Intelligent nursing method and system based on brain-computer interface

Abstract

The invention discloses an intelligent nursing method and system based on a brain-computer interface, and the method comprises the steps: setting an interactive interface: the interactive interface comprises a character input interface and a digital input interface, each interface is provided with a key position or pattern and a 24-key-position Sudoku layout, improves the input efficiency, and integrates a function entertainment interface and a ward management interface; setting visual stimulation signals, wherein the frequencies of the visual stimulation signals of the keys or patterns are different and are non-integer multiple fundamental frequencies; a visual stimulation signal of each key position or pattern is subjected to phase reversal alternating type PRCP coding setting, so that interference is reduced; sSVEP signal identification and processing: SSVEP signal notch filtering and FBCCA classification identification are carried out, interference is reduced, and the identification rate is improved. And according to the ALSFRS-R scoring range, progressive interaction mode conversion of eye movement, electroencephalogram and respiration is adopted.

Description

Technical Field

[0001] This invention relates to the field of medical system technology, and in particular to an intelligent care method and system based on brain-computer interface. Background Technology

[0002] The Brain-Computer Interface (BCI) intelligent care system for wards is a novel medical and nursing solution based on neuroscience and artificial intelligence technologies. It is specifically designed for scenarios such as intensive care units (ICUs), patients with motor neuron diseases (such as ALS), high-level paraplegia and other limb dysfunctions, neurorehabilitation departments, and long-term bedridden patients. It achieves integrated control of Chinese input, digital operation, and expression of daily living needs directly through neural signals. The system can decode brain signals through the BCI and convert them into language output, helping disabled patients and those with ALS achieve basic communication and daily living assistance. It can also be used in smart elderly care scenarios, ensuring safety while maintaining the individual's autonomy, making technology a guardian of dignified aging.

[0003] Brain-computer interface technologies used in the field of rehabilitation therapy and ward care mainly include the following three categories:

[0004] (1) Hybrid BCI system

[0005] Technical principle: Integrating SSVEP (steady-state visual evoked potential) and motor imagery (MI) dual-modal signals;

[0006] Clinical problems: The training period is as long as 28 days, the MI signal is susceptible to EMG interference (sensitivity decreases by 40%), and visual fatigue causes SSVEP signal attenuation (accuracy decreases by 35% after 2 hours of use).

[0007] (2) P300 speller

[0008] Technical principle: Utilizing the oddball paradigm to induce event-related potentials (N200 / P300 components).

[0009] Clinical challenges: Input speed is only 4-5 characters / minute, and the learning cycle for patients with cognitive impairment (such as Alzheimer's disease) is as long as 14.3 days, requiring the maintenance of high attention (NASA-TLX load score >80).

[0010] (3) SSVEP paradigm

[0011] Technical principle: Evoking brain electrical responses through visual stimulation at 8-30Hz;

[0012] Clinical problem: Ambient light interference causes a 50% decrease in SNR (when illuminance > 500 lux), and patients with advanced ALS cannot use it due to oculomotor paralysis;

[0013] (4) Existing care systems have limited functionality (supporting only basic commands). Summary of the Invention

[0014] The purpose of this invention is to provide an intelligent care method and system based on a brain-computer interface to solve one or more of the above-mentioned technical problems or defects.

[0015] To achieve this objective, the present invention adopts the following technical solution:

[0016] A brain-computer interface-based intelligent care method includes:

[0017] Setting up the user interface:

[0018] The interactive interface includes a text input interface and a number input interface, each with keys or icons;

[0019] Set visual stimulus signals:

[0020] The frequency of the visual stimulus signal for each key or pattern is different, and all of them are non-integer multiples of the fundamental frequency;

[0021] The visual stimulus signal of each key or pattern is set with phase-reversed alternating PRCP encoding.

[0022] SSVEP signal identification and processing:

[0023] SSVEP signals are used for FBCCA classification and recognition.

[0024] In some implementations, notch filtering is performed on the SSVEP signal during SSVEP signal identification and processing.

[0025] In some implementations, both the text input interface and the numeric input interface have 24 keys arranged in a matrix.

[0026] The text input interface can intelligently predict word frequency during input;

[0027] The numeric keypad on the numeric input interface uses a linear frequency distribution.

[0028] Both the text input interface and the numeric input interface have emergency call buttons;

[0029] The user interface includes a functional entertainment interface and a ward management interface;

[0030] An entertainment playback interface is located on the side of the text input interface and the number input interface.

