An individualized rehabilitation training prescription generation method based on evaluation results
By constructing a feature system of rehabilitation assessment results and generating individualized rehabilitation training prescriptions through rule-based reasoning, the problem of insufficient utilization of individual differences in rehabilitation training programs is solved, and the standardization and executability control of training programs are achieved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- CHANGCHUN UNIV OF TECH
- Filing Date
- 2026-05-08
- Publication Date
- 2026-06-05
AI Technical Summary
Existing rehabilitation training programs lack sufficient utilization of individual differences, inconsistent prescription generation rules, and adequate executability and dynamic updating capabilities. This results in a lack of unified constraints and coordination among training objectives, contraindications, and equipment capabilities, leading to insufficient executability of the generated results.
By acquiring rehabilitation assessment results, training history information, contraindication information, and equipment capability information, a prescription generation feature system is constructed. Candidate prescription parameter sets are generated using candidate training item screening and rule reasoning. A prescription update relationship graph is constructed, and individualized rehabilitation training prescriptions, parameter adjustment paths, and constraint interpretation information are output.
It achieves structured representation of individual differences in training subjects, standardized configuration of training programs, and executability control and continuous updating of prescription output.
Smart Images
Figure CN122157965A_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the fields of medical and health care information processing, rehabilitation training data processing and knowledge reasoning technology, and specifically relates to a method for generating individualized rehabilitation training prescriptions based on assessment results. Background Technology
[0002] In rehabilitation training applications, rehabilitation equipment, training terminals, and training management systems generate multi-dimensional data such as rehabilitation assessment results, training history records, training feedback data, contraindication and constraint information, and equipment capability parameters.
[0003] In existing technologies, training programs are usually developed manually by therapists or matched based on a single assessment level, which has the problems of insufficient utilization of individual differences and low standardization of prescription parameter configuration.
[0004] In addition, while some methods can recommend training items or give difficulty levels, there is still a lack of unified constraints and coordination between training objectives, taboo constraints and device capabilities, resulting in insufficient executability of the generated results.
[0005] Furthermore, existing methods typically only output the results of a single training scheme, making it difficult to simultaneously output parameter adjustment paths, constraint interpretation information, and dynamically updated results, which is not conducive to the continuous management and closed-loop adjustment of rehabilitation training schemes.
[0006] Therefore, there is an urgent need to provide a method for generating individualized rehabilitation training prescriptions based on assessment results, so as to achieve individualized prescription generation, constraint executability control, and dynamic updating of training prescriptions. Summary of the Invention
[0007] To address the problems of insufficient utilization of individual differences, inconsistent prescription generation rules, and insufficient executability and dynamic updating capabilities in existing technologies, this invention provides a method for generating individualized rehabilitation training prescriptions based on assessment results.
[0008] The technical solution adopted by this invention to solve the above-mentioned technical problems is as follows: by acquiring rehabilitation assessment results, training history information, training target information, contraindication and constraint information, and equipment capability information, a prescription generation feature system is constructed; on this basis, a candidate prescription parameter set is generated by using candidate training item screening and prescription rule reasoning, and a prescription update relationship graph is constructed by combining correlation analysis and statistical screening, thereby outputting the target individualized rehabilitation training prescription, parameter adjustment path, constraint interpretation information, and prescription suggestion information.
[0009] This invention provides a method for generating individualized rehabilitation training prescriptions based on assessment results. The method includes data acquisition and knowledge set construction, preprocessing and vector generation, prescription feature construction, candidate item screening, rule reasoning generation, candidate adjustment relationship statistics, prescription update relationship graph construction, feedback update feature generation, and target prescription output.
[0010] The method includes the following steps:
[0011] Step 1: Data Acquisition and Knowledge Set Construction
[0012] The system acquires rehabilitation assessment results, basic information, training history information, training objective information, contraindication and constraint information, rehabilitation equipment capability information, and prior information on prescription rules for the subjects to be trained, and constructs a prescription generation dataset and a prescription rule knowledge set.
[0013] Step 2, Preprocessing and Vector Generation:
[0014] The system performs missing value completion, outlier removal, dimension unification, normalization, and resampling on rehabilitation assessment results, training history information, contraindication and constraint information, and rehabilitation equipment capability information, and generates state feature vectors and constraint feature vectors.
[0015] Step 3: Prescription generation feature construction:
[0016] Based on rehabilitation assessment results, state feature vectors, constraint feature vectors, and training target information, a prescription generation feature system is constructed. The prescription generation feature system includes at least functional state features, training target features, and execution constraint features.
