data likelihood are calculated, and the first posterior probability, the second posterior probability, and the third posterior probability are further solved.

[0035] S104. Based on the first posterior probability, the second posterior probability, and the third posterior probability, determine the posterior probability of the fall accident of the target monitoring object.

[0036] In the embodiments of this application, the posterior probability of a fall accident refers to the probability of a fall event occurring that is finally calculated after fusing evidence of motion characteristics, physiological characteristics, and environmental characteristics.

[0037] In one feasible implementation, the multi-source posterior probabilities are fused using a multiplication rule to obtain the final posterior probability of the fall accident.

[0038] For example, the specific form of the above multi-source posterior probability fusion formula is as follows:

[0039] in, This represents the posterior probability of the fused fall accident. This is the first posterior probability. This is the second posterior probability. This is the third posterior probability.

[0040] By using the above formula, the posterior probabilities of the three dimensions are jointly calculated to determine the posterior probability of a fall accident of the target monitoring object.

[0041] S105, determine the fall accident level of the target monitoring object based on the probability value of the posterior probability of the fall accident, and execute the accident intervention strategy corresponding to the fall accident level.

[0042] In this embodiment, the fall accident level refers to the severity level of an alarm, classified according to the probability value of the posterior probability of the fall accident. The accident intervention strategy refers to the response measures taken for different fall accident levels, including alerts, confirmations, and coordinated rescue efforts.

[0043] In one feasible implementation, the calculated posterior probability of the fall accident is compared with a preset probability threshold, and the fall accident level is determined based on the comparison result, and the corresponding accident intervention strategy is executed.

[0044] For example, the probability threshold is set to 0.85. When the posterior probability of a fall accident exceeds 0.85, it is confirmed as a level 3 alarm, and a fall alarm is immediately triggered and the rescue process is initiated. When the posterior probability of a fall accident does not exceed 0.85 but there is an abnormal state, it is confirmed as a level 1 or level 2 alarm, and the corresponding level of reminder or confirmation strategy is executed.

[0045] It should be noted that the personal data of the target monitoring object involved in the above embodiments includes motion characteristic data, physiological characteristic data, and environmental characteristic data. Before performing the above-mentioned data acquisition, processing, and transmission steps, it must be done with the full knowledge and authorization of the target monitoring object. That is, only after obtaining explicit authorization from the target monitoring object for data collection and use can the corresponding sensors built into the wearable device be triggered to collect data and perform subsequent analysis. At the same time, in order to ensure the privacy and data integrity of the target monitoring object, all types of personal data collected are encrypted using preset encryption algorithms in the transmission link from the wearable device, edge computing gateway to cloud health management platform, and data anonymization is performed during storage to ensure that private data does not leave the site, prevent the risk of data leakage, and meet the compliance requirements for data security and privacy protection.

[0046] In summary, this application acquires motion characteristic data, physiological characteristic data, and environmental characteristic data of the target monitoring object, and combines this with the prior probability of a fall accident determined from historical fall accident experimental data. It then calculates the first data likelihood of the motion characteristic data, the second data likelihood of the physiological characteristic data, and the third data likelihood of the environmental characteristic data, thereby obtaining the first, second, and third posterior probabilities. This achieves a probabilistic fusion assessment of multi-dimensional evidence, avoiding the deficiency of relying on a single piece of evidence to accurately distinguish between normal motion and a real fall, reducing false alarm and false negative rates, and improving the accuracy of fall detection. Furthermore, based on the first, second, and third posterior probabilities, the posterior probability of a fall accident for the target monitoring object is determined, and the fall accident level is determined based on the probability value of the fall accident posterior probability. The corresponding accident intervention strategy is then executed, avoiding the deficiency of single alarm output in existing technologies. This achieves differentiated responses and precise interventions for different probability levels, improving the timeliness of rescue responses.

[0047] Please see Figure 2 , Figure 2 This is a flowchart illustrating the fall detection and early warning method provided in an embodiment of this application. Figure 2 As shown, the method in this application embodiment may include the following steps S201-S203: S201, extract the acceleration characteristic data, peak acceleration, and angular velocity change data of the target monitoring object from the motion characteristic data.

[0048] In this embodiment, acceleration characteristic data refers to data reflecting the acceleration change characteristics during the free fall phase. Peak acceleration refers to data reflecting the maximum amplitude of acceleration during the impact phase. Angular velocity change data refers to data reflecting the angular velocity change characteristics during the attitude rollover phase.

