Wheelchair control method and device and wheelchair

Abstract

The application discloses a wheelchair control method and device and a wheelchair, and relates to the field of wheelchair control, and comprises the following steps: acquiring a three-axis angular velocity signal and a three-axis acceleration signal generated according to head movement to identify a head action; extracting power spectrum densities of multiple frequency components of brain waves, and determining a concentration score according to the power spectrum densities; generating a wheelchair control instruction according to the head action, and controlling the wheelchair according to the wheelchair control instruction when the concentration score meets a control condition. The combination control of forehead brain electrical signals and head movement signals improves the accuracy and reliability of the instruction. The concentration score is calculated and used as a control condition, the wheelchair is prevented from responding when the user is not concentrating, the misoperation caused by fatigue and distraction is prevented, and the safety of the wheelchair use is improved.

Description

Technical Field

[0001] This invention relates to the field of wheelchair control, and in particular to a wheelchair control method, device, and wheelchair. Background Technology

[0002] Neurological diseases such as stroke often lead to limb motor dysfunction, and intelligent electric wheelchairs can improve patients' quality of life. Current intelligent wheelchair controls primarily use a single biosignal as the control source, such as head movement signals, resulting in a relatively limited control mode. Furthermore, they cannot effectively recognize the user's attention state; when the user is fatigued, inattentive, or in a poor mental state, the wheelchair may still respond to unconscious head movements, posing a safety hazard. Summary of the Invention

[0003] The purpose of this invention is to provide a wheelchair control method, device, and wheelchair. The combined control of prefrontal cortex electroencephalogram (EEG) signals and head movement signals improves the accuracy and reliability of commands. Calculating a focus score and using it as a control condition prevents the wheelchair from responding when the user is not paying attention, thus preventing misoperation due to fatigue or distraction and enhancing the safety of wheelchair use.

[0004] To solve the above-mentioned technical problems, the present invention provides a wheelchair control method, comprising:

[0005] Acquire prefrontal cortex EEG signals and triaxial angular velocity and triaxial acceleration signals generated by head movements, wherein the prefrontal cortex EEG signals include brain waves;

[0006] Head movements are identified based on the triaxial angular velocity signal and the triaxial acceleration signal, and the head movements include head tilting, left and right head turning, and head returning to center.

[0007] The power spectral density of multiple frequency components of the brainwave is extracted, and a focus score is determined based on the power spectral density. The focus score is positively correlated with the user's level of focus.

[0008] Wheelchair control commands are generated based on the head movements, and the wheelchair is controlled according to the wheelchair control commands when the attention score meets the control conditions.

[0009] On the other hand, identifying head movements based on the triaxial angular velocity signal and the triaxial acceleration signal includes:

[0010] Integrate the Z-axis angular velocity signal to obtain the cumulative Z-axis angle;

[0011] The absolute value of the cumulative Z-axis angle is compared with the head rotation threshold.

[0012] When the absolute value exceeds the head turning action threshold, the head action is determined to be a left head turning action or a right head turning action based on the direction of the cumulative angle of the Z-axis.

[0013] When the absolute value of the cumulative Z-axis angle gradually decreases from exceeding the head rotation threshold to less than the head rotation threshold, and the Z-axis angular velocity signal changes to within the zero threshold range, the head movement is determined to be a return-to-center movement.

[0014] Extract the instantaneous acceleration in the vertical direction from the Y-axis acceleration signal;

[0015] When the change in instantaneous acceleration exceeds the head-raising acceleration threshold, the head movement is determined to be a head-raising movement.

[0016] On the other hand, after identifying head movements based on the triaxial angular velocity signal and the triaxial acceleration signal, the process further includes:

[0017] When a head-raising or head-straightening motion is detected, or when no head movement exceeding a preset amplitude is detected within a preset time period, the static three-axis angular velocity signal and three-axis acceleration signal within the preset time period are acquired.

[0018] A new zero-point correction value is determined based on the static triaxial angular velocity signal and triaxial acceleration signal within the preset time period, and the current zero-point correction value is replaced with the new zero-point correction value. At the same time, the cumulative angle of the Z-axis is reset to zero.

[0019] On the other hand, the prefrontal cortex EEG signals also include blinking signals, and further include:

[0020] Multiple threshold levels are preset, each threshold level corresponds to a set of amplitude difference thresholds and time difference thresholds, and the threshold level is positively correlated with the amplitude difference thresholds and the time difference thresholds;

[0021] Determine the current threshold level, and within a preset calibration time, determine the blink signal from the prefrontal EEG signal at the current threshold level;

[0022] Determine whether a blink signal exists that simultaneously satisfies the amplitude difference threshold and the time difference threshold corresponding to the current threshold level;

[0023] If no blink signal matching the current threshold level is detected within the calibration period, the threshold level will be lowered to the next level.

[0024] If a blink signal that meets the current threshold level is detected within the calibration time, the amplitude difference threshold and time difference threshold corresponding to the current threshold level are used as the reference threshold for blink detection, and a valid blink signal is used as a confirmation instruction.

