Burn scar personalized rehabilitation intervention method based on edge computing

Abstract

The present application relates to the technical field of rehabilitation intervention method, specifically to a burn scar personalized rehabilitation intervention method based on edge computing, comprising the following steps: obtaining skin temperature and tendon compression data of traction path, matching time to identify traction asynchronous section, analyzing release rhythm lag and marking lag path, identifying interference section, reconstructing alternating decreasing release rhythm, pairing scar recovery and rehabilitation stage label, and obtaining individualized traction intervention structure content.In the present application, by synchronously comparing skin temperature and tendon compression changes, identifying rhythm response time difference positioning abnormal section, introducing path level rhythm lag analysis and contralateral interference discrimination, making traction rhythm adjustment based on multi-source physiological behavior correlation, reconstructing the path rhythm with conflict through nonlinear decreasing release, reducing interference risk, pairing the reconstructed rhythm with scar recovery status and rehabilitation stage, making traction control logic dynamically adapt to individual recovery changes, and improving intervention process stability and pertinence.

Description

Technical Field

[0001] This invention relates to the field of rehabilitation intervention methods, and in particular to a personalized rehabilitation intervention method for burn scars based on edge computing. Background Technology

[0002] The field of rehabilitation intervention techniques involves the intervention and treatment of functional impairments caused by factors such as trauma, surgery, and disease. Its core aspects include functional reconstruction, restoration of motor function, maintenance of tissue flexibility, and scar control. This technical field intervenes in the patient's rehabilitation process through physical stimulation, occupational activity training, and exercise guidance. It employs an intervention process based on assessment results, combined with continuous monitoring, to achieve closed-loop regulation of the rehabilitation pathway. It is characterized by clear methodologies and specific intervention procedures. Traditional edge computing-based personalized rehabilitation intervention for burn scars refers to addressing issues such as scar tissue hyperplasia, contracture, and functional impairment that occur after burn wound healing. This involves collecting epidermal thermal distribution data using an infrared thermal imager, joint range of motion data using a flexible strain sensor, and local humidity data using a resistive moisture-sensitive patch. A computing node is positioned near the data acquisition point, and a pre-defined statistical graph matching mechanism is used to classify the patient's condition based on scar morphology classification indicators and a rehabilitation movement standard scoring scale. This generates intervention elements such as rehabilitation movement timing, sequence, traction direction, or stimulation frequency. This approach typically combines local device deployment with real-time data processing to complete the intervention process formulation and instruction issuance.

[0003] Existing traction techniques for burn scar rehabilitation mostly rely on preset rhythms and fixed path control. In actual operation, it is difficult to synchronously reflect the dynamic differences between skin temperature changes and tendon force release. There is a time offset between the traction rhythm and physiological feedback, resulting in delayed traction release or insufficient retraction. Contralateral muscle groups are prone to reverse displacement and abnormal vibration interference during traction. Rhythm conflicts between different traction paths are difficult to identify and adjust in a timely manner. The correlation between rehabilitation intervention parameters and individual scar recovery stages is low, making it difficult to form a stable match, which in turn affects the consistency of traction safety and rehabilitation effect. Summary of the Invention

[0004] To address the technical problems existing in the prior art, embodiments of the present invention provide a personalized rehabilitation intervention method for burn scars based on edge computing; To achieve the above objectives, the present invention adopts the following technical solution: a personalized rehabilitation intervention method for burn scars based on edge computing, comprising the following steps: S1: Obtain skin temperature and tendon compression data in the traction path, match the time at the edge nodes, extract the difference between temperature and compression rhythm, identify the segment where the temperature rises earlier than the compression release, and obtain the traction asynchronous segment identification result. S2: Based on the traction asynchronous section identification results, extract the path number, obtain the corresponding traction rhythm control command, extract the release rate parameter and compare it with the lag benchmark, analyze the release rhythm delay state, and obtain the lag path traction command marking result. S3: Based on the lag path traction command marking results, extract the angular displacement and vibration data of the contralateral muscle group, compare the displacement direction with the traction direction, and obtain the results of the contralateral interference section. S4: Based on the time information of the results of the interference section on the opposite side and the traction command marking results of the lag path, compare the time intersection, identify the overlapping path and modify the release rhythm, replacing it with an alternating decreasing structure, to obtain the content of the reconstructed rhythm path release structure. S5: Based on the reconstructed rhythm path release structure content, extract scar recovery status labels, rehabilitation stage identifiers and response types, pair rhythm structures and labels, analyze the pairing relationship, and obtain individualized traction intervention structure content.

[0005] As a further aspect of the present invention, the traction asynchronous segment identification results include an asynchronous time interval set, a temperature and pressure mismatch segment identifier set, and a rhythm difference feature index set; the lag path traction command marking results include a lag traction path number set, a traction command marker bit set, and a release rate lag label set; the contralateral interference segment results include a interference traction cycle segment set, a reverse angular displacement offset segment identifier set, and a vibration frequency abnormal feature set; the reconstructed rhythm path release structure content includes a reconstructed target path number set, an alternating decreasing interval rhythm parameter set, and a rhythm structure replacement mapping relationship set; and the individualized traction intervention structure content includes an adapted traction path list, a recovery state and rhythm structure pairing table, and an individualized intervention parameter package.

[0006] As a further aspect of the present invention, the edge node refers to a computing node configured to execute real-time data processing and control command generation at the patient device end; The compression-release segment refers to the target time period during which the tendon transitions from a state of compression to a state of release within the tendon traction path.

