Panoramic intelligent rehabilitation management system for burn patients

Abstract

The invention discloses a panoramic intelligent rehabilitation management system for burn patients, and relates to the technical field of medical health informatization. Comprising a multi-target coupling modeling module, a multi-source state sensing module, a conflict identification calculation module, an intervention priority regulation and control module, a feedback calibration analysis module and a dynamic closed-loop control module, a rehabilitation stage multi-target model containing a coupling relationship among a skin healing target, a joint flexibility recovery target and a nerve regeneration target is constructed, a sensitive intervention window is identified according to a physiological change rule of each target in a rehabilitation period, and a target conflict mapping matrix with a time sequence attribute is generated. Through multi-target coupling modeling and a dynamic conflict regulation and control mechanism, intelligent cooperative control over multi-target intervention in the whole burn rehabilitation process is achieved, intervention safety and individual adaptability are improved, closed-loop optimization capacity is achieved, rehabilitation efficiency is remarkably improved, medical risks are reduced, and wide clinical application prospects are achieved.

Description

Technical Field

[0001] This invention relates to the field of medical and health information technology, specifically to a panoramic intelligent rehabilitation management system for burn patients. Background Technology

[0002] The panoramic intelligent rehabilitation management system for burn patients utilizes multi-source data collected from patients (such as wound healing progress, physiological indicators, psychological state, and functional assessment results) and big data technology to tailor personalized rehabilitation plans for burn patients. Based on the individual characteristics and recovery curve of each patient, it can intelligently predict key risk points, dynamically adjust rehabilitation intervention programs, and provide personalized treatment plans, psychological counseling, nutritional support, and functional training suggestions. This achieves precise support and collaborative management for burn patients throughout the entire cycle of hospitalization, rehabilitation, and follow-up, significantly improving rehabilitation efficiency and quality of life.

[0003] The existing technology has the following shortcomings: In the full-cycle rehabilitation management of burn patients, the system needs to simultaneously implement interventions and optimizations targeting multiple rehabilitation goals, such as skin healing, restoration of joint flexibility, and nerve regeneration. However, there is a dynamic inhibitory relationship between these goals, meaning that interventions for one goal may potentially interfere with another. For example, while training to promote joint mobility helps prevent joint stiffness and scar contractures, high-intensity training performed before the wound has fully healed or when tissue tension is unstable can easily lead to skin wound dehiscence, bleeding, or secondary infection, thus delaying the skin repair process. Simultaneously, frequent traction may cause secondary damage to superficial nerves in the early stages of regeneration, reducing the efficiency of nerve function recovery.

[0004] Existing rehabilitation systems often focus on single-goal or static assessments, lacking mechanisms for modeling the real-time interactions between different rehabilitation goals. This makes it difficult to dynamically identify functional conflicts and phased inhibition patterns between goals, leading to a lack of coordination and balance in the system's intervention recommendations. For example, directly calling standardized training templates without the system recognizing the risk to wound stability may generate exercise programs that are incompatible with the patient's current tissue condition, potentially causing wound rupture, delayed healing, or even medical malpractice.

[0005] The information disclosed in the background section is only intended to enhance the understanding of the background of this disclosure, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0006] The purpose of this invention is to provide a panoramic intelligent rehabilitation management system for burn patients to solve the problems mentioned in the background art.

[0007] To achieve the above objectives, the present invention provides the following technical solution: a panoramic intelligent rehabilitation management system for burn patients, comprising a multi-objective coupling modeling module, a multi-source state perception module, a conflict identification and calculation module, an intervention priority control module, a feedback calibration and analysis module, and a dynamic closed-loop control module. The multi-objective coupling modeling module constructs a phased multi-objective model of rehabilitation that includes the coupling relationship between skin healing objectives, joint flexibility recovery objectives, and nerve regeneration objectives. It identifies sensitive intervention windows based on the physiological change patterns of each objective during the rehabilitation cycle and generates an objective conflict mapping matrix with time-series attributes. The multi-source state perception module embeds the target conflict mapping matrix into the data processing channel to collect the patient's wound healing progress, tissue tension changes and joint activity load in each rehabilitation cycle, and constructs a state vector set that integrates time and structural stress dimensions. The conflict identification and calculation module constructs a conflict index calculation mechanism based on the state vector set. It combines the direction and rate of change of state indicators with the coupling sensitivity factor to generate a conflict index curve and marks potential conflict nodes by comparing it with the risk threshold. The intervention priority control module loads the target intervention priority decision logic onto the identified conflict nodes. Based on physiological dependence and the intensity of reverse intervention, it dynamically adjusts the intervention intensity, sequence, and time window of each target to complete the differentiated intervention configuration. The feedback calibration analysis module collects target response data after completing the differentiated intervention configuration, performs time-series backtracking and differential analysis, evaluates the state offset trend, and updates the target conflict mapping matrix. The dynamic closed-loop control module feeds back the updated target conflict mapping matrix and target response offset parameters to the multi-objective model, forming a dynamic closed-loop control mechanism that includes state perception, conflict prediction, intervention configuration and feedback calibration, thereby achieving the coordinated advancement of concurrent intervention for multiple objectives.

[0008] Preferably, the process of constructing a multi-objective model for rehabilitation stages includes the following steps: Establish a phased distribution model of skin healing goals, joint flexibility recovery goals, and nerve regeneration goals during the rehabilitation cycle, and divide the recovery stages according to the physiological indicators of each stage; Based on the impact of intervention behaviors for each rehabilitation goal within the same time period, determine the intensity of conflict or synergy between goal stages; A target conflict mapping matrix is ​​constructed based on the relationship between the time axis and the target combination, and the conflict intensity value in each time interval is marked. The target conflict mapping matrix is ​​used as the basis for intervention configuration and state matching, and conflict identification and dynamic correction are achieved by combining patient recovery data.

[0009] Preferably, the process of embedding the target conflict mapping matrix into the data processing channel specifically includes the following steps: The target conflict mapping matrix is ​​loaded into the data preprocessing process to identify the conflict intensity of each rehabilitation goal at different time periods; Collect data on wound healing progress, tissue tension changes, and joint activity load. All collected data were processed using a unified time standardization method to construct a state vector that includes wound closure rate, red area, wound temperature, tension, joint flexion angle, training time, number of training sessions, and pain score. The current state vector is compared with the target conflict mapping matrix to identify whether it is in a high-incidence range of target conflict, and intervention risk warning is given accordingly.

[0010] Preferably, based on comparing the state vector at the current time point with the target conflict mapping matrix, if it is identified that the current time point is in a high-incidence range where there is a moderate or greater conflict between the skin healing target and the joint flexibility recovery target, the system generates an intervention risk warning message and instructs to adjust the intervention intensity and training time of joint training or to suspend the intervention operation to prevent wound damage caused by excessive tension.

[0011] Preferably, the process of constructing the conflict index calculation mechanism includes the following steps: Extract the direction of change, amount of change, and rate of change per unit time of each dimension index in the continuous state vector; Based on the interference characteristics between skin healing goals, joint flexibility recovery goals, and nerve regeneration goals, a coupling sensitivity factor was set, and the conflict impact value of each goal pair was calculated by weighting it with the index change. The total conflict index at the current time point is generated by summing the conflict impact values ​​of all target pairs. The conflict index is compared with the risk threshold. If it exceeds the preset threshold, it is marked as a potential conflict intervention node.

