3D spine rehabilitation robot based on artificial intelligence and virtual reality

Abstract

We provide a spinal rehabilitation robot that utilizes artificial intelligence (AI) and virtual reality (VR) to analyze a patient's spinal movements in three dimensions, provide real-time feedback, and guide the correction of abnormal movement patterns. A spinal rehabilitation robot according to one aspect of the present invention comprises a shooting module configured to capture an image of a user's spinal movement, an analysis module configured to recognize an abnormal spinal movement pattern by analyzing the user's spinal posture from the captured image of the user's spinal movement, and a feedback module configured to provide feedback information to the user regarding the abnormal spinal movement pattern.

Description

Technology Field

[0001] The present invention relates to a spinal rehabilitation robot used for spinal rehabilitation treatment, and more specifically, to a spinal rehabilitation robot that utilizes artificial intelligence (AI) and virtual reality (VR) to analyze a patient's spinal movements in three dimensions, provides real-time feedback, and guides the correction of abnormal movement patterns. Background Technology

[0002] In modern society, spinal disorders are on the rise due to factors such as sedentary lifestyles, poor posture, and a lack of physical activity. In particular, the increased use of digital devices has exacerbated issues such as forward head posture, lower back pain, and scoliosis, highlighting the importance of rehabilitation treatment.

[0003] Traditional spinal rehabilitation has relied on manual procedures by physical therapists or the use of mechanical assistive devices to measure and correct limited movements. However, these traditional methods have limitations in precisely analyzing the patient's condition and make it difficult to provide individualized treatment.

[0004] Recently, with the active convergence of medical and IT technologies, new digital-based solutions are being introduced in the field of spinal rehabilitation. In particular, motion analysis systems utilizing 3D motion capture technology have been developed; however, most of these systems remain limited to simple data collection and reporting functions, showing limitations in providing real-time feedback or maximizing treatment effectiveness.

[0005] Furthermore, while some studies have demonstrated that rehabilitation therapy using virtual reality (VR) environments is effective in boosting patient motivation, VR-based rehabilitation systems still lack the capability to provide physical feedback and face many technical limitations that prevent their widespread use in actual medical settings. The problem to be solved

[0006] The present invention was devised to solve the aforementioned problems, and aims to provide a spinal rehabilitation robot that utilizes artificial intelligence (AI) and virtual reality (VR) to analyze a patient's spinal movement in three dimensions, provides real-time feedback, and guides the correction of abnormal movement patterns.

[0007] However, the technical problems that the present invention aims to solve are not limited to those described above, and other unmentioned problems will be clearly understood by those skilled in the art from the description of the invention below. means of solving the problem

[0008] A spinal rehabilitation robot according to one aspect of the present invention comprises a shooting module configured to capture an image of a user's spinal movement, an analysis module configured to recognize an abnormal spinal movement pattern by analyzing the user's spinal posture from the captured image of the user's spinal movement, and a feedback module configured to provide feedback information to the user regarding the abnormal spinal movement pattern.

[0009] Preferably, the imaging module may be configured to sense at least one of the anterior-posterior bending, lateral bending, rotation angle, and alignment state of the user's spine posture when capturing an image of the user's spine movement.

[0010] Preferably, the analysis module is configured to analyze the user's spinal movement in conjunction with the person's bilateral muscle activity and weight distribution data, and may be configured to generate a correction signal if the user's spinal movement is asymmetrical to the left or right based on the results of analyzing the user's spinal movement in conjunction with the person's bilateral muscle activity and weight distribution data.

[0011] Preferably, the analysis module may be configured to adjust the intensity of the correction signal according to a movement evaluation model having user fatigue, spinal pain level, and environmental factors as weighting factors.

[0012] Preferably, the feedback module may be configured to provide the feedback information based on a VR simulation that responds to the user's spine movement.

[0013] Preferably, the feedback module is configured to provide feedback information by adjusting the intensity of the VR simulation by assigning weights to environmental factors around the spinal rehabilitation device and the individual physical condition of the user, and the analysis module may be configured to optimize the user's spinal movement pattern by analyzing the user's spinal movement pattern according to the feedback information.

[0014] Preferably, the analysis module is configured to learn the user's spinal movements for a certain period, and can be configured to adjust the intensity of the correction signal when the user's movement speed is below a predetermined value during a specific movement. Effects of the invention

[0015] According to an embodiment of the present invention, by utilizing artificial intelligence (AI) and virtual reality (VR), the user's spinal movement can be analyzed in three dimensions and feedback information regarding abnormal spinal movement patterns can be provided in an immersive manner. Accordingly, there is an advantage in being able to provide more effective and optimized spinal rehabilitation guidelines to spinal patients.