[0007] As a further aspect of the present invention, the release rhythm delay state is determined to be a release rhythm delay state when the traction release rate is lower than the preset physiological response benchmark, that is, the retraction power of the traction action is insufficient and delayed. The alternating decreasing structure is a non-linear release rhythm pattern that inserts rhythmic steps with alternating strong and weak beats and progressively decreasing amplitudes into the original rhythm.

[0008] As a further aspect of the present invention, the specific steps of S1 are as follows: S101: Acquire skin surface temperature change data in the burn scar traction path collected by the infrared thermal imaging device, and simultaneously acquire tendon compression change data recorded by the pressure detection device. Align the data time in the edge computing node, split the aligned data into equal time periods, extract the temperature change rate and compression change rate within the time period, and obtain the temperature and pressure rhythm change dataset. S102: Based on the temperature and pressure rhythm change dataset, identify the intervals in a continuous period where the rate of temperature change always precedes the rate of compression change, extract the position index of the corresponding period in the time series, and obtain the time segment sequence of the temperature and pressure response being advanced. S103: Based on the time segment sequence of the temperature and pressure response advance, select the continuous time range, locate the corresponding segment in the traction path, mark the segment as temperature and pressure mismatch state, and obtain the traction asynchronous segment identification result.

[0009] As a further aspect of the present invention, the specific steps of S2 are as follows: S201: Based on the time interval in the traction asynchronous section identification result, and retrieving the corresponding traction path number, read the traction rhythm control instruction content associated with the traction path number in the edge computing node to obtain the traction rhythm instruction index. S202: Based on the traction rhythm command index, extract the traction release rate parameter in the traction rhythm control command, compare the traction release rate parameter with the preset traction lag judgment standard, identify the path number where the traction release rhythm exhibits lag characteristics, and obtain the release rhythm lag path number set. S203: Based on the correspondence between the release rhythm delay path number set and the traction path number, the traction path number is associated and limited to obtain the delay path traction command marking result.

[0010] As a further aspect of the present invention, the specific steps of S3 are as follows: S301: Obtain the traction path number and time cycle from the lag path traction instruction marking result, retrieve the data from the contralateral muscle group monitoring device that matches the time cycle, extract the angular displacement trajectory data and vibration change signal within the traction cycle, and obtain the traction cycle behavior dataset. S302: Based on the traction cycle behavior dataset, extract the angular displacement change direction and the main traction direction path, identify the time periods in which the direction changes are in opposite directions, extract the vibration frequency change sequence corresponding to the time period, and obtain the vibration trend mapping set. S303: Based on the vibration trend mapping set, match the angular displacement offset trend and vibration frequency change, locate the interference characteristic segment within the continuous traction cycle, and obtain the result of the interference segment on the opposite side.

[0011] As a further aspect of the present invention, the specific steps of S4 are as follows: S401: Based on the interference time period located in the interference segment results on the opposite side and the corresponding traction path number in the traction command marking results of the lag path, extract the intersection of the interference time period and the traction control cycle under the path number according to the time axis coordinate, extract the time overlap data frame in the intersection, and obtain the rhythm conflict positioning segment. S402: Based on the path number in the rhythm conflict positioning segment, obtain the original traction release rhythm sequence of the path, remove the original rhythm structure that matches the time-overlapping data frame in the sequence, insert an alternating decreasing interval rhythm structure at the corresponding position, and obtain the replaced traction rhythm structure. S403: Based on the replaced traction rhythm structure, update the traction control instruction content dataset corresponding to the path number in the edge node, write the new rhythm structure into the specified position in the original instruction set mapping table, and obtain the reconstructed rhythm path release structure content.

[0012] As a further aspect of the present invention, within the intersection of the interference time period located in the interference section results on the opposite side and the traction control cycle under the path number, the data frame intervals that fall within the frame range corresponding to the time boundary of the traction control cycle are identified as the start time point and the end time point, respectively. During the process of writing the replaced traction rhythm structure to the specified position in the original instruction set mapping table: the corresponding position of the rhythm node in the original control instruction content dataset is located by the path number, and the rhythm structure is replaced by equal-length segment overlay.

[0013] As a further aspect of the present invention, the specific steps of S5 are as follows: S501: Based on the corresponding traction path number in the reconstructed rhythm path release structure, by accessing the rehabilitation management interface, retrieve the individual scar recovery status label, rehabilitation stage identifier and traction response type, verify the time consistency of the label data, and obtain the path status label set. S502: Based on the path state label set, call the rhythm parameter sequence in the reconstructed rhythm path release structure content, match the rhythm parameter with the recovery state label item by item, identify the parameter association relationship in the matching process, and obtain the rhythm state correspondence set; S503: Based on the rhythm state correspondence set, determine the traction path number adaptation state, divide the correspondence range between rhythm parameters and recovery state, and obtain the individualized traction intervention structure content.

[0014] Compared with the prior art, the advantages and positive effects of the present invention are as follows: In this invention, abnormal segments are located by synchronously comparing changes in skin temperature and tendon compression, identifying rhythm response time differences, and introducing path-level rhythm lag analysis and contralateral interference discrimination. This allows traction rhythm adjustment to be based on multi-source physiological behavior correlation. By nonlinearly decreasing the release, conflicting path rhythms are reconstructed, reducing the risk of interference. The reconstructed rhythm is paired with the scar recovery status and rehabilitation stage, enabling the traction control logic to dynamically adapt to individual recovery changes and improve the stability and targeting of the intervention process. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 This is a schematic diagram of the steps of the present invention; Figure 2 This is a detailed schematic diagram of S1 of the present invention; Figure 3 This is a detailed schematic diagram of S2 of the present invention; Figure 4 This is a detailed schematic diagram of S3 of the present invention; Figure 5 This is a detailed schematic diagram of S4 of the present invention; Figure 6 This is a detailed schematic diagram of S5 of the present invention. Detailed Implementation

