Auxiliary rehabilitation method and system based on robot

Abstract

The invention discloses an auxiliary rehabilitation method and system based on a robot, and relates to the technical field of auxiliary rehabilitation. If the scene abnormality exceeds the expectation, the length of a rehabilitation observation period is constrained according to the scene abnormality, and the rehabilitation demand of a patient is identified after the rehabilitation state data of the patient is collected; a corresponding rehabilitation scheme is matched for the patient according to the rehabilitation requirements; after the rehabilitation robot executes the corresponding rehabilitation scheme, if the number of the abnormal sign indexes of the patient exceeds the expectation, a multi-stage alarm mechanism is triggered, and a multi-stage alarm instruction is sent to the outside; the rehabilitation robot automatically adjusts the environmental factors in the rehabilitation area, and if the rehabilitation target of the patient is still not achieved, the rehabilitation scheme is optimized according to the re-determined rehabilitation requirement, or the rehabilitation robot is intervened according to the detection data. The rehabilitation state of the patient is monitored and evaluated in real time, and the rehabilitation state is used as feedback to adjust the rehabilitation scheme and the control strategy of the rehabilitation robot.

Description

Technical Field

[0001] The present invention relates to the technical field of assisted rehabilitation, and in particular to a robot-based assisted rehabilitation method and system. Background Art

[0002] Robot-assisted rehabilitation is an advanced rehabilitation treatment method that integrates modern robotics technology and rehabilitation medicine theory. By using sophisticated robotic equipment, combined with professional rehabilitation medicine knowledge and personalized treatment plans, this method can provide patients with accurate, efficient and customized rehabilitation training services. Compared with traditional rehabilitation therapy, robot-assisted rehabilitation can more accurately simulate human movement, thereby more effectively promoting patients' muscle recovery, joint flexibility improvement and nerve function reconstruction. In addition, this method can significantly reduce the workload of rehabilitation therapists and improve the efficiency and quality of overall rehabilitation treatment. For patients, robot-assisted rehabilitation not only provides a more scientific and intelligent rehabilitation experience, but also accelerates the rehabilitation process, improves the quality of life, and creates favorable conditions for them to return to society and take care of themselves.

[0003] In a Chinese invention patent with application publication number CN114496159A, a robot-assisted rehabilitation method, system, device and medium are disclosed, the method comprising: collecting the user's EEG signal and starting timing; preprocessing the user's EEG signal to obtain a preprocessed EEG signal; performing feature extraction processing on the preprocessed EEG signal to obtain feature data; determining whether a training error-related potential occurs based on the feature data and a trained model; if so, sending a signal requiring assistance to the robot to control the robot to operate according to the signal requiring assistance; if not, determining whether the recorded time reaches a set time threshold; if the time threshold is reached, sending a signal requiring no assistance to the robot to control the robot to remain stationary; if the time threshold is not reached, continuing to collect the user's EEG signal; the method proposed by the present invention can achieve automatic assisted rehabilitation.

[0004] When the patient is in the rehabilitation state, considering the need to improve the patient's rehabilitation efficiency and save the caregiver's time, a rehabilitation robot is usually introduced in the patient's rehabilitation area to assist in rehabilitation. For example, the rehabilitation robot collects and monitors the patient's various physical signs and behavioral data in real time. When the patient's physical signs and behavioral data are abnormal, an alarm is promptly issued to the outside and the intervention of the guardian is introduced to ensure the patient's rehabilitation effect.

[0005] In the existing rehabilitation robot-assisted rehabilitation methods, medical staff usually develop personalized rehabilitation plans for patients in advance, which are then assisted in execution by the rehabilitation robot, and the patient's rehabilitation data is used as feedback to optimize and correct the rehabilitation plan until the patient reaches the expected rehabilitation goal. However, when there are certain interference conditions in the rehabilitation area, such as the noise, light conditions, temperature and humidity or other environmental factors in the rehabilitation area may be abnormal to a certain extent, it is difficult for the rehabilitation robot to adjust the current rehabilitation process and rehabilitation environment in real time based on the patient's physical signs and behavioral status data. Therefore, when controlling the rehabilitation robot to assist the patient in rehabilitation, the patient's rehabilitation process will be affected to a certain extent.

[0006] To this end, the present invention provides a robot-based assisted rehabilitation method and system. Summary of the invention

[0007] 1. Technical issues to be resolved

[0008] In view of the deficiencies of the prior art, the present invention provides a robot-based assisted rehabilitation method and system, which matches the corresponding rehabilitation plan for the patient according to the rehabilitation needs; after the rehabilitation robot executes the corresponding rehabilitation plan, if the number of abnormal physical signs of the patient exceeds the expected number, a multi-level alarm mechanism is triggered and a multi-level alarm instruction is issued to the outside; the rehabilitation robot automatically adjusts the environmental factors in the rehabilitation area, and if the patient's rehabilitation goal is still not achieved, the rehabilitation plan is optimized according to the re-determined rehabilitation needs, or the rehabilitation robot is intervened according to the detection data. The patient's rehabilitation status is monitored and evaluated in real time, and the rehabilitation plan and the control strategy of the rehabilitation robot are adjusted based on this feedback; thereby solving the technical problems recorded in the background technology.