[0031] In some implementations, the PRCP encoding is set as follows:

[0032] Odd frames: display positive phase;

[0033] Even-numbered frames: display negative phase; where the positive and negative phases are 180 degrees apart;

[0034] The phase reversal frequency fm is not an integer multiple of 50Hz or 60Hz.

[0035] In some implementations, the visual stimulus signal can be switched between a high-frequency SSVEP mode and a low-frequency SSVEP mode when setting the visual stimulus signal.

[0036] The visual stimulus signal for the key or pattern is set as a composite stimulus of flashing light spot and radial motion, and PCA is used to separate the components of SSVEP signal and SSVEP signal.

[0037] In some implementations, a hybrid interaction mode combining SSVEP and eye-tracking is also provided;

[0038] Prioritize SSVEP interaction mode;

[0039] When the SSVEP signal is lost, switch to eye-tracking interaction mode.

[0040] In some implementations, during eye-tracking interaction mode, an inertial measurement unit (IMU) is used to correct gaze point errors caused by head shift.

[0041] In eye-tracking interaction mode, blink frequency is monitored. If a blink interval > 5 seconds is detected, eye-tracking interaction mode is paused.

[0042] Eye-tracking signals undergo redundancy verification to reduce the false recognition rate;

[0043] Stare at the "Emergency Call" button for an extended period of time to make a voice call.

[0044] In some implementations, a progressive interaction mode is configured based on the patient's ALSFRS-R score range;

[0045] The progressive interaction mode is configured as follows: eye-tracking interaction mode → SSVEP interaction mode → breathing interaction mode;

[0046] Alternatively, a gradual setup can be implemented based on the stage of the disease;

[0047] Early-stage patients: Use SSVEP high-frequency mode;

[0048] For late-stage patients: use SSMVEP low-frequency mode or respiratory interaction mode.

[0049] In some implementations, the feature is that a CCA-wtCCA hybrid architecture is further provided to realize multi-stage signal processing;

[0050] The primary signal undergoes wtCCA processing;

[0051] The secondary verification signal undergoes cross-interaction mode correlation verification processing for CCA;

[0052] The final identification result is output after Bayesian decision fusion processing.

[0053] A brain-computer interface-based intelligent care system, applying the aforementioned intelligent care method, includes:

[0054] A screen is used to display the interactive interface;

[0055] The EEG acquisition module includes an EEG cap and a brain-computer interface; the EEG acquisition module is used to acquire EEG signals in real time, and the EEG signals include interactive signals;

[0056] The EEG signal analysis module is used for the recognition and processing of interactive signals to generate recognition results;

[0057] The control module executes operations based on the recognition results.

[0058] The beneficial effects of this invention are:

[0059] It integrates multiple interfaces such as text input, number input, lifestyle and entertainment, and ward management to improve functionality;

[0060] The text and number input uses a 24-key nine-grid layout. Each key is optimized with a harmonic-free frequency that is a non-integer multiple of the base frequency. It also features PRCP anti-interference coding design, dynamic brightness adjustment, reduced interference, improved recognition rate, and reduced visual fatigue.

[0061] SSVEP signal notch filtering, combined with FBCCA classification and recognition, reduces interference and improves recognition rate;

[0062] The SSVEP+ eye-tracking interaction mode is used to reduce fatigue;

[0063] It employs SSVEP high-frequency mode + SSVMVEP low-frequency mode to reduce visual fatigue;

[0064] Based on the ALSFRS-R score range, a progressive mode transition from eye movement to electroencephalography (EEG) to respiration was adopted, and a CCA-wtCCA hybrid architecture was used for cross-interaction mode correlation verification and multi-stage signal processing. Attached Figure Description

[0065] Figure 1 This is a flowchart of an intelligent care method based on a brain-computer interface according to the present invention;

[0066] Figure 2 This is a flowchart of the SSVEP and eye-tracking interaction mode of the present invention;

[0067] Figure 3 This is an example diagram of the text input interface of the present invention;

[0068] Figure 4 This is an example diagram of the digital input interface of the present invention;

[0069] Figure 5 This is one of the example diagrams of Pinyin input for the text input interface of the present invention;

[0070] Figure 6 This is the second example diagram of the Pinyin input method in the text input interface of the present invention;

[0071] Figure 7 This is an example diagram of the functional entertainment interface of the present invention;

[0072] Figure 8 This is a structural diagram of the system of the present invention;

[0073] The components include: 1. Screen; 2. EEG acquisition module; 21. EEG cap; 22. Brain-computer interface; 3. EEG signal analysis module; and 4. Control module. Detailed Implementation

[0074] The present invention will now be described in further detail with reference to the accompanying drawings.

