Intelligent old-age service platform operation method based on adaptive algorithm

Abstract

The invention relates to the technical field of smart old-age service, in particular to a smart old-age service platform operation method based on a self-adaptive algorithm, and the method comprises the steps: 1, collecting data in real time; 2, dynamically analyzing a service demand level; 3, quantifying the environment constraint, generating an environment influence factor, and calculating an indoor safety coefficient according to the indoor environment parameters; 4, executing a self-adaptive service decision; when the service demand level is emergency medical treatment and the environmental influence factor does not exceed the limit, starting on-site medical rescue; when the service demand level is emergency medical treatment and the environmental impact factor exceeds the limit, remote medical treatment and unmanned aerial vehicle distribution are activated; when the service demand level is daily nursing enhancement and the indoor safety coefficient reaches the standard, an enhanced indoor rehabilitation scheme is executed. The intelligent pension service platform not only can respond to the health requirements of the elderly in real time, but also can continuously optimize the service content, improve the comprehensiveness and adaptability of the pension service, and further promote the intelligent development of the pension industry.

Description

Technical Field

[0001] The present invention relates to the technical field of smart elderly care services, and in particular to a method for operating a smart elderly care service platform based on an adaptive algorithm. Background Art

[0002] With the development of smart elderly care service platforms, existing technologies mainly respond to the needs of the elderly through preset service templates. Typical implementation methods include: Static service mapping mechanism: Bind individual physiological indicators such as blood pressure and heart rate to fixed care levels, and trigger preset service packages when indicators are detected exceeding the threshold.

[0003] Environment-independent resource allocation: Service resources are allocated only based on the health status of the elderly, without considering the actual constraints of the external environment on service execution.

[0004] Data island processing mode: The physiological monitoring module, behavioral analysis module, and environmental perception module operate independently, and the output results of each subsystem are only used for local decision-making.

[0005] The above technologies expose the following serious problems in actual elderly care scenarios: (1) Static services lead to failure in responding to health mutations: Elderly individuals' health conditions can deteriorate suddenly (for example, a sudden stroke in a hypertensive patient), but existing platforms rely on periodic health assessments (typically daily). When a fall sensor triggers an alarm between assessments, the system still allocates basic care based on the pre-fall assessment results, failing to escalate to rehabilitation treatment based on real-time muscle tone loss data. A 2023 report from one nursing home revealed that this delayed response led to a 37% increase in the incidence of sequelae.

[0006] (2) Environmental factors causing large-scale service interruptions: Elderly care services are highly dependent on environmental conditions. Outdoor activities require real-time monitoring of PM2.5 concentrations (the tolerance threshold for elderly individuals with respiratory diseases is 30% lower than normal). Medical resource dispatch is impacted by traffic delays caused by rainfall (ambulance arrival times are extended by an average of 45 minutes during heavy rain). Existing technology lacks a quantitative correlation between environmental parameters and service availability, resulting in the cancellation of over 30% of outdoor services during extreme weather conditions. During community lockdowns, due to the lack of implementation of environmental adaptation plans (such as drone delivery instead of manual medication delivery), the medication non-availability rate for elderly individuals with chronic diseases reached as high as 25%.

[0007] Therefore, there is an urgent need for an operating method of a smart elderly care service platform based on an adaptive algorithm to solve the above problems. Summary of the Invention

[0008] Based on the above objectives, the present invention provides a method for operating a smart elderly care service platform based on an adaptive algorithm, comprising the following steps: Step 1: Real-time collection of the elderly’s physiological indicator data, behavioral activity data, and indoor and outdoor environmental parameter data; Step 2: Dynamically analyze the service demand level, assign weights based on the frequency of abnormal fluctuations in physiological indicators and the degree of behavioral activity deviation to generate a health index, and map the health index to different levels of service demand using personalized thresholds. Step 3: Quantify environmental constraints, generate environmental impact factors based on the historical impact patterns of outdoor environmental parameters, and calculate indoor safety factors based on indoor environmental parameters; Step 4: Execute adaptive service decision-making, inputting service demand level, environmental impact factors, and indoor safety factors into the decision-making model; When the service demand level is emergency medical and the environmental impact factors are within the limit, on-site medical rescue will be initiated; When the service demand level is emergency medical care and the environmental impact factor exceeds the limit, telemedicine and drone delivery are activated; When the service demand level is enhanced daily care and the indoor safety factor meets the standards, an enhanced indoor rehabilitation plan will be implemented.

[0009] Preferably, the step 1 comprises: Physiological indicator data include heart rate variability coefficient, body surface temperature gradient, and gait smoothness index, which are continuously collected at a fixed sampling frequency through wearable devices; Behavioral activity data is collected through an array of Bluetooth beacons deployed indoors to obtain location sequences, recording three-dimensional movement trajectories when the elderly stay in high-risk areas; The abnormal fluctuation identification method is to calculate the sliding standard deviation of physiological indicators over multiple consecutive sampling periods, compare it with the historical baseline value of the elderly person in the same period, and mark it as abnormal when it exceeds the set multiple of the historical baseline fluctuation range; The historical benchmark value is determined by analyzing the distribution of physiological indicator data of the elderly person in the same period of time within the past specific days, and the specific days are dynamically configured according to the severity of the elderly person's chronic disease.

[0010] Preferably, the method for adjusting the personalized threshold in step 2 includes: Obtain the opening records of the elderly's smart medicine box and calculate the ratio of the actual number of medications taken to the number of prescription requirements as the medication compliance rate; Establish a correlation model between medication adherence rate and the risk of health status mutation: when the adherence rate continues to fall below the set level, the emergency medical service trigger threshold is proportionally lowered; The threshold adjustment range is determined by historical data analysis: the average speed at which the health index deteriorates to a critical state under the same compliance rate level in the past is calculated, and a mapping rule is established based on the correlation between the deterioration speed and the threshold adjustment range.

