1. A method (700) comprising: i. collecting (701) one or more bio-signals from a user of a wearable electronic device (100); ii. identifying (702) one or more Electro-Magnetic (EM) noises present in a surrounding environment of the wearable electronic device (100) during the collection of the one or more bio-signals; iii. scaling (703) the one or more collected bio-signals, and the one or more identified EM noises; iv. analyzing (704) the one or more scaled bio-signals, and the one or more scaled EM noises by utilizing at least one Machine Learning (ML) model; v. generating (705) a compound noise signal (EMtotal) based on a result of the analysis; and vi. cancelling (706) the generated compound noise signal (EMtotal) from the one or more collected bio-signals, to enhance quality of the one or more collected bio-signals.

2. The method (700) as claimed in claim 1, i. wherein the one or more bio-signals comprise an Electroencephalography (EEG) signal, an Electrooculography (EOG) signal, an Electromyography (EMG) signal, and an Electrocardiography (ECG) signal; and ii. wherein the one or more EM noises comprise at least one of an ambient electromagnetic noise (EMambient), an EM noise from speakers (EMspeaker), an EM noise from internal device components (EMradio), and an impulsive noise (EMother).

3. The method (700) as claimed in claim 1, wherein identifying the one or more EM noises present in the surrounding environment of the wearable electronic device (100) comprises: i. detecting, via an EM sensor, the one or more EM noises, wherein the EM sensor comprises at least one of a loop antenna, a dipole antenna, an antenna array, and a Micro-Electro-Mechanical Systems (MEMS) based Inductor-Capacitor (LC) tank circuit; ii. quantifying the one or more EM noises detected by the EM sensor to generate one or more measurable parameters indicative of one or more EM noise characteristics; and iii. classifying the one or more quantified EM noises, based on a predefined classification mechanism, by utilizing at least one ML model, to identify the one or more EM noises present in the surrounding environment of the wearable electronic device (100).

4. The method (700) as claimed in claim 1, further includes integrating an auxiliary electrode probe placed in close proximity to one or more primary active electrodes to sample the electromagnetic noise impacting the one or more primary active electrodes. i. capturing, via the auxiliary electrode probe, real-time electromagnetic noise signals associated with interference sources that affect the primary active electrodes, wherein the captured real-time electromagnetic noise signals represent as external electromagnetic noise signals; ii. comparing the captured real-time electromagnetic noise signals with electromagnetic noise signals identified by an internal electromagnetic noise detection system within the wearable electronic device (100); iii. analyzing discrepancies between the external and internal electromagnetic noise signals to identify potential anomalies, including leakage of uncanceled noise or inconsistencies in hardware performance; and iv. dynamically adjusting one or more noise cancellation parameters of the wearable electronic device (100) based on the results of the comparison, to improve accuracy in biosignal isolation and enhance the overall quality of the one or more collected bio-signals.

5. The method (700) as claimed in claim 1, wherein, when the one or more EM noises comprise an ambient electromagnetic noise (EMambient), identifying the one or more EM noises comprises: i. determining a presence of one or more electronic devices within the surrounding environment that emit EM interference; and ii. identifying the ambient electromagnetic noise (EMambient) based on the one or more determined electronic devices, wherein the ambient electromagnetic noise (EMambient) is characterized as relatively constant over time and comprises the electromagnetic interference generated by the one or more determined electronic devices.

6. The method (700) as claimed in claim 1, wherein, when the one or more EM noises comprise an EM noise from speakers (EMspeaker), identifying the one or more EM noises comprises: i. determining one or more characteristics of an input audio signature and EM noise metadata, wherein the EM noise metadata is a descriptive file formatted according to a standard specification, wherein the EM noise metadata is designed to be paired with one or more audio files intended for playback, wherein the EM noise metadata comprises information associated with one or more ML models that weights pertinent to the EM noise emissions anticipated from a variety of speaker models and sound configurations during the playback of the one or more audio files; and ii. identifying the EM noise from speakers (EMspeaker) based on the one or more determined characteristics.

7. The method (700) as claimed in claim 6, further comprising: a. predicting EM noise generated by speaker movements during active Acoustic Noise Cancellation (ANC) operations in the wearable electronic device (100); b. analyzing, in real-time, an ANC input signal transmitted to the speaker, wherein the ANC signal includes an anti-noise waveform generated to cancel external acoustic noise; c. modeling the ANC-induced EM noise caused by rapid diaphragm oscillations and magnetic field fluctuations, using predefined speaker metadata and operational parameters, including input signal characteristics, driver efficiency, and coil dynamics; and d. incorporating a predicted ANC-induced EM noise based on the modeling into the EMspeaker classification, ensuring that a total electromagnetic noise profile accurately accounts for all speaker-related emissions, to improve biosignal isolation and enhance the overall quality of the one or more collected bio-signals.