[0016] In addition, various other additional effects may be achieved by various embodiments of the present invention. These various effects of the present invention are described in detail in each embodiment, or the description of effects that are easily understood by those skilled in the art is omitted. Brief explanation of the drawing

[0017] The following drawings attached to this specification illustrate preferred embodiments of the present invention and serve to further enhance understanding of the technical concept of the present invention together with the detailed description of the invention provided below; therefore, the present invention should not be interpreted as being limited only to the matters described in such drawings. FIG. 1 is a drawing showing a spinal rehabilitation robot according to one embodiment of the present invention. Figure 2 is a diagram exemplifying the method of analyzing user movement and posture and providing feedback by the spinal rehabilitation robot of Figure 1. FIG. 3 is a diagram showing an example of operation of a spinal rehabilitation robot according to a second embodiment of the present invention. FIG. 4 is a diagram showing an example of operation of a spinal rehabilitation robot according to the third embodiment of the present invention. FIG. 5 is a diagram showing an example of operation of a spinal rehabilitation robot according to the fourth embodiment of the present invention. Specific details for implementing the invention

[0018] Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the attached drawings. Prior to this, terms and words used in this specification and claims should not be interpreted as being limited to their ordinary or dictionary meanings, and should be interpreted in a meaning and concept consistent with the technical spirit of the present invention, based on the principle that the inventor can appropriately define the concept of the terms to best describe his invention.

[0019] Therefore, the embodiments described in this specification and the configurations illustrated in the drawings are merely the most preferred embodiments of the present invention and do not represent all of the technical ideas of the present invention; thus, it should be understood that various equivalents and modifications that can replace them may exist at the time of filing this application.

[0020] FIG. 1 is a drawing showing a spinal rehabilitation robot (10) according to one embodiment of the present invention, and FIG. 2 is a drawing exemplarily showing a method of analyzing user movement and posture and providing feedback by the spinal rehabilitation robot (10) of FIG. 1.

[0021] Referring to FIGS. 1 and 2, a spinal rehabilitation robot (10) according to one embodiment of the present invention may include a shooting module (100), an analysis module (200), and a feedback module (300). This spinal rehabilitation robot (10) may be a medical device for treating a user's scoliosis. In this spinal rehabilitation robot (10), scoliosis correction may be achieved by applying a pushing force or a pulling force to the spinal portion while the user is seated on the seating portion (S) and the user is gripping and pulling an exercise device (e.g., a handle) with their hand.

[0022] The above-described shooting module (100) may be configured to capture images of the user's spine movements. In one embodiment, the shooting module (100) may be a 3D camera. Additionally, the shooting module (100) may be configured to consist of multiple units to capture stereoscopic images of the user's spine movements at various positions.

[0023] Specifically, the shooting module (100) may be configured to sense at least one of the forward and backward bending, lateral bending, rotation angle, and alignment state of the user's spine posture when shooting an image of the user's spine movement.

[0024] The above analysis module (200) may be configured to recognize abnormal spinal movement patterns by analyzing the user's spinal posture from a video of the user's spinal movement captured.

[0025] In one embodiment, the analysis module (200) may include an AI model. The AI ​​model may be trained based on an unsupervised learning method, but is not limited thereto, and may also be trained based on supervised learning or semi-persistent learning methods. Additionally, the AI ​​model may be composed of one or more neural network layers, and each neural network may include one or more weights. The AI ​​model may perform learning / inference by performing operations between one or more weights and input data. As an example, the AI ​​model may perform image analysis based on a Deep Neural Network (DNN), a Convolutional Neural Network (CNN), etc.

[0026] The above feedback module (300) may be configured to provide feedback information to the user regarding the user's abnormal spinal movement pattern. In this case, the feedback information regarding abnormal spinal movement patterns may include information about movement patterns that the user needs to improve compared to normal spinal movement patterns.

[0027] Specifically, the feedback module (300) may be configured to provide feedback information based on a VR simulation that responds to the user's spinal movement. In one embodiment, the spinal rehabilitation robot (10) may be equipped with a VR simulation device (not shown). The user may receive feedback information regarding the user's abnormal spinal movement patterns through the VR simulation device by accessing a virtual reality (VR) environment using a headset device connected to the VR simulation device. Through this, the user may receive feedback information while feeling immersed during the spinal rehabilitation process.

