An electrical stimulation pulse signal selection device, equipment and medium

Abstract

This application discloses an electrical stimulation pulse signal selection device, apparatus, and medium, applicable to the field of medical rehabilitation. The device includes: an acquisition module for acquiring electromyographic (EMG) signals from electrodes on the healthy and affected sides; a screening module for feature screening of the EMG signals; a difference module for performing difference processing on the feature screening results corresponding to the healthy and affected sides; and a selection module for selecting an electrical stimulation pulse signal based on the difference processing result. By using the acquired EMG signals, feature screening is performed, and the feature screening results corresponding to the healthy and affected sides are processed by difference. Based on the difference processing result, a corresponding electrical stimulation pulse signal is selected to achieve treatment of the affected side. Selecting the electrical stimulation pulse signal after processing the actual measured data is more accurate than the traditional method of subjective pulse signal selection by physicians or patients, and avoids muscle over-fatigue, muscle spasms, etc.

Description

Technical Field

[0001] This application relates to the field of medical rehabilitation, and in particular to an electrical stimulation pulse signal selection device, equipment, and medium. Background Technology

[0002] With the development of medical technology, for some patients with muscle defects, such as hemiplegia, limb weakness, numbness, and pain, electrical stimulation is currently commonly used for treatment. This involves outputting pulse signals to stimulate the muscle groups to achieve rehabilitation. The selection of the pulse signal is generally based on the physician's subjective feeling or the patient's feedback.

[0003] Currently, the selection of pulse signals by physicians or patients largely depends on treatment experience. Pulse therapy is performed on patients without actual testing, which is considered blind. If a pulse signal that does not match the actual needs is selected, it can lead to muscle over-fatigue, muscle spasms, and other problems.

[0004] How to solve the problem of muscle group damage in patients caused by the current pulse regulation method being too subjective, not combined with actual situation and lacking data support is an urgent problem that needs to be solved by those skilled in the art. Summary of the Invention

[0005] The purpose of this application is to provide an electrical stimulation pulse signal selection device, equipment, and medium. By processing the actual measured data, the electrical stimulation pulse signal is selected, which is more accurate than the traditional method of selecting the pulse signal subjectively by the physician or patient. It will not cause muscle over-fatigue, muscle spasm, or other problems due to incorrect selection of electrical stimulation pulse signal.

[0006] To solve the above-mentioned technical problems, this application provides an electrical stimulation pulse signal selection device, comprising:

[0007] The acquisition module is used to acquire electromyographic signals from the healthy side electrode and the affected side electrode;

[0008] A filtering module is used to perform feature filtering on the electromyographic signals;

[0009] The difference module is used to perform difference processing on the feature screening results corresponding to the healthy side electrode and the affected side electrode;

[0010] The selection module is used to select an electrical stimulation pulse signal based on the result of the difference processing.

[0011] Preferably, the filtering module is specifically used for:

[0012] The usable electromyographic signals from the healthy side and the affected side were selected from the electromyographic signals.

[0013] The electromyographic signals from the healthy side and the affected side with the largest root mean square effective value are selected from the available electromyographic signals from both sides.

[0014] Preferably, the screening module filters out usable electromyographic signals from the healthy side and the affected side using a first feature algorithm, which includes:

[0015]

[0016]

[0017] Where x is the current position of the electromyographic signal, 1 indicates selecting the current electromyographic signal, 0 indicates discarding the current electromyographic signal, S(x) is the activation function, rand is a random number between 0 and 1, otherwise is a random number other than 0 to 1, d is the dimension of the search space, t is the number of iterations, and x d (t+1) represents the location of the electromyographic signal when the search space dimension is d and the number of iterations is t+1, x1 d (t) represents the position of electromyographic signal x1 when the search space dimension is d and the number of iterations is t, and x2 is the position of the electromyographic signal x1. d (t) represents the position of electromyographic signal x2 when the search space dimension is d and the number of iterations is t, and x3 is the position of the electromyographic signal x2. d (t) represents the position of the electromyographic signal x3 when the search space dimension is d and the number of iterations is t.

[0018] Preferably, the filtering module filters out the healthy-side electromyography (EMG) signal and the affected-side EMG signal with the largest root mean square effective value from the available healthy-side and affected-side EMG signals using a second feature algorithm. The second feature algorithm includes:

[0019]

[0020] Where L is the root mean square effective value (RMS) maximum value, and d is the dimension of the search space. d X represents the maximum effective root mean square value corresponding to a search dimension of d. L X is the electromyographic signal corresponding to the current maximum root mean square effective value. L d For the electromyographic signal corresponding to the maximum effective value of the current root mean square when the dimension of the search space is d, rand(0,1) is a random number between 0 and 1, R is a random number with an activity radius between 0.9 and 0, and r3 is a random number distributed between 0 and 1.

[0021] Preferably, the difference module is specifically used for:

[0022] The absolute value of the difference between the electromyographic signal on the healthy side with the electromyographic signal on the affected side that has the largest root mean square effective value is taken.

[0023] Preferably, the selection module is specifically used for:

[0024] The absolute value is input into the Hammerstein model, which consists of a static nonlinear element and a dynamic linear element connected in series, to obtain the output value.

[0025] The corresponding electrical stimulation pulse signal is selected based on the output value.

[0026] Preferably, it further includes:

[0027] The control module is used to control the display system to display the selection result of the electrical stimulation pulse signal.

[0028] To address the aforementioned technical problems, this application also provides an electrical stimulation pulse signal selection device, including a memory for storing a computer program;

[0029] A processor, when executing the computer program, performs the following steps:

[0030] Acquire electromyographic signals from the healthy and affected side electrodes;

[0031] Feature screening is performed on the electromyographic signals;

[0032] The difference between the feature screening results corresponding to the healthy side electrode and the affected side electrode is processed.

[0033] The electrical stimulation pulse signal is selected based on the result of the difference processing.

