Intelligent vestibular rehabilitation training system based on model recognition technology

Abstract

The invention discloses an intelligent vestibular rehabilitation training system based on a model recognition technology, which is characterized in that a lower cushion body and an upper cushion body are sequentially fixed on the upper surface of an air bag from bottom to top, and an array type pressure sensor is arranged between the lower cushion body and the upper cushion body. An air inlet of the air pump is communicated with air, an air outlet of the air pump is connected with an air inlet pipeline of the electromagnetic valve, an air outlet of the electromagnetic valve is communicated with one end of the inflation pipe, the other end of the inflation pipe is communicated with the air bag, and an air pressure sensor is arranged on the inflation pipe. The system comprises VR glasses, and an attitude sensor is arranged on the VR glasses. The controller is used for controlling the air pump to be started and controlling the electromagnetic valve to be switched on at the same time, so that the air pump pumps extracted air into the air bag through the inflation pipe to inflate the air bag. The controller is further used for conducting intelligent vestibular dysfunction recognition and vestibular rehabilitation training on the target patient standing on the upper pad body.

Description

Technical Field

[0001] The invention relates to the technical field of vestibular rehabilitation training, and in particular to an intelligent vestibular rehabilitation training system based on model recognition technology. Background Art

[0002] The vestibular system is a key part of the human body in maintaining a sense of balance. It provides spatial orientation information by sensing the movement, position changes, and acceleration of the head, and coordinates the balance and posture of the whole body. When the vestibular system is abnormal, it will directly affect the sense of balance. The patient may experience symptoms such as imbalance, unstable gait, abnormal posture, and incoordination of movement. In severe cases, it will increase the risk of falling. In addition, vestibular system abnormalities are often accompanied by problems such as motion sickness, vertigo, and spatial disorientation, which seriously affect the patient's daily life and social activities. Therefore, early diagnosis and timely rehabilitation treatment are very important, which can help patients with vestibular system abnormalities restore or compensate for vestibular function, improve balance, and reduce the risk of falls and other complications.

[0003] With the increasing trend of social aging and changes in lifestyle, the incidence of vestibular dysfunction has been rising year by year, especially among the elderly over 60 years old. Given my country's huge population base and increasingly serious aging problem, the demand for vestibular rehabilitation has become more urgent. However, traditional rehabilitation methods usually rely on drug therapy and physical therapy, which still have certain limitations in terms of personalization, targeting and rehabilitation efficiency. Summary of the invention

[0004] In view of the problems and shortcomings of the prior art, the present invention provides an intelligent vestibular rehabilitation training system based on model recognition technology.

[0005] The present invention solves the above technical problems through the following technical solutions:

[0006] The present invention provides a novel intelligent vestibular rehabilitation training system based on model recognition technology, which is characterized in that it includes an airbag, an array pressure sensor, a controller, an air pump, a solenoid valve, an air pressure sensor and an inflation tube, the upper surface of the airbag is provided with an array pressure sensor, the upper surface of the array pressure sensor is provided with an upper cushion body, the upper cushion body is fixedly connected to the airbag, the air inlet of the air pump is connected to the air, the air outlet of the air pump is connected to the air inlet pipeline of the solenoid valve, the air outlet of the solenoid valve is connected to and communicated with one end of the inflation tube, the other end of the inflation tube is connected to and communicated with the airbag, the inflation tube is provided with an air pressure sensor, the array pressure sensor, the air pump, the solenoid valve and the air pressure sensor are all electrically connected to the controller, the system also includes VR glasses, the VR glasses are provided with a posture sensor, and the VR glasses provide a virtual reality scene for the target patient, so that the patient can perform vestibular rehabilitation training through vision;

[0007] The controller is used to control the start of the air pump and control the conduction of the solenoid valve, so that the air pump pumps the extracted air into the airbag through the inflation tube to inflate the airbag, receives the air pressure in the airbag monitored in real time by the air pressure sensor, and when the air pressure reaches a preset air pressure value, controls the air pump to pause and controls the solenoid valve to disconnect;

[0008] The controller is also used to continuously collect the pressure matrix signal monitored by the array pressure sensor within a preset time and the posture signal monitored by the posture sensor through a set collection cycle for the target patient standing on the upper cushion body;

[0009] The controller is also used to pre-process the collected pressure matrix signal and posture signal, and perform feature extraction on the pre-processed pressure matrix signal and posture signal to extract pressure features and posture features;

[0010] The controller is also used to input the current air pressure value, pressure characteristics and posture characteristics into a target vestibular dysfunction model constructed based on a deep learning model for training and identification, so as to identify the vestibular dysfunction identification results of the target patient. The vestibular dysfunction identification results include no vestibular dysfunction, mild vestibular dysfunction, moderate vestibular dysfunction and severe vestibular dysfunction. When it is identified that the target patient has vestibular dysfunction, vestibular rehabilitation training is continued for the target patient. When it is identified that the target patient does not have vestibular dysfunction, vestibular rehabilitation training is no longer required for the target patient.