[0017] Step 4: Selection of candidate training projects:
[0018] Based on the prescription generation features, candidate training item matching is performed to generate a set of candidate training items and an item adaptation score.
[0019] Step 5: Rule-based reasoning generates candidate prescriptions:
[0020] Based on the candidate training item set, item adaptation score, and prescription rule knowledge set, prescription rule reasoning is performed to extract the training item selection logic, parameter mapping logic, and threshold conditions, and to generate a candidate prescription parameter set.
[0021] Step 6: Statistics on candidate adjustment relationships:
[0022] A correlation analysis was performed on historical training feedback data, review results, and prescription adjustment records, and a stability index of candidate adjustment relationships was statistically analyzed based on resampling.
[0023] Step 7: Constructing the prescription update relationship graph:
[0024] Candidate adjustment relationships that meet preset conditions are selected based on stability indicators, and a prescription update relationship diagram is constructed.
[0025] Step 8: Feedback and Update Feature Generation:
[0026] Obtain feedback data for the current training period of the target object and generate corresponding prescription update features.
[0027] Step 9, Target Prescription Output:
[0028] Based on the prescription update characteristics and combined with the prescription update relationship diagram, the candidate prescription parameter set is constrained, validated, and dynamically corrected, and the target individualized rehabilitation training prescription, parameter adjustment path, constraint explanation information, and prescription suggestion information are output.
[0029] Compared with the prior art, the present invention has the following beneficial effects: by constructing a prescription generation feature system based on the evaluation results, it is possible to achieve a structured representation of the individual differences of the subjects to be trained; by generating prescription parameters through candidate item screening and rule reasoning, it is possible to achieve a standardized configuration of the training scheme; and by stability screening, constraint verification and dynamic correction, it is possible to achieve executability control and continuous updating of prescription output. Attached Figure Description
[0030] Figure 1 This is a schematic diagram of the overall system architecture of the present invention.
[0031] Figure 2 This is a schematic diagram of the method flow of the present invention.
[0032] Figure 3 This is a schematic diagram of the prescription update relationship of the present invention.
[0033] Figure 4 This is a schematic diagram of the rule path and parameter determination logic of the present invention. Detailed Implementation
[0034] The present invention will now be described in detail with reference to the accompanying drawings.
[0035] like Figure 1 As shown, the system in this embodiment includes a data acquisition unit 101, a preprocessing unit 102, a feature construction unit 103, an item screening unit 104, a rule reasoning unit 105, a relationship mining unit 106, and a result output unit 107.
[0036] like Figure 2 As shown, the method flow of the present invention includes: data acquisition and knowledge set construction, preprocessing and vector generation, prescription feature construction, candidate item screening, rule reasoning generation, candidate adjustment relationship statistics, prescription update relationship graph construction, feedback update feature generation, and target prescription output.
[0037] In one embodiment, the rehabilitation assessment result vector can be represented as:
[0038] (1)
[0039] in, Represents a vector of rehabilitation assessment results; Indicates the first One evaluation indicator; This indicates the total number of evaluation indicators. The evaluation indicators include at least one or more of the following: trajectory accuracy, motion smoothness, motion efficiency, reaction time, and task completion ability.
[0040] In one embodiment, the state feature vector can be represented as:
[0041] (2)
[0042] in, Represents the state feature vector; Represents a state modeling function; Represents a vector of rehabilitation assessment results; Represents a basic information vector; This represents the training history information vector.
[0043] In one embodiment, the constrained feature vector can be represented as:
[0044] (3)
[0045] in, Represents the constraint eigenvector; Represents the constraint modeling function; Represents the taboo constraint information vector; This represents a vector containing information about the capabilities of rehabilitation equipment.
[0046] In one embodiment, the first in the training task library Each training item can be represented as:
[0047] (4)
[0048] in, Indicates the first One training program; Indicates the training item category; Indicates the type of training objective; Indicates the difficulty level; Indicates the applicable state range; This indicates the parameter boundary conditions.
[0049] In one embodiment, the training item category includes at least one or more of trajectory tracking training, target reach training, round-trip movement training, rhythmic response training, and fine control training.
[0050] In one embodiment, the set of candidate training items can be represented as:
[0051] (5)
[0052] in, Represents the set of candidate training items; This represents the function for filtering candidate items; Represents the state feature vector; Indicates training objective information; This represents the training task library.
[0053] In one embodiment, the first The item fit score for each candidate training item can be expressed as:
[0054] (6)
[0055] in, Indicates the first Project fit score for each candidate training item; This indicates the adaptation scoring function; Represents the state feature vector; Indicates training objective information; Indicates the first One training program.
