A method and system for controlling respiratory neuromuscular stimulation

Abstract

The application discloses a respiratory neuromuscular stimulation control method and system, and relates to the technical field of medical treatment. The method comprises the following steps: collecting body parameter information of a user; determining stimulation parameters of the user according to the body parameter information and parameters stored in a database; processing an initial muscle contraction signal obtained from a treatment terminal through a signal processing module before electric stimulation is performed on the body of the user, so as to obtain an initial muscle contraction filtered signal; wherein the initial muscle contraction signal is sensed by the treatment terminal attached to the body of the user; and controlling the current of the electric stimulation generated by the treatment terminal arranged on the body of the user in real time according to the stimulation parameters of the user, the initial muscle contraction filtered signal, a real-time muscle contraction filtered signal and real-time breathing conditions, through a self-adaptive current adjustment model which is trained and optimized. The application improves the efficiency of electric stimulation.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of medical treatment, in particular to a respiratory neuromuscular stimulation control method and system. BACKGROUND

[0002] Muscle atrophy is common in people after injury, after surgery, long-term bedridden, or after fitness abandonment. When muscle atrophy reaches a certain degree, many daily activities are difficult for people to complete.

[0003] Neuromuscular stimulation is a method of applying low-frequency pulse current to stimulate motor nerves and muscles to cause muscle contraction. When electrical stimulation acts on the muscle, the motor nerve is activated first. When the motor nerve is activated by electrical stimulation, a nerve impulse is generated, and the nerve impulse transmitted to the muscle can cause muscle contraction.

[0004] When treating a user with an electrical stimulation device, medical personnel usually set the stimulation time of the electrical stimulation device according to the severity of the user's respiratory muscle weakness or respiratory muscle fatigue. That is, in the prior art, the treatment of muscle atrophy with an electrical stimulation device is relatively single, and only the stimulation time is set according to medical experience before electrical stimulation treatment.

[0005] To address the above problems, no effective solutions have been proposed so far. SUMMARY

[0006] The present application provides a respiratory neuromuscular stimulation control method and system to solve the above technical problems.

[0007] The present application provides a respiratory neuromuscular stimulation control method, comprising:

[0008] Collecting body parameter information of a user; determining stimulation parameters of the user according to the body parameter information and parameters stored in a database;

[0009] Before electrical stimulation is performed on the user's body, an initial muscle contraction signal obtained from a treatment terminal is processed by a signal processing module to obtain an initial muscle contraction filtered signal; wherein the initial muscle contraction signal is sensed by the treatment terminal attached to the user's body;

[0010] According to the stimulation parameters of the user, the initial muscle contraction filtered signal, the real-time muscle contraction filtered signal, and the real-time breathing condition, the current of the electrical stimulation generated by the treatment terminal attached to the user's body is controlled in real time by a self-adaptive current adjustment model trained and optimized;

[0011] The real-time breathing condition includes the user's inhalation time and respiratory rate; the electrical stimulation current acts on the user's nerves or muscles, the muscles include single muscle groups or multiple muscle groups, and the nerves include phrenic nerves.

[0012] Furthermore, the adaptive current adjustment model is constructed and trained and optimized based on the following methods:

[0013] Acquiring historical user stimulation data, historical muscle contraction signals, historical respiration data, and real-time electrical stimulation current data; wherein the historical muscle contraction signals include initial muscle contraction signals and real-time stimulated muscle contraction signals; and the historical respiration data include initial respiration data and real-time stimulated respiration data;

[0014] determining a set of stimulus factors of the historical user stimulus data;

[0015] Determining a weight vector and a membership vector for each stimulus factor in the stimulus factor set;

[0016] Performing fuzzy transformation on the weight vector and the membership vector of each stimulus factor to obtain a first fuzzy evaluation matrix;

[0017] Performing fuzzy operations on the first fuzzy evaluation matrix and the weight vectors of the respective stimulus factors and performing normalization to obtain a second fuzzy evaluation matrix;

[0018] A deep neural network is constructed based on the second fuzzy evaluation matrix, including an input layer, a base network, and an adaptive weight layer; the base network includes six convolutional layers, three fully connected layers, and a pooling layer; wherein the convolutional layers are connected to the pooling layer, and the pooling layer is used to aggregate values ​​of adjacent regions; the three fully connected layers include a first fully connected layer, a second fully connected layer, and a third fully connected layer; the adaptive weight layer is used to complete adaptive adjustment of current parameters through back propagation;

