Old people falling alarm system and method based on multiple sensors

Abstract

The invention discloses an old people tumble alarm system and method based on multiple sensors, and relates to the technical field of intelligent monitoring, and the system collects the body data of old people through an intelligent bracelet and a sensor group, transmits the body data to a cloud platform through a wireless network, carries out the processing, and generates a state data group. The acceleration modulus length, the angular velocity modulus length, the gait cycle and the posture angle of the old people are calculated, so that the motion state and the falling risk are accurately judged, a falling risk assessment coefficient is calculated based on the data, old people falling assessment is performed according to the falling risk assessment coefficient and a preset falling risk threshold value, and an alarm is triggered and a guardian is notified when the risk exceeds the standard. The heart rate variability index and the blood oxygen content are analyzed, a comprehensive health index is obtained through calculation, possible body injury after falling is evaluated, and injury alarm information is generated. The system can accurately monitor the health state of the old people in real time, timely send out early warning and alarm information, and provide comprehensive fall prevention and health monitoring guarantee.

Description

Technical Field

[0001] The present invention relates to the field of intelligent monitoring technology, and in particular to a multi-sensor based elderly fall alarm system and method. Background Art

[0002] As the global aging problem becomes increasingly serious, the rapid growth of the elderly population has made the health management of the elderly a focus of social attention. Falls among the elderly are one of the main causes of their health problems, often accompanied by serious complications such as fractures and internal bleeding, and even life-threatening. With the advancement of science and technology, especially the development of multi-sensor technology, cloud computing and artificial intelligence, intelligent fall alarm systems have gradually become an effective solution. These systems can monitor the physical condition of the elderly in real time, accurately judge fall events, and provide health assessments through data analysis to help timely detect health risks after falls, create a safer living environment for the elderly, reduce the harm caused by falls, and reduce the care burden on families and society.

[0003] Although the existing elderly fall alarm system has made some progress in monitoring falls, there are still some shortcomings. First, most systems can only simply determine whether a fall has occurred, and their analysis dimensions are relatively single, failing to take into account the complex behavior patterns of the elderly, which may lead to false alarms and missed reports. Secondly, although the system can alarm when a fall occurs, it usually only focuses on whether the elderly have fallen, but ignores the injury assessment after the fall, such as potential health problems such as heart problems, fractures or concussions. In addition, the existing systems often do not have real-time health monitoring capabilities, and it is not easy to effectively distinguish whether the elderly have simply fallen or have been injured after falling, and it is not easy to provide personalized response measures based on the specific type of injury. This makes it possible for the elderly to fail to receive timely and appropriate treatment after falling, delaying the time for treatment. Summary of the invention

[0004] In view of the deficiencies in the prior art, the present invention provides a multi-sensor based fall alarm system and method for the elderly, which solves the problems in the above-mentioned background technology.

[0005] To achieve the above objectives, the present invention is implemented through the following technical solutions: a multi-sensor based elderly fall alarm system, including a data acquisition module, a cloud platform processing module, a posture analysis module, a fall detection module and a health analysis module;

[0006] The data acquisition module collects the elderly's physical data in real time through smart devices and sensor groups installed on the elderly's body, builds a cloud processing platform, establishes communication connections with the sensor group and the smart bracelet through a wireless network, and transmits the physical data to the cloud processing platform in real time;

[0007] The cloud platform processing module receives the body data in real time through the cloud processing platform, performs preprocessing, obtains the state data group, and constructs a data repository to store the state data group, the state data group includes a behavior state data group and a health state data group;

[0008] The posture analysis module is used to calculate the acceleration modulus |a|, the angular velocity modulus |w| and the gait period according to the acquired behavior state data group , and then calculate the posture angle θ of the elderly based on the acceleration modulus |a| and the angular velocity modulus |w|;

[0009] The fall detection module constructs a fall detection model through the cloud processing platform, and obtains the acceleration modulus |a|, angular velocity modulus |w|, posture angle θ, gait cycle The pressure yl is imported into the fall detection model to obtain the fall risk assessment coefficient dfx, and then the fall assessment of the elderly is performed with the preset fall risk threshold A;

[0010] The health analysis module is used to extract the heart rate RR and blood oxygen content xy of the health status data group when the elderly person is in a fall state during fall assessment, calculate the heart rate variability index HRV based on the heart rate RR, and build a health status model, import the heart rate variability index HRV and blood oxygen content xy into the health status model, calculate and obtain the comprehensive health index zhj, and perform health assessment on the elderly person with a preset health threshold B.

[0011] Preferably, the data acquisition module includes a habit data collection unit and a processing platform construction unit;

[0012] The habit data collection unit collects the elderly's physical data in real time through a sensor group installed on the elderly's head, the bottom of the insole and the inside of the smart device, and marks the collection timestamp;

[0013] The sensor group includes a pressure sensor, an acceleration sensor, a heart rate sensor, a blood oxygen sensor and a gyroscope, and the smart device is a smart bracelet;

[0014] The processing platform construction unit is used to construct a cloud processing platform, establish a communication connection between the cloud processing platform, the sensor group and the smart bracelet through a wireless network, and transmit the acquired body data to the cloud processing platform in real time.

[0015] Preferably, the cloud platform processing module includes a data processing unit and a data storage unit;

[0016] The data processing unit receives the body data in real time through the cloud processing platform, and performs denoising, filtering and dimensionless processing to obtain a state data group;

[0017] The state data group includes a behavior state data group and a health state data group;

[0018] The behavior state data group includes an acceleration vector a, an angular velocity vector w and a pressure yl;

[0019] The health status data group includes heart rate RR and blood oxygen content xy;

[0020] The data storage unit constructs a data repository through a cloud processing platform, and stores the acquired status data groups in the data repository after time-series integration according to the marked timestamps.