[0034] To address the aforementioned technical problems, this application also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, performs the following steps:

[0035] Acquire electromyographic signals from the healthy and affected side electrodes;

[0036] Feature screening is performed on the electromyographic signals;

[0037] The difference between the feature screening results corresponding to the healthy side electrode and the affected side electrode is processed.

[0038] The electrical stimulation pulse signal is selected based on the result of the difference processing.

[0039] The electrical stimulation pulse signal selection device provided in this application includes: an acquisition module for acquiring electromyographic signals from electrodes on the healthy and affected sides; a screening module for feature screening of the electromyographic signals; a difference module for performing difference processing on the feature screening results corresponding to the healthy and affected sides; and a selection module for selecting an electrical stimulation pulse signal based on the difference processing result. In this application, the acquired electromyographic signals are feature-screened, and the feature screening results corresponding to the healthy and affected sides are difference-processed. Based on the difference processing result, a corresponding electrical stimulation pulse signal is selected to treat the affected side. Selecting the electrical stimulation pulse signal after processing the actual measured data is more accurate than the traditional method of subjective selection by a physician or patient, and avoids situations such as muscle over-fatigue and muscle spasms caused by incorrect selection of the electrical stimulation pulse signal. Attached Figure Description

[0040] To more clearly illustrate the embodiments of this application, the accompanying drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0041] Figure 1 A structural diagram of the electrical stimulation pulse signal selection device provided in the embodiments of this application;

[0042] Figure 2 This is a structural diagram of the electromyography signal acquisition device provided in the embodiments of this application;

[0043] Figure 3 This is a structural diagram of the electrical stimulation parameter model provided in the embodiments of this application;

[0044] Figure 4 This is a structural diagram of the electrical stimulation pulse signal selection process provided in the embodiments of this application;

[0045] Figure 5 This is an overall structural diagram of the electrical stimulation pulse signal selection device provided in the embodiments of this application;

[0046] Figure 6 The flowchart of the gray wolf optimization algorithm provided in the embodiments of this application is shown.

[0047] Figure 7 This is a structural diagram of an electrical stimulation pulse signal selection device provided in another embodiment of this application. Detailed Implementation

[0048] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the protection scope of this application.

[0049] The core of this application is to provide an electrical stimulation pulse signal selection device that selects electrical stimulation pulse signals by processing the actual measured data. Compared with the traditional method of selecting pulse signals subjectively by doctors or patients, this method is more accurate and will not cause muscle over-fatigue, muscle spasms, or other problems due to incorrect selection of electrical stimulation pulse signals.

[0050] The acquisition, filtering, and selection operations in the electrical stimulation pulse signal selection device provided in this application can be implemented by a controller in a host computer. For example, the controller can be a Field Programmable Gate Array (FPGA) development board, whose parallel signal processing speed meets the design requirements. The microcontroller analyzes and processes the acquired data and generates the drive signal for the electrical stimulation pulse. Of course, it can also be implemented by other controllers besides FPGA, which is not limited in this application.

[0051] To enable those skilled in the art to better understand the present application, the present application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0052] Figure 1 A structural diagram of the electrical stimulation pulse signal selection device provided in the embodiments of this application is shown below. Figure 1 As shown, the device includes:

[0053] The acquisition module 10 is used to acquire electromyographic signals from the healthy side electrode and the affected side electrode.

[0054] Specifically, since most current functional electrical stimulation training systems do not consider the reference standard of electromyographic signals on the healthy and affected sides, the feedback after stimulation on the healthy and affected sides is usually different. Due to the difference in muscle activity, the healthy side generally reacts faster and more intensely than the affected side, exhibiting biased differential signals. In this embodiment, electromyographic signals are obtained by stimulating the healthy and affected sides in a natural synergistic mode under comparison, thus distinguishing the healthy and affected sides and facilitating subsequent difference processing. For example, the healthy and affected sides are guided to perform the same action simultaneously, such as clenching a fist, and the electromyographic signals are collected when both arms perform the same fist-clenching action. Figure 2 This is a structural diagram of the electromyography signal acquisition device provided in the embodiments of this application. Figure 2The number 1 in the diagram represents an electrode patch. One end of electrode patch 1 is connected to the electromyography (EMG) signal acquisition terminal, while the other end directly contacts the muscle being tested to acquire EMG signals. The EMG signal acquisition terminal utilizes an integrated 8-channel synchronous sampling data acquisition system (AD7606) to synchronously sample eight analog input channels. It also includes: an on-chip integrated input amplifier, overvoltage protection circuitry, a second-order analog anti-aliasing filter, an analog multiplexer, a digital filter, a 2.5V reference voltage source, a reference voltage buffer, and high-speed serial and parallel interfaces to complete the EMG signal acquisition. It should be noted that multiple electrode patches 1 can be used depending on the application requirements, but the two electrode patches 1 must be symmetrical. That is, one electrode patch 1 is placed on the healthy side of the patient's arm, and the other electrode patch 1 is placed on the affected side of the patient's arm, corresponding to the healthy side, forming a symmetrical relationship to ensure the accuracy of the acquired data.

[0055] As can be seen, in this implementation, the acquisition module 10 collects the electromyographic signals of the healthy and affected sides with bias differences through electrode patches, and the electromyographic signals of the healthy and affected sides are compared with each other to form a difference that facilitates subsequent analysis and processing.

[0056] The filtering module 11 is used to filter the electromyographic signals by features.

[0057] Specifically, in this embodiment, since the acquired electromyographic signals have numerous features, to help those skilled in the art better understand this embodiment, the following example is provided: For instance, the acquired electromyographic signals contain four features: A, B, C, and D. However, only feature A is relevant to subsequent difference processing and selection of electrical stimulation pulse signals. Therefore, a filtering scheme is set to filter out useless features, leaving only useful feature signals. If the feature data is not filtered, the original electromyographic signals contain a large amount of data, making them unsuitable for classification processing. Direct processing may lead to misleading results. It should be noted that the four features A, B, C, and D in this embodiment are only used to help those skilled in the art better understand this embodiment. In general, the number of electromyographic signals is enormous, and this does not mean that only these four features exist. This application does not impose any special limitations.

