A navigation brain-muscle interface control method and system

Abstract

This application discloses a navigation brain-muscle interface control method and system, relating to the field of medical device technology. The method includes: controlling an electromyography (EMG) acquisition device to collect multi-channel surface EMG signals and synchronous EEG signals when a target person performs navigation actions, obtaining raw EMG and EEG datasets covering the shoulder-neck to upper arm region; determining muscle activation regions based on the energy distribution of signals in each channel of the raw EMG dataset, and obtaining the activation timing relationship between channels based on the activation start time of the corresponding channels in the muscle activation regions; combining the muscle activation regions and the activation timing relationship between channels into an EMG feature combination, comparing the EMG feature combination with multiple pre-stored navigation command templates to obtain matching navigation control commands; determining whether there are any physiological abnormalities based on the raw EEG dataset, and controlling the output state of the navigation control commands based on the determination result. This application makes brain-muscle interface control more stable and more closely matches the user's actual abilities.

Description

Technical Field

[0001] This application relates to the field of medical device technology, and in particular to a navigation brain-muscle interface control method and system. Background Technology

[0002] When using wheelchairs, rehabilitation robots, or other assistive devices, people often need to use body movements to issue navigation commands; for example, they might move their shoulder to turn left or lean forward slightly to move forward. For patients with upper limb paralysis, the typical muscles in their hands or shoulders may no longer be able to exert force normally. Therefore, they often compensate by using atypical muscles such as those in their neck, trapezius, or even chest muscles to complete movements. This compensatory mechanism is a natural adaptive strategy and a genuine behavioral habit in their daily lives.

[0003] Current mainstream electromyography (EMG) navigation control methods typically involve having users repeatedly perform standard movements in a laboratory setting, following fixed instructions, such as "Please use your deltoid muscles to raise your arm." The system then records the surface EMG signals corresponding to this movement, using them as a template for future "raise arm" intention recognition. This method is designed with the assumption that the user will use fixed, anatomically typical muscles to perform the same movement. However, in real life, the same patient might use their trapezius muscles for left turns one day and switch to their pectoralis major muscles the next due to fatigue. The distribution and activation sequence of the EMG signals will change, making it impossible for the system to recognize the movement.

[0004] As a result, when upper limb paralyzed patients use atypical muscle compensation to perform navigation actions, the pattern of surface electromyography signals is mismatched with the preset template, causing navigation commands to be unable to be correctly recognized or to be completely lost. Summary of the Invention

[0005] In view of the aforementioned problems, this application is hereby filed.

[0006] Therefore, this application provides a navigation brain-muscle interface control method and system that can solve the problems mentioned in the background art.

[0007] To solve the above-mentioned technical problems, this application provides the following technical solution: In a first aspect, this application provides a navigation brain-muscle interface control method, comprising: controlling an electromyography (EMG) acquisition device to acquire multi-channel surface EMG signals and synchronous EEG signals when a target person performs navigation actions, thereby obtaining a raw EMG dataset and a raw EEG dataset covering the shoulder-neck to upper arm region; The muscle activation region is determined based on the energy distribution of each channel signal in the original electromyography dataset, and the activation timing relationship between channels is obtained based on the activation start time of the corresponding channel in the muscle activation region. The activation sequence relationship between the muscle activation region and the channel is combined into an electromyographic feature combination. The electromyographic feature combination is compared with a pre-stored navigation instruction template to obtain a matching navigation control instruction. The system determines whether any abnormal physiological events exist based on the original EEG dataset, and controls the output state of the navigation control commands based on the determination result.

[0008] Preferably, the control electromyography (EMG) acquisition device acquires multi-channel surface EMG signals and synchronous EEG signals when the target person performs navigation actions, including: The multi-channel surface electromyography (EMG) signals and synchronous electroencephalogram (EEG) signals collected when the target personnel perform navigation actions include: A multi-channel electrode array was deployed in the shoulder and neck area to the upper arm region of the target personnel. Guide the target personnel to perform navigation actions; Record multi-channel surface electromyography (EMG) and electroencephalography (EEG) signals during the execution of the navigation action to generate time-aligned raw EMG and raw EEG datasets; The original electromyography dataset was subjected to baseline correction and power frequency notch filtering, and the original electroencephalography dataset was subjected to electrooculography artifact removal.

[0009] Preferably, obtaining the activation timing relationship between channels includes: The energy of each channel in the original electromyography dataset is integrated by a sliding window to obtain the electromyography energy sequence of each channel over time. In the electromyographic energy sequence, the channel set corresponding to the local energy maxima is identified, and channels that are spatially continuous and whose local energy maxima occur in the same temporal neighborhood are clustered as muscle activation regions. For each channel in the muscle activation region, the activation start time is determined based on the rising edge of the electromyographic energy sequence, and the activation sequence relationship between channels is generated according to the relative order of the activation start times of each channel.

[0010] Preferably, the set of channels corresponding to the identified local maxima of energy includes: Traverse the electromyographic energy sequence of each channel and mark the points with energy values ​​greater than their adjacent time frames as local maxima. Channels that are spatially adjacent and whose local maxima time difference does not exceed a preset time neighborhood are grouped into the same cluster; For each channel in the muscle activation region, the activation initiation time is determined based on the rising edge of the electromyographic energy sequence, including: By tracing back from the local maximum point, the inflection point where the energy continues to rise is found as the activation start time. The inflection point is the starting position where the energy increases monotonically and the increase exceeds the preset ratio of the previous frame.

[0011] Preferably, comparing the electromyographic feature combination with a pre-stored navigation instruction template includes: The muscle activation region is represented as a continuous channel number interval, and the activation sequence relationship between the channels is represented as an ordered set of pairs; Load the navigation instruction template library. Each navigation instruction template corresponds to a type of navigation control instruction and contains at least one set of rules for combining allowed channel number ranges and ordered pairs. Traverse the navigation instruction template library and determine whether the electromyographic feature combination satisfies any combination rule.

[0012] Preferably, determining whether the electromyographic feature combination satisfies any combination rule includes: Determine whether the channel number range of the electromyographic feature combination is covered by the channel number range defined by the combination rule; Determine whether the ordered set of the electromyographic feature combinations is a subset of the ordered set of combinations defined by the combination rule.

[0013] Preferably, the determination of whether a physiological abnormality event exists includes: Frequency band power analysis was performed on the original EEG dataset to obtain the power distribution of multiple standard frequency bands in each channel; Physiological abnormal events are identified based on the power distribution and signal amplitude timing, including sudden high-amplitude slow wave events or widespread rhythm disorder events. If a physiological abnormality is detected, the current navigation control command to be output is interrupted and the device control state is set to standby. If no abnormal physiological event is detected, the navigation control command is allowed to be sent to the navigation execution device.

