Balancing table detection method and device combined with myoelectric sensing

Abstract

The invention provides a balancing stand detection method and device combined with myoelectricity sensing, and the method comprises the steps: applying platform disturbance in a preset direction, and synchronously collecting a myoelectricity signal and a platform motion parameter of a target muscle group of a user; and extracting the response duration and the muscle amplitude index of the electromyographic signal, and generating muscle group response characteristics. And outputting a neuromuscular function evaluation result based on the mapping relationship between the direction of the disturbance applied by the platform and the muscle group response characteristics. Wherein the response characteristics comprise that the reaction duration is used for quantifying the nerve conduction velocity. The muscle amplitude is used for quantifying the muscle power intensity. The bilateral muscle group response difference of the target muscle group is used to identify the collaborative imbalance. Whether the user muscle group is in an unbalanced state or not is detected through the method.

Description

Technical Field

[0001] This application relates to the field of intelligent machinery, and more particularly to a method and apparatus for detecting a balance platform that incorporates electromyography (EMG) sensing. Background Technology

[0002] The dynamic balance function assessment and training equipment is mainly used for the quantitative analysis of neuromuscular function in patients with vestibular dysfunction, post-stroke balance disorders, and sports injuries in the field of medical rehabilitation.

[0003] Current technology uses a three-dimensional force table to measure changes in plantar pressure during swaying, such as the trajectory of the center of gravity, and assesses balance ability by creating sudden disturbances through a moving platform. However, it has limitations: it can only observe the final swaying result and cannot detect the dynamic process of nerve control over muscles. Furthermore, it cannot distinguish whether balance disorders stem from slow nerve conduction, such as delayed signals from spinal cord injury, or from weak muscle strength, such as muscle atrophy. It also cannot detect compensatory behaviors like gripping handrails that mask true muscle weakness, and it is difficult to quantify muscle coordination imbalances, such as the disproportionate force distribution in the anteroposterior muscles of the lower leg in patients with vestibular disorders.

[0004] Therefore, there is a need for a balance platform detection method and device that combines electromyography to detect abnormal states in balance disorder. Summary of the Invention

[0005] In view of this, it is necessary to provide a balance platform detection method and device that combines electromyography to detect abnormal states in a state of balance impairment, in order to solve the above problems.

[0006] The above-mentioned balance table detection method and device combining electromyography (EMG) sensing integrates surface EMG sensing and synchronous triggering circuit in the balance table device. Under the time reference T0 of the platform disturbance, the reaction time and amplitude of the target muscle group are aligned and calculated. A mapping evaluation model is constructed to generate a one-to-one correspondence of the disturbance direction to the dominant muscle or antagonistic muscle response characteristics, thereby realizing the comprehensive quantification of nerve conduction velocity, muscle strength and muscle group coordination.

[0007] Furthermore, the system can simultaneously collect platform motion and multi-channel EMG data under a unified time reference, extract response features composed of reaction time and muscle amplitude, combine pattern recognition of dominant or antagonistic muscles and bilateral homologous muscles, output assessment results of nerve conduction efficiency, muscle strength and whether synergy is unbalanced, and further provide rehabilitation training suggestions. Attached Figure Description

[0008] Figure 1 This is a flowchart of the balance platform detection method combined with electromyography as described in this application. Detailed Implementation

[0009] The embodiments of this application will now be described with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments.

[0010] It should be noted that when a component is considered to be "connected" to another component, it can be directly connected to the other component or may also have an intervening component. When a component is considered to be "placed" on another component, it can be directly placed on the other component or may also have an intervening component. The terms "top," "bottom," "upper," "lower," "left," "right," "front," "back," and similar expressions used in this article are for illustrative purposes only.

[0011] Embodiments of this application provide a balance platform detection method incorporating electromyography (EMG) sensing, comprising the following steps:

[0012] The method involves applying a platform disturbance in a predetermined direction and simultaneously acquiring electromyographic signals of the user's target muscle groups and platform motion parameters; the method includes the following steps:

[0013] Apply platform perturbation in a predetermined direction and simultaneously collect electromyographic signals of the user's target muscle group and platform motion parameters;

[0014] Extract the reaction time and muscle amplitude indices of the electromyographic signals to generate muscle group response characteristics;

[0015] Based on the mapping relationship between the direction of the platform-applied perturbation and the muscle group response characteristics, the neuromuscular function assessment results are output.

[0016] The response features include:

[0017] The reaction time is used to quantify nerve conduction speed;

[0018] The muscle amplitude is used to quantify the intensity of muscle exertion.

[0019] Bilateral muscle group response differences in the target muscle group are used to identify synergistic imbalances.

[0020] Extract the reaction time and muscle amplitude indices of the electromyographic signals to generate muscle group response characteristics;

[0021] Based on the mapping relationship between the direction of the platform-applied perturbation and the muscle group response characteristics, the neuromuscular function assessment results are output.