[0058] As can be seen, by using the filtering module 11, the data in the electromyographic signal that are related to the subsequent process are filtered for features, while useless data is filtered out, which increases the speed and accuracy of data processing and avoids misleading results.

[0059] The difference module 12 is used to perform difference processing on the feature screening results corresponding to the healthy side electrode and the affected side electrode.

[0060] Specifically, in this embodiment, since electromyography (EMG) signals from both the healthy and affected sides need to be collected, the EMG signals from the healthy and affected sides, after being filtered by the filtering module, are then processed in the same channel for difference analysis. To help those skilled in the art better understand this embodiment, an example is given below: For instance, if A is the filtered feature signal in the healthy side EMG signal and B is the filtered feature signal in the affected side EMG signal, then the difference between A and B needs to be calculated, and the absolute value is taken. This absolute value is then input to the FPGA of the host computer for processing.

[0061] As can be seen, since there is a bias difference in the electromyographic signals of the healthy and affected sides, the result after difference processing is input into the FPGA.

[0062] Selection module 13 is used to select an electrical stimulation pulse signal based on the result of the difference processing.

[0063] Specifically, in this embodiment, after receiving the difference result, the FPGA will compare the difference result with the preset electrical stimulation pulse signal and select the treatment plan that best suits the current situation.

[0064] As can be seen, the electrical stimulation pulse signal selection device provided in this embodiment includes: an acquisition module 10 for acquiring electromyographic signals from the healthy side electrode and the affected side electrode; a screening module 11 for performing feature screening on the electromyographic signals; a difference module 12 for performing difference processing on the feature screening results corresponding to the healthy side electrode and the affected side electrode; and a selection module 13 for selecting an electrical stimulation pulse signal based on the difference processing result. In this application, the acquired electromyographic signals are used to perform feature screening, and the feature screening results corresponding to the healthy and affected sides are processed by difference processing. Based on the difference processing result, a corresponding electrical stimulation pulse signal is selected to achieve treatment of the affected side. Selecting an electrical stimulation pulse signal after processing the actual measured data is more accurate than the traditional method of subjectively selecting a pulse signal by a physician or patient, and avoids situations such as muscle over-fatigue and muscle spasms caused by incorrect selection of the electrical stimulation pulse signal.

[0065] Based on the above embodiments, as a preferred embodiment, the screening module 11 is specifically used for:

[0066] Screen out usable electromyographic signals from the healthy side and the affected side;

[0067] The healthy and affected side electromyographic signals with the largest root mean square effective values ​​were selected from the available healthy and affected side electromyographic signals.

[0068] Specifically, in this embodiment, after obtaining the available electromyographic signals from the healthy and affected sides, it is necessary to select the electromyographic signals with the most obvious characteristics. By analyzing the time-domain characteristics of the electromyographic signals, the root mean square effective value (RMS) is used as the evaluation criterion. That is, the larger the root mean square effective value, the more obvious the characteristics of the electromyographic signal. Therefore, it is necessary to select the electromyographic signals with the most obvious characteristics, which means it is necessary to select the electromyographic signals with the largest root mean square effective value.

[0069] It is evident that after initial feature screening of the electromyographic signals from the healthy and affected sides, a second screening is required. By analyzing the temporal characteristics of the electromyographic signals and using the root mean square effective value (RMS) as the evaluation criterion, the electromyographic signals with the most obvious features are selected to lay the groundwork for subsequent FPGA analysis and prevent redundant data from interfering with the analysis results.

[0070] Based on the above embodiments, as a preferred embodiment, the screening module 11 filters out usable electromyographic signals from the healthy side and the affected side using a first feature algorithm. The first feature algorithm includes:

[0071]

[0072]

[0073] Where x is the current position of the electromyographic signal, 1 indicates selecting the current electromyographic signal, 0 indicates discarding the current electromyographic signal, S(x) is the activation function, rand is a random number between 0 and 1, otherwise is a random number other than 0 to 1, d is the dimension of the search space, t is the number of iterations, and x d (t+1) represents the location of the electromyographic signal when the search space dimension is d and the number of iterations is t+1, x1 d (t) represents the position of electromyographic signal x1 when the search space dimension is d and the number of iterations is t, and x2 is the position of the electromyographic signal x1. d (t) represents the position of electromyographic signal x2 when the search space dimension is d and the number of iterations is t, and x3 is the position of the electromyographic signal x2. d (t) represents the position of the electromyographic signal x3 when the search space dimension is d and the number of iterations is t.

[0074] Specifically, in this embodiment, considering that binary encoding has stronger search capabilities and is simpler to operate than real number encoding, the first feature algorithm is used for filtering. 1 represents selecting the current electromyographic signal and 0 represents discarding the current electromyographic signal. Since the feature selection problem is to select or not select each feature, each feature subset is encoded as a binary string of 1 and 0. Therefore, all solutions are represented in the form of binary vectors, with only 1 or 0 calculation results, which is more convenient for FPGA search and processing.

