Physiological health state assessment method based on step number data

Abstract

The invention discloses a physiological health state assessment method based on step number data, and relates to the technical field of physiological health assessment, and the method comprises the steps: constructing a step number time set Step and a long-term stable baseline set Base, and reflecting a normal step number behavior mode formed by a user in a long-term natural life state; through generation of a disturbance event set Dist, continuous step number behavior abnormity generated under real conditions such as work intensity change, work and rest disruption or body state fluctuation can be identified; through extraction of a recovery section set Rec and a self-stabilization feature set Rest, the time required by a user to recover to a long-term stable state after disturbance, the stability degree of the recovery process and the stability keeping capacity after recovery can be quantified, so that the sub-health state with the step number reaching the standard but the recovery capacity declining is prevented from being misjudged as the normal health state, and the user experience is improved. The defect that in the prior art, the adjusting capacity of the physiological system is difficult to reflect only according to the step number or short-term statistics is effectively overcome.

Description

Technical Field

[0001] This invention relates to the field of physiological health assessment technology, specifically a method for assessing physiological health status based on step count data. Background Technology

[0002] With the widespread adoption of wearable devices and mobile terminals, assessing an individual's physiological health based on step count data is gradually becoming an important technological direction in health management and daily physiological monitoring. Step count data, as a behavioral parameter that can be acquired continuously over a long period, naturally reflects an individual's activity level, circadian rhythm, and behavioral trends in their real-life environment.

[0003] Existing methods for assessing physiological health based on step count data typically focus on analyzing total steps, adherence to targets, or short-term changes to determine whether an individual's activity level is sufficient. However, in reality, an individual's physiological health depends not only on the number of steps taken within a specific timeframe but also on whether their step count can return to a stable state within a certain period after disturbances such as work stress, disrupted sleep patterns, excessive exercise, or unexpected events. Current technologies generally lack continuous tracking and analysis of long-term changes in step count data, making it impossible to assess an individual's self-regulation and recovery capabilities after step count disruptions, thus failing to reflect the overall resilience of the physiological system. This lack of assessment means that individuals whose physiological functions are already declining may still be judged as having normal health despite meeting step count targets. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a method for assessing physiological health status based on step count data, thus solving the problems mentioned in the background section.

[0005] To achieve the above objectives, the present invention provides a method for assessing physiological health status based on step count data, comprising the following steps: S1. Collect step data generated by the user within a preset monitoring period, and perform time alignment processing on the step data to construct a step time set Step for describing the step change process; S2. Based on the step time set Step, extract step behavior features that meet the stability conditions, and construct a long-term stable baseline set Base to represent the user's normal step behavior pattern; S3. Compare the step time set Step with the long-term stable baseline set Base, identify the perturbation segment where the step behavior deviates from the long-term stable baseline set Base, and generate the corresponding perturbation event set Dist; S4. For the set of disturbance events Dist, construct a recovery process for step behavior regression of the long-term stable baseline set Base, and extract a self-stabilizing feature set Rest from the recovery process to represent the self-stabilizing capability of step behavior; S5. Generate a physiological resilience assessment result Eval based on the self-stabilizing feature set Rest to represent the long-term self-stabilizing capability of the user's physiological system, and output the physiological resilience assessment result Eval.

[0006] Preferably, S1 includes S11; S11. Within a preset monitoring period, the step count information generated by the user during daily activities is continuously collected through the step count acquisition unit in the user's portable terminal device. The collected step increments are arranged in chronological order and bound to the corresponding time information to generate a raw step sequence RawStep containing step values ​​and time stamps.

[0007] Preferably, S1 further includes S12; S12. Divide the preset monitoring period into continuous time intervals according to the preset time granularity, and map each step data in the original step sequence RawStep to the corresponding time interval; collect the step data falling into the same time interval, generate step statistics results corresponding to each time interval, and arrange them according to the order of the time intervals, thereby constructing a step time set Step that reflects the step change process under a unified time scale.

[0008] Preferably, S2 includes S21; S21. In the step time set Step, the continuous time segment is divided into sliding segments according to a preset time window length; within each time segment, based on the step value corresponding to each consecutive time point within the time segment, the step value between adjacent time points is compared one by one to calculate a step difference set reflecting the step change amplitude, and the number of occurrences of non-zero differences in the step difference set is counted to determine the frequency of step change within the time segment; based on the change amplitude reflected by the step difference set and the change frequency reflected by the number of occurrences of non-zero differences, the segment step stability Sds representing the step fluctuation level within the time segment is calculated; the segment step stability Sds corresponding to each time segment is compared with a preset stability threshold; When the stability of the number of steps in any time interval Sds meets the stability condition, the time interval is determined to be a stable step behavior interval, and all time intervals determined to be stable are gathered to form a stable interval set Sta. When the stability of the number of steps in a time interval, Sds, does not meet the stability condition, the time interval is determined to be an unstable step behavior interval.

[0009] Preferably, S2 further includes S22; S22. Step feature extraction processing is performed on each stable step behavior segment contained in the stable segment set Sta. Within each stable step behavior segment, based on the step value in the step time set Step corresponding to the segment, a segment step count statistical result reflecting the step distribution of the segment is calculated. The segment step count statistical results corresponding to each stable step behavior segment are aggregated across segments to form a long-term step count statistical result representing the overall distribution of the user's step behavior in multiple stable time segments. A long-term stable baseline set Base is constructed based on the long-term step count statistical result.

