A guided mirror rehabilitation system for hand movement recovery of hemiplegic patients

Abstract

The application provides a guided mirror rehabilitation system for hand movement recovery of hemiplegic patients, comprising a perception glove, a rehabilitation glove and an upper computer software platform; the perception glove comprises a sensor module, a flexible circuit module and an embedded electronic system, and is used for realizing collection of healthy side hand movement information; the rehabilitation glove comprises a sensor module, a flexible circuit module, an embedded electronic system, a port expansion module and a vibration tactile module, and is used for realizing collection of the movement information of the affected side hand; the upper computer software platform is used for receiving the hand movement information from the perception glove, and performing pattern recognition on the current hand gesture through a machine learning engine; the upper computer software platform generates a control signal according to the pattern recognition result, and sends the control signal to the rehabilitation glove; and the application can realize super-fine-grained hand movement information collection, patient subjective movement intention recognition and guided rehabilitation treatment strategy.

Description

Technical Field

[0001] This invention belongs to the field of rehabilitation equipment technology, and in particular relates to a guided mirror rehabilitation system for the recovery of hand movement in hemiplegic patients. Background Technology

[0002] Stroke-induced hemiplegia is a serious and common global health problem. Although recent medical advancements have significantly improved stroke survival rates, symptoms such as hemiplegia still require medical rehabilitation training to achieve motor function recovery. The S400 continuous passive hand training system proposed by Orange Elephant Medical Technology Co., Ltd. is widely used in the rehabilitation of hand motor disorders caused by neurological diseases. This device has a WIFI wireless information transmission module, adopts a task-based therapy approach, and supports single-finger rehabilitation training. However, the S400 continuous passive hand training system cannot perform mirror rehabilitation; in other words, it cannot incorporate the patient's subjective motor intentions into the rehabilitation process. The hand function rehabilitation robot designed by Shanghai Siyi Intelligent Technology Co., Ltd. is a commonly used device for the rehabilitation of hemiplegic patients in China. This device supports mechanical mirror rehabilitation training and is equipped with advanced virtual reality host computer software to increase patient motivation and participation. However, the human hand is extremely flexible, and the aforementioned devices can only achieve coarse-grained gesture recognition, such as detecting only finger flexion and extension movements, but not adduction-abduction movements. Currently available hand rehabilitation devices for hemiplegic patients have the following shortcomings:

[0003] (1) Most wearable upper limb motor rehabilitation devices currently available in the commercial and academic fields support coarse-grained tasks, such as wrist and forearm movements, but have weak capabilities in collecting fine-grained motor information of the finger joints; (2) Current hand motor rehabilitation systems are mostly focused on task therapy systems, which involve repeatedly moving the hand to a designated position through specific therapeutic tasks. However, this approach ignores the patient's subjective intention to rehabilitate their hand and cannot achieve rehabilitation through the patient's real-time movement intentions; (3) For patients with moderate to severe loss of hand motor ability, mechanically assisted rehabilitation methods, such as hand exoskeletons, are relatively effective. However, recent medical research indicates that for patients with moderate to severe loss of hand motor ability, non-mechanically assisted guided rehabilitation methods can help with the rehabilitation of the patient's motor nervous system. In addition, an ideal hand motor rehabilitation system should be comfortable, waterproof, and lightweight for home rehabilitation.

[0004] Therefore, this study aims to design a flexible wearable hand movement rehabilitation system capable of acquiring ultra-fine-grained hand movement information, recognizing the patient's subjective movement intentions, and implementing guided rehabilitation strategies. The hardware of the system consists of a sensing glove and a rehabilitation glove, worn on the healthy hand and the affected hand of a hemiplegic patient, respectively. Summary of the Invention

[0005] In view of this, the present invention aims to overcome the shortcomings of the above-mentioned problems in the prior art and proposes a guided mirror rehabilitation system for the recovery of hand movement in hemiplegic patients.

[0006] To achieve the above objectives, the technical solution of the present invention is implemented as follows:

[0007] A guided mirror rehabilitation system for hand movement recovery in hemiplegic patients includes a sensing glove, a rehabilitation glove, and a host computer software platform.

[0008] The sensing glove includes a sensor module, a flexible circuit module, and an embedded electronic system, used to collect motion information of the healthy side of the hand;

[0009] The rehabilitation glove includes a sensor module, a flexible circuit module, an embedded electronic system, a port expansion module, and a vibration tactile module, used to collect motion information of the affected hand;

[0010] The host computer software platform is used to receive hand movement information from the sensing glove, and to perform pattern recognition on the current gesture through a machine learning engine. The host computer software platform generates control signals based on the pattern recognition results and sends them to the rehabilitation glove.