[0021] Preferably, the posture analysis module includes a motion state analysis unit and a posture analysis unit;

[0022] The motion state analysis unit is used to analyze the motion state of the elderly according to the cloud processing platform, detect the acceleration change of the elderly through the accelerometer, judge the motion intensity and direction, and calculate the acceleration vector a=(a x , a y , a z ), calculate and obtain the acceleration modulus |a|, the specific formula is as follows: , capturing the instantaneous drastic acceleration change when falling, where a x 、a y and a z Represent the acceleration values ​​along the x-axis, y-axis and z-axis respectively, the x-axis represents the horizontal front-to-back direction, the y-axis represents the horizontal left-to-right direction, and the z-axis represents the vertical up-down direction;

[0023] The gyroscope is used to capture the angular velocity changes of the elderly's movements, identify sudden pitch and roll movements, and calculate the angular velocity vector w=(w x , w y , w z ), calculate and obtain the angular velocity modulus |w|, the specific formula is as follows: , where w x 、w y and w z Represent the angular velocity values ​​along the x-axis, y-axis and z-axis respectively, the x-axis represents the horizontal front-to-back direction, the y-axis represents the horizontal left-to-right direction, and the z-axis represents the vertical up-down direction;

[0024] The posture analysis unit is used to accurately identify the posture angle θ of the elderly in dynamic and static conditions, and the acceleration vector a=(a x , a y , a z ), calculate the pitch angle θ of the elderly person’s body x and roll angle θ y , the specific algorithm formula is: , , where arctan represents the inverse tangent function;

[0025] The time interval sj of each step is obtained by collecting the pressure of each two times by the pressure sensor, and the gait cycle of the elderly is calculated. , analyze the time it takes for an elderly person to land on one foot and then land on the same foot again. The specific formula is as follows: , where N represents the total number of steps in the monitoring period.

[0026] Preferably, the fall detection module comprises a fall analysis unit and a fall assessment unit;

[0027] The fall analysis unit constructs a fall detection model through a cloud processing platform, and then obtains the acceleration modulus |a|, angular velocity modulus |w|, posture angle θ, and gait cycle The pressure yl is imported into the fall detection model to obtain the fall risk assessment coefficient dfx and quantify the factors causing the elderly to fall;

[0028] The fall risk assessment coefficient dfx is calculated by the following formula;

[0029] ;

[0030] In the formula, e -b represents the exponential decay function, yl d Indicates the current pressure value, yl z It represents the pressure value of the elderly's normal contact with the ground. c1, c2 and c3 represent the weight coefficients of acceleration modulus |a|, angular velocity modulus |w| and posture angle θ respectively, and c1+c2+c3=1. The specific values ​​are set by the user.

[0031] Preferably, the fall assessment unit is used to count the distribution of fall events and normal activities in the characteristic dimensions of acceleration vector a, angular velocity vector w, posture angle θ and pressure yl based on a large number of historical behavior state data groups in the data storage library, and mark the fall events and normal activities, calculate the mean and standard deviation of the risk values ​​of normal activities and fall events by using statistical methods, preset the fall risk threshold A, and perform the fall assessment of the elderly with the obtained fall risk assessment coefficient dfx, analyze the current state of the elderly, and generate corresponding measures. The specific assessment plan is as follows;

[0032] When the fall risk assessment coefficient dfx> the fall risk threshold A, the elderly person is in a fall state, and a fall alarm message is generated. The cloud processing platform transmits the fall alarm message to the guardian user end through the wireless network and triggers a deep assessment.

[0033] When the fall risk assessment coefficient dfx ≤ the fall risk threshold A, the elderly person is in a normal state and maintains normal monitoring.

[0034] Preferably, the health analysis module is used to analyze the damage caused to the elderly by the fall when the elderly is in the fall state according to the fall assessment;

[0035] The health analysis module includes a health analysis unit and a depth analysis unit;

[0036] The health analysis unit is used to analyze the stress response of the elderly's heart after the elderly falls, construct a heart rate variability formula through the heart rate RR collected in real time by the heart rate sensor, and calculate the heart rate variability index HRV;

[0037] The heart rate variability index HRV is calculated by the following formula:

[0038] ;

[0039] Where M represents the total number of sampling points, RR i Represents the time interval from the i-1th heartbeat to the i-th heartbeat, RR i-1 Indicates the time interval from the i-2th heartbeat to the i-1th heartbeat;

[0040] By placing a blood oxygen sensor on the elderly's body, the fluctuation changes of the local blood oxygen content xy in the body can be obtained in real time to analyze the elderly's physical injuries.

[0041] Preferably, the in-depth analysis unit includes a comprehensive health analysis unit and a comprehensive health assessment unit;

[0042] The comprehensive health analysis unit is used to construct a health status model, based on the acquired heart rate variability index HRV and blood oxygen content xy, import the health status model for summary calculation, obtain the comprehensive health index zhj, and comprehensively analyze the physical health status of the elderly after the fall;

[0043] The comprehensive health index zhj is calculated by the following formula:

[0044] ;

[0045] Where, HRV z It represents a healthy heart rate variability value, 90% represents the lower limit of normal blood oxygen content in the human body, and 100% represents the oxygen saturation state of blood oxygen content.

[0046] Preferably, the comprehensive health assessment unit extracts a large amount of historical data from the data repository, obtains the average value and standard deviation of the heart rate variability index HRV of the elderly by statistical methods, and then associates the blood oxygen content xy with the lower limit of 90% of the normal blood oxygen content of the human body according to conventional clinical standards, and imports the health status model, outputs the value to preset the health threshold B, and performs the health assessment of the elderly with the obtained comprehensive health index zhj. The specific assessment scheme is as follows;

[0047] When the comprehensive health index zhj ≥ health threshold B, the elderly person’s fall did not cause physical injury;

[0048] When the comprehensive health index zhj is less than the health threshold B, the elderly person falls and causes physical injury. At this time, an injury alarm message is generated. The cloud processing platform transmits the injury alarm message to the guardian user terminal through the wireless network, reminding the guardian that the elderly person is injured and calling the emergency number immediately.