[0028] According to this embodiment of the present invention, artificial intelligence (AI) and virtual reality (VR) are utilized to analyze a user's spinal movement in three dimensions, and It can provide immersive feedback information on abnormal spinal movement patterns. Accordingly, this has the advantage of providing more effective and optimized spinal rehabilitation guidelines for spinal patients.

[0029] FIG. 3 is a diagram showing an example of operation of a spinal rehabilitation robot (10) according to a second embodiment of the present invention.

[0030] Referring to FIG. 3, the analysis module (200) can be configured to analyze the user's spinal movement in conjunction with the person's bilateral muscle activity and weight distribution data.

[0031] In particular, the analysis module (200) may be configured to generate a correction signal when the user's spinal movement is asymmetrical to the left and right, based on the result of analyzing the user's spinal movement in conjunction with the person's bilateral muscle activity and weight distribution data. At this time, the correction signal may be a type of feedback information.

[0032] Specifically, data on bilateral muscle activity and body weight distribution may vary depending on whether the user's spinal movement involves anterior-posterior flexion, lateral flexion, or torso rotation. This is because the user's vertical height or horizontal position differs during exercise for each type of spinal movement.

[0033] In one embodiment, when the user's spine movement is forward and backward bending, the analysis module (200) may generate the correction signal by giving more weight to the weight distribution data than to the person's bilateral muscle activity when generating the correction signal. This is because when the user's spine moves forward or backward, the center of gravity shifts toward the front of the body compared to lateral bending or torso rotation, resulting in a greater imbalance in the user's weight distribution than in lateral bending or torso rotation.

[0034] Additionally, when the user's spinal movement is lateral bending, the analysis module (200) may generate the correction signal by giving more weight to the bilateral muscle activity than to the person's weight distribution data when generating the correction signal. This is because when the user's spinal movement is lateral bending, the user's lateral muscle movement is more active than when it is anterior-posterior bending.

[0035] Additionally, when the user's spinal movement is torso rotation, the analysis module (200) may generate the correction signal by giving more weight to the bilateral muscle activity than to the person's weight distribution data when generating the correction signal. This is because when the user's spinal movement is torso rotation, the user's lateral muscle movement following the rotation is more active than when the user's spinal movement is forward or backward bending.

[0036] According to this embodiment, rather than simple spinal posture correction, a more accurate spinal correction guide can be provided to the user by analyzing the user's spinal movement in conjunction with data on bilateral muscle activity and body weight distribution.

[0037] FIG. 4 is a diagram showing an example of operation of a spinal rehabilitation robot (10) according to the third embodiment of the present invention.

[0038] Referring to FIG. 4, the analysis module (200) may be configured to adjust the intensity of the correction signal according to a movement evaluation model (a type of AI model) having the user's fatigue level, spinal pain level, and environmental factors as weighting factors. At this time, the environmental factors may include inclination information of the ground where the spinal rehabilitation robot (10) is located or ground condition information.

[0039] Specifically, the user's fatigue and spinal pain levels may vary depending on whether the spinal movement involves anterior-posterior bending, lateral bending, or torso rotation. This is because the user's vertical height or horizontal position during exercise differs for each type of spinal movement. Furthermore, the influence of environmental factors on the user's spinal movement may vary depending on whether the movement involves anterior-posterior bending, lateral bending, or torso rotation. This is because the user's vertical height or horizontal position during exercise differs for each type of spinal movement.

[0040] In one embodiment, when the user's spinal movement is forward-backward bending, the analysis module (200) may adjust the intensity of the correction signal by giving more weight to the user's fatigue and environmental factors than to the user's spinal pain level when adjusting the intensity of the correction signal. This is because when the user's spinal movement is forward-backward bending, the user's spinal movement becomes larger than when it is lateral bending or torso rotation, so the influence of user fatigue or environmental factors acts more significantly, whereas the pain applied to the user's spine is relatively smaller than when it is lateral bending or torso rotation.

[0041] Additionally, in one embodiment, when the user's spinal movement is lateral bending or torso rotation, the analysis module (200) may adjust the intensity of the correction signal by giving more weight to the user's spinal pain level than to the user's fatigue and environmental factors when adjusting the intensity of the correction signal. This is because when the user's spinal movement is lateral bending or torso rotation, the pain applied to the user's spine due to lateral or rotational movement is relatively greater than when the user's spine is forward or backward bending, while the influence of user fatigue or environmental factors is smaller.