[0009] (II) Technical solution

[0010] To achieve the above objectives, the present invention is implemented through the following technical solutions:

[0011] The robot-based assisted rehabilitation method includes collecting rehabilitation scene data in the rehabilitation area, generating scene abnormality from the rehabilitation scene data, and issuing a health assessment instruction to the outside if the scene abnormality exceeds expectations;

[0012] Constrain the length of the rehabilitation observation period based on the abnormality of the scene, identify the patient's rehabilitation needs after collecting the patient's rehabilitation status data, and match the patient with the corresponding rehabilitation plan based on the rehabilitation needs;

[0013] After the rehabilitation robot executes the corresponding rehabilitation plan, if the number of abnormal physical signs of the patient exceeds expectations, the knowledge graph of emergency treatment of abnormal physical signs will give an emergency treatment plan based on the abnormal physical signs, trigger a multi-level alarm mechanism and issue a multi-level alarm command to the outside;

[0014] After analyzing and obtaining the key factors that affect the patient's rehabilitation, the rehabilitation robot automatically adjusts the environmental factors in the rehabilitation area. If the patient's rehabilitation goal is still not achieved, a three-level alarm command is issued to the outside.

[0015] The rehabilitation program is optimized according to the re-determined rehabilitation needs, or the rehabilitation robot is comprehensively tested and corresponding test data is obtained, and intervention is performed on the rehabilitation robot according to the test data.

[0016] Furthermore, a sensor network is arranged in the rehabilitation area, and environmental parameters are collected using the sensor network. The acquired environmental parameters are summarized to generate a rehabilitation scene data set. The rehabilitation scene data in the rehabilitation scene data set is used as input, and the trained scene condition evaluation model is used to output the scene abnormality in the rehabilitation area.

[0017] Furthermore, after receiving the health assessment instruction, the length of the rehabilitation observation period cT is constrained according to the abnormality Co of several consecutive scenes recently, and the length of the rehabilitation observation period is adjusted according to the constraint conditions. The constraint method is as follows:

[0018]

[0019] Weight coefficient, 0≤α≤1, 0≤β≤1; n is the number of time nodes, Co i is the abnormality of the scene at the i-th time node, Co a is its average value.

[0020] Furthermore, the patient's physical signs, mental health data, and behavioral data in the current rehabilitation scenario are continuously monitored and collected, and summarized as the patient's rehabilitation status data; the patient's rehabilitation status data is used as input, and the trained rehabilitation needs identification model is used to identify rehabilitation needs, and the corresponding rehabilitation plan is matched to the patient based on the rehabilitation needs.

[0021] Furthermore, the rehabilitation robot executes a corresponding rehabilitation program in the patient rehabilitation area, and collects the patient's vital sign data after a preset rehabilitation observation period; the patient's vital sign data is used as input, and the abnormal vital sign index of the patient is identified using the trained abnormal index identification model;

[0022] If the number of abnormal vital signs of the patient exceeds expectations, a first-level alarm command will be issued to the outside.

[0023] Furthermore, after receiving the first-level alarm command, the emergency treatment of abnormal physical signs of rehabilitation patients is used as the target word, and a knowledge graph for emergency treatment of abnormal physical signs is pre-built; the abnormal physical sign indicators of patients during the rehabilitation observation period are collected, and the rehabilitation robot sends the emergency treatment plan to the cloud.

[0024] Furthermore, after executing the emergency response plan, if the number of consecutive first-level alarm instructions received exceeds the expected number, a risk value R is generated based on the state data of the first-level alarm instructions received. t ;

[0025] If the risk value R is obtained t If the risk threshold is exceeded, a secondary alarm command will be issued to the outside world, and the patient's recovery degree data, vital signs data and environmental conditions data will be recorded.

[0026] Furthermore, under dimensionless conditions, the risk value R is generated based on the status data of the first-level alarm instruction received. t The way is as follows:

[0027]

[0028] Where: T i is the time interval between the ith and i+1st alarms, N i is the number of abnormal indicators at the time of the i-th alarm, γ is the time attenuation coefficient, α and β are weight coefficients, and γ is the attenuation coefficient of the cumulative impact.

[0029] Furthermore, after receiving the secondary alarm instruction, the principal component analysis is used to analyze the environmental condition data in the rehabilitation area to obtain the key factors affecting the patient's rehabilitation, and the corresponding qualified interval is set for each key factor according to the degree of influence of the key factor;

[0030] During the rehabilitation observation period, various vital signs of the patient are collected as feedback data, and the feedback data is used as input. The trained patient rehabilitation evaluation model is used to output the rehabilitation value. If the obtained rehabilitation value is lower than the rehabilitation threshold, a three-level alarm instruction is issued to the outside.

