An intelligent anti-falling bed risk management system based on artificial intelligence

Abstract

The application discloses an intelligent anti-falling bed risk management system based on artificial intelligence, and relates to the technical field of artificial intelligence.The system comprises a management center, which is connected with a monitoring and collecting module, a transmission and processing module, a body position analysis module and a multidimensional supervision module.A static virtual space is constructed for bedridden old people, and the collected bedridden state data is marked with positions through the static virtual space based on the intelligent anti-falling bed to obtain a bedridden continuous track.The boundary distance is counted according to the bedridden continuous track and is converted into a real-time track fluctuation graph, the real-time track fluctuation graph is divided in range and the range proportion is counted to obtain a curve determination range.The system is simulated and diagnosed through the static virtual space to obtain an abnormal recognition state, and the bedridden old people are dynamically fed back and adjusted according to the abnormal recognition state based on the static virtual space to obtain anti-falling intervention measures.The system effectively reduces the occurrence of bed-falling accidents of the old people, improves the nursing efficiency and reduces the burden of nursing personnel.

Description

Technical Field

[0001] This invention relates to the field of artificial intelligence technology, specifically to an intelligent fall-prevention bed risk management system based on artificial intelligence. Background Technology

[0002] In modern society, hospitals, as vital providers of healthcare services, face an increasing number of patient safety issues. Among these, patient falls from bed are a common medical accident, causing not only physical but also psychological harm to patients. To reduce the risk of falls and improve patient safety, traditional fall prevention measures mainly include bed rails and nurse rounds; however, these methods have certain limitations in implementation.

[0003] In recent years, the rapid development of artificial intelligence (AI) technology has brought new opportunities to the medical field. Intelligent fall prevention bed risk management systems have emerged, based on AI technology, to achieve multiple functions such as continuous monitoring of the weight and nutrition of bedridden elderly, fall risk identification, timely risk notification, and active and passive risk intervention. This effectively solves the problems of incomplete monitoring, untimely warnings, and inadequate intervention found in similar products currently on the market. Therefore, this paper presents an AI-based intelligent fall prevention bed risk management system. Through sensors installed on the hospital bed, it collects patients' physiological and behavioral data in real time. Using AI algorithms, it analyzes the collected data in real time to identify patients' fall risks, providing targeted protective measures for medical staff. This effectively reduces the risk of patients falling from bed and improves the quality of medical services. Summary of the Invention

[0004] The purpose of this invention is to provide an intelligent fall protection bed risk management system based on artificial intelligence, so as to solve the problems of difficult control of intelligent fall protection bed risks and low safety of fall protection beds mentioned in the background art.

[0005] An intelligent fall prevention bed risk management system based on artificial intelligence includes a management center, which is connected to a monitoring and acquisition module, a transmission and processing module, a body position analysis module, and a multi-dimensional monitoring module.

[0006] The monitoring and acquisition module is used to collect bedridden elderly people's bedridden status data and build a bedridden data storage repository;

[0007] The transmission processing module is used to construct a quiet virtual space for bedridden elderly people, and to virtually transform and mark the smart anti-fall bed through the quiet virtual space to obtain the limited bed boundary, to map and identify the bed data storage repository, to obtain the bed model mark position point, and to continuously mark the bed model mark position point to obtain the continuous movement track of the bed.

[0008] The body position analysis module is used to perform boundary distance statistics based on the continuous movement of the bedridden person, obtain the boundary movement difference sequence, perform digital visualization of the boundary movement difference sequence, obtain the real-time movement fluctuation map, divide the real-time movement fluctuation map into ranges and count the range proportions to obtain the curve judgment range.

[0009] The multi-dimensional monitoring module is used to perform simulated diagnosis through a resting virtual space to obtain abnormal identification status. Based on the abnormal identification status in the resting virtual space, it dynamically adjusts the bedridden elderly to obtain virtual intervention measures, and adjusts and optimizes the virtual intervention measures to obtain fall prevention intervention measures.

[0010] Preferably, the process of collecting bedridden status data of elderly people includes:

[0011] Data collection and deployment are carried out on the intelligent fall protection bed to obtain motion capture terminals;

[0012] The motion capture device collects data on bedridden elderly people in real time to obtain bedridden status data, and timestamps the real-time collected bedridden status data to obtain the actual collection time point.

[0013] A database of bedridden elderly people was constructed to obtain a bedridden data repository. Based on the actual collection time, the obtained bedridden status data was uploaded to the bedridden data repository.

[0014] Preferably, the process of obtaining the defined bed boundary includes:

[0015] The space for bedridden elderly is constructed to obtain a virtual resting space, and the smart anti-fall bed is mapped to the virtual resting space to obtain an anti-fall bed model. Based on the virtual resting space, the bedridden elderly are virtually placed to obtain a virtual bed posture.

