KR20240036169A - Functional electrical stimulation therapy system using electromyogram signal - Google Patents

Abstract

A treatment method using an electrical stimulation treatment device that collects electromyography signals (EMG) generated in response to electrical stimulation from the body according to an embodiment of the present invention includes extracting a characteristic vector from the frequency domain of the electromyographic signal, and an artificial intelligence model. A step of distinguishing and detecting a voluntary muscle contraction signal and an involuntary muscle contraction signal from the characteristic vector extracted by applying a step of detecting the signal, removing the involuntary muscle contraction signal from the electromyogram signal according to the detection result, and the involuntary muscle contraction signal. A step of calculating the root mean square (RMS) of the electromyographic signal from which the muscle contraction signal has been removed, comparing the effective value and a threshold value, and generating a functional electrical stimulation signal to be applied to the body according to the comparison result. Includes.

Description

Functional electrical stimulation treatment device using electromyography signals and its treatment method {FUNCTIONAL ELECTRICAL STIMULATION THERAPY SYSTEM USING ELECTROMYOGRAM SIGNAL}

The present invention relates to a functional electrical stimulation treatment device, and more specifically, to an electrical stimulation treatment device that generates a functional electrical stimulation (FES) signal based on electromyography signals and a treatment method thereof.

Electromyography (EMG), which measures the degree of muscle activity by measuring the potential difference generated in muscle cells when the muscle is activated, is widely used not only in the medical field but also in the biomechanics field. EMG technology has been developed according to the configuration of electrodes that measure the potential difference of activated muscles, and the most commonly used type is EMG equipment in the form of electrodes attached to the skin surface. In addition, electrical stimulation technology is a technology that artificially induces muscle contraction by applying electrical stimulation to muscles in the form of constant current or constant voltage. Electrical stimulation technology has mainly been developed into functional electrical stimulation (FES) technology, which complements and replaces the function of weakened or lost muscles.

Functional electrical stimulation (FES) has generally been known to be the most effective rehabilitation treatment available in hospitals. For treatment using functional electrical stimulation (FES), rehabilitation experts apply electrical stimulation to the affected area while voluntary muscle contraction occurs. Rehabilitation experts perform the procedure by visually determining whether the patient is maintaining or starting muscle contraction and turning on the FES device. In general FES equipment, it is operated in such a way that electrical stimulation is generated when the user applies a certain amount of force.

Therefore, for rehabilitation treatment using FES, one rehabilitation expert has no choice but to handle one patient. Therefore, when there are a large number of patients, a shortage of manpower is inevitable even if FES equipment is used. Even in rehabilitation treatment in a home environment using FES devices, it is difficult to provide the most effective FES rehabilitation treatment without a rehabilitation expert. It is difficult for rehabilitation specialists to treat multiple patients because they must determine whether the patient is maintaining or starting voluntary muscle contraction. In conventional products or technologies, electrical stimulation is controlled to be output when the user applies a certain amount of force, so the equipment does not recognize whether the muscle is contracting after the electrical stimulation is output. Since a device is needed to automatically apply electrical stimulation when a patient's voluntary muscle contraction occurs, technology to analyze and judge the input signal is required. Therefore, there is a need for FES technology that allows patients to perform effective FES rehabilitation treatment on their own without a rehabilitation expert.

Republic of Korea Patent Publication No. 10-2018-0074597

The present invention is intended to solve the above-mentioned technical problems, and the aim of the present invention is to provide an electrical stimulation treatment device that generates functional electrical stimulation (FES) based on muscle stimulation signals. When muscles are stimulated by electrical stimulation, involuntary muscle contraction also occurs. The purpose of the present invention is to use preprocessing technology to distinguish involuntary muscle contraction from voluntary muscle contraction. The aim is to provide an effective FES treatment device using induced muscle contraction signals.

The electrical stimulation treatment device that collects electromyography signals (EMG) from the body and generates functional electrical stimulation signals according to an embodiment of the present invention extracts characteristic vectors from the electromyography signals to create voluntary muscle contraction signals and involuntary muscle contraction signals. A voluntary/involuntary contraction detection unit that detects the involuntary contraction signal separately, an involuntary contraction signal removal unit that removes the involuntary muscle contraction signal from the EMG signal according to the detection result, and an effective value of the EMG signal from which the involuntary muscle contraction signal is removed. It includes a muscle activity intensity calculation unit that calculates (RMS: Root Mean Square), and a functional electrical stimulation control unit that compares the effective value and the threshold value and generates the functional electrical stimulation signal to be applied to the body according to the comparison result. .

In this embodiment, the feature vector includes at least one of a percentile cumulative sum of spectra (PoSCS) and a log power spectrum detected in the frequency domain of the electromyography signal (EMG).

In this embodiment, the voluntary/involuntary contraction detection unit includes an artificial intelligence model that distinguishes the voluntary muscle contraction signal and the involuntary muscle contraction signal from the characteristic vector.

In this embodiment, the artificial intelligence model uses the LSTM algorithm.