[0017] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0018] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0019] Please see Figure 1 This invention provides a personalized rehabilitation intervention method for burn scars based on edge computing, comprising the following steps: S1: Acquire skin surface temperature change data in the burn scar traction path collected by the infrared thermal imaging device, and simultaneously retrieve tendon compression change data recorded by the pressure detection device. Perform time matching between skin temperature data and tendon compression data in the edge computing node, extract the difference interval between temperature change rhythm and compression response rhythm within the time period, identify continuous segments where the temperature rise rate is earlier than the tendon compression release response, and mark them as temperature-pressure mismatch state to obtain the traction asynchronous segment identification result. S2: Based on the time interval in the traction asynchronous section identification result, retrieve the traction path number corresponding to the time interval, extract the traction rhythm control instruction content associated with the number in the edge computing node, extract the traction release rate parameter in the traction rhythm control instruction, and compare the traction release rate parameter with the preset traction lag judgment standard. When the comparison result is located as the traction release rhythm delay state, the corresponding traction path number is obtained to mark the lag path traction instruction. S3: Based on the traction path number and time period corresponding to the traction command marking result of the lag path, obtain the angular displacement trajectory data and vibration change signal of the muscle group recorded by the contralateral muscle group monitoring device within the traction cycle, compare the angular displacement change direction with the main traction direction trajectory, extract the corresponding vibration frequency change information in the time period when the angular displacement direction shows a reverse offset trend, and jointly judge the angular displacement offset trend and vibration frequency change to identify the traction cycle segment with interference characteristics and obtain the contralateral interference segment result. S4: Based on the interference time period determined in the results of the interference section on the opposite side, compare the time axis intersection with the traction path number identified in the traction instruction marking results of the delayed path. When a time overlap relationship is detected, modify the traction release rhythm structure under the numbered path, replace the original rhythm mode with an alternating decreasing interval output rhythm, and replace it in the instruction set to obtain the reconstructed rhythm path release structure content. S5: Based on the modified traction path numbers in the reconstructed rhythm path release structure content, extract the individual scar recovery status labels, rehabilitation stage identifiers, and traction response types associated with the path numbers. Pair the label content with the adjusted rhythm structure item by item. Determine the fit status based on the parameter correspondence between the rhythm structure and the recovery status. Label the traction paths where the fit relationship is established to obtain the individualized traction intervention structure content.

[0020] The results of asynchronous traction segment identification include a set of asynchronous time intervals, a set of temperature and pressure mismatch segment identifiers, and a set of rhythm difference characteristic indicators. The results of lagging path traction command marking include a set of lagging traction path numbers, a set of traction command marker bits, and a set of release rate lag labels. The results of contralateral interference segment identification include a set of interference traction cycle segments, a set of reverse angular displacement offset segment identifiers, and a set of vibration frequency anomaly characteristics. The content of reconstructed rhythm path release structure includes a set of reconstructed target path numbers, an alternating decreasing interval rhythm parameter group, and a set of rhythm structure replacement mapping relationships. The content of individualized traction intervention structure includes a list of adapted traction paths, a pairing table of recovery state and rhythm structure, and an individualized intervention parameter package.

[0021] Please see Figure 2 The specific steps of S1 are as follows: S101: Acquire skin surface temperature change data in the burn scar traction path collected by the infrared thermal imaging device, and simultaneously acquire tendon compression change data recorded by the pressure detection device. Align the data time in the edge computing node, split the aligned data into equal time periods, extract the temperature change rate and compression change rate within the time period, and obtain the temperature and pressure rhythm change dataset. First, a high-frequency non-contact scan of the burn scar area is performed using an infrared thermal imaging device. The device's internal infrared focal plane array detector continuously records the thermal radiation distribution on the skin surface at a sampling frequency of 25 Hz. A microbolometer unit converts the far-infrared energy released by the human tissue into electrical signals and samples them in real time. By reconstructing the spatial matrix of the radiation intensity received by different pixel units, a continuous two-dimensional temperature distribution image is formed. The received radiation power signal is then converted into a precise temperature value sequence after analog-to-digital conversion, and subsequently written into a temporary buffer of the edge computing node for further processing. Simultaneously, a high-precision micro-pressure displacement sensor in the compressive stress detection device is invoked. This sensor detects the microscopic deformation of the tendon tissue under stress during traction using a piezoresistive strain gauge structure, and records the axial compression displacement of the tendon under traction at a period of 40 milliseconds, continuously acquiring a data stream of tendon compression changes. After receiving two heterogeneous data streams from the thermal imaging module and the pressure stress detection module, the edge computing node first extracts the microsecond-level timestamps from each data frame. It then performs time reference alignment using a unified clock synchronization mechanism, taking the initial trigger pulse of the infrared thermal imaging device as the global time reference. For pressure stress detection data with low sampling frequencies or transmission jitter, it performs linear interpolation compensation and time resampling to ensure complete overlap of the two data sequences on the same time axis. Subsequently, the aligned long sequence data is structurally split according to an equal-length 2-second interval. Within each segmented interval, first-order differential logic is used to calculate the temperature change rate. This involves taking the difference between the temperature value of the last frame and the temperature value of the first frame within the 2-second interval and dividing it by a fixed time interval of 2 seconds to obtain the temperature change rate. Similarly, the difference between the displacement value at the end and the displacement value at the beginning of the interval is divided by the time interval to obtain the compression change rate. For example, if the initial temperature is 33.20 degrees Celsius and the ending temperature is 33.80 degrees Celsius, the rate of temperature change is 0.30 degrees Celsius per second; if the corresponding tendon displacement increases from 1.50 mm to 2.10 mm, the rate of compression change is 0.30 mm per second. By performing segment-by-segment feature pairing mapping between the rate of temperature change and the rate of compression change for each time period, a temperature and pressure rhythm change dataset reflecting the coupling relationship between local skin thermodynamics and deep tissue mechanical response is finally constructed.