[0022] The response features include:

[0023] The reaction time is used to quantify nerve conduction speed;

[0024] The muscle amplitude is used to quantify the intensity of muscle exertion.

[0025] Bilateral muscle group response differences in the target muscle group are used to identify synergistic imbalances.

[0026] The following detailed description of some embodiments of this application is provided in conjunction with the accompanying drawings. Unless otherwise specified, the following embodiments and features can be combined with each other.

[0027] This embodiment provides a balance platform detection method and device that combines electromyography (EMG) sensing, which aims to quantitatively assess neuromuscular function and distinguish between three types of abnormalities: delayed nerve conduction, weak muscle strength, and synergistic imbalance. At the same time, it can achieve rehabilitation assessment and training feedback by combining precise intervention in the direction of perturbation and multi-channel EMG analysis.

[0028] Please see Figure 1 Embodiments of this application provide a balance platform detection method incorporating electromyography (EMG) sensing, the method comprising the steps of:

[0029] S10: Apply platform perturbation in a predetermined direction and simultaneously collect electromyographic signals of the user's target muscle group and platform motion parameters;

[0030] S20: Extract the reaction time and muscle amplitude indices of the electromyographic signals to generate muscle group response characteristics;

[0031] S30: Based on the mapping relationship between the direction of the platform-applied perturbation and the muscle group response characteristics, output the neuromuscular function assessment results;

[0032] The response features include:

[0033] The reaction time is used to quantify nerve conduction speed;

[0034] The muscle amplitude is used to quantify the intensity of muscle exertion.

[0035] Bilateral muscle group response differences in the target muscle group are used to identify synergistic imbalances.

[0036] In this embodiment, a controllable motion platform is used, with three-dimensional force sensors at its four corners to collect real-time data on plantar pressure distribution and center of gravity changes. The platform is driven by servo motors and generates displacement or tilting disturbances in preset directions, such as forward movement, backward movement, or clockwise tilting, by sending disturbance control signals through a controller.

[0037] Identify the target muscle groups for testing. The target muscle groups include homologous bilateral muscle groups, dominant muscle groups, and antagonistic muscle groups.

[0038] The controller sends a disturbance control signal to the drive motor and simultaneously sends a TTL trigger signal. The TTL trigger signal serves as the reference time T0 for the start of EMG acquisition. The EMG acquisition module synchronously records the raw EMG signals of the target muscle group. EMG acquisition modules such as Delsys or domestically produced 6-channel surface EMG acquisition devices are used.

[0039] Specifically, in S10, the user stands on a balance platform. After receiving a command, the controller drives the platform to move forward in a preset mode, such as a medium amplitude translation. Crucially, at the moment the controller's drive motor sends a disturbance control signal, the platform simultaneously generates a TTL pulse signal and sends it to the electromyography (EMG) acquisition module. The rising edge of this TTL pulse signal is marked as the time reference point T0, representing the absolute start time of the platform's physical disturbance. The EMG acquisition module uses T0 as a reference to begin acquisition and timestamp the signal, simultaneously acquiring the raw EMG signal of the target muscle group. The data acquisition time window is set from 200ms before T0 (the baseline of the resting state before the disturbance) to 1000ms after T0 to ensure coverage of the entire muscle response process.

[0040] Specifically, S20 is an analysis and processing unit that performs a series of processes on the acquired raw EMG signal to extract reliable features.

[0041] DC component removal eliminates the DC offset in the signal. Bandpass filtering, such as 20-500Hz, preserves the effective frequency range of the EMG signal while filtering out low-frequency motion artifacts and high-frequency noise. Full-wave rectification converts the AC signal into a unidirectional signal. Low-pass filtering, such as a 6Hz Butterworth filter, smooths the rectified signal, generating an EMG envelope signal, which facilitates the observation of activation trends and peak identification.

[0042] Based on the processed envelope signal, two core indicators are extracted: reaction time, also known as latency and muscle amplitude.

[0043] Reaction time (latency) is calculated from time T0 until the electromyographic envelope signal first exceeds the resting threshold. It is defined as the time interval from the average signal amplitude during the baseline period of 200 ms before the disturbance to three times the standard deviation, in milliseconds. This indicator directly quantifies the nerve conduction velocity from the occurrence of the disturbance to the onset of a muscle response.

[0044] Muscle amplitude is determined by identifying the maximum amplitude of the electromyographic envelope signal within the analysis time window, from T0 to T0+1000 ms. This amplitude can be further normalized to the percentage of an individual's maximum voluntary contraction to eliminate the influence of absolute strength differences between individuals, thereby quantifying muscle exertion intensity.

[0045] In addition, the system also calculates the dominant muscle / antagonist muscle ratio, such as the amplitude of the tibialis anterior muscle / gastrocnemius muscle during forward translation, and the symmetry index of the homologous bilateral muscle groups, such as the amplitude of the left tibialis anterior muscle / the amplitude of the right tibialis anterior muscle.

