Personalized physiological abnormality monitoring method, system and electronic device based on non-contact sensing

Abstract

The application relates to a personalized physiological abnormality monitoring method, system and electronic equipment based on non-contact sensing. The method comprises the following steps: acquiring real-time physiological parameter data of a target user provided by a non-contact sensing device; determining an individual baseline of the target user based on effective resting data in the past preset time length; calling age information and health condition labels of the target user, and generating upper and lower threshold values of the physiological parameters in combination with the individual baseline; monitoring real-time physiological data of the target user, and if it is confirmed that the real-time physiological data exceeds the upper and lower threshold values and that the real-time physiological data meets a preset continuous over-limit condition, an abnormality alarm signal is triggered. The application helps to realize accurate and intelligent monitoring of the physiological state of the elderly, effectively improves the accuracy and stability, and improves the safety guarantee level and emergency response efficiency in the intelligent elderly care scene.

Description

Technical Field

[0001] This invention relates to the field of smart elderly care technology, and more specifically, to a personalized physiological abnormality monitoring method, system, and electronic device based on non-contact sensing. Background Technology

[0002] With the increasing popularity of the concept of smart elderly care, intelligent health monitoring devices are being used more and more widely in home-based and community-based elderly care scenarios. Non-contact physiological monitoring technology, due to its characteristics of not requiring wearing and not affecting users' daily activities, is gradually becoming a research hotspot and important development direction in the field of smart elderly care. In existing technologies, smart chairs, smart beds, and other devices integrate non-contact sensing units such as radar and pressure sensors to continuously collect basic physiological parameters such as heart rate, respiratory rate, and blood oxygen saturation, providing convenience for the daily health monitoring of the elderly.

[0003] In the process of realizing this invention, the inventors discovered that existing non-contact physiological monitoring technologies have significant limitations. They typically use generalized and static abnormality judgment thresholds, ignoring the significant differences in physiological characteristics among different individuals, especially elderly people of different ages and with different underlying disease conditions. This leads to persistently high false alarm and false negative rates in the monitoring results. This not only reduces the clinical reliability of the system but may also lead to ineffective allocation of nursing resources or delayed handling of real risks, making it difficult to meet the core needs of modern smart elderly care for accurate identification and reliable early warning of physiological abnormalities. Summary of the Invention

[0004] Based on this, and in response to the above problems, the present invention provides a personalized physiological abnormality monitoring method, system, and electronic device based on non-contact sensing, thereby solving the problems mentioned in the background art.

[0005] In a first aspect, the present invention provides a personalized physiological abnormality monitoring method based on non-contact sensing, applied to the embedded control system of a smart elderly care chair. The method includes: acquiring real-time physiological parameter data of a target user provided by a non-contact sensing device; determining an individualized baseline for the target user based on effective resting data of a preset duration in the past; wherein the individualized baseline is a statistical model established based on physiological parameter data of the target user in a historical resting state; retrieving the target user's age information and health status label, and combining them with the individualized baseline to generate upper and lower limit thresholds for physiological parameters; monitoring the target user's real-time physiological data, and if it is confirmed that the real-time physiological data exceeds the upper and lower limit thresholds, and if it is confirmed that the real-time physiological data meets preset continuous over-limit conditions, then triggering an abnormality alarm signal.

[0006] Optionally, in this embodiment of the invention, before determining the individualized baseline of the target user based on valid resting data within a preset time period, the method further includes: receiving the limb activity intensity of the target user provided by an accelerometer; determining that the target user is in a preliminary resting state when the limb activity intensity is lower than a preset activity threshold; further detecting the ambient light intensity and electromagnetic interference level in the preliminary resting state; and marking the currently obtained physiological parameter data as valid resting data and including it in the dataset within the preset time period only when both the ambient light intensity and electromagnetic interference level meet preset environmental quality conditions. By using an accelerometer to determine the limb activity intensity to screen for the preliminary resting state, and combining this with dual verification of ambient light intensity and electromagnetic interference level, only physiological parameter data that meets the conditions is marked as valid resting data and included in the dataset. This effectively eliminates the influence of unstable factors such as user limb movement and environmental interference on baseline modeling, further improving the purity and reliability of the valid resting data. This allows the constructed individualized baseline to more accurately reflect the user's basic physiological state, providing more accurate data support for subsequent dynamic threshold generation and anomaly detection.

[0007] Optionally, in this embodiment of the invention, determining the individualized baseline of the target user based on effective resting data within a preset time period includes: determining the long-term mean μ and long-term standard deviation σ of physiological parameters based on the effective resting data; iteratively updating the long-term mean μ and long-term standard deviation σ over time using an exponentially weighted moving average algorithm; and constructing and determining the individualized baseline based on the updated long-term mean μ and long-term standard deviation σ. By calculating the long-term mean and standard deviation of the effective resting data and iteratively updating them using an exponentially weighted moving average algorithm, a dynamic individualized baseline is constructed. This allows for the gradual incorporation of the latest effective resting data while fully preserving the user's historical physiological characteristics, enabling the baseline to slowly and adaptively adjust with long-term changes in the user's physiological state. This avoids the shortcomings of static baselines, which cannot reflect changes in the user's circadian rhythm or long-term health trends, further improving the representativeness, stability, and timeliness of the baseline.