[0011] The positive progressive effect of the present invention is that the present invention simulates different standing surface hardness scenarios by making the patient stand on an inflated air bag, so that the patient can gradually adapt to vestibular rehabilitation training with different hardnesses, and at the same time combines VR glasses for visual rehabilitation training, thereby continuously training to improve the adaptability of the patient's brain, thereby promoting the recovery of the patient's vestibular system function. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] Figure 1-2 It is a structural schematic diagram of an intelligent vestibular rehabilitation training system based on model recognition technology in a preferred embodiment. DETAILED DESCRIPTION

[0013] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0014] Example 1

[0015] like Figure 1 and Figure 2 As shown, this embodiment provides an intelligent vestibular rehabilitation training system based on model recognition technology, which includes an airbag 1, a lower cushion body 2, an upper cushion body 3, an array pressure sensor 4, a base plate 5, a control box 6, an air pump 7, a solenoid valve 8, an air pressure sensor 9, an inflation tube 10, a controller 11, a sensor cable 12 and a power supply 13.

[0016] The upper surface of the airbag 1 is fixed with a lower cushion body 2 and an upper cushion body 3 from bottom to top, and an array pressure sensor 4 is arranged between the lower cushion body 2 and the upper cushion body 3. Among them, the upper surface of the airbag 1 is fixedly connected to the lower cushion body 2, and the lower cushion body 2 is provided with an array pressure sensor 4, and the upper surface of the array pressure sensor 4 is provided with an upper cushion body 3, and the lower cushion body 2 and the upper cushion body 3 are fixedly connected. The lower surface of the airbag 1 is a hard bottom surface, and the system also includes a bottom plate 5, and the hard bottom surface of the airbag 1 is fixed to the upper surface of the bottom plate 5. The airbag 1 is a highly elastic airbag 1, and different standing surface hardness scenes are simulated by inflating and deflating the airbag 1, so that the target patient can perform vestibular rehabilitation training through balance perception.

[0017] The system also includes a control box 6 that is separated from the airbag 1. The controller 11, power supply 13, air pump 7, solenoid valve 8 and air pressure sensor 9 are all fixedly placed in the control box 6. The air inlet of the air pump 7 is connected to the air, and the air outlet of the air pump 7 is connected to the air inlet pipeline of the solenoid valve 8. The air outlet of the solenoid valve 8 is connected to one end of the inflation tube 10. The other end of the inflation tube 10 is connected to the airbag 1 after passing through the control box 6. The inflation tube 10 is provided with an air pressure sensor 9. The array pressure sensor 4, air pump 7, solenoid valve 8 and air pressure sensor 9 are all electrically connected to the controller 11. Among them, the array pressure sensor 4 is connected to the controller 11 through the sensor cable 12 passing through the control box 6. The power supply 13 is electrically connected to the controller 11, and the power supply 13 supplies power to various electronic components.

[0018] The system also includes VR glasses, which are provided with posture sensors. The VR glasses provide virtual reality scenes for the target patients, allowing them to perform vestibular rehabilitation training through vision.

[0019] The controller 11 is used to identify and diagnose vestibular dysfunction for the target patient. First, the controller 11 controls the air pump 7 to start, and at the same time controls the solenoid valve 8 to be turned on, so that the air pump 7 pumps the extracted air into the airbag 1 through the inflation tube 10 to inflate the airbag 1, and receives the air pressure in the airbag 1 monitored in real time by the air pressure sensor 9. When the air pressure reaches a preset pressure value, the air pump 7 is controlled to pause, and the solenoid valve 8 is controlled to be disconnected.

[0020] The target patient is then made to stand on the upper cushion body 3. For the target patient standing on the upper cushion body 3, the controller 11 is also used to continuously collect the pressure matrix signal monitored by the array pressure sensor 4 within a preset time and the posture signal monitored by the posture sensor through a set collection cycle, and perform noise reduction filtering pre-processing on the collected pressure matrix signal and posture signal.