[0049] In one feasible implementation, acceleration feature data, peak acceleration, and angular velocity change data of the target monitoring object are extracted from motion feature data collected by the accelerometer and gyroscope built into the wearable device. Specifically, the acceleration feature data is extracted as the features of the free fall phase where the detected acceleration is less than 0.5g and the duration is between 0.5 and 2 seconds; the peak acceleration is extracted as the features of the impact phase where the detected acceleration is greater than 3g; and the angular velocity change data is extracted as the features of the attitude roll phase where the detected angular velocity is greater than 2 rad / s and the duration is 0.5 seconds.

[0050] S202, integrate the data likelihoods corresponding to the acceleration characteristic data, peak acceleration data, and angular velocity change data to obtain the first data likelihood.

[0051] In one feasible implementation, the probability of acceleration characteristic data, peak acceleration, and angular velocity change data under conditions of falling and not falling is determined based on historical fall accident experimental data. The data likelihood corresponding to each characteristic is integrated to obtain the first data likelihood.

[0052] For example, the specific form of the above acceleration evidence likelihood ratio formula is as follows:

[0053] in, For acceleration evidence likelihood ratio, H1 represents a fall event, H0 represents no fall event, A1 represents an observed event that detected the free fall phase, A2 represents an observed event that detected the impact peak, and A3 represents an observed event that detected attitude rollover.

[0054] Assuming the three events are independent, the joint likelihood of the fall occurring is:

[0055] Joint likelihood when no fall occurs:

[0056] The first data likelihood is obtained by multiplying and integrating the likelihoods of each feature.

[0057] S203, based on the first data likelihood and the prior probability of the fall accident, the probability is solved to obtain the first posterior probability.

[0058] In one feasible implementation, the first data likelihood and the prior probability of the fall accident are substituted into Bayes' theorem to obtain the first posterior probability.

[0059] For example, the specific form of the above posterior probability formula is as follows:

[0060] in, The first posterior probability based on acceleration evidence. The likelihood ratio corresponding to the first data likelihood. The prior probability of the fall accident is represented by the logit function, which represents the log-odds transformation. The first posterior probability is obtained by calculating the sum of the log-likelihood ratio and the log-odds of the prior probability, and then performing the inverse operation using the sigmoid function.

[0061] In summary, this application extracts multidimensional motion features from motion feature data and integrates them into acceleration evidence likelihood. By combining prior probability to solve for the first posterior probability, it achieves a quantitative assessment of the three typical kinematic stages during the fall: free fall, impact, and attitude rollover. This provides a reliable kinematic evidence basis for subsequent multi-source fusion.

[0062] Please see Figure 3 , Figure 3 This is a flowchart illustrating the fall detection and early warning method provided in an embodiment of this application. Figure 3 As shown, the method in this application embodiment may include the following steps S301-S303: S301 extracts the heart rate change rate, heart rate recovery curve, and instantaneous heart rate value of the target monitoring object from physiological characteristic data.

[0063] In this embodiment, the heart rate change rate refers to the rate of change of heart rate over a preset time period before the fall. The heart rate recovery curve refers to the trend curve of heart rate change over time over a preset time period after the fall. The instantaneous heart rate value refers to the heart rate measurement value at the moment the fall occurs.

[0064] In one feasible implementation, the heart rate variability rate, heart rate recovery curve, and instantaneous heart rate value of the target monitored object are extracted from the physiological characteristic data collected by the heart rate sensor built into the wearable device. Specifically, the heart rate variability rate is extracted as the rate of change of heart rate in the 10 seconds before the fall, the instantaneous heart rate value is extracted as the instantaneous heart rate value at the time of the fall, and the heart rate recovery curve is extracted as the heart rate recovery curve within 2 minutes after the fall. The above characteristic data are extracted by utilizing the human body's stress response to fear and pain.

[0065] S302 integrates the data likelihood corresponding to the heart rate change rate, heart rate recovery curve, and instantaneous heart rate value to obtain the second data likelihood.