[0014] Preferably, the identification of abnormal physiological events includes: When the signal amplitude of the frontal central channel exceeds the upper limit of the EEG signal amplitude of the user in the resting state, and the duration of the high amplitude state exceeds the upper limit of the duration of the high amplitude EEG of the user in the resting state, it is determined to be a sudden high amplitude slow wave event. When the power of the alpha rhythm in all leads is lower than the user's baseline alpha rhythm power at rest, and the power of the theta rhythm is higher than the user's baseline theta rhythm power at rest, it is considered a pervasive rhythm disorder event.

[0015] Preferably, when the electromyographic feature combination satisfies multiple combination rules, the navigation control command corresponding to the combination rule with the highest overlap between the channel number interval and the electromyographic feature combination interval is selected as the output result.

[0016] Secondly, this application also provides a navigation brain-muscle interface control system, including: a signal acquisition module, which controls the electromyography acquisition device to acquire multi-channel surface electromyography signals and synchronous electroencephalogram signals when the target person performs navigation actions, so as to obtain a raw electromyography dataset and a raw electroencephalogram dataset covering the shoulder and neck to upper arm region; The feature extraction module determines the muscle activation region based on the energy distribution of each channel signal in the original electromyography dataset, and obtains the activation timing relationship between channels based on the activation start time of the corresponding channel in the muscle activation region. The instruction matching module combines the activation timing relationship between the muscle activation region and the channel into an electromyographic feature combination, and compares the electromyographic feature combination with multiple pre-stored navigation instruction templates to obtain a matching navigation control instruction. The safety control module determines whether there are any abnormal physiological events based on the original EEG dataset, and controls the output state of the navigation control commands based on the determination result.

[0017] Thirdly, this application also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to perform the following steps: The electromyography (EMG) acquisition device is controlled to collect multi-channel surface EMG signals and synchronous EEG signals when the target person performs navigation actions, so as to obtain raw EMG datasets and raw EEG datasets covering the shoulder and neck to upper arm region. The muscle activation region is determined based on the energy distribution of each channel signal in the original electromyography dataset, and the activation timing relationship between channels is obtained based on the activation start time of the corresponding channel in the muscle activation region. The activation sequence relationship between the muscle activation region and the channel is combined into an electromyographic feature combination. The electromyographic feature combination is compared with a pre-stored navigation instruction template to obtain a matching navigation control instruction. The system determines whether any abnormal physiological events exist based on the original EEG dataset, and controls the output state of the navigation control commands based on the determination result.

[0018] Fourthly, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, performs the following steps: The electromyography (EMG) acquisition device is controlled to collect multi-channel surface EMG signals and synchronous EEG signals when the target person performs navigation actions, so as to obtain raw EMG datasets and raw EEG datasets covering the shoulder and neck to upper arm region. The muscle activation region is determined based on the energy distribution of each channel signal in the original electromyography dataset, and the activation timing relationship between channels is obtained based on the activation start time of the corresponding channel in the muscle activation region. The activation sequence relationship between the muscle activation region and the channel is combined into an electromyographic feature combination. The electromyographic feature combination is compared with a pre-stored navigation instruction template to obtain a matching navigation control instruction. The system determines whether any abnormal physiological events exist based on the original EEG dataset, and controls the output state of the navigation control commands based on the determination result.

[0019] Implementing this application will have the following beneficial effects: This application provides a navigation brain-muscle interface control method and system. This application avoids abnormal muscle recruitment caused by tension and unfamiliarity in hospital or laboratory environments by collecting multi-channel surface electromyography (EMG) signals and synchronous electroencephalogram (EEG) signals from target personnel performing real navigation actions in a familiar environment. This makes the recorded EMG patterns closer to the actual usage state. Based on this, the application automatically identifies the currently activated muscle area according to the energy distribution of the EMG signals and constructs EMG feature combinations by combining the activation sequence between channels. These combinations are then matched with multiple preset sets of effective navigation command templates. Even if the user activates atypical compensatory muscle groups such as the trapezius due to muscle weakness, as long as the activation pattern falls into any effective combination, the application can still correctly output navigation commands. This solves the problem of mismatch when traditional one-to-one templates are used in the face of compensatory strategies, making the control more stable and more in line with the user's actual ability.

[0020] This application does not rely on fixed anatomical locations or preset muscle identities. Instead, it represents muscle activation areas as continuous channel number intervals and activation sequence relationships as ordered pairs. It uses interval coverage and subset matching rules to distinguish commands, allowing the same navigation command to be triggered by multiple different electromyographic combinations. When multiple templates match simultaneously, the command with the highest overlap with the current activation area is selected first. This retains flexibility and avoids command ambiguity, enabling the effective identification of compensatory behaviors of different users and the same user in different states, thus improving the adaptability and usability of navigation control.

[0021] Before outputting navigation commands, this application simultaneously analyzes the EEG signal for any physiological abnormalities such as sudden high-amplitude slow waves or widespread rhythm disturbances. Once an abnormality is detected, the command output is immediately interrupted and the system enters a standby state to prevent accidental triggering of turning or acceleration actions when the user experiences spasms, confusion, or epileptic premonitions. This safety monitoring is entirely based on the objective characteristics of the raw EEG data and requires no additional sensors. It not only ensures the safety baseline for users with high-level paralysis when the device malfunctions, but also ensures that navigation operations are only performed when the physiological state is stable, making the entire control process both intelligent and reliable. Attached Figure Description

[0022] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the drawings used in the description of the embodiments of this application or related technologies will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0023] Figure 1 This is an overall flowchart of a navigation brain-muscle interface control method involved in this application; Figure 2 This is an application environment diagram of a navigation brain-muscle interface control method involved in this application; Figure 3 This is a schematic diagram of the overall structure of a navigation brain-muscle interface control system involved in this application; Figure 4 This is a computer device diagram of a navigation brain-muscle interface control method related to this application. Detailed Implementation

[0024] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0025] In one exemplary embodiment, such as Figure 1 As shown, a navigation brain-muscle interface control method is provided, including: S100: Controls the electromyography (EMG) acquisition device to acquire multi-channel surface EMG signals and synchronous EEG signals when the target person performs navigation actions, and obtains raw EMG datasets and raw EEG datasets covering the shoulder and neck to upper arm region; It should be noted that current navigation-related neuromuscular interface control methods mostly rely on collecting electromyographic (EMG) and electroencephalographic (EEG) signals from individuals performing standardized navigation actions in laboratory or hospital settings. However, in controlled environments, upper limb paralyzed patients often experience tension, fatigue, or insufficient activation of compensatory habits, resulting in surface EMG signals that fail to accurately reflect the activation patterns of commonly used muscle groups. For example, when given a command to raise the arm, a patient may temporarily use the trapezius muscle as compensation due to deltoid weakness. However, this compensatory pattern is suppressed in unfamiliar environments due to focused attention, leading to a lack of information about truly usable muscle groups in the collected sEMG data. Furthermore, during repetitive fixed movements, muscle fatigue can cause a shift in the activation area, leading to surface EMG signals being misinterpreted as corresponding to another navigation action, resulting in template mismatch.