[0075]

[0076]

[0077] Where x is the current position of the electromyographic signal, 1 indicates selecting the current electromyographic signal, 0 indicates discarding the current electromyographic signal, S(x) is the activation function, rand is a random number between 0 and 1, otherwise is a random number other than 0 to 1, d is the dimension of the search space, t is the number of iterations, and x d (t+1) represents the location of the electromyographic signal when the search space dimension is d and the number of iterations is t+1, x1 d (t) represents the position of electromyographic signal x1 when the search space dimension is d and the number of iterations is t, and x2 is the position of the electromyographic signal x1. d (t) represents the position of electromyographic signal x2 when the search space dimension is d and the number of iterations is t, and x3 is the position of the electromyographic signal x2. d (t) represents the position of the electromyographic signal x3 when the search space dimension is d and the number of iterations is t. Where x1(t) = x α (t)-A1·D α x2(t)=x β (t)-A2·D β x3(t)=x δ (t)-A3·D δ Where x1(t) is the position of the electromyographic signal x1 at iteration number t, x2(t) is the position of the electromyographic signal x2 at iteration number t, and x3(t) is the position of the electromyographic signal x3 at iteration number t. α (t) represents the position of α at iteration number t, x β (t) represents the position of β at iteration number t, x δ (t) represents the position of δ at iteration number t, and A1, A2, and A3 are convergence factors used in the equation, where the general formula is A = 2y·r1 - y.

[0078] The parameter y decreases linearly from 2 to 0, and T max It is the maximum number of iterations, and D is a random number distributed between 0 and 1. α D β D δ To use the equation, where D α =|C1x α (t)-x(t)|、D β =|C2x β (t)-x(t)|、D δ =|C3x δ (t)-x(t)|;

[0079] C = 2r², where r² is a random number between 0 and 1, and C is a random number between 0 and 2. Since the electromyographic signal itself has the above parameters, the formulas above can be used to calculate whether the current electromyographic signal feature will be selected. 1 indicates that the current electromyographic signal is selected, and 0 indicates that the current electromyographic signal is discarded. The selected electromyographic signal will continue to be filtered a second time using the second feature algorithm.

[0080] As can be seen, since electromyographic (EMG) signals contain a large number of features, the first feature algorithm selects EMG signals from both the healthy and affected sides, retaining those that meet the criteria for a second screening. This makes the selection of EMG signals more accurate and faster.

[0081] Based on the above embodiments, as a preferred embodiment, the screening module 11 filters out the available healthy-side electromyography (EMG) signals and the affected-side EMG signals with the largest root mean square effective value using a second feature algorithm. The second feature algorithm includes:

[0082]

[0083] Where L is the root mean square effective value (RMS) maximum value, and d is the dimension of the search space. d X represents the maximum effective root mean square value corresponding to a search dimension of d. L X is the electromyographic signal corresponding to the current maximum root mean square effective value. L d Let r be the electromyographic signal corresponding to the maximum effective value of the current root mean square when the search space has a dimension of d. rand(0,1) is a random number between 0 and 1, R is a random number with an activity radius between 0.9 and 0, and r3 is a random number distributed between 0 and 1.

[0084] Specifically, this embodiment employs an optimized gray wolf algorithm. To help those skilled in the art better understand this optimization algorithm, an example is given below. For instance, during a wolf pack's pursuit of prey, the final position of an individual gray wolf will fall within a circular area of ​​radius R. This circular area is determined by the positions of the three best gray wolves in the pack. The position of the prey can be estimated using the positions of these three optimal wolves. Therefore, the remaining gray wolves will randomly update their positions near the prey. First, a group of gray wolves is randomly initialized. Then, the wolves' fitness values ​​are evaluated, and α, β, and δ wolves are defined. In each iteration, the wolf pack is randomly divided into N / 2 pairs, where N is the population size. Afterward, there is competition between each pair of wolves. Wolves that obtain better fitness values ​​in the competition are called winners, and they are directly transferred to the new population. In contrast, the losers in the competition update their positions by learning from the winners and leaders. In this example, the fitness value is the root mean square effective value (RMS). The RMS value is used to evaluate the characteristic values ​​of the electromyography (EMG) signal; the larger the RMS value, the more obvious the features that the FPGA can acquire. Therefore, in the second feature algorithm of this embodiment...

[0085]

[0086] Where t is the iteration number, the activity radius R is a random number between 0 and 0.9, and T max The maximum number of iterations is given in the following formula:

[0087]

[0088] Where L is the root mean square effective value (RMS) maximum value, and d is the dimension of the search space. d X represents the maximum effective root mean square value corresponding to a search dimension of d. L X is the electromyographic signal corresponding to the current maximum root mean square effective value. L dLet r be the electromyographic signal corresponding to the maximum effective root mean square value when the search space dimension is d. rand(0,1) generates random numbers between 0 and 1, R is a random number between 0.9 and 0, and r3 is a random number distributed between 0 and 1. For this formula, in the leader enhancement step, if the newly generated leader provides a better fitness value, the current leader will be replaced; otherwise, the current leader will be retained for the next iteration. This algorithm is repeated until the maximum number of iterations is reached, and finally, the globally optimal α is known to be the optimal solution (the optimal feature subset). In this embodiment, this means that in the second screening, only the optimal features are retained. The optimal feature from the healthy side, i.e., the feature with the largest root mean square value, is retained, and the optimal feature from the affected side, i.e., the feature with the corresponding maximum root mean square value, is also retained. First, input the population size N and the maximum number of iterations T. max Initialize the wolf pack and parameters y, A, and C, and calculate the wolf pack's adaptability value F(x); define x α The wolf with the highest fitness in the population, x β The wolf, with the second highest fitness, is X. δ For the wolf with the third highest fitness, the number of iterations is defined as t = 1 to the maximum value T. max ,pass

[0089]

[0090] Calculate the activity radius. For competition between each wolf, i.e., competition between each feature, start with a quantity of 1 and increase to half the population size (N / 2). Randomly select two wolves x. k and x m Compare the fitness values ​​F(x) between the two. k ) and F(x m The larger one will be the winner (x). w The smaller one is considered the loser x1, if the fitness comparison result x k Greater than x m Then x w =x k , x1 = x m Conversely, x w =x1, x1 = x k The winner x w Add to the new population and remove two wolves from the total population. k and x m At this point, the position of the loser x1 needs to be updated. For the population starting from 1 and reaching half the population size N / 2, calculate x1(t) = x α (t)-A1·D α x2(t)=x β (t)-A2·D βx3(t)=x δ (t)-A3·D δ ,pass

