Description:FIELD OF THE INVENTION

[0001] The present invention relates to a brain-computer interface system for controlling devices using brain signals enabling real-time translation of motor imagery tasks into commands for external devices, thus classifying them using a neural network to generate real-time commands.

BACKGROUND OF THE INVENTION

[0002] Brain-computer interfaces (BCIs) enable direct communication between the brain and external devices by interpreting neural activity, offering promising applications in assistive technologies, rehabilitation, and human-computer interaction.

[0003] Traditional methods of brain-computer interface (BCI) operation involved collecting EEG signals followed by basic filtering and manual artifact rejection, which often resulted in poor signal clarity and limited classification accuracy for real-time control applications.

[0004] CN113268142A discloses a brain-computer interface system based on a direct flight time measurement technology, brain-computer interface wearable equipment and a control method thereof, and the system comprises a processor which is used for controlling the whole brain-computer interface system; a light source module, a photoelectric sensor, a trans-impedance amplifier, an analog-to-digital converter, a first height comparator and a second high-speed comparator. According to the system, a direct time-of-flight (DDoF) measurement technology is applied to the field of functional near-infrared brain-computer interfaces. By directly measuring the time difference between the incident ultrashort pulse near-infrared light and the emergent ultrashort pulse near-infrared light, the passing distance of the near-infrared light in the brain is calculated, and according to the scheme, the absolute value of the concentration of deoxyhemoglobin and oxyhemoglobin in the brain can be measured; and the position of the blood oxygen change of the brain can be directly measured.

[0005] CN219039708U discloses an electroencephalogram simulation system for a brain-computer interface system, which comprises a connecting device for externally connecting the brain-computer interface system; the transmission device is used for transmitting an externally input brain wave file; the signal conversion device is used for converting an externally input brain wave file into a first analog signal; the control output device is used for outputting the first analog signal to a brain-computer interface system through the connecting device; and the display device is used for displaying waveform information of the first analog signal. Compared with a common laboratory signal generator sold in the market, the signal generator can output analog signals and brain wave signals of a real human brain.

[0006] Conventionally, many systems lacked the ability to extract robust and high-resolution features across multiple frequency bands, thereby limiting the discrimination of complex motor imagery tasks. As a result, these existing systems demonstrated reduced performance in real-time scenarios, required frequent recalibration, and lacked generalization across different users and sessions.

[0007] In order to overcome the aforementioned drawbacks, there exists a need in the art to develop a system that requires to be capable of enhancing the quality of signal acquisition, performs effective noise filtering and enables adaptive and accurate classification of motor imagery tasks for controlling external devices in real time.

OBJECTS OF THE INVENTION

[0008] The principal object of the present invention is to overcome the disadvantages of the prior art.

[0009] An object of the present invention is to develop a brain-computer interface system that acquires raw brain signals enabling direct communication between the brain and external devices.

[0010] Another object of the present invention is to enhance the system’s ability to distinguish different motor imagery tasks effectively.

[0011] Yet another object of the present invention is to ensure that the brain-computer interface system operates with minimal recalibration and adapts to different users and sessions, thereby improving usability for long-term applications.

[0012] The foregoing and other objects, features, and advantages of the present invention will become readily apparent upon further review of the following detailed description of the preferred embodiment as illustrated in the accompanying drawings.

SUMMARY OF THE INVENTION

[0013] The present invention relates to a brain-computer interface system for controlling devices using brain signals enables real-time control of external devices using brain signals, thereby enhancing user-device interaction and supporting practical neurotechnology applications.

[0014] According to an embodiment of the present invention, a brain-computer interface system for controlling devices using brain signals, comprises an EEG signal acquisition unit with electrodes placed on a user's scalp to collect raw brain signals and uses multiple electrodes placed on the scalp to collect brain signals from different brain regions, a four class iterative filtering (FCIF) unit connected to the EEG signal acquisition unit to clean the raw brain signals by removing noise and iteratively removes noise by adjusting to different types of artifacts, and also processes EEG signals in real time to provide clean signals for immediate use in brain-computer interface applications.

[0015] According to another embodiment of the present invention, the system further comprises a four class filter bank common spatial pattern (FC-FBCSP) unit connected to the FCIF unit to pull out key features from the cleaned brain signals and analyzes multiple frequency bands to improve the separation of motor imagery tasks compared to single-frequency methods, a modified deep neural network (MDNN) classifier connected to the FC-FBCSP unit to analyze the features and classify motor imagery tasks adapts to different users and sessions with minimal recalibration, improving usability for long-term brain-computer interface applications and an output device connected to the MDNN classifier to turn the classified signals into commands for controlling devices.

[0016] While the invention has been described and shown with particular reference to the preferred embodiment, it will be apparent that variations might be possible that would fall within the scope of the present invention.

BRIEF DESCRIPTION OF THE DRAWINGS

[0017] These and other features, aspects, and advantages of the present invention will become better understood with regard to the following description, appended claims, and accompanying drawings where:
Figure 1 illustrates a flow chart of a brain-computer interface system for controlling devices using brain signals.