[0026] Therefore, this solution emphasizes collecting multi-channel surface electromyography (EMG) signals and synchronous electroencephalogram (EEG) signals from the target individual when performing navigation actions in a natural state in S100. Here, navigation actions refer to autonomous behaviors that simulate real-world navigation scenarios, such as maneuvering a wheelchair around obstacles or turning to enter a room, rather than isolated action commands from a laboratory setting, in order to reproduce the physiological response patterns of patients when actually using neuromuscular interface devices as closely as possible.

[0027] The electromyography (EMG) acquisition device is a multi-channel dry and wet electrode array attached to the shoulder, neck, and upper arm region to simultaneously capture sEMG and EEG signals. Navigation actions are performed autonomously by the user in a familiar environment, avoiding abnormal muscle recruitment caused by environmental interference. The raw EMG and raw EEG datasets are signal sets recorded synchronously within the same time window, ensuring the effectiveness of subsequent spatial and temporal feature alignment. The acquisition process is independent of the hospital setting; the signals directly originate from the actual navigation intent expression process, significantly improving the accuracy of subsequent muscle activation area recognition and command matching.

[0028] In some embodiments, step S100 involves acquiring multi-channel surface electromyography (EMG) signals and synchronous electroencephalogram (EEG) signals when the target person performs navigation actions in a natural state, including steps S110 to S140: Step S110: Deploy a multi-channel electrode array in the shoulder and neck to upper arm area of ​​the target person. The electrode array covers the body surface locations corresponding to the upper and middle trapezius muscles, the anterior, middle and posterior deltoid muscles, and the clavicular part of the pectoralis major muscle.

[0029] Understandably, the electrode array uses a ring or grid layout with no fewer than six channels. Each electrode is attached to the body surface area near the standard electromyographic (EMG) localization point to ensure coverage of typical and common compensatory muscle groups that may be involved in upper limb navigation movements. The electrode placement method follows the SENIAM EMG acquisition guidelines and is standard practice in this field, so it will not be elaborated upon here.

[0030] Step S120: Guide the target personnel to autonomously perform at least five basic navigation actions in a familiar environment, including forward, backward, left turn, right turn and stop, with each action repeated no less than three times.

[0031] It is understandable that navigation actions refer to the actual operational behaviors of the target person when using a wheelchair or mobility aid in daily life, such as steering with a handle, leaning forward to initiate forward movement, etc., rather than isolated muscle contraction commands from a laboratory setting. Repeated execution is used to capture the natural fluctuation range of electromyographic patterns under the same intention, providing robust samples for subsequent template construction.

[0032] Step S130: Simultaneously record multi-channel surface electromyography (EMG) and electroencephalography (EEG) signals during the execution of navigation actions to form time-aligned raw EMG and raw EEG datasets.

[0033] Understandably, surface electromyography (EMG) signals and electroencephalogram (EEG) signals are time-aligned through multi-channel synchronous sampling on the same acquisition device, with a sampling rate of no less than 1000 Hz and shared reference electrodes, ensuring that the two signals are strictly synchronized on the time axis, which facilitates subsequent cross-modal event correlation analysis.

[0034] Step S140: Baseline correction and power frequency notch processing are performed on the original electromyography (EMG) dataset, and electrooculography (EOG) artifact removal is performed on the original electroencephalography (EEG) dataset to obtain preprocessed EMG and EEG data.

[0035] Understandably, baseline correction uses the sliding window mean subtraction method, power frequency notch processing uses a 50Hz or 60Hz second-order IIR notch filter, and electrooculography artifact removal uses threshold-based sample removal or template regression. The above preprocessing methods are all conventional signal processing techniques and will not be elaborated here.

[0036] It should be noted that during the acquisition of EEG signals, the target person's eye movements, blinking, or eye muscle contractions will generate strong potential changes. These interference signals are called EEG artifacts. At the same time, the amplitude of EEG artifacts is usually much larger than the EEG signals generated by brain neural activity, and their frequency components overlap with the theta rhythm and alpha rhythm in EEG. If not processed, EEG artifacts can easily be misjudged as widespread rhythm disorders or sudden high-amplitude slow wave events, causing the safety control module to erroneously interrupt the navigation command output.

[0037] Therefore, this application performs oculomotor artifact removal on the raw EEG dataset during the signal preprocessing stage to preserve EEG signals generated by genuine neural activity in the cerebral cortex. Oculomotor artifact removal can employ a sample labeling and removal method based on amplitude thresholds. That is, when a channel signal shows a steep increase in amplitude within a short period of time and exceeds the normal EEG range, the data during this period is marked as contaminated by oculomotor activity and removed. Alternatively, independent component analysis can be used to separate and remove components highly correlated with eye movements from multi-channel EEG signals. These methods are routine techniques in the field of electroencephalography (EEG) testing and will not be elaborated upon here.

[0038] It should be noted that after the removal of ocular artifacts, transient interference caused by eye movements in the EEG signals entering the physiological abnormality event discrimination stage is effectively suppressed, thereby avoiding misjudgment triggered by normal physiological behaviors such as blinking or eye movement, making the safety control logic more stable and reliable.

[0039] Preferably, steps S110 to S140 involve collecting multi-channel electromyography and synchronous electroencephalography signals in a natural environment and performing standardized preprocessing. This effectively avoids sEMG data distortion caused by tension, movement distortion, or compensatory strategy inhibition in hospital or laboratory environments, providing high-quality input data for subsequent accurate identification of muscle activation areas and temporal relationships.

[0040] Preferably, step S100 obtains electromyography and electroencephalography data in real-world scenarios through the above sub-steps, which not only restores the physiological response patterns of upper limb paralyzed patients in actual navigation tasks, but also lays a reliable data foundation for the fusion control logic based on spatial topology, temporal characteristics and safety monitoring in S200–S400.