[0091]

[0092] Re-update the current x i The location of x needs to be reassessed. i The corresponding fitness value F(x) i ), x i A new population is added, and the wolf pack N with the highest fitness is retained from the merged population, where i can take the values ​​a, β, or δ, and x... α The wolf with the highest fitness in the population, x β The wolf, with the second highest fitness, is X. δ The wolf with the third highest fitness is selected. Update a, β, δ, and parameters y, A, and C. At this point, a leader enhancement strategy is adopted: for the number of leaders j = 1 to 3, through…

[0093]

[0094] Set the position of a to x a =X L (j=1) thus generating a new leader L j At this point, the new leader L is calculated. j The corresponding fitness value F(L) j ), X L Replace with x β (j=2) or x δ (j=3), according to the new leader L j Update a, β, and δ. At this point, a is used as the globally optimal output and is fed into the subsequent root mean square effective value algorithm for feature comparison processing.

[0095] It is evident that by setting the second feature algorithm, the optimal features are extracted from the available electromyographic signals, providing the optimal input for the root mean square effective value algorithm, and making feature comparison processing more convenient.

[0096] Based on the above embodiments, as a preferred embodiment, the difference module 12 is specifically used for:

[0097] The absolute value of the difference between the electromyographic signal on the healthy side with the electromyographic signal on the affected side, which has the largest root mean square effective value, is taken.

[0098] Specifically, in this implementation, the root mean square effective value is used as the standard for judging whether the electromyographic signal characteristics are obvious. The one with the most obvious characteristics, that is, the one with the largest root mean square effective value, is selected. Since there are two channels, the difference between the two largest root mean square effective values ​​in the channels on the healthy and affected sides is taken and the absolute value is obtained to complete the difference processing.

[0099] It is evident that after obtaining the electromyographic signal with the largest root mean square effective value from both channels, the absolute value of the difference between the two signals is taken due to their bias difference, thus laying the groundwork for the selection of subsequent electrical stimulation signals.

[0100] Based on the above embodiments, as a preferred embodiment, the selection module 13 is specifically used for:

[0101] The absolute value is input into the Hammerstein model, which consists of a static nonlinear element and a dynamic linear element connected in series, to obtain the output value.

[0102] Select the corresponding electrical stimulation pulse signal based on the output value.

[0103] Specifically, Figure 3 This is a structural diagram of the electrical stimulation parameter model provided in the embodiments of this application, such as... Figure 3 As shown, Z1 to Z N That is, the input quantity, which is obtained from X1 to X through a nonlinear element. N The output parameter y is obtained through a linear model. Feature comparison information, as a factor, first passes through the nonlinear stage of the Hammerstein model, and then the result is used as an input variable, substituted into a multi-input single-output model. This forms a Hammerstein model composed of a static nonlinear stage and a dynamic linear stage connected in series. The resulting output value is the output parameter of the electrostimulator. The FPGA can then select the appropriate electrostimulation pulse signal based on the corresponding parameters to complete the treatment of the patient. In addition, after selecting the electrical stimulation pulse signal, the pulse electrical stimulation unit completes the operation. This part consists of three parts: an FPGA development board, an AD9708 (8-bit, 125MSPS maximum conversion speed digital-to-analog converter), and a LabVIEW virtual instrument engineering platform host computer software. The graded results of the difference in electromyographic signals between the healthy and affected sides obtained after the data processing unit control the pulse stimulation output. The FPGA uses a read-only memory address (ROMIP) core to store four waveforms. After the FPGA processes and reads the waveform data in the read-only memory (ROM), it can output four modes: sine wave, square wave, triangle wave, and sawtooth wave, with frequencies ranging from 1kHz to 200kHz and different amplitude waveforms. The read data is then sent to the digital-to-analog converter, and after analog-to-digital conversion and low-pass filtering, the corresponding stimulation pulse is output and the corresponding output information is displayed in real time on the LabVIEW host computer software. The host computer can also select the corresponding treatment mode, which is convenient for patients or medical staff to analyze and refer to and independently select the mode output.

[0104] As can be seen, by inputting the absolute value obtained in the above embodiments into the corresponding model, the output value that can select the corresponding electrical stimulation pulse signal is obtained, and the electrical stimulation pulse signal is selected more accurately.

[0105] Figure 4 A structural diagram of the electrical stimulation pulse signal selection process provided in the embodiments of this application is shown below. Figure 4 As shown, for the electrodes on the healthy and affected sides included in the electromyography (EMG) electrodes, the corresponding data is collected by the surface EMG at the hardware platform, connected to the software platform via Bluetooth communication, and the above data is processed and then processed by the model to obtain the treatment plan.

[0106] Figure 5 This is an overall structural diagram of the electrical stimulation pulse signal selection device provided in the embodiments of this application, as shown below. Figure 5 As shown, the power supply module 14 supplies power to the electromyography (EMG) signal data acquisition unit 15, the Bluetooth communication unit 16, the host computer unit 17, the FPGA control unit 18, and the electrical stimulation output unit 19. The EMG signal data acquisition unit 15 includes distributed surface EMG electrodes 151 and a surface EMG acquisition module 152. The surface EMG electrodes 151 include healthy side EMG electrode 1, healthy side EMG electrode 2, affected side EMG electrode 1, and affected side EMG electrode 2. The surface EMG electrodes 151 and the surface EMG acquisition module 152 are connected to transmit the acquired EMG signals. The acquired EMG signals are transmitted to the host computer unit 17 via the Bluetooth communication unit 16. The data processing unit 171 in the host computer unit 17 processes the data and sends it to the model relationship unit 172 for model relationship processing. Finally, it sends the data to the treatment mode selection unit 173 to select the treatment mode. After the treatment mode is selected, the data is sent to the FPGA control unit 18, which controls the electrical stimulation output unit 19 to output electrical stimulation signals.