[0016] Upload the bed rest data repository to the resting virtual space, and use the resting virtual space to mark the boundaries of the fall-prevention bed model to obtain the defined bed boundaries.

[0017] Preferably, the process of continuously marking position points based on the bed model includes:

[0018] Obtain the bedridden data repository, classify and label the bedridden data repository, and obtain classification status data;

[0019] The bed rest monitoring cycle is set according to the bed rest data repository. Based on the bed rest monitoring cycle, the obtained classification status data is matched and marked in the fall prevention bed model to obtain the marked position points of the bed model.

[0020] Based on the bed rest monitoring cycle, the marked location points of the obtained bed model are linked over time to obtain the continuous movement trajectory of the bed resting person.

[0021] Preferably, the process of obtaining a real-time trajectory fluctuation map includes:

[0022] By statistically analyzing the differences in continuous movement patterns on the bed within a virtual space defined by the bed boundary, the difference in movement patterns at the boundary positions is obtained. Based on the continuous movement patterns on the bed, the difference in movement patterns at the boundary positions is sequentially combined to obtain a sequence of boundary movement pattern differences.

[0023] A two-dimensional rectangular coordinate system with respect to time is constructed. The boundary track difference sequence based on the continuous track of bed rest is uploaded to the two-dimensional rectangular coordinate system. Track difference change curve is generated according to the boundary track difference sequence, and the two-dimensional rectangular coordinate system containing the track difference change curve is marked as a real-time track fluctuation map.

[0024] Preferably, the process of dividing the real-time trajectory fluctuation map into ranges and calculating the range percentages includes:

[0025] Obtain a real-time movement fluctuation map, set an original reference axis for the real-time movement fluctuation map, upload the obtained original reference axis to the real-time movement fluctuation map, and perform movement monitoring on the real-time movement fluctuation map based on the obtained original reference axis to obtain the activity monitoring axis;

[0026] The range of the real-time trajectory fluctuation graph is statistically analyzed based on the activity monitoring axis and the original baseline axis to obtain the curve determination range.

[0027] Preferably, the process of conducting a simulated diagnosis through a virtual space while lying still includes:

[0028] Based on the range determined by the curve, a safety risk statistic is calculated from the real-time trajectory fluctuation map to obtain the regional proportion result;

[0029] By synchronously monitoring the virtual bed morphology in a quiet virtual space based on the regional proportion results, homomorphic change results are obtained.

[0030] Anomaly diagnosis is performed on the obtained homomorphic change results to obtain anomaly identification status.

[0031] Preferably, the process of obtaining fall protection interventions includes:

[0032] An intervention warning command is issued based on the anomaly identification status. The obtained intervention warning command is uploaded to the management center, which then sends the intervention warning command to the silent virtual space. The silent virtual space performs virtual intervention based on the anomaly identification status and obtains virtual intervention measures.

[0033] By using a virtual space to assess virtual intervention measures and obtain pre-adjustment intervention strategies, feedback adjustments are made to the virtual intervention measures based on the obtained intervention assessment results to obtain fall prevention intervention measures.

[0034] Compared with the prior art, the beneficial effects of the present invention are:

[0035] 1. Collect physiological movement data of bedridden elderly people in the intelligent fall-prevention bed and build a repository to store physiological movement data within a cycle. By collecting and storing the physiological movement data of the elderly, continuous analysis and decision-making can be carried out, thereby improving the accuracy of early warning and intervention. By constructing a virtual space to mark the location points of physiological movement data on the fall-prevention bed, the movement trajectory of the center of gravity and pressure peak of the bedridden elderly person on the fall-prevention bed can be obtained. The movement changes of the bedridden elderly person can be visualized, improving the targeted nature of care.

[0036] 2. By analyzing abnormal states of center of gravity movement and pressure peak movement, the system can identify potential bed fall risks in advance and issue timely warnings, thereby reducing the occurrence of bed fall accidents. Simulated rescue interventions for bedridden elderly can be conducted in virtual space to obtain virtual intervention measures. Testing different intervention measures in virtual space can identify the most effective bed fall prevention strategy, improve the success rate of interventions, and allow for adjustments and optimizations in virtual space to obtain the final bed fall prevention intervention measures, which is conducive to providing more personalized interventions. Attached Figure Description

[0037] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are 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.

[0038] Figure 1 This is a schematic diagram of the present invention. Detailed Implementation

[0039] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0040] like Figure 1 As shown, an intelligent fall prevention bed risk management system based on artificial intelligence includes a management center, which is connected to a monitoring and acquisition module, a transmission and processing module, a body position analysis module, and a multi-dimensional monitoring module.

