TWI880166B - Intelligent automated rehabilitation training device and method thereof - Google Patents

Abstract

An intelligent automated rehabilitation training device includes a data collecting module and a processing module. The data collecting module receive a first body condition data of a user at a first time point and the first body condition data includes a first heart rate variability. The processing module receives the first body condition data and has a decision making model. The decision making model is obtained by inputting a historical training data into a deep reinforcement learning model to execute a training process. The historical training data includes the training data of a plurality of training targets. The training data of each training target includes the heart rate variability of each training target performing an initial rehabilitation training each time and the feedback information corresponding thereto. The processing module inputs the first body condition data into the decision making model in order to generate a first rehabilitation prescription, such that the user can perform the first rehabilitation prescription after the first time point.

Description

Intelligent automated rehabilitation training device and method

This disclosure is about a rehabilitation training device, in particular, an intelligent automated rehabilitation training device. This disclosure also relates to an intelligent automated rehabilitation training method for the device.

An aging society has become a problem that all countries need to face, and the number of elderly people in each country is constantly increasing. Elderly people need to take in enough nutrition and do regular rehabilitation exercises to effectively improve frailty symptoms.

Existing rehabilitation training is usually carried out according to the doctor's rehabilitation prescription. The doctor needs to ask the elderly about their subjective feelings and adjust the rehabilitation prescription according to the elderly's feedback. Therefore, the elderly need to return for frequent consultations so that the doctor can understand the elderly's rehabilitation status and adjust the rehabilitation prescription.

Some existing rehabilitation training can be conducted through motion-sensing video games. Medical staff can adjust the difficulty of the motion-sensing video games based on the training effect of the elderly. However, this method requires medical staff to actively intervene and adjust, which not only consumes manpower resources, but also lacks efficiency.

Some existing rehabilitation training can be conducted through games. However, this method does not provide any rehabilitation prescription adjustment mechanism, so it is difficult to improve the training effect of rehabilitation training.

According to an embodiment of the present disclosure, a smart automated rehabilitation training device is proposed, which includes a data collection module and a processing module. The data collection module receives the first body state data of the user at a first time point, and the first body state data includes a first heart rate variability. The processing module receives the first body state data and has a decision model. The decision model is obtained by inputting historical training data into a deep reinforcement learning (DRL) model to execute a training program, and the historical training data includes the heart rate variability and corresponding feedback information of each time the user executes the initial rehabilitation prescription. Among them, the processing module inputs the first body state data into the decision model to generate a first rehabilitation prescription, so that the user can execute the first rehabilitation prescription after the first time point.

According to another embodiment of the present disclosure, a smart automated rehabilitation training method is proposed, which includes the following steps: inputting historical training data into a deep reinforcement learning model to execute a training procedure to obtain a decision model, wherein the historical training data includes the heart rate variability and corresponding feedback information of each time the user executes the initial rehabilitation prescription; receiving the user's first body state data at a first time point, wherein the first body state data includes a first heart rate variability; inputting the first body state data into the decision model to generate a first rehabilitation prescription; and allowing the user to execute the first rehabilitation prescription after the first time point.

1: Intelligent automated rehabilitation training device

11: Data collection module

12: Processing module

13: Input module

14: Display module

A1: The first rejuvenating prescription

A2: The second rehabilitation prescription

S1, S2: Health status

P1: First body status data

P2: Second body status data

S41~S47: Step flow

Figure 1 is a block diagram of the first embodiment of the intelligent automated rehabilitation training device disclosed herein.

Figure 2 is the first schematic diagram of the operating state of one embodiment of the intelligent automated rehabilitation training device disclosed herein.

Figure 3 is a second schematic diagram of the operating state of one embodiment of the intelligent automated rehabilitation training device disclosed herein.

Figure 4 is a flow chart of an embodiment of the intelligent automated rehabilitation training method disclosed herein.

The following will refer to the relevant drawings to illustrate the embodiments of the intelligent automated rehabilitation training device and method disclosed herein. For the sake of clarity and convenience of the drawings, the components in the drawings may be exaggerated or reduced in size and proportion. In the following description and/or patent application, when it is mentioned that an element is "connected" or "coupled" to another element, it may be directly connected or coupled to the other element or there may be an intervening element; and when it is mentioned that an element is "directly connected" or "directly coupled" to another element, there is no intervening element. Other words used to describe the relationship between elements or layers should be interpreted in the same way. For ease of understanding, the same elements in the following embodiments are illustrated with the same symbols.

