An intelligent assisted monitoring system for fracture in orthopedic care

Abstract

The present invention discloses an intelligent auxiliary monitoring system for orthopedic nursing, which includes the following steps: The system collects mechanical parameters, motion trajectory data and biological signals of fracture patients through multi-modal sensors, and uses deep learning algorithms to intelligently analyze the data to generate diagnostic results of fracture status and prediction data of recovery trends; The low-power wireless communication module realizes the real-time transmission of diagnostic data and supports remote system parameter adjustment at the same time; The personalized nursing plan generation module generates a rehabilitation plan, a load adjustment strategy and an auxiliary device configuration plan through an optimization algorithm based on the patient's biomechanical model and diagnostic results; The adaptive feedback adjustment module dynamically adjusts the data acquisition strategy and processing parameters to ensure the system's real-time response ability to the patient's state. The present invention is applicable to the field of intelligent auxiliary monitoring technology and can realize real-time monitoring, intelligent diagnosis and personalized rehabilitation guidance for fracture patients.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent assisted monitoring, and particularly to a fracture intelligent assisted monitoring system for orthopedic nursing. Background Art

[0002] With the development of medical technology and monitoring technology, real-time monitoring and assisted diagnosis of fracture patients in orthopedic nursing have become key requirements. During the rehabilitation process of fracture patients, it is necessary to closely monitor the fracture healing status, biomechanical changes, and rehabilitation training effects. However, there are still many deficiencies in the existing technology in the field of fracture monitoring and diagnosis, making it difficult to meet the personalized, real-time, and intelligent orthopedic nursing needs.

[0003] In the existing technology, traditional fracture monitoring systems usually rely on a single type of sensor or fixed monitoring parameters, supplemented by simple data analysis methods. Although these methods have, to a certain extent, solved the basic needs of fracture status monitoring, they have the following defects when facing complex clinical environments and diverse patient needs:

[0004] 1. Insufficient comprehensiveness of monitoring data: Existing systems mostly use a single sensor to collect data, lacking the collaborative ability of multi-modal sensors, and it is difficult to comprehensively capture multi-dimensional information such as mechanics, strain, and bioelectricity, resulting in insufficient accuracy and comprehensiveness of monitoring results.

[0005] 2. Limited intelligent diagnosis ability: Traditional monitoring systems are mostly based on rules or simple algorithms, and it is difficult to process complex and multi-dimensional fracture status data, unable to achieve high-precision prediction and personalized diagnosis of the fracture healing trend of patients.

[0006] 3. Lack of dynamic adjustment ability: Existing systems cannot dynamically adjust monitoring parameters and data processing strategies according to the real-time state changes of patients, resulting in a lag in the system's response to the rehabilitation needs of patients.

[0007] 4. High energy consumption of the communication module: The wireless communication module of existing systems often has high power consumption, is not suitable for long-term care scenarios, and lacks the ability to efficiently remotely transmit and provide real-time feedback on diagnostic data.

[0008] 5. Lack of personalization in the nursing plan: Traditional fracture nursing plans mostly use fixed templates and fail to generate personalized rehabilitation plans by combining the biomechanical characteristics and real-time state of patients, making it difficult to meet the diverse nursing needs of patients.

[0009] Therefore, how to provide a fracture intelligent assisted monitoring system for orthopedic nursing is an urgent problem that needs to be solved by those skilled in the art. Summary of the Invention

[0010] An object of the present invention is to propose an intelligent auxiliary monitoring system for orthopedic care. The present invention combines multi-modal sensor fusion technology, deep learning intelligent diagnosis technology, low-power wireless communication technology, and personalized care plan generation technology, and details the algorithms and methods for realizing real-time monitoring, diagnosis, and personalized rehabilitation plan generation for fracture patients, with the advantages of high comprehensiveness, strong real-time performance, low energy consumption, and high accuracy of care plans.

[0011] A fracture intelligent auxiliary monitoring system for orthopedic care according to an embodiment of the present invention includes the following steps:

[0012] S1. A multi-modal sensor fusion module that integrates an acceleration sensor, a strain sensor, and an optical fiber sensor, is used to collect the acceleration signal, strain distribution signal, and bioelectro-optical signal of a fracture patient, and processes the acceleration signal, strain distribution signal, and bioelectro-optical signal through a fusion algorithm to generate multi-modal fracture data;

[0013] S2. A data processing module is connected to the multi-modal sensor fusion module, is used to receive the multi-modal fracture data, and performs noise filtering, data normalization, and feature parameter extraction through a preprocessing algorithm, and outputs a set of feature parameters of the fracture state;

[0014] S3. An intelligent diagnosis module is connected to the data processing module, and a fracture state recognition model constructed based on a deep learning algorithm is used to analyze the set of feature parameters, and generate a diagnosis result of the fracture state and recovery trend prediction data;

[0015] S4. A low-power wireless communication module is connected to the intelligent diagnosis module, adopts a low-power wireless communication architecture, is used to transmit the diagnosis result of the fracture state and the recovery trend prediction data to a monitoring terminal in real time, and realizes dynamic adjustment of system parameters through two-way communication;

[0016] S5. A personalized care plan generation module is connected to the intelligent diagnosis module, combines the biomechanical model of the patient, the biomechanical model is constructed by the personalized characteristic data of the patient and the diagnosis result of the fracture state, the personalized characteristic data includes the bone density, weight distribution, fracture site characteristics, and muscle distribution information of the patient, and generates a set of personalized care plans through an optimization algorithm, including rehabilitation exercise plans, load adjustment strategies, and auxiliary device configuration parameters;

[0017] S6. An adaptive feedback adjustment module is connected to the multi-modal sensor fusion module and the data processing module, and is used to dynamically adjust the data acquisition strategy of the multi-modal sensor fusion module and the set of feature parameters of the data processing module based on the real-time acquired multi-modal fracture data and the change of the patient's state.