[0049] A multi-sensor based fall alarm method for the elderly comprises the following steps:

[0050] S1. Collect the elderly's physical data in real time through smart devices and sensor groups installed on the elderly's body, build a cloud processing platform, establish communication connections with the sensor group and smart bracelets through wireless networks, and transmit the physical data to the cloud processing platform in real time;

[0051] S2, the cloud processing platform receives the body data in real time, performs preprocessing, obtains the state data group, and builds a data repository to store the state data group, the state data group includes a behavior state data group and a health state data group;

[0052] S3, calculate the acceleration modulus |a|, angular velocity modulus |w| and gait period based on the acquired behavior state data set , and then calculate the posture angle θ of the elderly based on the acceleration modulus |a| and the angular velocity modulus |w|;

[0053] S4. Build a fall detection model based on the cloud processing platform, and obtain the acceleration modulus |a|, angular velocity modulus |w|, posture angle θ, and gait cycle The pressure yl is imported into the fall detection model to obtain the fall risk assessment coefficient dfx, and then the fall assessment of the elderly is performed with the preset fall risk threshold A;

[0054] S5. When the elderly person is in a falling state during fall assessment, the heart rate RR and blood oxygen content xy of the health status data group are extracted, the heart rate variability index HRV is calculated based on the heart rate RR, and a health status model is constructed. The heart rate variability index HRV and blood oxygen content xy are imported into the health status model, the comprehensive health index zhj is calculated, and the health assessment of the elderly person is performed with the preset health threshold B.

[0055] The present invention provides a multi-sensor based elderly fall alarm system and method, which has the following beneficial effects:

[0056] (1) The system collects the elderly's physical data in real time by installing multiple sensors such as pressure sensors, acceleration sensors, heart rate sensors, blood oxygen sensors and gyroscopes on different parts of the elderly's body, and transmits the data to the cloud processing platform through a wireless network. The cloud processing platform pre-processes the physical data, including denoising, filtering and dimensionless operations, to obtain status data groups to ensure the accuracy and timeliness of the data. These data include behavioral status data groups and health status data groups, which are used to evaluate the physical status of the elderly in real time.

[0057] (2) The system calculates the acceleration modulus |a| by using the acceleration vector a=(ax,ay,az) obtained by the acceleration sensor, and uses the gyroscope to capture the angular velocity changes of the elderly and calculate the angular velocity modulus |w|. The system also uses a pressure sensor to detect the time interval sj of each step and calculate the gait cycle In addition, combined with the acceleration modulus |a|, angular velocity modulus |w|, posture angle θ, gait period The fall detection model built by the cloud processing platform calculates the fall risk assessment coefficient dfx, and finally compares it with the preset fall risk threshold A to assess whether there is a fall risk. When the fall risk assessment coefficient dfx exceeds the fall risk threshold A, the system will generate a fall alarm message and immediately notify the guardian to handle it.

[0058] (3) After a fall occurs, the system can combine heart rate and blood oxygen data to further assess whether the elderly are injured. Through the heart rate data RR obtained by the heart rate sensor, the system calculates the heart rate variability index HRV to assess the health of the heart. At the same time, the blood oxygen content xy of the elderly is monitored in real time through the blood oxygen sensor to analyze the physical injuries that may be caused by the fall. Based on the heart rate variability index HRV and blood oxygen content xy, a health status model is constructed to calculate the comprehensive health index zhj, which is compared and evaluated with the preset health threshold B. If the comprehensive health index zhj is lower than the health threshold B, the system will consider that the elderly may be seriously injured and immediately trigger an injury alarm message, notify the guardian and call the emergency number. This multi-level risk assessment and health analysis mechanism ensures the efficient response of the system and the protection of the health and safety of the elderly. By accurately collecting and processing behavioral and health data, it can not only accurately identify whether the elderly have fallen, but also assess the degree of physical injury after the fall, and promptly remind the guardian to take measures. This not only improves the quality of life and safety of the elderly, but also provides strong technical support for families and caregivers to ensure that the physical and mental health of the elderly is fully protected. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] Figure 1 This is a flow chart of a multi-sensor based elderly fall alarm system of the present invention;

[0060] Figure 2 The present invention is a schematic diagram of the steps of a multi-sensor based fall alarm method for the elderly. DETAILED DESCRIPTION

[0061] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0062] Example 1, please refer to Figure 1 , the present invention provides a multi-sensor based elderly fall alarm system. To achieve the above purpose, the present invention is implemented through the following technical solutions: including a data acquisition module, a cloud platform processing module, a posture analysis module, a fall detection module and a health analysis module;

[0063] The data acquisition module collects the elderly's physical data in real time through smart devices and sensor groups installed on the elderly's body, builds a cloud processing platform, establishes communication connections with the sensor group and the smart bracelet through a wireless network, and transmits the physical data to the cloud processing platform in real time;

[0064] The cloud platform processing module receives the body data in real time through the cloud processing platform, performs preprocessing, obtains the state data group, and constructs a data repository to store the state data group, the state data group includes a behavior state data group and a health state data group;

[0065] The posture analysis module is used to calculate the acceleration modulus |a|, the angular velocity modulus |w| and the gait period according to the acquired behavior state data group , and then calculate the posture angle θ of the elderly based on the acceleration modulus |a| and the angular velocity modulus |w|;

[0066] The fall detection module constructs a fall detection model through the cloud processing platform, and obtains the acceleration modulus |a|, angular velocity modulus |w|, posture angle θ, gait cycle The pressure yl is imported into the fall detection model to obtain the fall risk assessment coefficient dfx, and then the fall assessment of the elderly is performed with the preset fall risk threshold A;

[0067] The health analysis module is used to extract the heart rate RR and blood oxygen content xy of the health status data group when the elderly person is in a fall state during fall assessment, calculate the heart rate variability index HRV based on the heart rate RR, and build a health status model, import the heart rate variability index HRV and blood oxygen content xy into the health status model, calculate and obtain the comprehensive health index zhj, and perform health assessment on the elderly person with a preset health threshold B.