[0021] The controller 11 is also used to determine whether the amplitude of the preprocessed posture signal within any set acquisition cycle is within the preset amplitude range corresponding to the current air pressure value. If it is yes, the posture signal and pressure matrix signal preprocessed within the set acquisition cycle are retained; if it is no, the posture signal and pressure matrix signal preprocessed within the set acquisition cycle are eliminated, and feature extraction is performed on the posture signal and pressure matrix signal preprocessed within each set acquisition cycle that are retained to extract posture features and pressure features.

[0022] In this embodiment, for the amplitude of the posture signal preprocessed within a set acquisition period, when the amplitude is within the preset amplitude range corresponding to the current air pressure value, the posture signal and pressure matrix signal preprocessed within the set acquisition period are retained; when the amplitude is not within the preset amplitude range corresponding to the current air pressure value, the posture signal and pressure matrix signal preprocessed within the set acquisition period are directly eliminated. This data processing method helps to provide accurate and optimized posture signals and pressure matrix signals, thereby enabling the extraction of accurate and optimized posture features and pressure features, thereby improving the vestibular dysfunction identification results.

[0023] A further optimized data processing method is: the controller 11 is also used to determine whether the amplitude of the preprocessed posture signal within the set acquisition period is within the preset amplitude range corresponding to the current air pressure value. If it is, the posture signal and pressure matrix signal preprocessed within the set acquisition period are retained as the retained signal; if it is not, the posture signal and pressure matrix signal preprocessed within the set acquisition period are used as the signal to be eliminated, and the cycle retention number of the retained set acquisition period is counted. When the cycle retention number reaches the preset cycle number, all the signals to be eliminated are eliminated; when the cycle retention number does not reach the preset cycle number, it is determined that the previous and next signals of the signal to be eliminated are both the first revised retention number of the retained signal, and when the first revised retention number + cycle When the period retention number reaches the preset number of cycles, for each signal to be eliminated whose front and back are both retained signals, the amplitude of the signal to be eliminated is replaced based on the average amplitude of the retained signals before and after the signal to be eliminated, and the replaced signal constitutes the retained signal, and the remaining signals to be eliminated are eliminated; when the first modified retention number + the period retention number has not reached the preset number of cycles, only one signal to be eliminated is retained from the continuous multiple signals to be eliminated, and for each only retained signal to be eliminated whose front and back are both retained signals, the amplitude of the only retained signal to be eliminated is replaced based on the average amplitude of the retained signals before and after the continuous signals to be eliminated, and the replaced signal constitutes the retained signal, until the period retention number reaches the preset number of cycles, and the remaining signals to be eliminated are eliminated. Feature extraction is performed on the pre-processed posture signals and pressure matrix signals within each set acquisition cycle retained to extract posture features and pressure features.

[0024] The controller 11 is also used to input the current air pressure value, pressure characteristics and posture characteristics into the target vestibular dysfunction model constructed based on the deep learning model for training and identification, so as to identify the vestibular dysfunction identification results of the target patient. The vestibular dysfunction identification results include no vestibular dysfunction, mild vestibular dysfunction, moderate vestibular dysfunction and severe vestibular dysfunction. When the target patient is identified as having vestibular dysfunction, vestibular rehabilitation training is continued for the target patient. When it is identified that the target patient does not have vestibular dysfunction, vestibular rehabilitation training is no longer required for the target patient.

[0025] The controller 11 is also used to perform vestibular rehabilitation training for patients with mild, moderate and severe vestibular dysfunction.

[0026] Specifically, when the controller 11 identifies that the target patient has mild vestibular dysfunction, the controller 11 continues to perform vestibular rehabilitation training combining balance perception and vision on the target patient using the current air pressure.

[0027] When the controller 11 identifies that the target patient has moderate vestibular dysfunction, the target patient is subjected to vestibular rehabilitation training with the increased air pressure. Further, the controller 11 controls the air pump 7 to start again, and controls the electromagnetic valve 8 to be turned on, so that the air pump 7 pumps the extracted air into the airbag 1 through the inflation tube 10 to inflate the airbag 1. During the inflation process, the pressure matrix signal monitored by the array pressure sensor 4 is continuously received, and the maximum pressure value is screened out therefrom. At the same time, the posture signal monitored by the posture sensor is continuously received to determine whether the change value of the maximum pressure value at the two previous and next acquisition time points is greater than the first preset pressure change value, and whether the change value of the posture signal at the two previous and next acquisition time points is greater than the first posture change value. If it is not, the air pump 7 is controlled to be paused, and the electromagnetic valve 8 is controlled to be disconnected to stop inflation when both are yes. The target patient is subjected to vestibular rehabilitation training combining balance perception and vision with the current air pressure.