[0022] S102: Based on the temperature and pressure rhythm change dataset, identify the intervals in a continuous period where the rate of temperature change always precedes the rate of compression change, extract the position index of the corresponding period in the time series, and obtain the time segment sequence of temperature and pressure response advance. First, the temperature and pressure rhythm change dataset stored in the memory image of the edge computing node is invoked, and a sequence phase difference identification program is executed to scan and analyze all data from equal time periods one by one. During this process, the slope inflection points in the temperature change rate sequence are retrieved first. By calculating local gradient changes and slope extrema, the critical moments when the temperature change rate reaches its peak or exceeds a preset slope threshold are located, and their corresponding millisecond-level time coordinates are precisely recorded. Simultaneously, the corresponding inflection point coordinates in the compression change rate sequence are retrieved within the same time period. By comparing the extreme positions in the displacement change rate curve, the key time points of the deep tissue mechanical response are determined. Then, the difference between the inflection points of the temperature change rate and the compression change rate is calculated to obtain the temperature and pressure response time difference between the two physiological responses. A lead judgment benchmark of 50 milliseconds is set. When the temperature and pressure response time difference is greater than 50 milliseconds and is positive, it indicates that the surface thermal change significantly precedes the deep compression change in time, and the period is determined to have a leading thermal response characteristic. The sliding window traversal logic is then executed to continuously scan the time series and check if there are six or more consecutive time periods that consistently meet the leading characteristic condition, i.e., the duration is not less than 12 seconds. If the condition is met, the consecutive time periods are identified as stable intervals where the rate of temperature change consistently precedes the rate of compression change, and the start and end frame position indices of these consecutive intervals in the total time series are extracted. For example, if it is detected that the peak time of temperature change between frames 500 and 800 consistently precedes the peak time of displacement change by approximately 80 milliseconds, then the numerical interval between frames 500 and 800 is recorded as a valid segment, and all detected index intervals are arranged in chronological order to form a complete sequence of temperature and pressure response advance time segments. Arranging all index intervals in chronological order yields the time segment sequence of temperature and pressure response advance.

[0023] S103: Based on the time segment sequence of temperature and pressure response advance, select the continuous time range, locate the corresponding segment in the traction path, mark the segment as temperature and pressure mismatch state, and obtain the traction asynchronous segment identification result; First, the index values ​​of each group in the time segment sequence of the advance temperature and pressure response are traversed, and temporal connectivity analysis is performed on these candidate segments. During this process, the interval duration between two adjacent time segments is calculated one by one, i.e., the temporal continuity is determined by comparing the millisecond-level time difference between the end index of the preceding segment and the start index of the following segment. This interval duration is then compared with a preset logical merging threshold of 200 milliseconds. When the interval is less than or equal to 200 milliseconds, the two time segments are considered to have continuous characteristics in the evolution of the physiological response, indicating that they may belong to the same continuous tissue response process. Therefore, the segment merging operation is automatically performed to integrate them into a longer continuous time range. After segment merging, the total duration of each continuous time range is further calculated, and reliability screening is performed. Short segments with a total duration of less than 1.5 seconds are easily affected by measurement noise, environmental interference, or random fluctuations due to their short duration and are therefore judged as instantaneous measurement noise and discarded. Only time ranges with longer durations and stable characteristics are retained as valid segments. Subsequently, using these filtered time ranges as indexes, the path space mapping coordinate table recorded inside the traction device is retrieved, and the physical space coordinate intervals corresponding to these time ranges are precisely located in the three-dimensional geometric model of the traction path. By jointly mapping time information with spatial coordinates, specific spatial regions where the thermodynamic response of the skin surface significantly precedes the displacement response of deep tissues during traction can be identified. These regions are formally labeled as thermo-pressure mismatch states, and their starting and ending physical positions are recorded, ultimately yielding complete identification results of asynchronous traction segments.

[0024] Please see Figure 3 The specific steps of S2 are as follows: S201: Based on the time interval in the traction asynchronous section identification result and the corresponding traction path number, read the traction rhythm control instruction content associated with the traction path number in the edge computing node to obtain the traction rhythm instruction index. First, the mismatch time intervals and their corresponding spatial coordinates recorded in the traction asynchronous segment identification results are read, and these time intervals are used as index keys to perform a precise retrieval operation in the real-time database of the edge computing node. By matching timestamps with equipment operation log records, the traction path number being executed within that time interval is located. Subsequently, by accessing the instruction stack structure of the traction control layer, all traction rhythm control instructions associated with the corresponding traction path number within that mismatch time interval are read. These instructions are mainly generated by the underlying drive control module, and their contents include key control variables such as the stepper motor's stepping pulse frequency parameters, the torque setting value of the drive current, and the dynamic acceleration and deceleration control slope. The extracted original control instructions are sorted in chronological order, and a complete instruction sequence snapshot is constructed. This snapshot highly overlaps with the detected asynchronous segment on the time coordinate, thus fully reconstructing the mechanical drive logic background of the equipment when the temperature and pressure mismatch occurs. For example, in a certain analysis, it was found that the drive frequency of path number 5 during the mismatch period gradually decreased from 250 Hz to 10 Hz every 50 milliseconds. This sequence can be used to accurately describe the trajectory of the traction rhythm change, thereby forming the corresponding traction rhythm command index.