[0008] In the above implementation process, the target user's age information and health status tags are retrieved based on the target user's identifier, and combined with an individualized baseline to generate upper and lower limit thresholds for anomaly detection. This includes: determining the target user's disease type based on the health status tags; retrieving the corresponding asymmetric adjustment coefficients from a preset rule base based on the age information and disease type, where the asymmetric adjustment coefficients include an upper limit adjustment coefficient k1 and a lower limit adjustment coefficient k2; and generating upper and lower limit thresholds based on the long-term mean μ, long-term standard deviation σ, and the asymmetric adjustment coefficients, wherein the upper limit threshold is: The lower limit threshold is: By combining the target user's age information and health status tags, and retrieving suitable asymmetric adjustment coefficients from a preset rule base, upper and lower thresholds are generated based on the mean and standard deviation of an individualized baseline. This allows for differentiated and asymmetric threshold configuration according to the physiological tolerance characteristics of users of different age groups and disease types. It avoids the shortcomings of traditional symmetric thresholds, which are insufficiently adaptable to the physiological fluctuations of special populations such as the elderly with chronic diseases and cardiovascular diseases. This makes the anomaly judgment logic more closely aligned with the individual physiological risk characteristics of users, further reducing false alarms and false negatives caused by "one-size-fits-all" symmetric threshold settings, and significantly improving the accuracy and personalized adaptation capabilities of anomaly identification.

[0009] In the above implementation process, after generating the upper and lower limit thresholds for anomaly detection, the following steps are also included: when the target user's activity intensity transitions from a non-resting state to a physiological recovery period, the current activity end time is obtained, where the physiological recovery period is the transition phase from an active state to a resting state; based on the activity end time, the numerical range of the upper and lower limit thresholds is temporarily expanded; the target user's limb activity intensity is continuously monitored, and when the target user is detected to have returned to a resting state, the upper and lower limit thresholds are restored to the values ​​determined based on an individualized baseline. By temporarily expanding the upper and lower limit threshold ranges for anomaly detection based on the activity end time when the user transitions from a non-resting state to a physiological recovery period, and automatically restoring the standard thresholds after the user returns to a resting state, the system can effectively adapt to the normal fluctuation range after physiological activity, avoid false alarms caused by physiological fluctuations after activity, further improve the rationality and scenario adaptability of anomaly detection, and significantly reduce the false alarm rate while ensuring the sensitivity of anomaly monitoring.

[0010] Optionally, in this embodiment of the invention, confirming that real-time physiological data meets a preset persistent over-limit condition includes: counting the number of over-limit sampling points of real-time physiological parameter data within a sliding time window; determining whether the number of over-limit sampling points reaches a preset point threshold, or determining whether the proportion of the number of over-limit sampling points to the total number of sampling points within the sliding time window exceeds a preset proportion threshold; if the point threshold is reached or the proportion threshold is exceeded, then the real-time physiological data is determined to meet the persistent over-limit condition. Determining the persistent over-limit condition by using the number or proportion of over-limit sampling points within the sliding time window can effectively filter out false alarms caused by instantaneous interference or brief fluctuations, ensuring that an alarm is triggered only when the physiological parameter abnormality reaches a certain level of persistence, significantly reducing the false alarm rate caused by occasional interference, and improving the reliability and stability of abnormal alarms.

[0011] Optionally, in this embodiment of the invention, monitoring real-time physiological parameters includes: using the heart rate parameter among the real-time physiological parameters as an initial trigger indicator for real-time comparison; when the initial trigger indicator exceeds the upper and lower thresholds, analyzing the respiratory rate parameter and blood oxygen saturation parameter among the real-time physiological parameters; if at least one of the respiratory rate parameter and blood oxygen saturation parameter meets a preset abnormal correlation condition, then increasing the anomaly confidence of the current monitoring result. By adopting a multi-physiological parameter linkage verification monitoring mechanism, using the heart rate parameter as the initial trigger indicator, and further analyzing the status of related parameters such as respiratory rate and blood oxygen saturation after the heart rate exceeds the limit, the anomaly confidence is increased only when multiple parameters are abnormally correlated. This effectively avoids misjudgment caused by environmental or equipment interference with a single indicator, and improves the anti-interference capability and reliability of anomaly identification.

[0012] Optionally, in this embodiment of the invention, the method further includes: receiving false alarm markers or confirmation markers for abnormal alarm signals; automatically adjusting the values ​​of asymmetric adjustment coefficients k1 or k2, or adjusting the preset time period used to establish an individualized baseline, based on the false alarm markers or confirmation markers; regenerating upper and lower limit thresholds based on the adjusted asymmetric adjustment coefficients, or updating the individualized baseline. By receiving false alarm / confirmation markers from users or caregivers for abnormal alarm signals, automatic iterative optimization of the asymmetric adjustment coefficients or baseline modeling time period is achieved. This enables continuous optimization of the monitoring strategy based on feedback from real scenarios, allowing the thresholds and baseline models to continuously adapt to the user's real physiological characteristics and scenario features, forming a closed-loop adaptive learning mechanism. This further improves the accuracy of abnormal identification during long-term use, continuously reduces the false alarm rate and missed alarm rate, and enhances the system's adaptability and long-term reliability.

