Simulated human body supporting device and human body supporting system for old-age nursing training

Abstract

The invention relates to the technical field of old-age nursing, and particularly discloses a simulated human body supporting device for old-age nursing training and a human body supporting system, the human body supporting device comprises a bed frame, a supporting panel, adjusting supporting legs and a data transmission interface, and a pressure sensor array is embedded in the supporting panel and used for collecting pressure distribution data. The system comprises a data acquisition module, a preprocessing module, a three-layer algorithm processing module, a cloud platform sharing module, a feedback display module and a training control module. According to the system, through a three-layer progressive data processing flow, firstly, the sliding risk of the dummy is detected based on pressure entropy change, then the dynamic balance ability of the nursing personnel is evaluated in combination with the sliding risk and the operation speed, and finally the response efficiency of the nursing personnel to sudden tumble is predicted based on balance ability data. According to the invention, the whole process optimization from basic risk early warning to advanced capability prediction is realized, and the pertinence, effectiveness and safety of the old-age nursing training are effectively improved.

Description

Technical Field

[0001] This invention relates to the field of elderly care technology, specifically to a simulated human body support device and human body support system for elderly care training. Background Technology

[0002] Elderly care training is a key component in improving the quality of nursing services. Simulated practical training relies on various training equipment, especially mannequin support devices and corresponding training systems. These systems typically include a bed structure for placing and securing the nursing mannequin, as well as integrated sensors and data processors to collect and analyze operational data during the training process. The aim is to train nursing staff's skills by simulating real nursing scenarios and to optimize training effectiveness with the help of increasingly advanced medical and health big data technology, ultimately providing data support and capacity assurance for actual elderly care work.

[0003] For example, application number "CN201710051203.7" discloses a voice simulation training system for elderly care workers. This system incorporates a voice simulation training system within a human training model. This system stores voice data from various training scenarios and can play different voice data based on set scenario requirements. This allows caregivers to experience more realistic scenarios during training using the human training model, effectively improving their caregiving skills and increasing the efficiency of skill development. However, existing elderly care training dummy support devices and systems still face a series of specific problems that urgently need to be solved in practical applications: Firstly, when simulating scenarios such as a slippery bed surface due to cleaning, sweat, or a humid environment, traditional devices... The lack of an effective real-time detection mechanism makes it impossible to accurately detect the potential slippage tendency of the dummy, leading to unexpected interruptions in training or even equipment damage. Secondly, even if the slippage risk can be detected, existing systems struggle to further assess the dynamic balance ability of caregivers when performing multitasking operations (such as simultaneously moving the dummy and observing the environment). This leaves training assessments superficial and unable to deeply quantify their practical stability in complex scenarios. Finally, existing technologies cannot scientifically predict the response efficiency of caregivers in dealing with emergencies such as sudden falls in the elderly based on basic balance ability data combined with historical medical and health big data. This results in a disconnect between training and real emergency scenarios, making it difficult to systematically improve the overall emergency response level of the nursing team. Summary of the Invention

[0004] (a) Technical problems to be solved

[0005] This invention provides a simulated human body support device and human body support system for elderly care training, which solves the problems mentioned in the background art.

[0006] (II) Technical Solution

[0007] To achieve the above objectives, the present invention is implemented through the following technical solution: a simulated human body support device for elderly care training, including a bed frame, a support panel fixedly installed on the top of the bed frame, an adjustable support leg fixedly connected to the bottom of the bed frame, and a data transmission interface fixedly installed on the top of the bed frame;

[0008] The bed frame is made of a lightweight aluminum alloy frame, which provides overall structural support; the support panel is made of high-strength composite material, the surface of the support panel is covered with an anti-slip coating, and the support panel is embedded with a pressure sensor array, which is arranged in a grid and covers the entire panel area to collect pressure distribution data in real time when simulating human body placement.

[0009] The adjustable legs are height-adjustable via electro-hydraulic means to ensure the bed surface is level. The data transmission interface is electrically connected to the pressure sensor array and uses a wireless transmission module for data transmission. The bed frame, support panel, adjustable legs, and data transmission interface are combined to form an integral bed structure. The support panel bears the weight of a simulated human body and collects data through the sensor array. The adjustable legs adapt to different ground conditions. The data transmission interface transmits the collected data to an external system unit in real time.

[0010] A simulated human body support system for elderly care training includes:

[0011] The data acquisition module is used to acquire pressure distribution data in real time from the pressure sensor array of the simulated human body support device, and to perform preliminary caching and format conversion;

[0012] The data preprocessing module, connected to the data acquisition module, filters, normalizes, and processes outliers on the acquired raw pressure data to generate a standardized data stream.

[0013] The first-layer algorithm processing module, connected to the data preprocessing module, calculates the slip risk index based on the standardized data stream using a dynamic rate of change algorithm, thereby enabling real-time detection of the dummy's slip risk.