[0011] The sensor module includes 6 IMU modules and 5 Flex sensors. The 5 IMU modules are located at the proximal phalanges of the 5 fingers to detect flexion-extension and adduction-abduction movements of the metacarpophalangeal joints. One IMU module is placed on the back of the hand to detect movements of the back of the hand. The 5 Flex sensors are placed at the proximal interphalangeal joints of each finger to detect their flexion-extension movements.

[0012] The flexible circuit module has an "island-bridge structure". Each active or passive device is welded to the "island" to achieve mechanical decoupling, and the "bridge" formed by the serpentine flexible circuit achieves high stretchability.

[0013] The embedded electronic system includes an MCU, a wireless transmission module, and a power management module;

[0014] The MCU is electrically connected to all IMU modules and Flex sensors on the corresponding gloves, and the MCU on the rehabilitation gloves is electrically connected to the vibration tactile module; the vibration tactile module is electrically connected to the MCU on the rehabilitation gloves through a port expansion module.

[0015] Furthermore, a first FFC connector is soldered to the back of the IMU module, which is connected to a second FFC connector on the flexible circuit module via a first flexible flat cable to achieve information transmission; the Flex sensor and the first FFC connector are both located on separate islands of the flexible circuit module.

[0016] Furthermore, the vibration tactile module is a vibration tactile actuator, which is a cylinder with a diameter of 6mm-12mm and a height of 1.5mm-5mm, and a maximum vibration frequency of 60Hz-300Hz.

[0017] Furthermore, the flexible circuit module is a three-layer flexible printed circuit board (FPCB) consisting of two copper layers and one polyimide film (Cu-Pi-Cu: 18um-12.5um-18um). The flexible circuit module has five branches: the thumb branch is led out separately, and the four-finger branches are led out as a whole and divided into four branches. Each branch is an elastically stretchable structure. The flexible circuit module, as well as the Flex sensor, the first FFC connector, or the vibration haptic module soldered on it, are all encapsulated in a 2mm thick cured silicone Ecoflex. The flexible circuit module is adhered to the back of the nylon glove through the cured silicone Ecoflex for easy wear by the user.

[0018] Furthermore, the MCU uses an STM32F103C8T6 chip, via I... 2 The C-communication module collects data from the IMU module and converts the voltage signals at both ends of the Flex sensor into digital signals through the MCU's digital-to-analog converter (ADC). The wireless transmission module, consisting of an NRF24L01 and its peripheral circuits, is used for wireless data transmission. The power management module is responsible for voltage regulation and converting the input voltage to 5V and 3.3V to power the active devices of the system.

[0019] The back of the substrate of the embedded electronic system is soldered with solder paste to a fourth FFC connector, which is connected to a third FFC connector fixed on the flexible circuit module via a second flexible flat cable to realize information transmission.

[0020] The port expansion module consists of a PCA9685 chip and its peripheral circuits. The port expansion module is used to expand the PWM output port GPIO of the MCU inside the rehabilitation glove.

[0021] Furthermore, each vibration haptic module is electrically connected to a lithium-ion battery, and each port of the port expansion module is connected to the gate G of a MOSFET. The source S and gate G of the MOSFET are connected to ground through an inductor, and the drain D of the MOSFET is connected to one end of the vibration haptic actuator. At the same time, a diode is connected in parallel across the two ends of the vibration haptic actuator. The lithium-ion battery converts the power supply voltage into the corresponding voltage through the power management module to power the MCU.

[0022] Furthermore, the embedded electronic system on the rehabilitation glove is equipped with a program download port, and the corresponding treatment tasks are downloaded to the MCU of the rehabilitation glove through the program download port using a J-Link emulator and debugger.

[0023] Furthermore, the host computer software platform integrates a serial port data receiving module, a data analysis platform, and a machine learning engine. The serial port receiving module establishes a connection with a module having NRF wireless communication receiving function via USB to receive hand movement information from the wireless communication module of the sensing glove. This information will be preprocessed by the data analysis platform, including data completion, data normalization, and feature extraction, and the hand movement information will be displayed in the form of curves. The machine learning engine will receive the movement information of the healthy hand and perform pattern recognition on the gestures performed by the patient's healthy hand. The host computer software platform will generate control signals based on the current pattern recognition results, and these control signals will be transmitted to the rehabilitation glove via NRF wireless communication.