[0025] S202: Based on the traction rhythm command index, extract the traction release rate parameter in the traction rhythm control command, compare the traction release rate parameter with the preset traction lag judgment standard, identify the path number of the traction release rhythm showing lag characteristics, and obtain the release rhythm lag path number set. First, the control parameters encapsulated in the traction rhythm command index sequence are analyzed, with a focus on extracting the traction release rate parameter. This parameter describes the force attenuation or displacement reduction per unit time at the traction end during the retraction release phase, and is a crucial dynamic indicator reflecting the mechanical release rhythm. A preset traction hysteresis judgment standard is invoked. This standard, based on the elastic recovery characteristics of biological tissues and the tendon's rebound limit, sets the absolute value of the release rate slope to be no less than 0.60 Newtons per second or its corresponding displacement change rate. Subsequently, a comparison operation logic is executed, calculating the difference between the real-time extracted traction release rate slope from the command sequence and this standard value, and comparing it with a threshold. When the real-time slope absolute value is less than 0.60, it indicates insufficient dynamic response speed of the machinery during the release phase, failing to release the accumulated stress within the tissue in a timely manner. Therefore, it is determined that the mechanical release rhythm of this path exhibits a significant dynamic hysteresis phenomenon. A full traversal check is performed on all participating traction paths, identifying all path numbers exhibiting hysteresis characteristics in their release rhythms. These path numbers are then uniformly summarized into a dynamic storage array. For example, when the actual release slope of path 5 is detected to be 0.45 Newtons per second, it is determined to be a release rhythm delay path because it is lower than the standard value of 0.60, and is recorded in the release rhythm delay path number set.

[0026] S203: Based on the correspondence between the release rhythm delay path number set and the traction path number, the traction path number is associated and limited to obtain the delay path traction command marking result; First, a logical mapping relationship is established between the release rhythm delay path number set and the global traction path configuration table. In the memory scheduling module of the edge computing node, the delay path number set is used as a filtering condition to limit subsequent optimization and compensation processing to only these traction paths with delay issues. For each identified traction path number, a Boolean type hysteresis attribute tag is added to its corresponding instruction control word structure or status register, and the value of the tag is set from the original 0 state to 1 state to clearly identify that the path has rhythm hysteresis characteristics. Through this tag-based association constraint mechanism, the target paths that need compensation and optimization can be quickly screened during subsequent algorithm execution, thereby avoiding unnecessary computational overhead or erroneous interference to normally operating paths and improving overall processing efficiency and stability. After completing the update and consistency verification of all path tag bits, the path number, mismatch start time, mismatch end time, and hysteresis status bit are integrated into a unified data structure and stored in the database in a structured form, finally forming the hysteresis path traction instruction tagging result.

[0027] Please see Figure 4 The specific steps of S3 are as follows: S301: Obtain the traction path number and time cycle from the traction instruction marking result of the lag path, retrieve the data of the contralateral muscle group monitoring device that matches the time cycle, extract the angular displacement trajectory data and vibration change signal within the traction cycle, and obtain the traction cycle behavior dataset. First, the affected traction path numbers and corresponding mismatch time periods are extracted from the traction command marking results of the lagging path. Based on this spatiotemporal information, physiological signal data synchronously collected and transmitted by the contralateral muscle group monitoring device within the same time window are retrieved. This monitoring device is installed symmetrically distributed between the patient's contralateral limb and the main traction area. A nine-axis inertial measurement unit acquires real-time information on the limb's posture changes in three-dimensional space, and a high-sensitivity piezoelectric vibration sensor captures subtle vibration changes in local muscle tissue under traction. Within a specified traction cycle, the angular velocity and acceleration data output by the inertial measurement unit are integrated and attitude-calculated to generate continuous angular displacement trajectory data. This trajectory data accurately reflects the swing amplitude and movement trend of the contralateral limb during traction due to nerve reflexes or muscle compensation. Simultaneously, the synchronously acquired vibration change signals are extracted, and a bandpass filtering algorithm is used to remove low-frequency body motion noise and high-frequency electromagnetic interference, retaining only the effective frequency components between 10 Hz and 60 Hz that reflect involuntary muscle tremors. Subsequently, the angular displacement trajectory data and vibration signal were sequentially spliced ​​together at millisecond-level sampling points, and joint feature extraction processing was performed on the two types of data to construct a comprehensive dataset that can simultaneously reflect the spatial motion behavior and muscle vibration activity of the contralateral limb. Finally, a traction cycle behavior dataset was obtained to describe the intensity and dynamic change characteristics of the compensatory response of the contralateral muscle group.

[0028] S302: Based on the traction periodic behavior dataset, extract the angular displacement change direction and the main traction direction path, identify the time periods in which the direction changes in opposite directions, extract the vibration frequency change sequence corresponding to the time period, and obtain the vibration trend mapping set. First, vector correlation analysis is performed on the multidimensional features of the traction cycle behavior dataset. The vector components of the contralateral limb movement direction in three-dimensional space are extracted from the angular displacement trajectory data and projected onto a unified spatial reference frame through coordinate transformation, thus obtaining a stable motion vector representation. Then, a dot product operation is performed between this motion vector and the pre-set traction direction vector of the main traction device. The relationship between the motion directions is evaluated by calculating the phase angle between the two. When the calculated angle is between 150 and 210 degrees, it indicates a significant counterbalancing relationship between the contralateral limb movement direction and the traction direction. This time period is identified as a potential counterbalancing movement segment. For these specific counterbalancing time periods, a Fast Fourier Transform algorithm is further invoked to perform frequency domain analysis on the original vibration signal within the corresponding time window. The main center frequency and corresponding power spectral density distribution are extracted from the power spectrum of the vibration signal, and a continuous vibration frequency variation sequence is generated accordingly. Subsequently, the extracted peak vibration frequency, average energy level, and corresponding motion direction offset angle are jointly correlated and mapped to construct a feature mapping table with time index as the core. This enables the limb's reverse motion displacement and vibration frequency change pattern to be bound and expressed within the same analytical framework, thereby obtaining a vibration trend mapping set that reflects the dynamic interference behavior characteristics of the contralateral muscle group.