[0018] Optionally, the S1 specifically includes:

[0019] S11. The multi-sensor collaborative acquisition unit includes a triaxial acceleration sensor , a micro strain sensor and a fiber Bragg grating sensor , which are respectively used to collect the acceleration signal, strain distribution signal and bioelectro-optical signal of the fracture site. Through the time synchronizer , the collected acceleration signal, strain distribution signal and bioelectro-optical signal are time-aligned, and a time series data set is output;

[0020] S12. The multi-modal signal preprocessing unit is connected to the multi-sensor collaborative acquisition unit, and the time series data set is denoised and normalized through a noise filtering algorithm and a normalization processing method to generate a processed normalized data set , where , , respectively represent the normalized values of the acceleration signal, strain distribution signal and bioelectro-optical signal;

[0021] S13. The multi-modal feature fusion unit is connected to the multi-modal signal preprocessing unit, and the normalized data set is fused based on a weighted feature fusion model to generate a comprehensive feature vector :

[0022] ;

[0023] Among them, , , are weight parameters, which are dynamically adjusted according to the fracture site, recovery stage and patient's individual needs, and satisfy the constraint condition ;

[0024] S14. The dynamic adaptive calibration unit is connected to the multi-modal feature fusion unit. Based on a real-time feedback mechanism and a deep learning optimization model, the weight parameters of the comprehensive feature vector and the signal fusion algorithm are adaptively optimized to correct the deviation caused by environmental interference and patient individual differences, and generate a calibrated fusion feature vector ;

[0025] S15. The fusion data generation unit is connected to the dynamic adaptive calibration unit, and the calibrated fusion feature vector is converted into multi-modal fracture data , including the acceleration signal, strain distribution signal and bioelectro-optical signal of the patient's fracture site.

[0026] Optionally, the S2 specifically includes:

[0027] S21. The data receiving unit is connected to the multi-modal sensor fusion module and is used to receive the multi-modal fracture data set , where represents the multi-modal fracture data collected at time , including acceleration signals, strain distribution signals, and bioelectro-optical signals;

[0028] S22. The noise filtering unit uses an adaptive filtering algorithm based on multi-dimensional spatio-temporal correlation to process each data item in the multi-modal fracture data set and removes random noise and non-linear interference through the following formula:

[0029] ;

[0030] where, represents the data item after noise filtering, is the noise component related to the signal, is the dynamic weight factor, which is optimized and adjusted according to the signal distribution characteristics;

[0031] S23. The normalization processing unit is connected to the noise filtering unit and is used to perform normalization processing on the noise-filtered data set and normalize the signal amplitude to the standard range :

[0032] ;

[0033] where, represents the normalized signal amplitude, and respectively represent the minimum and maximum values of the data set , and generate the normalized standardized data set ;

[0034] S24. The feature extraction unit is connected to the normalization processing unit and, based on the feature analysis model driven by deep learning, extracts features from the standardized data set to construct a feature parameter set , where each feature parameter represents specific mechanical properties, strain distribution, or bioelectrical activity properties related to the fracture state;

[0035] S25. The feature parameter output unit is connected to the feature extraction unit and is used to transfer the feature parameter set to the intelligent diagnosis module.

[0036] Optionally, the S3 specifically includes:

[0037] S31. Feature parameter receiving unit, connected to the data processing module, for receiving the set of feature parameters output by the data processing module , where each feature parameter represents the mechanical properties, strain distribution, and bio-signal properties related to the fracture state;

[0038] S32. Fracture state recognition model unit is connected to the feature parameter receiving unit, and constructs a multi-task neural network model based on deep learning , for simultaneously processing the fracture state classification and recovery stage prediction of the set of feature parameters . The output of the model includes the set of fracture state recognition results and the set of recovery stage classification results :

[0039] ;

[0040] Among them, is the weight parameter of the model, which is dynamically updated according to the training data;

[0041] S33. Adaptive model training unit is connected to the fracture state recognition model unit, and uses a loss function based on an adaptive optimization strategy to dynamically adjust the model weight parameter :

[0042] ;

[0043] Among them, represents the total loss, represents the total number of samples in the training dataset, and are the loss functions for fracture state classification and recovery stage prediction respectively, and are the true labels of the -th sample, represents the classification result output by the fracture state recognition model for the -th sample, is the recovery stage prediction result, and are the loss weight factors, used to balance the training priorities between the two tasks, satisfying ;

[0044] S34. Diagnostic report generation unit is connected to the fracture state recognition model unit, for generating a fracture diagnosis report according to the set of fracture state recognition results and the set of recovery stage classification results , where ​It includes fracture status description, recovery stage analysis, and potential risk assessment;

[0045] S35. A recovery trend prediction unit, based on the fracture diagnosis report and the patient's personalized characteristic data, where the personalized characteristic data includes the patient's bone density, weight distribution, fracture site characteristics, and muscle distribution information, and uses a time series prediction algorithm based on a recurrent neural network to generate a recovery trend prediction set , where represents the estimation of the fracture recovery status at a future time.

[0046] Optionally, the S4 specifically includes:

[0047] S41. A data encoding unit is connected to the intelligent diagnosis module and is used to receive the fracture status diagnosis result set , where the fracture status diagnosis result set includes the comprehensive analysis result of the fracture status recognition result set and the recovery trend prediction set, and compresses and formats the fracture status diagnosis result set through a low-power optimization encoding function to generate encoded data adapted for low-power transmission :

[0048] ;

[0049] where represents the optimized encoding function;

[0050] S42. A low-power transmission unit is connected to the data encoding unit and uses a dynamic power adjustment algorithm and a multi-protocol support architecture to transmit the encoded data in real time through a low-power wireless channel to a remote monitoring terminal. The wireless channel switches to use a long-distance low-power network protocol according to the communication scenario to achieve the optimal configuration of transmission distance and power consumption;

[0051] S43. A data decoding unit is deployed at the remote monitoring terminal and is used to receive the encoded data transmitted by the wireless channel , and decodes and restores the encoded data through a data decoding function to output the fracture status diagnosis result set and the recovery trend prediction set:

[0052] ;

[0053] where is the decoding function, which is consistent with the encoding function to ensure the integrity of data restoration;

[0054] S44. The two-way communication unit is connected to the low-power transmission unit and is used to support two-way data interaction between the monitoring terminal and the intelligent fracture-assisted monitoring system. The unit receives the parameter adjustment instruction set sent by the monitoring terminal and transfers the parameter adjustment instruction set to the intelligent diagnosis module for dynamically adjusting the weight parameters of the fracture state recognition model and other relevant configuration parameters;

[0055] S45. The energy consumption optimization unit is connected to the data encoding unit and the low-power transmission unit. Based on the real-time communication load, data transmission frequency, and channel state, it adjusts the communication power through a dynamic optimization algorithm and the data sending interval to achieve the optimal balance between energy consumption and data transmission quality:

[0056] ;

[0057] wherein, represents the energy consumption optimization function, is the data packet size, is the data sending interval.

[0058] Optionally, the S5 specifically includes:

[0059] S51. The biomechanical modeling unit is connected to the intelligent diagnosis module and is used to receive the patient's personalized feature data and the fracture state diagnosis result set to generate the patient's biomechanical model :

[0060] ;

[0061] wherein, represents the biomechanical modeling function;

[0062] S52. The requirement analysis unit is connected to the biomechanical modeling unit and is used to analyze the patient's rehabilitation requirements based on the biomechanical model and the recovery trend prediction set to generate a requirement vector , where each represents the weight of the rehabilitation requirement, and the requirement vector is used as the input for the optimization calculation;

[0063] S53. The multi-objective optimization unit is connected to the requirement analysis unit. Based on the multi-objective optimization algorithm, it combines the requirement vector , the recovery trend prediction set and the set of feasible care plans Perform optimization calculations, where the set of feasible nursing plans includes a nursing plan library predefined for different fracture types and recovery stages based on orthopedic clinical nursing standards and alternative plans for feasible nursing plans, and generate a set of personalized nursing plans , where is the final optimization result:

[0064] ;

[0065] Among them, is the optimization objective function, is the demand quantity, is the demand weight factor, is the th demand value in the demand vector, is the satisfaction function of the nursing plan for the demand, is the alternative plan for the feasible nursing plan;

[0066] S54. The dynamic plan adjustment unit is connected to the multi-objective optimization unit and is used to, based on the set of real-time updated fracture state diagnosis results and the set of recovery trend predictions , dynamically adjust the set of personalized nursing plans so that the plan continuously matches the patient's rehabilitation needs and status changes;

[0067] S55. The plan output unit is connected to the dynamic plan adjustment unit and is used to output the finally optimized set of personalized nursing plans to the monitoring terminal, including the rehabilitation exercise plan, the load adjustment strategy, and the auxiliary device configuration parameters.

[0068] Optionally, the S6 specifically includes:

[0069] S61. The real-time data acquisition and monitoring unit is connected to the multi-modal sensor fusion module and is used to receive the set of real-time acquired multi-modal fracture data , where represents the multi-modal fracture data acquired at time and includes mechanical property signals, strain property signals, and bioelectric signals;

[0070] S62. The state change trend analysis unit is connected to the data processing module and is used to, based on the set of real-time obtained characteristic parameters and the set of patient state change data , generate a state change trend vector :

[0071] ;

[0072] Among them, Indicates a trend analysis function, used to describe the dynamic association between patient characteristic parameters and state changes;

[0073] S63. An acquisition strategy optimization unit, connected to the real-time data acquisition and monitoring unit and the state change trend analysis unit, is used to dynamically adjust the multi-modal fracture data collected by the multi-modal sensor fusion module according to the state change trend vector The adjusted multi-modal fracture data is expressed as: ;

[0074] ;

[0075] where is an acquisition strategy optimization function, used to dynamically optimize the acquisition frequency, sensitivity, and data fusion strategy of the sensor;

[0076] S64. A characteristic parameter adjustment unit is connected to the data processing module, and is used to optimize the characteristic parameter set according to the adjusted multi-modal fracture data and the state change trend vector : ;

[0077] ;

[0078] where is the optimization objective function, is the total number of data items in the multi-modal fracture data set , is the total number of characteristic parameters in the characteristic parameter set , is the th original characteristic parameter, is the rd characteristic extraction parameter, is the th data item in the adjusted multi-modal fracture data set, is the mapping function;

[0079] S65. A feedback regulation output unit, connected to the acquisition strategy optimization unit and the characteristic parameter adjustment unit, is used to feedback the optimized acquisition parameter set and the characteristic extraction parameter set to the multi-modal sensor fusion module and the data processing module respectively, forming a closed-loop adaptive feedback system to achieve real-time dynamic adjustment.

[0080] The beneficial effects of the present invention are:

[0081] (1) By combining multi-modal sensor fusion technology and deep learning intelligent diagnosis algorithms, the present invention realizes the comprehensive collection and accurate analysis of mechanical parameters, movement trajectory data, and biological signals of fracture patients, enabling the system to dynamically adapt to the real-time state changes of patients, thereby effectively improving the accuracy of fracture state monitoring and diagnosis, especially the accurate identification of complex fracture states and recovery trends.

[0082] (2) Through a low-power wireless communication module, combined with multi-protocol support and dynamic power adjustment algorithms, the present invention realizes the efficient remote transmission of fracture diagnosis results and recovery trend data, while supporting two-way communication to dynamically adjust system parameters. This not only significantly reduces communication energy consumption and is suitable for long-term care scenarios, but also improves the response speed and adaptability of the system.

[0083] (3) Through the personalized care plan generation module, based on the biomechanical model and diagnostic data of patients, combined with optimization algorithms, the present invention dynamically generates rehabilitation exercise plans, load adjustment strategies, and auxiliary device configuration parameters, providing a highly personalized care plan, thereby significantly improving the scientificity and adaptability of the care plan.