[0024] Furthermore, the IMU module acquires the flexion-extension and adduction-abduction movements of the metacarpophalangeal joints. When the MCU of the affected hand calculates the angle difference at the corresponding position, the calculated angle difference is the data synthesized from the two movements, and the formula is:

[0025]

[0026] Among them Angle FE Angle represents the angle of finger flexion and extension movements captured by the IMU module. AA Angle represents the angle acquired by IMU module 10 during the adduction-abduction movement of the finger, and Angle represents the final angle difference.

[0027] Furthermore, when the angle difference is 90°, the vibration frequency is 100Hz, and the vibration frequency no longer increases with the increase of the angle difference; when the angle difference is in the range of 0 to 90°, the vibration frequency increases linearly.

[0028] Compared with existing technologies, the guided mirror rehabilitation system for hand movement recovery in hemiplegic patients described in this invention has the following advantages:

[0029] 1. This invention enables ultra-fine-grained hand motion acquisition. Utilizing multi-sensor information fusion technology, this invention acquires 41-dimensional information about hand movements by arranging an inertial measurement unit (IMU) module and Flex sensors. Notably, while finger flexion and extension movements have received considerable attention in hand movement recognition both domestically and internationally in commercial and academic fields, adduction-abduction movements are often overlooked. Therefore, this invention designs a small-volume IMU module based on the MPU6050 chip, using its roll angle to identify finger adduction-abduction movements, achieving fine-grained hand movement monitoring. Furthermore, the modular sensor arrangement design allows for flexible combinations of the number and types of sensors to suit different patients.

[0030] 2. This invention enables high-precision recognition of patients' subjective motor intentions. The invention develops a host computer software system with an integrated machine learning engine, which utilizes NRF wireless communication to receive hand movement information from a sensing glove and perform pattern recognition to determine the patient's current gesture.

[0031] 3. This invention is a guided mirror rehabilitation system. Compared with existing domestic technologies, it is the first to propose a guided mirror rehabilitation therapy device based on a vibration tactile actuator. This invention can generate corresponding control signals using the pattern recognition results of the aforementioned sensing glove, and transmit them to the rehabilitation glove via NRF wireless communication. Compared to the sensing glove, the rehabilitation glove has vibration tactile actuators positioned at the middle and proximal phalanges of each finger. When the movement of the corresponding joint position of the affected hand does not match the control signal, the vibration tactile actuator at the corresponding phalanx position will vibrate (this method of using the movement of the healthy hand to guide the affected hand to perform the same action is called mirror rehabilitation). Furthermore, the vibration frequency of the vibration tactile actuator in this application is adjusted through pulse width modulation (PWM), with more intense vibrations occurring when the hand gesture of the affected hand is more mismatched with the control signal.

[0032] 4. This invention designs a flexible circuit with an "island-bridge" structure that simultaneously features mechanical decoupling and high circuit stretchability. It consists of two copper layers and one polyimide film (Cu-Pi-Cu: 18um-12.5um-18um). All active and passive devices are arranged on the rectangular "island" structure, achieving mechanical decoupling between components. No components are arranged on the serpentine "bridge" structure; it only provides the stretching function. Using finite element analysis, at a yield strength of 0.3%, the maximum stretch length of the flexible circuit is 23.1%, higher than the maximum stretch rate of human skin (approximately 20%). Attached Figure Description

[0033] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an undue limitation of the invention. In the drawings:

[0034] Figure 1 This is a schematic diagram of the overall structure of the sensing glove of the present invention from one perspective.

[0035] Figure 2 This is a schematic diagram of the overall structure of the rehabilitation glove of the present invention from one perspective;

[0036] Figure 3 This is a front view of the overall structure of the sensing glove of the present invention;

[0037] Figure 4 This is a front view of the overall structure of the rehabilitation glove of the present invention;

[0038] Figure 5 This is a rear view of the overall structure of the sensing glove of the present invention;

[0039] Figure 6 This is a rear view of the overall structure of the rehabilitation glove of the present invention;

[0040] Figure 7 This is a schematic diagram of the flexible circuit of the sensing glove of the present invention;

[0041] Figure 8 This is a partial schematic diagram of the index finger area of ​​the rehabilitation glove of the present invention;

[0042] Figure 9 A schematic diagram of the overall operational framework of a guided mirror rehabilitation system for hand movement recovery in hemiplegic patients;

[0043] Figure 10 A schematic diagram illustrating the overall operation of the sensing glove (worn on the patient's healthy hand);

[0044] Figure 11 A schematic diagram illustrating the overall operation of the rehabilitation glove (worn on the patient's affected hand);

[0045] Figure 12 This is an illustration showing the effect of wearing the guided mirror rehabilitation device for hand movement recovery in hemiplegic patients according to the present invention;

[0046] Figure 13 This is the host computer software interface for a guided mirror rehabilitation system used to restore hand movement in hemiplegic patients.

