Intelligent fabric shoulder compensation monitoring system for rehabilitation treatment of stroke patient

Abstract

The invention discloses an intelligent fabric shoulder compensation monitoring system for rehabilitation treatment of a stroke patient. The intelligent fabric shoulder compensation monitoring system comprises a sensor, a fixing carrier part used for fixing the sensor, a signal acquisition module in signal connection with the sensor and a PC in signal connection with the signal acquisition module. The plurality of sensors are fixedly connected to the fixed carrier part, and the sensors are fabric stretching sensors; the fixing carrier piece is made of fabric, and the fixing carrier piece detachably wraps the single-side shoulder and the shoulder blade of the patient. According to the invention, the system can timely discover the compensatory exercise of the patient and remind the patient, so that the fitness compensatory is reduced and the exercise rehabilitation training quality is improved.

Description

Technical Field

[0001] This invention relates to the technical field of shoulder dynamic monitoring, and in particular to an intelligent fabric shoulder compensation monitoring system for the rehabilitation treatment of stroke patients. Background Technology

[0002] Related studies have shown that stroke patients can effectively improve their physical function and regain basic living abilities through rehabilitation training in the early stages after passing the critical period. In recent years, aerobic exercise rehabilitation training after stroke has received increasing attention from clinicians and researchers. There are three key areas of focus in the rehabilitation training process: effective assessment of movement impairment, accurate assessment of treatment effectiveness, and long-term rehabilitation training for sustained recovery in the chronic phase six months after stroke. These three aspects of assessment affect the overall effectiveness of stroke rehabilitation treatment. During the patient's rehabilitation process, due to decreased limb control on the affected side, patients often use muscle groups from the healthy side or the affected side to complete training tasks, resulting in compensatory movements. This causes patients to perform tasks using movement patterns they never had before, such as substituting functionally impaired joint freedom with more functional joint freedom, abnormal muscle activation patterns, abnormal movement sequences, and delayed or lagging movements of adjacent joints. Compensatory movements can seriously affect the effectiveness of rehabilitation training.

[0003] Therefore, from the perspective of the rehabilitation training sites, upper limb motor dysfunction is the main reason for the reduced quality of life of patients; from the perspective of the quality of rehabilitation training, compensatory behaviors during the training process are an important factor affecting the training effect. In order to improve the effectiveness of rehabilitation training, it is necessary to reduce compensatory behaviors, and the first priority is to monitor and detect compensatory behaviors in advance.

[0004] In order to better meet individual differences and achieve high-standard, quantifiable training results in the rehabilitation training process, rehabilitation robots have emerged in recent years to help patients with rehabilitation training. As a supplement to medical staff, this method provides patients with a multi-functional and automated rehabilitation training mode, and the results of various clinical trials have also proved its effectiveness.

[0005] Simultaneously, with the continuous innovation and upgrading of human body sensing and smart wearable devices, these technologies are being increasingly widely applied in the rehabilitation training of stroke patients. Their advantages are mainly reflected in accuracy, participation, and intelligence. These characteristics have led to the increasingly widespread use of this technology in various aspects of stroke treatment, including functional assessment, posture control, functional training, ADL monitoring, and compensation analysis. Not only can medical staff use the monitoring and data from wearable devices to make a comprehensive evaluation and thus a more objective judgment of the patient's health status, but patients themselves can also use the real-time feedback information provided by wearable devices to guide their various training processes, making the training more effective and improving the patient's enthusiasm and efficiency in training. Furthermore, analyzing and comparing the patient's rehabilitation training parameters can not only provide a basis for developing more targeted training programs but also be applied to research on the clinical treatment efficacy of stroke and the exchange of rehabilitation training techniques.

[0006] Limb movement monitoring during rehabilitation training is crucial for ensuring training effectiveness, and monitoring human posture using various technologies is currently the most challenging aspect. Traditional rehabilitation training is mostly conducted with the assistance of medical therapists and large-scale rehabilitation equipment. Monitoring compensatory behaviors during upper limb rehabilitation training for stroke patients primarily relies on manual and machine monitoring. Manual monitoring often depends on the experience of rehabilitation therapists, requiring significant manpower. Machine monitoring currently commonly uses optical motion capture systems and accelerometers. Optical motion capture systems are extremely expensive and have specific requirements regarding location, time, and environmental conditions, making widespread application difficult. While accelerometers are small and relatively easy to use, they have poor accuracy in capturing complex joint movements and lack comfort.