[0066] In one feasible implementation, based on historical fall accident experimental data and human stress response characteristics, the likelihood of heart rate change rate, heart rate recovery curve, and instantaneous heart rate value at the time of fall is calibrated. The data likelihoods corresponding to the above three characteristics are multiplied and integrated according to the conditional independence assumption to obtain a second data likelihood, which is used to characterize the likelihood of observing the current heart rate data under physiological characteristic evidence.

[0067] S303, based on the second data likelihood and the prior probability of the fall accident, the probability is solved to obtain the second posterior probability.

[0068] In one feasible implementation, the second data likelihood and the prior probability of the fall accident are substituted into Bayes' theorem to solve for the probability, thus obtaining the second posterior probability. The calculation process is consistent with the acceleration evidence solution process, which involves adding the log-likelihood ratio of the heart rate evidence to the log-probability of the prior probability and then converting it into a probability value, thereby obtaining the second posterior probability reflecting the association between heart rate abnormalities and the fall event.

[0069] In summary, this application utilizes the autonomic nervous system stress response of the human body during a fall by extracting heart rate-related stress response features and calculating the second posterior probability of physiological feature evidence. This effectively distinguishes between acceleration fluctuations caused by normal movement and the physiological impact of a real fall, further reducing the false alarm rate.

[0070] Please see Figure 4 , Figure 4 This is a flowchart illustrating the fall detection and early warning method provided in an embodiment of this application. Figure 4 As shown, the method in this application embodiment may include the following steps S401-S403: S401 extracts the altitude descent rate, total altitude descent, and altitude change curvature from environmental feature data.

[0071] In this embodiment, the rate of altitude descent refers to the change in relative altitude per unit time. The total altitude descent refers to the cumulative value of the relative altitude descent from the start to the end of detection. The curvature of altitude change refers to the degree of curvature of the altitude change curve over time.

[0072] In one feasible implementation, the altitude descent rate, total altitude descent, and altitude change curvature are extracted from environmental feature data collected by the built-in barometric altimeter of the wearable device. Specifically, the altitude descent rate is extracted as a rate feature greater than 2 m / s and lasting for more than 1 second; the total altitude descent is extracted as a cumulative descent feature greater than 3 m; and the altitude change curvature is extracted as a curvature feature with an approximately parabolic shape. The above altitude change features are accurately extracted using the 0.1-meter resolution of the barometric altimeter.

[0073] S402 integrates the data likelihoods corresponding to the altitude descent rate, the total altitude descent, and the altitude change curvature to obtain the third data likelihood.

[0074] In one feasible implementation, the likelihood of the rate of altitude descent, the total amount of altitude descent, and the curvature of altitude change at the time of the fall is calibrated based on experimental data from historical fall accidents. The data likelihoods corresponding to the above three features are multiplied and integrated according to the conditional independence assumption to obtain a third data likelihood, which is used to characterize the likelihood of observing the current pressure altitude change under environmental altitude feature evidence.

[0075] S403, based on the third data likelihood and the prior probability of the fall accident, the probability is solved to obtain the third posterior probability.

[0076] In one feasible implementation, the third data likelihood and the prior probability of the fall accident are substituted into Bayes' theorem to solve for the probability, thus obtaining the third posterior probability. The calculation process also uses the method of adding the log-likelihood ratio to the log-probability of the prior probability and then transforming it to obtain the third posterior probability reflecting the association between the abnormal descent in altitude and the fall event.

[0077] In summary, this application utilizes the significant difference in height changes between falls from height and falls on flat ground by extracting barometric altitude change features and calculating the third posterior probability of environmental feature evidence. This provides decisive evidence for fall detection in the spatial dimension, significantly improving detection accuracy and effectively distinguishing between falls from height and falls on flat ground.

[0078] In one feasible implementation, the execution steps include implementing an accident intervention strategy corresponding to the fall accident level, and further implementing specific measures.

[0079] If the fall accident is classified as Level 1, then a preset risk warning message will be pushed to the target monitoring object based on the monitoring device worn by the target monitoring object.

[0080] In this embodiment, the monitoring device refers to a smart bracelet terminal worn on the wrist of construction workers and integrating multi-dimensional physiological parameter sensors and alarm actuators. The first level refers to a yellow alarm level where the abnormal state detection has not reached the fall determination threshold, but workers still need to be alerted to safety.