[0107] Figure 6 The flowchart of the gray wolf optimization algorithm provided in the embodiments of this application is as follows: Figure 6 As shown,

[0108] S10: Begin.

[0109] S11: Single-channel data composed of electromyographic signals from the healthy and affected sides.

[0110] Because there is a bias difference in the electromyographic signals acquired from the healthy and affected sides, the electromyographic signals from the healthy and affected sides are separated and combined into a single channel for easy analysis.

[0111] S12: Feature extraction.

[0112] Feature extraction is performed using the aforementioned first feature algorithm to extract usable electromyographic signals.

[0113] S13: Original dataset.

[0114] The available electromyographic signals are used as the raw dataset to prepare the second feature algorithm.

[0115] S14: Perform feature selection.

[0116] The available electromyographic signals are filtered using the aforementioned second feature algorithm.

[0117] S15: Optimal feature subset.

[0118] The root mean square effective value (RMS) was used as the evaluation criterion, and the RMS value was taken as the optimal characteristic of the electromyographic signal.

[0119] S16: Classification by grade.

[0120] The electromyographic signals of the optimal feature are distinguished from the remaining electromyographic signals to ensure the independence of the optimal feature.

[0121] S17: End.

[0122] Based on the above embodiments, as a preferred embodiment, it further includes:

[0123] The control module is used to control the display system to display the selection results of the electrical stimulation pulse signal.

[0124] Specifically, in this example, a control module was added to enable physicians or patients to better understand the selection results of the electrical stimulation pulse signal. The control module will display the selected electrical stimulation pulse signal in the system.

[0125] Figure 7 A structural diagram of an electrical stimulation pulse signal selection device provided in another embodiment of this application is shown below. Figure 7 As shown, the electrical stimulation pulse signal selection device includes: a memory 20 for storing computer programs;

[0126] Processor 21 is configured to perform the following steps when executing a computer program:

[0127] Acquire electromyographic signals from the healthy and affected side electrodes;

[0128] Feature screening of electromyographic signals;

[0129] The difference between the feature screening results corresponding to the healthy side electrode and the affected side electrode was processed.

[0130] The electrical stimulation pulse signal is selected based on the result of the difference processing.

[0131] The electrical stimulation pulse signal selection device provided in this embodiment may include, but is not limited to, electrical stimulation physiotherapy machines.

[0132] The processor 21 may include one or more processing cores, such as a quad-core processor or an octa-core processor. The processor 21 may be implemented using at least one of the following hardware forms: Digital Signal Processor (DSP), Field-Programmable Gate Array (FPGA), or Programmable Logic Array (PLA). The processor 21 may also include a main processor and a coprocessor. The main processor, also known as the Central Processing Unit (CPU), is used to process data in the wake-up state; the coprocessor is a low-power processor used to process data in the standby state. In some embodiments, the processor 21 may integrate a Graphics Processing Unit (GPU), which is responsible for rendering and drawing the content to be displayed on the screen. In some embodiments, the processor 21 may also include an Artificial Intelligence (AI) processor, which handles computational operations related to machine learning.

[0133] The memory 20 may include one or more computer-readable storage media, which may be non-transitory. The memory 20 may also include high-speed random access memory and non-volatile memory, such as one or more disk storage devices or flash memory devices. In this embodiment, the memory 20 is used to store at least the following computer program 201, which, after being loaded and executed by the processor 21, can perform the aforementioned related steps. In addition, the resources stored in the memory 20 may also include an operating system 202 and data 203, etc., and the storage method may be temporary storage or permanent storage. The operating system 202 may include Windows, Unix, Linux, etc.

[0134] In some embodiments, the electrical stimulation pulse signal selection device may further include a display screen 22, an input / output interface 23, a communication interface 24, a power supply 25, and a communication bus 26.

[0135] Those skilled in the art will understand that Figure 7 The structure shown does not constitute a limitation on the electrical stimulation pulse signal selection device and may include more or fewer components than shown.

[0136] Finally, this application also provides an embodiment corresponding to a computer-readable storage medium. The computer-readable storage medium stores a computer program, which, when executed by a processor, performs the following steps:

[0137] Acquire electromyographic signals from the healthy and affected side electrodes;

[0138] Feature screening of electromyographic signals;

[0139] The difference between the feature screening results corresponding to the healthy side electrode and the affected side electrode was processed.

[0140] The electrical stimulation pulse signal is selected based on the result of the difference processing.

[0141] It is understood that if the methods in the above embodiments are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and executes all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0142] The above provides a detailed description of the electrical stimulation pulse signal selection device, apparatus, and medium provided in this application. The various embodiments in the specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the devices disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the descriptions are relatively simple; relevant parts can be referred to in the method section. It should be noted that those skilled in the art can make various improvements and modifications to this application without departing from the principles of this application, and these improvements and modifications also fall within the protection scope of the claims of this application.

[0143] It should also be noted that, in this specification, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

