KR102549103B1 - Development of artificial intelligence feedback training method and system for parkinson patient rehabilitation - Google Patents

Abstract

An artificial intelligence feedback training method and system development technology for rehabilitation of Parkinson's patients is disclosed. A feedback training method for rehabilitation performed by a feedback training system according to an embodiment includes collecting user gait data from sensor data sensed according to the user's gait; adjusting the level of difficulty for the collected gait data using an artificial intelligence model for gait pattern analysis and rehabilitation; and inducing a user's gait pattern to a normal person's gait pattern through a difficulty level adjustment signal generated based on the adjusted level of difficulty.

Description

Development of artificial intelligence feedback training method and system for rehabilitation of Parkinson's patients

The description below relates to a feedback training method and system for rehabilitation of a patient.

As nutritional hygiene improvement and advanced health and medical technology accelerate, geriatric diseases due to aging are emerging as an important challenge to overcome in order to create a healthy society among developed countries. There are various geriatric diseases, but among them, Parkinson's disease (PD) is a representative geriatric disease that must be prepared. Most Parkinson's disease treatments use drug therapy for the treatment of non-motor symptoms and rehabilitation using exercise therapy for the treatment of motor symptoms. Representative among them are cue training and feedback training. Cue training is one of the research methods for rehabilitation training for Parkinson's disease. It is a treatment that allows patients to continue to perform intentional movements by creating and repeating artificial cues under specific conditions. It refers to a treatment that induces normal gait using an auditory or visual display method. Conventional feedback training is a method in which a reference value is artificially set and feedback is received only when the normal gait is above, and despite the gait posture, walking speed, walking swing width, etc. vary from person to person, the reference value is constantly or manually adjusted. It not only reduces the effect of rehabilitation, but also has a disadvantage that can damage the reliability of rehabilitation equipment. In addition, although the current feedback training is producing good results as rehabilitation training, the performance of training is reduced due to insensitivity to the feedback itself due to the uniform feedback.

Spaulding SJ, Barber BB, Colby M, Cormack B, Mick T, Jenkins ME, Cueing and Gait Improvement Among People With Parkinson's Disease: A Meta-Analysis, Archives of Physical Medicine and Rehabilitation 94: 562-570, 2013 Suteerawattananon M, Morris GS, Etnyre BR, Jankovic J, Protas EJ, Effects of visual and auditory cues on gait in individuals with Parkinson's disease, J of the Neurological Sciences 219: p63-69, 2004.

It is possible to provide an artificial intelligence-based feedback training method and system in which reference values are automatically set based on various gait postures, walking speeds, and walking swing widths.

Artificial intelligence-based feedback that automatically updates new reference values to the rehabilitation aid device based on the rehabilitation training data in the built-in memory each time, and the updated information is applied directly to the rehabilitation aid device to maximize the rehabilitation of feedback training. Training methods and systems can be provided.

It is possible to provide a selective feedback method and system for generating vibration when a person does not walk more than a standard value for a certain period of time.

In addition, it is possible to provide a rehabilitation assistive device that can be freely attached and detached using a Bluetooth function and transmits walking speed and gait pattern to a mobile phone or wristwatch type display device so that the patient as well as the medical staff can be monitored in real time at the same time, and a feedback training method using the same. can

A feedback training method for rehabilitation performed by a feedback training system includes: collecting gait data of a user from sensor data sensed according to gait of a user wearing a rehabilitation assisting device; adjusting the level of difficulty for the collected gait data using an artificial intelligence model for gait pattern analysis and rehabilitation; and inducing a user's gait pattern to a normal person's gait pattern through a difficulty level adjustment signal generated based on the adjusted level of difficulty.

The adjusting step may include analyzing rehabilitation exercise information including difficulty level, gait speed, and gait recognition rate using the collected gait data, and obtaining current rehabilitation state information of the user through clinical results for the analyzed rehabilitation exercise information. Determining and adjusting the level of difficulty based on the determined user's current rehabilitation state information.

In the adjusting step, when the gyro sensor value rotating around the Z axis obtained from the gyro sensor during walking increases by exceeding the gyro threshold value as the rehabilitation exercise starts, accelerometer sensor values are collected and the number of times recorded, The method may include calculating an average acceleration when the gyro sensor value descends below a gyro threshold value, and determining whether the calculated average acceleration exceeds a level of difficulty set for the user.

In the adjusting step, when the calculated average acceleration exceeds the difficulty level set for the user, the effective number of steps and the number of simple steps are counted, and when the calculated average acceleration is less than the difficulty level set for the user, the number of simple steps is counted. steps may be included.

The adjusting step may include calculating a gait recognition rate through a ratio of the effective number of steps to the counted simple number of steps as the rehabilitation exercise ends, and when the calculated gait recognition rate exceeds the first reference gait recognition rate, the current It may include storing the difficulty level and generating a new difficulty level based on the stored current level of difficulty.

The adjusting step may include comparing the calculated gait recognition rate with a second reference gait recognition rate when the calculated gait recognition rate is less than the first reference gait recognition rate, and if the calculated gait recognition rate is less than the second reference gait recognition rate, the current It may include storing the difficulty level and generating a new difficulty level based on the stored current level of difficulty.

The collecting may include collecting sensor data generated during walking using an acceleration sensor or a gyro sensor worn on both ankles of the user based on a level of difficulty set for the user as the rehabilitation exercise starts, and collecting the collected sensor data. It may include calculating and processing to count the number of steps recognized as walking, and collecting the counted number of steps.

In the collecting step, only the highest acceleration data among the collected sensor data is stored in real time, and when the highest acceleration data is updated, new maximum acceleration data is stored again, and the maximum acceleration data exceeds the reference acceleration for gait detection. If so, counting the number of steps may be included.