[0011] Preferably, the method for generating the environmental impact factor in step 3 includes: Outdoor environmental parameters, including air quality index and precipitation probability, are obtained in real time from the meteorological service interface; Extract the correspondence between environmental parameter combinations and service interruption events from the platform's historical database. Service interruption events include the cancellation of outdoor activities and delayed arrival of medical resources. A machine learning algorithm is used to analyze the impact weights of different parameter combinations on service interruptions. The weight coefficient of the air quality index is determined by fitting its relationship with the frequency of respiratory disease attacks through regression analysis, and the weight coefficient of the precipitation probability is determined by analyzing the duration of traffic delays under different rainfall conditions.

[0012] Preferably, the method for determining the degree of deviation of the behavioral activity includes: Establishing a template for the elderly's daily behavior patterns: Analyzing historical location sequences through machine learning to extract thresholds for single bathroom stay duration and daily kitchen visit frequency. Real-time detection of behavioral anomalies: When the length of time spent in the bathroom exceeds the template threshold and the decrease in the gait stability index during the same period reaches a set ratio of the elderly person's historical fluctuation extreme value, a high-risk behavior flag is triggered; The determination benchmark value of the decline range of the gait stability index is calculated by calculating the dynamic range of the gait data of the elderly person during normal walking in the past week.

[0013] Preferably, when a resource conflict occurs in step 4, the method for resolving the conflict includes: When multiple elderly people trigger emergency medical services at the same time, calculate the change in the health index of each elderly person per unit time; Prioritize resource allocation rules: compare the difference between the rate of decline and the rate of increase of the health index, and prioritize the elderly with the largest difference; The health index change is calculated by the health index difference within a sliding time window, and the window length is dynamically set according to the historical median of the service response delay.

[0014] Preferably, the calculation method of the indoor safety factor in step 3 includes: Distributed pressure sensors placed on the ground are used to detect foot slip events. The slip event determination condition is: the local pressure change gradient exceeds the pressure change threshold under normal walking mode; The ground friction coefficient correction value is calculated based on the spatial distribution density of slip events. The higher the density, the greater the downward correction of the friction coefficient. By integrating the time series of indoor light intensity, when the duration of continuous low light exceeds the critical value of the elderly's visual adaptation ability, the safety factor is reduced in a step-by-step manner. The reduction range is determined by analyzing the correlation between historical fall events and light exposure duration.

[0015] Preferably, a closed-loop optimization step is further included to dynamically adjust the weight distribution in step 2 and the environmental impact factor reconstruction in step 3 according to the service execution effect, wherein the weight distribution adjustment method includes: Monitor the continuous deterioration trend of the health index before the start of the service. When the deterioration exceeds the expected control target of the service, increase the weight of the abnormal fluctuation frequency of physiological indicators in the calculation of the health index; Adjusting the weight of behavioral activity: Calculate the false alarm rate between behavioral risk indicators and actual health events. When the false alarm rate exceeds the platform's average false alarm level, reduce the weight of behavioral activity data in the health index. The weight adjustment range is dynamically calculated by establishing a mathematical model of health index prediction error and service parameter sensitivity.

[0016] Preferably, the method for reconstructing environmental impact factors includes: Collect the complete environment parameter combination and service execution results when the service fails, and establish a failure case library; Reconstructing the calculation model of environmental impact factors: Using an integrated learning algorithm to analyze the impact weights of sudden environmental variables in failure cases, focusing on enhancing the weights of parameter combinations that are not fully present in historical data; Model validation mechanism: The reconstructed model is applied to the simulated environment parameter combination. When the deviation between the predicted service failure rate and the actual failure rate exceeds the allowable error, the feature engineering strategy is readjusted.

[0017] Preferably, the activation method of the drone delivery includes: Real-time acquisition of community 3D building model data allows flight paths to be planned to avoid turbulent areas around high-rise buildings. Path safety is verified by analyzing the spatial distribution relationship between historical aircraft attitude data and wind speed. Drug temperature control guarantee: When the temperature sensor in the payload compartment detects that the temperature exceeds the drug storage threshold, the backup transportation plan is automatically switched. The switching logic includes: a: If the temperature control failure occurs before takeoff, activate the emergency transport vehicle equipped with cold chain equipment; b: If temperature control failure occurs during flight, command the drone to land at the nearest temperature control transfer station; The temperature storage threshold is dynamically set according to the drug chemical stability test data.

[0018] Beneficial effects of the present invention: 1. This invention dynamically analyzes service demand levels and combines real-time collected physiological indicator data, behavioral activity data, and environmental parameters to ensure timely adjustment of service responses to sudden changes in health status. In particular, in the event of abnormal fluctuations in physiological indicators and behavioral deviations, the calculation of a health index and dynamic adjustment of personalized thresholds enable rapid response based on real-time data.

[0019] 2. The present invention effectively quantifies the relationship between environmental parameters and service feasibility by introducing an environmental perception module and generating environmental impact factors based on historical data. Specifically, the platform monitors outdoor environmental parameters in real time and analyzes the impact on service execution based on environmental data. If the outdoor environment is not suitable for outdoor activities, the platform can enable alternative solutions such as telemedicine services or drone delivery of medicines to ensure that the elderly can still obtain necessary services in extreme weather or special environments. In addition, the reconstruction mechanism of environmental impact factors and machine learning algorithms further enhance the platform's adaptability to environmental changes and avoid large-scale service interruptions caused by environmental factors.

[0020] 3. This invention eliminates data silos by comprehensively processing the elderly's physiological, behavioral, and environmental data, enabling cross-module data fusion and collaborative decision-making. This cross-module collaborative analysis and decision-making not only improves the accuracy of health monitoring but also enables dynamic adjustment of service responses based on global data, ensuring the provision of precise and personalized elderly care services.