[0041] S200: Determine the muscle activation region based on the energy distribution of each channel signal in the original electromyography dataset, and obtain the activation timing relationship between channels based on the activation start time of the corresponding channel in the muscle activation region; It should be noted that existing electromyographic control methods typically use the sEMG amplitude or frequency domain energy of preset channels as criteria when recognizing user intentions, implicitly assuming that the user always uses typical muscles to complete the movement. However, upper limb paralysis patients often use atypical compensatory muscle groups to perform the same navigation movement due to nerve damage or muscle atrophy. If the discrimination is still based on fixed channels or fixed muscle templates, the following problems will occur: when the patient uses the trapezius muscle to compensate for a left turn, the discrimination result deviates from the actual intention because the energy peak appears in the upper shoulder rather than the outer shoulder; also, the activation range of the compensatory muscle group is diffuse, and the energy distribution spans multiple channels, making it impossible for traditional threshold methods to locate the effective activation center, resulting in lost commands.

[0042] Therefore, this scheme does not pre-define muscle identity in S200, but determines the muscle activation region based on the energy distribution of signals from each channel in the original electromyography dataset. Specifically, it calculates the absolute value of the sEMG signal integral or the RMS value of each channel within the action execution window, and defines the cluster of consecutive channels with the largest energy value as the current muscle activation region. This region reflects the surface projection location of the muscle group actually involved in the action, regardless of whether it is a typical anatomically oriented muscle.

[0043] Based on this, the activation timing relationship between channels is obtained according to the activation initiation time of the corresponding channels in the muscle activation region. The activation initiation time is determined by a dual threshold method: when the sEMG signal continuously exceeds three times the standard deviation of the baseline noise for a duration of more than 20 milliseconds, it is marked as activation initiation. Subsequently, the activation initiation times of each channel in the muscle activation region are compared to form timing relationships such as channel 3 earlier than channel 5, and channel 2 and channel 4 being synchronous. This relationship is determined by the neural drive sequence, and even if the muscle position changes, as long as the task is the same, the activation timing often remains stable.

[0044] Through the above processing, S200 transforms electromyographic signals from fixed-channel, fixed-muscle paradigms into dynamic, region-stable temporal feature expressions, providing a reliable, quantifiable, and physiologically reasonable input basis for subsequent instruction matching compatible with compensatory strategies.

[0045] In some embodiments, step S200 involves determining the muscle activation region based on the energy distribution of each channel signal in the original electromyography dataset, and obtaining the activation timing relationship between channels based on the activation start time of the corresponding channel in the muscle activation region, including steps S210 to S230: S210: Perform sliding window energy integration on the signals of each channel in the original electromyography dataset to obtain the electromyography energy sequence of each channel over time.

[0046] Specifically, the sliding window length is 50 milliseconds and the step size is 10 milliseconds. The absolute value of the integral of the surface electromyography signal of each channel is calculated within each window to form a time-aligned electromyographic energy sequence, which is used to characterize the intensity dynamics of muscle activity in the corresponding channel.

[0047] Furthermore, the sliding window energy integration is performed in parallel on all channels, and the output is a multi-channel electromyographic energy matrix, with rows corresponding to time frames and columns corresponding to channel numbers. The matrix element values ​​reflect the muscle activation intensity of the corresponding channel at the corresponding time frame.

[0048] Furthermore, the start time of the electromyographic energy sequence is aligned with the original electromyographic signal, ensuring that the temporal location of the activation event can be traced back to the original data, avoiding phase shifts introduced by processing delays.

[0049] It is easy to understand that the sliding window length and step size are set based on the typical activation duration of surface electromyography signals, so that the energy sequence can both capture transient activation initiation and smooth high-frequency noise.

[0050] Ideally, the sliding window energy integral uses the absolute value of the integral rather than the square value to avoid the excessive dominance of high-amplitude samples in energy estimation and to improve the sensitivity to weak compensatory muscle group signals.

[0051] Preferably, the electromyographic energy sequence is stored in the form of a floating-point array to preserve the dynamic range and facilitate accurate detection of rising edges and local maxima.

[0052] S220: Identify the set of channels corresponding to local maxima in electromyographic energy sequences, and cluster channels that are spatially continuous and whose local maxima occur in the same temporal neighborhood as muscle activation regions.

[0053] Specifically, for the electromyographic energy sequence of each channel, the energy value of the current frame is compared with the adjacent frames before and after. If the energy value of the current frame is greater than that of the adjacent frames, the current frame is marked as the local maximum energy point of the corresponding channel, and the time position and channel number of the local maximum energy point are recorded.

[0054] Furthermore, the energy local maxima of all channels are traversed, and points in spatially adjacent channels whose maxima time difference does not exceed a preset time neighborhood are grouped into the same cluster, with each cluster corresponding to a muscle activation region.

[0055] It should be noted that when analyzing electromyographic energy sequences, although there is a sequence of muscle activation in different channels, the time intervals are usually very short, generally within tens of milliseconds. This is because when the human body performs a continuous movement, the neural drive signals of the relevant muscle groups are fired almost continuously, and the activation process has physiological coordination and compactness.

[0056] It is easy to understand that the preset time neighborhood is determined based on the typical timescale of neuromuscular transmission and muscle group synergistic activation in the human body. The purpose is to ensure that the clustering results reflect real and coherent motor intentions, rather than scattered and irrelevant electrical activity. In actual implementation, this parameter can be calibrated and fixed using clinical data before the device leaves the factory, without requiring user adjustment. This application generally uses a 30-millisecond time neighborhood.

[0057] Furthermore, muscle activation regions are represented by channel number intervals, such as channel 3 to channel 5. The spatial continuity of muscle activation regions is predefined by the physical arrangement of the electrode array, ensuring that the clustering results conform to the anatomical continuity of muscle activation.

[0058] Ideally, the clustering process uses a connected component labeling algorithm to automatically merge channels that meet the spatiotemporal proximity condition into a single active region, avoiding the need to manually set the number of clusters.

[0059] Ideally, if the energy value of the boundary channel of the muscle activation region is lower than 50% of that of the channel inside the region, it should be pruned to ensure that the region focuses on the high activation core and improve the signal-to-noise ratio of the time series analysis.

[0060] S230: For each channel in the muscle activation region, the activation start time is determined based on the rising edge of the electromyographic energy sequence, and the activation sequence relationship between channels is generated according to the relative order of the activation start times of each channel.

[0061] Specifically, for the electromyographic energy sequence of each channel in the muscle activation area, the inflection point where the energy continuously increases is found by tracing back from the local maximum point. The inflection point is the starting position where the energy increases monotonically for multiple consecutive frames and the increase exceeds the preset ratio of the previous frame. In this application, the inflection point is generally defined as the energy increasing monotonically for three consecutive frames and the increase exceeds 10% of the previous frame.

[0062] Furthermore, the activation start times of each channel are sorted by numerical value to form a time sequence. For example, if the activation start time of channel 4 is 215 milliseconds and that of channel 5 is 230 milliseconds, then the time sequence relationship of channel 4 being earlier than that of channel 5 is generated.

