A control test method and system for verifying the effect of health protection measures on team performance improvement

Abstract

This invention discloses a controlled trial method and system for verifying the effect of health protection measures on team performance improvement. The method includes: synchronously collecting multimodal time series data; using a dynamic time warping algorithm to complete cross-device time alignment and using an isolated forest to remove outliers in real time; using a dual-mechanism health index model of "time decay + exponential evolution" to dynamically predict individual health trends; calculating team performance indicators based on task priority weighting and recording performance failure times; randomly grouping teams using team size, historical performance, and health baseline as three-dimensional balancing factors, with the experimental group receiving intelligent fitness and psychological assessment interventions; and constructing multiple linear regression and Cox proportional hazards models in parallel, using the health index as a time-dependent covariate to estimate the net effect of the intervention on the mean performance and failure risk. This invention significantly improves data synchronization accuracy and real-time cleaning, while shortening the management decision-making closed-loop time, providing enterprises with a digital health intervention solution with quantifiable ROI.

Description

Technical Field

[0001] This invention belongs to the interdisciplinary field of digital health and organizational behavior, and in particular relates to a controlled trial method and system for verifying the effect of health protection measures on team performance improvement. Background Technology

[0002] With the widespread adoption of knowledge-intensive and agile development models, team performance is no longer simply a measure of task completion speed, but a dynamic result of the coupling of "people-task-time." Numerous empirical studies have shown that members' physical health, psychological state, and work behavior have a significant impact on team delivery quality, innovation efficiency, and project risk. However, despite the rapid deployment of digital health management tools, enterprises still face three major pain points: First, data collection is fragmented. Physiological signals rely on wearable devices, while behavioral and performance data are scattered across project management software, version control systems, and time reporting platforms. The varying time granularity, sampling frequency, and semantic formats make it difficult to directly integrate heterogeneous data from "the same team, different timelines." Traditional methods, such as simple time window alignment or manual annotation, suffer from large errors and poor timeliness, failing to support real-time decision-making.

[0003] Second, health measurement is static. Existing technologies mostly use arithmetic averages or linear weighting of raw indicators such as heart rate and steps, ignoring the lag and rebound effects of health status, and failing to characterize the dynamic time lag of "intervention-response". As a result, after the implementation of health programs, the improvement in members' physical condition cannot be quantitatively verified at the performance level, and the ROI is difficult to convince management to continue investing.

[0004] Third, the technical models for effect evaluation are weak. Most mainstream approaches employ independent cross-sectional statistical models, such as t-tests or linear regression. Essentially, these techniques simplify the dynamic, continuous intervention process into a static binary variable (intervention / no intervention), and forcibly transform potentially right-censored performance failure time data into a single performance score point. This approach cannot characterize the time-varying effects of health status at the mathematical model level, nor can it handle the changing patterns of performance failure risk over time. This results in insufficient statistical power and unreliable evaluation conclusions in long-term, complex task scenarios.

[0005] To address the aforementioned pain points, the academic community has proposed algorithms such as DTW (Dynamic Time Warping) alignment, iForest anomaly detection, and survival analysis, but all of them remain at the stage of single-point technology verification. DTW is mostly used for voice recognition and action recognition, and has not yet been coupled with Bluetooth Low Energy (BLE) wearable streaming data in team scenarios; iForest typically runs on the server side in an offline batch processing mode, which cannot meet the motion intervention needs of real-time cleaning and instant feedback at the edge. While the Cox proportional hazards model can handle "time-to-event" data, it rarely incorporates "health index" as a time-dependent covariate into team performance failure studies, and it also fails to address confounding bias caused by imbalanced group baselines.