[0021] 4. This invention dynamically calculates the rate of change of the health index and prioritizes resource allocation based on the change in the elderly's health index, ensuring that those most in need of emergency services are prioritized when resources are limited. Through a sliding time window and dynamic adjustment mechanism, the platform can flexibly and in real time optimize resource allocation, avoiding service delays caused by resource conflicts and ensuring that every elderly person receives the services they need at the optimal time.

[0022] 5. This invention introduces a closed-loop optimization mechanism that dynamically adjusts the weight distribution and calculation model of environmental influencing factors during service execution based on service performance. Specifically, the platform monitors the changing trends of health indices. When it finds that service execution does not meet the expected control targets, it automatically adjusts the weights of abnormal fluctuations in physiological indicators and behavioral activity data. Through this self-adjustment mechanism, the platform can continuously optimize service strategies, improve service accuracy and responsiveness, and avoid problems such as poor service performance or service unsuitability for the elderly.

[0023] 6. This invention ensures that medications meet safety standards throughout transportation through drone delivery and temperature control measures. If temperature control fails during transportation, the system intelligently switches to a backup plan, ensuring timely and safe delivery under all circumstances. This mechanism effectively addresses medication delivery challenges during special circumstances, such as epidemic lockdowns, and reduces the risk of medication shortages for elderly individuals in these challenging circumstances. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0025] Figure 1 is a flow chart of the steps of the method of the present invention; Figure 2 A flowchart of the steps of the weight allocation adjustment method of the present invention; Figure 3 The figure is a flow chart of the steps of the method for reconstructing environmental impact factors according to the present invention. DETAILED DESCRIPTION

[0026] The present invention is described in detail below with reference to the accompanying drawings and specific embodiments. It is also noted that, to provide a more detailed description, the following embodiments are best and preferred embodiments, and those skilled in the art may employ alternative methods for implementing certain known technologies. Furthermore, the accompanying drawings are intended only to provide a more detailed description of the embodiments and are not intended to limit the present invention.

[0027] See Figure 1-Figure 3 The present invention provides a method for operating a smart elderly care service platform based on an adaptive algorithm, aiming to achieve dynamic, efficient, and environmentally adaptable personalized elderly care services. This method provides precise medical and nursing support for the elderly through real-time perception, multi-dimensional analysis, and adaptive response.

[0028] Specifically, wearable devices, smart home sensing devices, and camera recognition are used to collect physiological indicators such as heart rate, blood pressure, body temperature, blood oxygen, and blood sugar around the clock. Simultaneously, accelerometers, infrared sensors, and visual behavioral recognition are used to collect behavioral information such as walking frequency, range of activity, sitting and lying frequency, and circadian activity patterns. Furthermore, environmental sensing units are used to collect indoor temperature and humidity, air quality, and illumination, as well as external meteorological data such as temperature, air pollution index, and wind speed. This rich and timely data provides a comprehensive overview of the elderly's health and environmental status, providing a solid foundation for subsequent intelligent analysis.

[0029] Each physiological metric (such as heart rate and blood pressure) is compared to its historical normal range. If a value exceeds an individualized threshold multiple times within a short period of time, it is considered an "abnormal fluctuation." The number of fluctuations and their magnitude are used together as indicators to form an abnormality scoring matrix.

[0030] A reference behavior trajectory is established using historical behavioral activity models (such as average daily steps, active time, and time in bed). If current behavior deviates from the standard by more than a set threshold, it is considered abnormal. For example, a decrease in exercise volume of more than 40% for three consecutive days, or a disruption in circadian rhythms, triggers a deviation alert.

[0031] A weighted fusion algorithm combines physiological abnormality scores and behavioral deviation scores with a set weight (e.g., 6:4) to generate a comprehensive health index. The weight ratio can be dynamically adjusted based on individual health characteristics. For example, patients with chronic diseases may place greater emphasis on physiological fluctuations, while those recovering from hemiplegia may emphasize behavioral patterns.

[0032] Set multiple threshold intervals (such as 0-30 for safe status, 30-60 for observation status, 60-80 for enhanced care, and above 80 for emergency medical treatment), and map the health index to the corresponding service level.

[0033] This method fully integrates health status assessment at the physiological and behavioral levels. Compared with traditional static physical examination indicators, it is more in line with the risk change trends in the daily lives of the elderly and has highly personalized and dynamic response capabilities.

[0034] Machine learning algorithms are used to statistically model the impact of specific environmental conditions (such as heavy pollution and high temperatures) on physiological abnormalities in the elderly, generating an "environmental impact factor." This factor represents the feasibility and risk level of performing a specific service (such as outpatient medical care or home care) under current environmental conditions. Simultaneously, a comprehensive indoor safety factor is calculated based on real-time indoor environmental data. For example, a higher score indicates suitability for indoor rehabilitation activities when air quality is good, temperature and humidity are appropriate, and sunlight is sufficient.

[0035] The introduction of environmental impact factors and safety factors enables service decisions to fully consider the actual impact of external natural conditions, thereby improving the safety and execution efficiency of elderly care services.

[0036] The service demand level, environmental impact factor, and indoor safety factor obtained in the previous step are used as input variables, and a joint decision-making mechanism based on a rule base and an optimization algorithm is used to determine the optimal service strategy: If it is an emergency medical treatment and the external environment is suitable, immediately call nearby medical resources for on-site first aid.

[0037] If it is an emergency medical treatment and the external environment is dangerous, a remote medical video connection will be triggered and a drone will be called to quickly deliver first aid equipment and medicines to the scene.