The electrical stimulation treatment device that collects electromyography signals (EMG) generated from the body and generates functional electrical stimulation signals according to an embodiment of the present invention extracts characteristic vectors from the frequency domain of the EMG signals and applies an artificial intelligence model. A voluntary/involuntary muscle contraction detection unit that detects the voluntary muscle contraction signal by distinguishing it from the involuntary muscle contraction signal from the extracted characteristic vector, and a voluntary/involuntary muscle contraction signal that removes the involuntary muscle contraction signal from the electromyogram signal according to the detection result. An involuntary muscle contraction signal removal unit, a muscle activity intensity calculation unit that calculates the root mean square (RMS) of the EMG signal from which the involuntary muscle contraction signal has been removed, and compares the effective value and the threshold value, and provides the comparison result. It includes a functional electrical stimulation control unit that generates the functional electrical stimulation signal to be applied to the body.

In this embodiment, the feature vector includes at least one of a percentile cumulative sum of spectra (PoSCS) and a log power spectrum detected in the frequency domain of the electromyography signal (EMG).

In this embodiment, the involuntary muscle contraction signal removal unit removes the involuntary muscle contraction signal by attenuating a section of the EMG signal including the involuntary muscle contraction signal by 6 dB.

In this embodiment, the artificial intelligence model uses an artificial intelligence algorithm to distinguish between the involuntary muscle contraction signal and the voluntary muscle contraction signal from the electromyography signal.

In this embodiment, the involuntary muscle contraction signal removal unit includes a window unit that selects a window of the electromyography signal (EMG), a fast Fourier transform unit that processes the signal included in the selected window by fast Fourier transform, and the fast Fourier transform unit. It includes a magnitude and phase calculation unit that calculates the size and phase of the signal output from the conversion unit, a peak detection unit that detects a peak in the size of the signal, and a peak removal unit that filters the noise signal corresponding to the detected peak. .

The treatment method of the electrical stimulation treatment device that collects electromyography signals (EMG) from the body and generates functional electrical stimulation signals according to an embodiment of the present invention includes extracting characteristic vectors from the electromyography signals and dividing them into spontaneous muscle contraction signals and involuntary muscle contraction signals. A step of distinguishing and detecting a muscle contraction signal, removing the involuntary muscle contraction signal from the EMG signal according to the detection result, the root mean square (RMS) of the EMG signal from which the involuntary muscle contraction signal is removed. It includes calculating , comparing the effective value and the threshold, and generating the functional electrical stimulation signal to be applied to the body according to the comparison result.

A treatment method using an electrical stimulation treatment device that collects electromyography signals (EMG) generated in response to electrical stimulation from the body according to an embodiment of the present invention includes extracting a characteristic vector from the frequency domain of the electromyographic signal, and an artificial intelligence model. A step of distinguishing and detecting a voluntary muscle contraction signal and an involuntary muscle contraction signal from the characteristic vector extracted by applying a step of detecting the signal, removing the involuntary muscle contraction signal from the electromyogram signal according to the detection result, and the involuntary muscle contraction signal. A step of calculating the root mean square (RMS) of the electromyographic signal from which the muscle contraction signal has been removed, comparing the effective value and a threshold value, and generating a functional electrical stimulation signal to be applied to the body according to the comparison result. Includes.

According to the electrical stimulation treatment system according to an embodiment of the present invention, a functional electrical stimulation (FES) signal can be generated by distinguishing between voluntary and involuntary muscle contraction signals from electromyography signals (EMG) with high accuracy. Therefore, it is possible to implement an electrical stimulation treatment device that provides functional electrical stimulation (FES) with high accuracy without the need for rehabilitation experts or expensive equipment.

1 is a block diagram illustrating a sarcopenia diagnosis and treatment system according to an embodiment of the present invention.
Figure 2 is a block diagram illustrating an electrical stimulation treatment system according to an embodiment of the present invention.
FIG. 3 is a block diagram exemplarily showing the configuration of the electrical stimulation treatment system of FIG. 2.
Figure 4 is a flowchart showing a processing method in the frequency domain to extract percentile spectral cumulative sum (PoSCS) as an example of feature extraction.
Figure 5 is a graph showing a method of extracting percentile spectral cumulative sum (PoSCS).
Figure 6 shows probability density functions (PDFs) representing the results of extracting the percentile spectral cumulative sum (PoSCS) by frequency from the electromyography signal (EMG).
Figure 7 is a flowchart showing a learning method of the artificial intelligence calculation unit for separating voluntary muscle contraction signals and involuntary muscle contraction signals according to an embodiment of the present invention.
Figure 8 is a diagram briefly showing the structure of the LSTM algorithm for identifying voluntary muscle contraction signals and involuntary muscle contraction signals through sequential electromyography (EMG) data in the time domain of the present invention.
Figure 9 is a flowchart showing the actual operation and test operation of the voluntary/involuntary muscle contraction detection unit of the present invention.
FIG. 10 is a block diagram illustrating an involuntary muscle contraction signal removal unit shown in FIG. 3.
11 and 12 are graphs showing the results of frequency analysis of EMG data in the peak detection unit and peak suppression unit 1126.
Figure 13 is a graph showing a waveform with pre-processing in the inverse converter (red) and a waveform without pre-processing (black).
Figures 14 and 15 are diagrams showing the results of testing the performance of the electrical stimulation treatment system that generates functional electrical stimulation (FES) based on muscle stimulation signals of the present invention.