[0013] Secondly, the present invention also provides a personalized physiological abnormality monitoring system based on non-contact sensing, applied to the embedded control system of a smart elderly care chair, comprising: a data acquisition module for acquiring real-time physiological parameter data of the target user provided by a non-contact sensing device; an individualized baseline construction module for determining the individualized baseline of the target user based on effective resting data of a preset duration in the past; a dynamic threshold generation module for retrieving the target user's age information and health status label, and combining them with the individualized baseline to generate upper and lower limit thresholds for physiological parameters; and a real-time abnormality monitoring module for monitoring the target user's real-time physiological data, and triggering an abnormality alarm signal if it is confirmed that the real-time physiological data exceeds the upper and lower limit thresholds and that the real-time physiological data meets preset continuous over-limit conditions.

[0014] The personalized physiological abnormality monitoring method based on non-contact sensing described in this invention, through a complete process of target user identification acquisition, individualized baseline construction, dynamic threshold generation, and continuous over-limit monitoring, helps to achieve precise and intelligent monitoring of the physiological state of the elderly, effectively improves accuracy and stability, and enhances the level of safety and emergency response efficiency in smart elderly care scenarios.

[0015] According to the technical solution of the present invention, by obtaining the target user identifier and real-time physiological parameter data, and determining an individualized baseline based on the effective resting data within a preset time period, and then combining age information and health status labels to generate upper and lower limit thresholds, an alarm is triggered when the limits are exceeded and the continuous condition is met during real-time monitoring. This enables a technical shift from "general threshold judgment" to "individualized dynamic adaptation", changing the inherent defect of inaccurate abnormality judgment caused by ignoring individual differences in the existing technology, making the physiological monitoring logic more in line with the real user state, and significantly improving the accuracy and reliability of abnormality identification.

[0016] Meanwhile, the present invention constructs an individualized baseline based on effective resting data within a preset time period, which can fully reflect the physiological characteristics of users in a stable state, avoid the influence of accidental interference data on baseline modeling, and make the benchmark data more representative and individual-specific.

[0017] Furthermore, this invention retrieves the age information and health status tags of target users based on their identifiers, and generates upper and lower limit thresholds by combining them with individualized baselines. This enables differentiated threshold configurations for different age groups and different disease risk groups, giving the monitoring system the ability to adapt to individual physiological characteristics. It significantly reduces false alarms and missed alarms caused by "one-size-fits-all" threshold settings, and improves the personalization and clinical adaptability of the monitoring strategy. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of the present invention or the conventional technology, the drawings used in the description of the embodiments or the conventional technology 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 from these drawings without creative effort.

[0019] Figure 1 This is a flowchart of the personalized physiological abnormality monitoring method provided in the embodiments of the present invention; Figure 2 A flowchart of an effective resting data filtering method provided in an embodiment of the present invention; Figure 3 This is a flowchart of the individualized baseline construction method provided in the embodiments of the present invention; Figure 4This is a flowchart of the dynamic threshold generation method provided in the embodiments of the present invention; Figure 5 This is a flowchart of the threshold adaptive adjustment method provided in the embodiments of the present invention; Figure 6 A flowchart of the method for determining persistent over-limit conditions provided in the embodiments of the present invention; Figure 7 This is a flowchart of the multi-parameter anomaly monitoring method provided in the embodiments of the present invention; Figure 8 This is a flowchart of the alarm feedback and self-optimization method provided in the embodiments of the present invention; Figure 9 This is a schematic diagram of the structure of a personalized physiological abnormality monitoring system provided in another embodiment of the present invention. Detailed Implementation

[0020] The technical solutions of the embodiments of the present invention will now be described with reference to the accompanying drawings. For example, the flowcharts and block diagrams in the drawings illustrate the architecture, functions, and operations of possible implementations of systems, methods, and computer program products according to various embodiments of the present invention. In this regard, each block in the flowchart or block diagram may represent a module, program segment, or part of code, which contains one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than those marked in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagram and / or flowchart, and combinations of blocks in the block diagram and / or flowchart, can be implemented using a dedicated hardware-based system that performs the specified function or action, or can be implemented using a combination of dedicated hardware and computer instructions. In addition, the functional modules in the various embodiments of the present invention may be integrated together to form an independent part, or each module may exist separately, or two or more modules may be integrated to form an independent part.

[0021] Please refer to Figure 1 , Figure 1 This is a flowchart of a personalized physiological abnormality monitoring method provided in an embodiment of the present invention. The personalized physiological abnormality monitoring method includes: Step S100: Obtain real-time physiological parameter data of the target user provided by the non-contact sensing device.