[0038] If it is the intensive care stage and the indoor conditions meet the standards, the home rehabilitation plan will be automatically activated, including intelligent exercise guidance, dietary adjustments and spiritual comfort plans.

[0039] The decision-making mechanism can quickly switch service modes based on real-time changes in multiple factors, ensuring that the elderly can receive timely, accurate and safe service responses even in emergencies or bad weather.

[0040] This approach forms a closed-loop elderly care service process through a highly integrated perception, analysis, and response mechanism, combined with the generation of a dynamically weighted health index, environmental adaptability assessment, and adaptive decision-making logic. The key lies in dynamically mapping service needs through a data-driven approach, avoiding the delayed response and crude intervention issues of traditional models. This significantly enhances the intelligence and security capabilities of elderly care services.

[0041] In one possible implementation, the physiological indicators include heart rate variability coefficient, body surface temperature gradient, and gait smoothness index. These physiological indicators can effectively reflect the physical health status and risk changes of the elderly.

[0042] Among them, the heart rate variability coefficient: continuously collects heart rate data through wearable devices (such as heart rate monitors) and calculates the variability of heart rate. Heart rate variability (HRV) reflects the state of the autonomic nervous system of the elderly and can warn of potential heart disease or stress response.

[0043] Surface temperature gradient: Real-time monitoring of surface temperature changes through devices such as smart bracelets. By calculating the surface temperature gradient (for example, the rate of temperature fluctuation), possible infections, temperature regulation problems, or other health abnormalities can be identified.

[0044] Gait stability index: Use a device equipped with an accelerometer to track the gait of the elderly and analyze the stability and coordination of their steps to determine whether there is a risk of falling, gait abnormalities, and other problems.

[0045] These data are continuously collected by wearable devices at a fixed sampling frequency (for example, every second or every minute) to ensure real-time and accuracy, facilitating subsequent health monitoring and analysis.

[0046] Behavioral activity data is collected through an array of Bluetooth beacons deployed indoors. These beacons can accurately locate the elderly person's location, generating a sequence of locations that can be used to identify their behavior patterns and activity areas.

[0047] By deploying Bluetooth beacons, the system can identify in real time whether the elderly have entered preset high-risk areas (such as kitchens, bathrooms, etc.), and record their three-dimensional movement trajectories when they stay in these areas to ensure that activities in dangerous areas are monitored and intervened in a timely manner.

[0048] These data can reflect the spatial distribution of elderly people's activities in real time, helping to provide early warnings and take appropriate actions when the elderly enter dangerous areas.

[0049] The abnormal fluctuation identification method aims to detect potential health problems in a timely manner by monitoring fluctuations in physiological indicators.

[0050] First, the physiological indicator data for multiple consecutive sampling periods is extracted and its sliding standard deviation is calculated. This is to calculate the degree of fluctuation of the physiological indicator within a fixed time window. The sliding standard deviation can reflect the degree of fluctuation of the physiological indicator within that time period, thereby assessing the stability of the elderly person's physiological state.

[0051] Next, we compare historical baseline values. These values ​​are derived by analyzing the distribution of physiological indicator data for the elderly over the same period of time over a specific number of days. Specifically, these historical data represent the normal range of physiological fluctuations for elderly individuals in different health states. For example, if a patient's heart rate variability ranged from 20 to 40 over the same period of time over the past three days, fluctuations within this range are considered normal.

[0052] If the sliding standard deviation of a physiological indicator exceeds a set multiple of the historical baseline fluctuation range (for example, 1.5 or 2 times), it is marked as an abnormal fluctuation and an alert is triggered. This method can dynamically adapt to the individual differences of each elderly person by comparing it with historical data.

[0053] Historical baseline values ​​are dynamically configured based on the severity of the elderly's chronic conditions. For elderly individuals in relatively good health, the historical baseline value calculation may consider data from a shorter period of time. For those with chronic conditions, the reference period of historical data may be appropriately extended (e.g., 7 or 14 days) to better accommodate fluctuations in their physiological state.

[0054] This method can not only provide comprehensive health monitoring and real-time intervention for the elderly, but also can be flexibly adjusted according to individual differences to provide the elderly with more personalized and intelligent health management services.

[0055] In one possible implementation, the system automatically captures the elderly person's daily medication usage data by accessing the smart pill box's opening history. The smart pill box accurately records every opening, including a timestamp, and compares the data with the daily dose prescribed by the doctor.

[0056] Medication compliance rates are calculated by comparing prescription instructions with the actual number of times a patient opens the medication box. For example, if a doctor prescribes three doses of medication daily and the patient opens the medication box twice, the compliance rate for that day is 2 / 3, which is approximately 66.7%.

[0057] The compliance rate for multiple consecutive days can be calculated by sliding average to further eliminate occasional errors and improve assessment accuracy.

[0058] This method has the advantages of being non-invasive and automated, can accurately monitor medication behavior, and is the basis for subsequent health risk assessment.

[0059] Based on the results of long-term tracking of elderly people's health data, it is found that there is a significant correlation between medication compliance rate and sudden changes in health status. This method constructs a quantitative correlation model between compliance rate and sudden health events through historical data analysis: Set a risk threshold. For example, if the compliance rate is lower than 80%, the system will determine it as a potential risk state. If the compliance rate is lower than the set value for several consecutive days, the elderly person is considered to be in a high-risk period, and the trigger threshold for emergency medical services is proportionally lowered to increase the sensitivity of health warnings; The magnitude of the reduction in the ratio is linearly or nonlinearly related to the degree of deviation in the compliance rate. For example, when the compliance rate is below 60%, the trigger threshold can be lowered by 20%; when it is below 40%, the reduction is increased to 40%.

[0060] In this way, the daily medication behavior of the elderly is linked with the medical response mechanism to form a closed-loop feedback.