Please refer to Figures 1 and 2, which are block diagrams of an embodiment of the intelligent automated rehabilitation training device disclosed herein and a first schematic diagram of the operating state. As shown in Figure 1, the intelligent automated rehabilitation training device 1 includes a data collection module 11, a processing module 12, an input module 13 and a display module 14. The intelligent automated rehabilitation training device 1 can be an electronic device, such as a smart phone, a smart watch, a smart bracelet, a tablet computer, a personal computer, a notebook computer, etc. The data collection module 11, the input module 13 and the display module 14 are connected to the processing module 12.

As shown in FIG. 2, the data collection module 11 receives the first physical state data P1 of the user at the first time point. The first physical state data P1 includes the first heart rate variability; in addition, the first physical state data P1 also includes one or more of name, age, blood pressure, heart rate, blood sugar, body mass index (BMI), and health status S1. Among them, the health status S1 can be divided into several levels, such as frail, normal, or pre-frail; the level and quantity of the health status S1 can be defined according to actual needs, and the present disclosure is not limited to this. In one embodiment, the data collection module 11 can be a memory, such as a random access memory (RAM), a hard disk, or other similar devices. In another embodiment, the data collection module 11 can also be software, which can be used to control the storage of data.

The input module 13 can be used to receive the above-mentioned health status S1. The user can actively input his health status S1 through the input module 13. In one embodiment, the input module 13 can be a touch screen, a keyboard or other similar devices. In another embodiment, the input module 13 can also be software, which can be used to receive, analyze and convert the user's input signal.

Next, the processing module 12 receives the first body state data P1, and inputs the first body state data P1 into the decision model to generate the first re-enabling prescription A1, and transmits it to the display module 14. In one embodiment, the processing module 12 can be an application specific integrated circuit chip (ASIC), a digital signal processor (DSP), a digital signal processing device (DSPD), a programmable logic device (PLD), a field programmable logic gate array (FPGA), a central processing unit (CPU), a controller, a microcontroller, a microprocessor or other similar devices. In another embodiment, the processing module 12 can also be software, which can be used for data analysis and calculation.

Then, the display module 14 (such as a screen, etc.) displays the first rehabilitation prescription A1, so the user can obtain the first rehabilitation prescription A1 through the display module 14 and execute the first rehabilitation prescription A1 after the first time point to perform rehabilitation training.

Please refer to Figure 3, which is a second schematic diagram of the operating state of an embodiment of the intelligent automated rehabilitation training device disclosed herein. As shown in the figure, the data collection module 11 receives the second body state data P2 of the user at the second time point. The second body state data P2 includes a second heart rate variability; in addition, the second body state data P2 also includes one or more of name, age, blood pressure, heart rate, blood sugar, body mass index, and health status. Similarly, the user can input his health status S2 through the input module 13.

Next, the processing module 12 receives the second body state data P2, and inputs the second body state data P2 into the decision model to generate a second rehabilitation prescription A2, and transmits it to the display module 14.

Then, the display module 14 displays the second rehabilitation prescription A2, so the user can obtain the second rehabilitation prescription A2 through the display module 14 and execute the second rehabilitation prescription A2 after the second time point to perform rehabilitation training. In one embodiment, the first rehabilitation prescription A1 and the second rehabilitation prescription A2 are continuous rehabilitation prescriptions or discrete rehabilitation prescriptions.

Similarly, the intelligent automated rehabilitation training device 1 can continuously generate the third rehabilitation prescription, the fourth rehabilitation prescription, the fifth rehabilitation prescription, the sixth rehabilitation prescription, etc. through the same mechanism, so that the user can continuously and uninterruptedly carry out effective rehabilitation training. .

Therefore, before each rehabilitation training, the intelligent automated rehabilitation training device 1 analyzes and calculates the user's physical state data (including heart rate variability) before and after the rehabilitation training through the decision model to automatically generate a rehabilitation prescription that best matches the user's physical state data. Then, the user can perform rehabilitation training according to this rehabilitation prescription to improve the effect of rehabilitation training. In this way, the user does not need to frequently return to the doctor and can continue to effectively perform rehabilitation training in a different place (such as at home), so it can ensure that the user can perform rehabilitation training normally and achieve the best effect of rehabilitation training.