DETAILED DESCRIPTION OF THE INVENTION

[0018] The following description includes the preferred best mode of one embodiment of the present invention. It will be clear from this description of the invention that the invention is not limited to these illustrated embodiments but that the invention also includes a variety of modifications and embodiments thereto. Therefore, the present description should be seen as illustrative and not limiting. While the invention is susceptible to various modifications and alternative constructions, it should be understood, that there is no intention to limit the invention to the specific form disclosed, but, on the contrary, the invention is to cover all modifications, alternative constructions, and equivalents falling within the spirit and scope of the invention as defined in the claims.

[0019] In any embodiment described herein, the open-ended terms "comprising," "comprises,” and the like (which are synonymous with "including," "having” and "characterized by") may be replaced by the respective partially closed phrases "consisting essentially of," consists essentially of," and the like or the respective closed phrases "consisting of," "consists of, the like.

[0020] As used herein, the singular forms “a,” “an,” and “the” designate both the singular and the plural, unless expressly stated to designate the singular only.

[0021] The present invention relates to a brain-computer interface system for controlling devices using brain signals that enhances the reliability and usability of brain-controlled devices by dynamically filtering EEG signals, improving feature extraction across multiple frequency bands, and adapting classification performance to individual users, thereby facilitating consistent and efficient device control across varied environments and sessions.

[0022] Referring to Figure 1, a flow chart of a brain-computer interface system for controlling devices using brain signals is illustrated.

[0023] The system disclosed herein comprises a EEG (Electroencephalography) signal acquisition unit to detect and gather electrical activity generated by the brain. The EEG (Electroencephalography) unit consists of multiple electrodes that are placed non-invasively on the user’s scalp, following standardized placement systems like the 10-20 international system. Each electrode functions as a sensor and is positioned on specific regions of the scalp to monitor brain activity from different areas, such as the motor cortex, which is vital for motor imagery tasks like imagining hand movement. The brain generates electrical signals through neural activity. These signals are extremely weak (usually in the microvolt range). The electrodes detect these signals from the scalp surface.

[0024] The detected raw EEG signals, which include a mixture of brain activity and noise (like eye blinks or muscle movements), are sent from the electrodes to the EEG signal acquisition unit. Since EEG signals are very weak, the acquisition unit amplifies the signals and converts them from analog to digital form for further processing.

[0025] The EEG signal acquisition system utilizes multiple electrodes that are positioned on the scalp of the user to collect electrical activity generated by the brain. These electrodes are distributed across various regions of the scalp, targeting key brain areas such as the motor cortex, frontal lobe, and parietal lobe, which are known to produce relevant motor imagery signals. By capturing signals from different brain regions, the system improves the spatial resolution of the data and enhances the overall signal quality.

[0026] A Four Class Iterative Filtering (FCIF) unit is connected to the EEG signal acquisition unit and is responsible for cleaning the raw brain signals captured from the scalp. These raw signals often contain various types of noise and artifacts, such as eye blinks, muscle contractions, and ambient electrical interference, which can interfere with accurate interpretation. The FCIF unit applies a multistage iterative filtering approach, tailored for four specific motor imagery classes, to progressively remove unwanted components from the signal. Each iteration in the filtering process adapts to the signal's properties to preserve essential neural activity while discarding artifacts.

[0027] The Four Class Iterative Filtering (FCIF) unit iteratively removes noise by dynamically adjusting its filtering parameters in response to the characteristics of different types of artifacts. These artifacts may include ocular movements (like blinking), muscle activity, or other environmental interferences. The FCIF unit identifies the frequency and amplitude patterns of such artifacts and applies adaptive filters over multiple passes. Through each iteration, the unit gradually attenuates noise components while preserving the core neural signals related to motor imagery.

[0028] Further, the FCIF unit processes EEG signals in real time to ensure immediate availability of clean and artifact-free signals. This real-time operation is essential for interactive brain-computer interface applications where immediate response to user intent such as moving a cursor or controlling a robotic arm—is critical.

[0029] A Four-Class Filter Bank Common Spatial Pattern (FC-FBCSP) unit is a critical component in a brain-computer interface (BCI) system designed for motor imagery tasks, such as imagining moving the left hand, right hand, foot, or tongue. It processes cleaned brain signals, EEG, received from a preceding Fully-Connected Input Feature (FCIF) unit. The filter bank aspect involves splitting these signals into multiple frequency bands (e.g., alpha: 8–12 Hz, beta: 12–30 Hz) using bandpass filters, as different motor imagery tasks produce distinct patterns in specific frequency ranges. The Common Spatial Pattern (CSP) protocol is then applied to each frequency band to extract spatial features, which are mathematical transformations that maximize the differences in brain activity between the four classes, enhancing the ability to distinguish between tasks.