[0028] When the controller 11 identifies that the target patient has severe vestibular dysfunction, the target patient is subjected to vestibular rehabilitation training with the increased air pressure. Further, the controller 11 controls the air pump 7 to start again, and controls the solenoid valve 8 to be turned on, so that the air pump 7 pumps the extracted air into the airbag 1 through the inflation tube 10 to inflate the airbag 1. During the inflation process, the pressure matrix signal monitored by the array pressure sensor 4 is continuously received, and the maximum pressure value is screened out therefrom. At the same time, the posture signal monitored by the posture sensor is continuously received to determine whether the change value of the maximum pressure value at the two previous and next acquisition time points is greater than the second preset pressure change value, and whether the change value of the posture signal at the two previous and next acquisition time points is greater than the second posture change value. If it is not, the air pump 7 is controlled to be paused, and the solenoid valve 8 is controlled to be disconnected to stop inflation when both are yes. The target patient is subjected to vestibular rehabilitation training combining balance perception and vision with the current air pressure.

[0029] Example 2

[0030] On the basis of Example 1, an intelligent vestibular rehabilitation training system based on model recognition technology in this embodiment further has a control panel embedded in the control box 6 for inputting the identity information of the target patient, which is used to identify the target patient based on the input identity information, control the air pump 7 to start, and control the solenoid valve 8 to be turned on, so that the air pump 7 pumps the extracted air into the airbag 1 through the inflation tube 10 to inflate the airbag 1, receives the air pressure in the airbag 1 monitored in real time by the air pressure sensor 9, and determines whether the identified target patient has the input preset air pressure value. If yes, it determines whether the current air pressure reaches the preset air pressure value corresponding to the identified target patient. If yes, it controls the air pump 7 to pause and controls the solenoid valve 8 to disconnect. If no, it determines whether the current air pressure reaches the system default preset air pressure value. If yes, it controls the air pump 7 to pause and controls the solenoid valve 8 to disconnect, and stops inflation. The target patient is identified and diagnosed with vestibular dysfunction at the current air pressure.

[0031] The present invention controls the inflation of the airbag 1 through the controller 11, so that the target patient stands on the inflated airbag 1, collects the pressure matrix signal of the array pressure sensor 4 and the posture signal of the posture sensor, extracts the pressure characteristics and posture characteristics according to the collected signals, and inputs them into the target vestibular dysfunction model for intelligent identification and diagnosis. According to the diagnosis results, vestibular rehabilitation training combining balance perception and vision is performed for patients with vestibular dysfunction of different degrees. The process is to adjust the inflation amount in the airbag 1 through the control of the controller 11, simulate different standing surface hardness scenes, make the patient gradually adapt to the vestibular rehabilitation training of different hardness, and combine VR glasses for visual rehabilitation training at the same time, and then continuously train to improve the adaptability of the patient's brain, thereby promoting the recovery of the patient's vestibular system function.

[0032] Although the specific embodiments of the present invention are described above, those skilled in the art should understand that these are only examples, and the protection scope of the present invention is defined by the appended claims. Those skilled in the art may make various changes or modifications to these embodiments without departing from the principles and essence of the present invention, but these changes and modifications all fall within the protection scope of the present invention.