[0041] The monitoring and acquisition module is used to collect bedridden elderly people's bedridden status data and build a bedridden data storage repository;

[0042] The transmission processing module is used to construct a quiet virtual space for bedridden elderly people, and to virtually transform and mark the smart anti-fall bed through the quiet virtual space to obtain the limited bed boundary, to map and identify the bed data storage repository, to obtain the bed model mark position point, and to continuously mark the bed model mark position point to obtain the continuous movement track of the bed.

[0043] The body position analysis module is used to perform boundary distance statistics based on the continuous movement of the bedridden person, obtain the boundary movement difference sequence, perform digital visualization of the boundary movement difference sequence, obtain the real-time movement fluctuation map, divide the real-time movement fluctuation map into ranges and count the range proportions to obtain the curve judgment range.

[0044] The multi-dimensional monitoring module is used to perform simulated diagnosis through a resting virtual space to obtain abnormal identification status. Based on the abnormal identification status in the resting virtual space, it dynamically adjusts the bedridden elderly to obtain virtual intervention measures, and adjusts and optimizes the virtual intervention measures to obtain fall prevention intervention measures.

[0045] In practical applications, with the continuous development of technology, intelligent fall-prevention beds are increasingly widely used in the medical field. However, this also brings unavoidable risks during the use of intelligent fall-prevention beds. Therefore, it is necessary to monitor the fall-prevention risks of intelligent fall-prevention beds. By monitoring indirect indicators such as the elderly's dynamic weight data and their bed-keeping status, a comprehensive assessment of the elderly's overall condition and potential risks can be made. This provides the elderly with a safe and intelligent fall-prevention bed risk management solution, which has broad application prospects and significant social value. The monitoring and data collection module collects bedridden elderly's bedridden status data and constructs a bedridden data repository. The specific process includes:

[0046] Data collection and deployment are carried out on the intelligent fall protection bed to obtain motion capture terminals;

[0047] The data collection deployment refers to the installation of a data collection terminal, or motion capture terminal, on the smart anti-fall bed to collect all data information of the bedridden elderly while in bed. The motion capture terminal is installed at the four feet of the nursing bed. In this embodiment, high-sensitivity weighing sensors and pressure sensors are used as motion capture terminals for information collection. Specifically, to improve the comprehensiveness and accuracy of data collection, the motion capture terminal can also be installed on physiological monitoring devices at the head or feet of the nursing bed. This facilitates a more comprehensive analysis of the integrated data generated by the bedridden elderly to determine whether the bedridden elderly have a risk of falling out of bed, and whether the anti-fall bed reduces this risk.

[0048] The motion capture device collects data on bedridden elderly people in real time, obtains bedridden status data, timestamps the real-time collected bedridden status data to obtain the actual collection time point, and associates the obtained bedridden status data with the corresponding bedridden elderly people.

[0049] The real-time acquisition refers to the collection of action or movement data generated by the elderly on the smart fall-prevention bed through the deployed motion capture terminal to obtain bed rest status data. The bed rest elderly refers to the elderly who need to lie down on the smart fall-prevention bed. The bed rest status data includes physiological data and behavioral data. The physiological data includes, but is not limited to, heart rate, blood pressure, and respiratory rate. The behavioral data includes, but is not limited to, pressure change data and center of gravity shift data of the elderly on the nursing bed. The pressure change data includes, but is not limited to, pressure distribution data, pressure peak data, and pressure fluctuation frequency. The center of gravity shift data includes, but is not limited to, center of gravity movement data, weight data, and center of gravity height data.

[0050] A database of bedridden elderly people was constructed to obtain a bedridden data repository. Based on the actual collection time point, the obtained bedridden status data was uploaded to the bedridden data repository.

[0051] The database construction refers to the creation of a corresponding data repository for each bedridden elderly person, namely the bedridden data repository, which is used to store all bedridden status data of the bedridden elderly person during the bedridden period. The stored bedridden status data is uploaded in chronological order according to the actual collection time of each bedridden status data. Therefore, the bedridden status data in the bedridden data repository is also arranged in chronological order according to the actual collection time.

[0052] The transmission processing module is used to construct a virtual resting space for bedridden elderly people. Through this virtual space, the intelligent fall-prevention bed is virtually transformed and marked to obtain defined bed boundaries. The bed resting data repository is mapped and identified to obtain bed model marker location points. Based on these marker locations, continuous marking is performed to obtain the bed resting continuous movement track. The specific process includes:

[0053] A spatial structure is constructed for bedridden elderly people to obtain a virtual resting space. The smart anti-fall bed is then mapped onto the virtual resting space to obtain an anti-fall bed model. Based on the virtual resting space, the bedridden elderly people are virtually placed to obtain a virtual bed posture. Here, virtual placement means that the bedridden elderly people in the virtual resting space are also mapped onto the anti-fall bed model to obtain the virtual bed posture of the bedridden elderly people in the virtual resting space. This represents the virtual state of the bedridden elderly people in the anti-fall bed model, which is completely synchronized with the actions and behaviors of the bedridden elderly people in reality.