[0081] In one feasible implementation, when the fall accident level of the target monitoring object is determined to be Level 1 based on the probability value of the posterior probability of the fall accident, the alarm actuator of the monitoring wearable device performs a low-intensity alarm action. Specifically, it emits yellow light through an RGB three-color LED, the linear vibration motor performs low-frequency vibration at the first level of vibration intensity, and pushes preset risk warning information on the screen of the monitoring wearable device to remind the target monitoring object to pay attention to the current abnormal state, thereby achieving a safety warning without interfering with normal operations.

[0082] If the fall accident is classified as Level 2, then the system will push preset risk consultation information to the target monitoring object based on the monitoring device and obtain feedback information from the target monitoring object based on the monitoring device. In one feasible implementation, when the fall accident level of the target monitored object is determined to be Level 2 based on the probability value of the posterior probability of the fall accident, the alarm actuator of the monitoring wearable device performs a medium-intensity alarm action. Specifically, it emits orange light through an RGB three-color LED, the linear vibration motor performs medium-frequency vibration at the second vibration intensity, and pushes preset risk consultation information on the screen of the monitoring wearable device, indicating that there may be danger to the target monitored object and asking if it needs help. At the same time, it obtains the feedback information of the target monitored object in response to the risk consultation information based on the buttons on the monitoring wearable device. The feedback information includes operation instructions to confirm safety or request help.

[0083] If the fall accident level is Level 3, a preset fall warning sound will be broadcast based on the monitoring device, and the feedback operation signal of the target monitoring object will be obtained based on the monitoring device.

[0084] In this embodiment of the application, the third level refers to the red alarm level that confirms a fall accident and immediately initiates the rescue process.

[0085] In one feasible implementation, when the fall accident level of the target monitored object is determined to be Level 3 based on the probability value of the posterior probability of the fall accident, the alarm actuator of the monitoring device executes a high-intensity alarm action. Specifically, this is achieved by emitting red light through an RGB three-color LED, a linear vibration motor performing high-frequency strong vibration at the third vibration intensity level, and a speaker broadcasting a preset fall warning tone at a maximum volume of 90dB. The screen of the monitoring device displays "Fall detection! Automatic rescue call in 30 seconds" and starts a 30-second countdown. During the 30-second countdown, feedback operation signals from the target monitored object are obtained based on the buttons on the monitoring device. These feedback operation signals include "No problem," "Help," and "Wait." If a "No problem" signal is obtained, the system records this event as a false alarm and automatically adjusts subsequent detection thresholds to prevent frequent false alarms. If a "Help" signal is obtained, a Level 3 alarm is immediately triggered and the rescue process is initiated. If a "Wait" signal is obtained, the system continues to monitor physiological parameters and posture. If the condition does not improve after extending the confirmation time by 60 seconds, the alarm level is automatically upgraded to Level 3 and the rescue process is initiated.

[0086] In summary, this application achieves graded response to abnormal states, potential hazards, and confirmed fall accidents by dividing alarm levels into multiple levels based on the posterior probability of fall accidents and implementing differentiated accident intervention strategies, thus avoiding the waste of rescue resources or insufficient response caused by a single alarm mode.

[0087] In one feasible implementation, if the fall accident level is level three, the specific execution steps are performed after broadcasting a preset fall warning sound based on the monitoring device and obtaining the feedback operation signal of the target monitoring object based on the monitoring device.

[0088] Based on the target monitoring object and the corresponding fall accident level, generate assistance and rescue information; Send assistance and rescue information to wearable devices within a preset range of the target monitoring object, and then send the assistance and rescue information to a preset rescue service platform.

[0089] In one feasible implementation, after a fall alarm is triggered, the monitoring device vibrates and emits a warning sound. The screen displays a fall detection prompt and begins a 30-second countdown. Feedback signals from the target monitoring object are obtained. If the target monitoring object does not cancel the alarm within 30 seconds, assistance rescue information is generated based on the target monitoring object and the third level of response. This assistance rescue information includes the location and vital signs of the injured person. Subsequently, a multi-level linkage rescue is initiated. In the first-level response phase, the edge computing gateway broadcasts the assistance rescue information to the monitoring devices worn by other construction workers within a 30-meter range. Upon receiving the broadcast, the monitoring devices vibrate and display the location and distance of the injured person. In the second-level response phase, the edge computing gateway pushes the assistance rescue information to the team leader's mobile APP. In the third-level response phase, the assistance rescue information is pushed to the project department's safety supervisor and the emergency room's rescue service platform. In the fourth-level response phase, the assistance rescue information is automatically upgraded and pushed to the company headquarters and external medical institutions' rescue service platforms.