[0061] In order to ensure the scientificity and accuracy of threshold adjustment, this method further uses data-driven means to quantitatively design the adjustment range: Summarize historical health data of a large number of elderly people at different levels of compliance, with a particular focus on samples whose status deteriorated from stable to critical; These samples were analyzed for the rate of deterioration of their health indices (e.g., dramatic fluctuations in heart rate, blood pressure, and blood sugar within a few hours) to extract the statistical correlation between adherence rate and the rate of health deterioration; Establish a set of mapping rules or functional relationships that reflect the extent to which the original trigger threshold should be lowered at a specific level of adherence. This mapping rule can be derived through regression analysis or machine learning methods and regularly updated to reflect the latest clinical data.

[0062] This design ensures that threshold adjustment is no longer a fixed parameter, but a dynamic mechanism with learning ability and individual adaptability.

[0063] In one possible implementation, the platform first obtains real-time outdoor environmental parameters through a meteorological service interface. These parameters include the Air Quality Index (AQI) and precipitation probability. The AQI is a key indicator of air pollution levels, typically derived by monitoring the concentrations of various pollutants in the atmosphere. The precipitation probability indicates the likelihood of precipitation within a certain period of time. These two environmental parameters have a significant impact on smart elderly care services, particularly in the scheduling of outdoor activities and the dispatch of medical resources.

[0064] Secondly, the platform extracts the correlation between environmental parameter combinations and service disruption events from its historical database. This historical data helps the platform identify the impact of environmental changes on smart elderly care services. Specifically, service disruption events include the cancellation of outdoor activities and delayed arrival of medical resources. By analyzing historical data, it can be determined which situations, under specific environmental parameter combinations, are more likely to lead to service disruptions, thus providing a basis for subsequent intervention measures.

[0065] During the analysis, machine learning algorithms were used to process the combination of these environmental parameters and their weighted impact on service disruptions. The relationship between the air quality index and the frequency of respiratory illness attacks was fitted through regression analysis, deriving a weighted coefficient for the impact of the air quality index on the frequency of respiratory illness attacks. The weighted coefficient for the impact of precipitation probability was determined by analyzing the duration of traffic delays under different rainfall conditions, as these delays can directly affect the timely arrival of medical resources.

[0066] Through these steps, the platform can intelligently analyze and predict the likelihood of service interruptions based on real-time environmental data, providing more precise scheduling and adjustment measures for smart elderly care services. This approach can effectively reduce service interruptions caused by external environmental factors, improve service continuity and timeliness, ensure the smooth progress of elderly people's daily lives and medical needs, and enhance the platform's overall service quality and user satisfaction.

[0067] In one possible implementation, machine learning is used to analyze the elderly individual's historical location sequence data to create a daily behavior pattern template. This template aims to identify the elderly individual's normal daily activity patterns, focusing specifically on their behavioral characteristics in specific areas. By analyzing this historical data, key behavioral parameters can be extracted, such as a threshold for the duration of a single bathroom visit and a threshold for the daily frequency of visits to the kitchen area. These thresholds are automatically calculated based on the elderly individual's daily activity habits and accurately reflect their normal lifestyle patterns.

[0068] When the platform monitors in real time that the elderly's behavior deviates from the values ​​set in these templates, it triggers the detection mechanism for abnormal behavior. For example, if the time an elderly person stays in the bathroom exceeds the threshold set by the historical template, and the gait stability index shows a significant decrease at this time, this indicates that the elderly person may have an abnormality and immediate intervention measures may be needed. The gait stability index is a key factor in measuring the stability of the elderly when walking. When the gait stability index decreases by a set proportion of the elderly person's historical fluctuation extremes, the system will identify it as a high-risk behavior indicator and remind relevant personnel to pay further attention.

[0069] The gait stability index is determined by analyzing the elderly person's gait data from normal walking over the past week and calculating its dynamic range. Based on this, the system can determine whether there are significant abnormal fluctuations in the elderly person's current gait, thereby promptly identifying possible falls, movement disorders, or other health issues.

[0070] This method of determining the degree of behavioral deviation can help the platform identify health risks for seniors through intelligent analysis and real-time detection, especially those that are latent and haven't yet manifested themselves, such as falls or health emergencies. By promptly detecting abnormal behavior, the platform can respond quickly and take preventative measures, reducing the likelihood of accidents and improving the safety and health of seniors.

[0071] In one possible implementation, a smart elderly care service platform based on an adaptive algorithm can create resource conflicts when multiple elderly individuals simultaneously trigger emergency medical service requests. To address this, the platform uses a health index change calculation method to effectively resolve conflicts and allocate resources appropriately.

[0072] First, when multiple seniors request emergency medical services simultaneously, the platform calculates the change in each senior's health index per unit time. The health index reflects the senior's overall health status and typically includes a comprehensive assessment of physiological indicators such as heart rate, blood pressure, and body temperature. The change per unit time indicates fluctuations in the health index over a specific period, specifically the rate at which the index rises and falls. By calculating these changes, we can intuitively understand the fluctuations in each senior's health status and determine which patients should receive priority treatment.

[0073] The priority allocation rule to resolve resource conflicts is based on the difference between the rate of decline and the rate of improvement of each elderly individual's health index. Specifically, the platform compares the difference between the rate of decline and the rate of improvement of each elderly individual's health index and prioritizes resource allocation for those with the largest difference. The larger the difference, the more dramatic the change in the elderly individual's health status and the more urgent their need for medical resources, thus receiving priority treatment. This prioritized allocation method ensures the rationality of resource allocation and maximizes the safety of the elderly.

[0074] The change in the health index is calculated by taking the difference in the health index within a sliding time window. The length of the sliding window is dynamically set based on the median of the service's response latency history. This allows the platform to flexibly adjust the window size to accommodate the response speed requirements of different services. Shorter time windows enable faster responses to emergencies, while longer time windows smooth out health index fluctuations and better reflect changing health trends.