[0084] (4) Through the adaptive feedback adjustment module, the present invention realizes the real-time dynamic optimization of data collection strategies and feature extraction parameters, forming a closed-loop adjustment mechanism. This design ensures the efficient response ability of the system to changes in the patient's state and effectively solves the problem of insufficient dynamic adjustment ability in traditional methods.

[0085] (5) As a whole, the present invention provides a comprehensive perspective on fracture monitoring and care through multi-technology integration, enabling the automatic realization of patient monitoring, diagnosis, and care plan generation, reducing the dependence on manual intervention, and at the same time having high intelligence and adaptability, effectively improving the accuracy and efficiency of orthopedic care. BRIEF DESCRIPTION OF THE DRAWINGS

[0086] The drawings are used to provide a further understanding of the present invention and constitute a part of the specification. They are used together with the embodiments of the present invention to explain the present invention and do not constitute a limitation to the present invention. In the drawings:

[0087] Figure 1 is the overall framework diagram of a fracture intelligent auxiliary monitoring system for orthopedic care proposed by the present invention;

[0088] Figure 2 is the flowchart for constructing a fracture state recognition model based on deep learning in the intelligent diagnosis module of the present invention;

[0089] Figure 3 is the optimization calculation flowchart of the personalized care plan generation module of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0090] Now, the present invention will be further described in detail with reference to the accompanying drawings. These drawings are all simplified schematic diagrams, only illustrating the basic structure of the present invention in a schematic manner, so they only show the components related to the present invention.

[0091] Reference Figures 1-3 , an intelligent auxiliary monitoring system for orthopedic care fractures, comprising the following steps:

[0092] S1. A multi-modal sensor fusion module, integrating an acceleration sensor, a strain sensor, and an optical fiber sensor, is used to collect the acceleration signal, strain distribution signal, and bioelectro-optical signal of a fracture patient, and processes the acceleration signal, strain distribution signal, and bioelectro-optical signal through a fusion algorithm to generate multi-modal fracture data;

[0093] S2. The data processing module is connected to the multi-modal sensor fusion module, and is used to receive the multi-modal fracture data, and perform noise filtering, data normalization, and feature parameter extraction through a preprocessing algorithm, and output a set of feature parameters of the fracture state;

[0094] S3. The intelligent diagnosis module is connected to the data processing module, and a fracture state recognition model constructed based on a deep learning algorithm is used to analyze the set of feature parameters, and generate a diagnosis result of the fracture state and recovery trend prediction data;

[0095] S4. The low-power wireless communication module is connected to the intelligent diagnosis module, and adopts a low-power wireless communication architecture, and is used to transmit the diagnosis result of the fracture state and the recovery trend prediction data to the monitoring terminal in real time, and realize the dynamic adjustment of the system parameters through two-way communication;

[0096] S5. The personalized care plan generation module is connected to the intelligent diagnosis module, and combines the biomechanical model of the patient. The biomechanical model is constructed from the personalized characteristic data of the patient and the diagnosis result of the fracture state. The personalized characteristic data includes the patient's bone density, weight distribution, fracture site characteristics, and muscle distribution information, and generates a set of personalized care plans through an optimization algorithm, including a rehabilitation exercise plan, a load adjustment strategy, and auxiliary device configuration parameters;

[0097] S6. The adaptive feedback adjustment module is connected to the multi-modal sensor fusion module and the data processing module, and is used to dynamically adjust the data acquisition strategy of the multi-modal sensor fusion module and the set of feature parameters of the data processing module based on the real-time obtained multi-modal fracture data and the change of the patient's state.

[0098] In this embodiment, the S1 specifically includes:

[0099] S11. A multi-sensor collaborative acquisition unit, including a triaxial acceleration sensor , a micro strain sensor and fiber Bragg grating sensors are respectively used to collect the acceleration signal, strain distribution signal and bioelectro-optical signal of the fracture site. Through the time synchronizer to align the collected acceleration signal, strain distribution signal and bioelectro-optical signal in time, and output a time series data set ;

[0100] S12. The multi-modal signal preprocessing unit is connected to the multi-sensor collaborative acquisition unit. Through the noise filtering algorithm and normalization processing method, the time series data set is denoised and normalized to generate a processed normalized data set , where , , respectively represent the normalized values of the acceleration signal, strain distribution signal and bioelectro-optical signal;

[0101] S13. The multi-modal feature fusion unit is connected to the multi-modal signal preprocessing unit, and based on the weighted feature fusion model, the normalized data set is fused to generate a comprehensive feature vector :

[0102] ;

[0103] Among them, , , are weight parameters, which are dynamically adjusted according to the fracture site, recovery stage and patient's individual needs, and satisfy the constraint condition ;

[0104] S14. The dynamic adaptive calibration unit is connected to the multi-modal feature fusion unit. Based on the real-time feedback mechanism and deep learning optimization model, the weight parameters of the comprehensive feature vector and the signal fusion algorithm are adaptively optimized to correct the deviation caused by environmental interference and patient individual differences, and generate a calibrated fusion feature vector ;

[0105] S15. The fusion data generation unit is connected to the dynamic adaptive calibration unit, and converts the calibrated fusion feature vector into multi-modal fracture data , including the acceleration signal, strain distribution signal and bioelectro-optical signal of the patient's fracture site.

[0106] This embodiment realizes the comprehensive acquisition and standardized processing of multi-dimensional data of the fracture site through the multi-modal sensor collaborative acquisition unit and the signal preprocessing unit. Combining the weighted feature fusion model and the dynamic adaptive calibration unit, it effectively improves the accuracy and robustness of the comprehensive feature vector, thereby correcting the influence of environmental interference and patient individual differences on data quality.