[0012] Preferably, the process of configuring differentiated intervention for identified conflict nodes includes the following steps: Extract the physiological dependencies of conflicting target combinations and assign standardized dependency levels; Analyze the reverse interference intensity of the current intervention behavior on other targets and form an interference matrix; Calculate intervention priorities based on dependency level and interference intensity, and adjust the intervention intensity, execution order, and time parameters of the target. Based on the adjustment results, the intervention priority level, training intensity level, and intervention time window range are set for each objective.

[0013] Preferably, the process of analyzing the reverse interference intensity of the current intervention behavior on other targets and forming an interference matrix includes the following steps: For each target intervention, key physiological data were collected during the intervention period, including changes in skin tension, wound closure rate, changes in joint range of motion, and changes in nerve reflex scores. Compare the changes in key physiological data before and after the intervention to determine whether the intervention caused significant fluctuations in other target physiological indicators; Based on the direction and magnitude of the indicator changes, an interference intensity level is set, and a numerical weight value is assigned to each level. Using the target to be intervened as the row coordinate and the affected target as the column coordinate, a two-dimensional matrix is ​​constructed and the corresponding interference intensity values ​​are filled in to complete the construction of the interference matrix.

[0014] Preferably, the process of updating the target conflict mapping matrix includes the following steps: After completing the differentiated intervention configuration, response data for skin healing goals, joint flexibility recovery goals, and nerve regeneration goals were collected; Based on response data from three consecutive rehabilitation cycles, a temporal backtracking and differential analysis of the target state are performed to determine the direction and degree of deviation of the target state. The offset result is compared with the corresponding target combination conflict level in the original target conflict mapping matrix to determine whether the conflict level needs to be adjusted. Record the adjusted conflict level, corresponding time point, target pair number, reason for adjustment, and original offset data into the target conflict mapping matrix and save it as an updated version.

[0015] Preferably, the process of comparing the offset result with the corresponding target combination conflict level in the original target conflict mapping matrix to determine whether the conflict level needs to be adjusted includes the following steps: Extract the offset direction and offset magnitude of various status indicators of skin healing target, joint flexibility recovery target and nerve regeneration target within the current cycle; Identify the original conflict level of the target combination in the current conflict node and locate its corresponding position in the original target conflict mapping matrix; Based on the offset trend of the two targets within the same period, determine whether there are obvious opposing changes, one positive and one negative. If so, it is judged that there is a significant interference tendency. If both objectives show a positive improvement trend, it is determined that the actual conflict relationship is weaker than the set level, and the conflict level is downgraded. If there is a negative deviation, it is determined whether to upgrade the conflict level based on the degree of deviation and the dependency relationship between the objectives.

[0016] Preferably, the process of feeding back the updated target conflict mapping matrix and target response offset parameters into the rehabilitation phased multi-objective model includes the following steps: The updated target conflict mapping matrix is ​​imported into the internal control structure of the rehabilitation stage multi-objective model in a three-dimensional array structure to set the intervention priority and execution order. The response offset parameters of each rehabilitation goal in the previous intervention cycle are input into the state-aware structure, and the physiological stage labels of the rehabilitation goals in the current model are corrected accordingly. The intervention strategy configuration process is re-executed based on the updated conflict mapping level and stage label, and the output is an intervention plan containing specific target numbers, intervention parameters and warning thresholds. The intervention response after execution is incorporated into the data collection process of the next cycle, forming a closed-loop control mechanism consisting of conflict identification, status analysis, intervention configuration and response calibration.

[0017] The technical effects and advantages provided by the present invention in the above technical solution are as follows: This invention, by introducing a multi-objective coupling model and a time-series conflict mapping mechanism, breaks through the limitations of traditional rehabilitation programs that prioritize a single objective and control a fixed process. For the first time, it achieves a complete management chain across multiple objectives, encompassing "dynamic relationship modeling—real-time conflict identification—differentiated adjustment—effect feedback calibration—model update closed loop." Especially when dealing with potential intervention conflicts between objectives, this system can perceive changes in individual tissue states in real time. Through conflict index calculation and priority decision logic, it effectively avoids the problem of "physiological reverse interference" between intervention measures, thereby significantly improving the safety, individual adaptability, and clinical effectiveness of interventions. Furthermore, the continuous feedback mechanism enables model self-iterative optimization, significantly enhancing the system's adaptability to the diversity of patient states and the complexity of rehabilitation pathways. It truly achieves intelligent, closed-loop, and collaborative control of the entire burn rehabilitation process, improving rehabilitation efficiency while reducing medical risks, and possesses broad clinical application value and promising prospects for promotion. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0019] Figure 1 This is a schematic diagram of a panoramic intelligent rehabilitation management system for burn patients according to the present invention.

[0020] Figure 2This is a diagram illustrating the generation principle of the conflict index curve in a panoramic intelligent rehabilitation management system for burn patients.

[0021] Figure 3 This is a flowchart illustrating the update process of the target conflict mapping matrix in a panoramic intelligent rehabilitation management system for burn patients. Detailed Implementation

[0022] Exemplary embodiments will now be described more fully with reference to the accompanying drawings. However, these exemplary embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, they are provided so that the description of this disclosure will be more complete and fully convey the concept of the exemplary embodiments to those skilled in the art.

[0023] This invention provides, for example Figure 1 The panoramic intelligent rehabilitation management system for burn patients shown includes a multi-objective coupling modeling module, a multi-source state perception module, a conflict identification and calculation module, an intervention priority control module, a feedback calibration and analysis module, and a dynamic closed-loop control module. The multi-objective coupling modeling module constructs a multi-objective rehabilitation stage model that includes the coupling relationship between skin healing objectives, joint flexibility recovery objectives, and nerve regeneration objectives. Based on the physiological change patterns of each rehabilitation objective during the rehabilitation cycle, it identifies the sensitive intervention window for each objective and forms a target conflict mapping matrix with time-series attributes. To effectively model the interactions between various rehabilitation goals throughout the entire rehabilitation process for burn patients, a multi-goal modeling method for rehabilitation stages, focusing on skin healing, joint flexibility recovery, and nerve regeneration goals, is proposed. This method clarifies the interactive inhibition characteristics between each goal and their time-sensitive relationships, specifically including the following steps: A phased distribution model of skin healing goals, joint flexibility recovery goals, and nerve regeneration goals was established within the rehabilitation cycle. By analyzing the clinical rehabilitation pathways of a large sample of burn patients, combined with treatment literature and the experience of medical experts, the three types of rehabilitation goals were divided into several typical rehabilitation stages. Skin healing goals were divided into the wound debridement stage, granulation tissue formation stage, epidermal coverage stage, and scar remodeling stage. Key physiological parameters for each stage included wound closure degree, exudate secretion volume, wound color change, local temperature change, and local pain sensitivity. Joint flexibility recovery goals were divided into the early passive mobilization stage, the intermediate active assisted mobilization stage, and the late active weight-bearing training stage. Indicators of interest included joint range of motion (in degrees), resistance sensation during joint movement, patient subjective pain score, and muscle responsiveness. Nerve regeneration goals were divided into the nerve ending budding stage, axonal guidance stage, myelination stage, and functional remodeling stage. Corresponding indicators included local sensory threshold test results (such as two-point discrimination test distance), electrophysiological test results (such as action potential conduction velocity), voluntary motor function score, and reflex integrity assessment. The above-mentioned division of stages is based on time span and physiological characteristics, forming a standardized target stage model that covers the entire rehabilitation cycle.