[0006] Furthermore, while existing HR-Tech technologies disclose a general framework of "collection-cleaning-assessment," they lack the following key features: a) A dual-mechanism health index model, which simultaneously introduces two paths: "time decay integral" and "index evolution prediction", so that health status can reflect historical accumulation and predict inflection points in advance. b) A multi-layered fusion architecture is used to denoise the original sensing waveform at the pixel layer, perform dynamic importance weighting at the feature layer, and output interpretable health labels at the decision layer, thereby solving the model mismatch caused by sensor quality drift. c) The three-dimensional balanced random grouping algorithm incorporates team size, historical performance, and health baseline into the Mahalanobis distance minimization objective function, ensuring that the experimental group and the control group are on the same starting line of "health-performance" before the intervention, and avoiding the "healthy employee bias" from exaggerating the intervention effect. d) Parallel evaluation strategy, that is, multiple linear regression and Cox survival analysis are run simultaneously, which answers both "how much the mean performance has improved" and "how much the risk of performance failure has decreased", providing management with a dual perspective of short-term ROI and long-term risk control.

[0007] In summary, existing technologies have failed to provide an end-to-end, real-time, interpretable, and statistically rigorous solution to verify the causal effects of digital health interventions on team performance. This is precisely the technological bottleneck that this invention aims to overcome. Summary of the Invention

[0008] To address the shortcomings of the existing technology, this invention provides a controlled trial method for verifying the effect of health protection measures on team performance improvement, comprising the following steps: a) Collect multimodal time series data of team members synchronously via wearable sensors and team management API, including at least heart rate variability, steps, sleep depth, focus time, rest interval, milestone completion time and error rate; b) The dynamic time warping algorithm is used to align sequences with different sampling frequencies to a unified time axis. Then, sensor drift outliers are removed through isolated forest and the preprocessed dataset is obtained by Z-score normalization. c) On the preprocessed dataset, run the dual-mechanism health index model of "time decay + exponential evolution" to dynamically predict individual health trends and generate a health index sequence that is updated synchronously with health interventions; d) Calculate the team performance indicator sequence based on task priority weighting and record the performance failure time; e) Using team size, historical performance, and health baseline as three-dimensional balancing factors, the experimental group and the control group were randomly divided. The experimental group received smart fitness and psychological assessment intervention. f) Construct multiple linear regression and Cox proportional hazards models in parallel, using the health index as a time-dependent covariate, to estimate the net effect of the intervention on performance indicators and performance failure time, and output a statistical significance report including the risk ratio.

[0009] Among them, the dual-mechanism health index model integrates age, exercise intensity, diet and psychological state through index evolution terms, and strengthens the weight of recent physiological behavior data with decay coefficients.

[0010] Among them, the multi-level fusion framework dynamically weights the importance of heterogeneous sensor data at the pixel layer, feature layer and decision layer, and adjusts the feature weights in real time with data quality indicators.

[0011] The Cox model defines performance failure time as the moment when two consecutive milestones are overdue or the error rate exceeds a threshold for the first time.

[0012] Among them, the three-dimensional balancing algorithm adopts the principle of minimizing Mahalanobis distance to ensure that there are no significant differences between the experimental group and the control group in terms of size, historical performance, and health baseline vector.

[0013] Among them, the dynamic time warping algorithm introduces sensor confidence weights when backtracking the optimal path, reducing the distortion of alignment results by low-confidence sampling points.

[0014] Among them, intelligent fitness intervention includes adjusting the exercise intensity and rest intervals suggested by wearable devices in real time based on health indices.

[0015] The effectiveness evaluation report further provides the percentage decrease in the performance failure risk ratio corresponding to each unit increase in the health index during the intervention period.

[0016] All steps are completed within the edge-cloud collaborative computing framework. Data cleaning and real-time updates of health indices are performed at the edge, while Cox model parameter estimation and group balancing calculations are performed in the cloud.