[0054] The spatial construction refers to building a virtual space based on the ward space where the intelligent fall-prevention bed for the bedridden elderly is located, and mapping the intelligent fall-prevention bed into the virtual space to obtain a resting virtual space. The intelligent fall-prevention bed mapped into the virtual space is marked as a fall-prevention bed model. Here, the virtual space is a blank virtual space with the same size and structure as the space where the intelligent fall-prevention bed is located. "Mapping the intelligent fall-prevention bed into the virtual space" means converting the intelligent fall-prevention bed into a virtual three-dimensional model to obtain a three-dimensional intelligent fall-prevention bed template, which is the fall-prevention bed model. The structure and function of the fall-prevention bed are exactly the same as those of the real intelligent fall-prevention bed. It is just a virtual three-dimensional representation in the virtual space, which is convenient for studying the changes in the movement of the bedridden elderly during bed rest.

[0055] The obtained bed rest data repository is uploaded to the static virtual space. Based on the bed rest data repository, the fall prevention bed model is marked with boundaries through the static virtual space to obtain the limited bed boundaries.

[0056] The boundary markings indicate that, in the resting virtual space, the four edges of the smart fall-prevention bed are marked according to the four edges of the bed in the bed data repository. This marks the four edges of the bed as the defined bed boundaries. In this embodiment, the defined bed boundaries are highlighted with color in the resting virtual space, i.e., marked in red, so that the boundaries of the smart fall-prevention bed can be directly observed in the resting virtual space, making it easier to observe whether the elderly bedridden person's body movements indicate a tendency to fall out of bed.

[0057] Obtain a bed rest data repository, classify and label the obtained bed rest data repository to obtain classification status data, which includes center of gravity status data and pressure status data;

[0058] The classification label indicates that in the bedridden data repository, the bedridden state data is classified according to the pressure change data and center of gravity shift data included in the behavioral data, and the pressure state data corresponding to the pressure change data and the center of gravity state data corresponding to the center of gravity shift data are obtained. Thus, the bedridden data repository is classified according to the different categories of the classified state data.

[0059] The bed rest monitoring cycle is set according to the obtained bed rest data repository. The bed rest monitoring cycle is a period of time that includes several actual sampling time points, and the length of the time is the same as the length of time contained in the bed rest data repository.

[0060] Based on the bed rest monitoring cycle, the obtained classification status data is matched and marked in the fall prevention bed model to obtain the bed model marking position points, which include the center of gravity marking position points and the pressure marking position points;

[0061] It needs further explanation that, in the specific implementation process, the mapping mark represents marking the classification state data corresponding to each actual sampling time point in the fall-prevention bed model within the virtual resting space. The location points of the virtual bed posture state data in the fall-prevention bed model at each actual sampling time point are the bed model marking location points, representing the center of gravity position and pressure peak position at the corresponding time point. Based on the center of gravity state data and pressure state data included in the classification state data, the bed model marking location points include center of gravity marking location points and pressure marking location points. The classification state data is obtained by classifying the bed posture state data collected by the motion capture terminal. Therefore, based on each actual sampling time... The center of gravity and pressure data of bedridden elderly individuals at various points can be marked with markers at each actual sampling time point in the fall-prevention bed model. In this embodiment, the marked location points of the bed model are the center of gravity location point and the pressure peak location point, i.e., the center of gravity position and the pressure peak position. For example, the center of gravity offset data has a center of gravity marker location point at time t1, a corresponding center of gravity marker location point at time t2, and a corresponding center of gravity marker location point at time t3. The pressure peak at time t1 also has a corresponding pressure marker location point, the pressure peak at time t2 also has a corresponding pressure marker location point, and the pressure peak at time t3 also has a corresponding pressure marker location point.

[0062] Based on the bed rest monitoring cycle, the marked location points of the obtained bed model are linked over time to obtain the continuous bed rest trajectory, which includes the trajectory of center of gravity change and the trajectory of pressure peak change.

[0063] The time link refers to connecting the marked position points of the bed model according to the time sequence of the actual collection time points within the bed rest monitoring period. This obtains the trajectory of the center of gravity change and pressure peak change of the virtual bed rest form within the bed rest monitoring period, which is recorded as the bed rest continuous trajectory. It represents the trajectory of the changes in center of gravity and pressure data collected based on the changes in the movements of the bedridden elderly on the smart fall prevention bed within the bed rest monitoring period. The trajectory of the center of gravity change and the trajectory of the pressure peak change can be obtained, which is helpful in determining the movement tendency of the bedridden elderly.