[0025] On the other hand, the power spectral density of multiple frequency components of the brainwave is extracted, and a focus score is determined based on the power spectral density, including:

[0026] The DC component in the prefrontal cortex EEG signal was removed and bandpass filtered.

[0027] The theta wave, alpha wave, and beta wave with sequentially increasing frequencies, and their corresponding power spectral densities were separated from the filtered prefrontal EEG signal.

[0028] The focus score is determined based on the corresponding power spectral density, and the expression for the focus score is:

[0029] ;

[0030] Where F represents the focus score, For the power spectral density of the beta wave, The power spectral density of the theta wave, Let be the power spectral density of the alpha wave.

[0031] On the other hand, after determining the focus score based on the power spectral density, the process includes:

[0032] Determine the moving average of the focus score, the expression for which is:

[0033] ;

[0034] in, The moving average at the current time. This is the moving average value from the previous time step. The focus score calculated at the current moment. and For the preset weighting coefficients, and ;

[0035] The moving average value at the current moment is compared with a preset attention threshold. When the moving average value at the current moment is lower than the attention threshold, it is determined that there is insufficient attention.

[0036] Set the wheelchair drive enable signal to invalid, preventing the wheelchair from starting or stopping the currently executing wheelchair control command;

[0037] If the wheelchair is in motion, it will trigger a deceleration stop and generate a voice prompt.

[0038] On the other hand, wheelchair control commands are generated based on the head movements, including:

[0039] A turning command is generated based on the left and right head turning actions, and the duration of the turning command is positively correlated with the holding time of the left and right head turning angles;

[0040] An emergency stop command is generated based on the head-raising motion; the emergency stop command is used to control the wheelchair to decelerate and stop immediately.

[0041] A straight-ahead command is generated based on the return-to-center action. The straight-ahead command is used to cause the wheelchair to exit the turning state and resume straight-ahead travel.

[0042] On the other hand, it also includes:

[0043] Set a dropout threshold, which is lower than the attention threshold corresponding to the control condition;

[0044] When the focus score is lower than the detachment threshold, it is determined that the electrode has detached.

[0045] The user is prompted to check the electrode connections;

[0046] Keep wheelchair control commands disabled until the electrode detachment fault is resolved.

[0047] To address the aforementioned technical problems, the present invention also provides a wheelchair control device, comprising:

[0048] Memory, used to store computer programs;

[0049] A processor is used to implement the steps of the above-described wheelchair control method when executing the computer program.

[0050] To solve the above-mentioned technical problems, the present invention also provides a wheelchair, including the aforementioned wheelchair control device.

[0051] This application provides a wheelchair control method, device, and wheelchair, relating to the field of wheelchair control. The method includes acquiring triaxial angular velocity and triaxial acceleration signals generated based on head movements to identify head actions; extracting the power spectral density of multiple frequency components of the electroencephalogram (EEG) and determining an attention score based on the power spectral density; generating wheelchair control commands based on head movements; and controlling the wheelchair according to the commands when the attention score meets the control conditions. The combined control of prefrontal cortex EEG signals and head movement signals improves the accuracy and reliability of the commands. Calculating the attention score and using it as a control condition avoids wheelchair response when the user's attention is not focused, preventing misoperation due to fatigue or distraction, and improving the safety of wheelchair use. Attached Figure Description

[0052] To more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings used in the prior art and embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0053] Figure 1 A flowchart of a wheelchair control method provided by the present invention;

[0054] Figure 2 A schematic diagram of the structure of a wheelchair provided by the present invention;

[0055] Figure 3 A functional schematic diagram of a wheelchair provided by the present invention;

[0056] Figure 4 A schematic diagram of shortcut commands for a wheelchair provided by the present invention;

[0057] Figure 5 A schematic diagram of a mouse control interface for a wheelchair provided by the present invention;

[0058] Figure 6 A schematic diagram of a command interface for a wheelchair provided by the present invention;

[0059] Figure 7 A schematic diagram of a wheelchair's mobile interface provided by the present invention;

[0060] Figure 8 A calibration diagram of a wheelchair provided by the present invention;

[0061] Figure 9 This is a schematic diagram of the structure of a wheelchair control device provided by the present invention. Detailed Implementation

[0062] The core of this invention is to provide a wheelchair control method, device, and wheelchair. The combined control of prefrontal cortex EEG signals and head movement signals improves the accuracy and reliability of commands. Calculating a focus score and using it as a control condition prevents the wheelchair from responding when the user is not paying attention, thus preventing misoperation due to fatigue or distraction and enhancing the safety of wheelchair use.

[0063] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0064] Figure 1 A flowchart of a wheelchair control method provided by the present invention, the wheelchair control method comprising:

[0065] S11: Acquire prefrontal EEG signals and triaxial angular velocity and triaxial acceleration signals generated by head movements. Prefrontal EEG signals include brain waves.