[0017] The present invention also proposes a controlled trial system for verifying the effect of health protection measures on team performance improvement by performing the above method, comprising: The system includes a multimodal synchronous acquisition unit, a DTW alignment and isolated forest cleaning unit, a dual-mechanism health index calculation unit, a multi-level fusion weighting unit, a three-dimensional balanced random grouping unit, a parallel evaluation unit for multiple regression and Cox survival analysis, and an edge-cloud collaborative control unit for outputting a statistical report containing risk ratios. Through end-to-end innovation encompassing "data acquisition - time alignment - anomaly cleaning - dual-mechanism health index - three-dimensional balanced grouping - parallel statistical evaluation," the following quantifiable, reproducible, and implementable technical effects have been achieved: Data synchronization accuracy improved by 37%. Traditional time window alignment results in an average phase error of ±8.7 seconds when the sampling frequency difference across devices is greater than 2 times. Under the same experimental conditions, compared with the traditional linear interpolation alignment method based on a fixed time window, this invention adopts a BLE+DTW fusion strategy, which incorporates the Euclidean distance matrix and sensor confidence weights into the backtracking path, reducing the phase error to ±5.5 seconds. This ensures that heart rate variability and code submission events are aligned within seconds, providing a high-resolution time reference for subsequent causal inference.

[0018] Abnormal data cleaning real-time performance <200ms The lightweight iForest is deployed at the edge, retaining only ≤6 random trees and occupying 1.8MB of memory. It uses RSSI in BLE broadcast frames as prior weights to prune outlier paths in advance, reducing the inference latency of a single record from 820ms in the cloud to 178ms locally, achieving "collecting and cleaning on the spot" and avoiding dirty data from entering the health index model and causing misjudgments.

[0019] The prediction error of the health index (MAPE) decreased by 42%. In the dual-mechanism model, the time decay integral captures the 7-day historical inertia, while the exponential evolution term anticipates the 3-day trend. In a 28-day controlled experiment with 120 participants, the MAPE decreased from 18.4% in the traditional linear weighted model to 10.6%, and the recall rate for the 24-hour early warning of "overwork" increased from 0.68 to 0.85, providing HR with a decision-making window for timely scheduling of compensatory time off.

[0020] Baseline imbalance (standardized mean difference |SMD|) < 0.05 The three-dimensional balancing algorithm iterates and swaps 20 times after random assignment, minimizing the Mahalanobis distance between the experimental group and the control group in terms of size, historical performance, and healthy baseline vector. As a result, |SMD| decreased from 0.12 in the conventional randomization to 0.04, reaching the "near-random perfect" standard, significantly reducing confounding bias and making subsequent causal estimation more reliable.

[0021] The performance failure risk ratio decreased by 28%. The Cox model uses the health index as a time-dependent covariate. Over the 8-week intervention period, for every 1 unit increase in the health index in the experimental group, the risk ratio of performance failure decreased by 28% (HR=0.72, 95% CI: 0.58-0.89, p<0.01). This means that if the team's average health index increases by 2 units, the probability of project delays or major defects will be halved, saving the company approximately RMB 156,000 per 100 person-times in potential delay costs.

[0022] Short-term performance improved by an average of 12.4%. Multiple linear regression showed that, after controlling for health index, the intervention had a net effect of -12.4% on weighted task completion time (p<0.05) and a net effect of -9.7% on normalized error rate (p<0.05). After combining the two indicators, it was equivalent to saving 1.3 person-days per iteration cycle. A team of 100 people could free up about 326 person-days per year for innovative tasks or skills training.

[0023] Enhanced model interpretability The multi-level fusion framework outputs dynamic weights of feature importance at the decision-making level, allowing HR to view the percentage contribution of "heart rate variability" and "focus time" to the current health index in real time. When the weights fluctuate abnormally, the system alerts to sensor drift or sudden changes in task load, achieving "white-box" operation and maintenance and reducing algorithm skepticism.

[0024] System deployment costs reduced by 46% Through edge-cloud collaboration, 80% of the computation is completed at the edge nodes, while the cloud is only responsible for retraining and group balancing, saving cloud GPU time. For a company with 1,000 employees, the three-year TCO is reduced from RMB 1.22 million to RMB 660,000 for the traditional all-cloud solution, and the ROI cycle is shortened from 18 months to 9 months.

[0025] Improved employee privacy compliance Sensitive physiological data cleaning and health index calculation are completed on the edge side, and only the anonymization index and performance results are uploaded, which meets the minimum necessary principle of GDPR and the Personal Information Protection Act; after third-party penetration testing, the PII leakage risk score is "low", and it has passed the ISO 27701 certification audit.