[0090] In summary, this application generates assistance information after confirming a fall accident and sends it to nearby personnel wearing devices and multi-level rescue service platforms, enabling multi-level coordinated rescue efforts from people at the scene to managers at all levels. This ensures the rapid mobilization of rescue resources within the golden rescue time and improves the timeliness of the rescue response.

[0091] Please see Figure 5 , Figure 5 This is a flowchart illustrating the fall detection and early warning method provided in an embodiment of this application. Figure 5 As shown, the method in this application embodiment may include the following steps S501-S505: S501 determines the motion state of the target monitoring object based on motion feature data.

[0092] In this embodiment of the application, motion state refers to the current activity status of the target monitoring object as evaluated based on motion feature data, including working state and resting state.

[0093] In one feasible implementation, based on motion characteristic data collected by a wearable device, such as a smart bracelet, and its built-in accelerometer, gait analysis and motion intensity assessment are performed on the target monitored object to determine its motion state. When the acceleration data indicates that the target monitored object is walking, bending over, or carrying heavy physical loads, the motion state is determined to be a working state; when the acceleration data indicates that the target monitored object is stationary or asleep, the motion state is determined to be a resting state.

[0094] S502, when the motion state indicates the working state, determines the fatigue state level of the target monitoring object based on physiological characteristic data.

[0095] In this embodiment of the application, the fatigue level refers to the fatigue level classification of the target monitoring object based on the assessment of heart rate variability index.

[0096] In one feasible implementation, when the motion state indicates the working state, PPG signals are extracted from physiological characteristic data for HRV analysis. HRV time-domain, frequency-domain, and nonlinear indices are calculated. A fatigue index is calculated based on the offset ratio of each index relative to the individual baseline. The fatigue state level of the target monitoring object is determined based on the fatigue index. For example, the specific form of the fatigue index formula in step S502 above is:

[0097] in, The fatigue index. This is the standardized offset of the SDNN metric. The standardized offset of the RMSSD metric. The normalized offset of the LF / HF ratio. SD1 represents the standardized offset of the RR interval. SDNN is the standard deviation of the RR interval, RMSSD is the root mean square of the difference between adjacent RR intervals, LF / HF is the ratio of low-frequency power to high-frequency power, and SD1 is the minor axis of the Poincaré scatter plot. By calculating the standardized offset of each current indicator relative to the previous 7-day moving average baseline and substituting it into the formula, the fatigue index is obtained, and the fatigue index is mapped to five fatigue state levels: normal (0 to 20), mild fatigue (21 to 40), moderate fatigue (41 to 60), severe fatigue (61 to 80), and extreme fatigue (81 to 100).

[0098] S503, implement fatigue risk intervention strategies corresponding to the fatigue state level.

[0099] In the embodiments of this application, fatigue risk intervention strategy refers to reminders or mandatory rest measures taken for different fatigue levels.

[0100] In one feasible implementation, a corresponding fatigue risk intervention strategy is executed based on the calculated fatigue level. When the fatigue level is moderate or above, an alarm is triggered by the monitoring device, suggesting that the monitored person stop working and rest, thereby preventing construction accidents caused by excessive fatigue.

[0101] S504, when the motion state indicates the working state, determines the heatstroke risk level based on physiological and environmental characteristic data.

[0102] In this embodiment of the application, the heatstroke risk level refers to the risk classification of heatstroke occurrence derived from a comprehensive assessment of environmental heat stress and individual heat stress.

[0103] In one feasible implementation, when the activity state indicates the working state, the wet-bulb temperature (WBGT) is calculated based on environmental characteristic data, and body temperature and heart rate are obtained based on physiological characteristic data. Combined with a labor intensity assessment, the risk level of heatstroke is determined. For example, the specific form of the outdoor environmental WBGT calculation formula in step S504 above is:

[0104] Where WBGT is the wet-bulb black-bulb temperature. Natural wet-bulb temperature, For the temperature of the black ball, The value is the dry bulb temperature. The WBGT value is calculated by substituting the parameters obtained from the wristband's built-in sensor and the gateway's environmental sensor into the formula.

[0105] S505, implement the heatstroke risk intervention strategy corresponding to the fatigue level.