[0019] Performing feature extraction on the initial respiratory data and the real-time stimulated respiratory data to obtain an initial respiratory feature vector and a real-time stimulated respiratory feature vector;

[0020] Calculating the distance between the initial respiratory feature vector and the real-time stimulated respiratory feature vector;

[0021] The constructed deep neural network is optimized and trained according to the distance between the initial respiratory feature vector and the real-time stimulated respiratory feature vector, the initial muscle contraction signal, the real-time stimulated muscle contraction signal, the initial respiratory feature vector, the real-time stimulated respiratory feature vector and the real-time electrical stimulation current data until the loss function converges.

[0022] Furthermore, the real-time electrical stimulation current data includes current stimulation intensity, pulse width, pulse frequency, and stimulation duration; and the constructed deep neural network is optimized and trained according to the distance between the initial respiratory feature vector and the real-time stimulated respiratory feature vector, the initial muscle contraction signal, the real-time stimulated muscle contraction signal, the initial respiratory feature vector, the real-time stimulated respiratory feature vector, and the real-time electrical stimulation current data until the loss function converges, including:

[0023] A time series sample vector is constructed according to the distance between the initial respiratory feature vector and the real-time stimulated respiratory feature vector, the initial respiratory feature vector, and the real-time stimulated respiratory feature vector:

[0024] Inputting the time series sample vector into the input layer of the deep neural network to obtain a respiratory time series output dataset output by the first fully connected layer;

[0025] Performing non-negative matrix decomposition on the initial muscle contraction signal and the real-time stimulated muscle contraction signal based on the time series sample vector to obtain a signal non-negative matrix decomposition set;

[0026] Performing window processing on the signal non-negative matrix decomposition set to obtain time domain samples; performing short-time Fourier transform on the time domain samples to obtain frequency domain samples;

[0027] quantifying a first correlation between the frequency domain samples and the real-time electrical stimulation current data by means of information entropy and mutual information, and determining a first adaptive loss function according to a first adaptive adjustment function preset in the adaptive weight layer;

[0028] quantifying a second correlation between the frequency domain samples and the respiratory time series output data set by means of information entropy and mutual information, and determining a second adaptive loss function according to a second adaptive adjustment function preset in the adaptive weight layer;

[0029] The loss function is constructed according to the first adaptive loss function and the second adaptive loss function, and the deep neural network is optimized and trained until the loss function converges.

[0030] Furthermore, the physical parameter information includes age, joint mobility, muscle atrophy, spasticity parameters and lung function parameters; wherein the lung function parameters include vital capacity, forced vital capacity and forced expiratory volume in one second.

[0031] Furthermore, the treatment terminal includes multiple treatment electrodes; the multiple treatment electrodes include electrodes attached to the posterior edges of the left and right sternocleidomastoid muscles, the left and right subclavian pectoralis major muscles, the oblique abdominal muscles, the rectus abdominis muscles, the quadriceps femoris muscles of both lower limbs, the tibialis anterior muscles, the common peroneal nerve, the tibialis anterior muscles, the peroneus longus and brevis muscles, the quadriceps femoris muscles, the hip abductor muscles, and the wrist and finger extensor muscles.

[0032] The present application provides a respiratory nerve muscle stimulation control system, comprising a user stimulation parameter determination unit, an initial muscle contraction filter signal determination unit, and an electrical stimulation current real-time control unit; wherein,

[0033] The user stimulation parameter determination unit is used to collect the user's body parameter information; determine the user's stimulation parameters based on the body parameter information and the parameters stored in the database;

[0034] The initial muscle contraction filter signal determination unit is configured to process the initial muscle contraction signal obtained from the treatment terminal through a signal processing module before electrically stimulating the user's body to obtain an initial muscle contraction filter signal; wherein the initial muscle contraction signal is sensed by the treatment terminal attached to the user's body;

[0035] The electrical stimulation current real-time control unit is used to control the electrical stimulation current generated by the treatment terminal arranged on the user's body in real time according to the user's stimulation parameters, the initial muscle contraction filter signal, the real-time muscle contraction filter signal and the real-time breathing condition through an adaptive current adjustment model that has been trained and optimized; wherein the real-time breathing condition includes the user's inhalation time and breathing frequency; the electrical stimulation current acts on the user's nerves or muscles, the muscles include single muscle groups or multiple muscle groups, and the nerves include phrenic nerves.