[0047] Explanation of reference numerals in the attached figures

[0048] 1-IMU module; 2-First FFC connector; 3-First flexible flat cable; 4-Second FFC connector; 5-Flex sensor; 6-Flexible circuit module; 7-Third FFC connector; 8-Embedded electronic system; 9-Power management module; 10-MCU; 11-Wireless communication module; 12-Fourth FFC connector; 13-Vibration haptic actuator; 14-Port expansion module; 15-MOSFET; 16-Second flexible flat cable. Detailed Implementation

[0049] It should be noted that, unless otherwise specified, the embodiments and features described in the present invention can be combined with each other.

[0050] In the description of this invention, it should be understood that the terms "center," "longitudinal," "lateral," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicating orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the invention. Furthermore, the terms "first," "second," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined with "first," "second," etc., may explicitly or implicitly include one or more of that feature. In the description of this invention, unless otherwise stated, "a plurality of" means two or more.

[0051] In the description of this invention, it should be noted that, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art will understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0052] The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0053] The so-called mirror rehabilitation in this invention refers to a rehabilitation method that guides the affected hand to perform the same movements by acquiring the movements of the healthy hand. The so-called guided rehabilitation strategy refers to a method that does not provide force for hand rehabilitation, but uses a vibratory tactile actuator to generate vibrations to guide the movement of the affected hand. The closer the hand gestures are to each other, the lower the vibration frequency at the corresponding position, so as to achieve the purpose of using the healthy hand to guide the affected hand.

[0054] like Figure 1-9 As shown, this invention discloses a guided mirror rehabilitation system for hand movement recovery in hemiplegic patients, comprising a sensing glove, a rehabilitation glove, and a host computer software platform. The sensing glove includes a sensor module, a flexible circuit module 6, and an embedded electronic system 8, used to collect movement information from the healthy hand. The rehabilitation glove includes a sensor module, a flexible circuit module 6, an embedded electronic system 8, a port expansion module 14, and a vibration tactile module, used to collect movement information from the affected hand, compare it with gestures performed by the healthy hand, and then implement guided rehabilitation therapy through the vibration tactile module. The embedded electronic system 8 includes an MCU 10, a wireless transmission module 11, and a power management module 9.

[0055] Specifically, the sensor module in this invention includes six IMU modules 1 and five Flex sensors 5. The five IMU modules 1 are located at the proximal phalangeal joints of the five fingers, respectively, for detecting flexion-extension and adduction-abduction movements of the metacarpophalangeal joints. One IMU module 1 is placed on the back of the hand for detecting hand movements. The five Flex sensors 5 are placed at the proximal interphalangeal joints of each finger for detecting their flexion-extension movements. Each IMU module 1 provides 3-axis acceleration information and 3-axis angular velocity information, which are transmitted to the microcontroller unit (MCU) via IIC communication. Each Flex sensor 5 provides 1-dimensional deflection information, which is collected by analog-to-digital converter (ADC) and sent to the MCU 10.

[0056] Specifically, the IMU module 1 designed in this invention uses the MPU6050 chip as its core, with a size of only 12mm*12mm*1mm. It can collect 3-axis acceleration information and 3-axis angular velocity information, meaning that one IMU module 1 can collect 6-dimensional information. Furthermore, the raw data from the IMU module 1 can be calculated into angle data in the MCU 10 using a complementary filtering algorithm, facilitating the determination of the actual angles of finger flexion-extension and adduction-abduction movements. With six IMU modules 1 and five Flex sensors 5, a total of 6*6+5=41 dimensions of hand information can be collected, achieving fine-grained acquisition of hand information.

[0057] Specifically, the back of IMU module 1 is soldered with a first FFC connector 2 using solder paste, which is connected to a second FFC connector 4 on flexible circuit module 6 via a first flexible flat cable 3 to achieve information transmission. The two pins of Flex sensor 5 are soldered to a separate island structure of flexible circuit module 6 using low-temperature soldering technology.