The inducing may include displaying rehabilitation state information related to the gait speed and gait pattern including the adjusted level of difficulty on a display.

The inducing step displays the number of times counted as effective walking, the current difficulty level, and the result of the rehabilitation exercise including whether the difficulty level of the next rehabilitation exercise is increased or decreased based on the gait achievement rate and the current difficulty level. steps may be included.

The inducing may include transmitting rehabilitation state information related to a gait speed and a gait pattern including the adjusted difficulty level to the electronic device by being connected to the electronic device through wireless communication through a wireless communication module. .

The inducing may include generating a vibration when the user fails to achieve an effective gait while walking based on the gait speed and gait pattern including the adjusted level of difficulty.

In order to execute the feedback training method for rehabilitation performed by the feedback training system, a computer program stored in a computer readable storage medium collects user's gait data from sensor data detected according to the gait of the user wearing the rehabilitation assistance device. doing; adjusting the level of difficulty for the collected gait data using an artificial intelligence model for gait pattern analysis and rehabilitation; and inducing a user's gait pattern to a normal person's gait pattern through a difficulty level adjustment signal generated based on the adjusted level of difficulty.

The feedback training system includes at least one processor configured to execute computer readable instructions included in a memory, and the at least one processor is configured to perform a user's training based on sensor data sensed according to a gait of a user wearing a rehabilitation assistance device. Gait data is collected, the difficulty level of the collected gait data is adjusted using an artificial intelligence model for gait pattern analysis and rehabilitation, and the user's gait pattern is determined through a difficulty adjustment signal generated based on the adjusted level of difficulty. It can be induced by the gait pattern of a normal person.

The at least one processor analyzes rehabilitation exercise information including difficulty, gait speed, and gait recognition rate using the collected gait data, and the user's current rehabilitation state information through clinical results for the analyzed rehabilitation exercise information. It is possible to determine and adjust the level of difficulty based on the determined user's current rehabilitation state information.

Based on various gait postures, gait speeds, and gait swing widths, it is possible to guide the gait patterns of Parkinson patients to the gait patterns of normal people by analyzing the gait patterns of Parkinson patients and setting up rehabilitation algorithms through artificial intelligence that automatically sets the standard value. can

Based on the rehabilitation training data each time, a new reference value is automatically updated to the rehabilitation assisting device, and the updated information is directly applied to the rehabilitation assisting device, thereby maximizing the rehabilitation of the feedback training.

Instead of a one-size-fits-all feedback, it is possible to provide selective feedback that generates vibrations to the patient's rehabilitation assistance device when he or she does not walk more than a standard value for a certain period of time.

By providing a miniaturized and lightweight device through 3D printing technology, it is possible to provide a rehabilitation assistive device for Parkinson's disease improvement that can improve wearing convenience and portability and a feedback training method using the same.

In addition, it can be freely attached and detached using a wireless communication function, and the walking speed and walking pattern can be transmitted to an electronic device or a wrist watch type display device so that the patient as well as the medical personnel can be monitored in real time at the same time.

1 is a diagram for schematically explaining a rehabilitation assisting device for rehabilitation of a Parkinson's patient according to one embodiment.
2 is a diagram for explaining the operation of a rehabilitation assisting device for rehabilitation of a Parkinson's patient according to an embodiment.
3 is a schematic diagram showing the inside of a rehabilitation assisting device according to an embodiment.
4 to 11 are diagrams illustrating a control unit of a rehabilitation assisting device for rehabilitation of a Parkinson's patient according to an embodiment.
12 is a diagram for explaining an operation of adjusting difficulty using an artificial intelligence model according to an embodiment.
13 is a graph for explaining a relationship between a level of difficulty and a gait recognition rate, according to an embodiment.
14 is a flowchart illustrating a feedback training method for rehabilitation in a rehabilitation assisting device for rehabilitation of a Parkinson's patient according to an embodiment.
15 to 17 are detailed flowcharts illustrating a feedback training method for rehabilitation in a rehabilitation assisting device for rehabilitation of a Parkinson's patient according to an embodiment.

Hereinafter, an embodiment will be described in detail with reference to the accompanying drawings.

Parkinson's Disease (PD), one of the geriatric diseases, causes symptoms such as Freezing of Gait (FOG), step/shorter step, etc., resulting in activity restriction and further deteriorating quality of life. cause In the embodiment, an operation of inducing a Parkinson's patient's gait pattern to a normal person's gait pattern by artificial intelligence analyzing the gait pattern by itself and setting a rehabilitation algorithm based on a sensor that detects the current gait will be described.

1 is a diagram for schematically explaining a rehabilitation assisting device for rehabilitation of a Parkinson's patient according to one embodiment.

The feedback training system may be implemented in a rehabilitation assisting device. For example, the feedback training system may be operated in the form of a platform or program of a rehabilitation assistance device. Accordingly, the feedback training system may adjust the level of difficulty of the Parkinson's patient based on gait data related to the rehabilitation exercise performed by the rehabilitation assistance device.

The rehabilitation assisting device is a smart walking assisting device for assisting rehabilitation of Parkinson's patients, and may be composed of a sensor unit 110 and a control unit 120. Referring to FIG. 1 , it is a view showing that a user (eg, Parkinson's patient) 101 is wearing a rehabilitation assisting device.

The sensor unit 110 is composed of a plurality of pieces, and can be fixed to both ankles of the user 101 through protectors, and the control unit 120 is connected to the sensor unit 110 through wireless communication such as Bluetooth, and the user It can be worn in the form of a band (wrist watch type) on the body of (101), for example, on the wrist.