[0063] Furthermore, the temporal relationships are stored in ordered pairs, such as <channel 4, channel 5>, which means that the former is activated first and can be directly called for template matching.

[0064] It is easy to understand that the determination of the activation start time depends on the dynamic trend of the energy sequence rather than a fixed amplitude threshold, which can adapt to individual differences in the intensity of electromyographic signals among different users.

[0065] Preferably, the electromyographic energy sequence is smoothed by first-order differential before the rising edge start point detection to suppress false inflection points caused by sampling jitter and improve the stability of the start time determination.

[0066] Ideally, if the activation start time difference between two or more channels within the muscle activation area is less than 5 milliseconds, it is considered synchronous activation, generating a synchronous temporal relationship between channel A and channel B, which more accurately reflects the characteristics of neural coordinated discharge.

[0067] S300: Combine the activation timing relationship between muscle activation areas and channels into an electromyographic feature combination, compare the electromyographic feature combination with multiple pre-stored navigation instruction templates, and obtain matching navigation control instructions; It should be noted that existing electromyographic navigation control methods typically map the amplitude or frequency domain energy of a single channel directly to commands, or rely on the user to use fixed muscles to complete standard movements. This leads to control logic failure when faced with upper limb paralysis patients who activate compensatory muscle groups due to nerve damage. For example, when a patient uses the trapezius muscle instead of the deltoid muscle to perform a left turn, the judgment result deviates from the true intention because the activation position deviates from the preset channel. Furthermore, different users may use different muscle combinations for the same navigation movement, while traditional methods only support one-to-one mapping, failing to accommodate individual differences.

[0068] Therefore, this solution introduces multidimensional electromyographic feature combinations and redundant template matching in S300. Specifically, the muscle activation regions output by S200 are combined with the activation timing relationships between channels to form a structured electromyographic feature combination. This combination is not bound to specific muscle names, but rather uses the spatial location and timing logic of the electrode channels to jointly represent the user's current navigation intent.

[0069] The navigation instruction templates are pre-built rule bases for multiple valid combinations. Each template corresponds to a basic navigation instruction, such as forward or left turn, and each template contains at least one set of allowed combinations of muscle activation regions and their timing relationships. For example, the left turn instruction could correspond to one template requiring activation regions in channels 3 to 5 with a timing relationship where channel 4 precedes channel 5; or it could correspond to another template requiring activation regions in channels 1 to 2 with a timing relationship where channels 1 and 2 are synchronized. The corresponding navigation control instruction is output as long as the current electromyographic feature combination matches any template.

[0070] This solution breaks away from the rigid paradigm of fixing muscles to fixing commands, instead employing a multi-path, effective, and fault-tolerant design. This ensures that even if the user activates atypical compensatory muscle groups, the activation pattern can be correctly identified as long as it falls within a preset set of valid combinations. The matching process is based on deterministic rule comparison, requiring no online learning or model updates.

[0071] In some embodiments, step S300 combines the activation timing relationship between muscle activation regions and channels into an electromyographic feature combination, compares the electromyographic feature combination with multiple pre-stored navigation instruction templates, and obtains a matching navigation control instruction, including steps S310 to S330:

[0072] Step S310: Construct an electromyographic feature set, which includes an ordered set of channel number intervals and temporal relationships of muscle activation regions.

[0073] Specifically, step S310 includes steps S311 and S312:

[0074] Step S311: Represent the muscle activation area as a continuous channel number interval, such as channel 3 to channel 5.

[0075] It should be noted that by traversing the physical arrangement of the electrode array, the clustering channels output by S220 are arranged in ascending order of their numbers, and the smallest and largest numbers are used to form a closed interval as a structured representation of the muscle activation region.

[0076] Step S312: Represent the activation timing relationship between channels as an ordered set of pairs, such as <channel 4, channel 5> indicating that channel 4 is earlier than channel 5.

[0077] It should be noted that when traversing the timing sequence output by S230, for each pair of channels with a sequential relationship, an ordered pair is generated in which the preceding channel points to the succeeding channel. Channels that are activated synchronously are not generated into ordered pairs.

[0078] Preferably, step S311 compresses the muscle activation region into a channel number interval to reduce storage overhead while preserving spatial continuity constraints.

[0079] Preferably, step S312 expresses the temporal relationship in the form of ordered pairs, which facilitates the logical matching of subsequent template rules.

[0080] Step S320: Load the pre-stored navigation instruction template library. Each navigation instruction template corresponds to a type of navigation control instruction and contains at least one set of rules for combining allowed channel number ranges and ordered pairs.

[0081] Specifically, step S320 includes steps S321 and S322:

[0082] Step S321: Preset multiple valid combination rules for each type of basic navigation instruction. Each rule defines a channel number range and an ordered set of conditions for joint conditions.

[0083] It should be noted that the basic navigation commands include five categories: forward, backward, left turn, right turn, and stop. The left turn command can correspond to Rule 1: the channel number range is 3 to 5 and the ordered pair set contains <4,5>; or it can correspond to Rule 2: the channel number range is 1 to 2 and the ordered pair set is empty (indicating synchronous activation).

[0084] Step S322: Organize all combination rules into a template library according to navigation instruction categories. The template library is indexed by instruction type, and each index item contains one or more combination rules.

[0085] It should be noted that the template library is loaded into the memory during device initialization for real-time matching and calling, and the rules are stored in a structured data format.

[0086] Preferably, step S321, by pre-setting multiple sets of effective combination rules for the same navigation command, naturally accommodates the compensation strategies of different users.

[0087] Preferably, the template library structure in step S322 supports fast indexing and parallel matching, improving instruction discrimination efficiency.

[0088] Step S330: Compare the electromyographic feature combinations with the navigation instruction template library item by item. If a matching combination rule exists, output the corresponding navigation control instruction.

[0089] Specifically, step S330 includes steps S331 and S332: Step S321: Preset multiple valid combination rules for each type of basic navigation instruction. Each rule defines a channel number range and an ordered set of conditions for joint conditions.

[0090] It should be noted that the following conditions must be met to determine the matching of channel number intervals: the interval of the electromyographic feature combination must be completely contained within the regular interval; the following conditions must be met to determine the matching of ordered pairs: the ordered pair set of electromyographic feature combinations must be a subset of the regular ordered pair set.

[0091] Step S332: If at least one combination rule is satisfied, the navigation instruction category to which this combination rule belongs is taken as the matching result, and the corresponding navigation control instruction is output.

[0092] It should be noted that if multiple rules match simultaneously, the rule with the highest overlap between the channel number interval and the electromyographic feature combination interval will be selected first.