[0068] In this embodiment, the data acquisition module collects the physical data of the elderly in real time through the combination of intelligent devices and sensor groups, and transmits the data to the cloud processing platform through a wireless network. This module realizes seamless connection of data, ensures real-time performance and high precision, and provides a reliable data basis for subsequent posture analysis and fall detection. The cloud platform processing module ensures the efficient operation of the system by receiving, preprocessing and storing data in real time, and timely generates behavioral state data groups and health state data groups, providing sufficient support for further risk assessment and health analysis. In terms of fall detection, the system accurately calculates the movement state of the elderly through the posture analysis module, including acceleration modulus |a|, angular velocity modulus |w| and gait cycle, and then determines the posture angle θ to analyze whether the elderly have a risk of falling. The fall detection module constructs a fall detection model through the cloud processing platform, imports parameters such as acceleration modulus |a|, angular velocity modulus |w|, posture angle, gait cycle, etc. into the model, calculates the fall risk assessment coefficient dfx, and compares it with the preset fall risk threshold A. This effectively improves the accuracy of fall detection. Compared with traditional single sensor technology or judgment methods based on simple rules, it can more accurately capture the moment of fall and improve the timeliness and reliability of early warning. In addition, the health analysis module provides further health assessment after the elderly fall. It combines heart rate RR and blood oxygen content xy to calculate the heart rate variability index HRV and the comprehensive health index zhj to judge the health status of the elderly. If the comprehensive health index zhj is lower than the health threshold B, the system will immediately generate an injury alarm message to remind the guardian to take timely treatment measures. This module not only improves the ability to assess health risks after a fall, but also provides accurate reference data for emergency response. Compared with the existing technology, the system greatly improves the accuracy of fall risk detection by integrating multiple sensors and cloud-based intelligent analysis, and provides a more comprehensive and scientific guarantee for the health and safety of the elderly. The real-time, accuracy and intelligence of the system significantly enhance the effect of fall warning and health monitoring, making the treatment and prevention of falls in the elderly more efficient and timely, and significantly improving the quality of life and safety level of the elderly.

[0069] Example 2: This example is an explanation of Example 1. Figure 1 ,Specifically: the data acquisition module includes a habit data collection unit and a processing platform building unit;

[0070] The habit data collection unit collects the elderly's physical data in real time through a sensor group installed on the elderly's head, the bottom of the insole and the inside of the smart device, and marks the collection timestamp;

[0071] The sensor group includes a pressure sensor, an acceleration sensor, a heart rate sensor, a blood oxygen sensor and a gyroscope, and the smart device is a smart bracelet;

[0072] The processing platform construction unit is used to construct a cloud processing platform, establish a communication connection between the cloud processing platform, the sensor group and the smart bracelet through a wireless network, and transmit the acquired body data to the cloud processing platform in real time.

[0073] In this embodiment, the data acquisition module significantly improves the data acquisition capability and real-time transmission efficiency of the system through the close cooperation of the habit data acquisition unit and the processing platform construction unit. The habit data acquisition unit can comprehensively and in real time monitor the body data of the elderly by installing various sensors such as pressure sensors, acceleration sensors, heart rate sensors, blood oxygen sensors and gyroscopes on the head, bottom of the insole and smart bracelet of the elderly, and accurately mark the timestamp of each data acquisition. This refined data acquisition method ensures continuous monitoring of multi-dimensional and physiological states, allowing the system to obtain more useful personalized data and provide high-precision support for subsequent analysis. The processing platform construction unit ensures the real-time transmission and centralized processing of data by building a cloud processing platform, so that the collected body data can be instantly transmitted to the cloud platform for processing through a wireless network, eliminating delays and errors in the data transmission process. This design effectively improves the system response speed and accuracy, provides reliable technical support for real-time monitoring of the health of the elderly and timely warning, and has important practical significance, especially in improving the effect of fall warning and health assessment of the elderly.

[0074] Example 3: This example is an explanation of Example 2. Figure 1 ,Specifically: the cloud platform processing module includes a data processing unit and a data storage unit;

[0075] The data processing unit receives the body data in real time through the cloud processing platform, and performs denoising, filtering and dimensionless processing to obtain a state data group;

[0076] The state data group includes a behavior state data group and a health state data group;

[0077] The behavior state data group includes an acceleration vector a, an angular velocity vector w and a pressure yl;

[0078] The health status data group includes heart rate RR and blood oxygen content xy;

[0079] The data storage unit constructs a data repository through a cloud processing platform, and stores the acquired status data groups into the data repository after time-series integration according to the marked timestamps, thereby ensuring the synchronization of different data collected at the same timestamp.