Hereinafter, embodiments of the present invention will be described clearly and in detail so that a person skilled in the art can easily practice the present invention.

1 is a block diagram illustrating a sarcopenia diagnosis and treatment system according to an embodiment of the present invention. Referring to FIG. 1, the sarcopenia diagnosis and treatment system may include a sarcopenia diagnosis device 1000 and a sarcopenia treatment device 1100. The sarcopenia diagnosis device 1000 can diagnose the user's muscle quality (MQ) and sarcopenia. The sarcopenia diagnosis device 1000 can diagnose sarcopenia by measuring muscle strength, muscular endurance, etc. The sarcopenia diagnosis device 1000 according to an embodiment of the present invention obtains information related to muscle strength or muscular endurance using an electrical stimulation-based impact response signal (ES-based IR), and artificial By applying an intelligent learning model, sarcopenia can be diagnosed simply, quickly, and accurately. In particular, the sarcopenia diagnosis device 1000 of the present invention will use a multi-frequency impulse response signal (m-FIRS) as an electrical stimulation-based response signal (ES-based IR).

The sarcopenia treatment device 1100 can detect minute movements of muscles through electromyography (EMG) and apply electrical stimulation (ES) to assist the patient's movements. Additionally, the sarcopenia treatment device 1100 can provide effective rehabilitation training by applying electrical stimulation while voluntarily exerting force.

Figure 2 is a block diagram illustrating an electrical stimulation treatment system according to an embodiment of the present invention. Referring to FIG. 2, the electrical stimulation treatment system 1100 performs preprocessing through electromyography signals (EMG) obtained by applying electrical stimulation, and uses the preprocessed data to perform functional electrical stimulation (FES) to treat the patient. creates .

The electrical stimulation treatment system 1100 applies electrical stimulation (ES) to the patient's muscles or skin and generates functional electrical stimulation (FES) based on electromyography signals (EMG) provided in response to the electrical stimulation (ES). . Unlike general functional electrical stimulation (FES) signals, the electrical stimulation treatment system 1100 applies preprocessing technology to separate and remove involuntary muscle contraction signals from electromyography (EMG) signals. Of course, the collected electromyography (EMG) signals include electrical stimulation (ES) signals. The electrical stimulation treatment system 1100 extracts a voluntary muscle contraction signal by removing the electrical stimulation (ES) signal and the involuntary muscle contraction signal from the electromyography signal (EMG), and performs functional electrical stimulation ( FES) signal can be generated. Therefore, since the functional electrical stimulation (FES) signal of the present invention is generated based on electromyography signals (EMG), it will hereinafter be referred to as electromyography-based functional electrical stimulation (ECF: EMG-Controlled FES).

The electrical stimulation treatment system 1100 can control functional electrical stimulation (FES) based on electromyography signals (EMG) that measure muscle activity due to muscle contraction. The electrical stimulation treatment system 1100 can adjust the intensity of electrical stimulation according to the root mean square (RMS) size of the electromyography signal (EMG). Through this, the electrical stimulation treatment system 1100 can provide a rehabilitation treatment service in which the electrical stimulation is turned on when a certain amount of force is applied, and the electrical stimulation is turned off when the force falls below a certain level. In addition, the electrical stimulation treatment system 1100 can provide assistance by applying electrical stimulation to supplement the insufficient force when the force required for a specific movement cannot be applied.

The electrical stimulation treatment system 1100 of the present invention performs treatment of patients using electromyography-based functional electrical stimulation (ECF). For this purpose, pad-shaped electrodes may be used. The electrical stimulation pad 1111 may include an electrical stimulation pad that applies electrical stimulation (ES) and functional electrical stimulation (ECF) based on electromyography signals. For example, the electrical stimulation pad 1111 can be used in a wet form for one-time use or multiple use. Alternatively, the electrical stimulation pad 1111 may be manufactured using a dry, high-adhesive material to transmit the user's biological signals or electrical stimulation signals of the innervated muscles. For example, the electrical stimulation pad 1111 may be manufactured as a conductive dry adhesive electrode pad using carbon nanomaterial. The electrical stimulation measurement pad 1112 is used for electromyography (EMG) measurement. The electrical stimulation measurement pad 1112 may include an EMG sensor for sensing thigh electromyography. In addition, the reference pad 1113 is provided as an electrode pad to provide a ground level for the electrical stimulation pad 1111 or the electrical stimulation measurement pad 1112.

FIG. 3 is a block diagram exemplarily showing the configuration of the electrical stimulation treatment system of FIG. 2. Referring to FIG. 3, the electrical stimulation treatment system 1100 includes a voluntary/involuntary muscle contraction detection unit 1110, an involuntary muscle contraction signal removal unit 1120, a muscle activity intensity calculation unit 1130, and functional electrical stimulation. It may include a control unit 1140.

The voluntary/involuntary muscle contraction detection unit 1110 receives electromyography signals (EMG) collected in response to electrical stimulation (ES). The voluntary/involuntary muscle contraction detection unit 1110 can remove the electrical stimulation (ES) included in the input electromyography signal (EMG) and distinguish between the voluntary muscle contraction signal and the involuntary muscle contraction signal. It is difficult to distinguish between voluntary and involuntary muscle contraction signals based on signal amplitude alone. Therefore, an artificial intelligence (AI) model is needed to separate voluntary and involuntary muscle contraction signals.