[0107] In this embodiment, the S2 specifically includes:

[0108] S21. The data receiving unit is connected to the multi-modal sensor fusion module for receiving the multi-modal fracture data set , where represents the multi-modal fracture data collected at time , including acceleration signal, strain distribution signal and bioelectro-optical signal;

[0109] S22. The noise filtering unit uses an adaptive filtering algorithm based on multi-dimensional spatio-temporal correlation to process each data item in the multi-modal fracture data set , and removes random noise and non-linear interference through the following formula:

[0110] ;

[0111] where, represents the data item after noise filtering, is the noise component related to the signal, is the dynamic weight factor, which is optimized and adjusted according to the signal distribution characteristics;

[0112] S23. The normalization processing unit is connected to the noise filtering unit for normalizing the denoised data set , and normalizing the signal amplitude to the standard range :

[0113] ;

[0114] where, represents the normalized signal amplitude, and respectively represent the minimum and maximum values of the data set , generating the normalized standardized data set ;

[0115] S24. The feature extraction unit is connected to the normalization processing unit, and based on the deep learning-driven feature analysis model, extracts features from the standardized data set , constructs the feature parameter set , where each feature parameter represent specific mechanical properties, strain distributions, or bioelectrical activity characteristics related to the fracture state;

[0116] S25. The feature parameter output unit is connected to the feature extraction unit and is used to transfer the set of feature parameters to the intelligent diagnosis module.

[0117] Through the coordinated action of multi-modal data reception, noise filtering, normalization processing, and the feature extraction unit, this embodiment realizes the efficient preprocessing of complex fracture data. Combining with the feature analysis model based on deep learning, it accurately extracts specific parameters related to the fracture state, significantly improving the accuracy of data processing and the diagnostic efficiency.

[0118] In this embodiment, S3 specifically includes:

[0119] S31. The feature parameter receiving unit is connected to the data processing module and is used to receive the set of feature parameters output by the data processing module , where each feature parameter represents the mechanical properties, strain distributions, and bio-signal characteristics related to the fracture state;

[0120] S32. The fracture state recognition model unit is connected to the feature parameter receiving unit and constructs a multi-task neural network model based on deep learning for simultaneously processing the fracture state classification and recovery stage prediction of the set of feature parameters . The output of the model includes the set of fracture state recognition results and the set of recovery stage classification results :

[0121] ;

[0122] where, are the weight parameters of the model and are dynamically updated according to the training data;

[0123] S33. The adaptive model training unit is connected to the fracture state recognition model unit and dynamically adjusts the model weight parameters using a loss function based on an adaptive optimization strategy:

[0124] ;

[0125] where, represents the total loss, represents the total number of samples in the training dataset, and are the loss functions for fracture state classification and recovery stage prediction respectively, and are the The true label of the th sample, indicating the classification result output by the fracture status recognition model for the th sample, and is the loss weight factor, used to balance the training priorities between the two tasks, satisfying ;

[0126] S34. The diagnostic report generation unit is connected to the fracture status recognition model unit, and is used to generate a fracture diagnosis report according to the fracture status recognition result set and the recovery stage classification result set where the fracture diagnosis report includes fracture status description, recovery stage analysis, and potential risk assessment;

[0127] S35. The recovery trend prediction unit, based on the fracture diagnosis report and the patient's personalized feature data, where the personalized feature data includes the patient's bone density, weight distribution, fracture site characteristics, and muscle distribution information, uses a time series prediction algorithm based on a recurrent neural network to generate a recovery trend prediction set where represents the estimation of the fracture recovery status at a future time.

[0128] This embodiment combines a multi-task neural network model constructed by deep learning and a time series prediction algorithm of a recurrent neural network to achieve efficient processing of fracture status classification and recovery stage prediction. By dynamically optimizing the model weight parameters and generating a comprehensive fracture diagnosis report, not only the accuracy of the diagnosis result is improved, but also the recovery trend of the patient can be predicted, providing reliable data support for personalized care.

[0129] In this embodiment, the S4 specifically includes:

[0130] S41. The data encoding unit is connected to the intelligent diagnosis module, and is used to receive the fracture status diagnosis result set where the fracture status diagnosis result set includes the comprehensive analysis result of the fracture status recognition result set and the recovery trend prediction set, and compresses and formats the fracture status diagnosis result set through a low-power optimized encoding function to generate encoded data adapted for low-power transmission :

[0131] ;

[0132] where represents the optimized encoding function;

[0133] S42. The low-power transmission unit is connected to the data encoding unit. It adopts a dynamic power adjustment algorithm and a multi-protocol support architecture to transmit the encoded data in real time through a low-power wireless channel to the remote monitoring terminal. The wireless channel switches to use a long-distance low-power network protocol according to the communication scenario to achieve the optimal configuration of transmission distance and power consumption;

[0134] S43. The data decoding unit is deployed at the remote monitoring terminal and is used to receive the encoded data transmitted by the wireless channel and decode and restore the encoded data through a data decoding function to output a set of fracture status diagnosis results and a set of recovery trend predictions:

[0135] ;

[0136] Among them, is the decoding function, which maintains consistency with the encoding function to achieve the integrity of data restoration;

[0137] S44. The two-way communication unit is connected to the low-power transmission unit and is used to support two-way data interaction between the monitoring terminal and the fracture intelligent auxiliary monitoring system. The unit receives a set of parameter adjustment instructions sent by the monitoring terminal and transfers the set of parameter adjustment instructions to the intelligent diagnosis module to dynamically adjust the weight parameters of the fracture status recognition model and other relevant configuration parameters;

[0138] S45. The energy consumption optimization unit is connected to the data encoding unit and the low-power transmission unit. Based on the real-time communication load, data transmission frequency, and channel state, it adjusts the communication power and the data sending interval through a dynamic optimization algorithm to achieve the optimal balance between energy consumption and data transmission quality:

[0139] ;

[0140] Among them, represents the energy consumption optimization function, is the packet size, is the data sending interval.

[0141] The present invention realizes the efficient remote transmission of fracture state diagnosis results and recovery trend data by combining low-power optimized coding and dynamic power adjustment algorithms, supports multi-protocol switching to adapt to different communication scenarios, and ensures data integrity and real-time performance. Combined with an energy consumption optimization unit, it significantly reduces communication energy consumption, improves the applicability of the system in long-term care scenarios, and supports two-way communication to achieve dynamic parameter adjustment, effectively enhancing transmission efficiency and response capabilities.