[0080] In this embodiment, the cloud platform processing module greatly improves the data processing capability and accuracy of the system through the collaboration of the data processing unit and the data storage unit. The data processing unit is responsible for denoising, filtering and dimensionless processing of the real-time collected body data, thereby ensuring the accuracy and reliability of the data. This processing process not only improves the stability of the system, but also effectively eliminates external interference, making subsequent behavioral state analysis and health assessment more accurate. The construction of the state data group organically integrates the behavioral state data group and the health state data group, providing rich data support for the health risk assessment of the elderly. At the same time, the data storage unit ensures the consistency of multi-dimensional data collected at the same timestamp through time series integration and synchronous storage mechanism, which is crucial for subsequent data analysis and fall warning. This design enables the system to not only process and store massive data in real time, but also ensure the consistency and synchronization of data in later analysis, greatly improving the accuracy and timeliness of the system in fall detection and health assessment.

[0081] Example 4: This example is an explanation of Example 3. Figure 1 ,Specifically: the posture analysis module includes a motion state analysis unit and a posture analysis unit;

[0082] The motion state analysis unit is used to analyze the motion state of the elderly according to the cloud processing platform, detect the acceleration change of the elderly through the accelerometer, judge the motion intensity and direction, and calculate the acceleration vector a=(a x , a y , a z ), calculate and obtain the acceleration modulus |a|, the specific formula is as follows: , capturing the instantaneous drastic acceleration change when falling, where a x 、a y and a z Represent the acceleration values ​​along the x-axis, y-axis and z-axis respectively, the x-axis represents the horizontal front-to-back direction, the y-axis represents the horizontal left-to-right direction, and the z-axis represents the vertical up-down direction;

[0083] The gyroscope is used to capture the angular velocity changes of the elderly's movements, identify sudden pitch and roll movements, and calculate the angular velocity vector w=(w x , w y , w z ), calculate and obtain the angular velocity modulus |w|, the specific formula is as follows: , where w x 、w y and w z Represent the angular velocity values ​​along the x-axis, y-axis and z-axis respectively, the x-axis represents the horizontal front-to-back direction, the y-axis represents the horizontal left-to-right direction, and the z-axis represents the vertical up-down direction;

[0084] The posture analysis unit is used to accurately identify the posture angle θ of the elderly in dynamic and static conditions, and the acceleration vector a=(a x , a y , a z ), calculate the pitch angle θ of the elderly person’s body x and roll angle θ y , the specific algorithm formula is: , , where arctan represents the inverse tangent function;

[0085] The time interval sj of each step is obtained by collecting the pressure of each two times by the pressure sensor, and the gait cycle of the elderly is calculated. , analyze the time it takes for an elderly person to land on one foot and then land on the same foot again. The specific formula is as follows: , where N represents the total number of steps in the monitoring period.

[0086] In this embodiment, the motion state analysis unit detects the acceleration change and angular velocity change of the elderly in real time through the accelerometer and gyroscope, and collects the acceleration vector a=(a x , a y , a z ) and the angular velocity vector w=(w x , w y , w z ), and calculate the acceleration modulus |a| and angular velocity modulus |w| respectively, which can accurately identify the drastic changes in movement intensity, direction, and posture, especially the instantaneous drastic acceleration and angular velocity changes before a fall occurs, providing early warning. The posture analysis unit uses the acceleration vector a=(a x , a y , a z ), calculate the pitch angle θ of the elderly x and roll angle θ y, it can accurately identify the posture angle of the elderly in both dynamic and static states, further enhancing the system's recognition ability in complex environments. In addition, through the precise analysis of the gait cycle by the pressure sensor, the system can monitor the gait changes of the elderly and judge their walking stability, providing more dimensional data support for the prediction and early warning of fall risks. Overall, this multi-dimensional posture analysis not only improves the accuracy of fall detection, but also provides a more comprehensive and accurate basis for real-time monitoring and health early warning, significantly enhancing the intelligence level and reliability of the elderly safety monitoring system.

[0087] Example 5: This example is an explanation of Example 4. Figure 1 ,Specifically: the fall detection module includes a fall analysis unit and a fall assessment unit;

[0088] The fall analysis unit constructs a fall detection model through a cloud processing platform, and then obtains the acceleration modulus |a|, angular velocity modulus |w|, posture angle θ, and gait cycle The pressure yl is imported into the fall detection model to obtain the fall risk assessment coefficient dfx and quantify the factors causing the elderly to fall;

[0089] The fall risk assessment coefficient dfx is calculated by the following formula;

[0090] ;

[0091] In the formula, e -b represents the exponential decay function, yl d Indicates the current pressure value, yl z It represents the pressure value of the elderly's normal contact with the ground. c1, c2 and c3 represent the weight coefficients of acceleration modulus |a|, angular velocity modulus |w| and posture angle θ respectively, and c1+c2+c3=1. The specific values ​​are set by the user.

[0092] The fall assessment unit is used to count the distribution of fall events and normal activities in the characteristic dimensions of acceleration vector a, angular velocity vector w, posture angle θ and pressure yl according to a large number of historical behavior state data groups in the data storage library, and mark the fall events and normal activities, calculate the mean and standard deviation of the risk values ​​of normal activities and fall events by using statistical methods, preset the fall risk threshold A, and perform the fall assessment of the elderly with the obtained fall risk assessment coefficient dfx, analyze the current state of the elderly, and generate corresponding measures. The specific assessment plan is as follows;

[0093] When the fall risk assessment coefficient dfx> the fall risk threshold A, the elderly person is in a fall state, and a fall alarm message is generated. The cloud processing platform transmits the fall alarm message to the guardian user end through the wireless network and triggers a deep assessment.

[0094] When the fall risk assessment coefficient dfx ≤ the fall risk threshold A, the elderly person is in a normal state and maintains normal monitoring.