[0075] refer to Figure 1 A brain-computer interface-based intelligent care method includes:

[0076] Setting up the user interface:

[0077] The interactive interface includes a text input interface and a number input interface, each with keys or icons;

[0078] The text input interface is used to input Chinese or English characters; the number input interface is used to input numbers.

[0079] Set visual stimulus signals:

[0080] Each key or pattern visual stimulus signal has a different frequency, and all of them are non-integer multiples of the fundamental frequency; for example, 8.57Hz, 11.43Hz, 14.29Hz, etc., so that the fundamental frequency and harmonics avoid 50Hz and its multiples (50Hz, 100Hz, 150Hz), forming a flashing frequency without harmonic relationship, away from noise peaks, and reducing interference.

[0081] Each key or pattern's visual stimulus signal undergoes phase-reversal alternating PRCP encoding. PRCP (Pattern Recognition with Canonical Patterns) reduces interference from ambient light. PRCP encoding maintains an SNR >10dB for the SSVEP signal in strong light environments, improving recognition accuracy. PRCP encoding also allows for dynamic brightness adjustment (200-1000cd / m2), reducing visual fatigue and environmental interference.

[0082] The harmonic-free frequency design of non-integer multiples of the fundamental frequency and the PRCP coding design form a dual guarantee, reducing interference and improving the recognition rate;

[0083] SSVEP signal identification and processing:

[0084] SSVEP signals are used for FBCCA classification and identification;

[0085] FBCCA (Filter Bank Canonistic Correlation Analysis) is a method that uses a bandpass filter bank (e.g., (5,90), (14,90), (22,90), (30,90), (38,90)) to decompose the original signal into sub-bands of different frequencies. The correlation coefficient between each sub-band and the reference signal is then calculated and weighted, with the category of the reference signal showing the highest correlation being used for identification. Combining FBCCA with PRCP coding further reduces interference and improves the recognition rate.

[0086] Alternatively, a cascaded real-time decoding algorithm of CCA-wtCCA can be used to classify and identify SSVEP signals. CCA (Canonical Correlation Analysis) and wtCCA (Wavelet Transform CCA) are processed together, breaking through the single application mode and improving the accuracy of the identification results. Specifically, canonical correlation analysis (CCA) is used to extract the fundamental frequency and harmonic components of the target frequency steady-state visual evoked potential (SSVEP). Then, weighted canonical correlation analysis (wtCCA) is used to introduce individual EEG feature weights, i.e., weighting signals of different frequencies to enhance the contribution of the target frequency signal and suppress noise and other interference signals. Finally, the outputs of CCA and wtCCA are cascaded to output a more accurate identification result.

[0087] In SSVEP signal recognition and processing, notch filtering is applied to the SSVEP signal. Real-time notch filtering is achieved by setting a notch filter to remove interference signals at specific frequencies, especially power frequency interference (such as 50Hz or 60Hz) and its harmonics. The signal is then classified and identified, thereby reducing interference and improving the recognition rate.

[0088] refer to Figures 3 to 7 Both the text input interface and the number input interface have 24 keys arranged in a matrix, such as a 6×4 matrix layout, forming a 24-key nine-grid layout.

[0089] refer to Figure 3 The text input interface has 6×4 keys, including 8 letter keys, 5 number keys, page up and down keys, 4 punctuation mark keys, as well as keys for emergency call, interface switching, and function entertainment interface display; it also has an intelligent navigation bar, which includes a pinyin candidate area and a Chinese character candidate area; the number keys are used to select Chinese characters in the Chinese character candidate area.

[0090] Each key is configured with a unique frequency stimulus (8-15Hz, step size 0.5Hz), and the recognition rate is ensured through PRCP anti-interference coding design;

[0091] Pinyin candidate area: 17.5Hz / 19.5Hz orthogonal frequency pair setting to achieve zero accidental page turning;

[0092] Chinese character candidate area: 21.5Hz / 23.5Hz frequency encoding setting, and can be combined with a three-axis accelerometer (sensitivity ±0.5°). The three-axis accelerometer can be used to detect the user's head movement or other body movements, thereby providing additional input signals and enhancing the system's interactivity. It can also detect micro-head movements. The five-level candidate channel uses theta band (4-7Hz) energy threshold triggering, with a response time of <200ms.