[0031] Furthermore, if the third-level alarm command is not received, the patient's rehabilitation status data is re-collected, the patient's rehabilitation needs are re-identified, and the rehabilitation plan is optimized based on the re-determined rehabilitation needs, and the rehabilitation robot executes the optimized rehabilitation plan.

[0032] Furthermore, if a third-level alarm command is received, the rehabilitation robot is comprehensively inspected and corresponding detection data is obtained, which are summarized to generate a rehabilitation robot detection data set; a control coefficient is generated from the detection data of the rehabilitation robot detection data set, and if the control coefficient does not exceed a preset response threshold, the rehabilitation robot stops the current task, and the robot operation is guided or directly controlled through remote control.

[0033] The robot-based assisted rehabilitation system includes a scene analysis unit, which collects rehabilitation scene data in the rehabilitation area, generates scene abnormality from the rehabilitation scene data, and issues a health assessment instruction to the outside if the scene abnormality exceeds expectations;

[0034] The rehabilitation program generation unit constrains the length of the rehabilitation observation period according to the abnormality of the scene, identifies the patient's rehabilitation needs after collecting the patient's rehabilitation status data, and matches the patient with a corresponding rehabilitation program based on the rehabilitation needs;

[0035] Multi-level alarm unit: after the rehabilitation robot executes the corresponding rehabilitation plan, if the number of abnormal physical signs of the patient exceeds the expected number, the knowledge graph of emergency treatment of abnormal physical signs will give an emergency treatment plan based on the abnormal physical signs, trigger the multi-level alarm mechanism and issue a multi-level alarm command to the outside world;

[0036] The adjustment unit analyzes and obtains the key factors that affect the patient's rehabilitation. The rehabilitation robot automatically adjusts the environmental factors in the rehabilitation area. If the patient's rehabilitation goal is still not achieved, a three-level alarm command is issued to the outside.

[0037] The feedback processing unit optimizes the rehabilitation plan according to the re-determined rehabilitation needs, or performs a comprehensive test on the rehabilitation robot and obtains corresponding test data, and intervenes in the rehabilitation robot according to the test data.

[0038] (III) Beneficial effects

[0039] The present invention provides a robot-based assisted rehabilitation method and system, which has the following beneficial effects:

[0040] 1. Judge and evaluate the degree of abnormality in the rehabilitation area based on the abnormality of the scene. When the current rehabilitation area is not conducive to the patient's rehabilitation, adjust the environment in the rehabilitation area to avoid negative impact on the patient's rehabilitation process.

[0041] 2. Adjust the pre-set rehabilitation observation period, and readjust the patient's rehabilitation plan or the control strategy of the rehabilitation robot after the rehabilitation observation period, using the patient's environmental conditions as feedback to ensure the patient's rehabilitation effect.

[0042] 3. Based on data analysis and identification, determine the abnormality of the patient's current indicators, complete rehabilitation needs analysis and formulate rehabilitation plans. By formulating targeted rehabilitation plans and assisting with rehabilitation robots, the patient's rehabilitation efficiency and effectiveness can be improved.

[0043] 4. Identify abnormal physical signs and indicators, monitor and alarm the patient's rehabilitation effect, and intervene in time when the patient's rehabilitation effect is not good; use the patient's abnormal physical signs and indicators as feedback, and provide emergency treatment plans based on the abnormal physical signs emergency treatment knowledge map as a reference and executed or guided by the rehabilitation robot to ensure the patient's rehabilitation effect.

[0044] 5. By building a multi-level alarm mechanism and issuing alarms multiple times, the patient's rehabilitation status can be monitored and evaluated in real time when the patient is in the rehabilitation state, and this can be used as feedback to adjust the rehabilitation plan and the control strategy of the rehabilitation robot to ensure the patient's rehabilitation effect.

[0045] 6. Use rehabilitation robots to issue control commands, or use trained automatic control models to control various equipment in the rehabilitation area to reduce the interference of environmental conditions on patient rehabilitation. By evaluating the patient's rehabilitation effect, the reliability of the current rehabilitation process can be verified, and a three-level alarm command can be issued when the rehabilitation effect does not exceed expectations, enriching the current multi-level alarm mechanism.

[0046] 7. Re-detect and identify the patient's various status data. After the rehabilitation robot executes the new rehabilitation plan, it can switch the current rehabilitation stage and adjust the patient's rehabilitation stage. By constructing a control coefficient, the response effect of the rehabilitation robot is evaluated and the rehabilitation robot that executes the rehabilitation plan is adjusted. When the current rehabilitation effect is not good, positive feedback can be formed in time to ensure the patient's rehabilitation effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 This is a schematic diagram of the process of the robot-based assisted rehabilitation method of the present invention;

[0048] Figure 2 This is a schematic diagram of the structure of the robot-based assisted rehabilitation system of the present invention. DETAILED DESCRIPTION

[0049] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0050] See also Figure 1 The present invention provides a robot-based assisted rehabilitation method, comprising:

[0051] Step 1: Collect rehabilitation scene data in the rehabilitation area, generate scene abnormality from the rehabilitation scene data, and if the scene abnormality exceeds expectations, issue a health assessment instruction to the outside;

[0052] The step 1 includes the following contents:

[0053] Step 101: After the patient enters a rehabilitation state and the rehabilitation area of ​​the patient is determined, a sensor network, such as a light sensor, a temperature and humidity sensor, and a noise sensor, is arranged in the rehabilitation area. The sensor network is used to monitor and collect environmental parameters such as light, noise, temperature, humidity, and air quality in real time, and the acquired environmental parameters are aggregated to generate a rehabilitation scene data set.