[0058] Specifically, the flexible circuit module of the sensing glove is a three-layer flexible printed circuit board (FPCB) consisting of two copper layers and one polyimide film (Cu-Pi-Cu: 18um-12.5um-18um). The flexible circuit module is designed as an "island-bridge structure," with passive components (Flex sensor, second FFC connector) soldered onto it using a low-temperature soldering process, each located on a separate island, achieving mechanical decoupling between devices. It's worth noting that, compared to the sensing glove, the rehabilitation glove's flexible circuit module 6 also has vibration tactile actuators 13 attached using a low-temperature soldering process to perform guided rehabilitation tasks. There are a total of 10 vibration tactile actuators 13, with 5 positioned at the proximal phalanges of the rehabilitation glove and the other 5 positioned at the middle phalanges. The flexible circuit module has five branches: a separate thumb branch, and four finger branches that extend as a whole and then divide into four sub-branches. Each sub-branchineme is designed as an elastically stretchable structure, similar to a flexible spring, to ensure the stretchability of the flexible circuit. Furthermore, the flexible circuit module and the passive components soldered onto it are encapsulated in a 2mm thick layer of Ecoflex silicone to ensure sweat resistance. Finally, the flexible circuit module is adhered to the back of the nylon glove using a small amount of Ecoflex silicone for easy wear.

[0059] Specifically, the MCU10 is an STM32F103C8T6 chip, which is electrically connected to all IMU modules and Flex sensors on the sensing glove, and communicates with them via I... 2 The C-type communication collects data from the IMU module and uses the MCU10's digital-to-analog converter (ADC) to convert the voltage signal from the Flex sensor into a digital signal. The power management module 9 is used for voltage regulation and converts the external voltage to 5V and 3.3V respectively to power the active devices in the system. The MCU10 of the sensing glove establishes a connection with the computer through the wireless transmission module 11 and transmits data to the host computer software via serial port.

[0060] Specifically, similar to IMU module 1, the back of the embedded electronic system 8 is soldered with a fourth FFC connector 12, which can be connected to a third FFC connector 7 fixed on the flexible circuit module 6 via a second flexible flat cable 16 to achieve information transmission.

[0061] Specifically, the flexible circuit module of the rehabilitation glove is similar to that of the aforementioned sensory glove, but a "bridge" structure is added to the flexible circuit module below the fixed vibration tactile actuator 13 to mechanically decouple the components and facilitate soldering.

[0062] The rehabilitation glove features a vibration tactile module, with 10 modules positioned at the middle and proximal phalanges of each finger for guided hand movement rehabilitation. Each of the 10 modules is electrically connected to the MCU on the rehabilitation glove via a port expansion module 14. In addition to acquiring data, the MCU on the rehabilitation glove also compares and modulates the movement with that of the sensing glove, ensuring the movement of the rehabilitation glove closely mirrors that of the sensing glove. The vibration tactile actuator 13 in this invention is a cylinder with a diameter of 6mm-12mm and a height of 1.5mm-5mm, with a maximum vibration frequency of 60Hz-300Hz.

[0063] Specifically, the port expansion module 14, composed of a PCA9685 chip and its peripheral circuitry, is added to address the insufficient number of general purpose input / output (GPIO) ports on the MCU. The port expansion module 14 is used to expand the PWM output port GPIO of the MCU within the rehabilitation glove; to control the vibration tactile actuator 13, the MCU uses I... 2 C sends data to the port extender, addressing a specific GPIO of the port extender to program the duty cycle of the digital pulse width modulation (PWM) signal. These PWM signals are used to control the smooth change of the vibration frequency of the vibration haptic actuator 13. Since the output current of the port extender's GPIO is insufficient to drive a vibration haptic actuator normally, a metal-oxide-semiconductor field-effect transistor (MOSFET) is used as a voltage-controlled current source. This MOSFET meets the normal operating requirements of the vibration haptic actuator through a direct interface with the power supply. The specific connection method is as follows: Figure 8 All components within the dashed box are identical. Each vibration haptic actuator 13 is electrically connected to a lithium-ion battery. Each port of the port expansion module 14 is connected to the gate G of a MOSFET. The source S and gate G of the MOSFET are connected to ground via an inductor. The drain D of the MOSFET is connected to one end of the vibration haptic actuator. Simultaneously, diodes are connected in parallel across the two ends of the vibration haptic actuator 13 to meet the normal operation requirements of the vibration haptic actuator 13. The lithium-ion battery converts the power supply voltage into the corresponding voltage through the power management module 9 to power the MCU.

[0064] The present invention will now be described through specific embodiments.

[0065] This embodiment of the guided mirror rehabilitation device for hand movement recovery in hemiplegic patients includes an IMU module 1, a first FFC connector 2, a first flexible flat cable 3, a second FFC connector 4, a Flex sensor 5, a flexible circuit 6, a third FFC connector 7, an embedded electronic system 8, a power management module 9, an MCU 10, a wireless communication module 11, a fourth FFC connector 12, a vibration tactile actuator 13, a port expansion module 14, a MOSFET 15, and a second flexible flat cable 16.