Claims

1. An electrical stimulation pulse signal selection device, characterized by, The method comprises the following steps: obtaining electromyographic signals of a healthy side electrode and a diseased side electrode; screening the electromyographic signals; differencing the results of the screening of the healthy side electrode and the diseased side electrode; selecting an electrical stimulation pulse signal according to the results of the differencing; wherein the screening of the electromyographic signals comprises: screening usable healthy side electromyographic signals and usable diseased side electromyographic signals from the electromyographic signals, and screening the healthy side electromyographic signal and the diseased side electromyographic signal with the largest root mean square effective value from the usable healthy side electromyographic signals and the usable diseased side electromyographic signals; wherein the screening of the usable healthy side electromyographic signals and the usable diseased side electromyographic signals from the electromyographic signals, and the screening of the healthy side electromyographic signal and the diseased side electromyographic signal with the largest root mean square effective value from the usable healthy side electromyographic signals and the usable diseased side electromyographic signals comprises: Adopt the optimized grey wolf algorithm, take the wolf as the myoelectric signal feature, take the fitness value as the root mean square effective value, input the population number N and the maximum iteration number , initialize the wolf group and parameters 、 and , calculate the fitness value of the wolf group ; By Computing the active radius, starting from 1, to half the population size N / 2, randomly select two wolves and ; compare the fitness values between them and , the larger one is the winner , the smaller one is the loser If the fitness comparison result is greater , , otherwise , , add the winner to the new population and remove two wolves from the population and ; Update failures The position is calculated from a population size of 1 to half the population size N / 2. , , ;pass Re-update the current Reassess the location. Corresponding fitness value ,Will When a new population is added, the wolf pack N with the highest fitness is retained from the merged population. Desirable , , , The wolf is the most fit species in the population. The wolf has the second highest fitness level. It is the third most adaptable wolf; wherein ; is the position of the EMG signal at iteration is the position of the EMG signal at iteration is the position of the EMG signal at iteration is the position of the EMG signal at iteration is the position of the EMG signal at iteration is the position of the EMG signal at iteration is the position of the EMG signal at iteration is the position of the EMG signal at iteration is the position of the EMG signal at iteration , , is the convergence factor, calculated as , with parameters linearly decreasing from 2 to 0; , , , , is a random number distributed between 0 and 1;​​​​​​​​​ is the current position of the myoelectric signal, 1 indicates that the current myoelectric signal is selected, and 0 indicates that the current myoelectric signal is discarded, is an activation function, is a random number between 0 and 1, is a random number other than between 0 and 1, is the dimension of the search space, is the number of iterations, is the position of the myoelectric signal when the dimension of the search space is and the number of iterations is , is the position of the myoelectric signal when the dimension of the search space is and the number of iterations is , is the position of the myoelectric signal when the dimension of the search space is and the number of iterations is , is the position of the myoelectric signal when the dimension of the search space is and the number of iterations is , is the position of the myoelectric signal when the dimension of the search space is and the number of iterations is . A leader enhancement strategy is adopted, which is used when the number of leaders j = 1 to 3. ,set up Location (j=1) This generates a new leader. At this point, the new leader is calculated. The corresponding fitness value F ( ),Will Replace with (j=2) or (j=3), according to the new leader renew , , and parameters , and ; The as a global best output to a subsequent root-mean-square effective value algorithm; in, It is the maximum value of the root mean square effective value. For the dimensions of the search space, To search in unprecedented dimensions The corresponding root mean square effective value is the maximum value. The electromyographic signal corresponding to the current maximum root mean square effective value. For the dimension of the search space The electromyographic signal corresponding to the maximum value of the current root mean square effective value, in To generate random numbers between 0 and 1, A random number with an activity radius between 0.9 and 0. It is a random number distributed between 0 and 1.

2. The electrical stimulation pulse signal selection device of claim 1, wherein, the differencing module is specifically configured to: take the absolute value after differencing the healthy side electromyographic signal and the diseased side electromyographic signal with the largest root mean square effective value.

3. The electrical stimulation pulse signal selection apparatus of claim 2, wherein, The selection module is specifically configured to: input the absolute value into a Hammerstein model composed of a static nonlinear element and a dynamic linear element to obtain an output value; selecting the electrical stimulation pulse signal corresponding to the output value.

4. The electrical stimulation pulse signal selection apparatus of claim 3, wherein, Further comprising: a control module for controlling a display system to display the selection result of the electrical stimulation pulse signal.

5. An electrical stimulation pulse signal selection device, characterized by The method comprises the following steps: obtaining electromyographic signals of a healthy side electrode and a diseased side electrode; screening the electromyographic signals; differencing the results of the screening of the healthy side electrode and the diseased side electrode; selecting an electrical stimulation pulse signal according to the results of the differencing; wherein the screening of the electromyographic signals comprises: screening usable healthy side electromyographic signals and usable diseased side electromyographic signals from the electromyographic signals, and screening the healthy side electromyographic signal and the diseased side electromyographic signal with the largest root mean square effective value from the usable healthy side electromyographic signals and the usable diseased side electromyographic signals; wherein the screening of the usable healthy side electromyographic signals and the usable diseased side electromyographic signals from the electromyographic signals, and the screening of the healthy side electromyographic signal and the diseased side electromyographic signal with the largest root mean square effective value from the usable healthy side electromyographic signals and the usable diseased side electromyographic signals comprises: The computer readable storage medium stores a computer program, and the computer program is executed by the processor to realize the following steps: Adopt the optimized grey wolf algorithm, take the wolf as the myoelectric signal feature, take the fitness value as the root mean square effective value, input the population number N and the maximum iteration number , initialize the wolf group and parameters , and , calculate the fitness value of the wolf group ; By Computing the active radius, starting from 1, to half the population size N / 2, randomly select two wolves And ; Compare the fitness values between the two And The greater is the winner The smaller is the loser If the fitness comparison result Is greater than Then , Otherwise , Add the winner To the new population and remove two wolves from the population And ; Update failures The position is calculated from a population size of 1 to half the population size N / 2. , , ;pass Re-update the current Reassess the location. Corresponding fitness value ,Will When a new population is added, the wolf pack N with the highest fitness is retained from the merged population. Desirable , , , The wolf is the most fit species in the population. The wolf has the second highest fitness level. It is the third most adaptable wolf; wherein ; is the position of the EMG signal at iteration is the position of the EMG signal at iteration is the position of the EMG signal at iteration is the position of the EMG signal at iteration is a convergence factor, calculated using , parameters linearly decreasing from 2 to 0; , , , , is a random number distributed between 0 and 1;​​​​​​​​​​​​​​​​ is the current position of the myoelectric signal, 1 indicates that the current myoelectric signal is selected, and 0 indicates that the current myoelectric signal is discarded, is an activation function, is a random number between 0 and 1, is a random number other than between 0 and 1, is the dimension of the search space, is the number of iterations, is the position of the myoelectric signal when the dimension of the search space is and the number of iterations is , is the position of the myoelectric signal when the dimension of the search space is and the number of iterations is , is the position of the myoelectric signal when the dimension of the search space is and the number of iterations is , is the position of the myoelectric signal when the dimension of the search space is and the number of iterations is , is the position of the myoelectric signal when the dimension of the search space is and the number of iterations is . Taking the leader enhancement strategy, for the leader number j = 1 to 3, through , the position of (j = 1) is set from which a new leader is generated , at this time the fitness value F( ) corresponding to the new leader is calculated , replaces (j = 2) or (j = 3), and the parameters , and are updated according to the new leader . , , and . will be as a global best output into a subsequent root-mean-square effective value algorithm; wherein, is the maximum root mean square effective value, is the dimension of the search space, is the maximum root mean square effective value corresponding to the dimension of the search space, is the electromyographic signal corresponding to the current maximum root mean square effective value, is the electromyographic signal corresponding to the current maximum root mean square effective value when the dimension of the search space is is the electromyographic signal corresponding to the current maximum root mean square effective value when the dimension of the search space is is the electromyographic signal corresponding to the current maximum root mean square effective value when the dimension of the search space is is a random number between 0 and 1, is a random number between 0.9 and 0, is a random number distributed between 0 and 1.