[0106] In this application embodiment, the heatstroke risk intervention strategy refers to control measures such as hydration, rest, or work stoppage taken for different heatstroke risk levels.

[0107] In one feasible implementation, a corresponding heatstroke risk intervention strategy is executed based on the determined heatstroke risk level. When the level is "Caution," the monitoring device reminds the target to replenish water in time and take a break every 30 minutes; when the level is "Warning," the working time is forcibly shortened, the frequency of rest is increased, and mandatory water intake for electrolyte replenishment is implemented; when the level is "Danger," the monitoring device issues a stop-outdoor work instruction, requiring the target to move to a shady place or an air-conditioned room, thereby effectively preventing the occurrence of heatstroke.

[0108] In summary, this application comprehensively assesses fatigue level and heatstroke risk level by combining physiological and environmental characteristic data under working conditions, and implements corresponding intervention strategies. This solves the problem of single parameters in existing monitoring equipment, realizes comprehensive early warning and timely intervention for worker fatigue and heatstroke risk, and protects the life, health and safety of construction workers.

[0109] The following will combine Figure 6 This application provides a detailed description of the fall detection and early warning device provided in its embodiments. It should be noted that... Figure 6 The fall detection and early warning device in the present application is used to perform the functions described herein. Figures 1-5 The methods shown in the embodiments are for illustrative purposes only, illustrating the parts relevant to the embodiments of this application. For specific technical details not disclosed, please refer to this application. Figures 1-5In the embodiment shown, the fall detection and early warning device 600 may include a first data acquisition unit 601, a second data acquisition unit 602, a data processing unit 603, a posterior probability generation unit 604, and an accident strategy execution unit 605, as detailed below: The first data acquisition unit 601 is used to acquire historical fall accident experimental data and determine the prior probability of a fall accident based on the historical fall accident experimental data. The second data acquisition unit 602 is used to acquire motion characteristic data, physiological characteristic data and environmental characteristic data of the target monitoring object; The data processing unit 603 is used to obtain a first posterior probability based on the first data likelihood of the prior probability of the fall accident and the motion feature data, to obtain a second posterior probability based on the second data likelihood of the prior probability of the fall accident and the physiological feature data, and to obtain a third posterior probability based on the third data likelihood of the prior probability of the fall accident and the environmental feature data. The posterior probability generation unit 604 is used to determine the posterior probability of the fall accident of the target monitoring object based on the first posterior probability, the second posterior probability and the third posterior probability. The accident strategy execution unit 605 is used to determine the fall accident level of the target monitoring object based on the probability value of the posterior probability of the fall accident, and execute the accident intervention strategy corresponding to the fall accident level.

[0110] The fall detection and early warning device 600 provided in the above embodiments can realize the technical solutions described in the above fall detection and early warning method embodiments. The specific implementation principles of each module or unit can be found in the corresponding content in the above fall detection and early warning method embodiments, and will not be repeated here.

[0111] like Figure 7 As shown, the present invention also provides an electronic device 700. The electronic device 700 includes a processor 701, a memory 702, a display 703, and a data acquisition device 704. Figure 7 Only some components of the electronic device 700 are shown, but it should be understood that it is not required to implement all the components shown, and more or fewer components may be implemented instead.

[0112] In some embodiments, processor 701 may be a central processing unit (CPU), microprocessor, or other data processing chip, used to run program code stored in memory 702 or process data, such as the fall detection and early warning method of the present invention.

[0113] In some embodiments, processor 701 may be a single server or a group of servers. The server group may be centralized or distributed. In some embodiments, processor 701 may be local or remote. In some embodiments, processor 701 may be implemented on a cloud platform. In one embodiment, the cloud platform may include a private cloud, public cloud, hybrid cloud, community cloud, distributed cloud, intranet, multi-cloud, etc., or any combination thereof.

[0114] In some embodiments, memory 702 may be an internal storage unit of electronic device 700, such as a hard disk or memory of electronic device 700. In other embodiments, memory 702 may also be an external storage device of electronic device 700, such as a plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, etc. equipped on electronic device 700.

[0115] Furthermore, the memory 702 may include both internal storage units of the electronic device 700 and external storage devices. The memory 702 is used to store application software and various types of data installed on the electronic device 700.