[0014] The second-layer algorithm processing module, connected to the first-layer algorithm processing module, uses the sliding risk index sequence and the operation speed extracted from the stress data to calculate the dynamic balance ability score through the kinematic model algorithm, and evaluates the balance performance of nurses in multi-tasking operations.

[0015] The third-layer algorithm processing module connects to the second-layer algorithm processing module. Based on the dynamic equilibrium capability score sequence, it uses machine learning model algorithms to predict response efficiency and optimizes the model using historical data from the cloud platform data sharing module.

[0016] The cloud platform data sharing module connects the third-layer algorithm processing module and the data preprocessing module, providing storage, sharing, and model training services for medical and health big data, including historical nursing event data and algorithm updates;

[0017] The feedback and display module connects the third-layer algorithm processing module and the second-layer algorithm processing module, and outputs the sliding risk index, dynamic balance ability score and response efficiency prediction results to the user in a visual manner in real time.

[0018] The training control module connects all other modules, coordinates data flow and processing procedures, and enables the simulation and control of training scenarios.

[0019] Furthermore, the data acquisition module specifically includes a high-precision pressure sensor array interface, a data buffer unit, and a real-time clock unit. The pressure sensor array interface reads pressure values ​​from the sensor grid of the simulated human body support device at a rate of 10 samples per second. The data buffer unit uses a circular buffer to temporarily store raw data to prevent loss. The real-time clock unit adds a timestamp to each data sample to ensure timing consistency. The data acquisition module communicates with the data transmission interface of the simulated human body support device wirelessly or via wired means, collects pressure distribution data in real time, and forwards it to the data preprocessing module. At the same time, it monitors the sensor status and performs self-calibration to maintain data accuracy.

[0020] Furthermore, the data preprocessing module specifically includes a digital filter unit, a data normalization unit, and an anomaly detection unit. The digital filter unit uses a low-pass filter to remove high-frequency noise from the pressure data. The data normalization unit maps the pressure values ​​to the 0-1 range to eliminate the influence of dimensions. The anomaly detection unit identifies and removes outliers based on statistical methods. The data preprocessing module receives the raw data from the data acquisition module, outputs a standardized data stream to the first-layer algorithm processing module, and processes the real-time data stream through a sliding window mechanism to ensure data quality and consistency, providing reliable input for subsequent algorithm calculations.

[0021] Furthermore, the first-layer algorithm processing module specifically includes a spatial entropy calculation unit, a rate of change analysis unit, and a threshold comparison unit. The spatial entropy calculation unit calculates the spatial entropy value of the pressure distribution at each time point based on the standardized data stream. The rate of change analysis unit calculates the rate of change of entropy values ​​between adjacent time points and generates a sliding risk index. The threshold comparison unit dynamically detects sliding risk by calibrating thresholds through experiments. The first-layer algorithm processing module receives the output of the data preprocessing module, calculates the sliding risk index in real time, and triggers an early warning signal if the index exceeds the threshold. The index sequence is then transmitted to the second-layer algorithm processing module, while historical data is recorded for subsequent analysis.

[0022] Furthermore, the second-layer algorithm processing module specifically includes a pressure center trajectory calculation unit, an operation speed extraction unit, and a kinematic model unit. The pressure center trajectory calculation unit derives the movement path of the pressure center from the standardized data stream. The operation speed extraction unit calculates the average movement speed based on trajectory changes. The kinematic model unit combines the slip risk index and operation speed to generate a dynamic balance ability score through a weighted formula. The second-layer algorithm processing module receives the slip risk index sequence and pressure center data from the first-layer algorithm processing module, outputs the dynamic balance ability score to the third-layer algorithm processing module and the feedback and display module, and adapts to different training scenarios by adjusting the weight coefficients in real time to accurately assess the balance ability of nurses in multi-task operations.

[0023] Furthermore, the third-layer algorithm processing module specifically includes a machine learning model unit, a sequence data processing unit, and a prediction output unit. The machine learning model unit uses a pre-trained support vector machine or neural network model, the sequence data processing unit converts the dynamic balance ability score sequence into model input features, and the prediction output unit generates a predicted response efficiency value. The third-layer algorithm processing module receives the dynamic balance ability score sequence from the second-layer algorithm processing module, optimizes the model parameters using historical data from the cloud platform data sharing module, outputs the predicted response time to the feedback and display module, and periodically updates the model to maintain prediction accuracy, thereby achieving reliable prediction of the response efficiency of nursing staff in dealing with sudden fall events.

[0024] Furthermore, the cloud platform data sharing module specifically includes a data storage submodule, a model training submodule, and a data interface submodule. The data storage submodule uses a distributed database to store historical nursing event data such as fall records and response times. The model training submodule trains and optimizes the machine learning model of the third-layer algorithm processing module based on the stored data. The data interface submodule provides APIs for communication with other modules. The cloud platform data sharing module receives standardized data from the data preprocessing module and prediction results from the third-layer algorithm processing module to achieve data sharing and model updates, while ensuring data security and privacy protection. It also connects to external medical and health big data platforms via the Internet to enrich the dataset.