[0054] Step 102: After training the machine learning algorithm with the labeled sample data, a trained scene condition evaluation model is obtained; using the rehabilitation scene data as input, the trained scene condition evaluation model is used to evaluate the degree of abnormality of the rehabilitation environment in the rehabilitation area, and the corresponding scene abnormality is output;

[0055] Based on historical data and management expectations for patient rehabilitation, abnormal thresholds are pre-set; if the acquired scene abnormality exceeds the abnormal threshold, it means that the current environmental conditions are not conducive to the patient's rehabilitation. A rehabilitation robot can be introduced in the current rehabilitation scene to issue health assessment instructions to the outside world;

[0056] When using, combine the contents in steps 101 to 102:

[0057] After the patient enters the rehabilitation state, by collecting environmental condition data in the rehabilitation area and conducting comprehensive abnormality analysis, the degree of abnormality in the rehabilitation area is judged and evaluated based on the abnormality of the scene to see whether it will interfere with the patient's rehabilitation effect. When the current rehabilitation area is not conducive to the patient's rehabilitation, the environment in the rehabilitation area is adjusted, or the rehabilitation plan is optimized, etc., to avoid the negative impact of abnormal rehabilitation scenes on the patient's rehabilitation process.

[0058] In the existing rehabilitation robot-assisted rehabilitation methods, medical staff usually develop personalized rehabilitation plans for patients in advance, which are then assisted in execution by the rehabilitation robot, which uses the patient's rehabilitation data as feedback to optimize and modify the rehabilitation plan until the patient reaches the expected rehabilitation goal. However, when there are certain interference conditions in the rehabilitation area, such as the noise, light conditions, temperature and humidity or other environmental factors in the rehabilitation area, it is difficult for the rehabilitation robot to adjust the current rehabilitation process and rehabilitation environment in real time according to the patient's physical signs and behavioral status. As a result, when the rehabilitation robot is controlled to assist the patient in rehabilitation, the rehabilitation process will be affected to a certain extent.

[0059] Step 2: Constrain the length of the rehabilitation observation period based on the scene abnormality, identify the patient's rehabilitation needs after collecting the patient's rehabilitation status data, and match the patient with a corresponding rehabilitation plan based on the rehabilitation needs;

[0060] The step 2 includes the following contents:

[0061] Step 201: After receiving the health assessment instruction, the length of the rehabilitation observation period cT is constrained according to the abnormality Co of several consecutive scenes in the recent period, and the length of the rehabilitation observation period is adjusted according to the constraint conditions. The constraint method is as follows:

[0062]

[0063] Weight coefficient, 0≤α≤1, 0≤β≤1; n is the number of time nodes, Co i is the abnormality of the scene at the i-th time node, Co a is its average value;

[0064] When using it, considering that environmental conditions will have a certain interference on the patient's rehabilitation status and rehabilitation effect, in this scenario, the pre-set rehabilitation observation period can be adjusted, and the patient's rehabilitation plan or the control strategy of the rehabilitation robot can be readjusted after the rehabilitation observation period, so as to complete the preliminary construction of the feedback mechanism based on the patient's environmental conditions as feedback, and realize the guarantee of the patient's rehabilitation effect;

[0065] Step 202: The rehabilitation robot or wearable device continuously monitors and collects the patient's physical signs, such as heart rate, respiratory rate, and electromyography; mental health data, such as anxiety level and mood swings; and behavioral data in the current rehabilitation scenario, and generates patient rehabilitation status data after aggregation;

[0066] The machine learning algorithm is trained with the labeled sample data to obtain a trained rehabilitation needs recognition model;

[0067] Taking the patient's rehabilitation status data as input, the trained rehabilitation needs identification model is used to identify the patient's rehabilitation needs, and the corresponding rehabilitation plan is matched to the patient based on the rehabilitation needs;

[0068] When using, combine the contents in steps 201 and 202:

[0069] When introducing rehabilitation robots to assist patients in rehabilitation, wearable devices and imaging devices collect various data of the patient in the rehabilitation area, and judge the abnormality of the patient's current indicators based on data analysis and identification methods, complete the rehabilitation needs analysis and formulation of rehabilitation plans. By formulating targeted rehabilitation plans and assisting with rehabilitation robots, the patient's rehabilitation efficiency and effectiveness can be improved.