[0064] The body position analysis module is used to perform boundary distance statistics based on the continuous trajectories of bedridden individuals, obtain a boundary trajectories difference sequence, visualize the boundary trajectories difference sequence using a graphical method to obtain a real-time trajectories fluctuation graph, divide the real-time trajectories fluctuation graph into ranges and statistically analyze the range percentages to obtain the curve determination range. The specific process includes:

[0065] By performing differential statistics on the continuous movement of the bed in a resting virtual space based on the obtained defined bed boundaries, the difference in movement at the boundary positions is obtained. Based on the continuous movement of the bed, the difference in movement at the boundary positions is sequentially combined to obtain a sequence of boundary movement differences. The boundary movement differences include a sequence of center of gravity boundary movement differences and a sequence of pressure boundary movement differences.

[0066] Furthermore, the difference statistics represent the horizontal distance from each bed model marker point to the defined bed boundary in the continuous movement trajectory of the bed in the virtual space of rest, which is the boundary position movement difference. Specifically, if a bed model marker point in the continuous movement trajectory of the bed exceeds the defined bed boundary, the boundary position movement difference is recorded as a negative number, indicating that part of the elderly person's body has exceeded the boundary of the smart protective bed. If it is within the boundary of the smart protective bed, the boundary position movement difference is recorded as a positive number. The obtained boundary position movement differences are then combined according to the order of the continuous movement trajectory of the bed to obtain the boundary movement difference sequence, which is the arrangement of the position distance differences between each bed model marker point and the boundary of the smart fall prevention bed. Then, for the center of gravity change trajectory, there is a corresponding center of gravity boundary movement difference sequence, and for the pressure peak change trajectory, there is a corresponding pressure boundary movement difference sequence.

[0067] Specifically, in this embodiment, "calculating the horizontal distance from each bed model marker location point to the defined bed boundary in the continuous bed rest trajectory" means calculating the horizontal distance from the bed model marker location point to the nearest defined bed boundary. For the left and right bed boundaries, a median line is drawn in the middle. The horizontal distance between the bed model marker location point and the bed boundary on the side of the median line is calculated. The distance is not calculated repeatedly. That is, for each actual sampling time point, the relationship between the bed model marker location point and the median line determines which side of the linear bed boundary the bed model marker location point is used for calculation.

[0068] A two-dimensional rectangular coordinate system with respect to time is constructed. The obtained boundary track difference sequence is uploaded to the two-dimensional rectangular coordinate system based on the continuous track of bed rest. Track difference change curves are generated according to the obtained boundary track difference sequence. The track difference change curves include the center of gravity boundary track difference change curve and the pressure boundary track difference change curve. The two-dimensional rectangular coordinate system containing the track difference change curves is marked as a real-time track fluctuation map.

[0069] Furthermore, the horizontal axis of the constructed two-dimensional rectangular coordinate system represents the actual sampling time point. Each corresponding time point has a corresponding bed model marker position point, and the boundary position track difference between the bed model marker position point and the defined bed boundary is obtained. The boundary position track difference is marked in the two-dimensional rectangular coordinate system, and connected in the order of the boundary track difference sequence to obtain the track difference change curve. Then, the corresponding center of gravity boundary track difference change curve is obtained according to the center of gravity boundary track difference sequence, and the corresponding pressure boundary track difference change curve is obtained according to the pressure boundary track difference sequence. In the real-time track fluctuation diagram, there are two change curves, which respectively represent the distance difference between the center of gravity and pressure and the bed boundary. The movement changes of the bedridden elderly on the intelligent anti-fall bed can be intuitively observed through the real-time track fluctuation diagram.

[0070] Obtain a real-time trajectory fluctuation map and set an original reference axis for the real-time trajectory fluctuation map. The original reference axis is the midline of the left and right bed boundaries in the defined bed boundary, that is, the distance from the midline to the two bed boundaries of the fall protection bed is the original reference axis.

[0071] The obtained original baseline axis is uploaded to the real-time trajectory fluctuation map. Based on the obtained original baseline axis, the real-time trajectory fluctuation map is moved to obtain the activity monitoring axis.