The sensor unit 110 may collect gait data through sensor data sensed according to the gait of the user 101 . At this time, at least one or more sensors for detecting walking data of the user 101 may be configured in the sensor unit 110 . For example, various sensors such as an acceleration sensor and a gyro sensor may be included. The sensor unit 110 is attached to the ankle of the user, detects a walking pattern of the user 101 in a current state, and automatically transmits the detected pattern to the control unit 120 through wireless communication (eg, Bluetooth).

The control unit 110 stores the data of the sensor unit 110 transmitted through wireless communication in real time in the main microchip controller, and analyzes the data of the sensor unit 110 using a built-in artificial intelligence model and adjusts the level of difficulty. It can be provided so that it can be monitored in real time through a wristwatch type display.

The gait data analyzed by feedback training does not disappear when rehabilitation training is finished, but is stored in memory, and artificial intelligence can update indicators of continued rehabilitation training and finally induce walking in a normal gait pattern.

A vibration unit may be incorporated in the display of the control unit 120 in the shape of a wristwatch. When the user 101 walks without showing a normal gait pattern for a certain period of time, a stimulus is generated through the vibration unit to inform the user 101 that the current gait is not correct in real time, thereby inducing the user 101 to walk correctly. can Selective feedback may be provided to the user 101 by generating vibration when the user 101 does not walk more than a standard value for a certain period of time, rather than a uniform feedback. For example, the sensor unit 110 may sense that an acceleration equal to or less than a predetermined level of difficulty is generated based on sensor data of the user 101 . As the sensor unit 110 generates an acceleration of a predetermined degree of difficulty or less, the controller 120 may operate the vibration motor for a specific period of time (eg, 1 second).

For example, the sensor unit 110 may collect gait data using sensor data. The sensor unit 110 measures the acceleration sensor value using an acceleration sensor, calculates and processes it based on a self-developed gait detection algorithm, and obtains gait data (eg, gait posture, gait speed, gait swing width, etc.) may be detected, and the detected gait data may be transmitted to the control unit 120 worn on the wrist of the user 101 using short range wireless communication and displayed.

In addition, the feedback training system may analyze rehabilitation exercise information about gait data transmitted and received between the sensor unit 110 and the controller 120 by using gait pattern analysis and artificial intelligence for rehabilitation. For example, the feedback training system may obtain rehabilitation exercise information including a difficulty level of 5000, a gait speed of 2.3 lm/h, and a gait recognition rate of 95%. The feedback training system may analyze rehabilitation exercise information including difficulty, gait speed, and gait recognition rate based on gait data. The feedback training system may determine current rehabilitation state information for rehabilitation exercise information analyzed through clinical results. The feedback training system may adjust the level of difficulty based on the determined current rehabilitation state information. For example, the difficulty level may be adjusted to 6000.

The controller 120 may be wirelessly connected to the sensor unit 110 to receive and display information on steps of the user 101 . For example, the controller 120 may display the number of steps recognized by the user, difficulty level, and today's exercise result. For example, today's exercise results, such as a recognition count of 12345, a difficulty of 140, a gait achievement rate of 55%, and an increase in difficulty of the next exercise, may be displayed through the controller 120 .

2 is a diagram for explaining the operation of a rehabilitation assisting device for rehabilitation of a Parkinson's patient according to an embodiment.

The sensor unit 110 may transmit the current level of difficulty to the control unit 120 through wireless communication 210 . The sensor unit 110 may transmit data of the sensor unit 110 to the control unit 120 worn on the wrist of the user 101 until the rehabilitation of the gait exercise program is completed, and display the gait data in real time. In this way, by wirelessly connecting the sensor unit 110 worn on both ankles and the control unit 120 worn on the wrist, restrictions on the user 101's body movement can be reduced.

The feedback training system may analyze gait data collected at the end of rehabilitation (for example, the total number of steps and the number of steps above a reference value) using artificial intelligence, and may adjust the level of difficulty through the analysis.

The sensor unit 110 may count the number of steps by using acceleration data. In detail, the sensor unit 110 may count the number of times walking is detected. An operation of recognizing the number of walking steps equal to or greater than the reference value may be performed as follows. The sensor unit 110 may store only the maximum acceleration generated during walking in real time. When the maximum acceleration is updated, the sensor unit 110 may store the new maximum acceleration again. The sensor unit 110 may count the number of steps when the maximum acceleration exceeds the gait detection reference acceleration.

As another example, the sensor unit 110 may measure the acceleration of the X, Y, and Z axes in real time. The sensor unit 100 calculates an acceleration distance using the X, Y, and Z-axis acceleration obtained from the acceleration sensor, calculates an acceleration value using the calculated acceleration distance, and calculates an acceleration average using the calculated acceleration value. The number of steps may be counted when the average of the calculated accelerations is greater than the reference acceleration for gait detection.

As another example, the sensor unit 110 may measure acceleration and gyro sensor values using an acceleration sensor and a gyro sensor. At this time, data can be collected using a total of four axes: three X, Y, and Z axes of the accelerometer and one rotation axis (YAW) of the Z axis of the gyro sensor. The sensor unit 110 may count the number of steps by using the collected acceleration and gyro sensor values.

In addition, the sensor unit 110 collects the acceleration sensor value and calculates the number of times of collection when the gyro sensor value obtained from the gyro sensor during walking and rotating around the Z axis rises above a preset value or falls below a preset value. can be recorded Continuing, the sensor unit 110 calculates an average acceleration when the gyro sensor value during walking falls below a preset value or rises above a preset value, and when the average acceleration is equal to or greater than the preset value, the number of steps is counted as 1. can increase

On the other hand, after the rehabilitation assistance device counts the number of steps in the sensor unit 110 that detects the gait data, the counted number of steps, which is minimum data, is transmitted to the controller 120 and/or the electronic device 130 to be displayed. can Accordingly, the rehabilitation assisting apparatus has advantages in that battery life is reduced and data loss is reduced by minimizing data wireless transmission. That is, while receiving data from the acceleration sensor and the gyro sensor in real time within the sensor unit 110, when both the vector value and the gyro value exceed a standard, the number of steps data is set by an algorithm that counts the number of steps, and the control unit 120 can be transmitted wirelessly. In particular, since most of the rehabilitation assisting devices for rehabilitation of Parkinson's patients are used by elderly users, they can be displayed through the control unit 120 in the form of a simple wrist watch rather than the electronic device 130, which is difficult to use.