[0066] Existing single signal sources cannot simultaneously obtain the user's movement intention, cognitive state, and active confirmation commands, leading to high error rates or failure to identify safety hazards such as insufficient attention. Therefore, this application simultaneously collects two types of complementary signals: triaxial angular velocity and acceleration signals from the inertial measurement unit are used to analyze head posture, and electroencephalogram (EEG) signals from dual-lead electrodes in the prefrontal cortex are used to extract brain rhythms and electrooculogram artifacts.

[0067] like Figure 2 As shown, the head-mounted data acquisition terminal includes an inertial measurement unit and dual-lead EEG electrodes, a mobile processing unit (which can be a smartphone, tablet, or wheelchair control unit, i.e., the processor in this application), and the head-mounted data acquisition terminal and the mobile processing unit transmit data via Bluetooth. The mobile processing unit and the wheelchair control unit communicate via serial port.

[0068] In the head-mounted data acquisition terminal, the inertial measurement unit outputs triaxial angular velocity and triaxial acceleration at a fixed sampling rate. Dual-lead EEG electrodes are placed at positions FP1 and FP2 in the prefrontal cortex, with a reference electrode placed behind the ear or in the center of the forehead. The mobile processing unit receives the data stream in real time via Bluetooth. It should be noted that prefrontal EEG signals are essentially mixed signals, containing both rhythmic waves generated by the cerebral cortex and eye movements; the electrooculogram (EOG) signal generated by blinking can be extracted as a blinking event.

[0069] S12: Identify head movements based on triaxial angular velocity and triaxial acceleration signals. Head movements include head tilting, left and right head turning, and head returning to center.

[0070] Head posture control is the most intuitive way to operate a wheelchair, but traditional methods are prone to misinterpreting complex movements, such as turning the head while slightly raising it, as a single command, or residual steering after returning to center due to integral drift. This application analyzes three basic head movements: head up / down, left / right turn, and returning to center, avoiding command crosstalk and providing clean discrimination results for subsequent generation of turning, emergency stop, and straight-line commands.

[0071] Furthermore, this application distinguishes between left turn, right turn, straightening, and head tilt when recognizing head movements, avoiding false triggers caused by slight head swaying or complex movements. For example, if a user turns their head to observe the side and then straightens it, the system can accurately output a steering command and automatically resume straight driving after straightening.

[0072] S13: Extract the power spectral density of multiple frequency components of the brainwave and determine the focus score based on the power spectral density. The focus score is positively correlated with the user's level of focus.

[0073] In rehabilitation assistance scenarios, users may experience decreased attention due to fatigue, distraction, or illness. Allowing head movements to control the wheelchair under these circumstances greatly increases the risk of collisions or falls. Attention score is positively correlated with the user's level of focus; a higher score indicates a more focused user, making them more suitable for actively controlling the wheelchair.

[0074] The focus score determined by the power spectral density can stably and sensitively reflect the real-time changes in the user's attention. When the user is focused, the focus score remains high, while when the user is tired, closes their eyes, or is distracted, the focus score drops significantly.

[0075] S14: Generate wheelchair control commands based on head movements, and control the wheelchair according to the wheelchair control commands when the attention score meets the control conditions.

[0076] When controlling wheelchair movement, there must be clear head movements reflecting the user's intentions, while simultaneously maintaining a focus score within a safe range. This resolves the issue of accidental triggering and ensures that the user can only operate the wheelchair in a conscious and highly attentive state, thus improving safety. The conditions for wheelchair control are described below from two aspects:

[0077] Firstly, when a left or right head turn is detected, a turning command is generated. The duration of the turning command is positively correlated with the duration of the head turn. When a straightening motion is detected, a straight-ahead command is generated, causing the wheelchair to exit the turning state and resume straight-line travel. When a head-up motion is detected, an emergency stop command is generated, controlling the wheelchair to immediately decelerate and stop. This command has the highest priority and is unaffected by other conditions; even if attention is temporarily insufficient, an emergency stop will be executed to ensure safety.

[0078] Secondly, the control conditions are considered met only when the attention score exceeds the attention threshold; otherwise, an insufficient attention flag is output. If attention is insufficient, the wheelchair drive enable signal is disabled, and the wheelchair will not start even with head movements and blink confirmation. Only when both the blink confirmation command has been detected and the attention score meets the control conditions are wheelchair control commands generated based on head movements sent to the wheelchair control unit. These commands include, but are not limited to, forward, backward, turning, straight-line, and stop. Otherwise, the commands are intercepted or replaced by an emergency stop, ensuring that every wheelchair movement is confirmed by the user.

[0079] It should also be noted that after the brain-computer interface device is turned on, it will automatically perform gyroscope calibration and blink threshold calibration in sequence under voice and interface prompts. After calibration, head movements can control the cursor and blinking can achieve the left mouse button click function. The gyroscope calibration method is to prompt the user to look at the red cursor in the center of the screen for a preset time. The algorithm will calculate the average value of the gyroscope's three-axis acceleration and angular velocity within the preset time and set it to zero.