[0072] The motion monitoring refers to shifting the original reference axis closer to the horizontal axis in the real-time trajectory fluctuation graph, moving it to a position further away from the original reference axis. At the designated location, a new axis is obtained, denoted as the activity monitoring axis, where m is a positive integer greater than 1, set according to the width of the smart fall-prevention bed and the body shape of the bedridden elderly person. For example, if the bedridden elderly person lies flat with their body boundary centered on the center line of the smart fall-prevention bed, and the distance between their body boundary and the limit bed boundary is less than half of the original reference axis, then m can be selected as 2 or 3. This ensures that the bedridden elderly person has sufficient space to move around, while accurately identifying the tendency to fall within the safe activity range. For example, if the original reference axis is 8, the axis is moved to a position further away from the original reference axis. Location, i.e., the newly moved activity monitoring axis. Move it to a position 6 units away from the horizontal axis, 2 units away from the original reference axis.

[0073] Based on the obtained activity monitoring axis and the original baseline axis, the range proportion of the real-time movement fluctuation map is statistically analyzed to obtain the curve judgment range, which includes the appropriate activity range and the excessive activity range.

[0074] It should be further explained that, in the specific implementation process, the aforementioned range percentage statistics represent the division of the track difference change curve into regions based on the activity monitoring axis and the original reference axis in the real-time track fluctuation graph. The curves of different regions are the curve determination ranges. When the track difference change curve is located between the activity monitoring axis and the original reference axis, this part of the curve is recorded as the appropriate activity range, indicating that the distance from the limited bed boundary is within a safe range and there is no tendency to fall off the bed. When the track difference change curve is located below the activity monitoring axis, this part of the curve is recorded as the excessive activity range, indicating that the distance from the limited bed boundary is too close and there is a tendency to fall off the bed.

[0075] The multi-dimensional monitoring module is used to perform simulated diagnosis through a virtual space for bed rest, obtain abnormal identification status, dynamically adjust the bedridden elderly based on the abnormal identification status in the virtual space for bed rest, obtain virtual intervention measures, and adjust and optimize the virtual intervention measures to obtain fall prevention intervention measures. The specific process includes:

[0076] Based on the obtained curve determination range, the real-time trajectory fluctuation map is statistically analyzed for safety risks to obtain the regional proportion results, which include the proportion of safe range and the proportion of excessive range.

[0077] Furthermore, the safety risk statistics represent the proportion of the suitable activity range within the calculated curve range of the real-time movement fluctuation graph to the movement difference change curve, which is the area proportion result. The proportion of the movement difference change curve within the suitable activity range to the total movement difference change curve in the real-time movement fluctuation graph is recorded as the safe range proportion. This indicates that the higher the proportion of the movement difference change curve within the suitable activity range to the total movement difference change curve in the real-time movement fluctuation graph, the lower the safety risk for the bedridden elderly person on the smart anti-fall bed, and the bedridden elderly person's range of motion is within the safe activity range of the smart anti-fall bed. The proportion of the movement difference change curve within the excessive activity range to the total movement difference change curve in the real-time movement fluctuation graph is recorded as the excess range proportion, indicating that the changes in the center of gravity movement and pressure movement of the bedridden elderly person exceed the safe range, and the bedridden elderly person is at risk of falling out of bed.

[0078] By synchronously monitoring the virtual bed posture in the virtual space based on the area proportion results, homomorphic change results are obtained, including homomorphic changes and heteromorphic changes in center of gravity pressure.

[0079] The homomorphic monitoring refers to obtaining the regional proportion results within the corresponding bed rest monitoring period based on the motion data of the virtual bed rest form in each fall prevention bed model in the static virtual space. The regional proportion results are obtained by statistically analyzing the track difference change curves. The corresponding track difference change curves include the track difference change curve at the center of gravity boundary and the track difference change curve at the pressure boundary. It is necessary to analyze the same trend of change of the regional proportion results corresponding to the track difference change curve at the center of gravity boundary and the track difference change curve at the same time point. That is, within the bed rest monitoring period, the same regional proportion results in the track difference change curve at the center of gravity boundary and the track difference change curve at the pressure boundary are statistically analyzed.

[0080] When the regional proportions of the center of gravity boundary track difference change curve and the pressure boundary track difference change curve are both within the safe range, it indicates that the track change range of the center of gravity and pressure is within the safe activity range of the fall arrestor model, and the homomorphic change results of the center of gravity boundary track difference change curve and the pressure boundary track difference change curve are recorded as the center of gravity pressure homomorphic change.

[0081] When the regional proportions of the curves showing the difference in track difference between the center of gravity boundary and the pressure boundary are not all within the safe range, the homomorphic changes of these two curves are recorded as anomalous changes in center of gravity pressure. This indicates that the elderly person in bed is exhibiting significant movement on the intelligent anti-fall bed, i.e., an increased speed of movement towards the limited bed boundary, posing a risk of falling off the bed.