Claims

1. An intelligent vestibular rehabilitation training system based on model recognition technology, characterized in that: It includes an airbag, an array pressure sensor, a controller, an air pump, a solenoid valve, an air pressure sensor and an inflation tube. The upper surface of the airbag is provided with an array pressure sensor, the upper surface of the array pressure sensor is provided with an upper cushion body, the upper cushion body is fixedly connected to the airbag, the air inlet of the air pump is connected to the air, the air outlet of the air pump is connected to the air inlet pipeline of the solenoid valve, the air outlet of the solenoid valve is connected to and communicated with one end of the inflation tube, the other end of the inflation tube is connected to and communicated with the airbag, an air pressure sensor is arranged on the inflation tube, the array pressure sensor, the air pump, the solenoid valve and the air pressure sensor are all electrically connected to the controller, the system also includes VR glasses, the VR glasses are provided with a posture sensor, the VR glasses provide a virtual reality scene for the target patient, so that the patient can perform vestibular rehabilitation training through vision; The controller is used to control the start of the air pump and control the conduction of the solenoid valve, so that the air pump pumps the extracted air into the airbag through the inflation tube to inflate the airbag, receives the air pressure in the airbag monitored in real time by the air pressure sensor, and when the air pressure reaches a preset air pressure value, controls the air pump to pause and controls the solenoid valve to disconnect; The controller is also used to continuously collect the pressure matrix signal monitored by the array pressure sensor within a preset time and the posture signal monitored by the posture sensor through a set collection cycle for the target patient standing on the upper cushion body; The controller is also used to pre-process the collected pressure matrix signal and posture signal, and perform feature extraction on the pre-processed pressure matrix signal and posture signal to extract pressure features and posture features; The controller is also used to input the current air pressure value, pressure characteristics and posture characteristics into a target vestibular dysfunction model constructed based on a deep learning model for training and identification, so as to identify the vestibular dysfunction identification results of the target patient. The vestibular dysfunction identification results include no vestibular dysfunction, mild vestibular dysfunction, moderate vestibular dysfunction and severe vestibular dysfunction. When it is identified that the target patient has vestibular dysfunction, vestibular rehabilitation training is continued for the target patient. When it is identified that the target patient does not have vestibular dysfunction, vestibular rehabilitation training is no longer required for the target patient.

2. The intelligent vestibular rehabilitation training system based on model recognition technology according to claim 1, characterized in that: The controller is also used to continue vestibular rehabilitation training for the target patient at the current air pressure when it is identified that the target patient has mild vestibular dysfunction; The controller is also used to perform vestibular rehabilitation training on the target patient with increased air pressure when it is identified that the target patient has moderate vestibular dysfunction; The controller is also used to perform vestibular rehabilitation training on the target patient with increased air pressure when it is identified that the target patient has severe vestibular dysfunction.

3. The intelligent vestibular rehabilitation training system based on model recognition technology as claimed in claim 2, characterized in that: The controller is also used to control the air pump to start again when it is identified that the target patient has moderate vestibular dysfunction, and control the solenoid valve to be turned on, so that the air pump pumps the extracted air into the airbag through the inflation tube to inflate the airbag, and continuously receives the pressure matrix signal monitored by the array pressure sensor during the inflation process, screens out the maximum pressure value, and determines whether the change value of the maximum pressure value at two previous and subsequent acquisition time points is greater than the first preset pressure change value, and controls the continuous inflation when the value is negative, and controls the air pump to pause and controls the solenoid valve to be turned off when the value is positive; The controller is also used to control the air pump to start again when it is identified that the target patient has severe vestibular dysfunction, and to control the solenoid valve to be turned on, so that the air pump pumps the extracted air into the airbag through the inflation tube to inflate the airbag. During the inflation process, the pressure matrix signal monitored by the array pressure sensor is continuously received, the maximum pressure value is screened out, and it is determined whether the change value of the maximum pressure value between the two previous and subsequent collection time points is greater than the second preset pressure change value. If the answer is no, the air pump is controlled to continue to be inflated, and if the answer is yes, the air pump is controlled to be paused and the solenoid valve is controlled to be disconnected.

4. The intelligent vestibular rehabilitation training system based on model recognition technology according to claim 3, characterized in that: The controller is also used to control the air pump to start again when it is identified that the target patient has moderate vestibular dysfunction, and control the solenoid valve to be turned on, so that the air pump pumps the extracted air into the airbag through the inflation tube to inflate the airbag, and continuously receives the pressure matrix signal monitored by the array pressure sensor during the inflation process, screens out the maximum pressure value, and continuously receives the posture signal monitored by the posture sensor, and judges whether the change value of the maximum pressure value at the two previous and next collection time points is greater than the first preset pressure change value, and whether the change value of the posture signal at the two previous and next collection time points is greater than the first posture change value, and controls the air pump to pause and the solenoid valve to disconnect when both are yes, otherwise it controls to continue to inflate; The controller is also used to control the air pump to start again when it is identified that the target patient has severe vestibular dysfunction, and to control the solenoid valve to be turned on, so that the air pump pumps the extracted air into the airbag through the inflation tube to inflate the airbag. During the inflation process, the pressure matrix signal monitored by the array pressure sensor is continuously received, and the maximum pressure value is screened out therefrom. At the same time, the posture signal monitored by the posture sensor is continuously received to determine whether the change value of the maximum pressure value between the two previous and next collection time points is greater than the second preset pressure change value, and whether the change value of the posture signal between the two previous and next collection time points is greater than the second posture change value. When both are yes, the air pump is controlled to be paused and the solenoid valve is controlled to be disconnected. Otherwise, the inflation is controlled to continue.