[0142] In this embodiment, the S5 specifically includes:

[0143] S51. A biomechanical modeling unit, connected to the intelligent diagnosis module, for receiving patient personalized feature data and the fracture state diagnosis result set , and generating a biomechanical model of the patient :

[0144] ;

[0145] Wherein, represents the biomechanical modeling function;

[0146] S52. A requirement analysis unit is connected to the biomechanical modeling unit, and is used to analyze the patient's rehabilitation requirements based on the biomechanical model and the recovery trend prediction set , generating a requirement vector , where each represents the weight of the rehabilitation requirement, and the requirement vector is used as the input for the optimization calculation;

[0147] S53. A multi-objective optimization unit is connected to the requirement analysis unit. Based on the multi-objective optimization algorithm, the requirement vector , the recovery trend prediction set and the set of feasible care plans are subjected to optimization calculation. Among them, the set of feasible care plans includes a care plan library predefined for different fracture types and recovery stages based on orthopedic clinical care standards and alternative plans for feasible care plans, generating a set of personalized care plans , where is the final optimization result:

[0148] ;

[0149] Wherein, is the optimization objective function, is the number of requirements, is the requirement weight factor, is the th requirement value in the requirement vector, is the satisfaction function of the nursing plan for the needs is an alternative to the feasible nursing plan;

[0150] S54. The dynamic plan adjustment unit is connected to the multi-objective optimization unit, and is used to, based on the set of fracture status diagnosis results updated in real time and the set of recovery trend predictions , dynamically adjust the set of personalized nursing plans so that the plan continuously matches the patient's rehabilitation needs and status changes;

[0151] S55. The plan output unit is connected to the dynamic plan adjustment unit, and is used to output the finally optimized set of personalized nursing plans to the monitoring terminal, including the rehabilitation exercise plan, the load adjustment strategy, and the auxiliary device configuration parameters.

[0152] Through the combination of biomechanical modeling, rehabilitation needs analysis, and multi-objective optimization algorithms, the present invention realizes the dynamic generation and adjustment of the patient's personalized nursing plan, which can not only meet the diverse rehabilitation needs of the patient, but also adapt to the patient's status changes in real time, provide a highly accurate rehabilitation exercise plan, a load adjustment strategy, and auxiliary device configuration parameters, and effectively improve the scientificity and adaptability of nursing.

[0153] In this embodiment, the S6 specifically includes:

[0154] S61. The real-time data acquisition and monitoring unit is connected to the multi-modal sensor fusion module, and is used to receive the set of multi-modal fracture data collected in real time , where represents the multi-modal fracture data collected at time , including the mechanical property signal, the strain property signal, and the bioelectric signal;

[0155] S62. The state change trend analysis unit is connected to the data processing module, and is used to, based on the set of feature parameters obtained in real time and the set of patient state change data , generate the state change trend vector :

[0156] ;

[0157] wherein, represents the trend analysis function used to describe the dynamic association between the patient's feature parameters and the state change;

[0158] S63. The acquisition strategy optimization unit, which is connected to the real-time data acquisition and monitoring unit and the state change trend analysis unit, is used to, according to the state change trend vector , dynamically adjust the multi-modal fracture data collected by the multi-modal sensor fusion module , the adjusted multi-modal fracture data is expressed as:

[0159] ;

[0160] wherein, is the acquisition strategy optimization function, which is used to dynamically optimize the acquisition frequency, sensitivity and data fusion strategy of the sensor;

[0161] S64. The feature parameter adjustment unit is connected to the data processing module, and is used to optimize the feature parameter set and the state change trend vector according to the adjusted multi-modal fracture data :

[0162] ;

[0163] wherein, is the optimization objective function, is the multi-modal fracture data set the total number of data items, is the feature parameter set the total number of feature parameters in, is the th original feature parameter, is the th feature extraction parameter, is the th data item in the adjusted multi-modal fracture data set, is the mapping function;

[0164] S65. The feedback regulation output unit is connected to the acquisition strategy optimization unit and the feature parameter adjustment unit, and is used to feed back the optimized acquisition parameter set and the feature extraction parameter set to the multi-modal sensor fusion module and the data processing module respectively, to form a closed-loop adaptive feedback system and realize real-time dynamic adjustment.

[0165] In this embodiment, through real-time data acquisition, state change trend analysis and acquisition strategy optimization, the acquisition frequency, sensitivity and data fusion strategy of the multi-modal sensor are dynamically adjusted, and combined with the feature parameter optimization and feedback regulation mechanism, a closed-loop adaptive system is formed to realize real-time response and precise monitoring of the patient's state change, and greatly improve the accuracy and dynamic adaptation ability of the monitoring. Embodiment

[0166] To verify the feasibility and effectiveness of the present invention, the present invention is applied to the orthopedic rehabilitation center of a Class III Grade A hospital in a certain city. This rehabilitation center is responsible for the treatment and care of various fracture patients. The fracture types of the patients are complex, and the recovery needs are diverse. The traditional fracture monitoring and care methods are difficult to meet the requirements of real-time, precision, and personalization. This system is deployed in the rehabilitation center for real-time monitoring, intelligent diagnosis, and generation of personalized care plans for fracture patients.

[0167] In the application scenario, the system includes a multi-modal sensor fusion module, a data processing module, an intelligent diagnosis module, a low-power wireless communication module, a personalized care plan generation module, and an adaptive feedback adjustment module. The system is equipped with an integrated sensor device, including a mechanical sensor, a strain sensor, and a bioelectric sensor, which is worn on the fracture site of the patient to collect real-time fracture data of the patient. The intelligent diagnosis module uses deep learning algorithms to analyze the collected multi-modal data and generate a diagnosis result of the fracture state and prediction data of the recovery trend. The personalized care plan generation module generates a precise care plan according to the patient's diagnosis result and biomechanical model, including rehabilitation exercise guidance and auxiliary device configuration.