[0080] This application provides a wheelchair control method, relating to the field of wheelchair control. The method includes acquiring triaxial angular velocity and triaxial acceleration signals generated based on head movements to identify head actions; extracting the power spectral density of multiple frequency components of the electroencephalogram (EEG) and determining an attention score based on the power spectral density; generating wheelchair control commands based on head movements; and controlling the wheelchair according to the commands when the attention score meets the control conditions. The combined control of prefrontal cortex EEG signals and head movement signals improves the accuracy and reliability of the commands. Calculating the attention score and using it as a control condition avoids wheelchair response when the user's attention is not focused, preventing misoperation due to fatigue or distraction, and improving the safety of wheelchair use.

[0081] Based on the above embodiments:

[0082] In some embodiments, head movements are identified based on triaxial angular velocity signals and triaxial acceleration signals, including:

[0083] Integrate the Z-axis angular velocity signal to obtain the cumulative Z-axis angle;

[0084] Compare the absolute value of the cumulative angle along the Z-axis with the head rotation threshold;

[0085] When the absolute value exceeds the head turning action threshold, the head action is determined to be either a left head turning action or a right head turning action based on the direction of the cumulative angle along the Z-axis.

[0086] When the absolute value of the cumulative Z-axis angle gradually decreases from exceeding the head rotation threshold to less than the head rotation threshold, and the Z-axis angular velocity signal changes to within the zero threshold range, the head movement is determined to be a return-to-center movement.

[0087] Extract the instantaneous acceleration in the vertical direction from the Y-axis acceleration signal;

[0088] When the change in instantaneous acceleration exceeds the head-raising acceleration threshold, the head movement is determined to be a head-raising movement.

[0089] The Z-axis angular velocity signal is numerically integrated to obtain the cumulative Z-axis angle, which represents the cumulative angle of head rotation around the vertical axis. To prevent false triggering caused by noise, the integration result can be smoothed and filtered. The absolute value of the cumulative Z-axis angle is compared with a preset head-turning action threshold (e.g., 10°~15°). If the absolute value exceeds the threshold, the direction of rotation is determined as clockwise or counterclockwise based on the sign of the cumulative angle, thus identifying a left or right head-turning action. When the user's head returns from the turned position to the center position, the cumulative Z-axis angle gradually decreases, and the absolute value drops below the head-turning action threshold. Simultaneously, the instantaneous value of the Z-axis angular velocity signal is detected to rapidly decrease to the zero threshold range (e.g., ±5° / s), indicating that the head has stopped rotating and stabilized in the center position. This is then identified as a return-to-center action.

[0090] For head-up movements, based on the vertical component of the Y-axis acceleration signal, since the projection of gravitational acceleration on the Y-axis changes with the head-up / head-down angle, the Y-axis acceleration changes instantaneously when the head posture changes pitch. Therefore, the instantaneous acceleration of the Y-axis acceleration signal in the vertical direction is extracted, and its change is calculated, such as the difference between the current value and the previous sampling point or the fluctuation amplitude within a short time window. When the acceleration value exceeds a preset threshold, it is determined to be a head-up movement.

[0091] By calculating the three-axis angular velocity and three-axis acceleration signals, it can independently identify four actions: left turn, right turn, return to center, and head tilt. Even if the user tilts their head slightly during the turn, it will not trigger an emergency stop or straight-line command. When the user tilts their head to stop suddenly, it will not cause unnecessary steering due to slight head tilt.

[0092] like Figure 3 As shown, this application can realize three main and independent functions: wheelchair training for controlling the wheelchair, shortcut commands for expressing intentions, and a keyboard for typing. Shortcut commands and the keyboard can only be used when the wheelchair is stationary. The interface is as follows. Figure 4 and Figure 5 As shown, the usage involves moving the head to control the cursor position, and a normal blink performs a mouse click. The host computer provides corresponding voice prompts and Pinyin typing output. If the cursor cannot be moved to certain corners during this stage, the cursor position can be reset by clearly tilting the head back and looking at the center of the screen for a preset duration, such as 4 seconds. After resetting, the cursor can be easily moved to any location on the interface. The gyroscope controls the cursor by using the cursor's return to the center of the screen after gyroscope calibration as the origin, collecting x-axis and y-axis acceleration values ​​and mapping them to the current screen width and length.

[0093] If you click on wheelchair training, you will enter the command interface for controlling the wheelchair, such as... Figure 6 and Figure 7As shown, you can click "Forward" and then "Start" to enter the wheelchair movement interface, or click "Backward" and then "Start" to reverse. Entering the wheelchair movement interface will immediately start the wheelchair in both forward and reverse modes, but reversing will not allow turning. To turn while moving forward, the user can turn slightly in the target direction. During this process, the algorithm calculates the integral of the z-axis angular velocity to determine left or right head movement. Once a certain angle is reached, a turning command is triggered, controlling the wheelchair electrodes to achieve the turn. The turning process will continue as long as the user's head remains upright. Once the user has turned to the set angle and their head returns to upright at a constant speed, the z-axis angular velocity is quickly returned to zero, and a straight-ahead command is output. To stop the wheelchair, the user can tilt their head back to trigger an emergency stop function, which detects the y-axis acceleration.