[0024] Based on the established phase model of three types of objectives, this study further analyzes whether there are mutual influences between the intervention behaviors of each objective within the same time period. Using each phase as an analysis unit, and combining actual clinical intervention operations, potential conflict relationships between objectives are compared one by one. For example, during the granulation tissue formation stage when the skin wound is not yet fully closed, applying large-amplitude joint traction training may lead to excessive wound tension, resulting in dehiscence or bleeding. Therefore, a high-intensity conflict relationship is identified between the early stage of skin healing and the early stage of passive joint movement. Similarly, during the axonal guidance stage of nerve regeneration, excessively high frequency or heavy load on passive joint movement can easily cause micro-damage, affecting the stability of the nerve growth pathway. Therefore, a moderate-intensity conflict relationship is identified between the early stage of nerve regeneration and the moderate-intensity joint training stage. Furthermore, during the skin scar remodeling stage, scientifically applying tension helps improve scar texture, and there is a synergistic relationship with joint function training at this time; the conflict coefficient can be set to zero or even negative to indicate a promoting effect. Finally, for all combinations of objective phases, their physiological dependencies, potential inhibitory mechanisms, and possible synergistic relationships are clarified, providing a complete foundation for subsequent relationship modeling.

[0025] Based on the conflict and synergy relationships between target stages, the combination relationships among the three types of targets are organized into a time-related target conflict mapping matrix. This mapping table uses the actual rehabilitation timeline as the horizontal dimension and all target stage combinations as the vertical dimension. For each specific time point and target stage combination, a marker indicating the existence of a conflict relationship is entered (e.g., a conflict intensity value of 0 indicates no conflict, 1 indicates slight conflict, 2 indicates moderate conflict, 3 indicates high conflict, and -1 indicates a synergistic relationship). For example, between days 10 and 20 of rehabilitation, if the patient is in the epidermal coverage stage of the skin healing period and joint training has entered the active assisted activity stage, the conflict intensity value corresponding to this time interval might be 2, indicating a need to moderately control the intensity of joint training. This mapping table can be implemented using a two-dimensional array structure, facilitating rapid retrieval of conflict risks in subsequent operations.

[0026] The target conflict mapping matrix serves as the reference basis for subsequent intervention configuration and state identification, dynamically matching it with the individual patient's recovery progress. For example, when the actual wound closure rate reaches 85%, the tissue tension index is 3.2 N / cm, and the joint range of motion is 20 degrees, the system will compare the corresponding target stage combination and identify the moderate-intensity conflict relationship that may correspond to this state based on the mapping table. At this time, it will prompt medical staff to reduce the current joint training intensity or delay the training intervention to avoid interfering with the skin healing process. In addition, the mapping table has the characteristic of continuous updating. When unexpected conflict events occur in actual application (such as wound bleeding during training), the corresponding time point and target stage can be traced back to correct the conflict intensity value in the mapping table, thereby improving the accuracy of subsequent predictions and individual adaptability.

[0027] The purpose of this step is to provide a scientific, systematic, and quantifiable modeling foundation for potential conflicts between multiple interventions throughout the entire rehabilitation process of burn patients. Burn rehabilitation involves multiple physiological goals, including skin healing, joint function recovery, and nerve regeneration. These goals exhibit significant physiological dependencies and potential interference relationships at different rehabilitation stages. For example, excessive joint traction training during the unclosed wound stage may lead to wound tearing, bleeding, or even infection; similarly, frequent mechanical stimulation in the early stages of nerve regeneration may interfere with nerve growth pathways, resulting in delayed nerve function recovery. Therefore, before conducting concurrent multi-goal interventions, a structured model that clearly defines the interaction patterns between these goals must be established. This step extracts the evolutionary patterns of each rehabilitation goal in stages, identifies key physiological parameters and external intervention sensitivities, and constructs a conflict mapping table between goals by combining clinical pathways and patient data. This enables the prediction and quantitative expression of conflicts between multi-goal intervention behaviors. This mapping not only clarifies the sensitive windows of each intervention in the temporal dimension but also reveals the potential organizational risks generated by intervention behaviors in the spatial and mechanical dimensions. The execution of this step provides a foundation and core input for subsequent system intervention priority decisions, differentiated intervention intensity configuration, risk node identification, and closed-loop control, ensuring that the entire rehabilitation process can proceed safely, collaboratively, and efficiently, fundamentally solving the problem of imbalance in inter-goal coordination that is common in existing digital rehabilitation systems.

[0028] The multi-source state perception module embeds the target conflict mapping matrix into the data processing channel, collects wound healing progress data, tissue tension change data and joint activity load data of patients in each rehabilitation cycle, and constructs a multi-dimensional state vector set that integrates the time dimension and the structural stress dimension. To achieve dynamic tracking and risk assessment of the physiological state of burn patients during rehabilitation, a target conflict mapping matrix is ​​embedded into the rehabilitation data acquisition and processing workflow. A multi-dimensional state vector set, integrating time and structural stress dimensions, is constructed based on the patient's periodic recovery data to reflect the relationship between the current rehabilitation state and potential conflicts. This implementation includes the following steps:

[0029] The target conflict mapping matrix constructed in the previous stage will be embedded into the rehabilitation data processing workflow as a reference. This matrix, with time on the horizontal axis and the combination of rehabilitation goals on the vertical axis, indicates whether there is a conflict relationship and the degree of conflict between skin healing goals, joint flexibility recovery goals, and nerve regeneration goals within a specific time period. Each matrix cell records a specific conflict level value, taking values ​​of 0, 1, 2, or 3, where 0 represents no conflict, 1 represents mild conflict, 2 represents moderate conflict, and 3 represents severe conflict. This matrix is ​​stored in a structured table format and loaded into the preprocessing stage of the rehabilitation data processing workflow. During subsequent data analysis, it is used to determine whether the current time period belongs to a conflict-sensitive period between goals.

[0030] During the patient's actual rehabilitation process, several key physiological and mechanical data points were collected at the current rehabilitation stage. The collected data included the following three specific categories:

[0031] (1) Wound healing progress data, including the ratio of the closed area of ​​the wound to the initial total area of ​​the wound, wound images were recorded using standard medical image acquisition equipment, and the closed area was measured by image processing methods; the color change of the wound was assessed and recorded by clinicians by setting a color grading standard manual (e.g., red indicates new granulation tissue, yellow indicates exudate adhesion, and black indicates eschar area); the local temperature was measured by infrared thermometer to measure the surface temperature of different areas of the wound and to form a thermal distribution map; (2) Tissue tension change data, including the force value during skin stretching, were collected in real time using a flexible strain sensor attached to the periphery of the wound; the tension tolerance threshold was determined by clinical manual testing, and the maximum force value was recorded before the patient experienced significant discomfort or the wound bled. (3) Joint activity load data, including the maximum flexion angle of each joint during daily training, continuous training time, number of repetitions per training session, exercise frequency, and patient-reported pain score. All exercise parameters are collected through wearable inertial sensor devices and entered into the data terminal by medical staff after each training session. The pain score is filled in and confirmed using a 0 to 10 scale.

[0032] After collecting the above three types of data, the collected data were standardized using a unified time standard to ensure that all data were aligned based on the same time point, and a state vector at the current time point was constructed accordingly. The components of this state vector are fixed as follows: wound closure rate (percentage), area of ​​the red area of ​​the wound (in square centimeters), maximum wound temperature (in degrees Celsius), current skin tension (in Newtons per square centimeter), maximum joint flexion angle (in degrees), continuous training time (in minutes), number of training repetitions (in times), and subjective pain score (0-10 points). For example, at 8:00 AM on the 14th day of rehabilitation, if a patient's wound closure rate is 72%, the area of ​​the red region is 5.3 cm², the highest wound temperature is 36.8℃, the tension is 3.2 N / cm², the maximum knee flexion angle is 38 degrees, the training duration is 12 minutes, the number of repetitions is 18, and the pain score is 4, then the state vector at this time point would be expressed as: [72, 5.3, 36.8, 3.2, 38, 12, 18, 4]. Each dimension of data must undergo anomaly removal and missing data completion before generation to ensure the stability and accuracy of the state vector in subsequent processing.