[0093] The text input interface can intelligently predict word frequency during input; for example, by introducing an intelligent word frequency prediction engine, input efficiency can be improved by 40%.

[0094] refer to Figure 4 The numeric input interface has 24 keys arranged in a matrix, such as a 6×4 matrix layout, forming a 24-key nine-grid layout; it also has 10 numeric keys, punctuation keys, four software interaction keys, and other keys.

[0095] The numeric keys use a linear frequency distribution; for example, the 0-9 numeric keys use a linear frequency distribution of 10-19Hz. Through PRCP anti-interference coding design (minimum frequency interval 1.2Hz), the confirmation key (20Hz) integrates a dual verification mechanism.

[0096] Both the text input interface and the numeric input interface have interface switching keys and emergency call keys;

[0097] Interface switching key: 25Hz square wave stimulation enables switching between digital and text interfaces;

[0098] Emergency call button: 27Hz red and green dual-frequency alternating flashing (duty cycle 1:1) to ensure 100% recognition of emergency calls.

[0099] refer to Figure 7 The interactive interface also includes a functional entertainment interface and a ward management interface.

[0100] Functional entertainment interface: Used for daily life calls and entertainment, supporting calls for drinking water, eating, toilet, sleeping and washing, etc., and supporting entertainment such as WeChat, Weibo, Douyin, music, video, Baidu, etc.

[0101] Food / drink key: 28Hz red rectangle flashing (CIE1931 chromaticity coordinates x=0.67, y=0.33);

[0102] Toilet button: 26Hz orange square flashing (brightness 100cd / m2);

[0103] Washing / rinsing button: 22Hz sine wave stimulation (modulation depth 80%);

[0104] Sleep button: 24Hz slowly changing waveform (rise time 500ms);

[0105] Music key layout: 13Hz purple gradient grating (spatial frequency 0.5cpd);

[0106] Video key layout: 15Hz green horizontal scan line (motion speed 10° / s);

[0107] It also features an AI assistant: a built-in Natural Language Processing (NLP) engine enables voice / EEG control for functions such as weather inquiries and news reading. The AI ​​assistant can be configured in various interfaces.

[0108] Both the text input interface and the number input interface have an entertainment playback interface on the side, which is used to play or display entertainment items, such as Weibo, WeChat, Douyin, Baidu, etc. It is similar to placing multiple windows or areas in the same interface, so that you can watch entertainment content while inputting information.

[0109] Ward management interface: used to connect to the hospital's HIS (Hospital Information System), supporting functions such as medical order inquiry, medication reminder, and nursing assistant call; for example, the emergency call button allows for quick contact with medical staff in emergency situations via text-to-speech (TTS) or one-click alarm.

[0110] Therefore, the interactive interface integrates multiple functional interfaces, combining information input, entertainment, and hospital HIS (Hospital Information System) functions, thus improving functionality. Furthermore, both the text and number input interfaces feature a 24-key nine-grid layout, offering higher input efficiency compared to the existing 36-key design.

[0111] The PRCP encoding settings are as follows:

[0112] Odd frames: display positive phase;

[0113] Even-numbered frames: display negative phase; where the positive and negative phases are 180 degrees apart;

[0114] The phase reversal frequency fm is not an integer multiple of 50Hz or 60Hz.

[0115] Specifically, the dominant frequency (f) of the visual stimulus signal: for example, setting f = 10Hz;

[0116] Sideband frequency (f±fm): where fm is the phase reversal frequency. For example, setting fm = 0.5Hz means that the phase reversal occurs once every 2 seconds. By setting fm ≠ an integer multiple of 50Hz or 60Hz, the sideband frequency is kept away from 50Hz or 60Hz and its harmonics, thus avoiding spectral aliasing, reducing interference, and improving the recognition rate.

[0117] In setting the visual stimulus signal, the visual stimulus signal can switch between the SSVEP high-frequency mode and the SSVEP low-frequency mode;

[0118] The visual stimulus signal for the key or pattern is set as a composite stimulus of flashing light spot and radial motion, and PCA is used to separate the components of SSVEP signal and SSVEP signal.