[0036] Based on the embodiments provided by the present application, a user's body parameter information is collected; stimulation parameters for the user are determined based on the body parameter information and parameters stored in a database; before electrically stimulating the user's body, an initial muscle contraction signal obtained from a treatment terminal is processed by a signal processing module to obtain an initial muscle contraction filter signal; wherein the initial muscle contraction signal is sensed by a treatment terminal attached to the user's body; based on the user's stimulation parameters, the initial muscle contraction filter signal, the real-time muscle contraction filter signal, and the real-time breathing condition, an adaptive current adjustment model completed through training and optimization is used to control the current of the electrical stimulation generated by the treatment terminal disposed on the user's body in real time; wherein the real-time breathing condition includes the user's inhalation time and respiratory rate; the electrical stimulation current acts on the user's nerves or muscles, the muscles including single muscle groups or multiple muscle groups, and the nerves including phrenic nerves. This achieves timely adjustment of the current data of the electrical stimulation device based on the user's own physical condition and response to the electrical stimulation, rather than simply adjusting the action time of the electrical stimulation device before treatment begins based on the experience of medical staff, thereby improving the efficiency and accuracy of the electrical stimulation.

[0037] Specifically, it has the following beneficial effects: Collecting the user's body parameter information: It helps in personalized treatment: By collecting detailed body data of individuals, a customized treatment plan can be provided for each user; Improving treatment effect: Accurate body parameter information helps to more accurately assess treatment needs and adjust treatment plans; Enhancing safety: Understanding the user's physical condition can prevent potential treatment risks and ensure treatment safety; Determining the user's stimulation parameters based on body parameter information and parameters stored in the database: Optimizing treatment decisions: Combining historical data and real-time parameters, the stimulation parameters can be determined more scientifically to improve the effectiveness of treatment; Dynamic adjustment: The stimulation parameters can be dynamically adjusted according to changes in the user's condition to adapt to changes in treatment effect and user comfort; Processing the initial muscle contraction signal through the signal processing module: Improving signal quality: Filtering can remove noise and interference and extract more accurate muscle contraction signals; Precise control: Accurate signals are the basis for achieving precise control of electrical stimulation current, which helps to improve treatment effects; Using initial and real-time muscle contraction filtering signals Real-time control of the signal: Real-time feedback: Real-time monitoring of muscle contraction signals can adjust the electrical stimulation parameters in time to achieve closed-loop control; Adaptive adjustment: According to the changes in real-time signals, the adaptive current adjustment model can dynamically adjust the electrical stimulation to match the user's immediate needs; Consider real-time breathing conditions: Synchronous treatment: Synchronizing electrical stimulation with the user's breathing cycle can improve the comfort and effect of treatment; Personalized adjustment: Considering breathing frequency and inhalation time, more personalized plans can be customized for different users; The current of electrical stimulation acts on the user's nerves or muscles: Targeted treatment: Electrical stimulation can directly act on the target nerves or muscles to provide targeted treatment; Enhance muscle function: By stimulating specific muscle groups, muscle strength and function can be enhanced, and muscle atrophy or dysfunction can be improved; Real-time control of the electrical stimulation current generated by the treatment terminal: Precise dosage: Real-time control of current can ensure that the user receives treatment of appropriate intensity and avoid excessive or insufficient stimulation; Improve efficacy: Precise current control helps to improve treatment effects, reduce side effects, and enhance the user's overall treatment experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] The drawings described herein are used to provide a further understanding of the embodiments of the present invention and constitute a part of this application. The illustrative embodiments of this application and their descriptions are used to explain this application and do not constitute an improper limitation of this application. In the drawings:

[0039] Figure 1 is a flow chart of an optional respiratory neuromuscular stimulation control method according to an embodiment of the present application;

[0040] Figure 2 This is a structural diagram of an optional respiratory neuromuscular stimulation control system according to an embodiment of the present application.

[0041] The purpose, features and advantages of the present invention will be further described with reference to the accompanying drawings and in conjunction with the embodiments. DETAILED DESCRIPTION

[0042] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application.