[0042] According to this embodiment, since the intensity of the correction signal can be adjusted according to the user's fatigue, level of spinal pain, and environmental factors rather than simple spinal posture correction, an effective spinal correction guide can be provided to the patient while reducing the burden on the patient.

[0043] FIG. 5 is a diagram showing an example of operation of a spinal rehabilitation robot (10) according to the fourth embodiment of the present invention.

[0044] Referring to FIG. 5, the feedback module (300) may be configured to provide the feedback information by adjusting the intensity of the VR simulation by giving weights to environmental factors around the spinal rehabilitation device (10) and the individual physical condition of the user.

[0045] Specifically, environmental factors surrounding the spinal rehabilitation device (10) may include temperature information, humidity information, and hardness information of the floor surface surrounding the spinal rehabilitation device (10). Additionally, the individual physical condition of the user may include information on the user's muscle strength, flexibility, and spinal pain sensitivity.

[0046] In one embodiment, the feedback module (300) can adjust the intensity of the VR simulation by giving more weight to temperature and humidity information than to floor hardness information among environmental factors around the spinal rehabilitation device (10) when the user shows high fatigue according to the user's spinal movement, and giving more weight to muscle strength information than to user flexibility information and spinal pain sensitivity information among the user's individual physical conditions. This is, This is because when a user exhibits high fatigue, they are more likely to be affected by temperature and humidity relative to the hardness of the floor surface among environmental factors surrounding the spinal rehabilitation device (10), and more likely to be affected by the user's muscle strength than by flexibility and spinal pain sensitivity among the user's individual physical conditions.

[0047] Additionally, the feedback module (300) can adjust the intensity of the VR simulation by giving more weight to the hardness information of the floor surface than to the temperature information and humidity information among the environmental factors around the spinal rehabilitation device (10) when the user shows low fatigue according to the user's spinal movement, and giving more weight to the user's flexibility information and spinal pain sensitivity information than to the muscle strength information among the user's individual physical conditions. This is, This is because when the user shows low fatigue, there is a higher likelihood of being more affected by the hardness of the floor surface compared to the temperature and humidity among the environmental factors surrounding the spinal rehabilitation device (10), and there is a higher likelihood of being more affected by flexibility and spinal pain sensitivity than by the user's muscle strength among the user's individual physical conditions.

[0048] Additionally, the analysis module (200) may be configured to analyze the user's spine movement pattern according to the feedback information and optimize the user's spine movement pattern.

[0049] That is, the analysis module (200) can continuously optimize the user's spinal movement pattern by re-analyzing the user's spinal movement pattern based on the feedback information. Accordingly, the analysis module (200) can continuously improve and reflect a correction signal suitable for the user.

[0050] In one embodiment, the analysis module (200) may be configured to learn the user's spinal movements for a certain period. Additionally, the analysis module (200) may be configured to adjust the intensity of a correction signal when the user's movement speed in a specific movement is found to be below a predetermined value.

[0051] Specifically, the analysis module (200) can learn the user's spinal movement using an AI model for a certain period (e.g., one week or more) and adjust the strength of the correction signal.

[0052] In one embodiment, the analysis module (200) can automatically adjust the strength of the correction signal if the user continuously experiences difficulty with a specific movement or if the recovery speed is slow.

[0053] In addition, the analysis module (200) can guide medical staff to establish a user-customized treatment plan by notifying medical staff of the user's spinal movements learned over a certain period.

[0054] In another embodiment of the present invention, by integrally configuring a complex neural network-based analysis module and a feedback module operating in an AI-based edge computing environment, the real-time capability, precision, adaptability, and immersive feedback ability of a spinal rehabilitation robot can be dramatically improved. The system introduced in this embodiment runs on an NVIDIA Jetson Orin-based AI edge platform and is fully compatible with Python-based deep learning frameworks (PyTorch and TensorFlow). NVIDIA's TensorRT is used as an optimized inference engine for real-time processing, thereby ensuring high-speed parallel computing and energy-efficient processing performance.