[0119] The SSVEP high-frequency mode can be set to around 30-60Hz, and the SSVEP low-frequency mode can be set to around 8-15Hz.

[0120] Principal component analysis (PCA) is used for signal processing. By selecting specific principal components, SSVEP and SSMVEP signals can be separated. SSVEP signals are typically associated with specific visual stimulus frequencies, while SSMVEP signals are associated with motion-induced visual stimuli. Therefore, by setting high-frequency SSVEP and low-frequency SSMVEP modes, visual fatigue in patients can be reduced.

[0121] refer to Figure 2 The method also sets up a hybrid interaction mode of SSVEP and eye-tracking;

[0122] Prioritize SSVEP interaction mode;

[0123] When the SSVEP signal is lost, switch to eye-tracking interaction mode.

[0124] Therefore, the SSVEP and eye-tracking hybrid interaction mode can reduce patient fatigue.

[0125] Among them, the eye-tracking interaction mode can capture the trajectory of eye movements (such as pupil center / corneal reflection) through infrared optical tracking or video, and convert it into control commands (such as screen cursor movement, button clicks).

[0126] Therefore, the hybrid interaction mode of SSVEP and eye movement is beneficial to improving input efficiency and is also suitable for more patients.

[0127] In eye-tracking interaction mode, an inertial measurement unit (IMU) is used to correct gaze point errors caused by head shift.

[0128] In eye-tracking interaction mode, the blinking frequency is monitored. If a blinking interval > 5 seconds is detected, the eye-tracking interaction mode is paused to prevent or reduce fatigue and remind the user to rest.

[0129] Eye-tracking signals undergo redundancy checks to reduce the false recognition rate. For example, embedding a redundancy check module reduces the false recognition rate of eye-tracking signals by 42% (measured data) and also provides a stable transition for progressive interaction mode configuration. For example, while retaining the PRCP feature extraction framework, embedding a redundancy check module enables PRCP encoding to have fault tolerance. The redundancy check module can be a cyclic redundancy check, Hamming code, or repeat code design module.

[0130] Stare at the "Emergency Call" button for an extended period to make a voice call. This allows for quick contact with medical personnel via text-to-speech (TTS) or a one-button alarm in emergencies.

[0131] The method also includes: configuring multiple interactive modes progressively based on the patient's ALSFRS-R score range;

[0132] The progressive interaction mode is configured as follows: eye-tracking interaction mode → SSVEP interaction mode → breathing interaction mode;

[0133] Eye-tracking interaction mode: The eye movement trajectory (such as pupil center / corneal reflection) is captured by infrared optical tracking or video and converted into control commands (such as screen cursor movement, button click).

[0134] SSVEP Interaction Mode: Based on non-invasive EEG, electrodes collect SSVEP signals from the brain based on the visual flashing frequency and identify them as control commands.

[0135] Respiratory Interaction Mode: Respiratory rhythm is detected by pressure sensors or chest and abdominal movement monitoring and encoded as a switch signal.

[0136] ALSFRS-R is a standardized scale used to assess the functional status of patients with amyotrophic lateral sclerosis (ALS).

[0137] Low score (0-16 points): Patients have severe functional impairment. Eye-tracking interaction mode, SSVEP interaction mode and respiratory interaction mode can all be used as the main interaction mode. The specific choice depends on the patient's visual and respiratory function preservation.

[0138] Mid-to-high score range (17-48 points): Patients have some functional impairment, but still have some voluntary motor ability. Eye-tracking interaction mode and SSVEP interaction mode can be used as auxiliary interaction methods, and the respiratory interaction mode can be selected according to the degree of respiratory function impairment.

[0139] The method also includes a progressive setup based on the condition of early-stage and late-stage patients.

[0140] Early stage patients: SSVEP high-frequency mode (24 keys), for example, the 24 keys of both the text input interface and the numeric input interface use the SSVEP high-frequency mode;

[0141] For late-stage patients: Automatically switch between SSMVEP low-frequency mode (9 or 10 keys) or respiratory interaction mode assistance, such as using SSMVEP low-frequency mode for some or all keys in the text input interface and the numeric input interface, such as letter keys in the text input interface and number keys in the numeric input interface.