8. The method (700) as claimed in claim 1, wherein, when the one or more EM noises comprise an EM noise from internal device components (EMradio), identifying the one or more EM noises comprises: i. determining a presence of one or more electronic devices within the surrounding environment that emit EM interference, ii. determining, by utilizing the one or more ML models, one or more radio parameters comprise an operation characteristic of a radio circuit associated with the one or more determined electronic devices, data ingress and egress within the radio circuit, one or more radio transmission principles; and iii. identifying the EM noise from internal device components (EMradio) based on the one or more determined radio parameters.

9. The method (700) as claimed in claim 1, i. wherein an impulsive noise (EMother) associated with the one or more EM noises comprises at least one category of short-duration, high-amplitude bursts produced by one or more sources; and ii. wherein the one or more sources comprise at least one of one or more atmospheric events and one or more domestic appliances.

10. The method (700) as claimed in claim 1, wherein the at least one ML model is trained by: i. performing one or more data processes on the identified electromagnetic (EM) noises, wherein the data processes comprise data integration to consolidate various data sources, data cleaning to remove inaccuracies and inconsistencies, data reduction to minimize data volume while retaining essential information, data transformation to convert data into suitable formats for analysis, and data discretization to convert continuous data into discrete values for improved processing; ii. splitting the processed EM noises into at least one of a training set, validation set, and testing set, wherein the training set is used to train the ML model, the validation set is utilized for optimizing hyperparameters, and the testing set is reserved for unbiased performance evaluation of the trained ML model; iii. feeding the split data into the ML model for training, wherein the training process comprises model initialization to establish starting parameters, loss function selection to define error measurements, optimizer selection to determine algorithms for parameter updates, and hyperparameter tuning to refine model performance; and iv. performing evaluation processes to determine the optimal ML model by assessing performance metrics, conducting testing set evaluations, and executing error analysis to identify and correct prediction errors.

11. The method (700) as claimed in claim 1, further comprising: i. implementing the at least one ML model, comprising an edge computing module within the wearable electronic device (100), to enable real-time processing and analysis of bio-signals and identified EM noises; ii. Utilizing the at least one ML model comprising one or more transformer models to classify and predict noise patterns from the scaled bio-signals and EM noises; and iii. employing one or more advanced ML models comprising Large Language Models (LLMs), to analyze metadata and dynamically optimize noise cancellation parameters based on real-time environmental conditions and historical data; and iv. defining the compound noise signal (EMtotal) as a sum of all EM noises.

12. The method (700) as claimed in claim 1, wherein enhancing the quality of the one or more collected bio-signals comprises: i. assigning a dynamic weight to each identified EM noise, wherein the weight is continuously adjusted by the at least one ML model based on real-time data and contextual information; ii. continuously monitoring the surrounding environment to enable the at least one ML model to discern the significance of various identified EM noises; and iii. prioritizing one or more most impactful EM noise while effectively mitigating one or more less impactful EM noise, for efficient noise cancellation.

13. The method (700) as claimed in claim 1, wherein cancelling, by the wearable electronic device (100), the generated compound noise signal (EMtotal) from the one or more collected bio-signals comprising: i. performing at least one of : a) programmatically canceling the one or more EM noises from the one or more collected bio-signals in real-time based on at least one mechanism comprising an actual playback mechanism, a simulation-based mechanism, and a metadata-based mechanism; or b) canceling the one or more EM noises from the one or more collected bio-signals by utilizing one or more adaptive filtering mechanisms.

14. The method (700) as claimed in claim 13, wherein programmatically canceling the one or more EM noises from the one or more collected bio-signals in real-time based on the metadata-based mechanism comprising: i. associating an audio file with a metadata file that specifies one or more expected electromagnetic noise characteristics generated by one or more speaker and radio modules during a playback, a) wherein the metadata file describes the noise profile based on one or more audio signal properties comprising a frequency, an amplitude, and one or more anticipated interference patterns from one or more operations associated with the one or more speaker and radio modules; and ii. referencing, during the playback, the metadata file to predict one or more generated EM noises, to eliminate a real-time computation, and to reduce processing load by avoiding real-time calculations; a) wherein the metadata file enables an application of one or more pre-configured filters and adaptive algorithms to efficiently cancel one or more pre-defined EM noises, to ensure highly accurate noise cancellation tailored to a specific audio file and device configuration.