[0093] Furthermore, step S331 determines whether the combination of electromyographic features satisfies the combination rules, including steps A1 and A2: Step A1: Determine whether the channel numbering interval of the electromyographic feature combination is covered by the channel numbering interval of the combination rule.

[0094] Specifically, let the electromyographic feature combination interval be C. start C end The regular interval is R. start ,R end If R start ≤C start And C end ≤R end If so, then the coverage is deemed valid.

[0095] It's important to note that when the human body performs the same type of navigation movement, such as turning the shoulder to the left, the actual muscle positions involved in the effort may shift slightly due to postural adjustments, fatigue levels, or compensatory habits. Today, the middle trapezius muscle might be activated, but tomorrow it might be more pronounced in the upper part, corresponding to electrode channels changing from 4-5 to 3-4. If the exact same channels were required to be activated each time, the system would easily misjudge. Therefore, this invention uses a range-coverage approach; as long as the user's activation area falls entirely within the channel range allowed by the template, it is considered effective. This preserves the constraints on muscle spatial distribution while allowing users natural freedom of movement, making the control more realistic.

[0096] Step A2: Determine whether the ordered set of electromyographic feature combinations is a subset of the ordered set of combination rules.

[0097] Specifically, each ordered pair in the electromyographic feature combination is traversed. If all ordered pairs exist in the regular set of ordered pairs, then the subset is determined to be independent.

[0098] It's important to note that a complete navigation action typically involves multiple muscles working in a specific sequence, such as activating the trapezius first, followed by the deltoid. However, in practice, users may only fully trigger a portion of the temporal relationship due to fast movement, distraction, or muscle weakness. For example, the template might define two ordered pairs, <3,4> and <4,5>, but the user might only clearly express <4,5>. If a complete match is required, such a valid action would be rejected. This invention employs subset judgment logic; as long as the temporal relationship expressed by the user does not exceed the template's allowed range, even if only a portion is activated, it is considered a successful match. This design respects the complete differences in user action expression, avoids the rejection of valid commands due to missing details, and improves the smoothness and fault tolerance of daily use.

[0099] Preferably, step A1 uses interval inclusion determination instead of exact equality, allowing the user's active region to shift slightly within the template range, thus improving robustness.

[0100] Preferably, step A2 uses subset matching instead of perfect matching, allowing users to activate only some temporal features while still hitting templates containing more ordered pairs.

[0101] Preferably, steps S310 to S330 achieve highly fault-tolerant recognition of compensatory electromyographic patterns by using structured expression of electromyographic features, pre-setting multi-path effective templates, and employing interval coverage and subset matching rules.

[0102] Preferably, based on the above sub-steps, S300 solves the mismatch problem of traditional one-to-one mapping when facing atypical muscle activation, providing technical support for the stable operation of the navigation brain-muscle interface in real clinical scenarios.

[0103] S400: Determines whether there are any abnormal physiological events based on the original EEG dataset, and controls the output state of navigation control commands based on the determination results.

[0104] In some embodiments, step S400 involves determining whether a physiological abnormality event exists based on the original EEG dataset and controlling the output state of navigation control commands according to the determination result, including steps S410 to S430:

[0105] Step S410: Perform frequency band power analysis on the original EEG dataset to obtain the power distribution of each frequency band on each channel.

[0106] Specifically, step S410 includes steps S411 and S412: Step S411: Divide the original EEG dataset into consecutive frames according to time windows, with each frame lasting 500 milliseconds.

[0107] It should be noted that the time window slides in a non-overlapping or 50% overlap manner, covering all EEG data periods recorded in S100, ensuring complete coverage of the EEG state during the execution of navigation actions.

[0108] Step S412: Perform Fourier transform on each frame of EEG signal and calculate the average power of the four frequency bands: delta wave, theta wave, alpha wave, and beta wave. The frequency range of delta wave is 0.5 to 4 Hz, the theta wave is 4 to 7 Hz, the alpha wave is 8 to 13 Hz, and the beta wave is 14 to 30 Hz.

[0109] It should be noted that power calculations were performed separately at the frontal central lead and all leads to differentiate between focal and generalized abnormalities.

[0110] Preferably, step S411 uses a 500-millisecond time window, which balances the timeliness of event detection with frequency domain resolution.

[0111] Preferably, step S412 defines four standard frequency bands to avoid introducing non-physiologically relevant frequency bands and ensure that the power analysis results are clinically interpretable.

[0112] Step S420: Identify two types of abnormal physiological events based on power distribution and amplitude timing.

[0113] Specifically, step S420 includes steps S421 and S422: Step S421: Detect sudden high-amplitude slow wave events. The judgment condition is: the peak amplitude of any central channel exceeds 100 microvolts and the duration is greater than 300 milliseconds within two consecutive frames.

[0114] Specifically, the peak amplitude is determined by the maximum absolute value of the signal within each frame, with two consecutive frames corresponding to at least 600 milliseconds of sustained high amplitude within one second, excluding instantaneous interference.

[0115] It should be noted that the central frontal region is the most common origin of high-amplitude slow waves during epileptiform discharges or myoclonic seizures, and its signal characteristics have clear clinical significance. Limiting the judgment criteria to this region avoids false triggering caused by occipital alpha rhythms or temporal lobe artifacts. Furthermore, requiring the high-amplitude state to last for more than 300 milliseconds is because true pathological slow waves typically have a duration of hundreds of milliseconds or more, while transient noises such as electromyographic interference and electrode loosening often only last for tens of milliseconds. By combining spatial localization and temporal duration constraints, the specificity of event discrimination is effectively improved.

[0116] Step S422: Detect widespread rhythm disturbance events, the criteria for which are: the average power of the alpha wave in all leads decreases by more than 80% from the baseline and the average power of the theta wave increases by more than 50%, lasting for at least two frames.

[0117] Specifically, the baseline power is taken from the first 10 seconds of EEG data of the user in the S100 at rest to ensure individualized reference.

[0118] It's important to note that alpha rhythms are the dominant EEG rhythms in a conscious, closed-eye state; a sudden drop in their power typically reflects a decline in consciousness or severe inattention. Theta rhythms, on the other hand, have lower power in a conscious state, and abnormal increases in their power are often accompanied by cognitive overload or fatigue breakdown. When both rhythms simultaneously reverse direction and cover all leads, it indicates widespread instability in the overall brain function, rather than localized interference. Using the user's own resting-state data as a baseline accommodates individual differences across age and EEG backgrounds, avoiding misjudgments caused by uniform thresholds. Furthermore, requiring the changing state to persist for at least two frames is to eliminate transient power fluctuations caused by brief artifacts such as blinking and swallowing, ensuring the physiological continuity of the identified rhythm disturbances.