[0010] Preferably, S3 includes S31; S31. The step count values ​​corresponding to each time segment in the step count time set Step are compared segment by segment with the baseline step count distribution information contained in the long-term stable baseline set Base. During the comparison process, the difference value of the actual step count value in each time segment relative to the long-term stable baseline set Base is calculated to obtain the segment deviation degree Dvd, which is used to represent the degree of deviation of the step count behavior in the time segment. The segment deviation degree Dvd corresponding to each time segment is compared with a preset deviation threshold. When the deviation degree Dvd of any time segment exceeds the deviation threshold, the time segment is determined to be a step deviation segment, and all time segments determined to be step deviation segments are aggregated to form a deviation segment set Dev. When the deviation degree Dvd of a time segment does not exceed the deviation threshold, the time segment is determined to be a segment where the number of steps has not deviated.

[0011] Preferably, S3 further includes S32; S32. Merge multiple consecutive step deviation segments in the deviation segment set Dev that are less than a preset merging condition, and treat the merged consecutive deviation segments as the same step behavior disturbance event; record the start time segment, end time segment, and duration information for each step behavior disturbance event, and gather all the constructed step behavior disturbance events to form a disturbance event set Dist used to describe abnormal step behavior disturbances.

[0012] Preferably, S4 includes S41; S41. For each disturbance event in the disturbance event set Dist, determine the end time segment corresponding to the disturbance event, and use the end time segment as the start time segment of the recovery process; after the start time segment, analyze the continuous time segments in the step time set Step according to the same time window division method as in S21, and calculate the corresponding segment step stability Sds in each continuous time segment; when the segment step stability Sds corresponding to a consecutive preset number of time segments are all lower than the stability threshold, and the step behavior in the time segment falls back into the normal step behavior range represented by the long-term stable baseline set Base, the earliest time segment is determined as the recovery completion time segment; all continuous time segments from the recovery process start time segment to the recovery completion time segment are collected to form a recovery segment set Rec for representing the step behavior recovery process.

[0013] Preferably, S4 further includes S42; S42. After the set of recovery segments Rec is determined, all time segments contained in the set of recovery segments Rec are analyzed. Based on the step data in the step time set Step corresponding to each time segment and the corresponding segment step stability Sds, recovery features representing the self-stabilizing ability of step behavior are extracted. The recovery features include recovery duration features and recovery fluctuation features. Specifically, by counting the number of time segments in the recovery segment set Rec, a recovery duration feature is obtained, representing the time elapsed from the start time segment to the completion time segment of the recovery process. By analyzing the changes in the segment step stability Sds corresponding to each time segment in the recovery segment set Rec during the recovery process, a recovery fluctuation feature is obtained, representing whether the step behavior experiences repeated fluctuations during the recovery process. After the completion time segment, following the same time window division method as in S41, the corresponding segment step stability Sds is calculated for a predetermined number of consecutive time segments in the step time set Step, and compared with the stability threshold, resulting in a recovery stability feature representing the step behavior's ability to maintain a stable state after recovery. The recovery duration feature, recovery fluctuation feature, and recovery stability feature are combined to form a self-stabilizing feature set Rest, representing the self-stabilizing ability of the step behavior.

[0014] Preferably, S5 includes S51; S51. After obtaining the self-stabilizing feature set Rest, the recovery duration feature, recovery fluctuation feature, and recovery stability feature in the self-stabilizing feature set Rest are processed sequentially; wherein, the number of steps is determined according to the magnitude of the recovery duration feature to determine the time level required to complete the recovery. Based on the changes in the recovery fluctuation characteristics, determine whether the step behavior fluctuates repeatedly during the recovery process; Based on the recovery stability features, determine the ability of step behavior to maintain a stable state after recovery; and according to a preset mapping rule, uniformly map the processing results of the recovery duration features, the recovery fluctuation features, and the recovery stability features to generate a physiological resilience assessment result Eval that represents the user's physiological system's ability to self-recover and maintain stability over a long period of time; and then output the physiological resilience assessment result Eval as a physiological health status assessment result obtained based on step data.

[0015] This invention provides a method for assessing physiological health status based on step count data, which has the following beneficial effects: (1) By constructing the step time set Step and the long-term stable baseline set Base, the evaluation results can reflect the normal step behavior pattern formed by users in a long-term natural life state; by generating the perturbation event set Dist, it is possible to identify the continuous abnormal step behavior caused by real situations such as changes in work intensity, disruption of work and rest or fluctuations in physical condition; by extracting the recovery segment set Rec and the self-stabilizing feature set Rest, it is possible to quantify the time required for users to recover to a long-term stable state after the perturbation occurs, the smoothness of the recovery process, and the ability to maintain stability after recovery, thereby avoiding misjudging the sub-healthy state of meeting the step count target but declining recovery ability as a normal healthy state, making the physiological health status assessment based on step data more in line with real life scenarios and having higher interpretability and application value.

[0016] (2) By comparing the actual step count values ​​of each time segment in the step time set Step with the long-term stable baseline set Base segment by segment and calculating the segment deviation degree Dvd, the judgment of whether the step count behavior deviates from the normal state is based on the individual's long-term behavior pattern, thereby avoiding misjudgment caused by relying solely on fixed thresholds or simple comparisons. By merging the temporally continuous or near-continuous step count deviation segments in the deviation segment set Dev, a perturbation event set Dist is constructed, so that a continuous step count behavior anomaly can be identified as the same step count behavior perturbation event, rather than being split into multiple isolated anomalies.