The aforementioned decision model is obtained by inputting historical training data into the deep reinforcement learning model to execute the training process. The historical training data can be obtained by the wearable electronic device or mobile device equipped by the user. The historical training data includes the training data of multiple training subjects. The training data of each training subject includes the heart rate variability and corresponding feedback information before and after each execution of the initial rehabilitation prescription by each training subject.

When executing the training program, the initial rehabilitation prescription of each training object provided by the doctor and the physical state data including the heart rate variability of the training object before executing the rehabilitation prescription can be input into the deep reinforcement learning model; the aforementioned physical state data can also include one or more of name, age, blood pressure, heart rate, blood sugar, body mass index, and health status. In this embodiment, the heart rate variability can be LF HRV (or HF HRV). Generally speaking, when the training object feels bored, the value of its heart rate variability is higher; and when the training object feels anxious, the value of its heart rate variability is the lowest; when the training object reaches the flow state, the value of its heart rate variability is moderate. In another embodiment, the heart rate variability may be RMSSD, SDNN, pNN50 or VLF. The aforementioned health status may be input by the training subject himself or evaluated by a doctor, and may also be divided into several levels, such as frail, normal or pre-frail, which may be defined according to actual needs, and the present disclosure is not limited thereto.

Then, when the training subject completes the initial rehabilitation prescription, the deep reinforcement learning model receives the subjective feelings of the training subject after executing the initial rehabilitation prescription (which can be input by the training subject himself), and the subjective feelings can reflect the flow state of the training subject. In one embodiment, the subjective feelings can be divided into several levels, such as relaxed, slightly breathless, and very breathless. In another embodiment, the subjective feelings can be divided into anxiety, flow, and boredom; for example, subjective feeling F1=-1 means very breathless; subjective feeling F1=5 means slightly breathless; subjective feeling F1=-5 means relaxed. The level and quantity of subjective feelings can be defined according to actual needs, and this disclosure is not limited to this. The above definition is biased towards the training object feeling that the rehabilitation prescription has a certain exercise intensity to achieve a high score, not too easy (F1=-5), nor too strenuous (F1=-1), so as to achieve a better rehabilitation training effect.

The deep reinforcement learning model calculates feedback information based on the subjective feelings of the training subject after executing the initial rehabilitation prescription and the heart rate variability before executing the initial rehabilitation prescription, and obtains the physical condition data of the training subject after executing the initial rehabilitation prescription (part of the data can be input by the user, and part of the data can be measured by the device equipped by the user). The deep reinforcement learning model can obtain feedback information by calculating the weighted sum of the subjective feelings of the training subject after executing the initial rehabilitation prescription and the heart rate variability; in the above calculation, the weight of the subjective feelings of the training subject after executing the initial rehabilitation prescription can be less than the weight of the heart rate variability. For example, the weight of the subjective feeling F1 of the training subject after executing the initial rehabilitation prescription may be 0.4, and the weight of the first heart rate variability H1 may be 0.6, and the first feedback information R1 is 0.4*F1+0.6*H1. In another embodiment, the weight of the subjective feeling F1 of the training subject after executing the initial rehabilitation prescription may be greater than or equal to the weight of the first heart rate variability H1, which may be appropriately adjusted according to actual needs, and the present disclosure is not limited thereto. Similarly, the aforementioned initial rehabilitation prescription can be a continuous rehabilitation prescription (which can be, but is not limited to, riding an indoor exercise bike for 15 to 60 minutes...) or a discrete rehabilitation prescription (which can be, but is not limited to, riding an indoor exercise bike for 15 minutes, 30 minutes, 60 minutes...).

Next, the deep reinforcement learning model can continuously collect and store the physical state data (including heart rate variability) and corresponding feedback data of this training object before and after multiple executions of this initial rehabilitation prescription as training data for this training object. The deep reinforcement learning model collects and stores the training data of all training objects in the same way as the aforementioned historical training data.

Finally, the training process can be carried out based on the deep reinforcement learning model and the aforementioned historical training data to establish a decision model.