[0116] In some embodiments, display 703 may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, or an OLED (Organic Light-Emitting Diode) touchscreen. Display 703 is used to display information from electronic device 700 and to display a visual user interface. Components 701-704 of electronic device 700 communicate with each other via a system bus.

[0117] In one embodiment, when the processor 701 executes the fall detection warning program in the memory 702, the following steps can be implemented: Acquire historical fall accident experimental data and determine the prior probability of fall accidents based on the historical fall accident experimental data; Acquire motion characteristic data, physiological characteristic data, and environmental characteristic data of the target monitoring object; The first posterior probability is obtained based on the first data likelihood of the prior probability of the fall accident and the motion feature data; the second posterior probability is obtained based on the second data likelihood of the prior probability of the fall accident and the physiological feature data; and the third posterior probability is obtained based on the third data likelihood of the prior probability of the fall accident and the environmental feature data. Based on the first posterior probability, the second posterior probability, and the third posterior probability, the posterior probability of a fall accident of the target monitoring object is determined. The fall accident level of the target monitoring object is determined based on the probability value of the posterior probability of the fall accident, and the corresponding accident intervention strategy is executed.

[0118] It should be understood that when the processor 701 executes the fall detection warning program in the memory 702, in addition to the functions mentioned above, it can also perform other functions, as can be found in the description of the corresponding method embodiments above.

[0119] Furthermore, the embodiments of the present invention do not specifically limit the type of the electronic device 700 mentioned. The electronic device 700 can be a portable electronic device such as a tablet computer, a personal digital assistant (PDA), a wearable device, or a laptop computer. Exemplary embodiments of the portable electronic device include, but are not limited to, portable electronic devices running iOS, Android, Microsoft, or other operating systems. The aforementioned portable electronic device can also be other portable electronic devices, such as a laptop computer with a touch-sensitive surface (e.g., a touch panel). It should also be understood that in some other embodiments of the present invention, the electronic device 700 may not be a portable electronic device, but rather a desktop computer with a touch-sensitive surface (e.g., a touch panel).

[0120] Accordingly, this application also provides a computer-readable storage medium for storing computer-readable programs or instructions. When the programs or instructions are executed by a processor, they can implement the steps or functions of the fall detection and early warning methods provided in the above-described method embodiments.

[0121] Those skilled in the art will understand that all or part of the processes of the methods described in the above embodiments can be implemented by a computer program instructing related hardware (such as a processor, controller, etc.), and the computer program can be stored in a computer-readable storage medium. The computer-readable storage medium may be a disk, optical disk, read-only memory, or random access memory, etc.

[0122] The fall detection and early warning method, device, and electronic equipment provided by the present invention have been described in detail above. Specific examples have been used to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only for the purpose of helping to understand the method and core ideas of the present invention. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of the present invention. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A fall detection and early warning method, characterized in that, include: Acquire historical fall accident experimental data, and determine the prior probability of fall accidents based on the historical fall accident experimental data; Acquire motion characteristic data, physiological characteristic data, and environmental characteristic data of the target monitoring object; A first posterior probability is obtained based on the prior probability of the fall accident and the first data likelihood of the motion feature data; a second posterior probability is obtained based on the prior probability of the fall accident and the second data likelihood of the physiological feature data; and a third posterior probability is obtained based on the prior probability of the fall accident and the third data likelihood of the environmental feature data. Based on the first posterior probability, the second posterior probability, and the third posterior probability, the posterior probability of the fall accident of the target monitoring object is determined; The fall accident level of the target monitoring object is determined based on the probability value of the posterior probability of the fall accident, and the accident intervention strategy corresponding to the fall accident level is executed.

2. The method according to claim 1, characterized in that, The process of obtaining the first posterior probability based on the prior probability of the fall accident and the first data likelihood of the motion feature data includes: Extract the acceleration characteristic data, peak acceleration, and angular velocity change data of the target monitoring object from the motion characteristic data; The data likelihoods corresponding to the acceleration feature data, the peak acceleration, and the angular velocity change data are integrated to obtain the first data likelihood. The first posterior probability is obtained by solving the probability problem based on the first data likelihood and the prior probability of the fall accident.