[0033] The constructed current-time state vector is compared with the aforementioned target conflict mapping matrix to determine whether the current rehabilitation stage falls within a high-incidence range of conflict for a specific combination of targets. For example, if a patient is in the mid-stage of skin healing, and the tension level and joint range of motion in the state vector are significantly increased, while the matrix indicates a "moderate conflict" between skin and joint targets during this time period, subsequent intervention recommendations will adjust the intensity of joint training or postpone the training time to avoid excessive skin tension leading to wound tearing. The state vector is updated each time it is collected, and continuous state vectors form a rehabilitation trajectory curve for continuous monitoring of the evolution trend of rehabilitation status. By dynamically fusing structured state data with the conflict relationship between targets, real-time state expression tailored to individual patients is achieved, providing an operable, traceable, and feedback-enabled data foundation for subsequent risk identification, conflict index construction, and intervention regulation.

[0034] The main purpose of this step is to deeply integrate the previously established goal conflict mapping relationship with the patient's actual rehabilitation status data, constructing a multi-dimensional state expression that comprehensively, in real-time, and structurally reflects the patient's rehabilitation process. This provides a data foundation and real-time basis for subsequent conflict identification, risk prediction, and intervention decisions. During burn rehabilitation, the patient's physiological state constantly changes, involving multiple aspects such as wound healing progress, skin tension fluctuations, joint range of motion, tissue tolerance, and neurological function response. Each indicator not only reflects a single rehabilitation dimension but may also interact with other rehabilitation goals, creating potential intervention conflicts. Traditional rehabilitation monitoring methods often record isolated indicators, failing to effectively integrate multi-source information and match it with the intervention relationship between actual goals. This can easily lead to information fragmentation, delayed risk identification, and even inappropriate intervention risks. This step, by uniformly collecting key physiological and functional data of the patient in each rehabilitation cycle, clarifies the data collection methods, processing standards, and time synchronization mechanisms, integrating all information into a state vector with time and structural stress dimensions to form a complete and traceable rehabilitation status expression model. This model not only possesses the ability to connect multiple indicators in parallel horizontally, but also the ability to track temporal evolution vertically. It can pinpoint the patient's specific location within the target conflict mapping matrix based on their actual state, thereby determining whether there is an intervention conflict between multiple objectives. Through this mechanism, the rehabilitation process no longer relies on human experience-based judgment, but instead achieves an objective quantitative expression of the rehabilitation status and a risk structure mapping. This provides a solid foundation for subsequent conflict index analysis, priority control, and feedback adjustment, significantly enhancing the intelligence, adaptability, and personalization capabilities of the entire digital twin rehabilitation model.

[0035] The conflict identification and calculation module constructs a conflict index calculation mechanism based on the state vector set. It combines the direction and rate of change of state indicators between targets with the preset coupling sensitivity factor to generate a conflict index curve that reflects the intensity of interference between targets. By comparing it with the risk threshold model, it identifies and marks intervention nodes with potential conflicts. To quantitatively identify potential conflicts between skin healing goals, joint flexibility recovery goals, and nerve regeneration goals during the rehabilitation process of burn patients, a conflict index calculation mechanism is constructed based on continuous rehabilitation state data in a state vector set. A risk threshold is introduced to discriminate conflict outcomes, thereby accurately marking rehabilitation time points where potential goal interference may exist. This process mainly includes the following steps:

[0036] Key indicators for interference assessment were extracted. Specifically, in the state vector set constructed in the previous stage, each vector represented a point in time during the rehabilitation process, containing multiple defined dimensions such as wound closure rate (percentage), skin tension (Newtons per square centimeter), joint range of motion (degrees), neurological reflex score (out of 10), active training duration (minutes), and subjective pain score (out of 10). To identify the trends of these indicators over time, a complete set of state vector data was collected every 24 hours, comparing the direction (increase or decrease), magnitude (e.g., joint range of motion increased by 15 degrees), and rate of change per unit time (e.g., skin tension increased by 0.5 N / cm² / day) of each dimension between two consecutive time points. The key to this stage was identifying which indicators were changing rapidly and the potential tissue stress concentration or physiological system load changes caused by these trends.

[0037] An influence weight, or coupling sensitivity factor, is assigned based on the interference characteristics between each pair of rehabilitation goals. This factor is derived from a quantitative assessment of the physiological dependencies between goals and is determined through a combination of expert evaluation and clinical retrospective analysis. For example, if a patient undergoes large-angle joint traction training before the wound is fully closed, it is highly likely to cause skin tearing. In this case, the interference relationship between the skin healing goal and the joint flexibility goal is defined as "highly sensitive," and the coupling sensitivity factor is set to 0.9. If the patient is in the axonal growth phase and has a high tolerance for changes in muscle load, the coupling factor between the neural goal and the joint goal can be set to 0.4. Each coupling relationship is stored in tabular form during implementation, listing all goal combinations and their corresponding sensitivity factors. At the time point when the state vector changes, the change in each dimension index is multiplied by its corresponding coupling factor in a weighted manner to obtain the conflict influence value of each goal pair at that time point. By accumulating the influence values ​​of all goal pairs, a total conflict index is formed at that time point. This index is used to represent the overall interference intensity between multi-goal interventions in the current state.

[0038] The conflict index values ​​calculated chronologically are arranged sequentially to form a conflict index change curve. This curve plots the number of rehabilitation days on the x-axis and the conflict index value on the y-axis. For example, the conflict index is 0.25 on day 10, 0.35 on day 11, surges to 0.68 on day 12, and reaches 0.72 on day 13, indicating a continuous increase in the risk of interference between interventions and objectives during this stage. To facilitate the identification of high-risk states, this curve is compared with a preset risk threshold standard. The risk thresholds are established based on a large sample of historical rehabilitation data and include a low-risk threshold (below 0.3), a medium-risk threshold (0.3 to 0.6), and a high-risk threshold (above 0.6). This grading standard was developed by a medical team with rehabilitation expertise and has undergone multiple clinical validations in actual rehabilitation assessments.

[0039] A potential conflict intervention node is identified when the value of the conflict index curve exceeds the medium- or high-risk threshold at a certain point in time. When marking this node, the specific performance of each indicator in the state vector is also considered. For example, if the conflict index reaches 0.72 on day 15, and the state vector at that moment shows skin tension increasing from 3.0 N / cm² to 3.6 N / cm², joint range of motion increasing from 30 degrees to 45 degrees, and subjective pain score increasing from 2 to 5, it indicates a significant increase in mechanical stretching and subjective discomfort for the patient, suggesting an interference trend between the skin and joint targets. Once identified, this node is recorded in the intervention log with a precise time stamp (e.g., "Day 15, 08:00") for subsequent personalized training intensity adjustments, suspension of training plans, or introduction of other auxiliary treatment strategies. Multiple consecutive conflict nodes will form a "high-risk zone," alerting the clinical team to strengthen real-time monitoring and multi-target balance control during this period.