[0082] Anomaly diagnosis is performed on the obtained homomorphic change results to obtain anomaly identification status;

[0083] Furthermore, the abnormal diagnosis refers to identifying the isomorphic change result as an abnormal change in center of gravity pressure, obtaining an abnormal identification state, that is, identifying the isomorphic change result as an abnormal change in center of gravity pressure, and recording the virtual bed posture in this state as an abnormal identification state, indicating that the virtual bed posture has a risk of falling out of bed, and it is necessary to monitor the virtual bed posture in this state and give corresponding intervention measures to reduce the risk of falling out of bed.

[0084] Based on the obtained anomaly identification status, an intervention warning instruction is issued and uploaded to the management center. The management center then sends the intervention warning instruction to the silent virtual space. The silent virtual space performs virtual intervention based on the anomaly identification status and obtains virtual intervention measures.

[0085] It should be further explained that, in the specific implementation process, the intervention warning command indicates that the virtual bed rest form is in a dangerous state when it is in an abnormal identification state, triggering the warning mechanism, that is, issuing an intervention warning command. This is a command used to provide intervention and rescue, including the center of gravity marker position point and pressure marker position point of the virtual bed rest form in the abnormal identification state. The management center can obtain the position point of the abnormal identification state of the virtual bed rest form in the first time, that is, the position point of the abnormal action. The quiet virtual space performs virtual intervention according to the received intervention warning command. The virtual intervention means that the fall protection model is adjusted in the quiet virtual space to protect the virtual bed rest form from falling. By adjusting the height of the intelligent electric airbag installed on the side of the intelligent fall protection bed, the body parts of the virtual bed rest form that are moving, such as arms and thighs, are blocked. The raised intelligent electric airbag can effectively block arms and thighs that are close to the side of the bed.

[0086] Furthermore, since the boundary position tracking difference is obtained by calculating the horizontal distance from the marked position point of each bed model in the continuous bed rest tracking to the defined bed boundary, the intelligent electric airbag can be adjusted according to the spatial position of the boundary position tracking difference during virtual intervention. For example, the vertical distance from the boundary position to the bed surface, the horizontal distance from the intelligent electric airbag, and the distance from the head or foot of the bed can be obtained in the virtual space of resting. This yields a three-dimensional position point, which can be used to adjust the height, width, and distance from the head or foot of the bed. All of these are determined based on the three-dimensional position point of the boundary position tracking difference, enabling more intelligent and precise virtual intervention, recording the adjustment process, and obtaining virtual intervention measures.

[0087] By using a virtual space to assess virtual intervention measures and obtain pre-adjustment intervention strategies, feedback adjustments are made to the virtual intervention measures based on the obtained intervention assessment results to obtain fall prevention intervention measures.

[0088] It should be further explained that, in the specific implementation process, the intervention judgment means that after virtual intervention through the fall prevention bed model, the purpose of the virtual bed resting position is inquired. If the inquiry result is that the virtual bed resting position requires getting out of bed, the airbag device is closed after the virtual bed resting position is confirmed. The state of closing the airbag device is recorded as the pre-adjustment intervention strategy, which means that no further intervention adjustment is needed. This step is updated and recorded in the virtual intervention measures to obtain the fall prevention intervention measures, which are then applied to the elderly bedridden in reality. For the elderly who are initially judged to be in a fall-prone state, the intervention measures corresponding to the actual judgment results are continuously updated to provide timely intervention and care.

[0089] If the inquiry result does not indicate a need to get out of bed, then the abnormal identification state at this time is an unconscious tendency to fall out of bed in the virtual bed-sitting state. The intelligent electric airbag remains stationary, and the caregiver provides assistance in adjusting the virtual bed-sitting state to a comfortable and safe bed-sitting position. The physiological data of the bedridden elderly is continuously monitored through the motion capture device, and this process is recorded and added to the virtual intervention measures. The intervention measures for caregivers in the non-bed-sitting state are the fall prevention intervention measures.

[0090] The preferred embodiments of the present invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to specific implementations. Clearly, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention. The invention is limited only by the claims and their full scope and equivalents.