3. The method according to claim 1, characterized in that, The process of obtaining the second posterior probability based on the prior probability of the fall accident and the second data likelihood of the physiological characteristic data includes: Extract the heart rate change rate, heart rate recovery curve, and instantaneous heart rate value of the target monitoring object from the physiological characteristic data; The second data likelihood is obtained by integrating the heart rate change rate, the heart rate recovery curve, and the data likelihood corresponding to the instantaneous heart rate value. The second posterior probability is obtained by solving the probability problem based on the second data likelihood and the prior probability of the fall accident.

4. The method according to claim 1, characterized in that, The process of obtaining the third posterior probability based on the third data likelihood of the prior probability of the fall accident and the environmental feature data includes: Extract the altitude descent rate, total altitude descent, and altitude change curvature from the environmental feature data; The third data likelihood is obtained by integrating the data likelihood corresponding to the altitude descent rate, the total altitude descent, and the altitude change curvature. The third posterior probability is obtained by solving the probability problem based on the third data likelihood and the prior probability of the fall accident.

5. The method according to claim 1, characterized in that, The execution of the accident intervention strategy corresponding to the fall accident level includes: If the fall accident level is Level 1, then a preset risk warning message is pushed to the target monitoring object based on the monitoring device worn by the target monitoring object; If the fall accident level is Level 2, then based on the monitoring device, preset risk consultation information is pushed to the target monitoring object, and based on the monitoring device, feedback information from the target monitoring object is obtained; If the fall accident level is Level 3, a preset fall warning sound will be broadcast based on the monitoring device, and the feedback operation signal of the target monitoring object will be obtained based on the monitoring device.

6. The method according to claim 5, characterized in that, If the fall accident level is Level 3, then after broadcasting a preset fall warning sound based on the monitoring device and obtaining the feedback operation signal of the target monitoring object based on the monitoring device, the method further includes: Based on the target monitoring object and the corresponding fall accident level, generate assistance and rescue information; The assistance and rescue information is sent to wearable devices within a preset range of the target monitoring object, and the assistance and rescue information is sent to a preset rescue service platform.

7. The method according to claim 1, characterized in that, After acquiring the motion characteristic data, physiological characteristic data, and environmental characteristic data of the target monitoring object, the method further includes: Based on the motion feature data, the motion state of the target monitoring object is determined; When the motion state indicates the working state, the fatigue state level of the target monitoring object is determined based on the physiological characteristic data; Implement the fatigue risk intervention strategy corresponding to the fatigue state level.

8. The method according to claim 7, characterized in that, After determining the motion state of the target monitoring object based on the motion feature data, the method further includes: When the motion state indicates the working state, the heatstroke risk level is determined based on the physiological characteristic data and the environmental characteristic data; Implement the heatstroke risk intervention strategy corresponding to the fatigue state level.

9. A fall detection and early warning device, characterized in that, The device includes: The first data acquisition unit is used to acquire historical fall accident experimental data and determine the prior probability of a fall accident based on the historical fall accident experimental data. The second data acquisition unit is used to acquire motion characteristic data, physiological characteristic data, and environmental characteristic data of the target monitoring object; The data processing unit is configured to obtain a first posterior probability based on the prior probability of the fall accident and the first data likelihood of the motion feature data; obtain a second posterior probability based on the prior probability of the fall accident and the second data likelihood of the physiological feature data; and obtain a third posterior probability based on the prior probability of the fall accident and the third data likelihood of the environmental feature data. The posterior probability generation unit is used to determine the posterior probability of the fall accident of the target monitoring object based on the first posterior probability, the second posterior probability and the third posterior probability. The accident strategy execution unit is used to determine the fall accident level of the target monitoring object based on the probability value of the posterior probability of the fall accident, and execute the accident intervention strategy corresponding to the fall accident level.

10. An electronic device, characterized in that, The electronic device includes a data acquisition unit, a memory, a processor, and a policy executor, wherein, The data acquisition device is used to acquire historical fall accident experimental data; and to acquire motion characteristic data, physiological characteristic data, and environmental characteristic data of the target monitoring object. The memory is used to store executable program code; The processor is configured to determine the prior probability of a fall accident based on the historical fall accident experimental data; generate a first posterior probability, a second posterior probability, and a third posterior probability; determine the posterior probability of a fall accident of the target monitoring object based on the first posterior probability, the second posterior probability, and the third posterior probability; and determine the fall accident level of the target monitoring object based on the probability value of the posterior probability of the fall accident. The strategy executor is used to execute the accident intervention strategy corresponding to the fall accident level.

Images