[0046] In one specific embodiment, S30 includes the step of:

[0047] S31: Based on the platform interference in the predetermined direction, determine the dominant muscle group and the antagonistic muscle group of the target muscle group.

[0048] S32: Compare the difference in reaction time and the ratio of muscle amplitude between the dominant and antagonistic muscle groups;

[0049] S33: Determine whether the target muscle group is in an abnormal state.

[0050] In one specific embodiment, S30 further includes the step of:

[0051] S331: If no abnormal pattern is detected in the target muscle group, the target muscle group is determined to be in a normal state.

[0052] S332: If at least one abnormal pattern is detected in the target muscle group, the target muscle group is determined to be in an abnormal state.

[0053] S3321: Analyze abnormal patterns and provide the results of output neuromuscular function assessment.

[0054] Abnormal modes include:

[0055] The response amplitude of the homologous bilateral muscle groups is asymmetrical and exceeds the preset difference threshold;

[0056] The ratio of antagonistic muscle to activation intensity exceeds the preset coordination range;

[0057] In repetitive unidirectional perturbations, the rate of decrease in amplitude of the dominant muscle group was lower than the baseline value for muscle adaptive regulation.

[0058] S30 is specifically an analysis and processing unit with built-in evaluation logic that combines the extracted features with known physiological mapping relationships to identify abnormal patterns.

[0059] The first abnormal pattern is slow neural conduction. If the reaction time exceeds a preset threshold, such as 120ms set based on data from healthy individuals, it is determined that there is a delay in the neural conduction pathway.

[0060] The second abnormal pattern is insufficient muscle exertion. If the muscle amplitude %MVC is less than the preset intensity threshold such as 30%MVC, or the amplitude ratio of the bilateral homologous muscle groups exceeds the preset difference threshold, such as <0.7 or >1.3, i.e. the difference exceeds 30%, then it is judged as insufficient muscle strength or bilateral asymmetry.

[0061] The preset difference threshold can be taken from a large sample database of healthy people and clinical experience values.

[0062] To establish a normal distribution of the left-right amplitude ratio in a large-sample healthy population database, the threshold was set to the mean ± 1.96 times the standard deviation (corresponding to a 95% confidence interval). Clinical experience values, based on previous studies and clinical experience, conventionally define a ratio; for example, a difference of >25-30% in left-right amplitude is considered to indicate asymmetry.

[0063] The third abnormal pattern is synergistic imbalance. If the amplitude ratio of the dominant muscle to the antagonist muscle exceeds the preset coordination range, such as when the amplitude ratio of the tibialis anterior muscle to the gastrocnemius muscle is <2.0 during forward translation, it is judged as an abnormal muscle synergistic strategy.

[0064] In a healthy range of motion, the amplitude of the dominant muscle is significantly greater than that of the antagonist muscle. An abnormal range of motion indicates that this relationship between the dominant and antagonist muscles is disrupted. For example, during forward translation, if the amplitude of the gastrocnemius muscle is similar to or even greater than that of the tibialis anterior muscle, it is considered an imbalance.

[0065] The system can quantify this by calculating the amplitude ratio of the dominant muscle to the antagonist muscle. The healthy range of this ratio, i.e. the preset coordination range, also needs to be derived from data of healthy individuals. For example, when shifting forward, an amplitude ratio of tibialis anterior to gastrocnemius muscle >2.0 may be considered within the normal coordination range, while a value below this is considered an imbalance.

[0066] The fourth abnormal pattern is poor adaptability. In a repeated unidirectional perturbation test, such as five consecutive forward tilts, the decrease rate of amplitude of the dominant muscle group is calculated (Dr = (A1 - A5) / A1 × 100%). If Dr < the baseline value for muscle adaptive regulation, and if the baseline value for muscle adaptive regulation is taken from the decrease rate of healthy individuals, then the nervous system is judged to lack learning and adaptive capacity.

[0067] The baseline value for muscle adaptive regulation refers to the expected rate or magnitude of decrease in the amplitude of the muscle response under several consecutive identical perturbations. This needs to be defined through testing in healthy individuals.

[0068] The formula for calculating the rate of decline is (first amplitude - Nth amplitude) / first amplitude × 100%.

[0069] For example, in healthy individuals, after 5 tests, the decrease in amplitude of the dominant muscle is typically >40%. Therefore, 40% can be set as the baseline for muscle adaptation. If a patient's amplitude decrease is <40%, they are considered to have poor adaptation.

[0070] The system integrates all the judgment results to generate a quantitative neuromuscular function assessment report, and can provide personalized rehabilitation training suggestions based on the identified weak muscle groups or abnormal patterns.

[0071] Experiment 1 verifies the neuromuscular response characteristics under translational perturbation.

[0072] Experimental objective: To verify the effectiveness of the system and method in accurately quantifying the dynamic response characteristics of dominant and antagonistic muscle groups in healthy individuals under translational perturbations in different directions, including response latency and muscle amplitude.