[0017] (3) By taking the end time segment of each perturbation event in the set of perturbation events Dist as the starting point of the recovery process, and continuously calculating the segment step stability Sds under a unified time window, combined with the dual constraints of the stability threshold and the long-term stable baseline set Base, the time boundary of the actual completion of step behavior recovery can be accurately determined, thereby constructing a recovery segment set Rec with clear boundaries and practical significance. On this basis, by extracting recovery duration features, recovery fluctuation features, and recovery stability features from the recovery segment set Rec, and forming a self-stabilizing feature set Rest, the self-regulation ability of step behavior can be quantitatively characterized from three dimensions: recovery speed, recovery process stability, and post-recovery stability maintenance ability. By mapping the self-stabilizing feature set Rest to generate the physiological elasticity assessment result Eval, it is possible to identify user groups with significantly different recovery abilities when the total number of steps is similar. Attached Figure Description

[0018] Figure 1 This is a schematic diagram illustrating the steps of a physiological health status assessment method based on step count data according to the present invention. Detailed Implementation

[0019] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0020] Example 1 This invention provides a method for assessing physiological health status based on step count data. Please refer to [link / reference]. Figure 1 This includes the following steps: S1. Collect step data generated by the user within a preset monitoring period, and perform time alignment processing on the step data to construct a step time set Step for describing the step change process; S2. Based on the step time set Step, extract step behavior features that meet the stability conditions, and construct a long-term stable baseline set Base to represent the user's normal step behavior pattern; S3. Compare the step time set Step with the long-term stable baseline set Base, identify the perturbation segment where the step behavior deviates from the long-term stable baseline set Base, and generate the corresponding perturbation event set Dist; S4. For the set of disturbance events Dist, construct a recovery process for step behavior regression of the long-term stable baseline set Base, and extract a self-stabilizing feature set Rest from the recovery process to represent the self-stabilizing capability of step behavior; S5. Generate a physiological resilience assessment result Eval based on the self-stabilizing feature set Rest to represent the long-term self-stabilizing capability of the user's physiological system, and output the physiological resilience assessment result Eval.

[0021] In this embodiment, through steps S1 to S5, the present invention provides a physiological health status assessment method based on step count data. Without relying on additional hardware, it constructs a complete technical chain from step count behavior collection, long-term stable baseline modeling, disturbance identification, recovery process analysis, to physiological resilience assessment using only step count data. This method is no longer limited to judging the total number of steps or short-term changes. Instead, by constructing a step count time set (Step) and a long-term stable baseline set (Base), the assessment results can reflect the normal step count behavior patterns formed by users under long-term natural living conditions. Through the generation of a disturbance event set (Dist), it can identify persistent abnormal step count behavior caused by real-world situations such as changes in work intensity, disrupted sleep patterns, or fluctuations in physical condition. Furthermore, by extracting a recovery segment set (Rec) and a self-stabilizing feature set (Rest), the present invention can quantify the time required for a user to recover to a long-term stable state after a disturbance, the smoothness of the recovery process, and the ability to maintain stability after recovery. This avoids misjudging a sub-healthy state where step count meets the standard but recovery ability declines as a normal healthy state. For example, when daily step counts remain stable over a long period, if a user's recovery time for step count behavior is significantly prolonged or stability decreases after recovery due to continuous overtime work or disrupted lifestyle rhythms, the generated physiological resilience assessment result, Eval, can accurately reflect the changing trend of the physiological system's self-stabilizing ability, enabling early identification of chronic fatigue or decreased physiological resilience. Therefore, this invention effectively overcomes the shortcomings of existing technologies that rely solely on step counts or short-term statistics, which are insufficient to reflect the regulatory capacity of the physiological system. This makes step-count-based physiological health status assessments more consistent with real-life scenarios, possessing higher interpretability and application value.

[0022] Example 2 Specifically: S1 includes S11; S11. Within a preset monitoring period, the step count information generated by the user during daily activities is continuously collected through the step count acquisition unit in the user's portable terminal device. The step acquisition unit identifies the user's walking behavior based on the built-in motion perception mechanism, and records the corresponding step increment and occurrence time each time a valid walking behavior is identified; the collected step increments are arranged in order of occurrence time and bound with the corresponding time information to generate a raw step sequence RawStep containing step value and time stamp. It should be noted that: The RawStep sequence is a set of step data directly collected within the preset monitoring period. The data in the RawStep sequence has not undergone time normalization or statistical processing, and its time intervals are allowed to be uneven. By organizing the RawStep sequence in chronological order, the true chronological relationship of the user's steps can be completely preserved, thereby providing a basis for subsequent processing under a unified time scale and helping to avoid information loss due to differences in collection frequency.

[0023] S1 further includes S12; S12. Divide the preset monitoring period into continuous time intervals according to the preset time granularity, and map each step data in the original step sequence RawStep to the corresponding time interval; collect the step data falling into the same time interval, generate step statistics results corresponding to each time interval, and arrange them according to the order of the time intervals, thereby constructing a step time set Step that reflects the step change process under a unified time scale. It should be noted that: The step time set Step is a set of step data formed at a uniform time granularity. Each data item in the step time set Step corresponds to a continuous time interval within the preset monitoring period. By performing time alignment processing on the original step sequence RawStep to generate the step time set Step, the time deviation caused by inconsistent collection time intervals can be eliminated, making the step data comparable and continuous. This provides a unified data foundation for subsequent long-term stable baseline construction, disturbance event identification, and self-stabilization capability analysis.

[0024] In this embodiment, through the settings of steps S11 and S12, the present invention ensures the authenticity, continuity, and traceability of step behavior information during the step data collection and preprocessing stage. Specifically, by directly collecting and forming the raw step sequence RawStep within a preset monitoring period, the step increment and its occurrence time corresponding to each effective walking behavior are completely preserved. Even in real life, where users wear devices for intermittent recording, have uneven activity rhythms, or highly fragmented walking behaviors, the true time sequence of step generation can still be accurately reflected. For example, in scenarios such as daily commuting, short-term activities, or scattered walks at night, the raw step sequence RawStep can retain these irregular step behaviors without being ignored or merged due to inconsistent collection intervals. Furthermore, by performing time alignment processing on the raw step sequence RawStep, a step time set Step under a unified time scale is constructed, so that the originally unevenly distributed step data is mapped to a continuous time interval, thereby eliminating the time deviation caused by different collection frequencies and differences in device recording. This processing method ensures that the Step time set preserves the chronological order of the original step counts while possessing a continuous and comparable data structure. This provides a reliable and consistent data foundation for the subsequent construction of the long-term stable baseline set Base, the identification of the disturbance event set Dist, and the analysis of the self-stabilizing feature set Rest. Therefore, this invention effectively avoids the distortion of behavioral features caused by improper step data preprocessing at the data source level, making the physiological health status assessment based on step data closer to the user's real-life situation and possessing higher practical application value.