[0025] Furthermore, the feedback and display module specifically includes a visualization display unit, an audible and visual alarm unit, and a report generation unit. The visualization display unit displays the dynamic balance ability score from the second-layer algorithm processing module and its associated real-time sliding risk index and operation speed data on an LCD screen in real time. It also displays the response efficiency prediction results and prediction confidence from the third-layer algorithm processing module. The audible and visual alarm unit triggers sound and flashing warnings based on the high-risk status indicator signal output by the second-layer algorithm processing module. The report generation unit automatically generates a complete training assessment report, including skill gap analysis and training focus recommendations, based on the structured assessment report output by the third-layer algorithm processing module. The feedback and display module receives all the above output data through a data bus and presents it synchronously in a combination of graphics, numbers, text, and audible and visual signals. At the same time, it uploads the report data to the cloud platform data sharing module to achieve historical record storage and data analysis, providing nursing staff with comprehensive feedback from real-time operation guidance to long-term competency planning.

[0026] Furthermore, the training control module specifically includes a scenario simulation unit, a process coordination unit, and a user interaction unit. The scenario simulation unit dynamically generates training scenarios such as slippery surfaces or multi-task obstacles based on pressure data and algorithm results. The process coordination unit manages the data flow and execution order between modules to ensure real-time performance. The user interaction unit receives user input such as training parameter settings and start or stop commands. The training control module connects all other modules and coordinates the entire process from data acquisition to feedback through a central processor to achieve automated training management. It also adaptively adjusts the training difficulty based on real-time data to improve the effectiveness and safety of elderly care training.

[0027] (III) Beneficial Effects

[0028] This invention provides a simulated human body support device and system for elderly care training. It has the following beneficial effects:

[0029] (I) The simulated human body support device and system used in this elderly care training collect pressure distribution data of the simulated human body on the support bed in real time, and introduce dynamic change rate analysis of pressure distribution spatial entropy. It can accurately detect the tendency of the simulated human body to slide in special environments such as wet and slippery surfaces. Compared with the traditional method that relies on a single pressure threshold or manual observation, this method is more sensitive to small changes in pressure distribution and responds more promptly. Once the risk of sliding exceeds the preset threshold, the system immediately prompts the nursing staff to intervene through audible and visual alarms, or automatically activates the anti-slip mechanism of the bed surface. This effectively avoids training interruptions and equipment damage caused by the simulated human body sliding or falling during the training process, significantly improves the continuity and safety of training, and provides an accurate and reliable basic data stream for subsequent higher-level competency assessments.

[0030] (II) This elderly care training uses a simulated human body support device and system, which further integrates the operation speed information extracted from pressure data. By constructing a kinematic model that comprehensively considers risk and efficiency, it achieves an objective and quantitative assessment of the dynamic balance ability of nursing staff in a simulated multi-task operation environment. This assessment method breaks through the limitations of traditional subjective scoring and can truly reflect the physical stability and coordination of nursing staff when performing complex tasks such as moving a simulated human body while avoiding obstacles. The system provides real-time feedback of the assessment results in an intuitive scoring format and drives the training module to generate dynamic obstacle scenarios, thereby providing nursing staff with targeted reinforcement training basis, effectively improving their practical ability and safety level in dealing with multi-task needs in real nursing scenarios, and constituting an advanced optimization from risk prevention to ability improvement.

[0031] (III) This elderly care training uses a simulated human body support device and system. Utilizing time-series data of dynamic balance ability scoring and machine learning models trained on shared medical and health big data from a cloud platform, it achieves scientific prediction of the response efficiency of caregivers in the face of emergencies such as sudden falls in the elderly. This predictive model correlates performance during training with potential responses in real emergency scenarios, enabling early identification of slow response risks and generating personalized training improvement reports to guide training focus towards weaker areas. Simultaneously, the system supports continuous online updates of model parameters, ensuring the accuracy and adaptability of predictions. Ultimately, this training system not only improves routine nursing skills but also proactively enhances the emergency response capabilities and overall quality of care within the nursing team, forming a transition from basic operations to advanced early warning systems, significantly improving the systematic nature and effectiveness of elderly care training. Attached Figure Description

[0032] Figure 1 This is a schematic diagram of the overall structure of the present invention;

[0033] Figure 2 This is a schematic diagram of the support device of the present invention.

[0034] In the diagram: 1. Bed frame; 2. Support panel; 3. Adjustable legs; 4. Data transmission interface. Detailed Implementation

[0035] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0036] First embodiment: as follows Figures 1 to 2As shown, the present invention provides a technical solution: a simulated human body support device for elderly care training, including a bed frame 1, a support panel 2 fixedly installed on the top of the bed frame 1, an adjustable support leg 3 fixedly connected to the bottom of the bed frame 1, and a data transmission interface 4 fixedly installed on the top of the bed frame.