[0026] The management decision-making closed-loop time has been shortened from a week to a day. The traditional research-report-decision cycle averages 10.5 days; this invention automatically generates a visualized report containing risk ratio, confidence interval, and feature contribution on T+1 day. Management can adjust fitness plans or task assignments during daily stand-up meetings, forming a "data-insight-action" closed loop to continuously improve team resilience and delivery quality. Attached Figure Description

[0027] The above and other objects, features, and advantages of exemplary embodiments of the present disclosure will become readily apparent upon reading the following detailed description with reference to the accompanying drawings. In the drawings, several embodiments of the present disclosure are illustrated by way of example and not limitation, and like or corresponding reference numerals denote like or corresponding parts, wherein: Figure 1 This is a flowchart illustrating a controlled trial method for verifying the effect of health protection measures on team performance improvement according to an embodiment of the present invention; Figure 2 This is a schematic diagram illustrating a controlled trial system for verifying the effect of health protection measures on team performance improvement according to an embodiment of the present invention. Detailed Implementation

[0028] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this invention, and not all of them. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0029] The terminology used in the embodiments of this invention is for the purpose of describing particular embodiments only and is not intended to limit the invention. The singular forms “a,” “the,” and “the” as used in the embodiments of this invention and the appended claims are also intended to include the plural forms, and “multiple” generally includes at least two unless the context clearly indicates otherwise.

[0030] It should be understood that although the terms first, second, third, etc., may be used to describe... in the embodiments of the present invention, these... should not be limited to these terms. These terms are only used to distinguish... For example, first... may also be referred to as second... without departing from the scope of the embodiments of the present invention, and similarly, second... may also be referred to as first...

[0031] It should be understood that the term "and / or" used in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone. Additionally, the character " / " in this article generally indicates that the preceding and following related objects have an "or" relationship.

[0032] Depending on the context, the words “if” or “suppose” as used here can be interpreted as “when” or “in response to determination” or “in response to detection.” Similarly, depending on the context, the phrases “if determination” or “if detection (of the stated condition or event)” can be interpreted as “when determination” or “in response to determination” or “when detection (of the stated condition or event)” or “in response to detection (of the stated condition or event).”

[0033] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that an article or device that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such an article or device. Without further limitation, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the article or device that includes said element.

[0034] The optional embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0035] Example 1 like Figure 1 As shown, this invention discloses a controlled trial method for verifying the effect of health protection measures on team performance improvement, comprising the following steps: a) Collect multimodal time series data of team members synchronously via wearable sensors and team management API, including at least heart rate variability, steps, sleep depth, focus time, rest interval, milestone completion time and error rate; b) The dynamic time warping algorithm is used to align sequences with different sampling frequencies to a unified time axis. Then, sensor drift outliers are removed through isolated forest and the preprocessed dataset is obtained by Z-score normalization. c) On the preprocessed dataset, run the dual-mechanism health index model of "time decay + exponential evolution" to dynamically predict individual health trends and generate a health index sequence that is updated synchronously with health interventions; d) Calculate the team performance indicator sequence based on task priority weighting and record the performance failure time; e) Using team size, historical performance, and health baseline as three-dimensional balancing factors, the experimental group and the control group were randomly divided. The experimental group received smart fitness and psychological assessment intervention. f) Construct multiple linear regression and Cox proportional hazards models in parallel, using the health index as a time-dependent covariate, to estimate the net effect of the intervention on performance indicators and performance failure time, and output a statistical significance report including the risk ratio.

[0036] Among them, the dual-mechanism health index model integrates age, exercise intensity, diet and psychological state through index evolution terms, and strengthens the weight of recent physiological behavior data with decay coefficients.

[0037] Among them, the multi-level fusion framework dynamically weights the importance of heterogeneous sensor data at the pixel layer, feature layer and decision layer, and adjusts the feature weights in real time with data quality indicators.

[0038] The Cox model defines performance failure time as the moment when two consecutive milestones are overdue or the error rate exceeds a threshold for the first time.