For high-performance signal classification, in addition to a high-performance deep learning model, high-performance feature vectors are also important to improve model performance. Therefore, in the present invention, a sampling rate of 850Hz, a frame size of 320 samples, a shift size of 20 samples, and an FFT size of 512 can be applied for feature extraction. Since the frame size is 320 samples, 320 samples will be sequentially stored in the buffer. And, after 320 samples have elapsed, signals of 20 samples each can be sequentially stored in a buffer. After updating the buffer, feature vectors will be extracted using feature extraction techniques.

The involuntary muscle contraction signal removal unit 1120 removes the detected involuntary muscle contraction signal. The muscle activity intensity calculation unit 1130 calculates the RMS in a noise-removed state to intuitively determine how much force was applied.

The functional electrical stimulation control unit 1140 generates electromyography-based functional electrical stimulation (ECF). That is, the functional electrical stimulation control unit 1140 can turn on or off the application of functional electrical stimulation by comparing the RMS with a specific threshold. For example, the functional electrical stimulation control unit 1140 may apply functional electrical stimulation if the RMS is greater than or equal to the threshold, and may not apply functional electrical stimulation if it is smaller than the threshold. Alternatively, the functional electrical stimulation control unit 1140 may determine the intensity of functional electrical stimulation according to RMS. The functional electrical stimulation control unit 1140 can control the electrical stimulation so that it becomes stronger as the RMS increases and becomes weaker as the RMS decreases.

The electrical stimulation treatment system 1100 can control functional electrical stimulation (FES) based on electromyography signals (EMG) that measure muscle activity due to muscle contraction. The electrical stimulation treatment system 1100 can adjust the intensity of electrical stimulation according to the RMS size of the EMG. Through this, the electrical stimulation treatment system 1100 can provide a rehabilitation treatment service in which the electrical stimulation is turned on when a certain amount of force is applied, and the electrical stimulation is turned off when the force falls below a certain level. In addition, the electrical stimulation treatment system 1100 can provide assistance by applying electrical stimulation to supplement the insufficient force when the force required for a specific movement cannot be applied.

Figure 4 is a flowchart showing a processing method in the frequency domain to extract percentile spectral cumulative sum (PoSCS) as an example of feature extraction. Referring to FIG. 4, the Percentile of Spectral Cumulative Sum (PoSCS) can be extracted through frequency domain processing of the electromyography signal (EMG).

In step S110, a window in the time domain of the electromyography signal (EMG) to be converted to a frequency spectrum is selected. For example, a window of electromyography (EMG) signals may be selected on a sector-by-sector basis or a frame-by-frame basis.

In step S120, fast Fourier transform (FFT) and absolute value calculation are performed on the window of the electromyography signal (EMG) of the selected section.

In step S130, a cumulative spectrum sum (SCS) is extracted in the frequency domain based on the absolute value calculation result.

In step S140, normalization operation is performed.

In step S150, the percentile spectral cumulative sum (PoSCS) of each frequency is extracted based on the normalized data.

Figure 5 is a graph showing a method of extracting percentile spectral cumulative sum (PoSCS). Referring to FIG. 5, extraction of percentile spectral cumulative sum (PoSCS) may be exemplarily performed through the following process.

First, the magnitude is accumulated in the positive x-axis direction in the frequency domain, and then the max-normalization data is used. Next, the horizontal line is shifted from 0.05 to 0.30 in 0.05 increments based on the y-axis, and then the frequency bin of the contact point between the horizontal line and the cumulative spectral sum (SCS) is extracted as a feature. By extracting a 6-order feature vector for each frame, the use of spectral cumulative sum (SCS) can be used to effectively distinguish between involuntary muscle contraction and voluntary muscle contraction. This process is shown in the graph on the right, which shows the extracted features in frequency bins.

The noise component in the frequency domain that occurs due to involuntary muscle contraction has an abnormally bouncing value, unlike the frequency component of voluntary muscle contraction. Therefore, when involuntary muscle contraction and voluntary muscle contraction exist simultaneously, the characteristics of the cumulative spectral sum (SCS) appear differently than when only involuntary muscle contraction exists. Additionally, noise components in the frequency domain that occur due to involuntary muscle contraction appear differently depending on the frequency parameters of electrical stimulation (ES). The characteristics that stand out in the voluntary muscle contraction section differ depending on the electrical stimulation environment. Therefore, in order to construct a high-performance model for all electrical stimulation environments, multi-dimensional feature vectors must be utilized, as mentioned above. As a result, it can be seen that percentile spectral cumulative sum (PoSCS) appears prominently in the voluntary muscle contraction section.

Figure 6 is a probability density function (PDF) showing the results of extracting the percentile spectral cumulative sum (PoSCS) by frequency from the electromyography signal (EMG). Referring to FIG. 6, in the electromyography signal (EMG) in the time domain, involuntary muscle contraction and voluntary muscle contraction can be distinguished in each frequency band. However, it can be seen that there is an overlap. Therefore, it can be seen that separation calculation using artificial intelligence algorithm is necessary.