[0073] Experimental Methods: Ten healthy subjects were selected. Using the aforementioned balance platform detection system, standardized medium-amplitude dynamic forward and backward translational perturbations were applied to each subject. Electromyographic signals of the bilateral tibialis anterior and gastrocnemius muscles were acquired using the aforementioned TTL pulse precise synchronization mechanism. Data were processed according to the aforementioned signal processing method, and the mean response latency and muscle amplitude (normalized to %MVC) of each muscle in each perturbation were extracted.

[0074] The specific data is shown in the table below.

[0075] Table 1: Electromyographic response characteristics of healthy subjects under moderate amplitude translational perturbation (mean ± standard deviation)

[0076]

[0077] Experimental conclusion:

[0078] The experimental data shows that during forward translational perturbations, the tibialis anterior muscle, as the dominant muscle group, exhibits a significantly shorter reaction latency and a significantly larger muscle amplitude than the gastrocnemius muscle; during backward translational perturbations, the gastrocnemius muscle is the dominant muscle group. These results are entirely consistent with human physiological mechanisms, fully demonstrating that the system and method described in this invention can accurately and reliably capture and quantify the dynamic response characteristics of the neuromuscular system to external perturbations, providing a direct and effective technical means for objectively assessing neuromuscular function.

[0079] Experiment 2: Multidimensional identification of abnormal muscle coordination and adaptation under tilt perturbation.

[0080] Experimental objective: To demonstrate the comprehensive ability of the system and methods to identify multi-dimensional abnormal patterns such as abnormal muscle coordination strategies, delayed nerve conduction, and adaptive regulation disorders in different disease populations.

[0081] Experimental methods: Ten healthy volunteers were recruited as the normal control group (NC group), 18 patients with peripheral vestibular diseases (Vestibular group), 20 patients with balance disorders after spinal cord injury (SCI group), and 15 patients with hemiplegia after stroke (Stroke group).

[0082] Five repeated forward and backward tilting perturbation tests were performed on all subjects. Electromyographic signals of the bilateral tibialis anterior and gastrocnemius muscles were simultaneously acquired, and the changes in amplitude and latency across multiple tests, as well as inter-group differences, were analyzed. The tibialis anterior muscle is abbreviated as TA, and the lateral gastrocnemius muscle is abbreviated as GL.

[0083] The specific data is shown in the table below.

[0084] Table 2: Intergroup comparison of electromyographic response characteristics in the first forward tilt test (mean ± standard deviation)

[0085]

[0086] Note: The TA / GA amplitude ratio is used to assess muscle synergy; the affected / unaffected side comparison is used to assess bilateral symmetry.

[0087] Table 3: Muscle adaptation analysis during five forward tilt tests (decline rate of amplitude in the right tibialis anterior muscle)

[0088]

[0089]

[0090] Note: The descent rate Dr = (Amplitude of the first test - Amplitude of the fifth test) / Amplitude of the first test × 100%.

[0091] Experimental conclusion:

[0092] The above data validates the diagnostic efficacy of the system and method. The spinal cord injury group mainly exhibited an extremely prolonged reaction latency, clearly reflecting the pathological characteristics primarily based on impaired nerve conduction velocity.

[0093] The stroke group mainly showed significant asymmetry between the affected and healthy sides, effectively revealing the imbalance in bilateral muscle coordination control after stroke.

[0094] All disease groups showed decreased muscle adaptability, and the synergistic / antagonistic muscle amplitude ratio differed significantly from the healthy group, indicating abnormal synergistic strategies.

[0095] In summary, this invention can universally and accurately distinguish neuromuscular dysfunction patterns caused by different etiologies, including multiple dimensions such as nerve conduction velocity, muscle strength, bilateral synergy, and adaptability, providing a powerful quantitative assessment tool for precise clinical rehabilitation.

[0096] In one specific embodiment, step S10 includes:

[0097] S11: The motor receives a disturbance mode command from the controller, generates a motor control signal to drive the platform to move, and the controller collects raw electromyographic signals.

[0098] S12: When the physical movement of the detection platform begins, a TTL pulse signal is generated synchronously. The time when the TTL pulse signal is generated is marked as the time reference point T0.

[0099] In this implementation, electromyography (EMG) signals are crucial data reflecting the human muscle's response to external disturbances. If the EMG signals cannot be precisely aligned with the disturbance event, it is impossible to determine at what moment the muscle began to respond to the disturbance, making further accurate extraction impossible.

[0100] Reaction time indicates the time delay before muscle activation. Amplitude indicates whether sufficient muscle force is generated within a certain period of time.

[0101] TTL pulse signals are digital level-shifting signals commonly used for time synchronization control between different system modules. In this implementation, when the controller sends a disturbance command to the motor (the platform starts moving), it simultaneously sends a TTL pulse signal. Upon receiving this TTL signal, the electromyography (EMG) acquisition system uses its time point as T0. All subsequent EMG data timestamps are calculated based on this T0. In other words, the TTL signal serves as the anchor point for aligning the platform's physical disturbance with the EMG data.