15. The method (700) as claimed in claim 13, wherein programmatically canceling the one or more EM noises from the one or more collected bio-signals in real-time based on the metadata-based mechanism comprising: i. employing a unit test signal to validate identification and software-based removal of sounds, generated by a speaker, based on a metadata file; ii. detecting one or more deviations during the employment of the unit test signal to assess an accuracy of an audio cancellation; and iii. dynamically editing and updating the metadata file in response to detecting the one or more deviations, reflecting predefined acoustic characteristics under current usage conditions, for canceling the one or more EM noises from the one or more collected bio-signals in real-time.

16. The method (700) as claimed in claim 15, further comprising: i. storing the dynamically edited and updated metadata file in a user profile on a cloud server, for ensuring preservation and synchronization across one or more electronic devices; and ii. downloading and applying, upon synchronization across the one or more electronic devices, the saved metadata file, to ensure consistent audio cancellation across the one or more electronic devices associated with the user profile, a) wherein each metadata file is associated with a version, and each version indicates one or more changes made to the metadata file, ensuring the use of the most recent and accurate file while preserving previous versions for potential rollback, b) wherein each metadata file is stored in the cloud server along with one or more annotations related to one or more environmental conditions.

17. The method (700) as claimed in claim 1, wherein identifying one or more EM noises associated with the wearable electronic device (100) comprises: i. generating an acoustic metadata file associated with a speaker system within the wearable electronic device (100), wherein the acoustic metadata file includes: a) a frequency response profile defining output characteristics of the speaker system across a predefined frequency range; b) harmonic distortion data at one or more operating frequencies and amplitudes; c) one or more transient response times of the speaker system indicating stabilization delays after input signals; d) one or more phase response characteristics correlating input and output signals across key frequencies; and e) one or more speaker-specific operational profiles tailored for different playback modes; ii. associating the acoustic metadata file with audio input signals intended for playback as well as the inputs for Active acoustic Noise cancellation (ANC), to predict the speaker's acoustic output; iii. detecting a composite sound signal from within an ear canal, wherein the composite sound signal includes the predicted acoustic output, ambient noise, and one or more acoustic biosignals comprising a sound from the blood vessels in the vicinity of ear canal; iv. isolating, via the acoustic metadata file, the predicted acoustic output of the speaker system from the composite sound signal to enhance the accuracy of biosignal extraction; and v. dynamically updating the acoustic metadata file in response to detected deviations during real-time operation, ensuring accurate noise isolation and adaptive audio cancellation in varying conditions.

18. The method (700) as claimed in claim 1, comprising: i. determining whether a value of an ambient electromagnetic noise (EMambient) associated with the one or more EM noises is greater than a value of an electromagnetic noise threshold (EMthreshold); and ii. performing one of: a) collecting the one or more bio-signals from the user of the wearable electronic device (100) in response to determining that the value of the ambient electromagnetic noise (EMambient) is lower than the value of an electromagnetic noise threshold (EMthreshold); and b) transmitting an alert message to the user in response to determining that the value of the ambient electromagnetic noise (EMambient) is greater than the value of an electromagnetic noise threshold (EMthreshold), wherein the alert message indicates that an EM noise level is too high for collecting the one or more bio-signals and if the user wants to continue. c) Warning the user if the detected EM noise exceeds safe limits for human exposure, indicating a potential health risk.

19. A system (101) comprising: i. a memory (110); ii. a processor (120); iii. an Electro-Magnetic (EM) noise controller (121), operably connected to the memory (110) and the processor (120), configured to: a) collect one or more bio-signals from a user of the wearable electronic device (100); b) identify one or more Electro-Magnetic (EM) noises present in a surrounding environment of a wearable electronic device (100) during the collection of the one or more bio-signals; c) scale the one or more collected bio-signals, and the one or more identified EM noises; d) analyze the one or more scaled bio-signals and the one or more scaled EM noises by utilizing at least one Machine Learning (ML) model; e) generate a compound noise signal (EMtotal) based on a result of the analysis; and f) cancel the generated compound noise signal (EMtotal) from the one or more collected bio-signals, to enhance quality of the one or more collected bio-signals.