After extracting the percentile spectral cumulative sum (PoSCS) for each frequency, the activity density function (PDF) of the characteristic for the involuntary muscle contraction signal is expressed as curves (C11, C12, C13) at each frequency (10Hz, 60Hz, 90Hz). appear. And the probability density function of the characteristics of the voluntary muscle contraction signal appears as curves (C21, C22, C23) at each frequency (10Hz, 60Hz, 90Hz). According to the extraction results of percentile spectral cumulative sum (PoSCS), involuntary muscle contraction signals and voluntary muscle contraction signals at low frequencies have different averages, allowing relatively clear distinction. However, since there is some overlap between the extracted features, deep learning or artificial intelligence techniques for the extracted features are needed to provide high classification resolution at any frequency. In particular, in the present invention, the LSTM (Long Short Term Memory) algorithm, which provides the highest performance for long-term time series data, will be used.

Figure 7 is a flowchart showing a learning method of the artificial intelligence calculation unit for separating voluntary muscle contraction signals and involuntary muscle contraction signals according to an embodiment of the present invention. Referring to FIG. 7, the voluntary/involuntary muscle contraction detection unit 1110 (see FIG. 2) can use the input electromyography signal (EMG) to learn LSTM, a type of recurrent neural network (RNN). Through learning, high-resolution discrimination between voluntary and involuntary muscle contraction signals is possible.

In step S210, the voluntary/involuntary muscle contraction detection unit 1110 may collect electromyography (EMG) data. The voluntary/involuntary muscle contraction detection unit 1110 can apply electrical stimulation (ES) to the body muscles and measure electromyography (EMG) data.

In step S220, the voluntary/involuntary muscle contraction detection unit 1110 may analyze electromyography (EMG) data and extract characteristic vectors. The voluntary/involuntary muscle contraction detection unit 1110 may remove noise signals included in electromyography (EMG) data and then extract feature vectors related to muscle strength or muscular endurance.

In step S230, the voluntary/involuntary muscle contraction detection unit 1110 performs learning of an artificial intelligence (AI) model based on the characteristic vector. The voluntary/involuntary muscle contraction detection unit 1110 generates learning data for artificial intelligence learning. The voluntary/involuntary muscle contraction detection unit 1110 may create a learning database (DB) based on the characteristic vector (S231). The voluntary/involuntary muscle contraction detection unit 1110 may initialize the LSTM weight (S232). The voluntary/involuntary muscle contraction detection unit 1110 shuffles the learning database (DB). That is, the voluntary/involuntary muscle contraction detection unit 1110 may provide learning data to a fully connected neural network (FCNN) and process it through a learning operation (S233). The voluntary/involuntary muscle contraction detection unit 1110 can calculate the current LSTM model error (S234). The voluntary/involuntary muscle contraction detection unit 1110 determines whether the error (epoch) learned to date is smaller than the total error (total epoch) (S235). The voluntary/involuntary muscle contraction detection unit 1110 terminates (NO) if the epoch learned to date is not less than the total error (total epoch). On the other hand, if the epoch learned to date is less than the total epoch (YES), the voluntary/involuntary muscle contraction detection unit 1110 updates the LSTM weights (S236) and returns to step S233.

Figure 8 is a diagram briefly showing the structure of the LSTM algorithm for identifying voluntary muscle contraction signals and involuntary muscle contraction signals through sequential electromyography (EMG) data in the time domain of the present invention. Referring to FIG. 8, updating, or learning, of weights for electromyography (EMG) data provided sequentially in time is performed by the LSTM algorithm.

The structure of the LSTM algorithm consists of LSTM cells that process sequentially input data (Dt). Each LSTM cell decides how much past data to remember or discard based on its current state, reflects the current output in the result, and passes it to the next LSTM cell. For this function, one LSTM cell consists of a forget gate, an input gate, and an output gate to process the current input data (Dt).

Figure 9 is a flowchart showing the actual operation and test operation of the voluntary/involuntary muscle contraction detection unit of the present invention. Referring to FIG. 9, the voluntary/involuntary muscle contraction detection unit 1110 can distinguish between voluntary muscle contraction signals and involuntary muscle contraction signals using the LSTM learned in FIG. 8 based on the input electromyography signal (EMG). You can.

In step S310, the voluntary/involuntary muscle contraction detection unit 1110 may collect electromyography (EMG) data. The voluntary/involuntary muscle contraction detection unit 1110 can apply electrical stimulation (ES) to the body muscles and measure electromyography (EMG) data.

In step S320, the voluntary/involuntary muscle contraction detection unit 1110 may analyze electromyography (EMG) data and extract characteristic vectors. The voluntary/involuntary muscle contraction detection unit 1110 may remove noise electrical signals included in electromyography (EMG) data and then extract feature vectors related to muscle strength or muscular endurance.

In step S330, the voluntary/involuntary muscle contraction detection unit 1110 performs an LSTM operation based on feature vectors sequentially input in time series. In step S340, the parameters (Wo) of the output layer provided as a result of the LSTM operation are provided. The voluntary/involuntary muscle contraction detection unit 1110 provides an output value (y) using the parameter (Wo). In step S350, the voluntary/involuntary muscle contraction detection unit 1110 finally uses the output value (y) to perform classification using a threshold and outputs the result.