[0094] In some embodiments, after recognizing head movements based on triaxial angular velocity signals and triaxial acceleration signals, the method further includes:

[0095] When a head-raising or head-straightening motion is detected, or when no head movement exceeding a preset amplitude is detected within a preset time period, the static three-axis angular velocity signal and three-axis acceleration signal within the preset time period are acquired.

[0096] A new zero-point correction value is determined based on the static triaxial angular velocity signal and triaxial acceleration signal within a preset time period, and the current zero-point correction value is replaced with the new zero-point correction value. At the same time, the cumulative angle of the Z-axis is reset to zero.

[0097] In head attitude control based on inertial measurement units, the cumulative angle of the Z-axis obtained by integrating the Z-axis angular velocity will drift slowly over time. Even if the head has returned to the center position, the cumulative angle may not be zero. This drift mainly comes from the instability of the sensor's zero bias and temperature changes.

[0098] When a head-raising motion is detected, i.e., when the user tilts their head up or down, the user will typically maintain this head-raising posture briefly. During this time, the head is relatively static, and data collection immediately begins for a preset duration, such as 4 seconds, of static three-axis angular velocity and three-axis acceleration signals. To avoid transient interference during motion switching, it is preferable to use the data from the latter half of the duration, such as the last 2 seconds, to calculate the average value, which is then used as the zero bias for angular velocity and acceleration, respectively.

[0099] When the user's head returns to center after turning left or right, the head will also enter a brief period of stillness. Static data of a preset duration will be collected to calculate a new zero-point correction value.

[0100] If, within a preset time period, such as 6 seconds, no head movement exceeding a preset amplitude is detected (e.g., angular velocity change less than 5° / s and acceleration change less than 0.1g), it indicates the user may be in a relaxed or stationary state, and correction is triggered. It's important to note that this trigger condition does not apply while the wheelchair is in motion; it only applies when the wheelchair is stationary and the user is not actively controlling it.

[0101] To improve calculation accuracy, a correction is performed. Static triaxial angular velocity and triaxial acceleration signals are acquired over a preset duration, such as 4 seconds. Their average values ​​are calculated as new zero-point correction values, which then replace the currently used zero-point correction value. Simultaneously, all Z-axis accumulated angles obtained through integration are reset to zero, as the zero-point correction value has changed, rendering the old integration angles unreliable. After resetting, angle integration restarts with the new zero point as the reference, thus eliminating historical drift.

[0102] In some embodiments, the prefrontal cortex EEG signal further includes a blinking signal, and also includes:

[0103] Multiple threshold levels are preset, and each threshold level corresponds to a set of amplitude difference thresholds and time difference thresholds. The threshold level is positively correlated with the amplitude difference threshold and time difference threshold.

[0104] Determine the current threshold level and, within a preset calibration time, determine the blink signal from the prefrontal EEG signal at the current threshold level;

[0105] Determine whether a blink signal exists that simultaneously satisfies the amplitude difference threshold and the time difference threshold corresponding to the current threshold level;

[0106] If no blink signal matching the current threshold level is detected within the calibration period, the threshold level will be lowered to the next level.

[0107] If a blink signal that meets the current threshold level is detected within the calibration time, the amplitude difference threshold and time difference threshold corresponding to the current threshold level are used as the reference threshold for blink detection, and a valid blink signal is used as a confirmation instruction.

[0108] Considering the significant differences in the amplitude and duration of blink signals among different users, the weaker orbicularis oculi muscle in older adults results in a smaller difference in amplitude between peaks and troughs, while in younger people it is stronger. If a fixed threshold is used to detect blinks, it is easy to miss detections. If the threshold is too high, weak blinks cannot be identified or false detections may occur. If the threshold is too low, noise or unconscious minute eye movements may be misjudged as blinks.

[0109] Multiple threshold levels are pre-set, each corresponding to a set of parameters: the peak-to-trough amplitude difference threshold, for example, decreasing from 150μV to 50μV; and the peak-to-trough time difference threshold, for example, gradually widening from 100~250ms to 50~400ms. The threshold level is positively correlated with the amplitude difference threshold and the time difference threshold; the higher the level, the stricter the requirements, and the lower the level, the more lenient the requirements.

[0110] After calibration begins, the current threshold level is determined, typically starting with the highest level, i.e., the most stringent set of parameters. Within a preset calibration duration, such as 3 seconds, prefrontal cortex EEG signals are received, and blink waveforms are detected. By identifying peaks and troughs in the signal, the amplitude difference between them and the time difference between the peak and trough are calculated. Then, it is determined whether there is a blink signal that simultaneously satisfies the amplitude difference threshold and time difference threshold corresponding to the current threshold level. If at least one blink that meets the parameters of the current level is detected within the calibration duration, the calibration is immediately completed, and the amplitude difference threshold and time difference threshold corresponding to the current level are used as the baseline thresholds for subsequent blink detection for that user.