20. The system (101) as claimed in claim 19, wherein canceling the one or more EM noises from the one or more collected bio-signals in real-time based on a metadata-based mechanism comprising: i. associating an audio file with a metadata file that specifies one or more expected electromagnetic noise characteristics generated by one or more speaker and radio modules during a playback, a) wherein the metadata file describes the noise profile based on one or more audio signal properties comprising a frequency, an amplitude, and one or more anticipated interference patterns from one or more operations associated with the one or more speaker and radio modules; and ii. referencing, during the playback, the metadata file to predict one or more generated EM noises, to eliminate a real-time computation, and to reduce processing load by avoiding real-time calculations; a) wherein the metadata file enables an application of one or more pre-configured filters and adaptive algorithms to efficiently cancel one or more pre-defined EM noises, to ensure highly accurate noise cancellation tailored to a specific audio file and device configuration.

DESC:FIELD OF THE INVENTION

The present invention generally relates to the field of electronic devices, and more specifically relates to a method and a system for Electro-Magnetic (EM) noise detection and cancellation.
BACKGROUND

With the growing demand for health monitoring devices, there has been a parallel increase in the emission of electromagnetic (EM) radiation across various frequency ranges. Common electronic devices such as Wi-Fi modules, 5G antennas, mobile phones, smartwatches, routers, wireless chargers, microwaves, speakers, and compressors generate electromagnetic fields (EMF) that contribute to a significant rise in ambient EM noise.
This pervasive EM radiation, while sometimes harmless, can be potentially harmful, especially in the context of collecting low-power bio-signals like Electroencephalography (EEG), Electrooculography (EOG), Electromyography (EMG), and Electrocardiography (ECG). External EM noise can interfere with the quality of these bio-signals, leading to inaccurate readings and compromising both clinical diagnostics and research outcomes. Moreover, prolonged exposure to elevated levels of EM noise poses potential health risks to living organisms, including humans, necessitating careful mitigation strategies.
For instance, when sensitive bio signal circuitry, such as EEG, ECG,EMG, or EOG, is positioned in proximity to electromagnetic (EM) noise-generating devices such as audio speakers, radios (e.g., Bluetooth, Wi-Fi transmitters), or processors there is a considerable likelihood that EM noise may couple into the biosensing circuitry. This interference can manifest as disturbances, signal distortions, and unwanted artifacts within the recorded bio signals. Such artifacts degrade the quality and accuracy of biosensing, making it challenging to extract meaningful physiological information.
EM noise may originate from multiple sources, including devices within a wearable system itself, nearby electronic devices, or ambient environmental factors. These signals can overlap with or mask the weak bioelectric signals being measured, thereby reducing the reliability of the data. Given the highly sensitive nature of biosignal measurements, even minimal noise interference can result in significant errors or misinterpretations of the biosignal data.
To achieve precise and reliable biosensing, it is critical to employ a robust mechanism capable of detecting, isolating, and canceling out the EM noise in real time. This mechanism must not only address the EM noise from known internal sources, such as integrated processors or communication modules, but also adapt to dynamically changing ambient EM noise present in the surrounding environment.
There are certain existing methods or solutions currently available to detect and cancel the EM interference (i.e., EM noise). However, several problems are encountered in the existing methods or solutions, which are mentioned below. Existing methods or solutions for detecting and canceling EM noise are mainly intended for large-scale applications, like protecting rooms from external noise or shielding specific parts of electronic devices. These approaches are not compact or suitable for wearable electronic devices, and they fail to address the specific challenges of capturing bio-signals in active real-world environments, for example:
Scale and applicability: Most existing methods or solutions are built for stationary, large setups, making them impractical for portable wearable electronic devices. They do not effectively handle ongoing noise interference in smaller devices.
Specificity to bio-signal acquisition: Unlike standard electronics, devices that capture bio-signals, such as EEG machines, deal with very weak signals that are highly sensitive to external EM noise. Existing methods or solutions do not meet the strict needs for reducing noise to ensure clear signal quality.
Complex interference in wearable electronic devices: the wearable electronic devices face added complications because they include components like Bluetooth modules, which create unpredictable high-frequency noise, and speakers with magnetic diaphragms that generate significant EM noise during use.
Combination of use case challenges: In compact ear-worn electronic devices, the proximity of sensitive bio-signal electrodes and noise-producing components (like speakers and radio modules) leads to unique interference problems. This closeness increases noise contamination, making it hard for traditional shielding or cancellation methods to effectively isolate unwanted noise.
Thus, it is desired to address the above-mentioned disadvantages or other shortcomings or at least provide a useful alternative for the EM noise detection and cancellation.