FIG. 10 is a block diagram illustrating an involuntary muscle contraction signal removal unit shown in FIG. 3. Referring to FIG. 10, the involuntary muscle contraction signal removal unit 1120 includes a window unit 1121, a Fast Fourier Transform (FFT) 1122, and a magnitude and phase calculation unit 1123. , 1124), a peak detection unit 1125, a peak removal unit 1126, and an inverse FFT (IFFT) 1127.

The window unit 1121 performs windowing of an input signal (eg, EMG signal) in the time domain into a signal in the frequency domain. The window unit 1121 can shift in real time in units of 20 samples, configure a frame in units of 512 samples, and operate with an FFT size of 512. The fast Fourier transform unit 1122 performs Fourier transform, and the calculation units 1123 and 1124 calculate magnitude and phase. The peak detection unit 1125 and peak removal unit 1126 detect noise by detecting the peak of the waveform and perform peak suppression through substitution. The involuntary muscle contraction component appears in peak-shaped magnitude like an impulse. Accordingly, the peak detection unit 1125 detects an involuntary muscle contraction component whose magnitude appears like an impulse. The peak removal unit 1126 replaces the detected peak with an eps (=2.2204e-16) value and then performs IFFT to obtain a signal from which the involuntary muscle contraction component is removed. The inverse conversion unit 1127 performs inverse conversion using the magnitude of the waveform that has passed through the peak detection unit 1125 and the peak removal unit 1126 and the phase of the waveform calculated previously, and generates an output signal. .

Ultimately, the involuntary muscle contraction signal removal unit 1120 uses an adaptive noise suppression algorithm that detects and then removes the peak signal related to the involuntary muscle contraction signal in the frequency domain. As the frequency of electrical stimulation (ES) changes, the frequency component of the involuntary muscle contraction signal also changes. Using this adaptive noise suppression algorithm, involuntary muscle contraction signals of variable frequency can be effectively removed. When using a method with a fixed filter band, performance deviations may occur depending on the situation or user, so the involuntary muscle contraction signal removal unit 1120 using an adaptive noise suppression algorithm can provide stable performance. there is.

11 and 12 are graphs showing the results of frequency analysis of EMG data in the peak detection unit 1125 and the peak suppression unit 1126. Figure 13 is a graph showing a waveform (red) with pre-processing in the inverse converter 1127 and a waveform without pre-processing (black).

The involuntary muscle contraction signal removal unit 1120 can be implemented with an adaptive noise suppression algorithm that finds and removes peak signals related to involuntary muscle contraction (i.e., noise) in the frequency domain. there is. As the frequency of electrical stimulation changes, the frequency component of involuntary muscle contraction changes. The involuntary muscle contraction signal removal unit 1120 may adaptively remove the involuntary muscle contraction component in order to effectively remove it.

If a fixed filter is used, performance differences may occur depending on the situation or person. The involuntary muscle contraction signal removal unit 1120 is designed to solve this problem and can reduce performance deviations depending on the situation or person.

The involuntary muscle contraction signal removal unit 1120 may operate in the following manner. For example, the involuntary muscle contraction signal removal unit 1120 shifts in real time in units of 20 samples, configures a frame in units of 512 samples, and sets the FFT size to 512 to drive the algorithm. can do. The involuntary muscle contraction signal removal unit 1120 performs FFT on a predefined frame, calculates the magnitude and phase, and removes the involuntary muscle contraction component that appears like an impulse in the magnitude. You can detect peaks to find them.

The involuntary muscle contraction signal removal unit 1120 replaces the peak with an eps (= 2.2204e-16) value to remove the involuntary muscle contraction component and then performs IFFT to obtain a signal from which the involuntary muscle contraction component has been removed. can be obtained. The involuntary muscle contraction signal removal unit 1120 can finally classify the involuntary muscle contraction as an involuntary muscle contraction through an involuntary muscle contraction signal removal algorithm, and obtain a signal by removing the magnitude (magnitude) by 6 dB when it is not in the contraction section.

Referring to FIG. 13, the electrical stimulation treatment system 1100 using electromyography-based functional electrical stimulation (ECF) of the present invention can provide high removal efficiency of involuntary muscle contraction signals. Involuntary muscle contraction signals included in electromyography signals (EMG) in the time domain can be effectively removed by applying an adaptive noise suppression algorithm. Therefore, the electrical stimulation treatment system 1100 (see FIG. 1) of the present invention can generate functional electrical stimulation (ECF) based on low-noise voluntary muscle contraction signals. Therefore, highly reliable functional electrical stimulation treatment is possible without relying on experts or expensive devices.

Figures 14 and 15 are diagrams showing the results of testing the performance of the electrical stimulation treatment system that generates functional electrical stimulation (FES) based on muscle stimulation signals of the present invention. For testing the electrical stimulation treatment system 1100 of the present invention, electrical stimulation (ES) was applied at increasing frequencies from 10 Hz to 90 Hz in 5 Hz increments. And while electromyographic signals (EMG) were collected while electrical stimulation (ES) was provided, the measurement subject rested for 10 seconds, performed voluntary muscle contraction for 20 seconds, and rested for 10 seconds again. . A total of about 40 seconds of data is collected per person, and a database was constructed using data collected in the same way for a total of 6 people. The section in which voluntary muscle contraction was maintained was leveled to '1', and the section in which only involuntary muscle contraction existed was leveled to '0'.