[0022] In step S100 above, physiological signal acquisition is completed using the non-contact sensing device built into the smart elderly care chair. This process requires no device from the user and continuously and stably acquires real-time physiological parameter data of the target user without interfering with normal sitting posture. This provides raw data support for subsequent individualized baseline construction, dynamic threshold calculation, and anomaly detection. The non-contact sensing device may include a millimeter-wave radar sensor and a multi-wavelength optical sensor. The millimeter-wave radar is used to penetrate clothing to sense micro-movements in the human chest cavity to extract respiratory rate signals. The multi-wavelength optical sensor acquires heart rate signals based on the photoplethysmography (PPG) principle, and in some scenarios, can also simultaneously estimate blood oxygen saturation parameters. These signals are continuously acquired at a fixed frequency, and the raw signals undergo preliminary preprocessing to remove data with significant distortion, amplitude overflow, or insufficient confidence. This ensures that the real-time physiological parameter data entering subsequent processing is valid and usable, laying a high-quality data foundation for subsequent baseline modeling, threshold generation, and anomaly monitoring.

[0023] In practical implementation, the embedded control system of the smart elderly care chair uses a non-contact sensing device composed of a built-in millimeter-wave radar sensor and a multi-wavelength optical sensor to collect physiological signals without contact or perception while the user is in a seated position. The millimeter-wave radar outputs respiratory rate data by detecting periodic micro-movements of the chest cavity, while the multi-wavelength optical sensor collects blood flow fluctuation signals through dual channels of green and infrared light to obtain heart rate data and simultaneously outputs estimated blood oxygen saturation data. These real-time physiological parameters can be acquired at a fixed acquisition frequency of once per second. Simultaneously, the preprocessing module built into the sensing device filters, removes artifacts, and scores confidence levels on the raw signals, using only valid data with confidence levels higher than a preset standard as the target user's real-time physiological parameter data.

[0024] Step S200: Determine the individualized baseline of the target user based on the effective resting data of the preset duration in the past.

[0025] In step S200 above, the preset duration can be set to a continuous window of nearly 7 days, with at least 2 hours of effective resting data per day, and a rolling update mechanism is adopted, always based on the data of the most recent period. Effective resting data refers to high-quality physiological parameter data collected when the user is sitting still with no significant limb movement, and the ambient light and electromagnetic interference meet preset conditions, effectively eliminating data bias caused by motion interference and environmental noise. Based on the effective resting data, the long-term mean μ and long-term standard deviation σ of the physiological parameters are calculated, and the mean and standard deviation are iteratively updated using the exponentially weighted moving average (EWMA) algorithm, enabling the baseline to adaptively adjust as the user's health status slowly changes, forming an individualized baseline statistical model with μ and σ as the core, providing an accurate and stable personalized benchmark for subsequent dynamic threshold generation. In the specific implementation process, the embedded control system uses effective resting data from nearly 7 consecutive days, with no less than 2 hours per day, as a basis. It performs statistical calculations on the selected high-quality heart rate and respiratory rate data to obtain the long-term mean μ and long-term standard deviation σ of the corresponding physiological parameters. The system then uses an exponentially weighted moving average algorithm to iteratively update the long-term mean μ and long-term standard deviation σ over time according to a preset smoothing coefficient. This allows the baseline to continuously adapt to the slow changes in the user's physiological state. At the same time, the updated mean and standard deviation are used as core parameters to construct an individualized statistical baseline model, thereby obtaining an individualized baseline that can truly and stably represent the resting physiological characteristics of the target user.

[0026] Step S300: Retrieve the target user's age information and health status tags, and combine them with an individualized baseline to generate upper and lower limits for physiological parameters.

[0027] In step S300 above, by introducing the dual factors of user age and health status, a differential threshold is calculated for the individualized baseline. The embedded control system retrieves the corresponding age information and health status label from the local storage unit based on the target user identifier. Based on the health status label, it determines the user's disease type. Then, combining the age information and disease type, it retrieves the corresponding asymmetric adjustment coefficients k1 and k2 from the preset rule base. The calculation is based on the long-term mean μ and long-term standard deviation σ in the individualized baseline, with the upper limit threshold as follows: The lower limit threshold is: The calculation method generates upper and lower thresholds for abnormal judgment of corresponding physiological parameters, thereby obtaining a personalized dynamic threshold range that is adapted to the user's age characteristics and health status, providing accurate and reliable judgment criteria for subsequent real-time abnormal monitoring.

[0028] Step S400: Monitor the real-time physiological data of the target user. If it is confirmed that the real-time physiological data exceeds the upper and lower limit thresholds and that the real-time physiological data meets the preset continuous over-limit conditions, then trigger an abnormal alarm signal.

[0029] In step S400 above, real-time physiological parameters are continuously acquired using a high-frequency sampling method, and the parameters are compared in real time with dynamically generated personalized upper and lower limit thresholds. To avoid misjudgment caused by instantaneous interference or single fluctuations, a secondary verification can be performed using a continuous out-of-limit condition. The number of out-of-limit sampling points or the proportion of out-of-limit points are counted by using a sliding time window, and the parameter is only judged as a real abnormality when it continuously exceeds the threshold. At the same time, multi-parameter cross-validation can be combined, with heart rate as the initial trigger indicator, and parameters such as respiratory rate and blood oxygen saturation are linked to improve the confidence of the abnormality, ensuring that the alarm is real and effective. In the specific implementation process, the embedded control system continuously acquires real-time physiological data such as heart rate and respiratory rate of the target user at a fixed frequency. The real-time data is compared point by point with the currently effective upper and lower thresholds. When the real-time physiological data is detected to exceed the threshold range, the number of sampling points exceeding the limit or the proportion of exceeding the limit is further counted within a preset sliding time window to determine whether the preset continuous over-limit condition is met. At the same time, the heart rate is used as the initial trigger indicator and is linked with respiratory rate and blood oxygen saturation parameters for multi-parameter cross-validation to improve the confidence of the anomaly. When the real-time data both exceeds the threshold and meets the continuous over-limit condition, an abnormal alarm signal is immediately generated and triggered to achieve accurate, reliable, and low false alarm real-time monitoring and early warning of user physiological abnormalities.