[0025] Example 3 Specifically: S2 includes S21; S21. In the step time set Step, the continuous time segment is divided into sliding segments according to a preset time window length; within each time segment, based on the step value corresponding to each consecutive time point within the time segment, the step value between adjacent time points is compared one by one to calculate a step difference set reflecting the step change amplitude, and the number of occurrences of non-zero differences in the step difference set is counted to determine the frequency of step change within the time segment; based on the change amplitude reflected by the step difference set and the change frequency reflected by the number of occurrences of non-zero differences, the segment step stability Sds representing the step fluctuation level within the time segment is calculated; the segment step stability Sds corresponding to each time segment is compared with a preset stability threshold; When the stability of the number of steps in any time interval Sds meets the stability condition, the time interval is determined to be a stable step behavior interval, and all time intervals determined to be stable are gathered to form a stable interval set Sta. When the stability of the number of steps in a time segment, Sds, does not meet the stability condition, the time segment is judged as an unstable step behavior segment. It should be noted that: The segment step stability Sds is a stability result calculated based on the step change characteristics within the same time segment. Its calculation uses the step values ​​corresponding to multiple consecutive time points within the time segment as input data. During the calculation, a set of step difference values ​​reflecting the amplitude of step change is obtained by calculating the difference between the step values ​​corresponding to adjacent time points. Furthermore, a statistical result reflecting the frequency of step change is obtained by counting the number of times the step change occurs within this set of difference values. Based on this, the segment step stability Sds is generated by combining the amplitude and frequency of step change to represent the degree of step fluctuation within the time segment. The smaller the amplitude and the lower the frequency of step change, the lower the corresponding segment step stability Sds. By introducing the segment step stability Sds to perform segment-level analysis on the step time set Step, time segments with drastic step changes can be eliminated before constructing a long-term stable baseline, thereby improving the consistency and reliability of the stable segment set Sta. The stability threshold is a threshold parameter used to determine whether the step stability Sds of a segment meets the stability condition. Its setting is based on the statistical characteristics of the user's historical step data under long-term stable conditions. In the typical setting process, based on historical time segments in the step time set Step that have been determined to be in a stable state, statistical analysis is performed on the corresponding segment step stability Sds to obtain a stability distribution interval reflecting the normal fluctuation range of the user's step count. The upper limit or quantile value in the stability distribution interval is then determined as the stability threshold. By using a stability threshold adaptively generated based on historical stable states, the stability condition can be matched with the step behavior characteristics of different users, avoiding misjudgments of stable segments due to a uniform fixed threshold, thereby improving the accuracy and individual adaptability of the stable segment set Sta.

[0026] S2 further includes S22; S22. Step feature extraction processing is performed on each stable step behavior segment contained in the stable segment set Sta. Within each stable step behavior segment, based on the step value in the step time set Step corresponding to the segment, a segment step count statistical result reflecting the step distribution of the segment is calculated. The segment step count statistical results corresponding to each stable step behavior segment are aggregated across segments to form a long-term step count statistical result representing the overall distribution of the user's step behavior in multiple stable time segments. A long-term stable baseline set Base is constructed based on the long-term step count statistical result. It should be noted that: The long-term stable baseline set Base is a step behavior reference set constructed based on multiple stable step behavior segments. The long-term stable baseline set Base contains statistical data representing the distribution characteristics of the user's step behavior in a long-term stable state. By constructing the long-term stable baseline set Base using only the stable segment set Sta as the input data source, it is possible to avoid including short-term abnormal activities or occasional disturbances in the baseline model, making the long-term stable baseline set Base closer to the long-term step behavior pattern formed by the user under normal physiological conditions. This provides a stable and reliable reference basis for subsequent disturbance event identification and step self-stabilization capability analysis. The deviation threshold is a threshold parameter used to determine whether the degree of deviation of a segment (Dvd) constitutes a deviation in step count behavior. Its setting is based on the normal step count behavior distribution range represented by the long-term stable baseline set Base. In the typical setting process, statistical analysis is performed on the baseline step count distribution in the long-term stable baseline set Base to determine the allowable deviation interval reflecting the fluctuation range of normal step count behavior. The degree of deviation exceeding the allowable deviation interval is used as the basis for determining abnormal step count behavior, thereby determining the deviation threshold. By associating the deviation threshold with the user's own long-term stable baseline set Base, the step count deviation judgment can better conform to the individual's long-term behavioral characteristics, avoiding misjudging normal activity fluctuations as disturbance events and improving the reliability of identifying the disturbance event set Dist.