[0075] This prioritized resource allocation method ensures that those in critical condition receive priority treatment in emergencies by monitoring their health changes in real time, avoiding delays caused by resource conflicts. It also improves the efficiency of medical resource utilization, optimizes service quality, and enhances the health of the elderly.

[0076] In one possible implementation, indoor safety assessment is a crucial step in ensuring the reliability of the elderly's living environment during the operation of smart elderly care services. This method dynamically calculates indoor safety factors by fusing multi-source sensor data, enabling real-time assessment and intervention of potential risks.

[0077] First, distributed pressure sensors installed on the ground perform the core data collection function. Whenever an elderly person walks, the pressure distribution between the soles of their feet and the ground is continuously recorded and analyzed. If the pressure gradient in a specific area suddenly increases over a short period of time and exceeds the threshold for normal walking, it is identified as a slip event. This determination is based on the continuity and directionality of pressure during walking, as slip events are often accompanied by discontinuous, abnormal fluctuations.

[0078] Once a slip event is identified, the system calculates its spatial distribution density. Frequent slips in a specific area indicate insufficient friction in that area. Based on the slip density, the system then adjusts the floor friction coefficient downward for that area, with the adjustment increasing as the slip density increases. This adjustment mechanism allows the system to dynamically reflect environmental changes, such as increased slip risk caused by detergent residue or aging floors.

[0079] In addition to ground factors, indoor light intensity is also an important variable in assessing safety. The system analyzes the duration of low-light conditions by collecting time series of light intensity. When this duration exceeds the critical value of the visual adaptation ability of a specific elderly person (this value is obtained by individual visual ability assessment), the safety factor will be reduced in a step-by-step manner. The reduction rate is not fixed, but is derived by modeling the correlation between historical fall events and lighting conditions. For example, in a corridor area at night, if statistics show that the probability of falling in a low-light environment increases significantly, the reduction rate of the safety factor under low-light conditions here will be increased accordingly.

[0080] This approach has multiple beneficial effects. First, it can reflect the actual impact of the environment on elderly safety in real time, providing timely warnings, especially in high-risk scenarios. Second, it improves the accuracy of risk assessment by integrating personalized visual adaptation parameters with historical behavioral data. Third, the dynamic correction of the friction coefficient increases sensitivity to changes in ground conditions, which helps facilitate facility maintenance decisions. Overall, this approach significantly improves accident prevention capabilities and service accuracy in smart elderly care environments.

[0081] In one possible implementation, during the operation of the smart elderly care service platform, a closed-loop optimization process ensures continuous optimization and dynamic adaptation of services by monitoring and providing feedback on service performance. This process aims to improve the accuracy and real-time performance of health management by continuously adjusting the weights of different factors and reconstructing environmental influencing factors.

[0082] First, for the calculation of the health index, the platform will monitor the changing trends in the elderly's health status, especially the continued deterioration of the health index. When the health index shows an abnormal deterioration trend, and the deterioration exceeds the expected control target, the platform will increase the weight of the frequency of abnormal fluctuations in physiological indicators in the health index calculation. This is because when the health status deteriorates, abnormal fluctuations in physiological indicators are often an important early warning signal, which can reflect potential problems in the elderly's body at an early stage. Therefore, by increasing this weight, health changes can be captured more sensitively and the response speed of health management can be improved.

[0083] Secondly, to adjust the weighting of behavioral activity data, the platform calculates the false alarm rate between behavioral risk indicators and actual health events (such as falls and sudden illnesses). If the platform's false alarm rate exceeds the platform's average false alarm level, it indicates that the behavioral risk indicator may be overly sensitive. In this case, the platform will reduce the weight of behavioral activity data in the health index calculation. This adjustment can prevent excessive false alarms from distorting health assessment results, thereby ensuring more accurate and reliable health assessments.

[0084] The magnitude of these weight adjustments isn't static; rather, they're dynamically calculated based on a mathematical model that correlates health index prediction errors with service parameter sensitivity. Through continuous model training and optimization, the platform can precisely adjust the weights of different parameters based on real-time feedback, thereby improving the accuracy and adaptability of the entire service.

[0085] This closed-loop optimization can respond to health changes in real time, ensuring that the services provided by the platform always meet the health needs of the elderly; secondly, by dynamically adjusting weights and factors, the platform can make adaptive adjustments under different health conditions, avoiding service deviations caused by fixed strategies; finally, the weight adjustment method improves the accuracy of the service, reduces the false alarm rate, and improves the safety and health level of the elderly's daily life.

[0086] In one possible implementation, during the operation of a smart elderly care service platform based on an adaptive algorithm, environmental impact factor reconstruction improves the accuracy and robustness of services by continuously optimizing and adjusting the influence weights of environmental parameter combinations.

[0087] Specifically, after each service failure, the platform collects the complete environmental parameter combination and service execution results at the time, creating a failure case library. These failure cases include situations where the service execution did not achieve the expected results under different environmental conditions. By recording this data, the platform can form a detailed failure case library to facilitate subsequent analysis and model reconstruction. These cases provide valuable historical data for the subsequent reconstruction of environmental impact factors.

[0088] Next, the platform uses an ensemble learning algorithm to analyze the failure case database and reconstruct a calculation model for environmental influencing factors. Ensemble learning algorithms combine the predictions of multiple models, reducing the bias and error of a single model. During this process, the algorithm specifically focuses on combinations of environmental parameters that are underrepresented in historical data and increases the weighting of these combinations. This means that the platform dynamically adjusts the emphasis on these rare but critical factors, ensuring more accurate predictions of service success and failure in similar environments in the future.