[0119] Preferably, step S421 focuses on high-amplitude slow waves in the central frontal region to specifically capture focal abnormalities associated with motor spasticity or epileptic aura.

[0120] Preferably, step S422 effectively identifies pervasive EEG disturbances caused by decreased level of consciousness or cognitive breakdown through a combination criterion of alpha wave suppression and theta wave enhancement.

[0121] Step S430: Based on the identification results of physiological abnormal events, control the output state of navigation control commands.

[0122] Specifically, step S430 includes steps S431 and S432: Step S431: If a sudden high-amplitude slow wave event or a widespread rhythm disorder event is detected, the current navigation control command to be output is interrupted, and the device control state is set to standby.

[0123] It should be noted that the interrupt operation is completed within 100 milliseconds after the event is confirmed, ensuring a timely safety response.

[0124] Step S432: If no abnormal physiological event is detected, the current navigation control command is allowed to be sent to the navigation execution device.

[0125] It should be noted that the matching and discrimination of S300 must be completed before the instruction is sent, and S430 is only used as the final output gate.

[0126] Preferably, step S431 can trigger a vibration motor or buzzer to alert the user while interrupting the command, thereby improving the safety of human-computer interaction.

[0127] Preferably, step S432 ensures that navigation actions are performed only when the physiological state is stable, avoiding safety interruptions caused by brief artifacts and maintaining control continuity.

[0128] Preferably, steps S410 to S430 construct a lightweight, low-latency, and highly reliable physiological safety monitoring process through frequency division power analysis, dual-type abnormal event discrimination, and output gate control.

[0129] Ideally, the S400 does not rely on models or external inputs, but is based entirely on the objective characteristics of raw EEG signals, providing the necessary clinical safety assurance for navigation brain-muscle interfaces.

[0130] It should be understood that existing brain-muscle interface navigation control methods typically do not perform safety checks on the user's current physiological state when outputting electromyography (EMG) results. When a user experiences sudden myoclonus, epileptic-like discharges, or strong emotional fluctuations, the EEG signals may exhibit abnormal characteristics such as high amplitude, wide frequency range, or rhythmic irregularities. If navigation control commands are still output under these circumstances, it may cause the wheelchair to spin suddenly, accelerate, or stop unnecessarily, leading to collisions or falls. This is especially true for patients with high-level spinal cord injuries or amyotrophic lateral sclerosis (ALS), whose ability to autonomously intervene in the event of device malfunction is limited, making the safety risks even more pronounced.

[0131] Therefore, this solution introduces real-time identification and command output linkage control of physiological abnormal events based on raw EEG datasets in the S400. Specifically, without relying on external sensors or user feedback, it directly analyzes the raw EEG dataset synchronously acquired by the S100 to detect the presence of two typical physiological abnormal events. The first is a sudden high-amplitude slow-wave event, characterized by an amplitude exceeding 100 microvolts within 200 milliseconds and a duration greater than 300 milliseconds in the frontal central channel, commonly seen in epileptic aura or severe spasms. The second is widespread rhythm disturbance, characterized by a sudden drop in alpha rhythm power exceeding 80% across all leads, accompanied by an abnormal increase in theta wave power, where the alpha rhythm frequency ranges from 8 to 13 Hz and the theta wave frequency ranges from 4 to 7 Hz; this phenomenon is commonly seen in confusion or fatigue collapse.

[0132] Upon detection of any of the aforementioned events, the currently pending navigation control command is immediately interrupted, and the device control state is switched to standby mode. Local tactile or auditory cues may be triggered simultaneously. If no physiological abnormality is detected, navigation control commands are allowed to be sent normally to the navigation execution device. This processing is entirely based on the characteristics of EEG signals themselves, with criteria including clearly defined amplitude, duration, and frequency domain power change thresholds. The innovation lies in making physiological safety assessment a necessary prerequisite for navigation command output, forming a three-stage control logic of intent assessment, safety verification, and command execution, significantly improving the reliability and user trust of the neuromuscular interface in real clinical environments.

[0133] Preferably, this application completes signal acquisition, feature analysis, command matching, and security control entirely on the local device, without uploading the user's electromyography (EMG) or electroencephalography (EEG) data to a remote server, nor relying on network connections or cloud computing. In this way, the user's physiological data remains within the personal device, effectively protecting the privacy and security of sensitive health information. Simultaneously, all processing steps are completed within milliseconds, ensuring rapid response to navigation commands and preventing lag due to network latency or server load. More importantly, even in environments without network access at home, outdoors, or with limited power, the entire system can still operate stably, truly achieving offline availability and on-demand operation, allowing users with upper limb dysfunction to reliably control mobility aids in various life scenarios.

[0134] It should be noted that although the steps in the flowcharts of the above embodiments are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise expressly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the above embodiments may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0135] Based on the same inventive concept, this application also provides a navigation neuromuscular interface control system. The solution provided by this system is similar to the solution described in the above method; therefore, the specific limitations of one or more embodiments of the navigation neuromuscular interface control system provided below can be found in the limitations of the navigation neuromuscular interface control method described above, and will not be repeated here.

[0136] In one exemplary embodiment, such as Figure 3 As shown, a navigation brain-muscle interface control system is provided, comprising: The signal acquisition module controls the electromyography acquisition device to acquire multi-channel surface electromyography signals and synchronous electroencephalogram (EEG) signals when the target person performs navigation actions, and obtains raw electromyography datasets and raw EEG datasets covering the shoulder and neck to upper arm region. The feature extraction module determines the muscle activation region based on the energy distribution of each channel signal in the original electromyography dataset, and obtains the activation timing relationship between channels based on the activation start time of the corresponding channel in the muscle activation region. The instruction matching module combines the activation sequence relationship between muscle activation areas and channels into electromyographic feature combinations, compares the electromyographic feature combinations with multiple pre-stored navigation instruction templates, and obtains matching navigation control instructions. The safety control module determines whether there are any abnormal physiological events based on the original EEG dataset, and controls the output state of navigation control commands according to the judgment result.

[0137] The modules in the aforementioned navigation neuromuscular interface control system can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in the processor of a computer device in hardware form or independent of it, or stored in the memory of the computer device in software form, so that the processor can call and execute the corresponding operations of each module.

[0138] In one exemplary embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be as follows: Figure 4 As shown, the computer device includes a processor, memory, input / output interface, communication interface, display unit, and input device. The processor, memory, and input / output interface are connected via a system bus, and the communication interface, display unit, and input device are also connected to the system bus via the input / output interface. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The input / output interface is used for exchanging information between the processor and external devices. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, mobile cellular networks, Near Field Communication (NFC), or other technologies. When the computer program is executed by the processor, it implements a navigation neuromuscular interface control method. The display unit is used to form a visually visible image and can be a display screen, projection device, or virtual reality imaging device. The display screen can be an LCD screen or an e-ink screen. The input device of the computer device can be a touch layer covering the display screen, or buttons, trackballs, or touchpads set on the casing of the computer device, or external keyboards, touchpads, or mice, etc.