[0029] S303: Based on the vibration trend mapping set, the corresponding angular displacement offset trend and vibration frequency change are matched to locate the interference characteristic segment within the continuous traction cycle and obtain the result of the interference segment on the opposite side. First, interference level assessment is performed on each feature index in the vibration trend mapping set, and a pre-constructed interference identification matrix model is invoked. This matrix uses the angular displacement trend value as the horizontal axis index and the vibration frequency change as the vertical axis index, and comprehensively judges the interference intensity through a two-dimensional feature space. The critical value for angular displacement trend is set to 3.0 degrees, and the critical value for vibration frequency is set to 30 Hz. When the data items collected and calculated in real time exceed the above two critical values ​​simultaneously, it indicates that the contralateral limb not only shows obvious reverse compensatory movement in the spatial direction, but also that the muscle tremor activity is at a high frequency. If this state remains stable or shows a continuous upward trend within three consecutive sampling periods (approximately 120 milliseconds), then this time period is located as the effective interference feature performance segment in the continuous traction cycle. Subsequently, the physical start and end points of these segments are marked on a unified time axis, and spatial positioning is performed in conjunction with the corresponding traction path number information, thereby determining the specific location and timing of the kinetic energy interference of the contralateral muscle group on the main traction path, and finally forming a complete contralateral interference segment identification result.

[0030] Please see Figure 5 The specific steps of S4 are as follows: S401: Based on the interference time period located in the interference section results and the corresponding traction path number in the traction command marking results of the delayed path, extract the intersection of the interference time period and the traction control cycle under the path number according to the time axis coordinate, extract the time overlap data frame in the intersection, and obtain the rhythm conflict location segment. First, the start and end times of the interference recorded in the results of the interference segment on the opposite side are extracted, and the corresponding traction path number in the traction command marking results of the lagging path is read simultaneously. In the time series management module of the edge computing node, the two sets of sequences containing time information are projected onto the global synchronous time coordinate system, and the time axis coordinate intersection operation is performed. By performing interval overlap calculation on the two time series, the overlapping area between the time period of the interference on the opposite side and the control cycle of the traction path is found. For each identified intersection area, the data is further extracted cycle by cycle according to the minimum sampling period of the sensor (about 40 milliseconds). In each sampling period, key data such as the real-time feedback value of traction force, the current encoder position of the motor, and the vibration intensity of the interference on the opposite side are acquired synchronously, and these multi-source information are integrated into a unified data frame. Through this frame-level precision matching mechanism based on the sampling period, the micro-time segment where the physical conflict occurs between the change in traction rhythm and the compensation interference on the opposite side can be accurately located, thereby obtaining the rhythm conflict location segment.

[0031] S402: Based on the path number in the rhythm conflict location segment, obtain the original traction release rhythm sequence of the path, remove the original rhythm structure that matches the time-overlapping data frame in the sequence, insert an alternating decreasing interval rhythm structure at the corresponding position, and obtain the replaced traction rhythm structure. First, based on the traction path number identified in the rhythm conflict positioning segment, the original traction release rhythm sequence for that path is retrieved from the dynamic random access memory. Then, instruction removal processing is performed. The rhythm control words in the original sequence are precisely located according to the time index corresponding to the rhythm conflict positioning segment, and these control words are deleted from the original instruction sequence to cancel the original constant-speed release or continuous linear release control mode. At the time window position corresponding to the deleted instructions, a new alternating decreasing interval rhythm structure is re-inserted. This structure consists of a series of non-equal-interval drive pulse instructions, and its specific execution logic is as follows: First, a long-step release action with an amplitude of 1.2 times the original release amount is executed to quickly release part of the accumulated tension; then, a 20-millisecond zero-output pause interval is immediately set to give the mechanism a brief buffer; next, a short-step release action with an amplitude of 0.8 times the original release amount is executed, and in the subsequent continuous release process, the release amplitude decreases progressively by a ratio of 0.9. This nonlinear, intermittent, alternating release structure effectively breaks the mechanical resonance mode formed by the original continuous rhythm, thereby reducing the transmission and superposition effect of the interference signal on the opposite side in the tissue, and finally obtaining the replaced traction rhythm structure.

[0032] S403: Based on the replaced traction rhythm structure, update the traction control instruction content dataset corresponding to the path number in the edge node, write the new rhythm structure into the specified position in the original instruction set mapping table, and obtain the reconstructed rhythm path release structure content. First, the newly generated, replaced traction rhythm structure is repackaged into a binary control instruction packet conforming to the underlying driver protocol format and written into the data management module of the edge computing node. During the update process, the traction control instruction content dataset corresponding to the path number is modified. By adjusting the memory offset pointer in the instruction set mapping table, the entry address that originally pointed to the old continuous release control logic is redirected to the storage sector where the newly written alternating decreasing interval rhythm structure resides. Then, a Cyclic Redundancy Check (CRC) algorithm is executed to verify the integrity of the written instruction packet, ensuring that no bit-level errors occur during data writing and storage, and guaranteeing strict temporal continuity of the instruction stream during runtime switching. After completing all instruction updates and verifications, a set of release control content with adaptive anti-interference capabilities is generated, thus obtaining the reconstructed rhythm path release structure content.