[0142] Therefore, the interaction mode is progressively designed according to the patient's condition, which is beneficial for the patient's use, improves the applicability, and meets the full range (0-48) of ALSFRS-R. In addition, the fault-tolerant enhancement design of PRCP encoding retains the PRCP feature extraction framework and incorporates a redundancy check module, which enables PRCP encoding to have fault tolerance function, providing a stable transition for the adjustment or degradation of the interaction mode.

[0143] This method also sets up a CCA-wtCCA hybrid architecture to achieve multi-stage signal processing;

[0144] The primary signal undergoes wtCCA processing;

[0145] The secondary verification signal undergoes cross-interaction mode correlation verification processing for CCA;

[0146] The final identification result is output after Bayesian decision fusion processing.

[0147] Therefore, the hybrid architecture of CCA-wtCCA breaks through the traditional single application mode of CCA (Canonical Correlation Analysis) and wtCCA (Wavelet Transform CCA), constructing a hybrid analysis pipeline to improve the accuracy of recognition results. In multi-stage signal processing, setting up the hybrid architecture of CCA-wtCCA ensures that the system survival rate is improved to 91% when ALSFRS-R < 20.

[0148] For example, in multi-interaction mode hybrid settings such as progressive interaction mode, progressive settings for early and late stage patients, and SSVEP+eye-tracking hybrid interaction, a CCA-wtCCA hybrid architecture can be set up to realize cross-interaction mode correlation verification, multi-stage signal processing, and improve signal recognition rate.

[0149] To further improve the recognition rate, reduce interference, and enhance stability, the following settings can be made:

[0150] Lighting compensation: Ambient light sensor (ALS) monitors ambient illuminance in real time (0-100klux); equipped with color temperature detection unit (3000K-6500K); sampling frequency ≥10Hz;

[0151] Noise interference suppression: power frequency noise monitoring (50 / 60Hz±2Hz), electromyography artifact identification (30-200Hz), environmental electromagnetic noise scanning, etc.

[0152] Impedance monitoring acquires the values ​​of the brain-computer interface device's detectors in real time, performs real-time analysis and updates, and enables dynamic display and alarm of electrode contact impedance.

[0153] This method can also control the interaction mode;

[0154] 1. Multi-level command mapping: Supports single commands, combined commands, and consecutive commands.

[0155] For example, a single command: a basic need: such as "drink water" or "eat". An emergency call: an emergency call.

[0156] Combined commands for interactive entertainment: For example, intelligent search: search tool + input method typing + search terms + search commands

[0157] Continuous commands: text and number input, continuous page turning, and other operations.

[0158] 2. Context-aware interaction: Automatically adjusts control strategies based on application scenarios.

[0159] Task Dimension: Current operation context, dynamic adjustment of function priority: For example: Detecting that the user has not drunk water for a long time → pop-up reminder and pinning the "drink water" command to the top. When a video is playing, the SSVEP frequency of 12Hz is temporarily mapped from "call" to "pause".

[0160] Non-invasive sensing: The state is inferred through environmental sensors and user operation logs, without the need for additional wearable devices.

[0161] User perspective: Physiological and behavioral status. If eye tracking shows a distracted gaze, the duration will be automatically extended. If the EEG signal amplitude decreases, rest and suspension of non-emergency operations are recommended.

[0162] 3. Error detection and recovery: Abnormal operations are automatically identified and corrected.

[0163] Interaction logic anomalies: invalid command sequence (such as 5 consecutive "cancel" operations), conflicting commands (simultaneously requesting "sleep" and "watch TV"), high-frequency repetitive operations (the same command is triggered 5 times within 10 seconds).

[0164] Abnormal physiological signals: for example, signal-to-noise ratio that does not meet the standard, or target frequency energy that is lower than the baseline standard.

[0165] refer to Figure 8 A brain-computer interface-based intelligent care system, applying the aforementioned intelligent care method, includes:

[0166] Screen 1 is used to display the interactive interface;

[0167] The EEG acquisition module 2 includes an EEG cap 21 and a brain-computer interface 22; the EEG acquisition module 2 is used to acquire EEG signals in real time, and the EEG signals include interactive signals such as SSVEP; the EEG acquisition module 2 supports multi-channel signal acquisition, such as 8 channels.

[0168] The EEG signal analysis module 3 is used for the recognition and processing of interactive signals to generate recognition results; for example, performing classification recognition such as FBCCA and CCA-wtCCA.