[0089] The reconstructed environmental impact factor calculation model will be verified through a model validation mechanism. This validation process involves applying the new model to simulated environmental parameter combinations, predicting service failure rates, and comparing the predicted results with the actual failure rates. If the deviation between the predicted and actual failure rates exceeds the allowable error range, the platform will readjust the feature engineering strategy. This validation mechanism ensures that the reconstructed model can fully reflect the impact of environmental factors on service outcomes in real-world applications, thereby ensuring its accuracy and practicality.

[0090] By establishing a failure case library, the platform can learn from historical service failures, accumulate experience, and improve the system's adaptability. Secondly, the application of ensemble learning algorithms enables the platform to integrate multiple aspects of information, improving the stability and accuracy of predictions. Finally, the establishment of a model validation mechanism ensures continuous optimization of the model, allowing the platform to maintain high service quality and accuracy in an ever-changing environment. This process enables the platform to better cope with complex and changing environmental conditions, thereby providing more reliable and secure smart elderly care services.

[0091] In one possible implementation, in a smart elderly care service platform based on an adaptive algorithm, the activation method for drone delivery is a key component to ensure that the delivery process is efficient and safe.

[0092] Specifically, the platform first acquires real-time 3D building model data for the community. This data helps construct a detailed, three-dimensional map of the community environment. Based on this data, the platform can plan flight paths to avoid aircraft entering turbulent areas around high-rise buildings during flight. Turbulent areas around high-rise buildings often experience unstable air flow, which can affect drone flight stability. Therefore, avoiding these areas is essential to ensure flight safety. During path planning, the platform verifies the safety of the planned path by analyzing historical aircraft attitude data and the spatial distribution of wind speed. This analysis can identify potential risk areas based on actual flight data, thereby further optimizing flight routes.

[0093] In addition to flight path planning, ensuring the temperature control of medicines is a crucial step in drone delivery. Medicines are often very sensitive to temperature, and ensuring they remain within the appropriate temperature range during delivery is a key design priority for the platform. When the temperature sensor in the payload compartment detects that the temperature exceeds the threshold for storing medicines, the platform automatically initiates a backup transport plan. At this point, the system will take different emergency measures based on the time when the temperature control fails: If a temperature control failure occurs before takeoff, the platform will dispatch an emergency transport vehicle equipped with cold chain equipment to deliver the medicine. This strategy ensures that the medicine is effectively temperature-controlled during transportation, preventing temperature fluctuations that could cause the medicine to lose its effectiveness.

[0094] If temperature control failure occurs during flight, the platform will direct the drone to land at the nearest temperature control transfer station. This measure ensures that the medicines are protected by temperature control immediately, avoiding damage caused by temperature exceeding the range.

[0095] It's worth noting that the temperature thresholds for drug storage are dynamically set based on the drug's chemical stability test data. Since drug stability is closely linked to temperature, the platform sets specific temperature ranges for different drugs based on these test data, ensuring that each batch of drugs is shipped under optimal conditions.

[0096] By acquiring 3D building model data and historical flight data in real time, the platform can effectively plan flight routes and ensure the safety of the delivery process. Secondly, the drug temperature control assurance plan can be flexibly adjusted under different environmental changes to ensure temperature stability during drug transportation, thereby improving the reliability of delivery services. Finally, the temperature preservation threshold is dynamically set and adjusted according to the stability of the drug, so that each delivery can make the most appropriate decision based on actual needs, improving the accuracy and safety of delivery.

[0097] The present invention encompasses any alternatives, modifications, equivalents, and solutions that fall within the spirit and scope of the present invention. To provide a thorough understanding of the present invention, specific details are described in detail below in connection with the preferred embodiments of the present invention, but those skilled in the art will be able to fully understand the present invention without these detailed descriptions. Furthermore, to avoid unnecessary confusion regarding the essence of the present invention, well-known methods, processes, procedures, components, and circuits have not been described in detail.

[0098] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A method for operating a smart elderly care service platform based on an adaptive algorithm, characterized in that: The following steps are involved: Step 1: Real-time collection of the elderly’s physiological indicator data, behavioral activity data, and indoor and outdoor environmental parameter data; Step 2: Dynamically analyze the service demand level, assign weights based on the frequency of abnormal fluctuations in physiological indicators and the degree of behavioral activity deviation to generate a health index, and map the health index to different levels of service demand using personalized thresholds. Step 3: Quantify environmental constraints, generate environmental impact factors based on the historical impact patterns of outdoor environmental parameters, and calculate indoor safety factors based on indoor environmental parameters; Step 4: Execute adaptive service decision-making, inputting service demand level, environmental impact factors, and indoor safety factors into the decision-making model; When the service demand level is emergency medical and the environmental impact factors are within the limit, on-site medical rescue will be initiated; When the service demand level is emergency medical care and the environmental impact factor exceeds the limit, telemedicine and drone delivery are activated; When the service demand level is enhanced daily care and the indoor safety factor meets the standards, an enhanced indoor rehabilitation plan will be implemented.

2. The method for operating a smart elderly care service platform based on an adaptive algorithm according to claim 1, characterized in that: The step 1 comprises: Physiological indicator data include heart rate variability coefficient, body surface temperature gradient, and gait smoothness index, which are continuously collected at a fixed sampling frequency through wearable devices; Behavioral activity data is collected through an array of Bluetooth beacons deployed indoors to obtain location sequences, recording three-dimensional movement trajectories when the elderly stay in high-risk areas; The abnormal fluctuation identification method is to calculate the sliding standard deviation of physiological indicators over multiple consecutive sampling periods, compare it with the historical baseline value of the elderly person in the same period, and mark it as abnormal when it exceeds the set multiple of the historical baseline fluctuation range; The historical benchmark value is determined by analyzing the distribution of physiological indicator data of the elderly person in the same period of time within the past specific days, and the specific days are dynamically configured according to the severity of the elderly person's chronic disease.