[0037] The bed frame 1 is made of a lightweight aluminum alloy frame, which provides overall structural support; the support panel 2 is made of high-strength composite material, the surface of the support panel 2 is covered with an anti-slip coating, and the support panel 2 is embedded with a pressure sensor array, which is arranged in a grid and covers the entire panel area to collect pressure distribution data in real time when simulating human body placement.

[0038] The adjustable legs 3 are height-adjustable via electro-hydraulic means to ensure the bed surface is level. The data transmission interface 4 is electrically connected to the pressure sensor array and uses a wireless transmission module for data transmission. The bed frame 1, support panel 2, adjustable legs 3 and data transmission interface 4 are combined to form an overall bed structure. The support panel bears the weight of a simulated human body and collects data through the sensor array. The adjustable legs adapt to different ground conditions. The data transmission interface transmits the collected data to the external system unit in real time.

[0039] A simulated human body support system for elderly care training includes:

[0040] The data acquisition module is used to acquire pressure distribution data in real time from the pressure sensor array of the simulated human body support device, and to perform preliminary caching and format conversion;

[0041] The data preprocessing module, connected to the data acquisition module, filters, normalizes, and processes outliers on the acquired raw pressure data to generate a standardized data stream.

[0042] The first-layer algorithm processing module, connected to the data preprocessing module, calculates the slip risk index based on the standardized data stream using a dynamic rate of change algorithm, thereby enabling real-time detection of the dummy's slip risk.

[0043] The second-layer algorithm processing module, connected to the first-layer algorithm processing module, uses the sliding risk index sequence and the operation speed extracted from the stress data to calculate the dynamic balance ability score through the kinematic model algorithm, and evaluates the balance performance of nurses in multi-tasking operations.

[0044] The third-layer algorithm processing module connects to the second-layer algorithm processing module. Based on the dynamic equilibrium capability score sequence, it uses machine learning model algorithms to predict response efficiency and optimizes the model using historical data from the cloud platform data sharing module.

[0045] The cloud platform data sharing module connects the third-layer algorithm processing module and the data preprocessing module, providing storage, sharing, and model training services for medical and health big data, including historical nursing event data and algorithm updates;

[0046] The feedback and display module connects the third-layer algorithm processing module and the second-layer algorithm processing module, and outputs the sliding risk index, dynamic balance ability score and response efficiency prediction results to the user in a visual manner in real time.

[0047] The training control module connects all other modules, coordinates data flow and processing procedures, and enables the simulation and control of training scenarios.

[0048] The data acquisition module specifically includes a high-precision pressure sensor array interface, a data buffer unit, and a real-time clock unit. The pressure sensor array interface reads pressure values ​​from the sensor grid of the simulated human body support device at a rate of 10 samples per second. The data buffer unit uses a circular buffer to temporarily store raw data to prevent loss. The real-time clock unit adds a timestamp to each data sample to ensure timing consistency. The data acquisition module communicates with the data transmission interface of the simulated human body support device wirelessly or via wired means, collects pressure distribution data in real time, and forwards it to the data preprocessing module. At the same time, it monitors the sensor status and performs self-calibration to maintain data accuracy.

[0049] The data preprocessing module specifically includes a digital filter unit, a data normalization unit, and an anomaly detection unit. The digital filter unit uses a low-pass filter to remove high-frequency noise from the pressure data. The data normalization unit maps the pressure values ​​to the 0-1 range to eliminate the influence of dimensions. The anomaly detection unit identifies and removes outliers based on statistical methods. The data preprocessing module receives the raw data from the data acquisition module, outputs a standardized data stream to the first-layer algorithm processing module, and processes the real-time data stream through a sliding window mechanism to ensure data quality and consistency, providing reliable input for subsequent algorithm calculations.

[0050] The first-layer algorithm processing module specifically includes a spatial entropy calculation unit, a rate of change analysis unit, and a threshold comparison unit. The spatial entropy calculation unit calculates the spatial entropy value of the pressure distribution at each time point based on the standardized data stream. The rate of change analysis unit calculates the rate of change of entropy values ​​between adjacent time points and generates a sliding risk index. The threshold comparison unit dynamically detects sliding risk by calibrating thresholds through experiments. The first-layer algorithm processing module receives the output of the data preprocessing module, calculates the sliding risk index in real time, and triggers an early warning signal if the index exceeds the threshold. The index sequence is then transmitted to the second-layer algorithm processing module, and historical data is recorded for subsequent analysis.