[0030] Please refer to Figure 2 , Figure 2 The flowchart illustrates an effective resting data filtering method provided in an embodiment of the present invention; the effective resting data filtering method includes: Step S20: Receive the limb activity intensity of the target user provided by the accelerometer; when the limb activity intensity is lower than the preset activity threshold, determine that the target user is in a preliminary resting state.

[0031] In step S20 above, the accelerometer built into the seat collects the micro-motion signals of the user's limbs and quantifies the intensity of limb activity. This intensity is used as the first screening condition to determine whether the user is in a stable sitting posture, excluding fluctuations in physiological parameters caused by limb movements such as getting up, turning, or swaying. Subsequently, the accelerometer in the smart elderly care chair collects and transmits the target user's limb activity intensity data in real time and compares it with a preset activity threshold. When the limb activity intensity is determined to be lower than the preset activity threshold, it is confirmed that the target user is currently not making any obvious limb movements and is maintaining a stable sitting posture, and this state is then marked as a preliminary resting state.

[0032] Step S21: In the initial resting state, further detect the ambient light intensity and electromagnetic interference level; only when the ambient light intensity and electromagnetic interference level meet the preset environmental quality conditions, mark the currently obtained physiological parameter data as valid resting data and include it in the dataset within the preset time period.

[0033] In step S21 above, on the basis of the initial resting state, dual detection of environmental interference is added. After determining that the target user is in the initial resting state, the current ambient light intensity is detected by the ambient light sensor and the current electromagnetic interference level is obtained by the electromagnetic interference detection module. Only when the ambient light intensity and electromagnetic interference level meet the preset environmental quality conditions, the physiological parameter data collected by the current non-contact sensing device is marked as valid resting data and the valid resting data is included in the preset time period dataset used to construct an individualized baseline.

[0034] Please refer to Figure 3 , Figure 3 A flowchart of individualized baseline construction provided for embodiments of the present invention; the individualized baseline construction method includes: Step S30: Based on valid resting data, determine the long-term mean μ and long-term standard deviation σ of the physiological parameters.

[0035] In step S30 above, valid resting data within a preset time period are read, and statistical calculations are performed on physiological parameters such as heart rate and respiratory rate. The long-term mean μ of the corresponding physiological parameters is obtained by summing and averaging, and the long-term standard deviation σ is obtained by square root of variance, thus completing the determination of the initial statistical parameters of the individualized baseline.

[0036] Step S31: Iteratively update the long-term mean μ and long-term standard deviation σ over time using the exponentially weighted moving average algorithm.

[0037] In step S31 above, the embedded control system loads a preset smoothing coefficient, takes the latest collected effective resting data as input, and uses the exponentially weighted moving average algorithm to iteratively calculate the generated long-term mean μ and long-term standard deviation σ. The statistical parameters are continuously corrected according to the preset update rules to realize the dynamic update of the long-term mean μ and long-term standard deviation σ over time.

[0038] Step S32: Based on the updated long-term mean μ and long-term standard deviation σ, construct and determine the individualized baseline.

[0039] In step S32 above, a statistical model is constructed to characterize the user's personalized physiological characteristics, using the long-term mean μ and long-term standard deviation σ updated by the exponentially weighted moving average algorithm as the core components. This model is then used as the individualized baseline for the current target user and can be used for the dynamic generation of subsequent anomaly judgment thresholds.

[0040] Please refer to Figure 4 , Figure 4 This is a flowchart of a dynamic threshold generation method provided in an embodiment of the present invention; the dynamic threshold generation method includes: Step S40: Determine the disease type of the target user based on the health status label.

[0041] In step S40 above, the embedded control system classifies the user's health status according to the acquired target user health status label, and determines whether it includes common chronic diseases of the elderly such as hypertension, heart failure, and COPD. This provides a classification basis for subsequent matching of differentiated and personalized adjustment coefficients, ensuring that the threshold can adapt to the physiological tolerance characteristics of different disease groups.

[0042] Step S41: Based on age information and disease type, retrieve the corresponding asymmetric adjustment coefficient from the preset rule base. The asymmetric adjustment coefficient includes the upper limit adjustment coefficient k1 and the lower limit adjustment coefficient k2.

[0043] In step S41 above, the target user's age information and disease type are used as search conditions. A matching query is performed in the preset rule base to retrieve the asymmetric adjustment coefficient that is suitable for the age and disease type, and obtain independently configured upper limit adjustment coefficient k1 and lower limit adjustment coefficient k2. The upper limit adjustment coefficient k1 can be used to control the abnormal upper limit sensitivity, and the lower limit adjustment coefficient k2 can be used to control the abnormal lower limit sensitivity. The independent setting of the two can realize asymmetric threshold adjustment, which is more in line with the physiological fluctuation pattern of the elderly and chronic disease users.