The counted walking count may be displayed on the wrist display of the controller 120 and may be displayed on the electronic device 130 using wireless communication 210 so that a medical practitioner or a third party may monitor the patient's condition.

In this way, it can be transmitted to an electronic device 130 such as a smart phone using wireless communication to be monitored in real time by a medical person, and transmitted to the control unit 120 in the form of a wristwatch worn by the user so that the current state is displayed in real time. can be monitored.

The sensor unit 110 is connected to the electronic device 130 through wireless communication through a wireless communication module, and transmits the number of steps counted by the sensor unit 110 to the electronic device 130 so that the walking information can be displayed or analyzed. there is. In this way, the sensor unit 110 is connected to the electronic device 130 through wireless communication such as Bluetooth to provide information on steps, so that the user can more accurately analyze information on steps, and third parties such as medical staff It is possible to have the user identify and analyze information about the user's step.

Furthermore, the controller 120 may induce a rehabilitation exercise by providing sound according to steps or continuously providing sound and vibration according to a preset BPM. At this time, the controller 120 may set the type, size, BPM, etc. of sound and vibration inducing rehabilitation exercise, and may set the type, size, BPM, etc. of sound and vibration through the electronic device 130.

The electronic device 130 described herein includes a mobile phone, a smart phone, a laptop computer, a digital broadcasting terminal, personal digital assistants (PDA), a portable multimedia player (PMP), a navigation device, and a slate. PCs (slate PCs), tablet PCs (tablet PCs), ultrabooks (ultrabooks), digital TVs, desktop PCs, and the like may be included.

In addition, the wireless communication 210 is for wireless Internet access, and may be built into or external to the sensor unit 110 and the control unit 120 . As a wireless communication technology, a short-range communication module for short-distance communication may be used, and short-range communication technologies include Bluetooth??, RFID (Radio Frequency Identification), infrared communication (Infrared Data Association; IrDA), Ultra Wideband (UWB), ZigBee, Near Field Communication (NFC), Wi-Fi Direct, and the like may be used. In addition, Wireless LAN (WLAN), Wireless Fidelity (WiFi) Direct, Digital Living Network Alliance (DLNA), Wireless broadband (Wibro), World Interoperability for Microwave Access (Wimax), and High Speed Downlink Packet Access (HSDPA) may be used. there is. In the embodiment, the application of Bluetooth short-range communication will be described as an example.

According to the embodiment, by attaching an acceleration sensor and/or a gyro sensor to a general ankle support, an accurate gait pattern can be recognized regardless of the condition of the road surface or the ground slope, and can be automatically transmitted to a display or electronic device using wireless communication. .

In addition, the data of the accelerometer and/or gyro sensor transmitted through wireless communication is stored in real-time data in the main microchip controller, and it can be compared and analyzed with the gait of a normal person and provided to a wrist watch-type display for real-time monitoring. there is.

3 is a schematic diagram showing the inside of a rehabilitation assisting device according to an embodiment.

Referring to FIG. 3, a schematic diagram of the inside of the control unit and the sensor unit, the sensor unit includes a power jack, a power control toggle switch, a central processing unit, a battery, a wireless communication module, an acceleration sensor, and a tact switch for a motion end signal, and the control unit, It may include a display unit, a power jack, a power control toggle switch, a central processing unit, a battery, and a wireless communication module.

It can be composed of OLED display, 1.8mm DC power jack, power control toggle switch, Arduino Pro Mini, 500mAh 3.7V Li-Po battery, Xbee S2C communication module, MPU 6050 acceleration sensor, and pain motor.

And Figure 3a is a diagram showing a schematic diagram of the inside of the sensor unit. Referring to FIG. 3A , the sensor unit includes a power jack 320, a power control toggle switch 330, a central processing unit 340, a battery 350, a wireless communication module 360, and an acceleration sensor 390 according to an embodiment. ) can be made including. On the other hand, the sensor unit is configured with a hanger so that it can be easily installed without a separate ankle protector.

The power jack 320 is a connection terminal for charging the battery 350, and for example, a 1.8 mm DC power jack (female) for a charging terminal for charging a lithium polymer battery may be used.

The power control toggle switch 330 is a switch that serves to select ON/OFF of power.

The central processing unit (MCU, 340) is responsible for the operation of internal signals, and for example, an Arduino pro mini atmega328 3.3V may be used. Here, the central processing unit 340 may perform calculation and processing of counting the number of steps by using data collected by the acceleration sensor 390 .

The battery 350 serves to supply power to the sensor unit, and for example, a 3.7V 500mAh lithium polymer (Li-Po) battery may be used. In this way, when using a lithium polymer battery, it is light in weight, can be recharged and re-discharged, and has little additional cost.

The wireless communication module 360 can use a Digimesh standard wireless communication module used for 1:2 and 1:n communication, and can communicate with the wireless communication module of the display unit through this wireless communication module 360. In addition, it is possible to communicate with an external terminal (eg, a user terminal, a computer, etc.) through the wireless communication module 360 . For example, the wireless communication module 360 may be an XbeeS2C wireless communication module. In addition, when connecting the Xbee to the central processing unit (MCU), a protection module may be further included to stably supply voltage and exchange signals. For example, a regulated board may be used.