[0066] The IMU module 1 has dimensions of 12mm*12mm*1mm. The first FFC connector 2 is attached to the back of the IMU module 1 by a low-temperature soldering process and is connected to the second FFC connector 4 by a first flexible flat cable 3. The interface between the first FFC connector 2 and the second FFC connector 4 is a 6P interface, and the first flexible flat cable is also a 6P. The flexible circuit 6 is an "island-bridge structure". Active and passive devices are attached to the island position by low-temperature soldering, and the bridge is designed as a serpentine structure to ensure stretchability.

[0067] For the sensing glove without the vibration tactile actuator 13, only the second FFC connector 4 and the third FFC connector 7 are adhered to the upper surface of the island structure of the flexible circuit 6 using a low-temperature soldering process, while the Flex sensor is soldered to the lower surface of the island structure of the flexible circuit 6. For the rehabilitation glove, the second FFC connector 4, the third FFC connector 7, and the vibration tactile actuator 13 are all adhered to the upper surface of the island structure of the flexible circuit 6 using a low-temperature soldering process, while the Flex sensor is soldered to the lower surface of the island structure of the flexible circuit 6. The flexible circuit is generally spring-like, but at some soldered components, the spring structure becomes an island structure. This is done to mechanically decouple the sensors while ensuring the stability of the soldering. The IMU module 1 of the sensing glove is electrically connected to the IMU module 1 of the rehabilitation glove, and the IMU module 1 of the rehabilitation glove is also electrically connected to the vibration tactile actuator.

[0068] The fourth FFC connector 12 is adhered to the back of the embedded electronic system 8 using a low-temperature soldering process and is connected to the third FFC connector 7 via the second flexible flat cable 16. The power management module 9, MCU 10, and wireless communication module 11 are all adhered to the upper surface of the substrate of the embedded electronic system 8 using a low-temperature soldering process. The second flexible flat cable 16, the third FFC connector 7, and the fourth FFC connector 12 are all 25 pins.

[0069] In this embodiment, the vibration tactile actuator is a cylinder with a diameter of 7mm and a height of 2mm, and a maximum vibration frequency of 200Hz.

[0070] To verify the effectiveness of fine-grained recognition, pattern recognition was performed on the healthy side hand using machine learning algorithms, including SVM, DT, KNN, MLP-NN, and LSTM-CNN. Ten volunteers were selected, each repeating the gestures of the 26 English letters 10 times, and the data was collected. Taking LSTM-CNN as an example, the final classification accuracy reached 99.16%, achieving the goal of fine-grained detection and far exceeding the international cutting-edge research results of the same period. As shown in Table 1, the accuracy of gesture recognition using different algorithms is compared. The classification accuracy of the Long Short-Term Memory-Convolutional Neural Network (LSTM-CNN) deep learning algorithm can reach 99.16%, far exceeding the international cutting-edge research results of the same period.

[0071] Table 1

[0072]

[0073] like Figure 10-13 As shown, the working principle and workflow of this invention are as follows:

[0074] The healthy hand wears a sensory glove, and the affected hand wears a rehabilitation glove. When guided mirror rehabilitation training begins, the IMU module 1 of the sensory glove transmits triaxial acceleration and triaxial angular velocity information using IMU... 2 The C-type communication transmits information to the MCU10 of the sensing glove via the first flexible flat cable 3, the flexible circuit 6, and the second flexible flat cable 16. Similarly, the voltage changes of the two pins of the Flex sensor are transmitted to the MCU10 of the sensing glove via the flexible circuit 6 and the second flexible flat cable 16. The analog quantity is further converted into a digital quantity by the ADC inside the MCU10. The signal acquisition frequency of both the IMU module 1 and the Flex sensor 5 is 100Hz. The MCU of the sensing glove transmits the motion data of the healthy hand to the developed host computer software via the wireless communication module 11.

[0075] The host computer software was developed using Python and the PyCharm editor. This software integrates a serial port data receiving module, a data analysis platform, and a machine learning engine. The serial port receiving module connects via USB to a module with NRF wireless communication reception capabilities, receiving hand movement information from the sensing glove's wireless communication module 11. This information undergoes data preprocessing through the data analysis platform, including data completion, data normalization, and feature extraction, and the hand movement information is displayed as curves. Furthermore, the machine learning engine receives movement information from the healthy hand and performs pattern recognition on the gestures performed by the patient's healthy hand. Currently, the machine learning engine integrates SVM, DT, KNN, MLP-NN, and LSTM-CNN algorithms. The machine learning engine's output defaults to using the pattern recognition results from the LSTM-CNN algorithm. The host computer software generates control signals based on the current pattern recognition results, and these control signals are transmitted to the rehabilitation glove via NRF wireless communication.