Since the decision model is based on the physical state data and feedback information including heart rate variability for deep reinforcement learning training, the physical state data and feedback opinions (feedback information) of each trainee before and after the initial rehabilitation prescription are considered at the same time. In addition, the feedback opinions of this trainee integrate the subjective feelings of this trainee after the initial rehabilitation prescription and the heart rate variability of this trainee when the initial rehabilitation prescription is executed. In other words, when executing the initial rehabilitation prescription, the physical strength of this trainee must be able to achieve a flow state in order to obtain a higher subjective feeling score. For example, if the training subject has poor sleep quality, the training subject will still feel strenuous when executing the same rehabilitation prescription, so the subjective feeling score reported by the training subject is lower; on the contrary, if the training subject gets enough rest and his physical strength is significantly improved but the intensity of the rehabilitation prescription remains unchanged, the training subject will feel too relaxed and effortless when executing the rehabilitation prescription, so the subjective feeling score reported by the training subject is still lower.

Therefore, the decision model of this embodiment calculates based on the user's physical condition data and feedback information to generate the first rehabilitation prescription A1, and adjusts the first rehabilitation prescription A1 to generate the second rehabilitation prescription A2. In this way, the rehabilitation prescription provided by the intelligent automated rehabilitation training device 1 can meet the user's current comprehensive health status, allowing the user to feel a certain degree of burden when executing the rehabilitation prescription provided by the intelligent automated rehabilitation training device 1 but not too strenuous.

From the above, it can be seen that the intelligent automated rehabilitation training device 1 can effectively integrate the deep reinforcement learning technology based on artificial intelligence with the flow theory, and automatically adjust the rehabilitation prescription according to the subjective information including subjective feelings and the objective information including heart rate variability. Therefore, it is easy for users to achieve the flow state when performing rehabilitation training, which greatly improves the training effect of rehabilitation training, and can effectively improve the user's weakness symptoms.

In addition, users do not need to return to the hospital frequently and can continue to effectively perform rehabilitation training in a different place (such as at home), so it can ensure that users can perform rehabilitation training normally and enable rehabilitation training to achieve the best effect.

The aforementioned decision model can be obtained by inputting historical training data into a deep reinforcement learning (DRL) model based on artificial intelligence to execute a training procedure. The historical training data may include training data of multiple training subjects, and the training data of each training subject includes the heart rate variability and corresponding feedback information before and after each execution of the initial rehabilitation prescription by the training subject (and may also include data collected from several follow-up visits of each training subject). In one embodiment, the deep reinforcement learning model may adopt DQN (Deep Q-Learning Network) algorithm, A3C (Asynchronous Advantage Actor-Critic), etc. These data can be input into a computer device that has established this deep reinforcement learning model, and the historical training data can be input into the memory of this deep reinforcement learning model through this computer device to execute the training program. Therefore, the decision model generated by the above mechanism can effectively integrate the deep reinforcement learning technology based on artificial intelligence and the flow theory, and automatically adjust the rehabilitation prescription according to the subjective information including subjective feelings and the objective information including heart rate variability.

Of course, this embodiment is only used for illustration and does not limit the scope of the present disclosure. Equivalent modifications or changes made to the intelligent automated rehabilitation training device of this embodiment should still be included in the patent scope of the present disclosure.

It is worth mentioning that existing rehabilitation training is usually carried out according to the rehabilitation prescription of the doctor. Therefore, the elderly need to return for frequent consultations so that the doctor can adjust the rehabilitation prescription according to the elderly's feedback. Some existing rehabilitation training can be carried out through somatosensory video games. Medical staff can adjust the difficulty of somatosensory video games according to the training effect of the elderly. However, this method requires medical staff to actively intervene and adjust, which not only consumes manpower resources, but also lacks efficiency. On the contrary, according to the embodiment of the present disclosure, the intelligent automated rehabilitation training device can effectively integrate the deep reinforcement learning technology based on artificial intelligence with the flow theory, and automatically adjust the rehabilitation prescription according to the subjective information including subjective feelings and the objective information including heart rate variability, so that the user can easily achieve the flow state when performing rehabilitation training, greatly improving the training effect of rehabilitation training, and thus effectively improving the user's weakness symptoms.

In addition, the intelligent automated rehabilitation training device can effectively integrate the deep reinforcement learning technology based on artificial intelligence with the flow theory, and at the same time grasp the user's rehabilitation status and automatically adjust the rehabilitation prescription appropriately. Therefore, users do not need to return to the hospital frequently and can continue to effectively perform rehabilitation training in a different place (such as at home), so it can ensure that users can perform rehabilitation training normally and make rehabilitation training play the best effect.