[0111] If no blink signal matching the current level parameters is detected within the calibration time, the threshold level is automatically lowered to the next level, using a more lenient amplitude and time difference range, and the above detection process is repeated. The levels are lowered sequentially until a valid blink is detected or all levels have been traversed. If no valid blink is detected at any level, a calibration failure message is output, and the default threshold is retained as the baseline.

[0112] After calibration, blink signals are continuously detected from the prefrontal cortex EEG signals using this baseline threshold during real-time control. Each time a blink signal matching the baseline threshold is detected, it is considered a valid confirmation command, equivalent to a left mouse click or active confirmation.

[0113] In some embodiments, the power spectral density of multiple frequency components of the brainwave is extracted, and a focus score is determined based on the power spectral density, including:

[0114] The DC component in the prefrontal cortex EEG signal was removed and bandpass filtered.

[0115] The theta wave, alpha wave, and beta wave with sequentially increasing frequencies, and their corresponding power spectral densities were separated from the filtered prefrontal EEG signal.

[0116] The focus score is determined based on the corresponding power spectral density. The expression for the focus score is:

[0117] ;

[0118] Where F represents the focus score, For the power spectral density of the beta wave, The power spectral density of the theta wave, Let be the power spectral density of the alpha wave.

[0119] Considering that a user's level of focus is closely related to the energy of specific frequency bands in the prefrontal cortex EEG signals, and that beta waves are generally positively correlated with active attention, alertness, and cognitive processing, while theta and alpha waves are enhanced during relaxation, drowsiness, or when eyes are closed, this application employs a method of extracting and determining the power spectral density of the three waveforms separately, and finally determining the focus score based on the power spectral density.

[0120] The acquired prefrontal EEG signals underwent preprocessing. Since the original signal contained DC offset and low-frequency drift, the DC component needed to be removed. A bandpass filter was then used to retain the effective frequency band, filtering out power line interference and extremely high-frequency noise. The preprocessed signal was still a mixed rhythm. Next, a digital bandpass filter was used to separate it into three independent frequency band components: theta wave, alpha wave, and beta wave. For each frequency band component, its power spectral density was calculated and denoted as follows: , , Power spectral density reflects the energy distribution of a signal per unit frequency, and is usually taken as the average power within the effective frequency band as a representative value.

[0121] Specifically, the dual-lead EEG electrodes can acquire prefrontal cortex EEG signals FP1 and FP2 at a sampling rate of 250Hz. First, the EEG signals are averaged to remove the DC component. Then, a 2-40Hz bandpass filter is applied. Next, the bandpass filter separates the beta wave (13Hz-30Hz), alpha wave (8Hz-13Hz), and theta wave (4Hz-8Hz), and calculates the power spectral density for each. The focus score is calculated and updated every second, with each score based on the latest 3-second data. A higher focus score indicates that the focus-related rhythm is dominant relative to the relaxation / drowsiness rhythm.

[0122] In some embodiments, after determining the focus score based on the power spectral density, the process includes:

[0123] The moving average of the focus score is determined by the following expression:

[0124] ;

[0125] in, The moving average at the current time. This is the moving average value from the previous time step. The focus score calculated at the current moment. and For the preset weighting coefficients, and ;

[0126] The moving average at the current moment is compared with a preset attention threshold. If the moving average at the current moment is lower than the attention threshold, it is determined that there is insufficient attention.

[0127] Set the wheelchair drive enable signal to invalid, preventing the wheelchair from starting or stopping the currently executing wheelchair control command;

[0128] If the wheelchair is in motion, it will trigger a deceleration stop and generate a voice prompt.

[0129] Given that a single calculated focus score may fluctuate momentarily due to blinking, muscle artifacts, or brief distractions, directly using it for safety decisions could lead to frequent reversals of the enable signal, impacting user experience. Furthermore, a decline in attention is typically a continuous process rather than an instantaneous event, requiring assessment within a defined time window.

[0130] and For the preset weighting coefficients, and ,For example =0.8, =0.2. Then, the moving average... The attention threshold was compared to a preset threshold, which was experimentally calibrated and typically set to 60. When When the user's attention level falls below the specified threshold, it is determined that the current user is inattentive, and an inattentiveness flag is output. This flag controls the wheelchair drive enable signal. Once inattentiveness is confirmed, the wheelchair drive enable signal is disabled, thus preventing new wheelchair start commands. Even if the user makes head movements or blinks to confirm, the wheelchair will not respond. If the wheelchair is in motion, a deceleration and stop command is immediately triggered, causing the wheelchair to smoothly decelerate to a stop. At the same time, a voice prompt message is generated, such as "Please focus," reminding the user to regain attention.

[0131] In some embodiments, wheelchair control commands are generated based on head movements, including:

[0132] Turning commands are generated based on left and right head turning actions, and the duration of the turning commands is positively correlated with the duration of maintaining the left and right head turning angles.