[0070] Step 3: After the rehabilitation robot executes the corresponding rehabilitation plan, if the number of abnormal physical signs of the patient exceeds expectations, the knowledge graph for emergency treatment of abnormal physical signs will give an emergency treatment plan based on the abnormal physical signs, trigger a multi-level alarm mechanism and issue a multi-level alarm command to the outside;

[0071] The step three includes the following contents:

[0072] Step 301: The rehabilitation robot executes a corresponding rehabilitation program in the patient rehabilitation area, and collects the patient's vital sign data, such as heart rate, respiratory rate, blood sugar, blood oxygen and mobility data, after a preset rehabilitation observation period;

[0073] After training the convolutional neural network with the labeled sample data, a trained abnormal indicator recognition model is obtained;

[0074] Taking the patient's vital signs data as input, the trained abnormal indicator recognition model is used to identify the patient's abnormal vital signs. If the number of abnormal vital signs of the patient exceeds the expected number, a first-level alarm instruction is issued to the outside.

[0075] When in use, after executing the rehabilitation program, the system uses the patient's data as feedback to identify abnormal physical signs and indicators, monitor and alarm the patient's rehabilitation effect, and intervene in time when the patient's rehabilitation effect is not good;

[0076] Step 302: After receiving the first-level alarm instruction, the abnormal physical sign emergency treatment of the rehabilitation patient is used as the target word, and after deep retrieval and entity relationship building, a knowledge graph of abnormal physical sign emergency treatment is pre-built;

[0077] Collect the abnormal physical signs of patients during the rehabilitation observation period, and give emergency treatment plans based on the correspondence between the abnormal physical signs and emergency treatment plans. The rehabilitation robot will send the emergency treatment plans to patients, family members, medical staff, etc., or send the emergency treatment plans to the cloud.

[0078] When in use, taking into account the need for intervention when the patient's rehabilitation effect is poor, the patient's abnormal physical signs indicators are used as feedback, and an emergency treatment plan is given by the abnormal physical signs emergency treatment knowledge graph. This is used as a reference and executed or guided by the rehabilitation robot to ensure the patient's rehabilitation effect.

[0079] Step 303: After the rehabilitation robot, medical staff or rehabilitation robot executes the emergency treatment plan, if the number of consecutive first-level alarm instructions received exceeds the expectation, a risk value R is generated based on the state data of the first-level alarm instructions received. t , generated as follows:

[0080]

[0081] Where: T i is the time interval between the ith and i+1st alarms, N iis the number of abnormal indicators at the time of the i-th alarm, γ is the time attenuation coefficient, the value is between 0.01 and 1, α and β are weight coefficients, the values ​​are consistent with the previous values, γ is the attenuation coefficient of the cumulative impact, the value falls between 0.01 and 1, and e can be 2.713;

[0082] Pre-set risk thresholds based on historical data and patient recovery expectations;

[0083] If the risk value obtained exceeds the risk threshold, a secondary alarm command is issued to the outside to notify the patient, medical staff and family members, and record alarm-related data, such as the patient's recovery degree data, physical sign data and environmental condition data when the secondary alarm command is issued;

[0084] When using, combine the contents in steps 301 to 303:

[0085] Construct the risk value R based on the first-level alarm instruction t , according to the risk value R t Issue a secondary alarm command to complete the construction of a multi-level alarm mechanism. By building a multi-level alarm mechanism and issuing alarms multiple times, the patient's rehabilitation status can be monitored and evaluated in real time when the patient is in a rehabilitation state, and this can be used as feedback to adjust the rehabilitation plan and the control strategy of the rehabilitation robot to ensure the patient's rehabilitation effect.

[0086] Step 4: After analyzing and obtaining the key factors that affect the patient's rehabilitation, the rehabilitation robot automatically adjusts the environmental factors in the rehabilitation area. If the patient's rehabilitation goal is still not achieved, a three-level alarm command is issued to the outside;

[0087] The step 4 includes the following contents:

[0088] Step 401: After receiving the secondary alarm instruction, use principal component analysis to analyze the environmental condition data in the rehabilitation area to obtain key factors that affect the patient's rehabilitation;

[0089] Set the corresponding qualified interval for each key factor according to the influence of the key factor; train the machine learning algorithm with the labeled sample data to obtain the trained automatic control model;

[0090] The rehabilitation robot uses the trained automatic control model to automatically adjust the environmental factors in the rehabilitation area, such as reducing light intensity and noise to increase the comfort of the patient;

[0091] When in use, taking into account that the environmental conditions in the rehabilitation area may affect the patient's rehabilitation process, as a processing strategy, a rehabilitation robot is used to issue control instructions, or a trained automatic control model is used to control various equipment in the rehabilitation area, such as a humidifier or audio equipment in the rehabilitation area, to reduce the interference of environmental conditions on the patient's rehabilitation.