[0040] This step serves as a crucial technical support for the real-time identification, quantitative assessment, and risk node labeling of potential intervention conflicts among multiple rehabilitation goals in the full-cycle rehabilitation management of burn patients. It is an important intermediate link in achieving dynamic and collaborative control of multiple goals. In actual rehabilitation, skin healing, joint function recovery, and nerve regeneration often occur at different physiological stages, and their corresponding intervention methods and stimulation methods also differ and have potential mutual influences. For example, performing large-scale joint traction exercises before the wound is completely closed may lead to excessive skin tension and cause wound dehiscence; while frequent or high-load exercise stimulation in the early stages of nerve regeneration may negatively affect the still unstable neural structures. These potential risks cannot be predicted simply through static monitoring or empirical judgment; therefore, an operational and quantifiable mechanism is needed to systematically analyze the current intervention status.

[0041] This step extracts and compares key indicator changes in patients' continuous rehabilitation data to establish the correlation between the direction, magnitude, and rate of change of these indicators. Combined with the coupling sensitivity factor between objectives, a conflict index model is constructed to quantify the comprehensive intervention risk at each rehabilitation time point. Compared to traditional methods, this mechanism introduces a unified indicator, the "conflict index," which simplifies the complex interactions between multiple objectives into a single numerical curve, thereby dynamically tracking the risk evolution trend. Simultaneously, using preset multi-level risk threshold standards, the system can automatically identify which time periods are medium- to high-risk intervals and promptly mark "conflict nodes" as a basis for intervention optimization and adjustment.

[0042] This mechanism not only improves the efficiency and accuracy of risk identification during the rehabilitation process, but also lays a crucial foundation for subsequent intervention priority setting, differentiated intensity adjustment, and closed-loop feedback regulation, serving as a bridge to achieve intelligent rehabilitation control. In this process, rehabilitation pathways no longer rely on "general templates" based on human experience, but instead achieve refined, phased adjustments according to the individual patient's physiological response, significantly improving the scientific rigor, safety, and personalization of rehabilitation interventions.

[0043] The intervention priority control module, for the identified conflict nodes, loads the target intervention priority decision logic based on the physiological dependence between skin healing goals, joint flexibility recovery goals and nerve regeneration goals, as well as the intensity of reverse intervention interference, and dynamically adjusts the intervention intensity, intervention sequence and intervention time window length of each rehabilitation goal to complete the differentiated intervention weight configuration. To address potential conflicts among multiple intervention objectives in burn rehabilitation and ensure the effective advancement of each rehabilitation objective under reasonable physiological conditions, after identifying conflict points, a target intervention priority judgment mechanism is implemented based on the physiological dependencies and the intensity of adverse intervention interference between skin healing, joint flexibility restoration, and nerve regeneration objectives. This mechanism adjusts the intervention intensity, execution order, and time window length for each objective accordingly, thus achieving differentiated intervention weight allocation. This process includes the following steps:

[0044] At conflict points, specific goal combinations causing conflict are identified, and their physiological dependencies are extracted. Physiological dependency refers to whether effective intervention for one rehabilitation goal depends on the tissue stability or physiological maturity of another goal. For example, in the pre-closure stage of a wound, the skin tissue has not fully recovered its barrier function. At this time, passive traction training of the joints can easily lead to tissue tension concentration, causing wound dehiscence. Therefore, in this situation, skin healing goals physiologically take precedence over joint function goals. To ensure consistency, physiological dependency is standardized into a three-level evaluation: strong dependency (1.0), moderate dependency (0.7), and weak dependency (0.4). A fixed reference table is developed using literature data, empirical rules, and expert interviews, and the corresponding values ​​are applied at each conflict point.

[0045] For each pair of conflicting targets, the intensity of adverse intervention interference is further analyzed, i.e., the degree to which the intervention behavior of the current target has a negative impact on other targets. For example, when performing joint flexion and extension training, if skin tension is detected to increase from 2.5 N / cm² to 3.9 N / cm², accompanied by a decrease in wound closure rate from 74% to 71%, it indicates that joint training has a significant adverse effect on skin recovery, and the interference intensity is defined as high, assigned a value of 0.9. Conversely, if there is no significant decrease in the neurological function reflex score, it indicates that the training has a low interference on the neurological target, assigned a value of 0.3. All interference intensity data are organized into a two-dimensional matrix table in the form of target pairs, with rows representing the target to which intervention is applied, columns representing the affected target, and cells recording the interference intensity for easy retrieval in subsequent priority calculations.

[0046] Based on physiological dependence and interference intensity, intervention priority decisions are made. This decision-making process follows the logic of prioritizing the stability of highly dependent goals and suppressing interventions for goals with high interference intensity. In practice, if a goal has a "strong dependence" characteristic and is simultaneously affected by a "high interference" from another goal, interventions for the former must be implemented first, and interventions for the latter postponed. For example, in one patient, a high-intensity conflict was identified between skin healing and joint mobility goals between days 16 and 18. The skin goal was highly dependent, while the joint goal was highly interfering. In this case, joint training should be postponed for 48 hours, the traction angle reduced from the planned 45 degrees to 30 degrees, the training frequency reduced from twice daily to once daily, and the duration adjusted from 15 minutes to 8 minutes. These adjustments are clearly recorded in a table: intervention goal, original planned parameters, adjusted parameters, reasons for adjustment, and expected recovery milestones, ensuring the feasibility and trackability of the intervention plan.

[0047] Based on the above adjustments, differentiated intervention weight configurations are completed, clarifying the intervention priority, intensity level, and time window for each target within the specified time period. Intervention priorities are graded into "Level 1, Level 2, and Level 3," with Level 1 targets receiving full execution rights at conflict points, Level 2 targets allowing for moderate intervention adjustments, and Level 3 targets being temporarily deferred. Intervention intensity is categorized into "High, Medium, and Low" based on training intensity levels, specifically quantified by movement amplitude, repetition count, and workload per unit time. The intervention time window is represented by three sets of parameters: "Recommended Start Time," "Maximum Delay Time," and "Estimated Restart Time." For example, for a skin target (Level 1), with a medium intervention intensity, the recommended start time is 8:00 AM on the 16th, and the estimated completion time is 6:00 PM on the 18th; for a joint target (Level 3), with a low intervention intensity, training is postponed until the 19th. This configuration table allows for refined scheduling and personalized control of intervention tasks for each target.

[0048] The purpose of this step is to dynamically and differentiate the intervention strategies for each rehabilitation goal when significant conflicts are detected between different rehabilitation goals during multi-goal parallel rehabilitation. This is based on the physiological dependencies and mutual interference levels between skin healing, joint flexibility recovery, and nerve regeneration, ensuring the entire rehabilitation intervention process proceeds stably under safe, orderly, and individualized conditions. Burn rehabilitation is characterized by significant multi-stage and multi-goal features, with physiological dependencies and physical conflicts between different rehabilitation goals. For example, premature traction training before the skin wound is fully closed may cause wound tearing or bleeding; similarly, high-intensity mechanical stimulation during the early stages of nerve regeneration may disrupt regeneration pathways and delay functional recovery. Traditional rehabilitation management often fails to identify and dynamically adjust these conflicts between multiple goals in real time, easily leading to one-size-fits-all intervention plans, over-reliance on templates, and neglect of individual physiological differences.

[0049] This step, by constructing a clear "physiological dependency matrix" and "reverse intervention interference matrix," achieves quantitative identification of the relationships between conflicting targets. Based on this, intervention priority decision logic is loaded to specifically control whether, how, and when interventions for each target are executed. By dynamically adjusting parameters such as intervention intensity (e.g., reducing stretching angle), sequence (e.g., postponing neural training), and time window (e.g., setting a maximum delay time), it provides systematic and structured technical support for intervention optimization in conflict scenarios. This mechanism not only significantly reduces the medical risks caused by unreasonable concurrent interventions but also provides a traceable and executable basic strategy for subsequent feedback adjustments and closed-loop control. It is an important component of achieving a personalized intelligent rehabilitation pathway for burn rehabilitation "centered on the patient's physiological state."