In addition, the intelligent automated rehabilitation training device can grasp the user's rehabilitation status and automatically adjust the rehabilitation prescription appropriately, without the need for medical staff to actively intervene and adjust. Therefore, the intelligent automated rehabilitation training device can not only significantly reduce the consumption of human resources to reduce the burden on medical staff, but also effectively improve the efficiency of rehabilitation training and better meet the needs of practical applications. From the above, it can be seen that the intelligent automated rehabilitation training device according to the embodiment of the present disclosure can indeed have optimized performance and achieve excellent technical effects.

Please refer to Figure 4, which is a flow chart of an embodiment of the intelligent automated rehabilitation training method disclosed herein. As shown in the figure, the intelligent automated rehabilitation training method of this embodiment may include the following steps:

Step S41: Input historical training data into the deep reinforcement learning model to execute the training procedure to obtain a decision model. The historical training data includes training data of multiple training subjects. The training data of each training subject includes the heart rate variability and corresponding feedback information of each time the training subject executes the initial rehabilitation prescription.

Step S42: Receive the user's first physical state data at a first time point, the first physical state data including a first heart rate variability.

Step S43: Input the first body state data into the decision model to generate the first rehabilitation prescription.

Step S44: The user executes the first rehabilitation prescription after the first time point.

Step S45: Receive the user's second physical state data at a second time point after the first time point, the second physical state data including a second heart rate variability.

Step S46: Input the second body state data into the decision model to generate a second rehabilitation prescription.

Step S47: The user executes the second rehabilitation prescription after the second time point.

Of course, this embodiment is only used for illustration and does not limit the scope of the present disclosure. Equivalent modifications or changes made based on the intelligent automated rehabilitation training method of this embodiment should still be included in the patent scope of the present disclosure.

In summary, according to the embodiments disclosed herein, the intelligent automated rehabilitation training device can effectively integrate the deep reinforcement learning technology based on artificial intelligence with the flow theory, and automatically adjust the rehabilitation prescription according to the subjective information including subjective feelings and the objective information including heart rate variability, so that the user can easily achieve the flow state when performing rehabilitation training, greatly improving the training effect of rehabilitation training, and thus effectively improving the user's weakness symptoms.

In addition, the intelligent automated rehabilitation training device can effectively integrate the deep reinforcement learning technology based on artificial intelligence with the flow theory, and at the same time grasp the user's rehabilitation status and automatically adjust the rehabilitation prescription appropriately. Therefore, users do not need to return to the hospital frequently and can continue to effectively perform rehabilitation training in a different place (such as at home), so it can ensure that users can perform rehabilitation training normally and make rehabilitation training play the best effect.

In addition, the intelligent automated rehabilitation training device can grasp the user's rehabilitation status and automatically adjust the rehabilitation prescription appropriately, without the need for medical staff to actively intervene and adjust. Therefore, the intelligent automated rehabilitation training device can not only greatly reduce the consumption of human resources to reduce the burden on medical staff, but also effectively improve the efficiency of rehabilitation training and better meet the needs of actual applications.

Although the steps of the methods described in the present disclosure are shown and described in a particular order, the order of operation of each method can be changed, and some steps can be performed in the reverse order, or some steps can be performed simultaneously with other steps. In another embodiment, different steps can be implemented in an intermittent and/or alternating manner.

It should be noted that at least some steps of the method described in the present disclosure can be performed by software instructions stored on a computer-usable storage medium for execution by a computer (or processor). For example, an example of a computer program product includes a computer-usable storage medium for storing a computer-readable program.

Computer-usable storage media or computer-readable storage media can be electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems (or devices, equipment, etc.). Examples of non-transitory computer-usable and computer-readable storage media include semiconductor or solid-state memory, magnetic tape, portable floppy disk, random access memory (RAM), hard disk, and optical disk. Examples of optical disks include compact disk with read-only memory (CD-ROM), compact disk with rewritable memory (CD-R/W), and digital versatile disk (DVD).

In addition, each embodiment of the present disclosure (or each module of the system) can be implemented entirely in hardware, entirely in software, or in an implementation method that includes hardware and software elements. Regarding software embodiments, the software may include but is not limited to firmware, resident software, microcode, etc. Regarding hardware embodiments, the hardware can be in one or more application specific integrated circuit chips (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable logic gate arrays (FPGAs), central processing units (CPUs), controllers, microcontrollers, microprocessors, electronic devices, other electronic units designed to perform the functions described herein or combinations thereof.