5. The intelligent vestibular rehabilitation training system based on model recognition technology according to claim 1, characterized in that: The controller is also used to perform noise reduction filtering preprocessing on the collected pressure matrix signal and attitude signal, and to determine whether the amplitude of the preprocessed attitude signal within any set acquisition period is within a preset amplitude range corresponding to the current air pressure value. If the answer is yes, the attitude signal and pressure matrix signal preprocessed within the set acquisition period are retained; if the answer is no, the attitude signal and pressure matrix signal preprocessed within the set acquisition period are eliminated; and feature extraction is performed on the attitude signal and pressure matrix signal preprocessed within each set acquisition period that are retained to extract attitude features and pressure features.

6. The intelligent vestibular rehabilitation training system based on model recognition technology according to claim 5, characterized in that: The controller is also used to determine whether the amplitude of the preprocessed posture signal within any set acquisition cycle is within the preset amplitude range corresponding to the current air pressure value. If it is yes, the preprocessed posture signal and pressure matrix signal within the set acquisition cycle are retained as retained signals. If it is no, the preprocessed posture signal and pressure matrix signal within the set acquisition cycle are used as signals to be eliminated. The cycle retention number of the retained set acquisition cycle is counted. When the cycle retention number reaches the preset cycle number, all the signals to be eliminated are eliminated. When the cycle retention number does not reach the preset cycle number, it is determined that the previous and next signals of the signal to be eliminated are both the first revised retention number of the retained signal. When the first revised retention number + the cycle retention number reaches the preset When the number of cycles is reached, for each signal to be eliminated whose preceding and following are retained signals, the amplitude of the signal to be eliminated is replaced based on the average amplitude of the retained signals before and after the signal to be eliminated, and the retained signal is formed after the replacement, and the remaining signals to be eliminated are eliminated. When the first corrected retention number + the cycle retention number has not reached the preset number of cycles, only one signal to be eliminated is retained from multiple consecutive signals to be eliminated. For each signal to be eliminated whose preceding and following are retained signals, the amplitude of the only signal to be eliminated is replaced based on the average amplitude of the retained signals before and after the consecutive signals to be eliminated, and the retained signal is formed after the replacement, until the cycle retention number reaches the preset number of cycles, and the remaining signals to be eliminated are eliminated.

7. The intelligent vestibular rehabilitation training system based on model recognition technology according to claim 1, characterized in that: The system also includes a power supply, which is electrically connected to the controller. The system also includes a control box that is separated from the airbag. The controller, power supply, air pump, solenoid valve and air pressure sensor are all fixedly arranged in the control box. The array pressure sensor is connected to the controller wire via a sensor cable that passes through the control box. The airbag is connected to the air outlet of the solenoid valve via an inflation tube that passes through the control box.

8. The intelligent vestibular rehabilitation training system based on model recognition technology according to claim 7, characterized in that: The control box is embedded with a control panel for inputting the identity information of the target patient. The controller is used to identify the target patient based on the input identity information, control the start of the air pump, and control the conduction of the solenoid valve, so that the air pump pumps the extracted air into the airbag through the inflation tube to inflate the airbag, receives the air pressure in the airbag monitored in real time by the air pressure sensor, and determines whether the identified target patient has an input preset air pressure value. If so, it determines whether the current air pressure reaches the preset air pressure value corresponding to the identified target patient. If so, it controls the air pump to pause and the solenoid valve to disconnect. If not, it determines whether the current air pressure reaches the system default preset air pressure value. If so, it controls the air pump to pause and the solenoid valve to disconnect.

9. The intelligent vestibular rehabilitation training system based on model recognition technology according to claim 1, characterized in that: The system further comprises a bottom plate, the lower surface of the airbag is a hard bottom surface, and the hard bottom surface is fixed to the upper surface of the bottom plate.

10. The intelligent vestibular rehabilitation training system based on model recognition technology according to claim 1, characterized in that: A lower cushion body and an upper cushion body are fixed to the upper surface of the airbag in sequence from bottom to top, and an array-type pressure sensor is arranged between the lower cushion body and the upper cushion body.

Images