[0039] Among them, the three-dimensional balancing algorithm adopts the principle of minimizing Mahalanobis distance to ensure that there are no significant differences between the experimental group and the control group in terms of size, historical performance, and health baseline vector.

[0040] Among them, the dynamic time warping algorithm introduces sensor confidence weights when backtracking the optimal path, reducing the distortion of alignment results by low-confidence sampling points.

[0041] Among them, intelligent fitness intervention includes adjusting the exercise intensity and rest intervals suggested by wearable devices in real time based on health indices.

[0042] The effectiveness evaluation report further provides the percentage decrease in the performance failure risk ratio corresponding to each unit increase in the health index during the intervention period.

[0043] All steps are completed within the edge-cloud collaborative computing framework. Data cleaning and real-time updates of health indices are performed at the edge, while Cox model parameter estimation and group balancing calculations are performed in the cloud.

[0044] Example 2 The present invention proposes a controlled trial method for verifying the effect of health protection measures on team performance improvement, comprising the following steps: Data collection step: collecting physiological data, behavioral data and team performance data of team members to generate a raw dataset; Data synchronization steps: Time-align the time-series data in the original dataset to generate a synchronized dataset; Data preprocessing steps: Clean and standardize the synchronized dataset to generate a preprocessed dataset; Health index calculation steps: Calculate the health index of each member based on the preprocessed dataset to generate a health index sequence; Team performance calculation steps: Calculate team performance indicators based on the preprocessed dataset and generate a performance indicator sequence; Controlled trial design steps: Randomly assign the team to the experimental group and the control group. The experimental group receives health protection measures, while the control group does not. Generate group information. Effectiveness evaluation steps: Based on the health index sequence, performance indicator sequence, and grouping information, construct a statistical model to evaluate the effect of health protection measures on team performance and generate an effectiveness evaluation report; Specifically, the output of the data acquisition step serves as the input of the data synchronization step, the output of the data synchronization step serves as the input of the data preprocessing step, the output of the data preprocessing step serves as the input of the health index calculation step and the team performance calculation step, and the outputs of the health index calculation step, the team performance calculation step, and the controlled trial design step serve as the inputs of the effect evaluation step.

[0045] In the data collection step, physiological data includes heart rate variability, daily steps and sleep depth; behavioral data includes work focus time and rest intervals; team performance data includes project milestone completion time and code submission error rate; and the original dataset contains timestamps and multidimensional data streams.

[0046] The data synchronization step employs a dynamic time warping algorithm to align the time series data, whereby the dynamic time warping algorithm minimizes the cumulative distance between different sequences to eliminate time offset.

[0047] The dynamic time warping algorithm includes: calculating the Euclidean distance matrix between two time series, filling the cumulative distance matrix using a dynamic programming method, and determining the optimal curved path by backtracking to align sequences with different sampling frequencies to a unified time axis.

[0048] In the data preprocessing step, data cleaning includes identifying and removing outliers caused by sensor malfunctions or environmental noise, and standardization adopts a statistical Z-score normalization method to make each dimension of data have zero mean and unit variance.

[0049] The data cleaning process employs the isolated forest algorithm to detect and remove outliers. This algorithm identifies outliers in the data points by constructing a random tree structure.

[0050] The health index calculation step employs a health index model, which reflects the dynamic changes in health by fusing multimodal data and applying a time decay function. The health index model prioritizes the impact of recent data.

[0051] The health index model uses the following formula to calculate the health index H of member i at time t. i( t): Among them, H i(t) represents the health index of member i at time t, where t is the current time, τ is the integration variable, T is the observation time period, λ is the decay coefficient, β is the scaling parameter, θ is the threshold parameter, and x i (τ) represents the physiological data value of member i at time τ, and exp is an exponential function.

[0052] In the team performance calculation step, the team performance indicators include weighted average task completion time and normalized overall error rate, and the weighted average is based on the task priority.

[0053] In the controlled trial design steps, health safeguards include providing smart fitness equipment and regular mental health assessments, and grouping ensures baseline balance between groups based on team size and historical performance baselines.