[0007] Defects and shortcomings of existing technology:

[0008] In summary, the following limitations exist in monitoring motor posture during the rehabilitation training process for stroke patients:

[0009] 1. Manual observation: large empirical error.

[0010] Monitoring compensatory behaviors during upper limb rehabilitation training for stroke patients primarily relies on manual and machine monitoring. While these methods can effectively and quickly complete treatment on the affected side, they are costly, offer little autonomy, are monotonous and tedious, require significant human resources, and struggle to provide real-time and effective measurement and evaluation of rehabilitation outcomes. Furthermore, manual observation of movement angles requires specialized knowledge and experience and is subject to certain empirical errors.

[0011] 2. Light capture system: High cost (human, material, and financial resources).

[0012] Currently, machine monitoring mainly relies on optical capture systems and multi-axis accelerometers. Among these, optical capture systems are more mature than wearable technologies, but they are expensive, require specific locations, and are not user-friendly for patients with limited mobility.

[0013] 3. Accelerometer: It feels intrusive and out of place.

[0014] Accelerometers have limited ability to capture complex upper limb movements, especially multi-degree-of-freedom shoulder joint movements, and they can feel intrusive and uncomfortable for patients to wear or use for extended periods.

[0015] 4. Watches, wristbands and other wearable devices: poor comfort and lack of stability and accuracy in long-term monitoring.

[0016] Traditional physiological data collection is uncomfortable for users because they must wear multiple sensors or related devices. If only simple devices (such as smart bracelets and smartwatches) are worn, the collected data is inaccurate. Users may experience health problems while wearing body sensors, which can further lead to stress and other negative emotions. More seriously, the negative psychological effects can lead to some mental illnesses, especially when the patient feels lonely or depressed.

[0017] 5. Existing monitoring system: lacks long-term monitoring and effective evaluation, and has weak feedback function.

[0018] Assessment builds upon monitoring, allowing for more in-depth analysis. Monitoring is often a continuous process, while assessment is typically periodic, requiring greater analysis and reflection. In traditional rehabilitation training, therapists need to continuously monitor and provide feedback and adjustments to the patient's training progress. However, rehabilitation training cycles are generally long, and situations where therapists cannot provide timely feedback are inevitable. Furthermore, some standardized and basic training processes require systematic support and safeguards. Summary of the Invention

[0019] To address the shortcomings of existing technologies, the present invention aims to provide an intelligent fabric shoulder compensation monitoring system for stroke patient rehabilitation, which can promptly detect compensatory movements in patients, provide reminders, reduce compensatory movements, and improve the quality of exercise rehabilitation training. To achieve the above-mentioned objectives and other advantages of the present invention, an intelligent fabric shoulder compensation monitoring system for stroke patient rehabilitation is provided, comprising:

[0020] A sensor, a mounting carrier for fixing the sensor, a signal acquisition module connected to the sensor, and a PC connected to the signal acquisition module;

[0021] Multiple sensors are fixedly connected to the fixed carrier component, and the sensors are fabric tension sensors.

[0022] The fixation carrier is made of fabric and is detachably wrapped around one side of the patient's shoulder and scapula.

[0023] Preferably, the fixing carrier includes an intermediate component, the upper end of which is integrally connected to a first fixing component, and the lower end of which is integrally connected to a second fixing component. The second fixing component is fixed under one armpit, and the first fixing component passes through the other armpit for fixing.

[0024] Preferably, at least three sensors are fixedly attached to the fixed carrier.

[0025] Preferably, the fabric stretch sensor is a resistive fabric stretch braided tape incorporating silver fibers, fixed to a fixed carrier, and the shoulder movement angle is obtained by the length change of the fabric stretch sensor.

[0026] Preferably, the sensor is sewn onto the fixed carrier using a "Z" stitch.

[0027] Preferably, when the sensor and the fixed carrier are used as sensing media, the direction of motion is obtained by the stretching direction represented by the sensor's position, and the angle is obtained by the motion amplitude represented by the change in length.

[0028] Motion coordination is determined by analyzing the rate of stretching change of the sensors.

[0029] Preferably, the upper limb movement angle is obtained by spatial calculation using optical motion capture system technology. The limb angle data collected by the optical capture system and the fabric sensing stretch amount are used to establish a sequence prediction model through machine learning. The fabric sensing amount is then mapped to the limb movement direction and angle through the sequence prediction model, so as to realize a rehabilitation-meaning assessment of the compensation situation.