[0056] In one embodiment, the candidate prescription parameter set can be represented as:
[0057] (7)
[0058] in, Represents the set of candidate prescription parameters; This represents the function that generates prescription parameters; Represents the state feature vector; Represents the constraint eigenvector; Represents the set of candidate training items; Represents a set of prescription rule knowledge.
[0059] In one embodiment, a targeted individualized rehabilitation training prescription can be represented as:
[0060] (8)
[0061] in, This indicates a personalized rehabilitation training prescription with specific goals; Represents a set of training items; Indicates the training order; Indicates the training frequency; Indicates the duration of a single training session; Indicates training intensity; Indicates the difficulty of the task; Indicates the target size; Indicates rhythm and speed; Indicates the number of repetitions; Indicates a rest interval; Indicates the feedback mode; This indicates the update strategy.
[0062] In one embodiment, step S4 uses a task screening model based on rule matching and adaptation scoring to perform candidate training item screening; step S5 uses a rule generation model based on prescription rule knowledge set to perform prescription rule reasoning.
[0063] like Figure 4 As shown, the rule path includes: using training load constraint features or fatigue-related features as the first-level judgment condition, using trajectory accuracy features or task completion ability features as the second-level judgment condition, and using training target features as the third-level judgment condition, and generating training items, training order, training frequency, single training duration, training intensity, task difficulty, target size, rhythm speed, number of repetitions and rest intervals accordingly.
[0064] In one embodiment, when fatigue-related characteristics exceed a preset threshold, the training intensity is reduced and the rest interval is extended; when trajectory accuracy is poor and hit rate is low, training items with larger target size, slower pace, and simpler path are selected; when the completion rate continues to exceed a preset threshold, the task difficulty is increased or the target size is reduced.
[0065] In one embodiment, the stability index of the candidate adjustment relationship can be expressed as:
[0066] (9)
[0067] in, Indicates the first The feedback metric and the first Stability indicators among prescription adjustment items; This indicates that the correspondence is in The number of times it is retained in the resampling; This indicates the total number of resampling attempts.
[0068] In one embodiment, the prescription update relationship graph consists of candidate adjustment relationships that meet preset conditions, including a stable occurrence frequency threshold and a confidence threshold. The prescription update relationships include the relationship between increased training completion rate and increased task difficulty, the relationship between decreased trajectory error and increased pace speed, the relationship between increased fatigue indicators and decreased training intensity, the relationship between decreased hit rate and increased target size, and the relationship between improved motion stability and increased proportion of fine-control training items.
[0069] In one embodiment, the target prescription after constraint verification can be represented as:
[0070] (10)
[0071] in, This represents the target prescription after constraint verification and dynamic correction. This represents the constraint verification and prescription correction function; Represents the set of candidate prescription parameters; Represents the constraint eigenvector; This represents a prescription update relationship diagram.
[0072] In one embodiment, the prescription update result for the next training cycle can be represented as:
[0073] (11)
[0074] in, Indicates the first The updated prescription for each training cycle; This represents the prescription update function; Indicates the first The current prescription for each training cycle; Indicates the first Feedback data from each training cycle; Indicates the first The review results after the end of each training cycle; This represents a prescription update relationship diagram.
[0075] like Figure 3 As shown, the prescription update relationship includes the relationship between training completion rate, trajectory error, fatigue index, hit rate, and motion stability and the corresponding prescription adjustment items, and forms a relationship structure that can be used for dynamic updates after stability screening.
[0076] In one embodiment, the output includes a target individualized rehabilitation training prescription, parameter adjustment path, constraint interpretation information, and prescription suggestion information. The target individualized rehabilitation training prescription may be output in at least one form, such as a prescription parameter table, training task sequence, device configuration parameter set, or update instruction set. The prescription parameter table may include at least one or more of the following: training item set, training order, training frequency, single training duration, training intensity, task difficulty, target size, pace, number of repetitions, rest interval, feedback mode, and update strategy.
[0077] The output of this invention is the result of rehabilitation training parameter configuration, training management parameters, or equipment configuration parameters, and does not directly constitute a method for diagnosing or treating diseases.
[0078] This invention is not limited to the embodiments described above. Various modifications and improvements can be made by those skilled in the art without departing from the principles of this invention, and these modifications and improvements should also be considered within the scope of protection of this invention.