6. A computer-readable storage medium, characterized in that, obtaining electromyographic signals of a healthy side electrode and a diseased side electrode; screening the electromyographic signals; differencing the results of the screening of the healthy side electrode and the diseased side electrode; selecting an electrical stimulation pulse signal according to the results of the differencing; wherein the screening of the electromyographic signals comprises: screening usable healthy side electromyographic signals and usable diseased side electromyographic signals from the electromyographic signals, and screening the healthy side electromyographic signal and the diseased side electromyographic signal with the largest root mean square effective value from the usable healthy side electromyographic signals and the usable diseased side electromyographic signals; ​ The method comprises the following steps: screening available healthy-side electromyography signals and affected-side electromyography signals from the electromyography signals, and screening the healthy-side electromyography signal and the affected-side electromyography signal with the maximum root mean square effective value from the available healthy-side electromyography signals and affected-side electromyography signals, comprising the steps of: The method comprises the following steps: screening available healthy-side electromyography signals and affected-side electromyography signals from the electromyography signals, and screening the healthy-side electromyography signal and the affected-side electromyography signal with the maximum root mean square effective value from the available healthy-side electromyography signals and affected-side electromyography signals, comprising the steps of: The method comprises the following steps The optimized grey wolf algorithm is adopted, a wolf represents an electromyographic signal feature, a fitness value is an effective root mean square value, population quantity N and maximum iteration number are input , a wolf group and parameters are initialized , and , and the fitness value of the wolf group is calculated ; By computing the active radius, starting from 1, to half the population size N / 2, randomly select two wolves and ; compare the fitness values between the two and , the larger one is the winner , the smaller one is the loser if the fitness comparison result is greater than , then , , otherwise , add the winner to the new population and remove two wolves from the population and ; the loser of the update the position of the loser of the update, for a number starting at 1, to half the population size N / 2 , , by renewing the position of the current loser of the update, reevaluating the corresponding fitness value , adding the new population from the merged population, keeping the N fittest wolves from the merged population, where the , , , fittest wolf in the population, the second fittest wolf, the third fittest wolf wherein ; is the position of the myoelectric signal at iteration is the position of the myoelectric signal at iteration is the position of the myoelectric signal at iteration is the position of the myoelectric signal at iteration is the position of the myoelectric signal at iteration is the position of the myoelectric signal at iteration is the position of the myoelectric signal at iteration is the position of the myoelectric signal at iteration is the position of the myoelectric signal at iteration is the position of the myoelectric signal at iteration is the position of the myoelectric signal at iteration is the position of the myoelectric signal at iteration is the position of the myoelectric signal at iteration is the position of the myoelectric signal at iteration is the position of the myoelectric signal at iteration is the position of the myoelectric signal at iteration is the position of the myoelectric signal at iteration is the position of the myoelectric signal at iteration , , is a convergence factor, calculated using , parameters linearly decrease from 2 to 0; , , , , is a random number distributed between 0 and 1; This represents the current position of the electromyographic (EMG) signal. 1 indicates selecting the current EMG signal, and 0 indicates discarding the current EMG signal. For activation function, A random number between 0 and 1. For random numbers other than 0 and 1, For the dimensions of the search space, For the number of iterations, In the search space dimension The number of iterations is Location of electromyographic signals at that time For the search space dimension is The number of iterations is electromyography signals Location, For the search space dimension is The number of iterations is electromyography signals Location, For the search space dimension is The number of iterations is electromyography signals Location; Taking the leader enhancement strategy, for the leader number j = 1 to 3, through , setting The position (j = 1) of the new leader , at this time, the fitness value F( ) corresponding to the new leader , replace With (j = 2) or (j = 3), update , , , And parameters , And According to the new leader ; will be as a global best output into a subsequent root-mean-square effective value algorithm; wherein, is the maximum root mean square effective value, is the dimension of the search space, is the maximum root mean square effective value corresponding to the dimension of the search space, is the electromyographic signal corresponding to the current maximum root mean square effective value, is the electromyographic signal corresponding to the current maximum root mean square effective value when the dimension of the search space is is the electromyographic signal corresponding to the current maximum root mean square effective value when the dimension of the search space is is the electromyographic signal corresponding to the current maximum root mean square effective value when the dimension of the search space is is a random number between 0 and 1, is a random number between 0.9 and 0, is a random number distributed between 0 and 1.

Images