[0168] Ten patients in the inpatient ward of the rehabilitation center were selected for the test scenario, and the fracture sites included the femur, tibia, and humerus. During the test, the patients wore the system for 14 days of rehabilitation training and key data were recorded. The test contents included the performance of the system in real-time monitoring, diagnostic accuracy, care plan adaptability, and energy consumption management. The data summary is shown in Table 1.

[0169] Table 1 Comparison table of the performance of the intelligent fracture-assisted monitoring system

[0170] Index category Traditional nursing method The system of the present invention Fracture monitoring data coverage rate (%) 45.2 96.8 Accuracy rate of fracture status diagnosis (%) 72.5 95.3 Satisfaction rate of nursing plan (%) 65.7 93.4 Completion rate of rehabilitation training (%) 70.4 98.1 Average energy consumption of data transmission (mW) 56 23

[0171] It can be seen from the data in Table 1 that the system of the present invention is significantly superior to the traditional care method in many aspects. The coverage rate of fracture monitoring data has increased from 45.2% to 96.8%, indicating that the system can capture more comprehensive patient status information. The diagnostic accuracy of the fracture state has increased from 72.5% to 95.3%, benefiting from the precise analysis of complex data by deep learning algorithms. The satisfaction rate of patients with the care plan has increased significantly, from 65.7% to 93.4%, reflecting the scientificity and applicability of the personalized care plan. The completion rate of rehabilitation training has increased from 70.4% to 98.1%, indicating that the care plan is more operable and effective. The low-power communication module of the system reduces the average energy consumption of data transmission by 59%, which is very suitable for long-term monitoring.

[0172] In a specific application, a patient was tested for tibial fracture. After wearing the system, the sensor collected mechanical characteristic signals and bioelectrical data in real time, and the data processing module automatically filtered out noise and extracted characteristic parameters. The intelligent diagnosis module analyzed the data and indicated that the healing status of the patient's fracture site had reached the mid - stage and predicted that the healing trend would tend to be stable in the next 7 days. The personalized care plan generation module developed a low - load rehabilitation training plan for the patient and recommended adjusting the crutch fulcrum to reduce the stress on the fracture site. During the training process of the patient, the adaptive feedback adjustment module dynamically adjusted the acquisition sensitivity of the sensor and the characteristic extraction parameters according to the real - time data to ensure the accuracy of monitoring.

[0173] In summary, the intelligent fracture - assisted monitoring system for orthopedic care of the present invention effectively solves the problems of incomplete monitoring, inaccurate diagnosis, and inappropriate care plans in traditional care methods. It not only significantly improves the efficiency and effect of fracture care but also enhances the intelligence and adaptability of the system through multi - modal data fusion and real - time feedback adjustment mechanisms, providing an efficient and accurate solution for orthopedic care.

[0174] The above - mentioned is only the preferred specific implementation mode of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention, according to the technical solution and inventive concept of the present invention, making equivalent substitutions or changes should be covered within the protection scope of the present invention.

Claims

1. An intelligent auxiliary monitoring system for fractures used in orthopedic care, characterized in that: include: S1, multimodal sensor fusion module, integrating acceleration sensor, strain sensor and optical fiber sensor, used to collect acceleration signal, strain distribution signal and bio-electro-optical signal of fracture patients, and process the acceleration signal, strain distribution signal and bio-electro-optical signal through fusion algorithm to generate multimodal fracture data; S2, the data processing module is connected to the multimodal sensor fusion module, and is used to receive the multimodal fracture data, and perform noise filtering, data normalization and feature parameter extraction through a preprocessing algorithm, and output a feature parameter set of the fracture state; S3, the intelligent diagnosis module is connected with the data processing module, and the fracture state recognition model constructed based on the deep learning algorithm is used to analyze the characteristic parameter set and generate the diagnosis result of the fracture state and the recovery trend prediction data; S4, the low-power wireless communication module is connected to the intelligent diagnosis module, and adopts a low-power wireless communication architecture to transmit the diagnosis results of the fracture status and the recovery trend prediction data to the monitoring terminal in real time, and realize the dynamic adjustment of the system parameters through two-way communication; S5. The personalized nursing plan generation module is connected with the intelligent diagnosis module, and combined with the patient's biomechanical model, the biomechanical model is constructed by the patient's personalized feature data and the diagnosis result of the fracture status. The personalized feature data includes the patient's bone density, weight distribution, fracture site characteristics and muscle distribution information, and generates a personalized nursing plan set through an optimization algorithm, including a rehabilitation exercise plan, a load adjustment strategy and auxiliary device configuration parameters; S6, the adaptive feedback adjustment module is connected to the multimodal sensor fusion module and the data processing module, and is used to dynamically adjust the data acquisition strategy of the multimodal sensor fusion module and the characteristic parameter set of the data processing module based on the multimodal fracture data acquired in real time and the patient's state change; The S3 specifically includes: S31, a characteristic parameter receiving unit connected to the data processing module, used to receive the characteristic parameter set output by the data processing module , where each characteristic parameter Represents mechanical properties, strain distribution and biological signal characteristics related to fracture status; S32, the fracture state recognition model unit is connected to the characteristic parameter receiving unit, and a multi-task neural network model is constructed based on deep learning , used to process a set of feature parameters simultaneously The fracture status classification and recovery stage prediction of the model, the output of the model includes the fracture status recognition result set And the recovery phase classification result set : ； in, is the weight parameter of the model, which is dynamically updated according to the training data; S33, the adaptive model training unit is connected to the fracture state recognition model unit, and the model weight parameters are adjusted using a loss function based on an adaptive optimization strategy. To make dynamic adjustments: ； in, represents the total loss, represents the total number of samples in the training dataset, and are the loss functions for fracture status classification and recovery stage prediction, respectively. and For the The true labels of samples, The fracture state recognition model is targeted at The classification results of the sample output are To predict the outcome for the recovery phase, and is the loss weight factor, which is used to balance the training priority between the two tasks, satisfying ; S34, the diagnosis report generation unit is connected to the fracture state recognition model unit, and is used to identify the fracture state result set according to the fracture state recognition result set. And the recovery phase classification result set Generate fracture diagnosis report ,in Includes description of fracture status, analysis of recovery stages, and assessment of potential risks; S35, recovery trend prediction unit, based on fracture diagnosis report The patient's personalized feature data, including the patient's bone density, weight distribution, fracture site characteristics and muscle distribution information, uses a time series prediction algorithm based on a recursive neural network to generate a recovery trend prediction set ,in Represents an estimate of the fracture recovery state at a future time.