Claims
1. A method for generating individualized rehabilitation training prescriptions based on assessment results, characterized in that, Performed by electronic devices, the process includes: acquiring the rehabilitation assessment results, basic information, training history information, training target information, contraindication and constraint information, rehabilitation equipment capability information, and prior information on prescription rules for the subject to be trained; constructing a prescription generation dataset and a prescription rule knowledge set; performing missing value completion, outlier removal, dimension unification, normalization, or resampling processing on the rehabilitation assessment results, training history information, contraindication and constraint information, and rehabilitation equipment capability information to generate state feature vectors and constraint feature vectors; constructing a prescription generation feature system based on the rehabilitation assessment results, state feature vectors, constraint feature vectors, and training target information; selecting candidate training items and calculating item fit scores based on the prescription generation features; performing rule reasoning based on candidate training items, item fit scores, and the prescription rule knowledge set to generate a candidate prescription parameter set; performing correlation analysis on historical training feedback data, review results, and prescription adjustment records, and statistically analyzing the stability of candidate adjustment relationships to construct a prescription update relationship graph; generating prescription update features based on the feedback data of the current training cycle, and performing constraint verification and dynamic correction on the candidate prescription parameter set in conjunction with the prescription update relationship graph, outputting the target individualized rehabilitation training prescription and its parameter adjustment path, constraint explanation information, and prescription suggestion information.
2. The method for generating individualized rehabilitation training prescriptions based on assessment results according to claim 1, characterized in that, The prior information for the prescription rules is generated from at least one of the following sources: prescription experience rules given by rehabilitation physicians or therapists, historical rehabilitation training prescription records, rehabilitation equipment configuration rules, rehabilitation training task adaptation rules, and training safety restriction rules.
3. The method for generating individualized rehabilitation training prescriptions based on assessment results according to claim 1, characterized in that, The prescription generation feature system includes functional state features, training target features, and execution constraint features; the functional state features include trajectory accuracy features, motion stability features, motion efficiency features, reaction ability features, and task completion ability features; the training target features include target ability dimension features, priority training direction features, and stage target features. The execution constraint features include forbidden action features, joint range of motion constraint features, training load constraint features, and equipment capability boundary features.
4. The method for generating individualized rehabilitation training prescriptions based on assessment results according to claim 3, characterized in that: The trajectory accuracy feature is characterized by the trajectory error level; the motion smoothness feature is characterized by the degree of velocity fluctuation or acceleration change; the motion efficiency feature is characterized by path efficiency or task completion efficiency; the reaction capability feature is characterized by the prompt response delay; the task completion capability feature is characterized by the hit rate or completion rate; the target capability dimension feature is characterized by the category of control capability to be improved; the priority training direction feature is characterized by the preset training target weight; the forbidden action feature is characterized by the range of actions that are not allowed to be performed; the joint range of motion constraint feature is characterized by the allowed range of motion angles or displacements; and the device capability boundary feature is characterized by the range of displacement, speed, assistance, or resistance that the device can perform.
5. The method for generating individualized rehabilitation training prescriptions based on assessment results according to claim 1, characterized in that, The selection of candidate training projects adopts a task selection model based on rule matching and adaptation scoring, and the prescription rule reasoning adopts a rule generation model based on prescription rule knowledge set to perform reasoning.
6. The method for generating individualized rehabilitation training prescriptions based on assessment results according to claim 1, characterized in that, The training item selection logic and parameter mapping logic in the rule reasoning include: using training load constraint features or fatigue-related features as the first-level judgment features, using trajectory accuracy features or task completion ability features as the second-level judgment features, and using training target features as the third-level judgment features, and generating training items, training order, training frequency, single training duration, training intensity, task difficulty, target size, rhythm speed, number of repetitions and rest intervals accordingly.
7. The method for generating individualized rehabilitation training prescriptions based on assessment results according to claim 1, characterized in that, The screening criteria for the stability of the candidate adjustment relationship include a stable occurrence frequency threshold and a confidence threshold. The candidate adjustment relationship is retained in the prescription update relationship graph when it meets both the stable occurrence frequency threshold and the confidence threshold. The prescription update relationship graph includes at least one of the following relationships: the relationship between increased training completion rate and increased task difficulty, the relationship between decreased trajectory error and increased pace speed, the relationship between increased fatigue index and decreased training intensity, the relationship between decreased hit rate and increased target size, and the relationship between improved motion stability and increased proportion of fine control training items.
8. An electronic device, characterized in that, It includes a processor and a memory, wherein the memory stores a computer program, which, when executed by the processor, implements the method according to any one of claims 1 to 7.
9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method described in any one of claims 1 to 7.