[0030] Compared with the prior art, the beneficial effects of this invention are:

[0031] 1. This invention can replace manual monitoring to a certain extent. Based on fabric sensing data and deep learning algorithms, it can achieve high-precision capture of shoulder movements, intelligently monitor physical compensation during rehabilitation training, and does not require much manual intervention.

[0032] 2. The cost of motion capture in this invention is far lower than that of optical capture systems, and it has no special requirements for the operating environment, making it more acceptable to a wider range of patients.

[0033] 3. This invention uses fabric as a base material, offering high privacy and comfort, reducing the invasiveness and psychological stress caused to patients by various monitoring devices during rehabilitation training, making it suitable for long-term use during rehabilitation. For complex shoulder joints, smart fabric can achieve more precise activity monitoring, which is impossible with traditional accelerometers.

[0034] 4. This invention has an inherent advantage in supporting task-oriented repetitive motion monitoring and in long-term patient activity monitoring. It can monitor training effects and compensatory behaviors over a long period of time, and provide patients with more proactive and timely rehabilitation suggestions and training assessments based on the monitoring results. Attached Figure Description

[0035] Figure 1 This is a schematic diagram of the intelligent fabric shoulder compensation monitoring system for the rehabilitation treatment of stroke patients according to the present invention.

[0036] Figure 2 This is a front view of the intelligent fabric shoulder compensation monitoring system for rehabilitation treatment of stroke patients according to the present invention when worn.

[0037] Figure 3 This invention relates to a smart fabric shoulder compensation monitoring system for the rehabilitation treatment of stroke patients, and includes a fabric stretch sensor and a sewing effect diagram.

[0038] Figure 4 This is a schematic diagram of data transmission for the intelligent fabric shoulder compensation monitoring system for the rehabilitation treatment of stroke patients according to the present invention.

[0039] Figure 5 This is a diagram showing the arrangement of shoulder monitoring fabric sensors and the connection of fabric circuits in an intelligent fabric shoulder compensation monitoring system for the rehabilitation treatment of stroke patients, according to the present invention.

[0040] Figure 6 The following are rear and side views of the intelligent fabric shoulder compensation monitoring system for rehabilitation treatment of stroke patients according to the present invention when worn.

[0041] Figure 7 This is a schematic diagram of the training and processing of shoulder compensation monitoring fabric sensor data and the CTEN model of the intelligent fabric shoulder compensation monitoring system for the rehabilitation treatment of stroke patients according to the present invention. Detailed Implementation

[0042] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0043] Reference Figure 1 A smart fabric shoulder compensation monitoring system for the rehabilitation treatment of stroke patients, comprising:

[0044] The system comprises a sensor, a mounting carrier for fixing the sensor, a signal acquisition module connected to the sensor, and a PC connected to the signal acquisition module. The signal acquisition module uses an Arduino Mega development board, which can be powered by an external USB connection or a battery pack and has good expandability. It acquires and transmits the raw stretching signal through a microprocessor. The Arduino Mega is an Arduino development board designed based on the ATmega2560 chip, featuring 54 digital input / output pins, including 15 for PWM output, 16 analog input pins, 4 UART serial interfaces, a 16MHz crystal oscillator, a USB interface, a DC interface, an ICSP interface, and a microcontroller reset button. The Arduino Mega possesses all the basic pin functions of a microcontroller. In practical applications, simply connecting the development board to a computer via USB, an AC-DC adapter, or a battery pack will enable its normal operation. The Arduino Mega offers more I / O ports than the Arduino UNO, but its form factor and functionality are almost identical to the Arduino UNO. This allows developers to use all sensors compatible with the UNO board, and the programming is also universal, offering advantages such as quick learning and wide applicability. For this application, 5-8 analog input ports are required, which the Mega controller can meet, and it also has good scalability. The system uses a CTEN (Channel-Temporal Encoding Network) model based on a deep learning architecture using Transformer and LSTM to map fabric sensor signal data to actual joint angles. This model treats shoulder joint angle estimation as a multivariate time series regression problem, establishing a mapping relationship between the fabric sensor electrical signals and the actual joint angles.The model consists of three modules, and the results are shown in the figure: (1) Channel-wise Encoder (CwE), which adopts the encoder block in the original Transformer architecture, explicitly captures the correlation of each sensor channel at each time step through the attention mechanism, selectively emphasizes or suppresses specific channels for the signal change differences of resistive sensors in specific actions (some sensors are sensitive and some contain noise), and optimizes the sensor signal weight allocation; (2) Temporal-wise Encoder (TwE), which is built based on Long Short-Term Memory (LSTM) network, learns the temporal dynamic characteristics and hysteresis effect of sensor signals, solves the diversity and hysteresis problem of sensor data in the time dimension, and captures the temporal dependence of joint movement; (3) Shoulder joint angle representation decoder, which consists of two fully connected layers, decodes the feature representation output by the encoder into joint angle values, and uses the mean squared error (MSE) loss function to minimize the error between the predicted angle and the true angle.