Claims

1. An intelligent fall-prevention bed risk management system based on artificial intelligence, comprising a management center, characterized in that, The management center is connected to a monitoring and acquisition module, a transmission and processing module, a body position analysis module, and a multi-dimensional monitoring module. The monitoring and acquisition module is used to collect bedridden elderly people's bedridden status data and build a bedridden data storage repository; The transmission processing module is used to construct a quiet virtual space for bedridden elderly people, and to virtually convert and mark the smart anti-fall bed through the quiet virtual space to obtain the limited bed boundaries. The process of obtaining the defined bed boundaries includes: The space for bedridden elderly is constructed to obtain a virtual resting space, and the smart anti-fall bed is mapped to the virtual resting space to obtain an anti-fall bed model. Based on the virtual resting space, the bedridden elderly are virtually placed to obtain a virtual bed posture. Upload the bed rest data repository to the bed rest virtual space, and use the bed rest data repository to mark the boundaries of the fall prevention bed model in the bed rest virtual space to obtain the defined bed boundaries. Mapping and identifying the bed rest data repository to obtain bed model marker location points, and then continuously marking these bed rest location points to obtain continuous bed rest tracks, the process includes: Obtain the bedridden data repository, classify and label the bedridden data repository, and obtain classification status data; The bed rest monitoring cycle is set according to the bed rest data repository. Based on the bed rest monitoring cycle, the obtained classification status data is matched and marked in the bed fall prevention model to obtain the marked position points of the bed model. Based on the bed rest monitoring cycle, the marked location points of the obtained bed model are linked over time to obtain the continuous movement trajectory of the bed rester. The body position analysis module is used to perform boundary distance statistics based on the continuous movement of the bedridden person, obtain the boundary movement difference sequence, perform digital visualization of the boundary movement difference sequence, obtain the real-time movement fluctuation map, divide the real-time movement fluctuation map into ranges and count the range proportions to obtain the curve judgment range. The multi-dimensional monitoring module is used to perform simulated diagnosis through a resting virtual space to obtain abnormal identification status. Based on the abnormal identification status in the resting virtual space, it dynamically adjusts the bedridden elderly to obtain virtual intervention measures, and adjusts and optimizes the virtual intervention measures to obtain fall prevention intervention measures.

2. The intelligent fall-prevention bed risk management system based on artificial intelligence according to claim 1, characterized in that, The process of collecting bedridden status data from elderly people includes: Data collection and deployment are carried out on the intelligent fall-prevention bed to obtain motion capture data. The motion capture device collects data on bedridden elderly people in real time to obtain bedridden status data, and timestamps the real-time collected bedridden status data to obtain the actual collection time point. A database of bedridden elderly people was constructed to obtain a bedridden data repository. Based on the actual collection time, the obtained bedridden status data was uploaded to the bedridden data repository.

3. The intelligent fall-prevention bed risk management system based on artificial intelligence according to claim 1, characterized in that, The process of obtaining a real-time trajectory fluctuation graph includes: By statistically analyzing the differences in continuous movement patterns on the bed within a virtual space defined by the bed boundary, the difference in movement patterns at the boundary positions is obtained. Based on the continuous movement patterns on the bed, the difference in movement patterns at the boundary positions is sequentially combined to obtain a sequence of boundary movement pattern differences. A two-dimensional rectangular coordinate system with respect to time is constructed. The boundary track difference sequence based on the continuous track of bed rest is uploaded to the two-dimensional rectangular coordinate system. Track difference change curve is generated according to the boundary track difference sequence, and the two-dimensional rectangular coordinate system containing the track difference change curve is marked as a real-time track fluctuation map.

4. The intelligent fall-prevention bed risk management system based on artificial intelligence according to claim 1, characterized in that, The process of dividing the real-time trajectory fluctuation graph into ranges and calculating the range percentages includes: Obtain a real-time movement fluctuation map, set an original reference axis for the real-time movement fluctuation map, upload the obtained original reference axis to the real-time movement fluctuation map, and perform movement monitoring on the real-time movement fluctuation map based on the obtained original reference axis to obtain the activity monitoring axis; The range of the real-time trajectory fluctuation graph is statistically analyzed based on the activity monitoring axis and the original baseline axis to obtain the curve determination range.

5. The intelligent fall-prevention bed risk management system based on artificial intelligence according to claim 1, characterized in that, The process of conducting simulated diagnosis through a virtual space while lying still includes: Based on the range determined by the curve, a safety risk statistic is calculated from the real-time trajectory fluctuation map to obtain the regional proportion result; By synchronously monitoring the virtual bed morphology in a quiet virtual space based on the regional proportion results, homomorphic change results are obtained. Anomaly diagnosis is performed on the obtained homomorphic change results to obtain anomaly identification status.

6. The intelligent fall-prevention bed risk management system based on artificial intelligence according to claim 1, characterized in that, The process of obtaining fall protection interventions includes: An intervention warning command is issued based on the anomaly identification status. The obtained intervention warning command is uploaded to the management center, which then sends the intervention warning command to the silent virtual space. The silent virtual space performs virtual intervention based on the anomaly identification status and obtains virtual intervention measures. By using a virtual space to assess virtual intervention measures and obtain pre-adjustment intervention strategies, feedback adjustments are made to the virtual intervention measures based on the obtained intervention assessment results to obtain fall prevention intervention measures.

Images