[0027] In this embodiment, through steps S21 and S22, the present invention introduces a screening mechanism based on time segment stability in the step count behavior analysis stage, ensuring that the construction of a long-term stable baseline is based on reliable and representative step count behavior. Specifically, by introducing segment step count stability Sds into the step count time set Step, the amplitude and frequency of step count changes within continuous time segments are jointly evaluated, and time segments are distinguished by combining a stability threshold. This ensures that only time segments with smooth step count behavior changes and relatively consistent activity rhythms are included in the stable segment set Sta, thereby effectively eliminating segments with drastic step count fluctuations caused by temporary outings, occasional exercise, short-term fatigue, or abnormal work and rest schedules. For example, in a user's daily life, a sudden increase in steps caused by temporarily carrying heavy objects or short-term concentrated exercise on a particular day will not be mistakenly considered as part of the user's long-term normal activity pattern. Furthermore, by constructing a long-term stable baseline set Base based solely on the stable segment set Sta, the obtained baseline can truly reflect the long-term step count behavior distribution characteristics formed by the user under normal physiological conditions, rather than a simple historical average level. Building upon this foundation, by individually associating deviation thresholds with a long-term stable baseline set (Base), subsequent judgments of step count deviations can fully consider the differences in activity habits and rhythms among different users, avoiding misclassification of normal fluctuations consistent with an individual's long-term behavioral characteristics as disturbance events (Dist). Thus, this invention significantly improves the accuracy and individual adaptability of step count behavior analysis in the stable behavior screening and long-term baseline construction stages, providing a reliable and stable reference basis for subsequent disturbance identification and recovery process analysis, making physiological health status assessments based on step count data more closely reflect real-life conditions.

[0028] Example 4 Specifically: S3 includes S31; S31. The step count values ​​corresponding to each time segment in the step count time set Step are compared segment by segment with the baseline step count distribution information contained in the long-term stable baseline set Base. During the comparison process, the difference value of the actual step count value in each time segment relative to the long-term stable baseline set Base is calculated to obtain the segment deviation degree Dvd, which is used to represent the degree of deviation of the step count behavior in the time segment. The segment deviation degree Dvd corresponding to each time segment is compared with a preset deviation threshold. When the deviation degree Dvd of any time segment exceeds the deviation threshold, the time segment is determined to be a step deviation segment, and all time segments determined to be step deviation segments are aggregated to form a deviation segment set Dev. When the deviation degree Dvd of the time segment does not exceed the deviation threshold, the time segment is determined to be a segment where the number of steps has not deviated. It should be noted that: The segment deviation degree Dvd is a calculated result representing the difference between the actual step count behavior within a certain time segment and the normal step count behavior pattern represented by the long-term stable baseline set Base. Its calculation is based on the comparison between the actual step count value within the time segment and the corresponding baseline step count distribution in the long-term stable baseline set Base. During the calculation, by analyzing the magnitude of the difference between the actual step count value and the baseline step count distribution, a deviation result reflecting whether the step count behavior within the time segment significantly deviates from the normal state is obtained. When the segment deviation degree Dvd is high, it indicates that the step count behavior within the time segment has significantly deviated from the long-term stable state. By introducing the segment deviation degree Dvd to perform segment-level comparison of the step count time set Step, the specific segments where abnormal changes in step count behavior occur can be accurately identified in the time dimension, providing a reliable basis for the construction of subsequent disturbance events.

[0029] S3 further includes S32; S32. Merge multiple consecutive step deviation segments in the Dev deviation segment set that are less than a preset merging condition, and treat the merged consecutive deviation segments as the same step behavior disturbance event; record the start time segment, end time segment, and duration information for each step behavior disturbance event, and gather all the constructed step behavior disturbance events to form a disturbance event set Dist used to describe abnormal step behavior disturbances; It should be noted that: The Dist set of disturbance events is a collection of events consisting of one or more step behavior disturbance events. Each disturbance event is formed by merging continuous or near-continuous step deviation segments, representing the state of a user's step behavior continuously deviating from the long-term stable baseline set Base within a certain time range. By integrating multiple adjacent step deviation segments into a complete disturbance event, it is possible to avoid incorrectly splitting the same physiological state change into multiple isolated anomalies, thereby more accurately reflecting the continuous changes in step behavior caused by external loads, changes in work and rest, or fluctuations in physiological state in real life. The Dist set of disturbance events serves as important input data for subsequently constructing the step recovery process and analyzing the step self-stabilization capability, thus elevating the assessment of physiological health status from single-point anomaly identification to dynamic analysis based on time continuity.

[0030] In this embodiment, through steps S31 and S32, the present invention achieves a shift from single-point deviation judgment to continuous behavioral disturbance recognition in the step count behavior anomaly identification stage. Specifically, by comparing the actual step count values ​​of each time segment in the step count time set Step with the long-term stable baseline set Base segment by segment, and calculating the segment deviation degree Dvd, the judgment of whether step count behavior deviates from the normal state is based on the individual's long-term behavioral pattern, thereby avoiding misjudgments caused by relying solely on fixed thresholds or simple comparisons. For example, for users with high long-term activity levels, a short-term increase in step count will not be incorrectly identified as abnormal; only when the segment deviation degree Dvd continuously exceeds the deviation threshold will the relevant time segment be included in the deviation segment set Dev. Furthermore, by merging temporally continuous or near-continuous step count deviation segments in the deviation segment set Dev, a disturbance event set Dist is constructed, so that a continuous step count behavior anomaly can be identified as a single step count behavior disturbance event, rather than being split into multiple isolated anomaly points. For example, when a user's step count declines over several days due to continuous overtime work, irregular sleep patterns, or physical discomfort, this invention can identify this process as a continuous disturbance event, rather than a series of scattered abnormal records. Therefore, the generated disturbance event set Dist more closely reflects the actual process of physiological state changes in real life, providing accurate and continuous time boundaries for the subsequent construction of the recovery segment set Rec and the extraction of the self-stabilizing feature set Rest, thus giving the physiological health status assessment stronger dynamism and real-world interpretability.