[0102] The lack of a unified reference for the platform startup time and the start time of EMG acquisition can lead to data misalignment. It can also cause the actual response to be advanced or delayed, resulting in misjudgments of slow or fast nerve conduction. Furthermore, the EMG analysis window may not cover the actual response time, leading to analysis failure. Different perturbations cannot be compared consistently because each platform movement is not synchronized with the EMG acquisition start point.

[0103] In this embodiment, it should be noted that the platform uses a programmable servo motor as the driving unit. The motor receives disturbance mode instructions from the controller, such as displacement step, velocity step, or sine sweep, through the power drive module. The controller generates motor control signals, such as PWM, pulse + direction, or CANopen servo messages, based on preset disturbance direction and amplitude parameters, to drive the platform to produce translation or tilting motion around the axis in a predetermined direction.

[0104] The controller establishes a hardware synchronization link with the EMG acquisition module. Before the experiment begins, the controller first initializes the sampling task and opens the raw EMG signal channel, with an example sampling rate ≥2kHz and A / D resolution ≥16bit. To achieve time zero-point consistency, the platform's mechanical side monitors the initial moment of the platform's physical motion in real time via a motor encoder, gyroscope, or displacement sensor. When a platform pose quantity, such as displacement or angular displacement, is detected to exceed the static noise threshold and persist for more than a minimum duration window (≥2ms) to eliminate jitter, the controller immediately outputs a TTL pulse signal (e.g., 5V level) with a pulse width of, for example, 10ms. The moment this pulse is emitted is marked as the time reference point T0 and simultaneously written to the acquisition buffer as a trigger marker. To reduce the uncertainty of propagation delay, the TTL pulse is directly connected to the digital trigger input port of the EMG front end via a hardware interrupt, avoiding jitter caused by software queues. Simultaneously, the encoder count value corresponding to T0 and the controller system clock timestamp are recorded on the motor side for post-experiment verification.

[0105] Throughout the process, the controller continuously acquires raw electromyographic (EMG) signals and aligns the T0 data with the same sampling time reference, achieving strict synchronization between platform disturbance events and EMG data. If necessary, a warm-up calibration can be applied before platform startup. Several no-load triggers are performed to determine the system's inherent delay, establish startup criteria thresholds such as a displacement threshold of 0.2–0.5 mm or an angle threshold of 0.02–0.05°, and define a de-jitter time window. These criteria are then written into the firmware parameter table of this embodiment to ensure consistency and repeatability across different batches of devices.

[0106] By issuing a TTL trigger as soon as the physical motion of the platform is detected, and defining a hardware-level synchronization mechanism T0 based on this trigger, the time registration error between the disturbance event and the electromyography sampling is significantly reduced, so that the systematic error of reaction time measurement is mainly dominated by the subject's physiological response rather than the device delay.

[0107] Compared to using only the moment the control command is issued as the zero point, this embodiment uses the actual start-up moment of the platform as T0, effectively avoiding deviations caused by mechanical start-up delays due to motor dead zones, transmission backlash, and load inertia, thus improving the accuracy and batch comparability of reaction time determination. Furthermore, the use of a hardware interrupt-driven TTL pulse with a fixed pulse width ensures a clear trigger edge and minimal jitter (typically <±0.5ms), facilitating the EMG front-end to write T0 onto the same sampling time axis. Combined with the setting of the start-up threshold and debouncing criteria, false triggering caused by minor vibrations or electromagnetic interference can be suppressed, thereby improving synchronization reliability.

[0108] In summary, this embodiment achieves high-precision synchronous acquisition of platform disturbance events and raw electromyographic signals without increasing the burden on the subjects. This provides a stable and consistent time reference for subsequent calculations of reaction time and muscle amplitude based on T0, thereby improving the measurement sensitivity and repeatability of this method in the quantification of nerve conduction velocity and the assessment of muscle exertion intensity. In this embodiment alone, millisecond-level time resolution and consistent synchronization accuracy across multiple channels within a single experiment can be achieved.

[0109] In one specific embodiment, step S20 includes:

[0110] S21: Extract the raw electromyographic signal during the effective time period T(t);

[0111] S22:T0-200ms≤T(t)≤T0+1000ms;

[0112] S23: Sequentially execute DC component elimination, 20-500Hz bandpass filtering, full-wave rectification, and 6Hz low-pass filtering to generate electromyographic envelope signals, thereby generating electromyographic signals.

[0113] In this embodiment, it should be noted that: using the time reference point T0 as the alignment reference, the original electromyographic signals of each channel are truncated according to a fixed window to obtain the effective time period T(t), the range of which is T0. - 200ms to T0+1000ms.