[0055] The core artificial intelligence architecture operating in such high-performance edge environments can be broadly classified into six categories. These consist of ① GNN (Graph Neural Network), ② PINN (Physics-Informed Neural Network), ③ DT (Digital Twin), ④ RL (Reinforcement Learning), ⑤ FL (Federated Learning), and ⑥ XAI (eXplainable AI); additionally, ⑦ RAG (Retrieval-Augmented Generation) is also implemented in a convergent manner to enhance patient communication and immersive feedback functions. While each technology can operate independently, it interacts organically as a system to contribute to the analysis of user spinal movements, the provision of feedback, and long-term learning and optimization.

[0056] First, the analysis module utilizes a GNN to precisely model the spatial and temporal interrelationships between the user's spine and surrounding joint data. Since spinal movement is not limited to rotation or bending along a single axis but is organically connected to joints throughout the body, such as the shoulders, pelvis, knees, and ankles, these complex relationships cannot be accurately explained solely by linear regression models or one-dimensional time-series analysis. Accordingly, in this embodiment, spinal-centered multi-joint interconnected movement data is represented in a graph form through a GNN structure, and by quantifying the characteristics of each node (joint point) and edge (motion path), abnormal movement patterns or inter-joint disharmony can be detected in real time. For example, if excessive elevation of the left shoulder is abnormally linked to left rotation of the pelvis, the GNN can infer this imbalance of edge weights and identify it as an abnormal pattern, which can be interpreted as an early sign of scoliosis.

[0057] Such GNN-based interpretation gains greater reliability when combined with physically valid analysis. To this end, the present invention simultaneously introduces a PINN structure to integrate human biomechanical constraints into the neural network learning process. By directly incorporating quantitative physics-based models—such as Newton's laws of motion, the center of rotation of human joints, range of motion limitations, and ligament elastic moduli—into the neural network's loss function, PINN automatically suppresses joint movements that cannot occur in reality (e.g., hyperextension, bending) while continuously verifying whether the predicted movements match the actual patient's kinematics. Through this, the analysis module can identify risk movements that appear normal on the surface but actually involve high joint stress at an early stage, and provides customized analysis by adjusting internal PINN weights according to the individual patient's body type, flexibility, and pain sensitivity.

[0058] These analysis results are reflected in the digital twin model in real time. In this embodiment, the digital twin is not a simple visualization model, but a three-dimensional virtual replica capable of dynamic simulation based on the user's biometric data. When the user is connected to the robot and performs rehabilitation movements, the movements of each joint and spine are projected onto the virtual model in real time. This digital twin provides medical staff with both quantitative explanations and intuitive judgments by visually highlighting problem areas identified by the analysis module (e.g., excessive left rotation, lumbar kyphosis, etc.). For example, if movements deviating from the center of spinal rotation occur three or more times consecutively, the corresponding area is highlighted in a red hue and displayed on the digital twin, which is also shared with the user in real time.

[0059] This structure enables real-time adaptive correction. By applying reinforcement learning (RL), the system evaluates the results of the user's repetitive actions using a reward function and dynamically adjusts the feedback strategy. For example, if the user consistently makes errors in a specific action, or if fatigue increases or speed decreases during repetitive performance, the system maintains the user's learning continuity by lowering the intensity of feedback for that action or replacing it with an easier inducing action. This RL structure goes beyond simple correction repetition to present an optimized learning path for the user and adjusts the sequence and frequency of individual actions to match the user's recovery speed.

[0060] Furthermore, in this embodiment, model performance is enhanced by independently training diverse user data from a central server based on Federated Learning (FL). Individual user data is trained on the corresponding local equipment, and only encrypted weights for privacy protection are transmitted to the server, which integrates them to reconstruct a common model. This approach addresses privacy issues while collecting diversified data from users of different ages, body types, and medical histories, and significantly contributes to strengthening the generalization capabilities of the entire system. In particular, sub-models specialized for specific disease groups (e.g., patients with kyphosis or herniated discs) can be trained to enable medical professionals to make more precise judgments.

[0061] This advanced AI-based analysis and learning structure goes beyond simple technical functions and necessarily incorporates Explainable AI (XAI) to secure the trust of patients and medical staff and increase the likelihood of clinical adoption. The rationale behind the analysis module's decisions is explained through visualization techniques such as SHAP and Grad-CAM. For example, if a correction instruction is issued for a specific movement, the reason—specifically which part of the spine and which joint linkages caused the problem—can be displayed in the form of a visual heatmap; this data can also be utilized as a basis for medical staff judgment. Instead of simply being told that they are "wrong," patients receive specific information, such as "the center of rotation is off-center because the shoulder height is tilted 4 degrees relative to the pelvis." This allows them to enhance their awareness of their own posture and actively participate in corrective movements.