[0092] Step 402: Collect various vital sign data of the patient as feedback data during the rehabilitation observation period, such as heart rate, respiratory rate, blood sugar, blood oxygen and mobility data;

[0093] The machine learning algorithm is trained with the labeled sample data to obtain the trained patient rehabilitation evaluation model; the feedback data is used as input, and the trained patient rehabilitation evaluation model is used to evaluate the patient's current rehabilitation status and output the rehabilitation value, and the patient's rehabilitation degree can be verified based on the rehabilitation value;

[0094] The rehabilitation threshold is pre-set based on historical data and expectations for patient rehabilitation management; if the obtained rehabilitation value is lower than the rehabilitation threshold, it means that the patient has not achieved the expected effect after the intervention of the rehabilitation robot and the implementation of the targeted rehabilitation plan. At this time, a three-level alarm command is issued to the outside;

[0095] When using, combine the contents in steps 401 and 402:

[0096] After making adaptive adjustments to the key influencing factors within the environmental conditions and completing the current rehabilitation process, the reliability of the current rehabilitation process can be verified by evaluating the patient's rehabilitation effect. At this time, when the rehabilitation effect does not exceed expectations, a third-level alarm command is issued, enriching the current multi-level alarm mechanism.

[0097] Step 5: Optimize the rehabilitation plan based on the re-determined rehabilitation needs, or conduct a comprehensive test on the rehabilitation robot and obtain corresponding test data, and intervene in the rehabilitation robot based on the test data;

[0098] The step five includes the following contents:

[0099] Step 501: If the third-level alarm instruction is not received, after re-collecting the patient's rehabilitation status data, the rehabilitation needs identification model after training is used to identify the patient's rehabilitation needs again, and the rehabilitation plan is optimized according to the re-determined rehabilitation needs, and the rehabilitation robot executes the optimized rehabilitation plan;

[0100] When in use, when the rehabilitation effect in the current stage reaches the expected effect, the patient's various status data are re-detected and identified, and after the rehabilitation robot executes the new rehabilitation plan, the current rehabilitation stage is switched and the rehabilitation stage of the patient's rehabilitation is adjusted;

[0101] Step 502: If a third-level alarm instruction is received, a comprehensive inspection of the rehabilitation robot is performed and corresponding inspection data, such as control accuracy data response speed, etc., are obtained, and the inspection data set of the rehabilitation robot is generated after being summarized;

[0102] The control coefficient Kcp is generated from the rehabilitation robot detection data set detection data in the following way: linear normalization is performed on the control accuracy Jo and the response speed Fo, and the corresponding data values ​​are mapped to the interval [0,1] according to the following method:

[0103]

[0104] Weight coefficient: 0≤F 1 ≤1,0≤F 2 ≤1 and F 2 +F 1 =1;Jo i is the control accuracy of the i-th test node, Jo a is the mean value of control accuracy, Jo b is the qualified standard value of control accuracy; Fo i is the response speed of the ith test node, Fo a is the mean response speed, Fo b is the qualified standard value of response speed;

[0105] If the control coefficient Kcp does not exceed the preset response threshold, the rehabilitation robot stops the current task and performs robot operation guidance or direct control through remote control;

[0106] When using, combine the contents in steps 501 and 502:

[0107] As another aspect of feedback, different from the switching of rehabilitation stages, when the rehabilitation effect fails to meet expectations, it is considered that there may be deficiencies in the control strategy or control method of the rehabilitation robot. By constructing a control coefficient to evaluate the response effect of the rehabilitation robot, this can be used as a reference to adjust the rehabilitation robot that executes the rehabilitation plan. When the current rehabilitation effect is not good, positive feedback can be formed in time to ensure the patient's rehabilitation effect.

[0108] The construction method of the knowledge graph of emergency treatment of abnormal physical signs of rehabilitation patients can refer to the following content:

[0109] Clarify the goal of building a knowledge graph

[0110] First, we need to clarify the goal of building the knowledge graph, which is to represent and reason about the knowledge of emergency treatment of abnormal signs of rehabilitation patients. This includes but is not limited to the basic information of the patient, description of abnormal signs, emergency treatment measures, and evaluation of treatment effects.

[0111] Data collection and preprocessing

[0112] Data sources: Collect data on emergency treatment of abnormal physical signs of rehabilitation patients published by rehabilitation hospitals, rehabilitation centers, and relevant scientific research institutions, including case reports, expert experience, clinical guidelines, etc. Data preprocessing: Clean, remove duplicates, and format the collected data to ensure data accuracy and consistency.

[0113] Knowledge Extraction

[0114] Entity extraction: Identify entities related to the emergency treatment of abnormal signs of rehabilitation patients from the preprocessed data, such as the patient's name, age, gender, disease type, name of abnormal signs, emergency treatment measures, etc. Relationship extraction: Use natural language processing and other technologies to extract the relationship between entities from the text, such as "patient A suffers from disease B", "the abnormal sign of disease B is C", "the emergency treatment measure for abnormal sign C is D", etc. Attribute extraction: Extract attributes from entities, such as the patient's vital signs, the severity of abnormal signs, and the specific steps of emergency treatment measures.