[0114] The left side of the window is reserved for 200ms to characterize the resting baseline, while the right side of the window, 1000ms, covers the short-latency reflex induced by the perturbation and the subsequent voluntary / coordinated contraction. To ensure the comparability of subsequent response duration and muscle amplitude,

[0115] This embodiment employs a unified preprocessing pipeline: DC component elimination, 20-500Hz bandpass filtering, full-wave rectification, 6Hz low-pass filtering to generate electromyographic envelope signals, baseline and threshold setting, and data quality control.

[0116] In the DC component elimination process, the truncated segment is subjected to mean reduction processing and a first-order high-pass filter of 0.5–1 Hz is applied to suppress electrode polarization drift and low-frequency baseline drift, ensuring that the envelope is not affected by the slow variation trend.

[0117] In 20–500 Hz bandpass filtering, linear-phase FIR or zero-phase IIR, such as a forward-backward strategy, is preferred to avoid systematic shifts in response time determination due to group delay. A lower cutoff of 20 Hz suppresses motion artifacts and baseline fluctuations, while an upper cutoff of 500 Hz suppresses high-frequency electrical noise. If necessary, power frequency notches such as 50 / 60 Hz and their harmonics can be inserted before and after the bandpass as options to adapt to different electromagnetic environments.

[0118] In full-wave rectification, the absolute value of the bandpass output is taken to convert bipolar electromyography into a unipolar energy metric, which facilitates subsequent gradual envelope estimation and threshold comparison.

[0119] In generating the electromyography envelope using a 6Hz low-pass filter, a 6Hz cutoff low-pass filter such as a linear-phase FIR or a zero-phase IIR is employed to obtain a linear envelope signal, preserving the gradual variation trend of muscle activation intensity and suppressing high-frequency jitter. To avoid filtering boundary effects, short-time mirror extensions are added at both ends of the truncated segment during processing, and the extended intervals are removed after output.

[0120] In baseline and threshold settings, the envelope mean μ0 and standard deviation σ0 are calculated within the resting interval [T0-200ms, T0-50ms]. The resting threshold θ is defined as μ0 + k·σ0, where k can be 2–3, preset by the equipment manufacturer or experimentally, providing a unified baseline for subsequent reaction time and peak detection. Suspected motion artifacts, such as sudden large fluctuations and low-frequency fluctuations, are labeled using a combined amplitude-frequency criterion and used for result reliability assessment without altering the master data.

[0121] In data quality control, if the σ0 of a channel exceeds the preset range during the resting period, or if the entire signal exhibits saturation or dropout, and the criteria for saturation or dropout include shear rate, sample shear ratio, or duration of consecutive zero values, the signal is marked as requiring retesting or entering a robust statistical process, such as median substitution or channel weighting, to ensure reliable evaluation.

[0122] Through the above steps, an envelope signal that is strictly aligned with T0 and has controlled noise is obtained, which serves as the sole input for calculating reaction time and muscle amplitude.

[0123] A preprocessing strategy consisting of a fixed event alignment window, a zero-phase filter chain, and a statistical threshold combination can stably recover the electromyographic activation contour induced by perturbation without introducing additional time delay bias.

[0124] Compared to methods that rely solely on threshold detection of the original waveform, this embodiment significantly improves the signal-to-noise ratio and peak resolution by using an energy envelope formed by a 20–500 Hz bandpass and full-wave rectification, and a 6 Hz low-pass filter, while reducing interference from electromyographic coherent ringing on peak calculation.

[0125] Zero-phase processing reduces the systematic error of group delay on reaction time to near zero, so that the measured reaction time mainly reflects the lag of real neural conduction and central / reflex pathways rather than algorithmic lag.

[0126] Based on the adaptive threshold setting θ = μ0 + k·σ0 for the resting segment before T0, the detection sensitivity remains consistent under different subjects, different electrode impedances, and different environmental noise levels. Meanwhile, mirror continuation and boundary processing avoid edge attenuation and spurious peaks caused by short-window filtering.

[0127] The overall effect is that in repeated measurements on the same subject, the repeatability of reaction time is improved, the jitter caused by preprocessing is significantly reduced, and the relative error of muscle amplitude peak is reduced. This provides stable and comparable input data for subsequent neuromuscular function assessment based on mapping relationship. For the processing chain described in this embodiment only, robust detection of reaction initiation can be achieved on a millisecond time scale, and consistent amplitude-time quantization accuracy can be maintained across multiple channels in a single experiment.

[0128] In one specific embodiment, step S30 further includes:

[0129] S1: Calculate the reaction time from time T0 until the electromyographic envelope signal exceeds the resting threshold;

[0130] S2: Based on the linear envelope signal, the reaction time and muscle amplitude are derived. The reaction time is the time from the moment the synchronization signal is triggered until the envelope signal first exceeds the resting threshold. The muscle amplitude is the maximum amplitude value in the envelope signal.

[0131] Using the actual start-up instant T0 of the platform as the time reference, the linear envelope signal of each electromyographic channel is processed within the window of [T0-200ms, T0+1000ms].