The artificial intelligence model uses all extracted characteristics as input, and in addition, the initialization of the artificial intelligence model uses random initialization, and fine-tuning uses backpropagation. , the number of fully connected layers was set to 1 and the number of units was also set to 1. In addition, the Adaptive Momentum Estimation (Adam) method was used as an optimization algorithm to determine the weight update method. In addition, the cost function is binary cross entropy, the activation function is hyperbolic tangent, the number of cells is 3, and the number of hidden units per cell is 128. , 64, 32 were used.

Referring to Figure 14, the removal performance of involuntary muscle contraction signals in the case of using LSTM, the artificial intelligence model of the present invention, and in the case of applying general artificial intelligence models (SVM, ANN, DNN) is shown in the table. It is shown as According to the test results for a total of two groups (Set1, Set2), the total accuracy (TA) was the best when using the LSTM model, at 90.01% and 82.82%, respectively.

Referring to Figure 15, the results of the frequency-specific performance (AUC: Area under the Curve) evaluation of artificial intelligence models for two groups (Set1, Set2) are shown graphically. It was observed that the reliability (AUV) was highest across all experimental frequencies when using the LSTM model for each of the two groups (Set1, Set2).

The above-described details are specific embodiments for carrying out the present invention. In addition to the above-described embodiments, the present invention will also include embodiments that can be simply changed or easily changed in design. In addition, the present invention will also include technologies that can be easily modified and implemented using the embodiments. Therefore, the scope of the present invention should not be limited to the above-described embodiments, but should be determined by the claims and equivalents of the present invention as well as the claims described later.

Claims (20)