[0031] Example 5 Specifically: S4 includes S41; S41. For each disturbance event in the disturbance event set Dist, determine the end time segment corresponding to the disturbance event, and use the end time segment as the start time segment of the recovery process; after the start time segment, analyze the continuous time segments in the step time set Step according to the same time window division method as in S21, and calculate the corresponding segment step stability Sds in each continuous time segment; when the segment step stability Sds corresponding to a consecutive preset number of time segments are all lower than the stability threshold, and the step behavior in the time segment falls back into the normal step behavior range represented by the long-term stable baseline set Base, the earliest time segment is determined as the recovery completion time segment; all continuous time segments from the recovery process start time segment to the recovery completion time segment are collected to form a recovery segment set Rec for representing the step behavior recovery process; It should be noted that: The determination of the recovery completion time segment is based on the same stability determination mechanism as step S21, that is, by calculating the segment step stability Sds corresponding to the continuous time segment and comparing it with the stability threshold. By reusing the stability determination conditions consistent with the stable segment selection in the recovery process analysis, it can be ensured that the technical meaning of stable step behavior remains consistent in different steps, avoiding the deviation in recovery completion time identification due to inconsistent determination standards, thereby improving the interpretability and consistency of the recovery segment set Rec construction results. During the determination of the recovery completion time segment, it is required that the stability of the step count Sds corresponding to a consecutive preset number of time segments is lower than the stability threshold. The purpose is to avoid misjudging a temporary stable state of step count behavior as a true recovery completion state. Since there are often periods of reduced fluctuations but not yet fully recovered during the transition of step count behavior from a disturbed state to a stable state, if the recovery completion is determined based solely on the stability of a single time segment, it is easy to misidentify temporary relief as stable recovery. By introducing a consecutive preset number of time segments as the determination condition, it is possible to ensure that the step count behavior remains stable in the time dimension, thereby improving the reliability of the recovery completion determination and avoiding premature termination of the recovery process analysis. At the same time, it is required that the step count behavior within the time period falls back into the normal step count behavior range represented by the long-term stable baseline set Base. The purpose is to ensure that the step count behavior not only tends to stabilize in terms of fluctuation, but also returns to the normal behavior pattern formed by the user in the long term at the overall level. Only when the step count behavior conforms to the normal distribution range reflected by the long-term stable baseline set Base can it be considered that the user's activity state has substantially returned from the disturbed state to the long-term stable state, thereby avoiding misjudging the state that is continuously at an abnormal level but has small fluctuations as the recovery completed. By simultaneously introducing the continuous stability determination condition and the baseline regression determination condition, the recovery completion state can be jointly constrained from two dimensions: temporal continuity and behavioral consistency. This ensures that the recovery completion determination has both process stability and state rationality, thereby more accurately depicting the true recovery of user step behavior after disturbance. This joint determination mechanism provides reliable and consistent boundary conditions for the subsequent construction of the recovery segment set Rec and the extraction of the self-stabilizing feature set Rest, making the analysis results of the step behavior self-stabilization capability more interpretable and practically valuable.

[0032] S4 further includes S42; S42. After the set of recovery segments Rec is determined, all time segments contained in the set of recovery segments Rec are analyzed. Based on the step data in the step time set Step corresponding to each time segment and the corresponding segment step stability Sds, recovery features representing the self-stabilizing ability of step behavior are extracted. The recovery features include recovery duration features and recovery fluctuation features. Specifically, by counting the number of time segments in the recovery segment set Rec, a recovery duration feature is obtained, representing the time elapsed from the start time segment to the completion time segment of the recovery process. By analyzing the changes in the segment step stability Sds corresponding to each time segment in the recovery segment set Rec during the recovery process, a recovery fluctuation feature is obtained, representing whether the step behavior experiences repeated fluctuations during the recovery process. After the completion time segment, following the same time window division method as in S41, the corresponding segment step stability Sds is calculated for a predetermined number of consecutive time segments in the step time set Step, and compared with the stability threshold, resulting in a recovery stability feature representing the step behavior's ability to maintain a stable state after recovery. The recovery duration feature, recovery fluctuation feature, and recovery stability feature are combined to form a self-stabilizing feature set Rest, representing the self-stabilizing ability of the step behavior. It should be noted that: The self-stabilizing feature set Rest is a feature set extracted based on the recovery segment set Rec and the stability of step behavior after the recovery completion time segment. The recovery segment set Rec is explicitly defined in step S41 as all consecutive time segments from the start time segment of the recovery process to the recovery completion time segment. The recovery duration feature is determined based on the number of time segments in the recovery segment set Rec, representing the time span required for step behavior to complete regression; a shorter recovery duration indicates higher self-regulation efficiency of step behavior. The recovery fluctuation feature is determined based on the changes in segment step stability Sds corresponding to each time segment in the recovery segment set Rec, representing whether step behavior fluctuates repeatedly during the recovery process. If segment step stability Sds increases multiple times during the recovery process, it indicates instability in the recovery process. The recovery stability feature is determined based on the comparison between the segment step stability Sds corresponding to a preset number of consecutive time segments after the recovery completion time segment and the stability threshold, representing the ability of step behavior to maintain a stable state after completing regression. By extracting the various recovery features without changing the boundary definition of the recovery segment set Rec, and incorporating them into the self-stabilizing feature set Rest, the self-stabilizing ability of step behavior can be comprehensively characterized from multiple aspects such as time span, process stability, and post-recovery stability while maintaining the consistency of object relationships. This provides a clear, reliable, and interpretable input basis for the subsequent generation of the physiological resilience assessment result Eval.