[0043] Alternatively, as Figure 1 As shown, the present application provides a respiratory neuromuscular stimulation control method, comprising:

[0044] S101, collecting user's body parameter information; determining the user's stimulation parameters based on the body parameter information and parameters stored in a database;

[0045] S102, before electrically stimulating the user's body, processing the initial muscle contraction signal obtained from the treatment terminal by a signal processing module to obtain an initial muscle contraction filtered signal; wherein the initial muscle contraction signal is sensed by the treatment terminal attached to the user's body;

[0046] S103, based on the user's stimulation parameters, initial muscle contraction filter signal, real-time muscle contraction filter signal and real-time breathing conditions, the adaptive current adjustment model completed through training and optimization is used to control the electric stimulation current generated by the treatment terminal set on the user's body in real time; wherein the real-time breathing conditions include the user's inhalation time and breathing frequency; the electric stimulation current acts on the user's nerves or muscles, the muscles include single muscle groups or multiple muscle groups, and the nerves include phrenic nerves.

[0047] Based on the embodiments provided by the present application, the user's body parameter information is collected; the user's stimulation parameters are determined based on the body parameter information and the parameters stored in the database; before electrically stimulating the user's body, the initial muscle contraction signal obtained from the treatment terminal is processed by a signal processing module to obtain an initial muscle contraction filter signal; wherein the initial muscle contraction signal is sensed by the treatment terminal attached to the user's body; based on the user's stimulation parameters, the initial muscle contraction filter signal, the real-time muscle contraction filter signal, and the real-time breathing condition, the adaptive current adjustment model completed by training and optimization controls the current of the electrical stimulation generated by the treatment terminal set on the user's body in real time; wherein the real-time breathing condition includes the user's inhalation time and respiratory rate; the electrical stimulation current acts on the user's nerves or muscles, the muscles including single muscle groups or multiple muscle groups, and the nerves including phrenic nerves. Thus, the electrical stimulation device can be adjusted in time according to the user's own physical condition and response to the electrical stimulation, thereby improving the efficiency and accuracy of the electrical stimulation.

[0048] In particular, the following benefits are achieved: collecting user's body parameter information: helping individualized treatment: by collecting detailed body data of individuals, customized treatment plans can be provided for each user; improving treatment effectiveness: accurate body parameter information helps more accurately assess treatment needs and adjust treatment programs; enhancing safety: understanding the user's physical condition can prevent potential treatment risks and ensure treatment safety; determining user's stimulation parameters based on body parameter information and parameters stored in the database: optimizing treatment decisions: combining historical data and real-time parameters, stimulation parameters can be determined more scientifically to improve the effectiveness of treatment; dynamically adjusting: stimulation parameters can be dynamically adjusted according to changes in user conditions to adapt to changes in treatment effectiveness and user comfort; processing initial muscle contraction signals through the signal processing module: improving signal quality: filtering can remove noise and interference to extract more accurate muscle contraction signals; precise control: accurate signals are the basis for precise control of electrical stimulation current, which helps improve treatment effectiveness; using initial and real-time muscle contraction filtered signals for real-time control: real-time feedback: real-time monitoring of muscle contraction signals can allow timely adjustment of electrical stimulation parameters to achieve closed-loop control; adaptive adjustment: based on changes in real-time signals, the adaptive current adjustment model can dynamically adjust electrical stimulation to match the user's immediate needs; considering real-time breathing conditions: synchronized treatment: synchronizing electrical stimulation with the user's breathing cycle can improve the comfort and effectiveness of treatment; personalized adjustment: considering breathing frequency and inhalation time, more personalized solutions can be customized for different users; the current of the electrical stimulation acts on the user's nerves or muscles: targeted treatment: electrical stimulation can directly act on target nerves or muscles, providing targeted treatment; enhance muscle function: by stimulating specific muscle groups, muscle strength and function can be enhanced, improving muscle atrophy or dysfunction; real-time control of the current of the electrical stimulation generated by the treatment terminal: precise dosage: real-time control of current can ensure that users receive appropriate intensity treatment, avoiding excessive or insufficient stimulation; improve efficacy: precise current control helps improve treatment effectiveness, reduce side effects, and enhance the overall treatment experience of users.