[0132] Calculate the mean μ0 and standard deviation σ0 in the resting interval [T0-200ms, T0-50ms], and set the resting threshold θ = μ0 + k·σ0 (k approximately 2–3).

[0133] Within the range [T0-10ms, T0+1000ms], find the moment when the envelope first crosses θ and remains continuously for ≥8–15ms. If the threshold is only crossed momentarily, ignore it. Refine tonset using linear interpolation.

[0134] Reaction time RT = tonset - T0. Muscle amplitude Amax is the maximum value of the envelope within [tonset, T0 + 1000ms], and the maximum value is the peak value after slightly smoothing the envelope.

[0135] If the entire window does not exceed the threshold, it is recorded as "no response". Synchronous output quality flags include excessive noise, saturation, etc.

[0136] Using the actual mechanical start-up T0 as the zero point and refining the starting point through interpolation, the RT measurement error is small, reaching the millisecond level. Furthermore, by using a combination of adaptive threshold and continuous hold as criteria, it can filter out sporadic noise and spurious triggers, reducing false starting points. The peak value is obtained based on a smooth envelope, resulting in better consistency between channels and between repeated tests. No response or quality anomalies are individually marked, facilitating retesting and result interpretation.

[0137] In one specific embodiment, a disturbance mode is defined as applying a platform disturbance in a predetermined direction. This disturbance mode includes a translation mode (forward and backward movement) and a tilting mode (up and down tilting around an axis). Here, T0 is the starting moment when the motor begins to drive the platform to translate or tilt physically.

[0138] In this embodiment, it should be noted that there are two types of disturbances: forward and backward translation, which refers to the linear movement of the platform along the forward and backward direction, and tilting around the axis, which refers to the pitching of the platform.

[0139] The unified time reference T0 is defined as the moment when the motor actually starts driving and causes the platform to produce a detectable change in displacement or tilt.

[0140] Once a startup exceeding a tiny threshold and lasting for an extremely short time is detected, the controller immediately sends a TTL pulse as a marker to precisely align the platform disturbance with the electromyography acquisition.

[0141] A balance platform detection device incorporating electromyography (EMG) sensing mainly consists of the following components: an electric platform, a three-dimensional force sensing module, an EMG acquisition module, a controller, and an analysis and processing unit.

[0142] A balancing platform is used to apply disturbances in a controllable direction;

[0143] The three-dimensional force sensing module is used to collect data on the movement of the center of gravity.

[0144] The electromyography (EMG) acquisition module is used to acquire surface EMG signals of the target muscle group.

[0145] The controller is connected to the platform and the electromyography (EMG) acquisition module. It is used to coordinate and control disturbances and trigger synchronous acquisition.

[0146] The analysis and processing unit is used to extract the reaction time and muscle amplitude of the electromyographic signal and output the neuromuscular function assessment results.

[0147] Three-dimensional force sensing modules are installed at each of the four corners of the balance platform. The electromyography (EMG) acquisition module is a 6-channel surface EMG acquisition unit, with channels allocated to at least the TA, GL, and rectus femoris muscles. The controller uses a TTL trigger signal to align the EMG module with the platform's disturbance events. The analysis and processing unit includes a functional module for generating rehabilitation training suggestions, providing target muscle group strengthening plans based on the evaluation results.

[0148] The aforementioned balance platform detection device incorporating electromyography (EMG) sensing is applied to any of the steps described above, and the connections between the components of this device are existing technology. Therefore, they will not be described in detail here.

[0149] Therefore, the above-mentioned balance table detection method and device combining electromyography (EMG) sensing integrates surface EMG sensing and synchronous triggering circuit in the balance table device. Under the time reference T0 of the platform disturbance, the reaction time and amplitude of the target muscle group are aligned and calculated. A mapping evaluation model is constructed to generate a one-to-one correspondence of the disturbance direction to the dominant muscle or antagonistic muscle response characteristics, thereby realizing the comprehensive quantification of nerve conduction velocity, muscle strength and muscle group coordination.

[0150] Furthermore, the system can simultaneously collect platform motion and multi-channel EMG data under a unified time reference, extract response features composed of reaction time and muscle amplitude, combine pattern recognition of dominant or antagonistic muscles and bilateral homologous muscles, output assessment results of nerve conduction efficiency, muscle strength and whether synergy is unbalanced, and further provide rehabilitation training suggestions.

[0151] The above description is merely an embodiment of this application. It should be noted that those skilled in the art can make improvements without departing from the inventive concept of this application, but these improvements all fall within the protection scope of this application.