[0050] The feedback calibration analysis module collects response data for each rehabilitation goal after completing the differentiated intervention configuration, performs time-series retrospective and differential analysis, evaluates the state deviation trend after multi-goal intervention, and updates the goal conflict mapping matrix based on the deviation direction and degree. To achieve closed-loop evaluation of the effectiveness of multi-objective rehabilitation intervention configurations and further correct the conflict relationships between different rehabilitation objectives within a specific time period, a dynamic updating method for the mapping matrix based on response data differential analysis and temporal backtracking is proposed. After completing an intervention configuration, this method collects the actual response status of each rehabilitation objective of the patient, determines whether the current intervention has caused a shift in the objective status, and uses the analysis results to update the objective conflict mapping matrix to enhance the model's adaptability to the real rehabilitation process. This implementation includes the following steps:

[0051] Based on the completion point of the rehabilitation cycle set in the differentiated intervention plan, comprehensive data collection was conducted on the status responses to skin healing goals, joint flexibility recovery goals, and nerve regeneration goals. The data collection specifically included the following three aspects:

[0052] (1) Skin healing target: collect wound closure rate, take pictures of the wound using standardized image acquisition equipment and obtain the closure area by combining the wound image recognition model; measure skin tension, collect pressure data per unit area in the traction area using an implantable flexible tension sensor; extract the proportion of red tissue, and statistically analyze the proportion of pixels occupied by the red area according to the image color level partition to reflect the degree of wound activity. (2) Joint flexibility target: The maximum flexion and extension angle of the joint is collected by a wearable angle sensor; the number of repetitions and total training time of a single training session are obtained by a motion recorder; the patient's subjective pain experience is marked on a scale of 0-10. (3) Nerve regeneration goals: Collect superficial tactile sensitivity response time, and record the patient's first perception time using cotton ball stimulation; record nerve conduction velocity, and use electromyography to detect the signal transmission rate in the conduction path; assess active motor response score, and quantify the amplitude and stability of movement using a movement quality standard scale. All data are collected by designated personnel, with a collection cycle of once every 24 hours, and must be performed continuously for no less than 3 cycles.

[0053] Based on the response data of each rehabilitation goal at different time points, time-series backtesting and differential analysis are performed. The specific procedure is as follows: the latest value of each goal at the end of the current cycle is directly compared with the value of the previous cycle to calculate the difference for each indicator, and the positive or negative direction and magnitude of the difference are determined. For example, if the wound closure rate increases from 76% on day 15 to 78% on day 16, it is a positive shift; if skin tension increases from 3.2 N / cm² to 3.8 N / cm², it is a negative shift, suggesting possible excessive stress on the tissue; if the joint flexion-extension angle increases from 42 degrees to 50 degrees and the pain score decreases from 5 to 3, it indicates good functional recovery; if the nerve reaction time increases from 1.2 seconds to 1.6 seconds, it is considered an increased risk of nerve function impairment. The results of differential analysis are not directly used to adjust intervention parameters, but rather to determine whether the current conflict mapping relationship is accurate.

[0054] Compare the above offset results with the previously constructed target conflict mapping matrix. If the original target pair conflict level in the mapping matrix is ​​"moderate conflict," but both targets show positive offsets during this round of intervention, it indicates that there is no negative interference between them in the current cycle, and the original setting is too conservative, so the conflict level should be lowered. If one target shifts significantly positively and the other shifts significantly negatively, it indicates that there is substantial interference between the target pair, and the current conflict level may be too low, so the conflict weight should be increased. For example, if from day 16 to day 18, the increase in skin closure rate slows down, tension continues to rise, while joint range of motion improves significantly and pain is reduced, it indicates that the current training has caused too much tension interference to the skin, and the conflict level between the skin and joint targets in this cycle should be increased from "moderate" to "high." Conversely, if both improve simultaneously, the conflict level should be decreased from "moderate" to "mild."

[0055] The adjusted conflict levels are written into the corresponding time period and target combination position in the target conflict mapping matrix and saved as the latest version to guide the differentiated intervention configuration for the next cycle. The update operation is performed according to the following standards: after every three periodic response data collections, an offset judgment and mapping matrix update operation is performed; the matrix update content must include time point markers, target pair numbers, old levels, new levels, a description of the adjustment reason, and the original offset data. All update records are stored in tabular form for subsequent intervention optimization and strategy review.

[0056] The purpose of this step is to establish a dynamic feedback pathway between the implementation effect of multi-objective interventions and the target conflict relationship model. This ensures that the interaction between various rehabilitation goals during burn rehabilitation is no longer a static setting, but can be periodically updated and optimized based on the patient's individual actual response. This guarantees that the rehabilitation intervention path continuously adapts to changes in physiological state, thereby improving the scientific rigor, timeliness, and safety of multi-objective collaborative interventions. Specifically, after completing one round of differentiated intervention configuration, the system collects response data for multiple rehabilitation goals, such as skin healing, joint flexibility, and nerve regeneration. By comparing data at continuous time points, it determines the direction and degree of change in each indicator, such as whether wound closure rate is accelerating or slowing down, whether tissue tension is increasing or decreasing, whether joint range of motion is improving, and whether nerve conduction function is declining. These actual trends represent the real impact of intervention measures on different rehabilitation goals. By performing differential analysis on these changes, it is possible to identify whether a particular goal has experienced a negative shift due to intervention conflict.

[0057] Once a target state deviates from the expected recovery trajectory, the system will retrospectively examine the intervention configuration and conflict mapping relationship within the current cycle to determine if there was a misjudgment in the original conflict assessment. If two targets initially defined as having low conflict exhibit significant negative interference after parallel intervention, the system can automatically adjust the conflict level of that target pair, enhancing the avoidance mechanism for future interventions. Conversely, if the targets do not show significant conflict, or even improve together, their conflict weight can be reduced, allowing for earlier or more intense parallel interventions. This mechanism of reversing the target relationship model based on actual therapeutic effects enables the system to iteratively update with each patient's rehabilitation response, achieving a shift from experience-driven to data-driven approaches, significantly improving the individualized accuracy and long-term synergistic stability of rehabilitation interventions. This step plays a crucial feedback regulation role in the entire digital twin rehabilitation control logic, serving as a fundamental link in achieving closed-loop control, adaptive optimization, and continuous improvement in rehabilitation quality.

[0058] The dynamic closed-loop control module feeds back the updated target conflict mapping matrix and target response offset parameters to the rehabilitation stage multi-objective model, forming a dynamic closed-loop control mechanism that includes state perception, conflict prediction, intervention configuration, and feedback calibration, so as to achieve coordinated and stable advancement of multi-objective concurrent interventions throughout the rehabilitation process of burn patients. To achieve dynamic and coordinated intervention and control among skin healing goals, joint flexibility recovery goals, and nerve regeneration goals during the rehabilitation process of burn patients, after completing the identification of conflict relationships and response offset analysis of the previous cycle, the updated target conflict mapping matrix and response offset data are used as input parameters and fed back to the rehabilitation stage multi-objective model to establish a closed-loop control process with four functions: state perception, conflict prediction, intervention configuration, and feedback calibration. This specific implementation includes the following steps:

[0059] The updated target conflict mapping matrix is ​​imported into the structural hierarchy of the rehabilitation phased multi-target model. This mapping matrix is ​​stored as a three-dimensional array, with each dimension corresponding to a rehabilitation time node, target combination number, and conflict intensity level. The target combination number explicitly indicates a pair of two targets, such as "skin healing—joint mobility" or "joint mobility—nerve regeneration." The conflict intensity level is represented by integers (e.g., 1 for low conflict, 2 for moderate conflict, and 3 for high-intensity conflict). Each conflict level cell includes detailed source information, including the corresponding time period, difference analysis data from the previous period, and a description of the indicators that caused the shift during the intervention, ensuring that the matrix not only reflects the current target relationships but also has traceability. By embedding this matrix into the model's internal control structure, the system can automatically call it before each intervention plan is generated and set the intervention priority and execution order accordingly.