[0049] Further, the adaptive current adjustment model is constructed and trained based on the following methods:

[0050] Obtain historical user stimulation data, historical muscle contraction signals, historical breathing data, and real-time electrical stimulation current data; wherein the historical muscle contraction signals include initial muscle contraction signals and real-time stimulation muscle contraction signals; the historical breathing data includes initial breathing data and real-time stimulation breathing data;

[0051] Determine a set of stimulation factors of the historical user stimulation data;

[0052] Determine the weight vector and membership degree vector of each stimulation factor in the set of stimulation factors;

[0053] The weight vector and the membership degree vector of each stimulating factor are subjected to fuzzy transformation to obtain a first fuzzy evaluation matrix;

[0054] The first fuzzy evaluation matrix is subjected to fuzzy operation with the weight vector of each stimulating factor and normalized to obtain a second fuzzy evaluation matrix;

[0055] A deep neural network is constructed according to the second fuzzy evaluation matrix, including an input layer, a basic network, and an adaptive weight layer; the basic network includes six convolutional layers, three fully connected layers, and a pooling layer; the convolutional layer is connected with the pooling layer, and the pooling layer is used for aggregating values of adjacent regions; the three fully connected layers include a first fully connected layer, a second fully connected layer, and a third fully connected layer; the adaptive weight layer is used for completing adaptive adjustment of the current parameter through back propagation;

[0056] Feature extraction is performed on the initial respiratory data and the real-time stimulating respiratory data to obtain an initial respiratory feature vector and a real-time stimulating respiratory feature vector;

[0057] The distance between the initial respiratory feature vector and the real-time stimulating respiratory feature vector is calculated;

[0058] The deep neural network constructed is subjected to optimization training according to the distance between the initial respiratory feature vector and the real-time stimulating respiratory feature vector, the initial muscle contraction signal, the real-time stimulating muscle contraction signal, the initial respiratory feature vector, the real-time stimulating respiratory feature vector, and the real-time electric stimulation current data, until the loss function converges.

[0059] Based on the embodiments provided in this application, the following beneficial effects are achieved: obtaining historical and real-time data: data-driven decision-making: by analyzing historical and real-time data, the adaptive current adjustment model can be constructed and optimized more accurately; enhancing the generalization ability of the model: combining historical and real-time data, the model can adapt to changes in different users and at different time points; determining the set of stimulus factors: comprehensiveness: ensuring that the model considers all relevant stimulus factors and improves the accuracy of predicting treatment effects; customized treatment: identifying key stimulus factors and providing customized treatment parameters for each user; determining weight vectors and membership vectors: quantifying influence: quantifying the influence of different stimulus factors through weights and membership, making model adjustment more precise. Increase precision; Prioritization: Clarify which factors are more important, so as to give them higher priority when optimizing the model; Perform fuzzy transformation: Handle uncertainty: Fuzzy logic can handle uncertainty and ambiguity in data, and improve the robustness of the model; Flexibility: Allow the model to make effective decisions when faced with inaccurate or partial information; Adaptability: The adaptive weight layer enables the model to adjust in real time based on feedback, improving the accuracy of current control; Feature extraction and distance calculation: Key information identification: Extracting key features from breathing data helps the model better understand the user's breathing pattern; Dynamic tracking: Calculating the distance between feature vectors enables the model to track dynamic changes in user status.

[0060] Furthermore, the real-time electrical stimulation current data includes the current stimulation intensity, pulse width, pulse frequency, and stimulation duration; and the constructed deep neural network is optimized and trained according to the distance between the initial respiratory feature vector and the real-time stimulated respiratory feature vector, the initial muscle contraction signal, the real-time stimulated muscle contraction signal, the initial respiratory feature vector, the real-time stimulated respiratory feature vector, and the real-time electrical stimulation current data until the loss function converges, including:

[0061] Construct a time series sample vector based on the distance between the initial respiratory feature vector and the real-time stimulated respiratory feature vector, the initial respiratory feature vector, and the real-time stimulated respiratory feature vector:

[0062] Input the time series sample vector into the input layer of the deep neural network to obtain the respiratory time series output dataset output by the first fully connected layer;

[0063] Based on the time series sample vector, the initial muscle contraction signal and the real-time stimulated muscle contraction signal are subjected to non-negative matrix decomposition to obtain a signal non-negative matrix decomposition set;