[0115] Knowledge Fusion

[0116] Entity alignment: Align the same entities in data from different sources to ensure that the entities in the knowledge graph are unique. Relationship integration: Integrate the extracted relationships to eliminate redundancy and conflict and form a complete relationship network. Attribute filling: Fill and improve the attributes of the entity to ensure the integrity and accuracy of the knowledge graph.

[0117] Knowledge graph construction

[0118] Ontology construction: Based on the knowledge domain of emergency treatment of abnormal signs of rehabilitation patients, an ontology model is constructed to define the types and hierarchical structures of entities, relationships, and attributes. Knowledge representation: The knowledge in the knowledge graph is represented in the form of triples (entity-relationship-entity) or (entity-attribute-attribute value). Knowledge storage: Select a suitable database or graph database to store the knowledge graph to ensure efficient retrieval and reasoning of knowledge.

[0119] Quality Assessment and Optimization

[0120] Accuracy assessment: The accuracy of the knowledge in the knowledge graph is assessed through methods such as manual sampling detection to ensure the correctness of the knowledge. Consistency assessment: Check whether the relationships in the knowledge graph are consistent to avoid self-contradictions. Completeness assessment: Evaluate the completeness of the knowledge graph to ensure that the main knowledge points of emergency treatment of abnormal signs of rehabilitation patients are covered. Optimization: Based on the evaluation results, the knowledge graph is optimized, including correcting errors and supplementing missing knowledge points.

[0121] See also Figure 2 The present invention provides a robot-based assisted rehabilitation system, comprising:

[0122] The scene analysis unit collects the rehabilitation scene data in the rehabilitation area, generates the scene abnormality degree from the rehabilitation scene data, and issues a health assessment instruction to the outside if the scene abnormality degree exceeds expectations;

[0123] The rehabilitation program generation unit constrains the length of the rehabilitation observation period according to the abnormality of the scene, identifies the patient's rehabilitation needs after collecting the patient's rehabilitation status data, and matches the patient with a corresponding rehabilitation program based on the rehabilitation needs;

[0124] Multi-level alarm unit: after the rehabilitation robot executes the corresponding rehabilitation plan, if the number of abnormal physical signs of the patient exceeds the expected number, the knowledge graph of emergency treatment of abnormal physical signs will give an emergency treatment plan based on the abnormal physical signs, trigger the multi-level alarm mechanism and issue a multi-level alarm command to the outside world;

[0125] The adjustment unit analyzes and obtains the key factors that affect the patient's rehabilitation. The rehabilitation robot automatically adjusts the environmental factors in the rehabilitation area. If the patient's rehabilitation goal is still not achieved, a three-level alarm command is issued to the outside.

[0126] The feedback processing unit optimizes the rehabilitation plan according to the re-determined rehabilitation needs, or performs a comprehensive test on the rehabilitation robot and obtains corresponding test data, and intervenes in the rehabilitation robot according to the test data.

[0127] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0128] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0129] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only some logical function divisions. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0130] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0131] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A robot-based assisted rehabilitation method, characterized in that: include, Collect rehabilitation scene data in the rehabilitation area, generate scene abnormality from the rehabilitation scene data, and issue a health assessment instruction to the outside if the scene abnormality exceeds expectations; Constrain the length of the rehabilitation observation period based on the abnormality of the scene, identify the patient's rehabilitation needs after collecting the patient's rehabilitation status data, and match the patient with the corresponding rehabilitation plan based on the rehabilitation needs; After the rehabilitation robot executes the corresponding rehabilitation plan, if the number of abnormal physical signs of the patient exceeds expectations, the knowledge graph of emergency treatment of abnormal physical signs will give an emergency treatment plan based on the abnormal physical signs, trigger a multi-level alarm mechanism and issue a multi-level alarm command to the outside; After analyzing and obtaining the key factors that affect the patient's rehabilitation, the rehabilitation robot automatically adjusts the environmental factors in the rehabilitation area. If the patient's rehabilitation goal is still not achieved, a three-level alarm command is issued to the outside. The rehabilitation program is optimized according to the re-determined rehabilitation needs, or the rehabilitation robot is comprehensively tested and corresponding test data is obtained, and intervention is performed on the rehabilitation robot according to the test data.

2. The robot-based assisted rehabilitation method according to claim 1, characterized in that: Arrange a sensor network in the rehabilitation area and use the sensor network to collect environmental parameters; The acquired environmental parameters are summarized to generate a rehabilitation scene data set. The rehabilitation scene data in the rehabilitation scene data set is used as input, and the trained scene condition evaluation model is used to output the scene abnormality within the rehabilitation area.

3. The robot-based assisted rehabilitation method according to claim 2, characterized in that: After receiving the health assessment instruction, the length of the rehabilitation observation period cT is constrained according to the abnormality degree Co of several consecutive scenes, and the length of the rehabilitation observation period is adjusted according to the constraint conditions. The constraint method is as follows: Weight coefficient, 0≤α≤1, 0≤β≤1; n is the number of time nodes, Co i is the abnormality of the scene at the i-th time node, Co a is its average value.