[0076] The rehabilitation glove receives control signals from the host computer software via the wireless communication module 11 and transmits the information to the MCU of the rehabilitation glove; simultaneously, the rehabilitation glove also performs its own information acquisition: the IMU module 1 of the rehabilitation glove transmits the three-axis acceleration and angular velocity information using I... 2 Communication between the two pins of the Flex sensor and the MCU of the rehabilitation glove is achieved through the first flexible flat cable 3, the flexible circuit 6, and the second flexible flat cable 16. The voltage changes of the two pins are also transmitted to the MCU of the rehabilitation glove via the flexible circuit 6 and the second flexible flat cable 16. The MCU's internal ADC further converts the analog signal into a digital signal. The MCU of the rehabilitation glove then transmits the signal via the GPIO port at a frequency of 10Hz through the I / O port. 2 The C communication sends data to the port expansion module 14, and the port expansion module 14 adjusts the vibration frequency of the 10 vibration haptic actuators 13 by adjusting the duty cycle of the PWM signal of the specified GPIO port.

[0077] The MCU of the rehabilitation glove compares the angle difference between the control signal and the corresponding joint of the affected hand. The greater the mismatch, the higher the vibration frequency of the vibration tactile actuator 13 at the corresponding position of the rehabilitation glove; conversely, the lower the vibration frequency. Specifically, when the angle difference is 90°, the vibration frequency is 100Hz, and the vibration frequency no longer increases with the increase of the angle difference. Within the angle difference range of 0-90°, the vibration frequency increases linearly. As the movement position of the corresponding joint of the affected hand gradually approaches that of the healthy hand, the vibration frequency of the vibration tactile actuator 13 at the corresponding position gradually decreases until it drops to 0Hz. At this point, the gestures of the affected hand and the healthy hand are consistent, and the guided mirror rehabilitation training is completed. It is worth noting that the IMU module 1 collects the flexion-extension and adduction-abduction movements of the metacarpophalangeal joints. When the MCU of the affected hand calculates the angle difference at the corresponding position, the calculated angle difference is the data synthesized from the two movements, as shown in the formula:

[0078]

[0079] Among them Angle FE Angle represents the angle of finger flexion and extension movements captured by the IMU module. AA Angle represents the angle of the finger during adduction and abduction, acquired by IMU module 10, and Angle represents the final angle difference.

[0080] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A guided mirror rehabilitation system for hand movement recovery in hemiplegic patients, characterized in that: include The sensing glove, worn on the healthy hand of the patient, includes a sensor module, a flexible circuit module, and an embedded electronic system to collect motion information of the healthy hand. The rehabilitation glove, worn on the affected hand of the patient, includes a sensor module, a flexible circuit module, an embedded electronic system, a port expansion module, and a vibration tactile module, used to collect motion information and provide tactile feedback for the affected hand; The host computer software platform is used to receive hand movement information from the sensing glove and perform pattern recognition on the current gesture through a machine learning engine. The host computer software platform generates control signals based on the pattern recognition results and sends them to the rehabilitation glove. The sensor module includes 6 IMU modules and 5 Flex sensors. The 5 IMU modules are located at the proximal phalanges of the 5 fingers to detect flexion and extension movements and adduction-abduction movements of the metacarpophalangeal joints. The 1 IMU module is placed on the back of the hand to detect movements of the back of the hand. Five Flex sensors are positioned at the proximal interphalangeal joint of each finger to detect its flexion and extension movements; The flexible circuit module has an island-bridge structure, with each active or passive device welded to the island to achieve mechanical decoupling, and the bridge formed by the serpentine flexible circuit achieving high stretchability. The embedded electronic system includes an MCU, a wireless transmission module, and a power management module; The MCU is electrically connected to all IMU modules and Flex sensors on the corresponding gloves, and the MCU on the rehabilitation gloves is electrically connected to the vibration tactile module; the vibration tactile module is electrically connected to the MCU on the rehabilitation gloves through a port expansion module.

2. The guided mirror rehabilitation system for hand movement recovery in hemiplegic patients according to claim 1, characterized in that: The IMU module has a first FFC connector soldered to its back, which is connected to a second FFC connector on the flexible circuit module via a first flexible flat cable to transmit information; the Flex sensor and the first FFC connector are both located on separate islands of the flexible circuit module.