[0030] By combining the filter bank and CSP approaches, the FC-FBCSP unit generates a robust set of features that capture both frequency-specific and spatial characteristics of the brain signals. These features are essential for accurately identifying the user’s intended motor imagery task, as they highlight unique patterns in brain activity associated with each imagined movement. The extracted features are fed into a classifier to determine which of the four tasks the user is performing, enabling applications like controlling devices or prosthetics through thought alone.

[0031] The FC-FBCSP (Four Class Filter Bank Common Spatial Pattern) unit processes the cleaned EEG signals by dividing them into multiple frequency bands rather than analyzing just one frequency band. Each frequency band captures different aspects of the brain's electrical activity related to motor imagery tasks (such as imagining moving the left hand or right hand). By analyzing multiple frequency bands, the unit can better distinguish between different mental commands, improving the accuracy and reliability of identifying which specific motor imagery task the user is performing.

[0032] The output device receives the classification results from the Modified Deep Neural Network (MDNN) classifier, which identifies the user’s intended motor imagery tasks based on brain signals. This device converts these classified signals into actionable commands, enabling real-time control of external devices such as moving a computer cursor, operating a robotic arm, or controlling other assistive technologies.

[0033] The Modified Deep Neural Network (MDNN) classifier described in the present invention is designed to adapt to different users and usage sessions with minimal need for recalibration. Traditional brain-computer interface systems often require extensive and repeated calibration for each user, as brain signal patterns can vary significantly between individuals and even across different sessions for the same person. This invention addresses that limitation by employing an MDNN classifier capable of learning and adjusting to new brain signal patterns without requiring complete retraining.

[0034] Such adaptability is achieved through techniques like transfer learning or continual learning, enabling the classifier to maintain high accuracy and responsiveness even as signal characteristics shift due to factors like electrode repositioning or user fatigue.

[0035] The present invention works best in the following manner, where the brain-computer interface (BCI) system is configured for real-time control of external devices through interpretation of brain signals. The user first wears an EEG signal acquisition unit comprising multiple electrodes positioned over the scalp, enabling the collection of raw electrical signals generated by neural activity from different brain regions. These raw EEG signals, which often include noise such as eye blinks, facial muscle activity, or external interference, are passed to the Four Class Iterative Filtering (FCIF) unit. The FCIF unit processes the incoming data in real time and removes artifacts while preserving essential neural features relevant to motor imagery, using adaptive filtering iterations that refine the quality of the signal.

[0036] In continuation, the filtered, clean signals are then forwarded to the Four Class Filter Bank Common Spatial Pattern (FC-FBCSP) unit, applying spatial and frequency-based decomposition across multiple bandpass filters to extract discriminative features corresponding to different motor imagery tasks (e.g., imagining left hand, right hand, feet, or tongue movement). The extracted features are sent to the Modified Deep Neural Network (MDNN) classifier, which has been trained to associate different EEG feature patterns with specific user intentions. The MDNN classifier uses its layered structure to identify the intended motor imagery task with high accuracy. Once classification is completed, the output is transmitted to an external output device, such as a robotic arm, wheelchair interface, or computer cursor.

[0037] Although the field of the invention has been described herein with limited reference to specific embodiments, this description is not meant to be construed in a limiting sense. Various modifications of the disclosed embodiments, as well as alternate embodiments of the invention, will become apparent to persons skilled in the art upon reference to the description of the invention. , Claims:1) A brain-computer interface system for controlling devices using brain signals, comprising:

i) an EEG signal acquisition unit with electrodes placed on a user's scalp to collect raw brain signals;
ii) a four class iterative filtering (FCIF) unit connected to the EEG signal acquisition unit to clean the raw brain signals by removing noise, such as eye blinks and muscle movements;
iii) a four class filter bank common spatial pattern (FC-FBCSP) unit connected to the FCIF unit to pull out key features from the cleaned brain signals across multiple frequency bands to identify different motor imagery tasks;
iv) a modified deep neural network (MDNN) classifier connected to the FC-FBCSP unit to analyze the features and classify motor imagery tasks, like imagining left or right hand movement; and
v) an output device connected to the MDNN classifier to turn the classified signals into commands for controlling devices, like moving a cursor or robotic arm, in real time.

2) The device as claimed in claim 1, wherein the EEG signal acquisition system uses multiple electrodes placed on the scalp to collect brain signals from different brain regions for better signal quality.

3) The system as claimed in claim 1, wherein the FCIF unit iteratively removes noise by adjusting to different types of artifacts, ensuring the cleaned EEG signals keep important neural activity for accurate motor imagery detection.

4) The system as claimed in claim 1, wherein the FCIF unit processes EEG signals in real time to provide clean signals for immediate use in brain-computer interface applications.

5) The system as claimed in claim 1, wherein the FC-FBCSP unit analyzes multiple frequency bands to improve the separation of motor imagery tasks compared to single-frequency methods.

6) The system as claimed in claim 1, wherein the MDNN classifier adapts to different users and sessions with minimal recalibration, improving usability for long-term brain-computer interface applications.