[0033] S5 includes S51; S51. After obtaining the self-stabilizing feature set Rest, the recovery duration feature, recovery fluctuation feature, and recovery stability feature in the self-stabilizing feature set Rest are processed sequentially; wherein, the number of steps is determined according to the magnitude of the recovery duration feature to determine the time level required to complete the recovery. Based on the changes in the recovery fluctuation characteristics, determine whether the step behavior fluctuates repeatedly during the recovery process; Based on the recovery stabilization characteristics, the ability of step count behavior to maintain a stable state after recovery is determined; and according to a preset mapping rule, the processing results of the recovery duration characteristics, the recovery fluctuation characteristics, and the recovery stabilization characteristics are uniformly mapped to generate a physiological resilience assessment result Eval, which represents the user's physiological system's ability to self-recover and maintain stability over a long-term scale; and then the physiological resilience assessment result Eval is output as a physiological health status assessment result based on step count data. It should be noted that: The physiological resilience assessment result Eval is generated based on the recovery duration feature, recovery fluctuation feature and recovery stability feature contained in the self-stabilizing feature set Rest. The determination of each type of recovery feature comes from the statistical analysis of the number of specific time segments or the stability judgment result. Specifically, the recovery duration feature is determined by statistically analyzing the number of time segments in the recovery segment set Rec, from the start time segment to the completion time segment of the recovery process. This number of time segments directly reflects the time span experienced by the step behavior to complete the regression, thus indicating the speed of step behavior recovery. The recovery fluctuation feature is determined by analyzing the changes in the segment step stability Sds corresponding to each time segment in the recovery segment set Rec during the recovery process. When the segment step stability Sds shows multiple increases or large fluctuations during the recovery process, it is determined that the recovery process has repeated fluctuations, thus reflecting the stability of the step behavior recovery process. The recovery stability feature is determined by calculating the segment step stability Sds corresponding to a preset number of consecutive time segments after the recovery completion time segment and determining whether it is continuously lower than the stability threshold. When consecutive time segments meet the stability condition, it indicates that the step behavior can maintain a stable state after recovery. When generating the physiological resilience assessment result Eval, the results corresponding to the above-mentioned recovery characteristics are mapped to the same assessment result according to a preset mapping logic. Among them, a higher level of physiological resilience corresponds to a shorter recovery time, smaller fluctuations in the recovery process, and stronger stability after recovery; conversely, a lower level of physiological resilience corresponds to a longer recovery time, significant fluctuations in the recovery process, or weaker stability after recovery. By generating the physiological resilience assessment result Eval based on the above-mentioned clearly sourced recovery characteristics and using a consistent mapping logic, it is possible to avoid relying on experience judgment or a single indicator to assess physiological health status, so that the assessment result has clear data basis and stable judgment logic.

[0034] In this embodiment, through steps S41, S42, and S51, the present invention introduces a capability-level analysis mechanism based on the recovery process in the assessment of physiological health status. This extends the assessment results from whether an abnormality occurs to whether the individual possesses the ability to self-recover and maintain long-term stability after the abnormality occurs. Specifically, by using the end time segment of each perturbation event in the set of perturbation events Dist as the starting point of the recovery process, and continuously calculating the segment's step stability Sds within a unified time window, combined with the dual constraints of the stability threshold and the long-term stable baseline set Base, the time boundary for the actual completion of step behavior recovery can be accurately determined, thereby constructing a clearly defined and realistically meaningful set of recovery segments Rec. For example, when a user's step behavior declines due to continuous overtime work, physical discomfort, or disrupted lifestyle, the present invention can distinguish between a state of temporary relief but still at an abnormal level and a state of having truly returned to a long-term normal activity pattern, avoiding the misjudgment of superficial stability as complete recovery. Building upon this foundation, by extracting recovery duration, fluctuation, and stability features from the recovery segment set Rec, and forming a self-stabilizing feature set Rest, the self-regulation ability of step-counting behavior can be quantitatively characterized from three dimensions: recovery speed, recovery process stability, and post-recovery stability. Furthermore, by mapping the self-stabilizing feature set Rest to generate a physiological resilience assessment result Eval, this invention can identify user groups with significantly different recovery abilities despite similar total step counts. For example, it can distinguish between individuals who can recover quickly in the short term and maintain long-term stability, and those who recover slowly and are prone to further fluctuations, thus reflecting the differences in the long-term self-stabilizing ability of the physiological system. Therefore, this invention effectively overcomes the shortcomings of existing technologies that rely solely on step count levels or the frequency of abnormal occurrences to reflect the regulatory ability of the physiological system, making step-counting-based physiological health status assessments more consistent with the actual patterns of physiological state changes, and possessing higher predictive value and health management guidance significance.

[0035] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended technical solutions and their equivalents.

Claims

1. A method for evaluating a physiological health state based on step data, characterized by: The method comprises the following steps: S1, collecting step data generated by a user in a preset monitoring period, and performing time alignment processing on the step data to construct a step time set for describing a step change process; S2, based on the step time set Step, extracting step behavior characteristics satisfying a stability condition to construct a long-term stable baseline set Base representing normal step behavior patterns of the user; S3, comparing the step time set Step with the long-term stable baseline set Base to identify disturbance segments of step behavior deviating from the long-term stable baseline set Base, and generating a corresponding disturbance event set Dist; S4, constructing a recovery process of step behavior returning to the long-term stable baseline set Base for the disturbance event set Dist, and extracting a self-stabilization feature set Rest representing the self-stabilization ability of step behavior from the recovery process; S5, generating a physiological elasticity evaluation result Eval representing the long-term self-stabilization ability of the physiological system of the user based on the self-stabilization feature set Rest, and outputting the physiological elasticity evaluation result Eval.

2. The physiological health status evaluation method based on step data according to claim 1, characterized in that: The S1 comprises S11; S11, in a preset monitoring period, continuously collecting step information generated by a user in daily activities through a step acquisition unit in a terminal device carried by the user; Arranging the collected step increments in chronological order and binding them with corresponding time information to generate an original step sequence RawStep containing step values and time markers. 3.The physiological health state evaluation method based on step data according to claim 2, characterized in that: The S1 further comprises S12; S12, dividing the preset monitoring period into continuous time intervals according to a preset time granularity, and mapping each step data in the original step sequence RawStep to a corresponding time interval; Collecting step data falling into the same time interval to generate step statistical results corresponding to each time interval one by one, and arranging them in chronological order to construct a step time set Step reflecting the step change process under a unified time scale.