3. The method for operating a smart elderly care service platform based on an adaptive algorithm according to claim 1, characterized in that: The method for adjusting the personalized threshold in step 2 includes: Obtain the opening records of the elderly's smart medicine box and calculate the ratio of the actual number of medications taken to the number of prescription requirements as the medication compliance rate; Establish a correlation model between medication adherence rate and the risk of health status mutation: when the adherence rate continues to fall below the set level, the emergency medical service trigger threshold is proportionally lowered; The threshold adjustment range is determined by historical data analysis: the average speed at which the health index deteriorates to a critical state under the same compliance rate level in the past is calculated, and a mapping rule is established based on the correlation between the deterioration speed and the threshold adjustment range.

4. The method for operating a smart elderly care service platform based on an adaptive algorithm according to claim 1, characterized in that: The method for generating the environmental impact factor in step 3 includes: Outdoor environmental parameters, including air quality index and precipitation probability, are obtained in real time from the meteorological service interface; Extract the correspondence between environmental parameter combinations and service interruption events from the platform's historical database. Service interruption events include the cancellation of outdoor activities and delayed arrival of medical resources. A machine learning algorithm is used to analyze the impact weights of different parameter combinations on service interruptions. The weight coefficient of the air quality index is determined by fitting its relationship with the frequency of respiratory disease attacks through regression analysis, and the weight coefficient of the precipitation probability is determined by analyzing the duration of traffic delays under different rainfall conditions.

5. The method for operating a smart elderly care service platform based on an adaptive algorithm according to claim 2, characterized in that: The method for determining the degree of deviation of the behavioral activity includes: Establishing a template for the elderly's daily behavior patterns: Analyzing historical location sequences through machine learning to extract thresholds for single bathroom stay duration and daily kitchen visit frequency. Real-time detection of behavioral anomalies: When the length of time spent in the bathroom exceeds the template threshold and the decrease in the gait stability index during the same period reaches a set ratio of the elderly person's historical fluctuation extreme value, a high-risk behavior flag is triggered; The determination benchmark value of the decline range of the gait stability index is calculated by calculating the dynamic range of the gait data of the elderly person during normal walking in the past week.

6. The method for operating a smart elderly care service platform based on an adaptive algorithm according to claim 1, characterized in that: When a resource conflict occurs in step 4, methods for resolving the conflict include: When multiple elderly people trigger emergency medical services at the same time, calculate the change in the health index of each elderly person per unit time; Prioritize resource allocation rules: compare the difference between the rate of decline and the rate of increase of the health index, and prioritize the elderly with the largest difference; The health index change is calculated by the health index difference within a sliding time window, and the window length is dynamically set according to the historical median of the service response delay.

7. The method for operating a smart elderly care service platform based on an adaptive algorithm according to claim 1, characterized in that: The calculation method of the indoor safety factor in step 3 includes: Distributed pressure sensors placed on the ground are used to detect foot slip events. The slip event determination condition is: the local pressure change gradient exceeds the pressure change threshold under normal walking mode; The ground friction coefficient correction value is calculated based on the spatial distribution density of slip events. The higher the density, the greater the downward correction of the friction coefficient. By integrating the time series of indoor light intensity, when the duration of continuous low light exceeds the critical value of the elderly's visual adaptation ability, the safety factor is reduced in a step-by-step manner. The reduction range is determined by analyzing the correlation between historical fall events and light exposure duration.

8. The method for operating a smart elderly care service platform based on an adaptive algorithm according to claim 4, characterized in that: It also includes a closed-loop optimization step, dynamically adjusting the weight distribution of step 2 and the environmental impact factor reconstruction of step 3 according to the service execution effect, wherein the weight distribution adjustment method includes: Monitor the continuous deterioration trend of the health index before the start of the service. When the deterioration exceeds the expected control target of the service, increase the weight of the abnormal fluctuation frequency of physiological indicators in the calculation of the health index; Adjusting the weight of behavioral activity: Calculate the false alarm rate between behavioral risk indicators and actual health events. When the false alarm rate exceeds the platform's average false alarm level, reduce the weight of behavioral activity data in the health index. The weight adjustment range is dynamically calculated by establishing a mathematical model of health index prediction error and service parameter sensitivity.

9. The method for operating a smart elderly care service platform based on an adaptive algorithm according to claim 8, characterized in that: The method for reconstructing the environmental impact factors includes: Collect the complete environment parameter combination and service execution results when the service fails, and establish a failure case library; Reconstructing the calculation model of environmental impact factors: Using an integrated learning algorithm to analyze the impact weights of sudden environmental variables in failure cases, focusing on enhancing the weights of parameter combinations that are not fully present in historical data; Model validation mechanism: The reconstructed model is applied to the simulated environment parameter combination. When the deviation between the predicted service failure rate and the actual failure rate exceeds the allowable error, the feature engineering strategy is readjusted.

10. The method for operating a smart elderly care service platform based on an adaptive algorithm according to claim 4, characterized in that: The activation method of the drone delivery includes: Real-time acquisition of community 3D building model data allows flight paths to be planned to avoid turbulent areas around high-rise buildings. Path safety is verified by analyzing the spatial distribution relationship between historical aircraft attitude data and wind speed. Drug temperature control guarantee: When the temperature sensor in the payload compartment detects that the temperature exceeds the drug storage threshold, the backup transportation plan is automatically switched. The switching logic includes: a: If the temperature control failure occurs before takeoff, activate the emergency transport vehicle equipped with cold chain equipment; b: If temperature control failure occurs during flight, command the drone to land at the nearest temperature control transfer station; The temperature storage threshold is dynamically set according to the drug chemical stability test data.

Images