[0139] Those skilled in the art will understand that Figure 4 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0140] In one embodiment, a computer device is also provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above method embodiments.

[0141] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon that, when executed by a processor, implements the steps in the above method embodiments.

[0142] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above method embodiments.

[0143] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0144] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, artificial intelligence (AI) processors, etc., and are not limited to these.

[0145] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0146] The above embodiments are merely illustrative of several implementation methods of this application, and their descriptions are relatively specific and detailed. However, they should not be construed as limiting the scope of this application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A method of navigating a brain-muscle interface control, the method comprising: include: The electromyography (EMG) acquisition device is controlled to collect multi-channel surface EMG signals and synchronous EEG signals when the target person performs navigation actions, so as to obtain raw EMG datasets and raw EEG datasets covering the shoulder and neck to upper arm region. The muscle activation region is determined based on the energy distribution of each channel signal in the original electromyography dataset, and the activation timing relationship between channels is obtained based on the activation start time of the corresponding channel in the muscle activation region. The activation sequence relationship between the muscle activation region and the channel is combined into an electromyographic feature combination. The electromyographic feature combination is compared with a pre-stored navigation instruction template to obtain a matching navigation control instruction. The system determines whether any abnormal physiological events exist based on the original EEG dataset, and controls the output state of the navigation control commands based on the determination result.

2. The navigation brain-muscle interface control method as described in claim 1, characterized in that: The multi-channel surface electromyography (EMG) signals and synchronous electroencephalogram (EEG) signals collected when the target personnel perform navigation actions include: A multi-channel electrode array was deployed in the shoulder and neck area to the upper arm region of the target personnel. Guide the target personnel to perform navigation actions; Record multi-channel surface electromyography (EMG) and electroencephalography (EEG) signals during the execution of the navigation action to generate time-aligned raw EMG and raw EEG datasets; The original electromyography dataset was subjected to baseline correction and power frequency notch processing, and the original electroencephalography dataset was subjected to electrooculography artifact removal processing.

3. The method of claim 2, wherein: The obtained inter-channel activation timing relationship includes: The energy of each channel in the original electromyography dataset is integrated by a sliding window to obtain the electromyography energy sequence of each channel over time. In the electromyographic energy sequence, the channel set corresponding to the local energy maxima is identified, and channels that are spatially continuous and whose local energy maxima occur in the same temporal neighborhood are clustered as muscle activation regions. For each channel in the muscle activation region, the activation start time is determined based on the rising edge of the electromyographic energy sequence, and the activation sequence relationship between channels is generated according to the relative order of the activation start times of each channel.

4. The method of claim 3, wherein: The set of channels corresponding to the identified local maxima of energy includes: Traverse the electromyographic energy sequence of each channel and mark the points with energy values ​​greater than their adjacent time frames as local maxima. Channels that are spatially adjacent and whose local maxima time difference does not exceed a preset time neighborhood are grouped into the same cluster; For each channel in the muscle activation region, the activation initiation time is determined based on the rising edge of the electromyographic energy sequence, including: By tracing back from the local maximum point, the inflection point where the energy continues to rise is found as the activation start time. The inflection point is the starting position where the energy increases monotonically and the increase exceeds the preset ratio of the previous frame.

5. The method of navigating a brain-machine interface of claim 4, wherein: The comparison of the electromyographic feature combination with the pre-stored navigation instruction template includes: The muscle activation region is represented as a continuous channel number interval, and the activation sequence relationship between the channels is represented as an ordered set of pairs; Load the navigation instruction template library. Each navigation instruction template corresponds to a type of navigation control instruction and contains at least one set of rules for combining allowed channel number ranges and ordered pairs. Traverse the navigation instruction template library and determine whether the electromyographic feature combination satisfies any combination rule.

6. The method of navigating a brain-machine interface of claim 5, wherein: Determining whether the electromyographic feature combination satisfies any combination rule includes: Determine whether the channel number range of the electromyographic feature combination is covered by the channel number range defined by the combination rule; Determine whether the ordered set of the electromyographic feature combinations is a subset of the ordered set of combinations defined by the combination rule.

7. The method of claim 1, wherein: The determination of whether a physiological abnormality event exists includes: Frequency band power analysis was performed on the original EEG dataset to obtain the power distribution of multiple standard frequency bands in each channel; Physiological abnormal events are identified based on the power distribution and signal amplitude timing, including sudden high-amplitude slow wave events or widespread rhythm disorder events. If a physiological abnormality is detected, the current navigation control command to be output is interrupted and the device control state is set to standby. If no abnormal physiological event is detected, the navigation control command is allowed to be sent to the navigation execution device.

8. The method of navigating a brain-machine interface of claim 7, wherein: The identification of abnormal physiological events includes: When the signal amplitude of the frontal central channel exceeds the upper limit of the EEG signal amplitude of the user in the resting state, and the duration of the high amplitude state exceeds the upper limit of the duration of the high amplitude EEG of the user in the resting state, it is determined to be a sudden high amplitude slow wave event. When the power of the alpha rhythm in all leads is lower than the user's baseline alpha rhythm power at rest, and the power of the theta rhythm is higher than the user's baseline theta rhythm power at rest, it is considered a pervasive rhythm disorder event.

9. The navigation brain-muscle interface control method as described in claim 6, characterized in that: When the electromyographic feature combination satisfies multiple combination rules, the navigation control command corresponding to the combination rule with the highest overlap between the channel number interval and the electromyographic feature combination interval is selected as the output result.

10. A navigation brain-muscle interface control system, employing the navigation brain-muscle interface control method as described in any one of claims 1 to 9, characterized in that, include: The signal acquisition module controls the electromyography acquisition device to acquire multi-channel surface electromyography signals and synchronous electroencephalogram (EEG) signals when the target person performs navigation actions, and obtains raw electromyography datasets and raw EEG datasets covering the shoulder and neck to upper arm region. The feature extraction module determines the muscle activation region based on the energy distribution of each channel signal in the original electromyography dataset, and obtains the activation timing relationship between channels based on the activation start time of the corresponding channel in the muscle activation region. The instruction matching module combines the activation timing relationship between the muscle activation region and the channel into an electromyographic feature combination, and compares the electromyographic feature combination with multiple pre-stored navigation instruction templates to obtain a matching navigation control instruction. The safety control module determines whether there are any abnormal physiological events based on the original EEG dataset, and controls the output state of the navigation control commands based on the determination result.

Images