[0095] In this embodiment, the fall detection module significantly improves the accuracy and response speed of elderly people's fall monitoring through the collaboration of the fall analysis unit and the fall assessment unit. and pressure yl, and combined with the fall detection model built by the cloud processing platform, the system can quantify the fall risk assessment coefficient dfx, so as to more accurately assess the fall risk of the elderly. In particular, by adjusting the weight coefficients of the acceleration modulus |a|, the angular velocity modulus |w| and the posture angle θ, the system realizes individualized risk assessment for different elderly people, further improving the sensitivity and accuracy of fall detection. In addition, the fall assessment unit calculates the mean and standard deviation of the risk value through statistical analysis of historical behavior state data, combined with the distribution characteristics of normal activities and fall events, and sets the fall risk threshold A to ensure that the fall event can be captured sensitively. When the fall risk assessment coefficient dfx exceeds the fall risk threshold A, the system will immediately generate an alarm message and notify the guardian, further triggering a deep assessment to ensure that the elderly can get help in time. The advantage of this system lies in its highly personalized and accurate fall warning mechanism, which greatly reduces the risk of false alarms and missed alarms, improves the efficiency of safety management of the elderly, and provides guardians with more reliable and timely support, effectively protecting the quality of life and life safety of the elderly.

[0096] Example 6: This example is an explanation of Example 5. Figure 1 ,Specifically: the health analysis module is used to analyze the harm caused to the elderly by the fall assessment;

[0097] The health analysis module includes a health analysis unit and a depth analysis unit;

[0098] The health analysis unit is used to analyze the stress response of the elderly's heart after a fall, especially for elderly people with heart disease, which may cause heart problems, and to construct a heart rate variability formula through the heart rate RR collected in real time by the heart rate sensor to calculate and obtain the heart rate variability index HRV;

[0099] The heart rate variability index HRV is calculated by the following formula:

[0100] ;

[0101] Where M represents the total number of sampling points, RR i Represents the time interval from the i-1th heartbeat to the i-th heartbeat, RRi-1 Indicates the time interval from the i-2th heartbeat to the i-1th heartbeat;

[0102] By placing a blood oxygen sensor on the elderly's body, the fluctuations of the local blood oxygen content xy in the body can be obtained in real time to analyze the elderly's physical injuries. Falls that cause fractures will cause local bleeding at the fracture site, affecting the reduction of red blood cells and hemoglobin in the blood and the decrease of local blood oxygen saturation. Falls that hit the head and cause concussion and intracranial hemorrhage will directly affect the brain's demand and supply of oxygen, resulting in insufficient blood oxygen levels in the brain area.

[0103] The in-depth analysis unit includes a comprehensive health analysis unit and a comprehensive health assessment unit;

[0104] The comprehensive health analysis unit is used to construct a health status model, based on the acquired heart rate variability index HRV and blood oxygen content xy, import the health status model for summary calculation, obtain the comprehensive health index zhj, and comprehensively analyze the physical health status of the elderly after the fall;

[0105] The comprehensive health index zhj is calculated by the following formula:

[0106] ;

[0107] Where, HRV z Indicates a healthy heart rate variability value. 90% indicates the lower limit of normal blood oxygen content in the human body. A blood oxygen content below 90% means insufficient oxygen supply and a threat to physical health. 100% indicates the oxygen saturation state of the blood oxygen content.

[0108] The comprehensive health assessment unit extracts a large amount of historical data from the data repository, obtains the average value and standard deviation of the heart rate variability index HRV of the elderly through statistical methods, and then associates the blood oxygen content xy with the lower limit of 90% of the normal blood oxygen content of the human body according to conventional clinical standards, and imports the health status model, outputs the value to the preset health threshold B, and performs health assessment of the elderly with the obtained comprehensive health index zhj. The specific assessment plan is as follows;

[0109] When the comprehensive health index zhj ≥ health threshold B, the elderly person’s fall did not cause physical injury;

[0110] When the comprehensive health index zhj is less than the health threshold B, the elderly person falls and causes physical injury. At this time, an injury alarm message is generated. The cloud processing platform transmits the injury alarm message to the guardian user terminal through the wireless network, reminding the guardian that the elderly person is injured and calling the emergency number immediately.

[0111] In this embodiment, the health analysis module realizes a comprehensive assessment of the health status of the elderly after a fall through the comprehensive application of heart rate sensors and blood oxygen sensors, as well as the health status model of the deep analysis unit. Its special advantage is that it can accurately calculate the heart rate variability index HRV and monitor the blood oxygen content xy in real time, and combine the dynamic changes of the two to construct a comprehensive health index zhj, so as to analyze whether the fall causes serious health damage. Compared with the monitoring method of a single parameter, the system realizes the integrated analysis of multi-dimensional data by importing HRV and blood oxygen content xy into the health status model. Especially in the case where a fall may cause fractures, local bleeding and craniocerebral injury, the system can accurately capture the local decrease in blood oxygen saturation and abnormal heart rate fluctuations, timely assess risks, and automatically generate injury alarm information when the comprehensive health index zhj is lower than the health threshold B. This mechanism significantly improves the perception of health risks after a fall, not only provides clear health assessment information for guardians, but also wins valuable time for medical response, and effectively reduces the potential threat of fall events to the health of the elderly.