[0054] In the effect evaluation step, the statistical model is a multiple linear regression model, with health protection measures as binary independent variables, team performance indicators as dependent variables, and health index series as covariates, and the least squares method is used to estimate the model parameters.

[0055] The effectiveness evaluation step also includes calculating statistical significance indicators, including p-values ​​and confidence intervals, and the generated effectiveness evaluation report provides a quantitative analysis of the impact of health protection measures on performance.

[0056] The data acquisition step is achieved through wearable sensors and team management software. The wearable sensors use Bluetooth Low Energy technology to transmit real-time physiological data, and the team management software integrates behavioral and performance data through an application programming interface.

[0057] The health index calculation step further employs an index evolution model as a variant of the health index model, wherein the health index H... i (t) is represented as , Let t be the initial health index, a be the rate of change in health, and t be the time variable. Furthermore, a integrates factors such as age, exercise intensity, diet, and mental state as adjustment parameters.

[0058] The data preprocessing step further includes a feature selection and extraction sub-step, which fuses data from the pixel layer, feature layer, and decision layer through multi-level fusion, and uses a dynamic weighting method based on feature importance to dynamically adjust feature weights according to data quality indicators.

[0059] The effectiveness evaluation step further integrates a survival analysis model, constructing a health index as a time-dependent covariate based on time-to-event data. The survival analysis model uses a Cox proportional hazards model to assess the long-term impact and risk ratio of health protection measures on team performance.

[0060] Example 3 like Figure 2 As shown, this invention also proposes a controlled trial system for verifying the effect of health protection measures on team performance improvement, comprising: The system includes a multimodal synchronous acquisition unit, a DTW alignment and isolated forest cleaning unit, a dual-mechanism health index calculation unit, a multi-level fusion weighting unit, a three-dimensional balanced random grouping unit, a parallel evaluation unit for multiple regression and Cox survival analysis, and an edge-cloud collaborative control unit for outputting a statistical report containing risk ratios.

[0061] The system adopts a three-layer data processing architecture to achieve robust feature fusion: Signal layer: The original sensor waveforms are preprocessed by filtering, denoising, etc., to obtain a clean physiological signal time series.

[0062] Feature Layer: Extracts statistical features (such as SDNN for heart rate variability, daily average steps) and time / frequency domain features from the preprocessed signal. This layer introduces a dynamic weighting strategy: calculates a real-time quality confidence score for each feature (e.g., based on signal-to-noise ratio or sensor wearing status), and uses this score as the weight for that feature in subsequent calculations.

[0063] Model layer: The weighted feature vector is input into the dual-mechanism health index model and the team performance calculation model to generate the final health index and performance indicators.

[0064] About the edge-cloud collaborative computing framework: Edge side (deployed on personal smart devices or team local servers): responsible for performing tasks with high real-time requirements, including: multimodal data acquisition, DTW time alignment, isolated forest anomaly detection, forward computation of health index models, and real-time intervention feedback based on health indexes (such as adjusting exercise prompts).

[0065] Cloud: Responsible for performing computationally intensive or global data-required tasks, including: periodic training and updating of isolated forest and health index models, execution of three-dimensional balanced random grouping algorithms, parameter estimation and statistical testing of multiple linear regression and Cox proportional hazards models, and generation of full-cycle analysis reports.

[0066] Data Flow: The edge server periodically uploads anonymized and aggregated health index sequences, performance indicators, and feature quality reports to the cloud. The cloud then distributes updated model parameters to the edge server. This architecture reduces the uploading of sensitive raw data, lowers bandwidth and cloud computing costs, and ensures real-time performance.

[0067] Example 4 This disclosure provides a non-volatile computer storage medium storing computer-executable instructions that can perform the steps described in the above embodiments.

[0068] It should be noted that the computer-readable medium described in this disclosure can be a computer-readable signal medium or a computer-readable storage medium, or any combination thereof. A computer-readable storage medium can be, for example,—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this disclosure, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In this disclosure, a computer-readable signal medium can include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium can be any computer-readable medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to: wires, optical fibers, RF (radio frequency), etc., or any suitable combination thereof.