Claims

1. A balance platform detection method combined with electromyographic sensing, characterized in that, The method comprises the steps of: applying a platform disturbance in a predetermined direction, and synchronously collecting the electromyogram of the target muscle group of the user and the platform motion parameters; extracting the response duration and muscle amplitude indicators of the electromyogram to generate muscle group response characteristics; outputting the neuromuscular function evaluation result based on the mapping relationship between the direction of the platform disturbance and the muscle group response characteristics; wherein the response characteristics comprise: the response duration is used to quantify the nerve conduction velocity; the muscle amplitude is used to quantify the muscle strength; the bilateral muscle group response difference of the target muscle group is used to identify the coordination imbalance.

2. The balance platform test method with electromyography sensing according to claim 1, characterized in that, The "outputting the neuromuscular function evaluation result based on the mapping relationship between the direction of the platform disturbance and the muscle group response characteristics" comprises the steps of: determining the dominant muscle group of the target muscle group and the antagonistic muscle group of the target muscle group according to the platform disturbance in the predetermined direction; comparing the difference value of the response duration and the ratio of the muscle amplitude of the dominant muscle group and the antagonistic muscle group; determining whether the target muscle group is in an abnormal state.

3. The balance platform test method with electromyography sensing of claim 1, wherein, The "outputting the neuromuscular function evaluation result based on the mapping relationship between the direction of the platform disturbance and the muscle group response characteristics" comprises the steps of: if the target muscle group is detected without abnormal patterns, it is determined that the target muscle group is in a normal state; if the target muscle group is detected with at least one abnormal pattern, it is determined that the target muscle group is in an abnormal state; analyzing the abnormal patterns and outputting the neuromuscular function evaluation result. The abnormal patterns comprise: the response amplitude asymmetry of the homologous bilateral muscle group exceeds the preset difference threshold; the activation intensity ratio of the antagonistic muscle group exceeds the preset coordination range; in repeated disturbance in the same direction, the amplitude decline rate of the dominant muscle group in the repeated disturbance in the same direction is lower than the muscle adaptability adjustment reference value.

4. The balance platform test method combined with electromyographic sensing according to claim 1, characterized in that, The "applying a platform disturbance in a predetermined direction, and synchronously collecting the electromyogram of the target muscle group of the user and the platform motion parameters" step comprises: the motor receives the disturbance mode instruction from the controller, generates the motor control signal to drive the platform motion, and the controller collects the original electromyogram; detecting the start of the platform physical motion, and synchronously generating a TTL pulse signal, the time when the TTL pulse signal is generated is marked as a time reference point T0.

5. The method of claim 4, wherein the method further comprises: The "extracting the response duration and muscle amplitude indicators of the electromyogram" step comprises: intercepting the original electromyogram in the effective time period T(t); T0-200ms≤T(t)≤T0+1000ms; in sequence, performing direct current component elimination, 20-500Hz band pass filtering, full wave rectification and 6Hz low pass filtering to generate an electromyogram envelope signal, to generate the electromyogram.

6. The balance platform test method combined with electromyographic sensing according to claim 1, characterized in that, The "outputting the neuromuscular function evaluation result based on the mapping relationship between the direction of the platform disturbance and the muscle group response characteristics" step further comprises: calculating the response duration from T0 time to the time when the electromyogram envelope signal exceeds the resting threshold; based on the linear envelope signal, the response duration and the muscle amplitude are obtained, the response duration is the time period from the synchronous signal trigger time to the first time when the envelope signal exceeds the resting threshold, and the muscle amplitude is the maximum amplitude in the envelope signal.

7. The method of claim 1, wherein the method further comprises: The "applying a platform disturbance in a predetermined direction, and synchronously collecting the electromyogram of the target muscle group of the user and the platform motion parameters" comprises: the disturbance mode comprises a translation mode of moving forward and backward and an inclination mode of tilting up and down around an axis. Wherein, the T0 is the physical motion starting time when the motor starts to drive the platform to translate or tilt.

8. A balance platform testing device incorporating electromyographic sensing, characterised in that, The application is applied to the balance platform detection method of the myoelectric sensor, and the device comprises: A balance platform is used to apply controllable direction disturbance; A three-dimensional force sensing module is used to collect the center of gravity motion data; A myoelectric collection module is used to collect the surface myoelectric signal of the target muscle group; A controller is connected with the platform and the myoelectric collection module, and is used to coordinate and control the disturbance and trigger the synchronous collection; An analysis processing unit is used to extract the reaction time and muscle amplitude of the myoelectric signal, and output the neuromuscular function evaluation result.

9. The balance platform detection device combined with myoelectric sensing according to claim 8, characterized in that, The balance platform is provided with three-dimensional force sensing modules at four corners respectively; The myoelectric collection module is a 6-channel surface myoelectric collection unit, and the channel distribution at least includes the tibialis anterior muscle, the lateral gastrocnemius muscle and the rectus femoris muscle.

10. The balance platform detection device combined with myoelectric sensing of claim 8, wherein, The controller controls the myoelectric module and the platform disturbance event time alignment through the TTL trigger signal; The analysis processing unit comprises a function module for generating rehabilitation training suggestions, and gives the target muscle group strengthening scheme according to the evaluation result.

Images