[0112] Example 7, please refer to Figure 2 , a multi-sensor based elderly fall alarm method, comprising the following steps:

[0113] S1. Collect the elderly's physical data in real time through smart devices and sensor groups installed on the elderly's body, build a cloud processing platform, establish communication connections with the sensor group and smart bracelets through wireless networks, and transmit the physical data to the cloud processing platform in real time;

[0114] S2, the cloud processing platform receives the body data in real time, performs preprocessing, obtains the state data group, and builds a data repository to store the state data group, the state data group includes a behavior state data group and a health state data group;

[0115] S3, calculate the acceleration modulus |a|, angular velocity modulus |w| and gait period based on the acquired behavior state data set , and then calculate the posture angle θ of the elderly based on the acceleration modulus |a| and the angular velocity modulus |w|;

[0116] S4. Build a fall detection model based on the cloud processing platform, and obtain the acceleration modulus |a|, angular velocity modulus |w|, posture angle θ, and gait cycle The pressure yl is imported into the fall detection model to obtain the fall risk assessment coefficient dfx, and then the fall assessment of the elderly is performed with the preset fall risk threshold A;

[0117] S5. When the elderly person is in a falling state during fall assessment, the heart rate RR and blood oxygen content xy of the health status data group are extracted, the heart rate variability index HRV is calculated based on the heart rate RR, and a health status model is constructed. The heart rate variability index HRV and blood oxygen content xy are imported into the health status model, the comprehensive health index zhj is calculated, and the health assessment of the elderly person is performed with the preset health threshold B.

[0118] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A multi-sensor based fall alarm system for the elderly, characterized by: It includes data acquisition module, cloud platform processing module, posture analysis module, fall detection module and health analysis module; The data acquisition module collects the elderly's physical data in real time through smart devices and sensor groups installed on the elderly's body, builds a cloud processing platform, establishes communication connections with the sensor group and the smart bracelet through a wireless network, and transmits the physical data to the cloud processing platform in real time; The cloud platform processing module receives the body data in real time through the cloud processing platform, performs preprocessing, obtains the state data group, and constructs a data repository to store the state data group, the state data group includes a behavior state data group and a health state data group; The posture analysis module is used to calculate the acceleration modulus |a|, the angular velocity modulus |w| and the gait period according to the acquired behavior state data group , and then calculate the posture angle θ of the elderly based on the acceleration modulus |a| and the angular velocity modulus |w|; The fall detection module constructs a fall detection model through the cloud processing platform, and obtains the acceleration modulus |a|, angular velocity modulus |w|, posture angle θ, gait cycle The pressure yl is imported into the fall detection model to obtain the fall risk assessment coefficient dfx, and then the fall assessment of the elderly is performed with the preset fall risk threshold A; The health analysis module is used to extract the heart rate RR and blood oxygen content xy of the health status data group when the elderly person is in a fall state during fall assessment, calculate the heart rate variability index HRV based on the heart rate RR, and build a health status model, import the heart rate variability index HRV and blood oxygen content xy into the health status model, calculate and obtain the comprehensive health index zhj, and perform health assessment on the elderly person with a preset health threshold B.

2. The multi-sensor based fall alarm system for the elderly according to claim 1 is characterized by: The data acquisition module includes a habit data acquisition unit and a processing platform construction unit; The habit data collection unit collects the elderly's physical data in real time through a sensor group installed on the elderly's head, the bottom of the insole and the inside of the smart device, and marks the collection timestamp; The sensor group includes a pressure sensor, an acceleration sensor, a heart rate sensor, a blood oxygen sensor and a gyroscope, and the smart device is a smart bracelet; The processing platform construction unit is used to construct a cloud processing platform, establish a communication connection between the cloud processing platform, the sensor group and the smart bracelet through a wireless network, and transmit the acquired body data to the cloud processing platform in real time.

3. The multi-sensor based fall alarm system for the elderly according to claim 2, characterized in that: The cloud platform processing module includes a data processing unit and a data storage unit; The data processing unit receives the body data in real time through the cloud processing platform, and performs denoising, filtering and dimensionless processing to obtain a state data group; The state data group includes a behavior state data group and a health state data group; The behavior state data group includes an acceleration vector a, an angular velocity vector w and a pressure yl; The health status data group includes heart rate RR and blood oxygen content xy; The data storage unit constructs a data repository through a cloud processing platform, and stores the acquired status data groups in the data repository after time-series integration according to the marked timestamps.

4. The multi-sensor based fall alarm system for the elderly according to claim 3 is characterized by: The posture analysis module includes a motion state analysis unit and a posture analysis unit; The motion state analysis unit is used to analyze the motion state of the elderly according to the cloud processing platform, detect the acceleration change of the elderly through the accelerometer, judge the motion intensity and direction, and calculate the acceleration vector a=(a x , a y , a z ), calculate and obtain the acceleration modulus |a|, the specific formula is as follows: , capturing the instantaneous drastic acceleration change when falling, where a x 、a y and a z Represent the acceleration values ​​along the x-axis, y-axis and z-axis respectively, the x-axis represents the horizontal front-to-back direction, the y-axis represents the horizontal left-to-right direction, and the z-axis represents the vertical up-down direction; The gyroscope is used to capture the angular velocity changes of the elderly's movements, identify sudden pitch and roll movements, and calculate the angular velocity vector w=(w x , w y , w z ), calculate and obtain the angular velocity modulus |w|, the specific formula is as follows: , where w x 、w y and w z Represent the angular velocity values ​​along the x-axis, y-axis and z-axis respectively, the x-axis represents the horizontal front-to-back direction, the y-axis represents the horizontal left-to-right direction, and the z-axis represents the vertical up-down direction; The posture analysis unit is used to accurately identify the posture angle θ of the elderly in dynamic and static conditions, and the acceleration vector a=(a x , a y , a z ), calculate the pitch angle θ of the elderly person’s body x and roll angle θ y , the specific algorithm formula is: , , where arctan represents the inverse tangent function; The time interval sj of each step is obtained by collecting the pressure of each two times by the pressure sensor, and the gait cycle of the elderly is calculated. , analyze the time it takes for an elderly person to land on one foot and then land on the same foot again. The specific formula is as follows: , where N represents the total number of steps in the monitoring period.