[0051] During operation, the simulated human body is first placed on a simulated human body support device. A high-precision pressure sensor array is embedded within the support panel of this device, collecting real-time pressure distribution data of the dummy in a grid pattern. This data is then transmitted to the system via a data transmission interface. The data acquisition module receives this raw data at a rate of 10 samples per second, performs buffering and format conversion to ensure data integrity, and then the data preprocessing module performs low-pass filtering to remove noise, normalization to eliminate dimensional influences, and outlier detection to remove errors, generating a standardized data stream. Finally, the first-layer algorithm processing module calculates the spatial entropy of the pressure distribution and its rate of change based on this data stream. When the rate of change exceeds the experimentally calibrated threshold, a slip risk index is output, triggering an audible and visual alarm in the feedback and display module. This achieves real-time detection and early warning of slip risk, effectively preventing the simulated human body from slipping or falling due to slippery surfaces during training, improving the continuity of training and equipment safety. Specifically:

[0052] A mannequin used for elderly care training is stably placed on a support device. The support panel is made of high-strength composite material with a non-slip coating and incorporates a high-precision pressure sensor array. This array is uniformly distributed in a grid pattern, with each grid point measuring 1cm x 1cm, for real-time acquisition of pressure distribution data generated by the mannequin. The sampling frequency is set to 10Hz to ensure high timeliness and integrity of the data. The data acquisition module then receives the raw pressure data via a wireless transmission interface and performs initial buffering and format conversion. The buffer unit uses a ring buffer structure to prevent data loss. A timestamp is added to each data sample to ensure temporal consistency of the data sequence. The data preprocessing module then performs three steps on the raw data: First, a low-pass filter is applied to remove high-frequency noise, with the filter cutoff frequency set to 5Hz to retain valid signals. Second, normalization is performed, mapping the raw pressure values ​​to the 0-1 range to eliminate dimensional differences. The calculation formula is... Where P is the original pressure value, and The third step, using the minimum and maximum values ​​of the current sampling window, detects outliers using the Z-score method, removing data points exceeding ±3 standard deviations to generate a standardized data stream. Finally, the first-layer algorithm processing module calculates the sliding risk index based on this data stream. Specific steps include: first, calculating the spatial entropy of the pressure distribution at each time point, using the formula:

[0053]

[0054] in This represents the proportion of the normalized pressure value of the i-th grid point to the total pressure, where n is the total number of grid points. Then, the rate of change of entropy values ​​between adjacent time points is calculated using the following formula:

[0055]

[0056] The time interval Δt is 0.1 seconds (corresponding to 10Hz sampling). When the rate of change exceeds the experimental calibration threshold of 0.5 / s, the slip risk index is marked as a high-risk state, and the audible and visual alarm of the feedback and display module is immediately triggered, prompting the nursing staff to adjust the position of the dummy or take anti-slip measures. This embodiment can effectively prevent the dummy from slipping or falling due to wet surfaces during training through real-time data analysis and dynamic entropy change detection, improve the continuity of training and equipment safety, and provide a reliable data basis for subsequent evaluation.

[0057] Second embodiment: as follows Figures 1 to 2 As shown, the second-layer algorithm processing module specifically includes a pressure center trajectory calculation unit, an operation speed extraction unit, and a kinematic model unit. The pressure center trajectory calculation unit derives the movement path of the pressure center from the standardized data stream. The operation speed extraction unit calculates the average movement speed based on trajectory changes. The kinematic model unit combines the slip risk index and operation speed to generate a dynamic balance ability score through a weighted formula. The second-layer algorithm processing module receives the slip risk index sequence and pressure center data from the first-layer algorithm processing module, outputs the dynamic balance ability score to the third-layer algorithm processing module and the feedback and display module, and adapts to different training scenarios by adjusting the weight coefficients in real time to accurately assess the balance ability of nurses in multi-task operations.

[0058] During operation, based on the slip risk index sequence calculated in the first embodiment, the second-layer algorithm processing module first extracts the movement trajectory of the pressure center from the same pressure data and calculates the operation speed as input. Then, combining the slip risk index and operation speed, a weighted calculation is performed by the kinematic model unit to generate a dynamic balance ability score. The weight coefficients are dynamically adjusted according to the real-time training scenario to optimize accuracy. This score is then transmitted to the feedback and display module in real time and displayed to the nursing staff in the form of a visual chart. At the same time, the training control module simulates multi-task obstacle scenarios, such as virtual projected obstacles, requiring the nursing staff to maintain balance while moving the dummy. Finally, the system records the score sequence for subsequent analysis. Specifically:

[0059] First, based on the sliding risk index sequence and standardized pressure data stream calculated in the first embodiment, the second-layer algorithm processing module initiates the dynamic equilibrium assessment process. The sliding risk index sequence comes from the real-time output of the first-layer algorithm processing module, while the standardized pressure data stream is provided by the data preprocessing module. Next, the pressure center movement trajectory is extracted from the same pressure data. The specific calculation steps include: based on the gridded pressure distribution data, each grid point has two-dimensional coordinates. and normalized pressure value (Using the normalization formula of the first embodiment) Calculation), coordinates of the pressure center ( , Calculated using the weighted average formula:

[0060]

[0061] Where n is the total number of grid points, the operating speed is then calculated based on the pressure center coordinate sequence. That is, the Euclidean distance of the pressure center movement at adjacent time points divided by the time interval Δt (consistent with the first embodiment, Δt = 0.1 seconds), calculated using the following formula:

[0062]

[0063] Next, the sliding risk index and operating speed are normalized to eliminate the influence of dimensions. The sliding risk index is normalized using... ,in and The minimum and maximum values ​​within the current time window are used to normalize the operation speed. ,in and A reference range is set based on historical data, and then the dynamic balance ability score is calculated through kinematic model units using a weighted combination formula:

[0064]

[0065] The weighting coefficients α and β are optimized using training data (typical values ​​α=0.6, β=0.4) to reflect the relative impact of slippage risk and operation speed on balance ability. The scoring results are transmitted to the feedback and display module in real time. Finally, the training control module simulates multi-task obstacle scenarios based on the dynamic balance ability score, such as generating moving obstacles through virtual projection, requiring nurses to maintain balance while moving the dummy. At the same time, the system records the scoring sequence for subsequent analysis and optimization. The beneficial effect of this embodiment is that it can objectively quantify the dynamic balance ability of nurses in complex multi-task operations, provide data support for personalized training, improve operational efficiency and safety, and help reduce operational errors through real-time feedback.

[0066] Third embodiment: as follows Figures 1 to 2As shown, the third-layer algorithm processing module specifically includes a machine learning model unit, a sequence data processing unit, and a prediction output unit. The machine learning model unit uses a pre-trained support vector machine or neural network model. The sequence data processing unit converts the dynamic balance ability score sequence into model input features. The prediction output unit generates a predicted response efficiency value. The third-layer algorithm processing module receives the dynamic balance ability score sequence from the second-layer algorithm processing module, optimizes the model parameters using historical data from the cloud platform data sharing module, outputs the predicted response time to the feedback and display module, and updates the model regularly to maintain prediction accuracy, thereby achieving reliable prediction of the response efficiency of nursing staff in dealing with sudden fall events.

[0067] The cloud platform data sharing module specifically includes a data storage submodule, a model training submodule, and a data interface submodule. The data storage submodule uses a distributed database to store historical nursing event data such as fall records and response times. The model training submodule trains and optimizes the machine learning model of the third-layer algorithm processing module based on the stored data. The data interface submodule provides APIs for communication with other modules. The cloud platform data sharing module receives standardized data from the data preprocessing module and prediction results from the third-layer algorithm processing module to achieve data sharing and model updates, while ensuring data security and privacy protection. It also connects to external medical and health big data platforms via the Internet to enrich the dataset.

[0068] The feedback and display module specifically includes a visualization display unit, an audible and visual alarm unit, and a report generation unit. The visualization display unit displays the dynamic balance ability score from the second-layer algorithm processing module and its associated real-time sliding risk index and operation speed data on an LCD screen in real time. It also displays the response efficiency prediction results and prediction confidence from the third-layer algorithm processing module. The audible and visual alarm unit triggers sound and flashing warnings based on the high-risk status indicator signal output by the second-layer algorithm processing module. The report generation unit automatically generates a complete training assessment report based on the structured assessment report output by the third-layer algorithm processing module, including skill gap analysis and training focus recommendations. The feedback and display module receives all the above output data through a data bus and presents it synchronously in a combination of graphics, numbers, text, and audible and visual signals. At the same time, it uploads the report data to the cloud platform data sharing module to realize historical record storage and data analysis, providing nursing staff with comprehensive feedback from real-time operation guidance to long-term competency planning.

[0069] The training control module specifically includes a scenario simulation unit, a process coordination unit, and a user interaction unit. The scenario simulation unit dynamically generates training scenarios such as slippery surfaces or multi-task obstacles based on pressure data and algorithm results. The process coordination unit manages the data flow and execution order between modules to ensure real-time performance. The user interaction unit receives user input such as training parameter settings and start or stop commands. The training control module connects all other modules and coordinates the entire process from data acquisition to feedback through a central processor to achieve automated training management. It also adaptively adjusts the training difficulty based on real-time data to improve the effectiveness and safety of elderly care training.

[0070] During operation, based on the dynamic balance ability score sequence output in the second embodiment, the third-layer algorithm processing module first uses a pre-trained machine learning model. This model is trained using historical nursing event data (such as fall records and response times) provided by the cloud platform data sharing module. Then, the dynamic balance ability score sequence is input into the model, and converted into feature vectors by the sequence data processing unit to predict the nurses' response efficiency in sudden fall events, i.e., outputting the predicted response time. The prediction results are then used by the feedback and display module to generate a detailed evaluation report, including improvement suggestions and training priorities. Simultaneously, the cloud platform data sharing module periodically updates the model parameters, using new data to optimize prediction accuracy. Finally, the system integrates the prediction results for training adjustments. Specifically:

[0071] First, based on the Dynamic Balance Capability Score (DBS) sequence output in the second embodiment, which is generated and transmitted in real time by the second-layer algorithm processing module (where DBS calculation uses the formula in the second embodiment and the sequence data is cached in chronological order), the third-layer algorithm processing module initiates the response efficiency prediction process. This begins with feature extraction from the DBS sequence, specifically calculating the sequence's statistical features, such as the mean.