[0062] Finally, the feedback module integrates a Retrieval-Augmented Generation (RAG) structure to dynamically generate patient-specific verbal feedback content. For example, when the analysis module recognizes patterns of improvement or deterioration in specific exercises, it searches relevant historical data and general medical guidelines to generate coaching messages in natural language. These messages can be provided via voice guidance or text display, and the difficulty and length of the sentences are automatically adjusted based on the user's language proficiency, emotional response, and level of concentration. Sentences such as, “Your left lumbar rotation has significantly improved today. Try increasing the right rotation angle a little more in the next session,” serve as a more effective motivational tool than repetitive voice instructions and help patients better understand their own recovery plans.

[0063] This integrated system represents an evolution from existing rehabilitation robot systems centered on static treatment guidance or simple repetitive correction. By implementing biomechanics-based neural network learning, real-time feedback adjustment, user-customized interfaces, and data-driven medical support on a single continuous platform, it can bring substantial changes to the medical field. In particular, since this embodiment of the invention operates based on edge computing, it can operate without delay even in remote environments and offers the advantage of providing effective patient-customized treatment while meeting medical information protection requirements without the need for external network connections.

[0064] As described above, although the present invention has been explained by limited embodiments and drawings, the present invention is not limited thereto, and it is obvious that various modifications and variations are possible within the scope of the technical spirit of the present invention and the equivalent scope of the claims described below by those skilled in the art to which the present invention belongs.

[0065] Meanwhile, although terms indicating directions such as up, down, left, right, front, and back have been used in this invention, these terms are used merely for convenience of explanation, and it is obvious to those skilled in the art that they may vary depending on the position of the object or the position of the observer. Explanation of the symbols

[0066] 10 : Spinal rehabilitation robot 100 : Shooting module 200: Analysis Module 300: Feedback Module

Claims

Claim 1 A shooting module configured to capture video of the user's spinal movement; an analysis module configured to recognize abnormal spinal movement patterns by analyzing the user's spinal posture from the captured video of the user's spinal movement; A spinal rehabilitation robot characterized by comprising a feedback module configured to provide feedback information regarding the abnormal spinal movement pattern of the user to the user, wherein the analysis module is configured to analyze the user's spinal movement in conjunction with data on the user's bilateral muscle activity and weight distribution, and, based on the result of analyzing the user's spinal movement in conjunction with data on the user's bilateral muscle activity and weight distribution, if the user's spinal movement is asymmetrical to the left and right, the analysis module is configured to generate the correction signal by giving more weight to the weight distribution data than to the user's bilateral muscle activity when generating the correction signal, if the user's spinal movement is lateral bending, the analysis module is configured to generate the correction signal by giving more weight to the bilateral muscle activity than to the user's weight distribution data when generating the correction signal, and if the user's spinal movement is torso rotation, the analysis module is configured to generate the correction signal by giving more weight to the bilateral muscle activity than to the user's weight distribution data when generating the correction signal. Claim 2 A spinal rehabilitation robot according to claim 1, characterized in that the above-mentioned imaging module is configured to sense at least one of the anterior-posterior bending, lateral bending, rotation angle, and alignment state of the user's spinal posture when capturing an image of the user's spinal movement. Claim 3 delete Claim 4 A spinal rehabilitation robot according to claim 1, characterized in that the analysis module is configured to adjust the intensity of a correction signal according to a movement evaluation model having user fatigue, spinal pain level, and environmental factors as weighting factors. Claim 5 A spinal rehabilitation robot according to claim 1, characterized in that the feedback module is configured to provide feedback information based on a VR simulation that responds to the user's spinal movement. Claim 6 A spinal rehabilitation robot according to claim 5, wherein the feedback module is configured to provide feedback information by adjusting the intensity of the VR simulation by assigning weights to environmental factors surrounding the spinal rehabilitation robot and the individual physical conditions of the user, and the analysis module is configured to optimize the user's spinal movement pattern by analyzing the user's spinal movement pattern according to the feedback information. Claim 7 A spinal rehabilitation robot according to claim 1, characterized in that the analysis module is configured to learn the user's spinal movements for a certain period, and is configured to adjust the intensity of a correction signal when the user's movement speed in a specific movement is below a predetermined value.

Images