5. The multi-sensor based fall alarm system for the elderly according to claim 4, characterized in that: The fall detection module includes a fall analysis unit and a fall assessment unit; The fall analysis unit constructs a fall detection model through a cloud processing platform, and then obtains the acceleration modulus |a|, angular velocity modulus |w|, posture angle θ, and gait cycle The pressure yl is imported into the fall detection model to obtain the fall risk assessment coefficient dfx and quantify the factors causing the elderly to fall; The fall risk assessment coefficient dfx is calculated by the following formula; ； In the formula, e -b represents the exponential decay function, yl d Indicates the current pressure value, yl z It represents the pressure value of the elderly's normal contact with the ground. c1, c2 and c3 represent the weight coefficients of acceleration modulus |a|, angular velocity modulus |w| and posture angle θ respectively, and c1+c2+c3=1. The specific values ​​are set by the user.

6. The multi-sensor based fall alarm system for the elderly according to claim 5, characterized in that: The fall assessment unit is used to count the distribution of fall events and normal activities in the characteristic dimensions of acceleration vector a, angular velocity vector w, posture angle θ and pressure yl according to a large number of historical behavior state data groups in the data storage library, and mark the fall events and normal activities, calculate the mean and standard deviation of the risk values ​​of normal activities and fall events by using statistical methods, preset the fall risk threshold A, and perform the fall assessment of the elderly with the obtained fall risk assessment coefficient dfx, analyze the current state of the elderly, and generate corresponding measures. The specific assessment plan is as follows; When the fall risk assessment coefficient dfx> the fall risk threshold A, the elderly person is in a fall state, and a fall alarm message is generated. The cloud processing platform transmits the fall alarm message to the guardian user end through the wireless network and triggers a deep assessment. When the fall risk assessment coefficient dfx ≤ the fall risk threshold A, the elderly person is in a normal state and maintains normal monitoring.

7. The multi-sensor based fall alarm system for the elderly according to claim 6, characterized in that: The health analysis module is used to analyze the damage caused to the elderly by the fall when the elderly is in a fall state; The health analysis module includes a health analysis unit and a depth analysis unit; The health analysis unit is used to analyze the stress response of the elderly's heart after the elderly falls, construct a heart rate variability formula through the heart rate RR collected in real time by the heart rate sensor, and calculate the heart rate variability index HRV; The heart rate variability index HRV is calculated by the following formula: ； Where M represents the total number of sampling points, RR i Represents the time interval from the i-1th heartbeat to the i-th heartbeat, RR i-1 Indicates the time interval from the i-2th heartbeat to the i-1th heartbeat; By placing a blood oxygen sensor on the elderly's body, the fluctuation changes of the local blood oxygen content xy in the body can be obtained in real time to analyze the elderly's physical injuries.

8. The multi-sensor based fall alarm system for the elderly according to claim 7, characterized in that: The in-depth analysis unit includes a comprehensive health analysis unit and a comprehensive health assessment unit; The comprehensive health analysis unit is used to construct a health status model, based on the acquired heart rate variability index HRV and blood oxygen content xy, import the health status model for summary calculation, obtain the comprehensive health index zhj, and comprehensively analyze the physical health status of the elderly after the fall; The comprehensive health index zhj is calculated by the following formula: ； Where, HRV z It represents a healthy heart rate variability value, 90% represents the lower limit of normal blood oxygen content in the human body, and 100% represents the oxygen saturation state of blood oxygen content.

9. The multi-sensor based fall alarm system for the elderly according to claim 8, characterized in that: The comprehensive health assessment unit extracts a large amount of historical data from the data repository, obtains the average value and standard deviation of the heart rate variability index HRV of the elderly through statistical methods, and then associates the blood oxygen content xy with the lower limit of 90% of the normal blood oxygen content of the human body according to conventional clinical standards, and imports the health status model, outputs the value to the preset health threshold B, and performs health assessment of the elderly with the obtained comprehensive health index zhj. The specific assessment plan is as follows; When the comprehensive health index zhj ≥ health threshold B, the elderly person’s fall did not cause physical injury; When the comprehensive health index zhj is less than the health threshold B, the elderly person falls and causes physical injury. At this time, an injury alarm message is generated. The cloud processing platform transmits the injury alarm message to the guardian user terminal through the wireless network, reminding the guardian that the elderly person is injured and to call the emergency number immediately.

10. A multi-sensor based elderly fall alarm method, applied to a multi-sensor based elderly fall alarm system according to any one of claims 1 to 9, characterized in that: The following steps are involved: S1. Collect the elderly's physical data in real time through smart devices and sensor groups installed on the elderly's body, build a cloud processing platform, establish communication connections with the sensor group and smart bracelets through wireless networks, and transmit the physical data to the cloud processing platform in real time; S2, the cloud processing platform receives the body data in real time, performs preprocessing, obtains the state data group, and builds a data repository to store the state data group, the state data group includes a behavior state data group and a health state data group; S3, calculate the acceleration modulus |a|, angular velocity modulus |w| and gait period based on the acquired behavior state data set , and then calculate the posture angle θ of the elderly based on the acceleration modulus |a| and the angular velocity modulus |w|; S4. Build a fall detection model based on the cloud processing platform, and obtain the acceleration modulus |a|, angular velocity modulus |w|, posture angle θ, and gait cycle The pressure yl is imported into the fall detection model to obtain the fall risk assessment coefficient dfx, and then the fall assessment of the elderly is performed with the preset fall risk threshold A; S5. When the elderly person is in a falling state during fall assessment, the heart rate RR and blood oxygen content xy of the health status data group are extracted, the heart rate variability index HRV is calculated based on the heart rate RR, and a health status model is constructed. The heart rate variability index HRV and blood oxygen content xy are imported into the health status model, the comprehensive health index zhj is calculated, and the health assessment of the elderly person is performed with the preset health threshold B.

Images