[0072]

[0073] Where T is the size of the time window (typically T=50 corresponds to 5 seconds of data), and the standard deviation is:

[0074]

[0075] To measure the volatility of the sequence, the coefficient of DBS(t) with respect to time t is fitted by linear regression to form an eigenvector. Then, the feature vectors are input into a pre-trained machine learning model, which is trained using historical care event data (such as fall records and response times) provided by the cloud platform's data sharing module. The model uses linear regression, and the formula for predicting response efficiency (PRE), i.e., predicting response time, is as follows:

[0076]

[0077] Where the weight vector The bias term b is obtained by optimizing historical data using the least squares method. After the model outputs the PRE value, the system generates a detailed evaluation report through the feedback and display module, including improvement suggestions and training priorities. Finally, the cloud platform data sharing module regularly updates the model parameters based on newly collected data to ensure prediction accuracy. At the same time, the training control module adaptively adjusts the training difficulty according to the PRE results. The beneficial effect of this embodiment is that it can assess the emergency response capabilities of nursing staff in advance through sequence feature extraction and machine learning prediction, optimize training content to cope with emergencies in real-world environments, and thus improve the overall quality and efficiency of nursing care.

[0078] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0079] 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 alterations 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 claims and their equivalents.

Claims

1. A simulated human body support device for elderly care training, comprising a bed frame (1), characterized in that: A support panel (2) is fixedly installed on the top of the bed frame (1), an adjustable leg (3) is fixedly connected to the bottom of the bed frame (1), and a data transmission interface (4) is also fixedly installed on the top of the bed frame (1). The bed frame (1) is made of a lightweight aluminum alloy frame, the support panel (2) is made of high-strength composite material, the surface of the support panel (2) is covered with an anti-slip coating, and the support panel (2) is embedded with a pressure sensor array. The pressure sensor array in the support panel (2) is arranged in a grid and covers the entire panel area to collect pressure distribution data in real time when simulating human body placement. The adjustable support leg (3) is height-adjustable via electro-hydraulic means. The data transmission interface (4) is electrically connected to the pressure sensor array. The data transmission interface (4) uses a wireless transmission module for data transmission. The bed frame (1), support panel (2), adjustable support leg (3) and data transmission interface (4) are combined to form an integral bed structure.

2. A simulated human body support system for elderly care training, using the simulated human body support device for elderly care training as described in claim 1, characterized in that: include: The data acquisition module is used to acquire pressure distribution data in real time from the pressure sensor array of the simulated human body support device, and to perform preliminary caching and format conversion; The data preprocessing module, connected to the data acquisition module, filters, normalizes, and processes outliers on the acquired raw pressure data to generate a standardized data stream. The first-layer algorithm processing module, connected to the data preprocessing module, calculates the slip risk index based on the standardized data stream using a dynamic rate of change algorithm, thereby enabling real-time detection of the dummy's slip risk. The second-layer algorithm processing module, connected to the first-layer algorithm processing module, uses the sliding risk index sequence and the operation speed extracted from the stress data to calculate the dynamic balance ability score through the kinematic model algorithm, and evaluates the balance performance of nurses in multi-tasking operations. The third-layer algorithm processing module connects to the second-layer algorithm processing module. Based on the dynamic equilibrium capability score sequence, it uses machine learning model algorithms to predict response efficiency and optimizes the model using historical data from the cloud platform data sharing module. The cloud platform data sharing module connects the third-layer algorithm processing module and the data preprocessing module, providing storage, sharing, and model training services for medical and health big data, including historical nursing event data and algorithm updates; The feedback and display module connects the third-layer algorithm processing module and the second-layer algorithm processing module, and outputs the sliding risk index, dynamic balance ability score and response efficiency prediction results to the user in a visual manner in real time. The training control module connects all other modules, coordinates data flow and processing procedures, and enables the simulation and control of training scenarios.

3. The simulated human body support system for elderly care training according to claim 2, characterized in that: The data acquisition module specifically includes a high-precision pressure sensor array interface, a data buffer unit, and a real-time clock unit. The pressure sensor array interface reads pressure values ​​from the sensor grid of the simulated human body support device at a rate of 10 samples per second. The data buffer unit uses a circular buffer to temporarily store raw data to prevent loss. The real-time clock unit adds a timestamp to each data sample to ensure timing consistency. The data acquisition module communicates with the data transmission interface of the simulated human body support device wirelessly or via wired means, collects pressure distribution data in real time, and forwards it to the data preprocessing module. At the same time, it monitors the sensor status and performs self-calibration to maintain data accuracy.