[0044] This pre-defined rule base is categorized and configured according to two dimensions: age stratification and health status / disease type. The age stratification includes at least two ranges: under 75 years old and 75 years old and above. The disease type includes at least common elderly health conditions such as hypertension, heart failure, chronic obstructive pulmonary disease (COPD), and no underlying diseases. For different combinations, the rule base pre-defines a unique upper limit adjustment coefficient k1 and a lower limit adjustment coefficient k2, with k1 and k2 taking independent values. This achieves asymmetric adjustment of the upper and lower thresholds, adapting to the physiological tolerance and abnormal risk sensitivity of different age groups and health statuses, making abnormal judgments more consistent with clinical monitoring logic. For example, for users over 75 years old with cardiovascular diseases such as hypertension and heart failure, the rule base configures a stricter upper limit adjustment coefficient k1; for COPD users with weaker respiratory function, a relatively lenient lower limit adjustment coefficient k2 is configured to avoid misjudging normal physiological fluctuations as abnormalities.

[0045] Step S42: Generate upper and lower limit thresholds based on the long-term mean μ, long-term standard deviation σ, and asymmetric adjustment coefficient.

[0046] In step S42 above, the long-term mean μ and long-term standard deviation σ in the individualized baseline are used as the basis for calculation. Combined with the obtained upper limit adjustment coefficient k1 and lower limit adjustment coefficient k2, the upper limit threshold is as follows: The lower limit threshold is: The calculation formulas are performed to generate upper and lower limit thresholds for judging abnormal physiological parameters applicable to the current target user.

[0047] Please refer to Figure 5 , Figure 5 This is a flowchart of a threshold adaptive adjustment method provided in an embodiment of the present invention; the threshold adaptive adjustment method includes: Step S50: When the target user's activity intensity changes from a non-resting state to a physiological recovery period, obtain the current activity end time.

[0048] In step S50 above, the intensity of the target user's limb activity is monitored in real time by a non-contact sensing device (including accelerometer and millimeter-wave radar). When the user returns to the seat from a non-resting state such as getting up or walking and enters the physiological recovery period, the current time is immediately obtained and recorded as the end time of the activity.

[0049] It should be noted that the physiological recovery period is the transitional phase where the target user returns to a seated position from a non-resting state (such as getting up, walking, or engaging in limb activity), but their physiological parameters such as heart rate and respiratory rate have not yet returned to their resting baseline levels. During this phase, the intensity of the user's limb activity has significantly decreased and meets the initial resting condition criteria, but due to the influence of previous physical activity, physiological parameters such as heart rate and respiratory rate still exhibit normal physiological fluctuations. If standard individualized baseline thresholds are used directly for judgment, these normal fluctuations are easily misjudged as abnormal. To avoid this problem, after detecting that the user has entered the physiological recovery period, an automatic temporary threshold expansion mechanism is activated to accommodate normal physiological changes after activity; the physiological recovery period is considered to have ended when the user's limb activity stabilizes and physiological parameters return to resting levels.

[0050] Step S51: Based on the activity end time, temporarily expand the numerical range of the upper and lower threshold values.

[0051] In step S51 above, starting from the end time of the activity, the upper and lower thresholds of the currently effective abnormal judgment are temporarily expanded according to the preset strategy during the physiological recovery period. By increasing the upper limit adjustment coefficient or directly widening the threshold range, the normal physiological fluctuations of the user's heart rate and respiratory rate after the activity are adapted to avoid false alarms triggered by normal fluctuations, and the monitoring sensitivity and anti-interference ability are taken into account.

[0052] Step S52: Continuously monitor the intensity of the target user's limb activity. When the target user is detected to have returned to a resting state, restore the upper and lower limit thresholds to the values ​​determined based on the individualized baseline.

[0053] In step S52 above, the intensity of the target user's limb activity can be continuously monitored during the temporary threshold expansion period. When the intensity of the user's limb activity is detected to be lower than the preset activity threshold and stabilizes again in a resting state, the currently effective upper and lower limit thresholds are immediately restored to the standard values ​​calculated by the individualized baseline and the asymmetric adjustment coefficient.

[0054] Please refer to Figure 6 , Figure 6 This is a flowchart of a method for determining persistent exceedance conditions provided in an embodiment of the present invention; the method for determining persistent exceedance conditions includes: Step S60: Count the number of out-of-limit sampling points of real-time physiological parameter data within the sliding time window.

[0055] In step S60 above, the embedded control system segments and buffers the real-time physiological parameter sampling points according to a preset sliding time window, compares the value of each sampling point in the window with the corresponding upper and lower limit thresholds, marks the sampling points that exceed the limit one by one, and completes the quantity statistics.

[0056] Step S61: Determine whether the number of out-of-limit sampling points reaches the preset point threshold, or determine whether the proportion of the number of out-of-limit sampling points to the total number of sampling points in the sliding time window exceeds the preset proportion threshold.

[0057] In step S61 above, the number of out-of-limit sampling points obtained by statistics can be compared with the preset number of points threshold. At the same time, the proportion of the number of out-of-limit sampling points to the total number of sampling points in the sliding time window can be calculated, and this proportion can be compared with the preset proportion threshold to complete the two-dimensional condition judgment.