The acceleration sensor 390 may collect the degree of acceleration and rotation occurring during walking. As the acceleration sensor 390, a 6-axis sensor including 3-axis acceleration and 3-axis gyro may be used, and for example, an MPU 6050 acceleration sensor may be used.

4 to 11 are diagrams illustrating a control unit of a rehabilitation assisting device for rehabilitation of a Parkinson's patient according to an embodiment. And Figure 3b is a diagram showing a schematic diagram of the inside of the control unit.

4 to 11 and FIG. 3B , the control unit according to an embodiment includes a display unit 310, a power jack 320, a power control toggle switch 330, a central processing unit 340, and a battery 350. ) and a wireless communication module 360. In addition, the control unit may further include a tact switch 370 for a motion end signal and a vibration motor 380. On the other hand, the control unit is configured in a wristwatch type so that the user can easily attach and detach it.

The display unit 310 may receive and display walking signals collected by the sensor unit and calculated and processed through the central processing unit of the sensor unit. For example, an OLED display may be used.

The power jack 320 is a connection terminal for charging the battery 350. For example, a 1.8 mm DC power jack (female) for a charging terminal for charging a lithium polymer battery may be used.

The power control toggle switch 330 is a switch that serves to select ON/OFF of the power of the display unit.

The central processing unit (MCU, 340) is responsible for the operation of internal signals, and for example, an Arduino pro mini atmega328 3.3V may be used.

The battery 350 serves to supply power to the display unit, and for example, a 3.7V 500mAh Li-Po battery may be used.

The wireless communication module 360 may be a Digimesh standard wireless communication module used for 1:2 and 1:n communication. Through this wireless communication module 360, it is possible to communicate with the wireless communication module of the sensor unit and communicate with an external terminal (eg, a user terminal, a computer, etc.). For example, the wireless communication module 360 may be an XbeeS2C wireless communication module. In addition, when connecting the Xbee to the central processing unit (MCU), a protection module may be further included to stably supply voltage and exchange signals. For example, a regulated board may be used.

The tact switch 370 for the motion end signal may perform the function of an external switch interrupt according to an algorithm.

The vibration motor 380 may transmit a vibration signal when an effective gait is not achieved during walking.

12 is a diagram for explaining an operation of adjusting difficulty using an artificial intelligence model according to an embodiment.

The feedback training system can perform a process of analyzing rehabilitation exercise information, a process of determining the current user's condition through clinical results, and a process of adjusting difficulty by using an artificial intelligence model. At this time, the artificial intelligence model is configured for gait pattern analysis and rehabilitation, and may be built based on various networks such as CNN, RNN, and LSTM and learned to adjust the difficulty of the user's rehabilitation exercise. For example, an artificial intelligence model can be learned through a data set for gait pattern analysis and rehabilitation, and result data can be output by inputting new input data for gait pattern analysis and rehabilitation to the learned artificial intelligence model. . In this way, gait data collected through rehabilitation exercises may be input to the learned artificial intelligence model, and the level of difficulty of the user's gait data may be adjusted through the learned artificial intelligence model.

The feedback training system analyzes rehabilitation exercise information including difficulty level, gait speed, and gait recognition rate using the collected gait data, and determines the user's current rehabilitation status information through clinical results for the analyzed rehabilitation exercise information. The level of difficulty may be adjusted based on the current rehabilitation state information of the user.

As the rehabilitation exercise begins, the feedback training system collects sensor data generated during walking using the acceleration sensor or gyro sensor worn on both ankles of the user based on the level of difficulty set for the user, and calculates and processes the collected sensor data. Thus, the number of steps recognized as walking may be counted, and the counted number of steps may be collected. The feedback training system stores only the highest acceleration data among the collected sensor data in real time, stores new maximum acceleration data again when the highest acceleration data is updated, and when the maximum acceleration data exceeds the standard acceleration for gait detection, gait number of times can be counted. The feedback training system may analyze gait data collected at the end of rehabilitation (eg, the total number of gaits and the number of gaits above a reference value) using an artificial intelligence model and adjust the level of difficulty.

As the rehabilitation exercise begins, the feedback training system collects the acceleration sensor value, records the number of recordings, and collects the acceleration sensor value when the gyro sensor value rotating around the Z axis obtained from the gyro sensor during walking exceeds the gyro limit value and rises. When the sensor value descends below the gyro threshold value, an average acceleration may be calculated, and it may be determined whether the calculated average acceleration exceeds a preset level of difficulty.

The feedback training system may count the number of valid steps and simple steps when the calculated average acceleration exceeds the difficulty level set for the user, and count the number of simple steps when the calculated average acceleration is less than the difficulty level set for the user. The feedback training system calculates the gait recognition rate through the ratio of the effective number of steps to the simple number of steps counted as the rehabilitation exercise is completed, and the calculated gait recognition rate exceeds the first reference gait recognition rate (eg, 80%). In this case, the current difficulty level may be stored, and a new difficulty level may be generated based on the stored current level of difficulty. In addition, when the calculated gait recognition rate is less than the first reference gait recognition rate (eg, 80%), the feedback training system compares the calculated gait recognition rate with the second reference gait recognition rate (eg, 75%), and calculates When the determined gait recognition rate is less than the second reference gait recognition rate (eg, 75%), the current difficulty level may be stored, and a new difficulty level may be created based on the stored current level of difficulty.

The feedback training system can adjust the level of difficulty for gait data collected by using an artificial intelligence model for gait pattern analysis and rehabilitation. The feedback training system may induce a user's gait pattern to a normal person's gait pattern through a difficulty adjustment signal generated based on the adjusted difficulty level. The difficulty adjustment signal generated for the difficulty adjusted by the controller may be transmitted to the sensor unit. The sensor unit may provide feedback of the user's gait pattern based on the difficulty adjustment signal.