[0060] The response offset parameters of each rehabilitation goal collected in the previous intervention cycle are synchronously input into the model's state-aware structure. These offset parameters, presented as data sets, include: the direction of change in the target indicator (e.g., increase or decrease in closure rate), the value of the indicator change (expressed as a percentage or absolute value), and the rate of change per unit time (e.g., 1.6% change in closure rate every 24 hours), supplemented by human clinical evaluation results (e.g., physician-marked wound redness score, impression of changes in movement angle, and observation of neural responses). All parameters are archived chronologically for the model to identify whether a transition has occurred in the physiological state of the current rehabilitation stage. If the rate of increase in skin closure rate decreases and tension increases for three consecutive cycles, the system will determine that the patient's skin recovery has entered the "tissue stabilization phase" and accordingly correct the physiological stage label of the skin target in the current model.

[0061] Based on the fusion of the two feedback results mentioned above, the differentiated intervention strategy configuration process is re-executed. This process, while invoking the latest conflict mapping level, combines the state stage label to schedule the intervention intensity. For example, in a real-world scenario: when the model identifies neural regeneration as still in the "early regeneration stage," and the conflict level between it and the joint movement target is "high-intensity conflict," the generated intervention suggestion will limit the joint training amplitude to no more than 30 degrees, adjust the training frequency from twice daily to once daily, and require the neurophysiological monitoring equipment to be on; simultaneously, neural training will be set to only be initiated when daily skin tension is below 2.5 N / cm² and the post-activity neural response score is higher than the baseline value. All intervention suggestions clearly include the target number, type of action to be performed, parameter configuration, warning threshold, and data source, and are stored in the plan output table.

[0062] The implemented intervention response is reintegrated into the data collection process of the next intervention cycle, continuously advancing the closed-loop control process consisting of four steps: conflict identification, status analysis, intervention configuration, and response calibration. This closed-loop control process cycles every 72 hours, completing a full multi-objective analysis and optimization update within this cycle. To ensure the consistency and stability of the control results, the system compares the difference between the current intervention plan and the results of the previous cycle at the beginning of each control cycle. If the target deviation remains stable within ±2% for two cycles, it is allowed to enter a low-frequency adjustment mode; if fluctuations occur, it switches to a high-frequency adjustment mode, performing target response analysis daily. This mechanism can automatically adjust the frequency of intervention strategy updates according to the patient's actual recovery progress, enhancing the responsiveness and rhythm adaptability of the control strategy.

[0063] The purpose of this step is to establish a closed-loop control mechanism that runs through the entire burn rehabilitation process and has the ability to perceive, adjust dynamically and optimize feedback in real time. This mechanism enables the coordinated and concurrent intervention of three rehabilitation goals—skin healing, restoration of joint flexibility and nerve regeneration—at different stages of the patient's rehabilitation. It also solves the technical problems in traditional rehabilitation interventions, such as "fixed intervention templates", "uncontrolled conflict of goals" and "inability to adapt to individual differences".

[0064] Specifically, after the multi-target differentiated intervention configuration is completed, the patient's various rehabilitation indicators will exhibit real physiological responses, such as changes in wound closure speed, fluctuations in tension, improvement or deterioration in joint flexibility, and the progress of nerve conduction function recovery. These changes will be fed back as continuous numerical values ​​through data acquisition devices. By analyzing the temporal sequence and evaluating the deviations of these response data, the system can not only determine whether a certain target has achieved the expected recovery trajectory, but also identify whether certain interventions are causing negative effects on other targets. Subsequently, these "target response deviation parameters" and the "target conflict mapping matrix" updated in this cycle through the conflict index calculation mechanism are fed back to the core model.

[0065] The model itself uses "rehabilitation stages" as its timeline, constructing a rehabilitation logic structure that evolves continuously over time and with changing states. By inputting the two feedback variables mentioned above into the model, the system can dynamically identify which rehabilitation stage it is currently in, whether the current conflict between various goals is intensifying or alleviating, and automatically adjust the intervention plan for the next cycle based on this: such as adjusting the order of interventions for each goal (e.g., strengthening skin stability before joint stretching), intensity (reducing the frequency of interventions for high-risk goals), and time window (e.g., postponing neurotraining to the next stage), among other key parameters.

[0066] Each round of this closed-loop process involves completing the entire sequence from "state perception → conflict prediction → difference configuration → response feedback → control correction" again, and then continuing in the next round, demonstrating continuous adaptability and the ability to update at each stage. In this way, the entire rehabilitation process is no longer a static "preset process," but a dynamic system that adjusts in real time and evolves intelligently. Its technical effectiveness lies in significantly improving patient rehabilitation efficiency, reducing the risk of complications, and achieving truly personalized, multi-objective, and comprehensive refined rehabilitation management. This is one of the essential characteristics that distinguishes the "digital twin rehabilitation control technology" of this invention from traditional methods.

[0067] The panoramic intelligent rehabilitation management system for burn patients constructed using the above scheme can achieve precise and coordinated intervention at different stages of multiple rehabilitation goals, such as skin healing, joint flexibility recovery, and nerve regeneration, with significant beneficial effects. This invention breaks through the limitations of traditional rehabilitation programs that prioritize single goals and control fixed processes by introducing a multi-goal coupling model and a time-series conflict mapping mechanism. For the first time, it realizes a complete management chain of "dynamic relationship modeling—real-time conflict identification—differentiated adjustment—effect feedback calibration—model update closed loop" among multiple goals. Especially when dealing with potential intervention conflicts between goals, this system can perceive changes in individual tissue states in real time. Through conflict index calculation and priority decision logic, it effectively avoids the problem of "physiological reverse interference" between intervention measures, thereby significantly improving the safety, individual adaptability, and clinical effectiveness of interventions. Furthermore, the continuous feedback mechanism enables model self-iterative optimization, significantly enhancing the system's adaptability to the diversity of patient states and the complexity of rehabilitation pathways. It truly achieves intelligent, closed-loop, and coordinated control of the entire burn rehabilitation process, improving rehabilitation efficiency while reducing medical risks, and possesses broad clinical application value and promising prospects for promotion.