2. The intelligent auxiliary monitoring system for fractures used in orthopedic care according to claim 1 is characterized in that: The S1 specifically includes: S11, multi-sensor collaborative acquisition unit, including a three-axis acceleration sensor , Micro strain sensor and Fiber Bragg Grating Sensors , which are used to collect the acceleration signal, strain distribution signal and bio-electro-optical signal of the fracture site, respectively, through the time synchronizer Time-align the collected acceleration signals, strain distribution signals and bio-electro-optical signals, and output a time series data set ; S12, the multimodal signal preprocessing unit is connected to the multi-sensor collaborative acquisition unit, and the time series data set is processed by a noise filtering algorithm and a normalization processing method. Perform denoising and standardization to generate a processed standardized data set ,in , , They represent the normalized values ​​of acceleration signal, strain distribution signal and bio-electro-optical signal respectively; S13, the multimodal feature fusion unit is connected to the multimodal signal preprocessing unit, and the standardized data set is processed based on the weighted feature fusion model Fusion to generate comprehensive feature vector : ； in, , , is the weight parameter, which is dynamically adjusted according to the fracture site, recovery stage and individual needs of the patient to meet the constraints ; S14, the dynamic adaptive calibration unit is connected with the multimodal feature fusion unit, based on the real-time feedback mechanism and deep learning optimization model, to calculate the comprehensive feature vector The weight parameter The signal fusion algorithm is adaptively optimized to correct the deviation caused by environmental interference and individual differences of patients, and generate a calibrated fusion feature vector ; S15, the fusion data generation unit is connected to the dynamic adaptive calibration unit, and the calibrated fusion feature vector Conversion to multimodal fracture data , including acceleration signals, strain distribution signals and bio-electro-optical signals at the patient's fracture site.

3. The intelligent auxiliary monitoring system for fractures used in orthopedic care according to claim 1 is characterized in that: The S4 specifically includes: S41, the data encoding unit is connected to the intelligent diagnosis module to receive the fracture status diagnosis result set The fracture status diagnosis result set includes the comprehensive analysis results of the fracture status recognition result set and the recovery trend prediction set, and the encoding function is optimized through low power consumption. Compress and format the fracture status diagnosis result set to generate coded data suitable for low-power transmission : ； in, represents the optimized encoding function; S42, the low-power transmission unit is connected to the data encoding unit, adopts a dynamic power adjustment algorithm and a multi-protocol support architecture, and transmits the encoded data in real time through a low-power wireless channel To the remote monitoring terminal, the wireless channel switches to use a long-distance low-power network protocol according to the communication scenario , achieving the optimal configuration of transmission distance and power consumption; S43, the data decoding unit is deployed in the remote monitoring terminal to receive the coded data transmitted through the wireless channel , through the data decoding function Decode and restore the encoded data, and output the fracture status diagnosis result set and recovery trend prediction set: ； in, is the decoding function, and the encoding function Maintain consistency to achieve data restoration integrity; S44, the two-way communication unit is connected to the low-power transmission unit to support two-way data interaction between the monitoring terminal and the fracture intelligent auxiliary monitoring system, and the unit receives the parameter adjustment instruction set sent by the monitoring terminal , and pass the parameter adjustment instruction set to the intelligent diagnosis module to dynamically adjust the weight parameters of the fracture state recognition model ; S45, the energy consumption optimization unit is connected with the data encoding unit and the low power transmission unit, and adjusts the communication power through a dynamic optimization algorithm based on the real-time communication load, data transmission frequency and channel status. and data transmission interval , to achieve the optimal balance between energy consumption and data transmission quality: ； in, represents the energy consumption optimization function, is the packet size, The data sending interval.

4. The intelligent auxiliary monitoring system for fractures used in orthopedic care according to claim 3 is characterized in that: The S5 specifically includes: S51, biomechanical modeling unit, connected to the intelligent diagnosis module, for receiving patient personalized characteristic data and fracture status diagnosis result set , generating a biomechanical model of the patient : ； in, represents the biomechanical modeling function; S52, the demand analysis unit is connected with the biomechanical modeling unit for and recovery trend forecast set , analyze the patient's rehabilitation needs and generate a demand vector , where each represents the weight of rehabilitation demand, and the demand vector is used as the input of the optimization calculation; S53, the multi-objective optimization unit is connected to the demand analysis unit, based on the multi-objective optimization algorithm, the demand vector , restore trend forecast collection and a collection of possible care options Perform optimization calculations, where the feasible care plan set includes a library of care plans for different fracture types and recovery stages based on orthopedic clinical care standards, as well as alternatives to feasible care plans, to generate a personalized care plan set ,in For the final optimization result: ； in, To optimize the objective function, is the demand quantity, is the demand weight factor, is the first The demand value, is the satisfaction function of the nursing plan for the needs, Alternatives for viable care options; S54, the dynamic scheme adjustment unit is connected with the multi-objective optimization unit to be used for the diagnosis result set of the fracture status based on real-time update and recovery trend forecast set , a collection of personalized care plans Make dynamic adjustments to ensure that the program continues to match the patient's rehabilitation needs and status changes; S55, the program output unit is connected to the dynamic program adjustment unit to collect the final optimized personalized nursing program Output to the monitoring terminal includes rehabilitation exercise plan, load adjustment strategy and auxiliary device configuration parameters.

Images