13 is a graph for explaining a relationship between a level of difficulty and a gait recognition rate, according to an embodiment.

As the AI model is used, the recognition rate may decrease, but even if the recognition rate decreases, the difficulty may increase. In other words, it can be determined that the Parkinson's patient is performing intensive rehabilitation exercise. If the level of difficulty (reference value) is set too high, the user's gait posture, gait speed, and gait swing width are all different, so in the case of severe Parkinson's patients, it may be unreasonable to ask a normal person to walk from the beginning, and the achievement will be reduced. It can lead to a lack of self-confidence. Therefore, as the difficulty setting is adjusted with the rehabilitation training data, as seen in the simulated rehabilitation diary, when the gait recognition rate above the difficulty level by rehabilitation day rises and falls at about 80%, the difficulty level also rises and falls according to the recognition rate, but overall points upward Not only does it boost your confidence, but it can also help you maximize the effectiveness of your rehabilitation.

14 is a flowchart illustrating a feedback training method for rehabilitation in a rehabilitation assisting device for rehabilitation of a Parkinson's patient according to an embodiment.

A feedback training operation for rehabilitation through the sensor unit 110 and the controller 120 configured in the rehabilitation assistance device will be described.

Device power may be input to each of the sensor unit 110 and the control unit 120 (1410 and 1420). Power of each of the sensor unit 110 and the control unit 120 may be switched from off to on.

The sensor unit 110 may transmit the user's difficulty (gait difficulty) to the controller 120 (1421). The sensor unit 110 may transmit the current user's level of difficulty to the controller 120 through wireless communication such as Bluetooth. The control unit 120 may wait until the reception of the level of difficulty from the sensor unit 110 is completed (1411).

Based on the difficulty of the user, the sensor unit 110 may sense that the rehabilitation exercise is repeated until the rehabilitation of the walking exercise program is completed (1422). The sensor unit 110 may transmit gait data generated until rehabilitation of the gait exercise program is completed to the controller 120 . The controller 120 may output gait data generated from the sensor unit 110 to a display in real time (1412). As shown in FIG. 12, the controller 120 analyzes and rehabilitates gait data collected at the end of rehabilitation of the gait exercise program (for example, the total number of steps and the number of steps equal to or greater than a reference value) using an artificial intelligence model. You can adjust the difficulty for The adjusted level of difficulty may be transmitted to the sensor unit 120 and stored when the device shuts down. The switch may be pressed through an external interrupt configured in the control unit 120 (1413). The control unit 120 may transmit a difficulty adjustment signal generated based on the collected gait data to the sensor unit 110 (1414). The sensor unit 110 may adjust the level of difficulty through a level of difficulty adjustment signal based on the exercise result (1423).

The power of each device of the sensor unit 110 and the control unit 120 may be terminated (1415 and 1424). Power to each of the sensor unit 110 and the control unit 120 may be switched from on to off.

15 to 17 are detailed flowcharts illustrating a feedback training method for rehabilitation in a rehabilitation assisting device for rehabilitation of a Parkinson's patient according to an embodiment.

Referring to FIG. 15 , device power may be input to each of the sensor unit 110 and the control unit 120 (1410 and 1420). Power of each of the sensor unit 110 and the control unit 120 may be switched from off to on.

The sensor unit 110 may retrieve the patient's current difficulty level from the internal storage. The sensor unit 110 may display the stored level of difficulty after the past rehabilitation exercise. The sensor unit 110 may transmit the difficulty to the controller 120 in the form of [SOS] + difficulty + [EOS].

The controller 120 may wait for xbee communication in a state in which the sensor difficulty level is not received. The controller 120 may determine the reception history of the sensor difficulty based on the data recorded in the string. The control unit 120 may receive data from the sensor unit 110 through xbee communication when the reception history of the sensor difficulty does not exist.

The controller 120 may extract 'EOS' data when a sensor difficulty reception history exists. The controller 120 may store the received data in a string, return the string as a number, and store the level of difficulty. The control unit 120 may delete the sensor difficulty reception history and make the character string empty. The controller 120 may determine whether the sensor difficulty level is received.

In addition, the controller 120 may extract 'SOS' data when there is no sensor difficulty reception history. The controller 120 may determine the sensor difficulty reception history based on the [SOS] data. At this time, if a sensor difficulty reception history exists, data may be re-received from the sensor unit 110 .

The sensor unit 110 may detect that a rehabilitation exercise starts from the user. The sensor unit 110 may collect sensor data (eg, gyro sensor value) sensed according to the user's walking. The sensor unit 110 may determine whether a gyro sensor value exceeds a gyro threshold value. When the gyro sensor value exceeds the gyro threshold value, the sensor unit 110 may calculate the sum of the accelerations by summing the acceleration distance values. The sensor unit 110 may increase the number of recordings through the calculated sum of accelerations. When the gyro sensor value does not exceed the gyro threshold value, the sensor unit 110 may determine whether the gyro sensor value is less than the gyro threshold value. The sensor unit 110 may calculate an average acceleration when the gyro sensor value is less than the gyro threshold value. At this time, the acceleration (value) may be calculated using the acceleration distance. The average acceleration may be derived by calculating the sum of the accelerations/the number of recordings. The sensor unit 110 may set the sum of the accelerations and the number of recordings to zero.

Referring to FIG. 16 , the sensor unit 110 may determine whether the average acceleration exceeds the degree of difficulty. When the average acceleration exceeds the level of difficulty, the sensor unit 110 may transmit ''X' data to the controller 120 by increasing the effective number of steps and the number of simple steps, and when the average acceleration does not exceed the level of difficulty 'A' data may be transmitted to the controller 120 by simply increasing the number of steps.