[0058] Step S62: If the number of points reaches the threshold or exceeds the proportion threshold, the real-time physiological data is determined to meet the continuous over-limit condition.

[0059] In step S62 above, "OR logic" is used for comprehensive judgment, that is, if either the number of sampling points or the proportion condition is met, it can be identified as a continuous abnormality. That is, based on the judgment result of the two conditions, when the number of sampling points exceeding the limit reaches the preset number of sampling points threshold, or the proportion of exceeding the limit exceeds the preset proportion threshold, the current real-time physiological parameters are immediately determined to meet the preset continuous exceeding condition.

[0060] Please refer to Figure 7 , Figure 7 This is a flowchart of a multi-parameter anomaly monitoring method provided in an embodiment of the present invention; the multi-parameter anomaly monitoring method includes: Step S70: Use the heart rate parameter in the real-time physiological parameters as the initial trigger index for real-time comparison.

[0061] In step S70 above, the embedded control system prioritizes the heart rate parameter with high sensitivity and fast response as the core trigger index. It continuously compares the real-time collected heart rate data with the currently dynamically generated upper and lower limit thresholds, using this as the pre-trigger condition for multi-parameter joint verification. This ensures timely anomaly identification while reducing unnecessary multi-parameter computation overhead.

[0062] Step S71: When the initial trigger index exceeds the upper and lower limit thresholds, analyze the respiratory rate parameter and blood oxygen saturation parameter in the real-time physiological parameters.

[0063] In step S71 above, when the heart rate parameter, which is the initial trigger indicator, is determined to exceed the upper and lower limit thresholds, the respiratory rate parameter and blood oxygen saturation parameter within the same sampling period are immediately acquired and analyzed simultaneously, and the multi-parameter joint verification process is entered.

[0064] Step S72: If at least one of the respiratory rate parameter and blood oxygen saturation parameter meets the preset abnormal correlation condition, then increase the abnormal confidence of the current monitoring result.

[0065] In step S72 above, the respiratory rate parameter and blood oxygen saturation parameter are compared with their respective preset normal ranges or association rules. If at least one of the parameters meets the preset abnormal association conditions, it is determined to be a multi-parameter synergistic abnormality, and the confidence of the current physiological abnormality monitoring results is increased accordingly.

[0066] Please refer to Figure 8 , Figure 8 The flowchart below shows the alarm feedback and self-optimization method provided in the embodiments of the present invention; the alarm feedback and self-optimization method includes: Step S80: Receive a false alarm flag or confirmation flag for the abnormal alarm signal.

[0067] In step S80 above, the embedded control system provides an alarm result marking interface to receive marking operations from nursing staff, family members or users for the triggered abnormal alarm signals. The marking types are divided into two categories: false alarm marking and confirmation marking, which are used to characterize whether the alarm is an invalid false alarm or a real valid abnormality, respectively, and record the corresponding marking results and occurrence time, thereby providing real scene supervision signals for subsequent model adaptive optimization.

[0068] Step S81: Based on the false alarm flag or the confirmation flag, automatically adjust the value of the asymmetric adjustment coefficient k1 or k2, or adjust the preset time period used to establish the individualized baseline.

[0069] In step S81 above, a differentiated optimization strategy is executed according to the type of the received marker: if it is a false alarm marker, the asymmetric adjustment coefficients k1 and k2 are appropriately relaxed to expand the normal threshold range, or the preset time period used for baseline modeling is adjusted to optimize the baseline fitting effect; if it is a confirmed marker, the coefficients are maintained or moderately tightened to ensure monitoring sensitivity and achieve dynamic optimization based on real usage feedback.

[0070] Step S82: Regenerate the upper and lower limit thresholds based on the adjusted asymmetric adjustment coefficients, or update the individualized baseline.

[0071] In step S82 above, new upper and lower limit thresholds are recalculated and generated based on the adjusted asymmetric adjustment coefficient, or effective resting data are re-screened and individualized baselines are updated based on the adjusted preset time period, so that subsequent monitoring is carried out using the optimized standard.

[0072] Please refer to Figure 9 , Figure 9 This is a schematic diagram of the structure of the personalized physiological abnormality monitoring system 9 provided in an embodiment of the present invention. The system includes a data acquisition module 10, an individualized baseline construction module 20, a dynamic threshold generation module 30, and a real-time abnormality monitoring module 40. The data acquisition module 10 is used to acquire real-time physiological parameter data of the target user provided by a non-contact sensing device. The individualized baseline construction module 20 is used to determine the individualized baseline of the target user based on valid resting data of a preset duration. The dynamic threshold generation module 30 is used to retrieve the target user's age information and health status label, and, in conjunction with the individualized baseline, generate upper and lower limit thresholds for physiological parameters. The real-time abnormality monitoring module 40 is used to monitor the real-time physiological data of the target user; if the real-time physiological data exceeds the upper and lower limit thresholds and meets preset continuous exceedance conditions, an abnormality alarm signal is triggered.