[0033] Please see Figure 6 The specific steps of S5 are as follows: S501: Based on the corresponding traction path number in the reconstructed rhythm path release structure, by accessing the rehabilitation management interface, retrieve the individual scar recovery status label, rehabilitation stage identifier and traction response type, verify the time consistency of the label data, and obtain the path status label set. First, based on the traction path number identified in the reconstructed rhythm path release structure, the patient's personal digital pathology record is retrieved through the hospital's internal rehabilitation management information interface. Individual scar recovery status tags corresponding to the traction path area are extracted from the record database, including key physiological indicators such as scar thickness measured by ultrasound elastography, local vascular density index, and tissue elastic modulus. Simultaneously, the patient's current rehabilitation stage (e.g., proliferative or maturing stage) and historical traction response type (e.g., hypersensitive or sluggish response) are also retrieved. Then, a time consistency verification logic is executed, comparing the most recent update time of these tags with the current execution time to confirm the data's validity and timeliness. All verified physiological and pathological indicators undergo unified structured integration processing, ensuring that each traction path corresponds to a clear set of biological background attributes, thus obtaining a complete set of path status tags.

[0034] S502: Based on the path state label set, call the rhythm parameter sequence in the reconstructed rhythm path release structure content, match the rhythm parameter with the recovery state label item by item, identify the parameter association relationship during the matching process, and obtain the rhythm state correspondence set; First, the rhythm parameter sequence from the reconstructed rhythm path release structure is retrieved. This sequence includes key parameters such as frequency values, decreasing ratio coefficients, and pause intervals in the alternating release structure. Then, these rhythm parameters are matched item by item with various physiological indicators in the path state label set, and parameter correlation identification is performed. During the analysis, a multivariate Pearson correlation analysis algorithm is used to quantify the correlation between variables such as scar tissue elastic modulus, vascular density, and scar thickness and control parameters such as release frequency and pause intervals. A correlation strength threshold of 0.80 is set; when the absolute value of the calculated correlation coefficient exceeds this threshold, a significant statistical correlation is considered to exist between the rhythm parameter configuration and the corresponding scar recovery state label. All parameter pairs that meet the conditions are recorded, and a decision support library reflecting the mapping law between "pathological features—motor control parameters" is constructed, thus obtaining the rhythm state correspondence set.

[0035] S503: Based on the rhythm state correspondence set, determine the traction path number adaptation state, divide the correspondence range between rhythm parameters and recovery state, and obtain the individualized traction intervention structure content. First, a comprehensive analysis of the feature pairings in the rhythm state correspondence set is performed, and quantitative discrimination processing of the traction path number adaptation status is executed. Based on the correlation coefficient and a pre-set adaptation range, the current traction control scheme is divided into three levels: highly adapted, dynamically optimized, and unsuitable. A correlation coefficient greater than 0.90 is considered highly adapted, indicating a high match between the current control parameters and the scar recovery state; a correlation coefficient between 0.80 and 0.90 is considered dynamically optimized, indicating further optimization potential; and a correlation coefficient below 0.80 is considered unsuitable, requiring readjustment of the rhythm parameters. For path combinations within the highly adapted or dynamically optimized range, specific suggested value ranges for rhythm parameters are further defined, and a correspondence between the recovery state label value range and the rhythm parameters is established. For example, when the elastic modulus of the scar tissue is detected to be in the range of 45 to 55 kPa, a personalized control criterion of a release reduction ratio of 0.85 is automatically established. By integrating the adaptation logic and dynamic adjustment rules of all traction pathways, a complete control scheme is ultimately generated that can be adjusted in a closed loop according to the patient's current recovery progress and real-time physiological feedback, thus forming an individualized traction intervention structure.

[0036] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A personalized rehabilitation intervention method for burn scars based on edge computing, characterized in that, Includes the following steps: S1: Obtain skin temperature and tendon compression data in the traction path, match the time at the edge nodes, extract the difference between temperature and compression rhythm, identify the segment where the temperature rises earlier than the compression release, and obtain the traction asynchronous segment identification result. S2: Based on the traction asynchronous section identification results, extract the path number, obtain the corresponding traction rhythm control command, extract the release rate parameter and compare it with the lag benchmark, analyze the release rhythm delay state, and obtain the lag path traction command marking result. S3: Based on the lag path traction command marking results, extract the angular displacement and vibration data of the contralateral muscle group, compare the displacement direction with the traction direction, and obtain the results of the contralateral interference section. S4: Based on the time information of the results of the interference section on the opposite side and the traction command marking results of the lag path, compare the time intersection, identify the overlapping path and modify the release rhythm, replacing it with an alternating decreasing structure, to obtain the content of the reconstructed rhythm path release structure. S5: Based on the reconstructed rhythm path release structure content, extract scar recovery status labels, rehabilitation stage identifiers and response types, pair rhythm structures and labels, analyze the pairing relationship, and obtain individualized traction intervention structure content.

2. The personalized rehabilitation intervention method for burn scars based on edge computing according to claim 1, characterized in that, The traction asynchronous segment identification results include an asynchronous time interval set, a temperature and pressure mismatch segment identifier set, and a rhythm difference characteristic index set. The lag path traction command marking results include a lag traction path number set, a traction command marker bit set, and a release rate lag label set. The contralateral interference segment results include a interference traction cycle segment set, a reverse angular displacement offset segment identifier set, and a vibration frequency abnormality characteristic set. The reconstructed rhythm path release structure content includes a reconstructed target path number set, an alternating decreasing interval rhythm parameter set, and a rhythm structure replacement mapping relationship set. The individualized traction intervention structure content includes an adapted traction path list, a recovery state and rhythm structure pairing table, and an individualized intervention parameter package.