[0064] Perform window processing on the signal non-negative matrix decomposition set to obtain time domain samples; perform short-time Fourier transform on the time domain samples to obtain frequency domain samples;

[0065] quantifying a first correlation between the frequency domain samples and the real-time electrical stimulation current data by means of information entropy and mutual information, and determining a first adaptive loss function according to a first adaptive adjustment function preset in the adaptive weight layer;

[0066]

[0067] in, is the first adaptive loss function; U is the number of adaptive weight layers; u is the number of adaptive weight layers; Strat1 u is the first strategy of the u-th adaptive weight layer; Ajust1() is the first adaptive adjustment function; α1 is the regularization parameter; e is a natural constant; relev1 is the first correlation quantity;

[0068] quantifying a second correlation between the frequency domain samples and the respiratory time series output data set by means of information entropy and mutual information, and determining a second adaptive loss function according to a second adaptive adjustment function preset in the adaptive weight layer;

[0069] A loss function is constructed according to the first adaptive loss function and the second adaptive loss function, and the deep neural network is optimized and trained until the loss function converges.

[0070] Based on the embodiments provided in this application, the following beneficial effects are achieved: constructing a time series sample vector: by combining the initial and real-time breathing and muscle contraction signals, a comprehensive time series sample vector can be created, which helps the deep neural network capture the dynamic changes in the user's physiological state; inputting the deep neural network: inputting the time series sample vector into the network enables the network to receive rich physiological state information starting from the input layer, laying the foundation for subsequent feature learning and state prediction; output of the first fully connected layer: the respiratory time series output data set output by the first fully connected layer can provide the network with preliminary breathing pattern features, which is helpful for subsequent analysis and processing; non-negative matrix decomposition: performing non-negative matrix decomposition on the initial and real-time stimulation muscle contraction signals can extract the key features of muscle activity, which helps to understand the intrinsic structure of muscle activity; window processing and short-time Fourier transform: through window processing and short-time Fourier transform, time domain samples can be converted into frequency domain samples , which helps to analyze the frequency characteristics of muscle activity and provide a basis for adjusting the frequency of electrical stimulation; quantification of information entropy and mutual information: using information entropy and mutual information to quantify the correlation between frequency domain samples and electrical stimulation current data, the relationship between different physiological signals and electrical stimulation effects can be evaluated, providing guidance for optimizing electrical stimulation parameters; determination of adaptive loss function: the first adaptive loss function and the second adaptive loss function determined according to the first correlation quantity enable the model to automatically adjust the sensitivity to outliers during training, thereby improving the robustness of the model; optimization training of deep neural networks: by combining the first and second adaptive loss functions for network optimization training, it can be ensured that the network continuously adjusts itself during training to adapt to changes in the user's physiological state until the loss function converges; real-time current control: real-time control of the electrical stimulation current, so that the electrical stimulation can be dynamically adjusted according to the user's real-time physiological state, thereby improving the treatment effect and user comfort.

[0071] Together, these steps form a closed-loop adaptive current adjustment system that intelligently adjusts to the user's real-time physiological data to achieve personalized and optimized electrical stimulation therapy. This approach can significantly improve the effectiveness and safety of electrical stimulation therapy.

[0072] Further,

[0073] in, is the second adaptive loss function; U is the number of adaptive weight layers; u is the number of adaptive weight layers; Strat2 u is the second strategy of the u-th adaptive weight layer; Ajust2() is the second adaptive adjustment function; α2 is the regularization parameter; e is a natural constant; relev2 is the second related quantity.

[0074] Furthermore, the distance between sample points is calculated based on the following formula:

[0075]

[0076] Where distan(X,Y) represents the distance between the initial respiratory feature vector X and the real-time stimulated respiratory feature vector Y; m is the dimension of the initial respiratory feature vector; n is the dimension of the real-time stimulated respiratory feature vector; x i represents the coordinate value of the initial respiratory feature vector X in the dimension numbered i; j is the coordinate value of the real-time stimulated respiratory feature vector Y in the dimension numbered j; t is a controllable variable parameter.

[0077] Furthermore, the body parameter information includes age, joint mobility, muscle atrophy, spasticity parameters and lung function parameters; wherein, the lung function parameters include vital capacity, forced vital capacity and forced expiratory volume in one second.