[0045] Multiple sensors are fixed to the fixed carrier, and these sensors are fabric tension sensors. For shoulder movement, joint motion manifests as skin deformation on the human body surface. Correspondingly, in close-fitting clothing, the tensile strain of the fabric can indirectly reflect changes in limb angles; therefore, length is chosen as the variable for external motion perception. As the system's data acquisition medium, the fabric sensor has the ability to convert shoulder movement into quantifiable physical parameters, typically sensing changes in length, pressure, and temperature.

[0046] The fixation carrier is made of fabric and is detachably designed to wrap around one side of the patient's shoulder and scapula. The fixation carrier is a one-piece structure. It includes a middle component, with a first fixation assembly integrally connected to its upper end and a second fixation assembly integrally connected to its lower end. The second fixation assembly is secured under one armpit, and the first fixation assembly passes under the other armpit for further fixation. Both the first and second fixation assemblies are secured under both armpits using Velcro, making the device easy to put on and take off, lightweight, breathable, and providing a good fit.

[0047] Considering that stroke patients often suffer from hemiplegia, and that daily upper limb rehabilitation training is primarily unilateral, the wearable prototype for the upper limb is designed as a one-piece shoulder wrap for easy on and off. Figure 2 , Figure 6 As shown, a breathable and elastic fabric is selected, and Velcro is used to fix it under both armpits. The increased back area can wrap around the scapula and support more precise monitoring of the movement of the humerus and scapula.

[0048] Furthermore, at least three sensors are fixedly attached to the fixed carrier. In this application, the fabric sensor consists of several fabric tension sensors. Based on the movement principle of the shoulder joint and objective constraints, the sensors are rationally arranged. The fiber material of the wires connecting the fabric tension sensors and the controller is selected as a flexible, low-resistance, and small-diameter No. 6 composite fiber material with a resistance of 100Ω / M. The positions of the fabric sensors and the fabric circuit connection materials are as follows... Figure 5 As shown, each sensor is connected in series with a 220Ω load resistor, and then they are all connected in parallel to a 5V power supply circuit. A pin is led out from the non-grounded end of the sensor and connected to the analog port of the Arduino board to collect the voltage change across the two ends caused by the resistance change of the fabric sensor due to the change in length. The actual resistance of the stretch sensor is then obtained through the voltage divider formula. This resistance is used for subsequent machine learning with the angle calculated by the optical motion capture system to obtain a model between the resistance change and the actual direction of movement angle.

[0049] Furthermore, the fabric stretch sensor employs a resistive fabric stretch braided tape incorporating silver fibers, fixed to a fixed carrier. It exhibits near-linear resistance change above 117%, a wide resistance variation range, high sensitivity, low hysteresis, high cycle stability, lightweight design, comfort, economy, and scalability. The shoulder movement angle is obtained through the change in length of the fabric stretch sensor. Based on the characteristics of shoulder movement, the sensor is sewn to the fixed carrier using a "Z" stitch to achieve optimal motion capture.

[0050] Furthermore, when the sensor and the fixed carrier are used as sensing media, the direction of motion is obtained by the stretching direction represented by the sensor's position, and the angle is obtained by the motion amplitude represented by the change in length.

[0051] Motion coordination is derived by analyzing the rate of stretching change of sensors. Monitoring shoulder compensatory activities includes monitoring the degrees of freedom, range of motion, and motion coordination in terms of external form. Considering the principles of shoulder movement, for the degrees of freedom and range of motion, when using smart fabric as the sensing medium, the direction of movement is obtained by the stretching direction represented by the placement position, and the angle is obtained by the range of motion represented by the change in length. The motion coordination index can be derived post-hocly through analysis of the rate of stretching change of the fabric sensor.