3. The personalized rehabilitation intervention method for burn scars based on edge computing according to claim 1, characterized in that, The edge node refers to a computing node configured to execute real-time data processing and control command generation at the patient device end; The compression-release segment refers to the target time period during which the tendon transitions from a state of compression to a state of release within the tendon traction path.

4. The personalized rehabilitation intervention method for burn scars based on edge computing according to claim 1, characterized in that, The release rhythm delay state is determined when the traction release rate is lower than the preset physiological response benchmark, that is, the retraction power of the traction action is insufficient and delayed. The alternating decreasing structure is a non-linear release rhythm pattern that inserts rhythmic steps with alternating strong and weak beats and progressively decreasing amplitudes into the original rhythm.

5. The personalized rehabilitation intervention method for burn scars based on edge computing according to claim 1, characterized in that, The specific steps of S1 are as follows: S101: Acquire skin surface temperature change data in the burn scar traction path collected by the infrared thermal imaging device, and simultaneously acquire tendon compression change data recorded by the pressure detection device. Align the data time in the edge computing node, split the aligned data into equal time periods, extract the temperature change rate and compression change rate within the time period, and obtain the temperature and pressure rhythm change dataset. S102: Based on the temperature and pressure rhythm change dataset, identify the intervals in a continuous period where the rate of temperature change always precedes the rate of compression change, extract the position index of the corresponding period in the time series, and obtain the time segment sequence of the temperature and pressure response being advanced. S103: Based on the time segment sequence of the temperature and pressure response advance, select the continuous time range, locate the corresponding segment in the traction path, mark the segment as temperature and pressure mismatch state, and obtain the traction asynchronous segment identification result.

6. The personalized rehabilitation intervention method for burn scars based on edge computing according to claim 1, characterized in that, The specific steps of S2 are as follows: S201: Based on the time interval in the traction asynchronous section identification result, and retrieving the corresponding traction path number, read the traction rhythm control instruction content associated with the traction path number in the edge computing node to obtain the traction rhythm instruction index. S202: Based on the traction rhythm command index, extract the traction release rate parameter in the traction rhythm control command, compare the traction release rate parameter with the preset traction lag judgment standard, identify the path number where the traction release rhythm exhibits lag characteristics, and obtain the release rhythm lag path number set. S203: Based on the correspondence between the release rhythm delay path number set and the traction path number, the traction path number is associated and limited to obtain the delay path traction command marking result.

7. The personalized rehabilitation intervention method for burn scars based on edge computing according to claim 1, characterized in that, The specific steps for S3 are as follows: S301: Obtain the traction path number and time cycle from the lag path traction instruction marking result, retrieve the data from the contralateral muscle group monitoring device that matches the time cycle, extract the angular displacement trajectory data and vibration change signal within the traction cycle, and obtain the traction cycle behavior dataset. S302: Based on the traction cycle behavior dataset, extract the angular displacement change direction and the main traction direction path, identify the time periods in which the direction changes are in opposite directions, extract the vibration frequency change sequence corresponding to the time period, and obtain the vibration trend mapping set. S303: Based on the vibration trend mapping set, match the angular displacement offset trend and vibration frequency change, locate the interference characteristic segment within the continuous traction cycle, and obtain the result of the interference segment on the opposite side.

8. The personalized rehabilitation intervention method for burn scars based on edge computing according to claim 1, characterized in that, The specific steps of S4 are as follows: S401: Based on the interference time period located in the interference segment results on the opposite side and the corresponding traction path number in the traction command marking results of the lag path, extract the intersection of the interference time period and the traction control cycle under the path number according to the time axis coordinate, extract the time overlap data frame in the intersection, and obtain the rhythm conflict positioning segment. S402: Based on the path number in the rhythm conflict positioning segment, obtain the original traction release rhythm sequence of the path, remove the original rhythm structure that matches the time-overlapping data frame in the sequence, insert an alternating decreasing interval rhythm structure at the corresponding position, and obtain the replaced traction rhythm structure. S403: Based on the replaced traction rhythm structure, update the traction control instruction content dataset corresponding to the path number in the edge node, write the new rhythm structure into the specified position in the original instruction set mapping table, and obtain the reconstructed rhythm path release structure content.

9. The personalized rehabilitation intervention method for burn scars based on edge computing according to claim 8, characterized in that, During the extraction of the time-overlapping data frames: In the intersection of the interference time period located in the interference segment results on the opposite side and the traction control cycle under the path number, the data frame intervals that fall within the frame range corresponding to the time boundary of the traction control cycle are identified as the start time point and the end time point. During the process of writing the replaced traction rhythm structure to the specified position in the original instruction set mapping table: the corresponding position of the rhythm node in the original control instruction content dataset is located by the path number, and the rhythm structure is replaced by equal-length segment overlay.

10. The personalized rehabilitation intervention method for burn scars based on edge computing according to claim 1, characterized in that, The specific steps of S5 are as follows: S501: Based on the corresponding traction path number in the reconstructed rhythm path release structure, by accessing the rehabilitation management interface, retrieve the individual scar recovery status label, rehabilitation stage identifier and traction response type, verify the time consistency of the label data, and obtain the path status label set. S502: Based on the path state label set, call the rhythm parameter sequence in the reconstructed rhythm path release structure content, match the rhythm parameter with the recovery state label item by item, identify the parameter association relationship in the matching process, and obtain the rhythm state correspondence set; S503: Based on the rhythm state correspondence set, determine the traction path number adaptation state, divide the correspondence range between rhythm parameters and recovery state, and obtain the individualized traction intervention structure content.

Images