[0078] Furthermore, the treatment terminal includes multiple treatment electrodes; the multiple treatment electrodes include electrodes attached to the posterior edges of the left and right sternocleidomastoid muscles, the left and right subclavian pectoralis major muscles, the oblique abdominal muscles, the rectus abdominis muscles, the quadriceps femoris muscles of both lower limbs, the tibialis anterior muscles, the common peroneal nerve, the tibialis anterior muscles, the peroneus longus and brevis muscles, the quadriceps femoris muscles, the hip abductor muscles, and the wrist and finger extensor muscles.

[0079] Alternatively, as Figure 2 As shown, the present application provides a respiratory nerve muscle stimulation control system, including a user stimulation parameter determination unit 201, an initial muscle contraction filter signal determination unit 202 and an electrical stimulation current real-time control unit 203; wherein,

[0080] The user stimulation parameter determination unit 201 is used to collect the user's body parameter information; determine the user's stimulation parameters based on the body parameter information and the parameters stored in the database;

[0081] The initial muscle contraction filter signal determining unit 202 is configured to process the initial muscle contraction signal obtained from the treatment terminal using a signal processing module before electrically stimulating the user's body to obtain an initial muscle contraction filter signal; wherein the initial muscle contraction signal is sensed by the treatment terminal attached to the user's body;

[0082] The electrical stimulation current real-time control unit 203 is used to control the electrical stimulation current generated by the treatment terminal set on the user's body in real time according to the user's stimulation parameters, initial muscle contraction filter signal, real-time muscle contraction filter signal and real-time breathing conditions through an adaptive current adjustment model that has been trained and optimized; wherein the real-time breathing conditions include the user's inhalation time and breathing frequency; the electrical stimulation current acts on the user's nerves or muscles, the muscles include single muscle groups or multiple muscle groups, and the nerves include phrenic nerves.

[0083] Furthermore, the body parameter information includes age, joint mobility, muscle atrophy, spasticity parameters and lung function parameters; wherein, the lung function parameters include vital capacity, forced vital capacity and forced expiratory volume in one second.

[0084] Furthermore, the treatment terminal includes multiple treatment electrodes; the multiple treatment electrodes include electrodes attached to the posterior edges of the left and right sternocleidomastoid muscles, the left and right subclavian pectoralis major muscles, the oblique abdominal muscles, the rectus abdominis muscles, the quadriceps femoris muscles of both lower limbs, the tibialis anterior muscles, the common peroneal nerve, the tibialis anterior muscles, the peroneus longus and brevis muscles, the quadriceps femoris muscles, the hip abductor muscles, and the wrist and finger extensor muscles.