[0068] The foregoing has only described certain exemplary embodiments of the present invention by way of illustration. Undoubtedly, those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the foregoing drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

Claims

1. A panoramic intelligent rehabilitation management system for burn patients, characterized in that, It includes a multi-objective coupling modeling module, a multi-source state perception module, a conflict identification and calculation module, an intervention priority control module, a feedback calibration and analysis module, and a dynamic closed-loop control module. The multi-objective coupling modeling module constructs a phased multi-objective model of rehabilitation that includes the coupling relationship between skin healing objectives, joint flexibility recovery objectives, and nerve regeneration objectives. It identifies sensitive intervention windows based on the physiological change patterns of each objective during the rehabilitation cycle and generates an objective conflict mapping matrix with time-series attributes. The multi-source state perception module embeds the target conflict mapping matrix into the data processing channel to collect the patient's wound healing progress, tissue tension changes and joint activity load in each rehabilitation cycle, and constructs a state vector set that integrates time and structural stress dimensions. The conflict identification and calculation module constructs a conflict index calculation mechanism based on the state vector set. It combines the direction and rate of change of state indicators with the coupling sensitivity factor to generate a conflict index curve and marks potential conflict nodes by comparing it with the risk threshold. The intervention priority control module loads the target intervention priority decision logic onto the identified conflict nodes. Based on physiological dependence and the intensity of reverse intervention, it dynamically adjusts the intervention intensity, sequence, and time window of each target to complete the differentiated intervention configuration. The feedback calibration analysis module collects target response data after completing the differentiated intervention configuration, performs time-series backtracking and differential analysis, evaluates the state offset trend, and updates the target conflict mapping matrix. The dynamic closed-loop control module feeds back the updated target conflict mapping matrix and target response offset parameters to the multi-objective model, forming a dynamic closed-loop control mechanism that includes state perception, conflict prediction, intervention configuration and feedback calibration, thereby achieving the coordinated advancement of concurrent intervention for multiple objectives.

2. The panoramic intelligent rehabilitation management system for burn patients according to claim 1, characterized in that, The process of constructing a phased, multi-objective model for rehabilitation includes the following steps: Establish a phased distribution model of skin healing goals, joint flexibility recovery goals, and nerve regeneration goals during the rehabilitation cycle, and divide the recovery stages according to the physiological indicators of each stage; Based on the impact of intervention behaviors for each rehabilitation goal within the same time period, determine the intensity of conflict or synergy between goal stages; A target conflict mapping matrix is ​​constructed based on the relationship between the time axis and the target combination, and the conflict intensity value in each time interval is marked. The target conflict mapping matrix is ​​used as the basis for intervention configuration and state matching, and conflict identification and dynamic correction are achieved by combining patient recovery data.

3. The panoramic intelligent rehabilitation management system for burn patients according to claim 1, characterized in that, The process of embedding the target conflict mapping matrix into the data processing channel specifically includes the following steps: The target conflict mapping matrix is ​​loaded into the data preprocessing process to identify the conflict intensity of each rehabilitation goal at different time periods; Collect data on wound healing progress, tissue tension changes, and joint activity load. All collected data were processed using a unified time standardization method to construct a state vector that includes wound closure rate, red area, wound temperature, tension, joint flexion angle, training time, number of training sessions, and pain score. The current state vector is compared with the target conflict mapping matrix to identify whether it is in a high-incidence range of target conflict, and intervention risk warning is given accordingly.

4. The panoramic intelligent rehabilitation management system for burn patients according to claim 3, characterized in that, Based on the comparison between the state vector at the current time point and the target conflict mapping matrix, if it is identified that the current time point is in a high-incidence range where there is a moderate or greater conflict between the skin healing target and the joint flexibility recovery target, the system generates an intervention risk warning message and instructs to adjust the intervention intensity and training time of joint training or to suspend the intervention operation to prevent wound damage caused by excessive tension.

5. The panoramic intelligent rehabilitation management system for burn patients according to claim 1, characterized in that, The process of constructing the conflict index calculation mechanism includes the following steps: Extract the direction of change, amount of change, and rate of change per unit time of each dimension index in the continuous state vector; Based on the interference characteristics between skin healing goals, joint flexibility recovery goals, and nerve regeneration goals, a coupling sensitivity factor was set, and the conflict impact value of each goal pair was calculated by weighting it with the index change. The total conflict index at the current time point is generated by summing the conflict impact values ​​of all target pairs. The conflict index is compared with the risk threshold. If it exceeds the preset threshold, it is marked as a potential conflict intervention node.

6. The panoramic intelligent rehabilitation management system for burn patients according to claim 1, characterized in that, The process of configuring differentiated intervention for identified conflict nodes includes the following steps: Extract the physiological dependencies of conflicting target combinations and assign standardized dependency levels; Analyze the reverse interference intensity of the current intervention behavior on other targets and form an interference matrix; Calculate intervention priorities based on dependency level and interference intensity, and adjust the intervention intensity, execution order, and time parameters of the target. Based on the adjustment results, the intervention priority level, training intensity level, and intervention time window range are set for each objective.

7. The panoramic intelligent rehabilitation management system for burn patients according to claim 6, characterized in that, The process of analyzing the reverse interference intensity of the current intervention behavior on other targets and forming an interference matrix includes the following steps: For each target intervention, key physiological data were collected during the intervention period, including changes in skin tension, wound closure rate, changes in joint range of motion, and changes in nerve reflex scores. Compare the changes in key physiological data before and after the intervention to determine whether the intervention caused significant fluctuations in other target physiological indicators; Based on the direction and magnitude of the indicator changes, an interference intensity level is set, and a numerical weight value is assigned to each level. Using the target to be intervened as the row coordinate and the affected target as the column coordinate, a two-dimensional matrix is ​​constructed and the corresponding interference intensity values ​​are filled in to complete the construction of the interference matrix.

8. The panoramic intelligent rehabilitation management system for burn patients according to claim 1, characterized in that, The process of updating the target conflict mapping matrix includes the following steps: After completing the differentiated intervention configuration, response data for skin healing goals, joint flexibility recovery goals, and nerve regeneration goals were collected; Based on response data from three consecutive rehabilitation cycles, a temporal backtracking and differential analysis of the target state are performed to determine the direction and degree of deviation of the target state. The offset result is compared with the corresponding target combination conflict level in the original target conflict mapping matrix to determine whether the conflict level needs to be adjusted. Record the adjusted conflict level, corresponding time point, target pair number, reason for adjustment, and original offset data into the target conflict mapping matrix and save it as an updated version.

9. A panoramic intelligent rehabilitation management system for burn patients according to claim 8, characterized in that, The process of comparing the offset result with the corresponding target combination conflict level in the original target conflict mapping matrix to determine whether the conflict level needs to be adjusted includes the following steps: Extract the offset direction and offset magnitude of various status indicators of skin healing target, joint flexibility recovery target and nerve regeneration target within the current cycle; Identify the original conflict level of the target combination in the current conflict node and locate its corresponding position in the original target conflict mapping matrix; Based on the offset trend of the two targets within the same period, determine whether there are obvious opposing changes, one positive and one negative. If so, it is judged that there is a significant interference tendency. If both objectives show a positive improvement trend, it is determined that the actual conflict relationship is weaker than the set level, and the conflict level is downgraded. If there is a negative deviation, it is determined whether to upgrade the conflict level based on the degree of deviation and the dependency relationship between the objectives.

10. The panoramic intelligent rehabilitation management system for burn patients according to claim 1, characterized in that, The process of feeding the updated target conflict mapping matrix and target response offset parameters back into the rehabilitation phase multi-objective model includes the following steps: The updated target conflict mapping matrix is ​​imported into the internal control structure of the rehabilitation stage multi-objective model in a three-dimensional array structure to set the intervention priority and execution order. The response offset parameters of each rehabilitation goal in the previous intervention cycle are input into the state-aware structure, and the physiological stage labels of the rehabilitation goals in the current model are corrected accordingly. The intervention strategy configuration process is re-executed based on the updated conflict mapping level and stage label, and the output is an intervention plan containing specific target numbers, intervention parameters and warning thresholds. The intervention response after execution is incorporated into the data collection process of the next cycle, forming a closed-loop control mechanism consisting of conflict identification, status analysis, intervention configuration and response calibration.

Images