[0133] An emergency stop command is generated based on the head-raising motion. The emergency stop command is used to control the wheelchair to decelerate and stop immediately.

[0134] The straight-ahead command is generated based on the return-to-center action. The straight-ahead command is used to remove the wheelchair from the turning state and resume straight-ahead driving.

[0135] Turning command generation: When a left or right head turn is detected, a corresponding turning command is generated. The duration of this turning command is positively correlated with the duration of the left or right head turn; the longer the user maintains the head turn, the longer the turning command is output. Furthermore, the turning radius is also positively correlated with the size of the head turn: the larger the head turn, the greater the differential drive ratio of the wheelchair, and the smaller the turning radius. The turning command continues to be output as long as the user's head remains turned and has not returned to a neutral position; once the user's head returns to a neutral position, the turning command stops.

[0136] Emergency stop command generation: When a head-up movement is detected, an emergency stop command is generated immediately. The emergency stop command has the highest priority. Regardless of whether the wheelchair is currently moving straight, turning, or stationary, the wheelchair drive motor is immediately controlled to decelerate to zero at maximum deceleration, and the drive enable is locked until the user completes the confirmation process again.

[0137] Straight-line command generation: When a return-to-center action is detected, and the wheelchair returns from the turning position to the center, a straight-line command is generated. This command is used to exit the turning state and resume straight-line travel. Specifically, it monitors whether the Z-axis angular velocity signal quickly returns to zero, and combines this with the confirmation that the Z-axis accumulated angle has dropped below a threshold. Once the return-to-center is confirmed, the straight-line command is sent. At this point, the wheelchair will move forward in a straight line at the currently set speed without any yaw.

[0138] When generating wheelchair control commands, this application establishes a positive correlation between head turning angle, holding time, turning radius, and duration, allowing users to precisely control the wheelchair's steering behavior by fine-tuning their head posture, much like driving a vehicle.

[0139] In some embodiments, it also includes:

[0140] Set a dropout threshold, which is lower than the attention threshold corresponding to the control condition;

[0141] When the focus score is lower than the detachment threshold, it is determined that the electrode has detached.

[0142] The user is prompted to check the electrode connections;

[0143] Keep wheelchair control commands disabled until the electrode detachment fault is resolved.

[0144] The quality of prefrontal cortex EEG signals depends on good contact between the electrodes and the skin. In actual use, due to head movements, sweating, or unintentional touch by the user, one or two electrodes may gradually loosen or completely detach. When electrodes detach, the amplitude of the acquired signal decreases, and the attention score will be abnormally low. Judging solely by an attention score below the attention threshold may mislead users into believing they are not concentrating, while ignoring the contact problem with the device itself.

[0145] To improve control accuracy, this application pre-sets two thresholds: an attention threshold and a dropout threshold, where the dropout threshold is lower than the attention threshold. During real-time control, a moving average of the attention score is continuously calculated. .when When the attention threshold is lower than the attention threshold, first determine whether it is also lower than the dropout threshold.

[0146] like If the value is greater than or equal to the detachment threshold but less than the attention threshold, it is determined that the user is not paying enough attention, an attention deficiency flag is output, and the aforementioned safety enable / disable and voice prompts are executed.

[0147] like If the value is below the detachment threshold, it is determined to be an electrode detachment or severe contact failure, and an electrode detachment fault indicator is output. The user is immediately prompted to check the electrode connection via voice, a pop-up window, or flashing indicator lights. Simultaneously, until the electrode detachment fault is resolved, wheelchair control commands remain disabled; that is, even if the user makes head movements or blinks for confirmation, the wheelchair will not respond to prevent malfunctions due to unreliable signals.

[0148] To restore control of the wheelchair, the quality of the EEG signal will be continuously monitored. Once the user has reattached the electrodes and the signal has returned to normal, the moving average of the attention score will rise above the detachment threshold, automatically clearing the fault marker and allowing for recalibration of blinks or restoration of normal control mode.

[0149] Figure 9 This is a schematic diagram of a wheelchair control device provided by the present invention. The wheelchair control device includes:

[0150] Memory 21 is used to store computer programs;

[0151] The processor 22 is used to implement the steps of the above-described wheelchair control method when executing a computer program.

[0152] The description of the wheelchair control device provided in this application is similar to that in the above embodiments and will not be repeated here.

[0153] The present invention also provides a wheelchair, including the wheelchair control device described above.

[0154] The wheelchair provided in this application is described in the above embodiments and will not be repeated here.

[0155] It should also be noted that, in this specification, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0156] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0157] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for controlling a wheelchair, characterized in that, include: Acquire prefrontal cortex EEG signals and triaxial angular velocity and triaxial acceleration signals generated by head movements, wherein the prefrontal cortex EEG signals include brain waves; Head movements are identified based on the triaxial angular velocity signal and the triaxial acceleration signal, and the head movements include head tilting, left and right head turning, and head returning to center. The power spectral density of multiple frequency components of the brainwave is extracted, and a focus score is determined based on the power spectral density. The focus score is positively correlated with the user's level of focus. Wheelchair control commands are generated based on the head movements, and the wheelchair is controlled according to the wheelchair control commands when the focus score meets the control conditions.