[0085] The above are only preferred embodiments of the present invention and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention description and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A respiratory neuromuscular stimulation control system, characterized in that: The system comprises a user stimulation parameter determination unit, an initial muscle contraction filter signal determination unit, and an electrical stimulation current real-time control unit; wherein the user stimulation parameter determination unit is used to collect the user's body parameter information; and determine the user's stimulation parameters according to the body parameter information and the parameters stored in the database; The initial muscle contraction filter signal determination unit is configured to process the initial muscle contraction signal obtained from the treatment terminal through a signal processing module before electrically stimulating the user's body to obtain an initial muscle contraction filter signal; wherein the initial muscle contraction signal is sensed by the treatment terminal attached to the user's body; The electrical stimulation current real-time control unit is used to control the electrical stimulation current generated by the treatment terminal provided on the user's body in real time according to the user's stimulation parameters, the initial muscle contraction filter signal, the real-time muscle contraction filter signal, and the real-time breathing condition through a trained and optimized adaptive current adjustment model; wherein the real-time breathing condition includes the user's inhalation time and respiratory rate; the electrical stimulation current acts on the user's nerves or muscles, the muscles including single muscle groups or multiple muscle groups, and the nerves including phrenic nerves; The adaptive current regulation model is constructed and trained based on the following methods: Acquiring historical user stimulation data, historical muscle contraction signals, historical respiration data, and real-time electrical stimulation current data; wherein the historical muscle contraction signals include initial muscle contraction signals and real-time stimulated muscle contraction signals; and the historical respiration data include initial respiration data and real-time stimulated respiration data; determining a set of stimulus factors of the historical user stimulus data; Determining a weight vector and a membership vector for each stimulus factor in the stimulus factor set; Performing fuzzy transformation on the weight vector and the membership vector of each stimulus factor to obtain a first fuzzy evaluation matrix; Performing fuzzy operations on the first fuzzy evaluation matrix and the weight vectors of the respective stimulus factors and performing normalization to obtain a second fuzzy evaluation matrix; Constructing a deep neural network according to the second fuzzy evaluation matrix, including an input layer, a basic network, and an adaptive weight layer; Performing feature extraction on the initial respiratory data and the real-time stimulated respiratory data to obtain an initial respiratory feature vector and a real-time stimulated respiratory feature vector; Calculating the distance between the initial respiratory feature vector and the real-time stimulated respiratory feature vector; According to the distance between the initial respiratory feature vector and the real-time stimulated respiratory feature vector, the initial muscle contraction signal, the real-time stimulated muscle contraction signal, the initial respiratory feature vector, the real-time stimulated respiratory feature vector and the real-time electrical stimulation current data, the constructed deep neural network is optimized and trained until the loss function converges, including: constructing a time series sample vector according to the distance between the initial respiratory feature vector and the real-time stimulated respiratory feature vector, the initial respiratory feature vector and the real-time stimulated respiratory feature vector: inputting the time series sample vector into the input layer of the deep neural network to obtain a respiratory time series output data set output by the first fully connected layer; performing non-negative matrix decomposition on the initial muscle contraction signal and the real-time stimulated muscle contraction signal based on the time series sample vector to obtain a signal non-negative matrix decomposition set; performing window processing on the signal non-negative matrix decomposition set to obtain time domain samples; performing short-time Fourier transform on the time domain samples to obtain frequency domain samples; wherein, determining the first adaptive loss function according to the first adaptive adjustment function preset in the adaptive weight layer; ； in, is the number of adaptive weight layers; is the number of the adaptive weight layer; For the The first strategy of layer-adaptive weight layer; is the first adaptive adjustment function; 1 is the regularization parameter; is a natural constant; is the first related quantity; The second correlation between the frequency domain samples and the respiratory time series output data set is quantified by means of information entropy and mutual information, and a second adaptive loss function is determined according to a second adaptive adjustment function preset in the adaptive weight layer: ； in, is the number of adaptive weight layers; is the number of the adaptive weight layer; For the The second strategy of layer-adaptive weight layer; is the second adaptive adjustment function; is the regularization parameter; is a natural constant; is the second related quantity; The loss function is constructed according to the first adaptive loss function and the second adaptive loss function, and the deep neural network is optimized and trained until the loss function converges.

2. The respiratory nerve muscle stimulation control system according to claim 1, characterized in that: The basic network includes six convolutional layers, three fully connected layers and a pooling layer; wherein the convolutional layer is connected to the pooling layer, and the pooling layer is used to aggregate the values ​​of adjacent areas; the three fully connected layers include a first fully connected layer, a second fully connected layer and a third fully connected layer; the adaptive weight layer is used to complete the adaptive adjustment of the current parameters through back propagation.

3. The respiratory nerve muscle stimulation control system according to claim 2, characterized in that: The real-time electrical stimulation current data includes stimulation intensity, pulse width, pulse frequency and stimulation duration of the current.

4. The respiratory nerve muscle stimulation control system according to claim 1, characterized in that: The physical parameter information includes age, joint mobility, muscle atrophy, spasticity parameters and lung function parameters; wherein the lung function parameters include vital capacity, forced vital capacity and forced expiratory volume in one second.

5. The respiratory nerve muscle stimulation control system according to claim 1, characterized in that: The treatment terminal includes multiple treatment electrodes; the multiple treatment electrodes include electrodes attached to the posterior edges of the left and right sternocleidomastoid muscles, the left and right subclavian pectoralis major muscles, the oblique abdominal muscles, the rectus abdominis muscles, the quadriceps femoris muscles of both lower limbs, the tibialis anterior muscles, the common peroneal nerve, the tibialis anterior muscles, the peroneus longus and brevis muscles, the quadriceps femoris muscles, the hip abductor muscles, and the wrist and finger extensor muscles.

Images