4. The physiological health status evaluation method based on step data according to claim 3, characterized in that: The S2 comprises S21; S21, in the step time set Step, slidingly dividing continuous time segments according to a preset time window length; in each time segment, comparing step values between adjacent time points one by one based on step values corresponding to each continuous time point in the time segment, calculating a step difference set reflecting step change amplitudes, and counting the number of non-zero difference values in the step difference set to determine the frequency of step change in the time segment; according to the change amplitude reflected by the step difference set and the change frequency reflected by the number of non-zero difference values, calculating a segment step stability Sds representing the step fluctuation level in the time segment; comparing the segment step stability Sds corresponding to each time segment with a preset stability threshold; When the segment step stability Sds of any time segment meets the stability condition, the time segment is determined as a stable step behavior segment, and all time segments determined as stable are collected to form a stable segment set Sta; When the segment step stability Sds of any time segment does not meet the stability condition, the time segment is determined as an unstable step behavior segment.

5. The method of claim 4, wherein the method further comprises: The S2 further comprises S22; S22, performing step feature extraction processing on each stable step behavior segment contained in the stable segment set Sta, and in each stable step behavior segment, based on the step value in the step time set Step corresponding to the segment, calculating the segment step statistical result reflecting the step distribution of the segment; The segment step statistical result corresponding to each stable step behavior segment is processed across segments to form a long-term step statistical result for representing the overall distribution of the user's step behavior in multiple stable time segments, and a long-term stable baseline set Base is constructed based on the long-term step statistical result.

6. The physiological health status evaluation method based on step data according to claim 5, characterized in that: The S3 comprises S31; S31, comparing the step value corresponding to each time segment in the step time set Step with the baseline step distribution information contained in the long-term stable baseline set Base segment by segment; in the comparison process, the difference value of the actual step value in each time segment relative to the long-term stable baseline set Base is calculated to obtain the segment deviation degree Dvd for representing the deviation degree of the step behavior of the time segment; the segment deviation degree Dvd corresponding to each time segment is compared with a preset deviation threshold value; When the segment deviation degree Dvd of any time segment exceeds the deviation threshold value, the time segment is determined as a step deviation segment, and all time segments determined as step deviation segments are collected to form a deviation segment set Dev; When the segment deviation degree Dvd of any time segment does not exceed the deviation threshold value, the time segment is determined as a step non-deviation segment.

7. The method of claim 6, wherein the method further comprises: The S3 further comprises S32; S32, performing merging processing on multiple step deviation segments in the deviation segment set Dev that are continuous to each other or have a time interval less than a preset merging condition, and taking the merged continuous deviation segment as a same step behavior disturbance event; The starting time segment, ending time segment and duration information of each step behavior disturbance event are recorded, and all constructed step behavior disturbance events are collected to form a disturbance event set Dist for describing the abnormal disturbance of the step behavior. 8.The method of claim 7, wherein the method further comprises: The S4 comprises S41; S41, for each disturbance event in the disturbance event set Dist, determining the ending time segment corresponding to the disturbance event, and taking the ending time segment as the starting time segment of the recovery process; after the starting time segment, the continuous time segments in the step time set Step are analyzed according to the same time window division manner as in S21, and the corresponding segment step stability Sds is calculated in each continuous time segment; When the segment step stability Sds corresponding to a continuous preset number of time segments are all lower than the stability threshold, and the step behavior in the time segments falls back into the normal step behavior range represented by the long-term stability baseline set Base, the earliest time segment is determined as the recovery completion time segment; all continuous time segments from the recovery process starting time segment to the recovery completion time segment are collected to form a recovery segment set Rec representing the step behavior recovery process. 9.The physiological health state evaluation method based on step data according to claim 8, characterized in that: The S4 further includes S42; S42, after the recovery segment set Rec is determined, analyzing all time segments included in the recovery segment set Rec, and based on the step data in the step time set Step corresponding to each time segment and the segment step stability Sds corresponding thereto, extracting a recovery feature representing the self-stabilization ability of the step behavior; The recovery feature includes a recovery duration feature and a recovery fluctuation feature; wherein the recovery duration feature representing the length of time experienced by the step behavior from the recovery process starting time segment to the recovery completion time segment is obtained by counting the number of time segments included in the recovery segment set Rec; the recovery fluctuation feature representing whether the step behavior exists repeated fluctuations in the recovery process is obtained by analyzing the changes of the segment step stability Sds corresponding to each time segment in the recovery segment set Rec in the recovery process; the recovery stability feature representing the ability of the step behavior to maintain a stable state after recovery is completed is obtained by continuing to calculate the segment step stability Sds corresponding to a continuous preset number of time segments in the step time set Step after the recovery completion time segment according to the same time window division manner in S41, and comparing with the stability threshold; and the recovery duration feature, the recovery fluctuation feature and the recovery stability feature are collected to form a self-stabilization feature set Rest representing the self-stabilization ability of the step behavior.

10. The method of claim 9, wherein the method further comprises: The S5 includes S51; S51, after obtaining the self-stabilization feature set Rest, sequentially processing the recovery duration feature, the recovery fluctuation feature and the recovery stability feature in the self-stabilization feature set Rest; wherein the time level required for the step behavior to complete recovery is determined according to the size of the recovery duration feature; whether the step behavior exists repeated fluctuations in the recovery process is determined according to the changes of the recovery fluctuation feature; the duration capability of the step behavior to maintain a stable state after recovery is completed is determined according to the recovery stability feature; and the processing results of the recovery duration feature, the recovery fluctuation feature and the recovery stability feature are uniformly mapped according to a preset mapping rule to generate a physiological elasticity evaluation result Eval representing the self-recovery and stable maintenance ability of the user's physiological system in a long-term scale; and the physiological elasticity evaluation result Eval is output as a physiological health status evaluation result based on the step data.

Images