4. The simulated human body support system for elderly care training according to claim 3, characterized in that: The data preprocessing module specifically includes a digital filter unit, a data normalization unit, and an anomaly detection unit. The digital filter unit uses a low-pass filter to remove high-frequency noise from the pressure data. The data normalization unit maps the pressure values ​​to the 0-1 range to eliminate the influence of dimensions. The anomaly detection unit identifies and removes outliers based on statistical methods. The data preprocessing module receives the raw data from the data acquisition module, outputs a standardized data stream to the first-layer algorithm processing module, and processes the real-time data stream through a sliding window mechanism.

5. The simulated human body support system for elderly care training according to claim 4, characterized in that: The first-layer algorithm processing module specifically includes a spatial entropy calculation unit, a rate of change analysis unit, and a threshold comparison unit. The spatial entropy calculation unit calculates the spatial entropy value of the pressure distribution at each time point based on the standardized data stream. The rate of change analysis unit calculates the rate of change of entropy values ​​between adjacent time points and generates a sliding risk index. The threshold comparison unit dynamically detects sliding risk by calibrating thresholds through experiments. The first-layer algorithm processing module receives the output of the data preprocessing module, calculates the sliding risk index in real time, and triggers an early warning signal if the index exceeds the threshold. The index sequence is then transmitted to the second-layer algorithm processing module, and historical data is recorded for subsequent analysis.

6. The simulated human body support system for elderly care training according to claim 5, characterized in that: The second-layer algorithm processing module specifically includes a pressure center trajectory calculation unit, an operation speed extraction unit, and a kinematic model unit. The pressure center trajectory calculation unit derives the movement path of the pressure center from the standardized data stream. The operation speed extraction unit calculates the average movement speed based on trajectory changes. The kinematic model unit combines the slip risk index and operation speed to generate a dynamic balance ability score through a weighted formula. The second-layer algorithm processing module receives the slip risk index sequence and pressure center data from the first-layer algorithm processing module, outputs the dynamic balance ability score to the third-layer algorithm processing module and the feedback and display module, and adapts to different training scenarios by adjusting the weight coefficients in real time to accurately assess the balance ability of nurses in multi-task operations.

7. A simulated human body support system for elderly care training according to claim 6, characterized in that: The third-layer algorithm processing module specifically includes a machine learning model unit, a sequence data processing unit, and a prediction output unit. The machine learning model unit uses a pre-trained support vector machine or neural network model. The sequence data processing unit converts the dynamic balance capability score sequence into model input features. The prediction output unit generates a predicted response efficiency value. The third-layer algorithm processing module receives the dynamic balance capability score sequence from the second-layer algorithm processing module, optimizes the model parameters using historical data from the cloud platform data sharing module, outputs the predicted response time to the feedback and display module, and periodically updates the model to maintain prediction accuracy.

8. The simulated human body support system for elderly care training according to claim 7, characterized in that: The cloud platform data sharing module specifically includes a data storage submodule, a model training submodule, and a data interface submodule. The data storage submodule uses a distributed database to store historical nursing event data such as fall records and response times. The model training submodule trains and optimizes the machine learning model of the third-layer algorithm processing module based on the stored data. The data interface submodule provides APIs for communication with other modules. The cloud platform data sharing module receives standardized data from the data preprocessing module and prediction results from the third-layer algorithm processing module to achieve data sharing and model updates, while ensuring data security and privacy protection. It also connects to external medical and health big data platforms via the Internet to enrich the dataset.

9. A simulated human body support system for elderly care training according to claim 8, characterized in that: The feedback and display module specifically includes a visualization display unit, an audible and visual alarm unit, and a report generation unit. The visualization display unit displays the dynamic balance ability score from the second-layer algorithm processing module and its associated real-time sliding risk index and operation speed data on an LCD screen in real time. It also displays the response efficiency prediction results and prediction confidence from the third-layer algorithm processing module. The audible and visual alarm unit triggers sound and flashing warnings based on the high-risk status indicator signal output by the second-layer algorithm processing module. The report generation unit automatically generates a complete training assessment report, including skill gap analysis and training focus recommendations, based on the structured assessment report output by the third-layer algorithm processing module. The feedback and display module receives all the above output data through a data bus and presents it synchronously in a combination of graphics, numbers, text, and audible and visual signals. At the same time, it uploads the report data to the cloud platform data sharing module to achieve historical record storage and data analysis, providing nursing staff with comprehensive feedback from real-time operation guidance to long-term competency planning.

10. A simulated human body support system for elderly care training according to claim 9, characterized in that: The training control module specifically includes a scenario simulation unit, a process coordination unit, and a user interaction unit. The scenario simulation unit dynamically generates training scenarios such as slippery surfaces or multi-task obstacles based on pressure data and algorithm results. The process coordination unit manages the data flow and execution order between modules to ensure real-time performance. The user interaction unit receives user input such as training parameter settings and start or stop commands. The training control module connects all other modules, coordinates the entire process from data acquisition to feedback through a central processing unit, and adaptively adjusts the training difficulty based on real-time data.

Images