3. The guided mirror rehabilitation system for hand movement recovery in hemiplegic patients according to claim 1, characterized in that: The vibration tactile module is a vibration tactile actuator, which is a cylinder with a diameter of 6mm-12mm and a height of 1.5mm-5mm, and a maximum vibration frequency of 60Hz-300Hz.

4. The guided mirror rehabilitation system for hand movement recovery in hemiplegic patients according to claim 1, characterized in that: The flexible circuit module is a three-layer flexible printed circuit board (FPCB) consisting of two copper layers (Cu) and one polyimide film (Pi), with Cu-Pi-Cu ratios of 18µm-12.5µm-18µm. The flexible circuit module has five branches: a separate thumb branch and four branch lines that extend out as a whole. Each branch is a resilient and expandable structure. The flexible circuit module, along with the Flex sensor, the first FFC connector, or the vibration haptic module soldered onto it, are all encapsulated in a 2mm thick cured silicone Ecoflex. The flexible circuit module is adhered to the back of the nylon glove via the cured silicone Ecoflex for easy wear by the user.

5. A guided mirror rehabilitation system for hand movement recovery in hemiplegic patients according to claim 1, characterized in that: The MCU uses an STM32F103C8T6 chip, which collects data from the IMU module via I2C communication and converts the voltage signals at both ends of the Flex sensor into digital signals via the MCU's digital-to-analog converter (ADC). The wireless transmission module is composed of an NRF24L01 and its peripheral circuits for wireless data transmission. The power management module is responsible for voltage regulation and converting the input voltage to 5V and 3.3V to power the active devices of the system. The back of the substrate of the embedded electronic system is soldered with solder paste to a fourth FFC connector, which is connected to a third FFC connector fixed on the flexible circuit module via a second flexible flat cable to realize information transmission. The port expansion module consists of a PCA9685 chip and its peripheral circuits. The port expansion module is used to expand the PWM output port GPIO of the MCU inside the rehabilitation glove.

6. A guided mirror rehabilitation system for hand movement recovery in hemiplegic patients according to claim 1, characterized in that: Each vibration haptic module is electrically connected to a lithium-ion battery. Each port of the port expansion module is connected to the gate G of a MOSFET. The source S and gate G of the MOSFET are connected to ground through an inductor. The drain D of the MOSFET is connected to one end of the vibration haptic actuator. At the same time, a diode is connected in parallel across the two ends of the vibration haptic actuator. The lithium-ion battery converts the power supply voltage into the corresponding voltage through the power management module to power the MCU.

7. A guided mirror rehabilitation system for hand movement recovery in hemiplegic patients according to claim 1, characterized in that: The embedded electronic system on the rehabilitation glove is equipped with a program download port. The corresponding treatment tasks are downloaded to the MCU of the rehabilitation glove through the program download port using a J-Link emulator and debugger.

8. A guided mirror rehabilitation system for hand movement recovery in hemiplegic patients according to claim 1, characterized in that: The host computer software platform integrates a serial port data receiving module, a data analysis platform, and a machine learning engine. The serial port receiving module connects to a module with NRF wireless communication receiving function via USB to receive hand movement information from the wireless communication module of the sensing glove. This information will be preprocessed by the data analysis platform, including data completion, data normalization, and feature extraction, and the hand movement information will be displayed in the form of curves. The machine learning engine will receive the movement information of the healthy hand and perform pattern recognition on the gestures performed by the patient's healthy hand. The host computer software platform will generate control signals based on the current pattern recognition results, and these control signals will be transmitted to the rehabilitation glove via NRF wireless communication.

9. A guided mirror rehabilitation system for hand movement recovery in hemiplegic patients according to claim 1, characterized in that: The IMU module acquires the flexion-extension and adduction-abduction movements of the metacarpophalangeal joints. When the MCU of the affected hand calculates the angle difference at the corresponding position, the calculated angle difference is the data synthesized from the two movements, and the formula is: ； Among them Angle FE Angle represents the angle of finger flexion and extension movements captured by the IMU module. AA Angle represents the angle acquired by IMU module 10 during the adduction-abduction movement of the finger, and Angle represents the final angle difference.

10. A guided mirror rehabilitation system for hand movement recovery in hemiplegic patients according to claim 9, characterized in that: When the angle difference is 90°, the vibration frequency is 100Hz, and the vibration frequency no longer increases with the increase of the angle difference. When the angle difference is in the range of 0~90°, the vibration frequency increases linearly.

Images