[0052] Furthermore, in order to achieve a rehabilitative assessment of the compensation situation, data processing is required between the fabric-sensor variables and the compensation monitoring indicators. The relationship between these two needs to be determined by using optical motion capture system technology to spatially calculate the upper limb movement angle, establish a sequence prediction model by machine learning based on the limb angle data collected by the optical capture system and the fabric-sensor stretching amount, and map the fabric-sensor quantities into limb movement directions and angles through the sequence prediction model, thereby achieving a rehabilitative assessment of the compensation situation.

[0053] An optical motion capture system was used to synchronously acquire shoulder joint angles. Based on anatomical knowledge, optical motion capture markers were placed, and coordinate systems for the sternoclavicular joint, scapulothoracic joint, and glenohumeral joint were established for spatial calculation of joint angles. The angle data and sensor resistance data from the training samples underwent data quality preprocessing, resampling, and filtering. Using the CTEN model, the data acquired by the fabric sensor was correlated with the data from the optical motion capture system, achieving a mapping from fabric sensor monitoring values ​​to the scapular and humeral movement angles. Validation on a test dataset showed an average absolute angle error (RMSE) of 2.59°, verifying the feasibility of this invention for monitoring compensatory activities in complex shoulder joints.

[0054] The number of devices and processing scale described herein are for simplification of the invention. Applications, modifications, and variations of this invention will be readily apparent to those skilled in the art. Although embodiments of the invention have been disclosed above, they are not limited to the applications listed in the specification and embodiments. It can be applied to various fields suitable for this invention, and further modifications can be readily implemented by those skilled in the art. Therefore, without departing from the general concept defined by the claims and their equivalents, this invention is not limited to the specific details and illustrations shown and described herein.

Claims

1. A smart fabric shoulder compensation monitoring system for rehabilitation treatment of stroke patients, characterized in that, include: A sensor, a mounting carrier for fixing the sensor, a signal acquisition module connected to the sensor, and a PC connected to the signal acquisition module; Multiple sensors are fixedly connected to the fixed carrier component, and the sensors are fabric tension sensors. The fixation carrier is made of fabric and is detachably wrapped around one side of the patient's shoulder and scapula.

2. The intelligent fabric shoulder compensation monitoring system for rehabilitation treatment of stroke patients as described in claim 1, characterized in that, The fixing carrier includes an intermediate component, the upper end of which is integrally connected to a first fixing component, and the lower end of which is integrally connected to a second fixing component. The second fixing component is fixed under one armpit, and the first fixing component passes through the other armpit for fixing.

3. The intelligent fabric shoulder compensation monitoring system for rehabilitation treatment of stroke patients as described in claim 2, characterized in that, At least three sensors are fixedly attached to the fixed carrier.

4. The intelligent fabric shoulder compensation monitoring system for rehabilitation treatment of stroke patients as described in claim 1, characterized in that, The fabric stretch sensor uses a resistive fabric stretch braided tape incorporating silver fibers, which is fixed to a fixed carrier. The movement angle of the shoulder is obtained by measuring the length change of the fabric stretch sensor.

5. The intelligent fabric shoulder compensation monitoring system for rehabilitation treatment of stroke patients as described in claim 3, characterized in that, The sensor is sewn onto the fixed carrier using a "Z" shaped stitch.

6. The intelligent fabric shoulder compensation monitoring system for rehabilitation treatment of stroke patients as described in claim 1, characterized in that, When the sensor and the fixed carrier are used as sensing media, the direction of motion is obtained by the stretching direction represented by the sensor's position, and the angle is obtained by the motion amplitude represented by the change in length. Motion coordination is determined by analyzing the rate of stretching change of the sensors.

7. The intelligent fabric shoulder compensation monitoring system for rehabilitation treatment of stroke patients as described in claim 1, characterized in that, Using optical motion capture system technology, the upper limb movement angle is obtained through spatial calculation. The limb angle data collected by the optical capture system and the fabric sensing stretch amount are used to establish a sequence prediction model through machine learning. The fabric sensing amount is then mapped to the limb movement direction and angle through the sequence prediction model, so as to realize a rehabilitation-meaning assessment of the compensation situation.

Images