[0073] The above description is merely an embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Various modifications can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A personalized physiological abnormality monitoring method based on non-contact sensing, characterized in that, An embedded control system for intelligent elderly care chairs, the method comprising: Acquire real-time physiological parameter data of the target user provided by a non-contact sensing device; The individualized baseline of the target user is determined based on effective resting data of a preset duration in the past; wherein, the individualized baseline is a statistical model established based on the physiological parameter data of the target user in the historical resting state; Retrieve the target user's age information and health status tags, and combine them with the individualized baseline to generate upper and lower limits for physiological parameters; The system monitors the target user's real-time physiological data. If the real-time physiological data exceeds the upper and lower thresholds and meets the preset continuous over-limit conditions, an abnormal alarm signal is triggered.

2. The method according to claim 1, characterized in that, Before determining the individualized baseline of the target user based on valid resting data within a preset time period, the method further includes: The system receives the limb activity intensity of the target user from the accelerometer; when the limb activity intensity is lower than a preset activity threshold, the system determines that the target user is in a preliminary resting state. In the initial resting state, the ambient light intensity and electromagnetic interference level are further detected; only when the ambient light intensity and electromagnetic interference level meet the preset environmental quality conditions, the currently obtained physiological parameter data are marked as the valid resting data and included in the dataset within the preset time period.

3. The method according to claim 1, characterized in that, The process of determining the individualized baseline of the target user based on valid resting data within a preset time period includes: Based on the effective resting data, the long-term mean μ and long-term standard deviation σ of the physiological parameters are determined; The long-term mean μ and long-term standard deviation σ are iteratively updated over time using an exponentially weighted moving average algorithm. The individualized baseline is constructed and determined based on the updated long-term mean μ and long-term standard deviation σ.

4. The method according to claim 3, characterized in that, The step of retrieving the target user's age information and health status tags based on the target user identifier, and combining this with the individualized baseline to generate upper and lower thresholds for anomaly detection, includes: The disease type of the target user is determined based on the health status label; Based on the age information and the disease type, the corresponding asymmetric adjustment coefficient is retrieved from the preset rule base. The asymmetric adjustment coefficient includes an upper limit adjustment coefficient k1 and a lower limit adjustment coefficient k2. Based on the long-term mean μ, the long-term standard deviation σ, and the asymmetric adjustment coefficient, the upper and lower thresholds are generated, wherein the upper threshold is: The lower limit threshold is: .

5. The method according to claim 4, characterized in that, After generating the upper and lower thresholds for anomaly detection, the method further includes: When the target user's activity intensity changes from a non-resting state to a physiological recovery period, obtain the current activity end time; Based on the activity end time, temporarily expand the numerical range of the upper and lower threshold values; The intensity of the target user's limb activity is continuously monitored. When the target user is detected to have returned to a resting state, the upper and lower limit thresholds are restored to the values ​​determined based on the individualized baseline.

6. The method according to claim 1, characterized in that, The confirmation that the real-time physiological data meets the preset continuous exceedance condition includes: Count the number of out-of-limit sampling points of the real-time physiological parameter data within the sliding time window; Determine whether the number of out-of-limit sampling points reaches a preset point threshold, or determine whether the proportion of the number of out-of-limit sampling points to the total number of sampling points in the sliding time window exceeds a preset proportion threshold. If the number of points reaches the threshold or exceeds the proportion threshold, the real-time physiological data is determined to meet the continuous over-limit condition.

7. The method according to claim 1, characterized in that, The monitoring of the real-time physiological parameters includes: The heart rate parameter among the real-time physiological parameters is used as the initial trigger index for real-time comparison. When the initial trigger index exceeds the upper and lower limit thresholds, analyze the respiratory rate parameter and blood oxygen saturation parameter in the real-time physiological parameters; If at least one of the respiratory rate parameter and blood oxygen saturation parameter meets the preset abnormal correlation condition, the abnormality confidence of the current monitoring result is increased.

8. The method according to claim 4, characterized in that, The method further includes: Receive false alarm flags or confirmation flags for the abnormal alarm signals; Based on the false alarm flag or the confirmation flag, the value of the asymmetric adjustment coefficient k1 or k2 is automatically adjusted, or the preset time period used to establish an individualized baseline is adjusted. The upper and lower limit thresholds are regenerated based on the adjusted asymmetric adjustment coefficients, or the individualized baseline is updated.

9. A personalized physiological abnormality monitoring system based on non-contact sensing, applied to the embedded control system of a smart elderly care chair, characterized in that, include: The data acquisition module is used to acquire real-time physiological parameter data of the target user provided by the non-contact sensing device; The individualized baseline construction module is used to determine the individualized baseline of the target user based on effective resting data of a preset duration in the past; The dynamic threshold generation module is used to retrieve the target user's age information and health status tags, and combine them with the individualized baseline to generate upper and lower limit thresholds for physiological parameters. The real-time anomaly monitoring module is used to monitor the real-time physiological data of the target user. If it is confirmed that the real-time physiological data exceeds the upper and lower limit thresholds and that the real-time physiological data meets the preset continuous over-limit conditions, an anomaly alarm signal is triggered.

10. An electronic device, characterized in that, The electronic device includes a memory and a processor, the memory storing program instructions, and the processor executing the steps of the method according to any one of claims 1-8 when running the program instructions.

Images