[0069] The aforementioned computer-readable medium may be included in the aforementioned electronic device; or it may exist independently and not assembled into the electronic device.

[0070] Computer program code for performing the operations of this disclosure can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (AN) or a wide area network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0071] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated 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 diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0072] The units described in the embodiments of this disclosure can be implemented in software or hardware. The names of the units are not, in some cases, intended to limit the specific unit.

[0073] The preferred embodiments of the present invention have been described above to make the spirit of the present invention clearer and easier to understand, and are not intended to limit the present invention. All modifications, substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope summarized by the appended claims.

Claims

1. A controlled trial method for verifying the effect of health protection measures on team performance improvement, characterized in that, Includes the following steps: a) Synchronously collect multimodal time-series data of team members via wearable sensors and team management API; b) A dynamic time warping algorithm is used to obtain a preprocessed dataset; c) On the preprocessed dataset, run the "time decay + exponential evolution" dual-mechanism health index model to dynamically predict individual health trends and generate a health index sequence that is updated synchronously with health interventions. d) Calculate the team performance indicator sequence based on task priority weighting and record the performance failure time; e) Using team size, historical performance, and health baseline as three-dimensional balancing factors, the experimental group and the control group were randomly divided. The experimental group received smart fitness and psychological assessment intervention. f) Construct multiple linear regression and Cox proportional hazards models in parallel, using the health index as a time-dependent covariate, to estimate the net effect of the intervention on performance indicators and performance failure time, and output a statistical significance report including the risk ratio.

2. The method as described in claim 1, characterized in that, The multimodal time series data mentioned in a) includes at least heart rate variability, steps, sleep depth, focus time, rest interval, milestone completion time, and error rate.

3. The method as described in claim 1, characterized in that, Specifically, b) includes: using a dynamic time warping algorithm to align sequences with different sampling frequencies to a unified time axis, further removing sensor drift outliers through isolated forests, and obtaining a preprocessed dataset through Z-score normalization.

4. The method as described in claim 3, characterized in that, The dynamic time warping algorithm introduces sensor confidence weights when backtracking the optimal path, reducing the distortion of alignment results caused by low-confidence sampling points.

5. The method as described in claim 1, characterized in that, The dual-mechanism health index model described in c) integrates age, exercise intensity, diet, and psychological state through an index evolution term, and strengthens the weight of recent physiological behavior data with a decay coefficient.

6. The method as described in claim 1, characterized in that, In the Cox proportional hazards model, a performance failure event is defined as: the actual completion time of a project milestone first exceeds the predetermined deadline, or the task error rate first exceeds a predetermined threshold; the performance failure time is the moment when the performance failure event occurs.

7. The method as described in claim 1, characterized in that, The three-dimensional balancing factors mentioned in step e) are specifically team size, which represents team attributes; historical performance baseline, which represents historical performance; and health baseline, which represents the status of members. The three-dimensional balancing algorithm uses the three-dimensional balancing factors to form a feature vector and performs group optimization using the Mahalanobis distance minimization principle.

8. The method as described in claim 1, characterized in that, The effectiveness evaluation report further provides the percentage decrease in the performance failure risk ratio corresponding to each unit increase in the health index during the intervention period.

9. The method as described in claim 1, characterized in that, The intelligent fitness intervention includes adjusting the exercise intensity and rest intervals prompted by wearable devices in real time based on health indices.

10. A controlled trial system for verifying the effect of health protection measures on team performance improvement by performing the method of any one of claims 1 to 9, comprising: The system includes a multimodal synchronous acquisition unit, a DTW alignment and isolated forest cleaning unit, a dual-mechanism health index calculation unit, a multi-level fusion weighting unit, a three-dimensional balanced random grouping unit, a parallel evaluation unit for multiple regression and Cox survival analysis, and an edge-cloud collaborative control unit for outputting a statistical report containing risk ratios.

Images