4. The robot-based assisted rehabilitation method according to claim 3, characterized in that: Continuously monitor and collect the patient's physical signs, mental health data, and behavioral data in the current rehabilitation scenario, and summarize them as the patient's rehabilitation status data; use the patient's rehabilitation status data as input, use the trained rehabilitation needs identification model to identify rehabilitation needs, and match the patient with the corresponding rehabilitation plan based on rehabilitation needs; In the patient rehabilitation area, the rehabilitation robot executes the corresponding rehabilitation plan and collects the patient's vital sign data after a preset rehabilitation observation period; using the patient's vital sign data as input, the trained abnormal indicator recognition model is used to identify the patient's abnormal vital sign indicators; If the number of abnormal vital signs of the patient exceeds expectations, a first-level alarm command will be issued to the outside.

5. The robot-based assisted rehabilitation method according to claim 4, characterized in that: After receiving the first-level alarm command, the emergency treatment of abnormal physical signs of rehabilitation patients is used as the target word, and a knowledge graph for emergency treatment of abnormal physical signs is built in advance; the abnormal physical sign indicators of patients during the rehabilitation observation period are collected, and the rehabilitation robot sends the emergency treatment plan to the cloud.

6. The robot-based assisted rehabilitation method according to claim 5, characterized in that: After executing the emergency response plan, if the number of consecutive first-level alarm instructions received exceeds the expected number, a risk value R is generated based on the state data of the first-level alarm instructions received. t ; If the risk value R is obtained t If the risk threshold is exceeded, a secondary alarm command will be issued to the outside world, and the patient's recovery degree data, vital signs data and environmental conditions data will be recorded.

7. The robot-assisted rehabilitation method according to claim 6, characterized in that: Under dimensionless conditions, the risk value R is generated based on the status data of the first-level alarm command received. t The way is as follows: Where: T i is the time interval between the ith and i+1st alarms, N i is the number of abnormal indicators at the time of the i-th alarm, γ is the time attenuation coefficient, α and β are weight coefficients, and γ is the attenuation coefficient of the cumulative impact.

8. The robot-based assisted rehabilitation method according to claim 7, characterized in that: After receiving the secondary alarm command, the principal component analysis is used to analyze the environmental condition data in the rehabilitation area to obtain the key factors that affect the patient's rehabilitation, and the corresponding qualified interval is set for each key factor according to the degree of influence of the key factor; During the rehabilitation observation period, various vital signs of the patient are collected as feedback data, and the feedback data is used as input. The trained patient rehabilitation evaluation model is used to output the rehabilitation value. If the obtained rehabilitation value is lower than the rehabilitation threshold, a three-level alarm instruction is issued to the outside.

9. The robot-based assisted rehabilitation method according to claim 8, characterized in that: If the third-level alarm command is not received, the patient's rehabilitation status data is collected again, the patient's rehabilitation needs are identified again, the rehabilitation plan is optimized according to the re-determined rehabilitation needs, and the rehabilitation robot executes the optimized rehabilitation plan; If a third-level alarm command is received, the rehabilitation robot is comprehensively inspected and corresponding inspection data is obtained, which are summarized to generate a rehabilitation robot inspection data set; a control coefficient is generated from the inspection data of the rehabilitation robot inspection data set. If the control coefficient does not exceed the preset response threshold, the rehabilitation robot stops the current task and the robot operation is guided or directly controlled through remote control.

10. A robot-based assisted rehabilitation system, characterized in that: include, The scene analysis unit collects the rehabilitation scene data in the rehabilitation area, generates the scene abnormality degree from the rehabilitation scene data, and issues a health assessment instruction to the outside if the scene abnormality degree exceeds expectations; The rehabilitation program generation unit constrains the length of the rehabilitation observation period according to the abnormality of the scene, identifies the patient's rehabilitation needs after collecting the patient's rehabilitation status data, and matches the patient with a corresponding rehabilitation program based on the rehabilitation needs; Multi-level alarm unit: after the rehabilitation robot executes the corresponding rehabilitation plan, if the number of abnormal physical signs of the patient exceeds the expected number, the knowledge graph of emergency treatment of abnormal physical signs will give an emergency treatment plan based on the abnormal physical signs, trigger the multi-level alarm mechanism and issue a multi-level alarm command to the outside world; The adjustment unit analyzes and obtains the key factors that affect the patient's rehabilitation. The rehabilitation robot automatically adjusts the environmental factors in the rehabilitation area. If the patient's rehabilitation goal is still not achieved, a three-level alarm command is issued to the outside. The feedback processing unit optimizes the rehabilitation plan according to the re-determined rehabilitation needs, or performs a comprehensive test on the rehabilitation robot and obtains corresponding test data, and intervenes in the rehabilitation robot according to the test data.

Images