The above description is for illustrative purposes only and is not intended to be limiting. Any other equivalent modifications or changes that do not depart from the spirit and scope of this disclosure should be included in the scope of the attached patent application.

1: Intelligent automated rehabilitation training device

11: Data collection module

12: Processing module

13: Input module

14: Display module

Claims (14)

A smart automatic rehabilitation training device includes: a data collection module for receiving a first physical state data of a user at a first time point, the first physical state data including a first heart rate variability; and a processing module for receiving the first physical state data and having a decision model, the decision model is obtained by inputting a historical training data into a deep reinforcement learning model to execute a training program, the historical training data including training of a plurality of training objects The training data of each training subject includes the heart rate variability and corresponding feedback information of each training subject executing an initial rehabilitation prescription, wherein the training program calculates the weighted sum of the subjective feeling and the heart rate variability of each training subject after executing the initial rehabilitation prescription each time to obtain the feedback information; wherein the processing module inputs the first body state data into a decision model to generate a first rehabilitation prescription, so that the user executes the first rehabilitation prescription after the first time point. The intelligent automated rehabilitation training device as described in claim 1, wherein the first physical condition data includes one or more of name, age, blood pressure, heart rate, blood sugar, body mass index, and health status. The intelligent automated rehabilitation training device as described in claim 1, wherein the data collection module is used to receive a second physical state data of the user at a second time point after the first time point, the second physical state data includes a second heart rate variability, and the processing module inputs the second physical state data into the decision model to generate a second rehabilitation prescription, so that the user executes the second rehabilitation prescription after the second time point. The intelligent automated rehabilitation training device as described in claim 3, wherein the second physical status data includes one or more of name, age, blood pressure, heart rate, blood sugar, body mass index, and health status. The intelligent automated rehabilitation training device as described in claim 3, wherein the first rehabilitation prescription and the second rehabilitation prescription are continuous rehabilitation prescriptions or discrete rehabilitation prescriptions. The intelligent automated rehabilitation training device as described in claim 1, wherein the weight of the subjective feeling after executing the initial rehabilitation prescription is less than the weight of the heart rate variability. The intelligent automated rehabilitation training device as described in claim 6, wherein the subjective feeling after executing the initial rehabilitation prescription corresponds to the user's flow state. A smart automated rehabilitation training method includes: inputting a historical training data into a deep reinforcement learning model to execute a training program to obtain a decision model, wherein the historical training data includes training data of a plurality of training subjects, and the training data of each training subject includes the heart rate variability and corresponding feedback information of each execution of an initial rehabilitation prescription by the training subject, and the feedback information is calculated by the training program. The weighted sum of the subjective feeling and the heart rate variability of each training subject after executing the initial rehabilitation prescription is obtained; a first body state data of a user at a first time point is received, and the first body state data includes a first heart rate variability; the first body state data is input into a decision model to generate a first rehabilitation prescription; and the user is made to execute the first rehabilitation prescription after the first time point. The intelligent automated rehabilitation training method as described in claim 8, wherein the first physical condition data includes one or more of name, age, blood pressure, heart rate, blood sugar, body mass index, and health status. The intelligent automated rehabilitation training method as described in claim 8 further comprises: receiving a second body state data of the user at a second time point after the first time point, the second body state data comprising a second heart rate variability; inputting the second body state data into the decision model to generate a second rehabilitation prescription; and enabling the user to execute the second rehabilitation prescription after the second time point. The intelligent automated rehabilitation training method as described in claim 10, wherein the second physical condition data includes one or more of name, age, blood pressure, heart rate, blood sugar, body mass index, and health status. The intelligent automated rehabilitation training method as described in claim 10, wherein the first rehabilitation prescription and the second rehabilitation prescription are continuous rehabilitation prescriptions or discrete rehabilitation prescriptions. The intelligent automated rehabilitation training method as described in claim 8, wherein the weight of the subjective feeling after executing the initial rehabilitation prescription is less than the weight of the heart rate variability. The intelligent automated rehabilitation training method as described in claim 13, wherein the subjective feeling after executing the initial rehabilitation prescription corresponds to the user's flow state.

Images