2. The wheelchair control method as described in claim 1, characterized in that, Identifying head movements based on the triaxial angular velocity signal and the triaxial acceleration signal includes: Integrate the Z-axis angular velocity signal to obtain the cumulative Z-axis angle; The absolute value of the cumulative Z-axis angle is compared with the head rotation threshold. When the absolute value exceeds the head turning action threshold, the head action is determined to be a left head turning action or a right head turning action based on the direction of the cumulative angle of the Z-axis. When the absolute value of the cumulative Z-axis angle gradually decreases from exceeding the head rotation threshold to less than the head rotation threshold, and the Z-axis angular velocity signal changes to within the zero threshold range, the head movement is determined to be a return-to-center movement. Extract the instantaneous acceleration in the vertical direction from the Y-axis acceleration signal; When the change in instantaneous acceleration exceeds the head-raising acceleration threshold, the head movement is determined to be a head-raising movement.

3. The wheelchair control method as described in claim 1, characterized in that, After recognizing head movements based on the triaxial angular velocity signals and the triaxial acceleration signals, the method further includes: When a head-raising or head-straightening motion is detected, or when no head movement exceeding a preset amplitude is detected within a preset time period, the static three-axis angular velocity signal and three-axis acceleration signal within the preset time period are acquired. A new zero-point correction value is determined based on the static triaxial angular velocity signal and triaxial acceleration signal within the preset time period, and the current zero-point correction value is replaced with the new zero-point correction value. At the same time, the cumulative angle of the Z-axis is reset to zero.

4. The wheelchair control method as described in claim 1, characterized in that, The prefrontal cortex EEG signals also include blinking signals, and further include: Multiple threshold levels are preset, each threshold level corresponds to a set of amplitude difference thresholds and time difference thresholds, and the threshold level is positively correlated with the amplitude difference thresholds and the time difference thresholds; Determine the current threshold level, and within a preset calibration time, determine the blink signal from the prefrontal EEG signal at the current threshold level; Determine whether a blink signal exists that simultaneously satisfies the amplitude difference threshold and the time difference threshold corresponding to the current threshold level; If no blink signal matching the current threshold level is detected within the calibration period, the threshold level will be lowered to the next level. If a blink signal that meets the current threshold level is detected within the calibration time, the amplitude difference threshold and time difference threshold corresponding to the current threshold level are used as the reference threshold for blink detection, and a valid blink signal is used as a confirmation instruction.

5. The wheelchair control method as described in claim 1, characterized in that, Extracting the power spectral density of multiple frequency components of the EEG, and determining the focus score based on the power spectral density, including: The DC component in the prefrontal cortex EEG signal is removed and bandpass filtered. The theta wave, alpha wave, and beta wave with sequentially increasing frequencies, and their corresponding power spectral densities were separated from the filtered prefrontal EEG signal. The focus score is determined based on the corresponding power spectral density, and the expression for the focus score is: ； Where F represents the focus score, For the power spectral density of the beta wave, The power spectral density of the theta wave, Let be the power spectral density of the alpha wave.

6. The wheelchair control method as described in claim 1, characterized in that, After determining the focus score based on the power spectral density, the following is included: Determine the moving average of the focus score, the expression for which is: ； in, The moving average at the current time. This is the moving average value from the previous time step. The focus score calculated at the current moment. and For the preset weighting coefficients, and ; The moving average value at the current moment is compared with a preset attention threshold. When the moving average value at the current moment is lower than the attention threshold, it is determined that there is insufficient attention. Set the wheelchair drive enable signal to invalid, preventing the wheelchair from starting or stopping the currently executing wheelchair control command; If the wheelchair is in motion, it will trigger a deceleration stop and generate a voice prompt.

7. The wheelchair control method as described in claim 1, characterized in that, Generate wheelchair control commands based on the head movements, including: A turning command is generated based on the left and right head turning actions, and the duration of the turning command is positively correlated with the holding time of the left and right head turning angles; An emergency stop command is generated based on the head-raising motion; the emergency stop command is used to control the wheelchair to decelerate and stop immediately. A straight-ahead command is generated based on the return-to-center action. The straight-ahead command is used to cause the wheelchair to exit the turning state and resume straight-ahead travel.

8. The wheelchair control method according to any one of claims 1 to 7, characterized in that, Also includes: Set a dropout threshold, which is lower than the attention threshold corresponding to the control condition; When the focus score is lower than the detachment threshold, it is determined that the electrode has detached. The user is prompted to check the electrode connections; Keep wheelchair control commands disabled until the electrode detachment fault is resolved.

9. A control device for a wheelchair, characterized in that, include: Memory, used to store computer programs; A processor, configured to implement the steps of the wheelchair control method as described in any one of claims 1 to 8 when executing the computer program.

10. A wheelchair, characterized in that, Includes the wheelchair control device as described in claim 9.

Images