The control unit 120 may wait for xbee communication. The control unit 120 may receive communication data transmitted from the sensor unit 110 . The controller 120 may increase the effective number of steps when the received communication data is 'X' data, and increase the simple number of steps when the received communication data is 'A' data. The controller 120 may determine whether to operate the vibration motor according to an increase in the simple number of steps. The controller 120 may record the operation start time and operate the vibration motor. The controller 120 may stop the operation of the vibration motor as 1 second has elapsed. The controller 120 may determine not to operate the vibration motor. The controller 120 may display the level of difficulty and the number of steps on the display, and may wait for xbee communication.

Referring to FIG. 17 , an external switch interrupt may be generated in the control unit 120 . The controller 120 may transmit 'INT' data to the sensor unit 110 . The controller 120 may calculate the total number of steps by adding the simple number of steps and the effective number of steps. The controller 120 may calculate a gait recognition rate based on the total number of gaits. The gait recognition rate may be derived by calculating (effective number of steps * 100)/total number of steps. The controller 120 may display the gait result through the display.

The sensor unit 110 may receive 'INT' data from the control unit 120 . The sensor unit 110 may determine whether the data is 'INT' data. If the data is not 'INT' data, the sensor unit 110 may determine whether the gyro sensor value exceeds the gyro threshold value. If the data is 'INT' data, the sensor unit 110 may determine that the rehabilitation exercise is not in a starting state. In other words, the user's rehabilitation exercise may end.

The sensor unit 110 may calculate a gait recognition rate. The gait recognition rate may be calculated through (effective number of steps * 100)/simple number of steps. It may be determined whether the gait recognition rate calculated by the sensor unit 110 exceeds 80%. When the gait recognition rate exceeds 80%, the sensor unit 110 may store the current level of difficulty in a space without data. The sensor unit 110 may generate a new difficulty level by adding 100 to the stored current difficulty level, and may store the generated new difficulty level.

When the gait recognition rate is less than 75%, the sensor unit 110 may store the current level of difficulty in a space without data. The sensor unit 110 may generate a new difficulty level by adding 100 to the stored current difficulty level, and may store the generated new difficulty level. The sensor unit 110 may shut down the device.

The devices described above may be implemented as hardware components, software components, and/or a combination of hardware components and software components. For example, devices and components described in the embodiments may include, for example, a processor, a controller, an arithmetic logic unit (ALU), a digital signal processor, a microcomputer, a field programmable gate array (FPGA) , a programmable logic unit (PLU), microprocessor, or any other device capable of executing and responding to instructions. A processing device may run an operating system (OS) and one or more software applications running on the operating system. A processing device may also access, store, manipulate, process, and generate data in response to execution of software. For convenience of understanding, there are cases in which one processing device is used, but those skilled in the art will understand that the processing device includes a plurality of processing elements and/or a plurality of types of processing elements. It can be seen that it can include. For example, a processing device may include a plurality of processors or a processor and a controller. Other processing configurations are also possible, such as parallel processors.

Software may include a computer program, code, instructions, or a combination of one or more of the foregoing, which configures a processing device to operate as desired or processes independently or collectively. The device can be commanded. Software and/or data may be any tangible machine, component, physical device, virtual equipment, computer storage medium or device, intended to be interpreted by or provide instructions or data to a processing device. can be embodied in Software may be distributed on networked computer systems and stored or executed in a distributed manner. Software and data may be stored on one or more computer readable media.

The method according to the embodiment may be implemented in the form of program instructions that can be executed through various computer means and recorded on a computer readable medium. The computer readable medium may include program instructions, data files, data structures, etc. alone or in combination. Program commands recorded on the medium may be specially designed and configured for the embodiment or may be known and usable to those skilled in computer software. Examples of computer-readable recording media include magnetic media such as hard disks, floppy disks and magnetic tapes, optical media such as CD-ROMs and DVDs, and magnetic media such as floptical disks. - includes hardware devices specially configured to store and execute program instructions, such as magneto-optical media, and ROM, RAM, flash memory, and the like. Examples of program instructions include high-level language codes that can be executed by a computer using an interpreter, as well as machine language codes such as those produced by a compiler.

As described above, although the embodiments have been described with limited examples and drawings, those skilled in the art can make various modifications and variations from the above description. For example, the described techniques may be performed in an order different from the method described, and/or components of the described system, structure, device, circuit, etc. may be combined or combined in a different form than the method described, or other components may be used. Or even if it is replaced or substituted by equivalents, appropriate results can be achieved.

Therefore, other implementations, other embodiments, and equivalents of the claims are within the scope of the following claims.

Claims (15)