[0169] Control module 4 executes operations based on the recognition results. Control module 4 incorporates a microcontroller or similar processing chip structure.

[0170] Of course, sensors such as infrared optical trackers, pressure sensors, inertial sensors, and ambient light sensors are all connected to control module 4.

[0171] Comparison table of prior art and the technical solution of this invention

[0172]

[0173]

[0174] The above description only discloses some embodiments of the present invention. For those skilled in the art, various modifications and improvements can be made without departing from the inventive concept of the present invention, and these all fall within the scope of protection of the invention.

Claims

1. A brain-computer interface-based intelligent care method, characterized in that, include: Setting up the user interface: The interactive interface includes a text input interface and a number input interface, each with keys or icons; Set visual stimulus signals: The frequency of the visual stimulus signal for each key or pattern is different, and all of them are non-integer multiples of the fundamental frequency; The visual stimulus signal of each key or pattern is set with phase-reversed alternating PRCP encoding. SSVEP signal identification and processing: SSVEP signals are used for FBCCA classification and recognition.

2. The intelligent care method based on brain-computer interface according to claim 1, characterized in that, In SSVEP signal recognition and processing, notch filtering is applied to the SSVEP signal.

3. The intelligent care method based on brain-computer interface according to claim 1, characterized in that, Both the text input interface and the number input interface have 24 keys arranged in a matrix; The text input interface can intelligently predict word frequency during input; The numeric keys on the numeric input interface are arranged with a linear frequency distribution. Both the text input interface and the number input interface are equipped with emergency call buttons; The interactive interface also includes a functional entertainment interface and a ward management interface; Both the text input interface and the number input interface have an entertainment playback interface on their sides.

4. The intelligent care method based on brain-computer interface according to claim 1, characterized in that, The PRCP encoding setting method: Odd frames: display positive phase; Even-numbered frames: display negative phase; where the positive and negative phases are 180 degrees apart; The phase reversal frequency fm is not an integer multiple of 50Hz or 60Hz.

5. The intelligent care method based on brain-computer interface according to claim 1, characterized in that, In setting the visual stimulus signal, the visual stimulus signal can switch between the SSVEP high-frequency mode and the SSVEP low-frequency mode; The visual stimulus signals for the keys or patterns are set as follows: A combined stimulation modality of scintillation spot and radial motion was employed, and PCA was used to separate the components of the SSVEP and SSVMVEP signals.

6. The intelligent care method based on brain-computer interface according to claim 1, characterized in that, It also features a hybrid interaction mode combining SSVEP and eye tracking; Prioritize SSVEP interaction mode; When the SSVEP signal is lost, switch to eye-tracking interaction mode.

7. The intelligent care method based on brain-computer interface according to claim 6, characterized in that, In eye-tracking interaction mode, an inertial measurement unit (IMU) is used to correct gaze point errors caused by head shift. In eye-tracking interaction mode, blink frequency is monitored. If a blink interval > 5 seconds is detected, eye-tracking interaction mode is paused. Eye-tracking signals undergo redundancy verification to reduce the false recognition rate; Stare at the emergency call button for an extended period of time to make a voice call.

8. The intelligent care method based on brain-computer interface according to claim 1, characterized in that, Configure a progressive interactive mode based on the patient's ALSFRS-R score range; The progressive interaction mode is configured as follows: eye-tracking interaction mode → SSVEP interaction mode → breathing interaction mode; Alternatively, a gradual setup can be implemented based on the stage of the disease; Early-stage patients: Use SSVEP high-frequency mode; For late-stage patients: use SSMVEP low-frequency mode or respiratory interaction mode.

9. A brain-computer interface-based intelligent care method according to claim 5, 6, or 8, characterized in that, A hybrid CCA-wtCCA architecture is also set up to implement multi-stage signal processing; The primary signal undergoes wtCCA processing; The secondary verification signal undergoes cross-interaction mode correlation verification processing for CCA; The final identification result is output after Bayesian decision fusion processing.

10. A brain-computer interface-based intelligent care system, employing the intelligent care method of claim 1, the system comprising: A screen is used to display the interactive interface; EEG acquisition module, including EEG cap and brain-computer interface; The EEG acquisition module is used to acquire EEG signals in real time, and the EEG signals include interactive signals; The EEG signal analysis module is used to identify and process interactive signals to generate recognition results; the control module executes operations based on the recognition results.

Images