In the electrical stimulation treatment device that collects electromyography signals (EMG) from the body and generates functional electrical stimulation signals:
a voluntary/involuntary contraction detection unit that extracts a characteristic vector from the electromyogram signal and detects it by distinguishing between a voluntary muscle contraction signal and an involuntary muscle contraction signal;
an involuntary contraction signal removal unit that removes the involuntary muscle contraction signal from the EMG signal according to the detection result;
a muscle activity intensity calculator that calculates a root mean square (RMS) value of the EMG signal from which the involuntary muscle contraction signal has been removed; and
An electrical stimulation treatment device comprising a functional electrical stimulation control unit that compares the effective value and the threshold value and generates the functional electrical stimulation signal to be applied to the body according to the comparison result. According to claim 1,
The characteristic vector is an electrical stimulation treatment device including at least one of a percentile cumulative sum of spectra (PoSCS) and a log power spectrum detected in the frequency domain of the electromyography signal (EMG). According to claim 1,
The voluntary/involuntary contraction detection unit is an electrical stimulation treatment device including an artificial intelligence model that distinguishes the voluntary muscle contraction signal and the involuntary muscle contraction signal from the characteristic vector. According to claim 3,
The artificial intelligence model is an electrical stimulation treatment device that uses the LSTM algorithm. In the electrical stimulation treatment device that collects electromyography signals (EMG) generated from the body and generates functional electrical stimulation signals:
A voluntary/involuntary muscle contraction detection unit that extracts a characteristic vector from the frequency domain of the EMG signal and detects a voluntary muscle contraction signal and an involuntary muscle contraction signal from the extracted characteristic vector by applying an artificial intelligence model;
an involuntary muscle contraction signal removal unit that removes the involuntary muscle contraction signal from the EMG signal according to the detection result;
a muscle activity intensity calculator that calculates a root mean square (RMS) value of the EMG signal from which the involuntary muscle contraction signal is removed; and
An electrical stimulation treatment device comprising a functional electrical stimulation control unit that compares the effective value and the threshold value and generates the functional electrical stimulation signal to be applied to the body according to the comparison result. According to claim 5,
The characteristic vector is an electrical stimulation treatment device including at least one of a percentile cumulative sum of spectra (PoSCS) and a log power spectrum detected in the frequency domain of the electromyography signal (EMG). According to claim 5,
The involuntary muscle contraction signal removal unit attenuates the section of the EMG signal containing the involuntary muscle contraction signal by 6 dB to remove the involuntary muscle contraction signal. According to claim 5,
The artificial intelligence model is an electrical stimulation treatment device that distinguishes the involuntary muscle contraction signal and the voluntary muscle contraction signal from the electromyography signal using an artificial intelligence algorithm. According to claim 5,
The involuntary muscle contraction signal removal unit:
a window unit that selects a window of the electromyography signal (EMG);
a fast Fourier transform unit that processes signals included in the selected window by fast Fourier transform;
a magnitude and phase calculation unit that calculates the magnitude and phase of the signal output from the fast Fourier transform unit, respectively;
a peak detection unit that detects a peak in the magnitude of the signal; and
An electrical stimulation treatment device including a peak removal unit that filters a noise signal corresponding to the detected peak. In a treatment method using an electrical stimulation treatment device that collects electromyography signals (EMG) from the body and generates functional electrical stimulation signals:
Extracting a characteristic vector from the electromyogram signal and detecting it separately into a voluntary muscle contraction signal and an involuntary muscle contraction signal;
removing the involuntary muscle contraction signal from the EMG signal according to the detection result;
Calculating the root mean square (RMS) of the EMG signal from which the involuntary muscle contraction signal has been removed; and
A treatment method comprising comparing the effective value and the threshold value and generating the functional electrical stimulation signal to be applied to the body according to the comparison result. According to claim 10,
The electrical stimulation treatment device,
A voluntary/involuntary muscle contraction detection unit that extracts a characteristic vector from the frequency domain of the EMG signal and detects a voluntary muscle contraction signal and an involuntary muscle contraction signal from the extracted characteristic vector by applying an artificial intelligence model;
an involuntary muscle contraction signal removal unit that removes the involuntary muscle contraction signal from the EMG signal according to the detection result;
a muscle activity intensity calculator that calculates a root mean square (RMS) value of the EMG signal from which the involuntary muscle contraction signal is removed; and
A treatment method comprising a functional electrical stimulation control unit that compares the effective value and the threshold value and generates the functional electrical stimulation signal to be applied to the body according to the comparison result. According to claim 11,
The characteristic vector includes at least one of a percentile cumulative sum of spectra (PoSCS) and a log power spectrum detected in the frequency domain of the electromyography signal (EMG). According to claim 11,
A treatment method in which the involuntary muscle contraction signal removal unit removes the involuntary muscle contraction signal by attenuating the section of the EMG signal containing the involuntary muscle contraction signal by 6 dB. According to claim 11,
A treatment method in which the artificial intelligence model distinguishes between the involuntary muscle contraction signal and the voluntary muscle contraction signal from the electromyography signal using an artificial intelligence algorithm. According to claim 11,
The involuntary muscle contraction signal removal unit:
a window unit that selects a window of the electromyography signal (EMG);
a fast Fourier transform unit that processes signals included in the selected window by fast Fourier transform;
a magnitude and phase calculation unit that calculates the magnitude and phase of the signal output from the fast Fourier transform unit, respectively;
a peak detection unit that detects a peak in the magnitude of the signal; and
A treatment method comprising a peak removal unit that filters a noise signal corresponding to the detected peak. In a treatment method using an electrical stimulation treatment device that collects electromyography signals (EMG) generated in response to electrical stimulation from the body:
extracting a feature vector from the frequency domain of the electromyogram signal;
A step of detecting and distinguishing between voluntary muscle contraction signals and involuntary muscle contraction signals from the characteristic vector extracted by applying an artificial intelligence model;
removing the involuntary muscle contraction signal from the electromyography signal according to the detection result;
Calculating the root mean square (RMS) of the EMG signal from which the involuntary muscle contraction signal has been removed; and
A treatment method comprising comparing the effective value and the threshold value and generating a functional electrical stimulation signal to be applied to the body according to the comparison result. The electrical stimulation treatment device,
a voluntary/involuntary contraction detection unit that extracts a characteristic vector from the electromyogram signal and detects it by distinguishing between a voluntary muscle contraction signal and an involuntary muscle contraction signal;
an involuntary contraction signal removal unit that removes the involuntary muscle contraction signal from the EMG signal according to the detection result;
a muscle activity intensity calculator that calculates a root mean square (RMS) value of the EMG signal from which the involuntary muscle contraction signal has been removed; and
A treatment method comprising a functional electrical stimulation control unit that compares the effective value and the threshold value and generates the functional electrical stimulation signal to be applied to the body according to the comparison result. According to claim 17,
The characteristic vector includes at least one of a percentile cumulative sum of spectra (PoSCS) and a log power spectrum detected in the frequency domain of the electromyography signal (EMG). According to claim 17,
The voluntary/involuntary contraction detection unit includes an artificial intelligence model that distinguishes the voluntary muscle contraction signal and the involuntary muscle contraction signal from the characteristic vector. In a treatment method using an electrical stimulation treatment device that collects electromyography signals (EMG) generated in response to electrical stimulation from the body:
The electrical stimulation treatment device,
A voluntary/involuntary muscle contraction detection unit that extracts a characteristic vector from the frequency domain of the EMG signal and detects a voluntary muscle contraction signal and an involuntary muscle contraction signal from the extracted characteristic vector by applying an artificial intelligence model;
an involuntary muscle contraction signal removal unit that removes the involuntary muscle contraction signal from the EMG signal according to the detection result;
a muscle activity intensity calculator that calculates a root mean square (RMS) value of the EMG signal from which the involuntary muscle contraction signal is removed; and
A functional electrical stimulation control unit that compares the effective value and the threshold value and generates the functional electrical stimulation signal to be applied to the body according to the comparison result,
The treatment method of the electrical stimulation treatment device is,
extracting a feature vector from the frequency domain of the electromyogram signal;
A step of detecting and distinguishing between voluntary muscle contraction signals and involuntary muscle contraction signals from the characteristic vector extracted by applying an artificial intelligence model;
removing the involuntary muscle contraction signal from the electromyography signal according to the detection result;
Calculating the root mean square (RMS) of the EMG signal from which the involuntary muscle contraction signal has been removed; and
A treatment method comprising comparing the effective value and the threshold value and generating a functional electrical stimulation signal to be applied to the body according to the comparison result.

Images