In the feedback training method for rehabilitation performed by the feedback training system,
collecting gait data of a user from sensor data sensed according to gait of a user wearing a rehabilitation assistance device;
adjusting the level of difficulty for the collected gait data using an artificial intelligence model for gait pattern analysis and rehabilitation; and
Inducing a user's gait pattern to a normal person's gait pattern through a difficulty adjustment signal generated based on the adjusted difficulty level
including,
The adjusting step is
Using the collected gait data, rehabilitation exercise information including difficulty, gait speed, and gait recognition rate is analyzed, and current rehabilitation state information of the user is determined through clinical results for the analyzed rehabilitation exercise information, and the determined The level of difficulty is adjusted based on the user's current rehabilitation state information, and when the gyro sensor value rotating around the Z axis obtained from the gyro sensor during walking rises beyond the gyro limit value as the rehabilitation exercise begins, the acceleration sensor value is collected and the number of times is recorded, when the gyro sensor value falls below the gyro threshold value, an average acceleration is calculated, it is determined whether the calculated average acceleration exceeds a difficulty level set for the user, and the calculated average acceleration counting the effective number of steps and the number of simple steps when the value exceeds the difficulty level set for the user, and counting the number of simple steps when the calculated average acceleration is less than the level of difficulty set for the user;
including,
The inducing step is
Generating vibration through a vibration unit built into a wrist watch-type display when the user fails to achieve an effective gait when walking based on the gait speed and gait pattern including the adjusted level of difficulty.
Feedback training method for rehabilitation comprising a. delete delete delete According to claim 1,
The adjusting step is
When the rehabilitation exercise is completed, a gait recognition rate is calculated through a ratio of the effective number of steps to the counted simple number of steps, and when the calculated gait recognition rate exceeds a first reference gait recognition rate, a current difficulty level is stored. Generating a new difficulty level based on the stored current difficulty level
Feedback training method for rehabilitation comprising a. According to claim 1,
The adjusting step is
When the calculated gait recognition rate is less than the first reference gait recognition rate, the calculated gait recognition rate is compared with the second reference gait recognition rate, and when the calculated gait recognition rate is less than the second reference gait recognition rate, the current difficulty level is stored; Generating a new difficulty level based on the stored current difficulty level
Feedback training method for rehabilitation comprising a. According to claim 1,
The collecting step is
As the rehabilitation exercise starts, sensor data generated during walking is collected using the acceleration sensor or gyro sensor worn on both ankles of the user based on the difficulty level set for the user, and the collected sensor data is calculated and processed to walk Counting the number of steps recognized as , and collecting the counted number of steps
Feedback training method for rehabilitation comprising a. According to claim 7,
The collecting step is
Among the collected sensor data, only the highest acceleration data is stored in real time, when the highest acceleration data is updated, new maximum acceleration data is stored again, and when the maximum acceleration data exceeds the standard acceleration for gait detection, the number of steps is counted. step to count
Feedback training method for rehabilitation comprising a. According to claim 1,
The inducing step is
Displaying rehabilitation state information related to the gait speed and gait pattern including the adjusted level of difficulty on a display
Feedback training method for rehabilitation comprising a. According to claim 9,
The inducing step is
Displaying the number of times counted as effective walking, the current difficulty, and displaying the result of the rehabilitation exercise, including whether the difficulty of the next rehabilitation exercise is raised or lowered based on the gait achievement rate and the current difficulty
Feedback training method for rehabilitation comprising a. According to claim 1,
The inducing step is
Connecting to an electronic device through wireless communication through a wireless communication module and transmitting rehabilitation state information related to a walking speed and a walking pattern including the adjusted difficulty level to the electronic device.
Feedback training method for rehabilitation comprising a. delete In a computer program stored in a computer readable storage medium to execute a feedback training method for rehabilitation performed by a feedback training system,
collecting gait data of a user from sensor data sensed according to gait of a user wearing a rehabilitation assistance device;
adjusting the level of difficulty for the collected gait data using an artificial intelligence model for gait pattern analysis and rehabilitation; and
Inducing a user's gait pattern to a normal person's gait pattern through a difficulty adjustment signal generated based on the adjusted difficulty level
including,
The adjusting step is
Using the collected gait data, rehabilitation exercise information including difficulty, gait speed, and gait recognition rate is analyzed, and current rehabilitation state information of the user is determined through clinical results for the analyzed rehabilitation exercise information, and the determined The level of difficulty is adjusted based on the user's current rehabilitation state information, and when the gyro sensor value rotating around the Z axis obtained from the gyro sensor during walking rises beyond the gyro limit value as the rehabilitation exercise begins, the acceleration sensor value is collected and the number of times is recorded, when the gyro sensor value falls below the gyro threshold value, an average acceleration is calculated, it is determined whether the calculated average acceleration exceeds a difficulty level set for the user, and the calculated average acceleration counting the effective number of steps and the number of simple steps when the value exceeds the difficulty level set for the user, and counting the number of simple steps when the calculated average acceleration is less than the level of difficulty set for the user;
including,
The inducing step is
Generating vibration through a vibration unit built into a wrist watch-type display when the user fails to achieve an effective gait when walking based on the gait speed and gait pattern including the adjusted level of difficulty.
A computer program stored in a computer readable storage medium comprising a. In the feedback training system,
at least one processor configured to execute computer readable instructions contained in memory;
including,
The at least one processor,
The user's gait data is collected from sensor data detected according to the gait of the user wearing the rehabilitation assistive device;
Adjusting the level of difficulty for the collected gait data using an artificial intelligence model for gait pattern analysis and rehabilitation,
Inducing a user's gait pattern to a normal person's gait pattern through a difficulty adjustment signal generated based on the adjusted difficulty level
including,
The at least one processor analyzes rehabilitation exercise information including difficulty level, gait speed, and gait recognition rate using the collected gait data, and present rehabilitation state information of the user through clinical results for the analyzed rehabilitation exercise information. Determines, adjusts the difficulty based on the determined user's current rehabilitation state information, and as the rehabilitation exercise starts, the gyro sensor value rotating around the Z axis obtained from the gyro sensor during walking exceeds the gyro limit value When ascending, the acceleration sensor value is collected and the number of times is recorded, and when the gyro sensor value descends below the gyro threshold value, an average acceleration is calculated and it is determined whether the calculated average acceleration exceeds the difficulty level set for the user. and if the calculated average acceleration exceeds the difficulty level set for the user, counts the effective number of steps and simple number of steps, and if the calculated average acceleration is less than the level of difficulty set for the user, counts the number of simple steps, and adjusts When the user fails to achieve an effective gait based on the gait speed and gait pattern including the difficulty level, vibration is generated through a vibration unit built into a wrist watch-type display.
Feedback training system, characterized in that. delete

Images