SUMMARY

The present invention provides a method and/or system for detecting and canceling Electro Magnetic (EM) noise in wearable electronic devices, particularly those used for bio-signal acquisition (e.g., EEG, EMG, ECG). The invention addresses limitations in current EM artifact reduction solutions, which are unsuitable for compact, wearable devices due to their inability to mitigate dynamic, multi-source EM noise in real time.
The disclosed method involves collecting bio-signals while identifying EM noise from one or more internal device components, ambient environments, and impulsive sources. Using specialized sensors and Machine Learning (ML) models, the disclosed method and /or system classifies, scales, and analyzes the bio-signals and EM noise, generating a compound noise signal. This signal is adaptively canceled in real time to ensure high-fidelity bio-signal acquisition.
The disclosed method and/or system further integrates metadata-based mechanisms, edge computing, and real-time calibration to improve noise prediction and mitigation. By leveraging advanced hardware and software, including ML models and signal processing algorithms, the invention ensures precise EM noise cancellation even in complex, real-world conditions, enabling reliable health monitoring and diagnostics through wearable technology.

BRIEF DESCRIPTION OF THE DRAWINGS

These and other features, aspects, and advantages of the present invention will become better understood when the following detailed description is read with reference to the accompanying drawings in which like characters represent like parts throughout the drawings, wherein:
FIG. 1 illustrates a block diagram of a wearable electronic device for Electro-Magnetic (EM) noise detection and cancellation, according to an embodiment as disclosed herein;
FIG. 2 is a flow diagram illustrating a method for the EM noise detection and cancellation, according to an embodiment as disclosed herein;
FIG. 3 is a flow diagram illustrating a method for environmental diagnosis for the EM noise detection and cancellation, according to an embodiment as disclosed herein;
FIG. 4 illustrates operations for training a Machine Learning (ML) model for electromagnetic (EM) noise detection and cancellation, according to an embodiment as disclosed herein;
FIG. 5 illustrates an EM noise sensing system for EM signal detection (Antenna), according to an embodiment as disclosed herein;
FIG. 6 is a flow diagram illustrating a method for an adaptive sensor resolution adjusting for the EM noise detection and cancellation, according to an embodiment as disclosed herein; and
FIG. 7 is a flow diagram illustrating a method for the EM noise detection and cancellation, according to another embodiment as disclosed herein.
Further, skilled artisans will appreciate that elements in the drawings are illustrated for simplicity and may not have necessarily been drawn to scale. For example, the flow charts illustrate the method in terms of the most prominent steps involved to help to improve understanding of aspects of the present invention. Furthermore, in terms of the construction of the device, one or more components of the device may have been represented in the drawings by conventional symbols, and the drawings may show only those specific details that are pertinent to understanding the embodiments of the present invention so as not to obscure the drawings with details that will be readily apparent to those of ordinary skill in the art having the benefit of the description herein.

DETAILED DESCRIPTION OF FIGURES

For the purpose of promoting an understanding of the principles of the invention, reference will now be made to the embodiment illustrated in the drawings and specific language will be used to describe the same. It will nevertheless be understood that no limitation of the scope of the invention is thereby intended, such alterations and further modifications in the illustrated system, and such further applications of the principles of the invention as illustrated therein being contemplated as would normally occur to one skilled in the art to which the invention relates.
It will be understood by those skilled in the art that the foregoing general description and the following detailed description are explanatory of the invention and are not intended to be restrictive thereof.
Reference throughout this specification to “an aspect”, “another aspect” or similar language means that a particular feature, structure, or characteristic described in connection with the embodiment is included in at least one embodiment of the present invention. Thus, appearances of the phrase “in an embodiment”, “in one embodiment”, “in another embodiment”, and similar language throughout this specification may, but do not necessarily, all refer to the same embodiment.
The terms “comprise”, “comprising”, or any other variations thereof, are intended to cover a non-exclusive inclusion, such that a process or method that comprises a list of steps does not include only those steps but may include other steps not expressly listed or inherent to such process or method. Similarly, one or more devices or sub-systems or elements or structures or components proceeded by “comprises... a” does not, without more constraints, preclude the existence of other devices or other sub-systems or other elements or other structures or other components or additional devices or additional sub-systems or additional elements or additional structures or additional components.
The embodiments herein and the various features and advantageous details thereof are explained more fully with reference to the non-limiting embodiments that are illustrated in the accompanying drawings and detailed in the following description. Descriptions of well-known components and processing techniques are omitted so as to not unnecessarily obscure the embodiments herein. Also, the various embodiments described herein are not necessarily mutually exclusive, as some embodiments can be combined with one or more other embodiments to form new embodiments. The term “or” as used herein, refers to a non-exclusive or unless otherwise indicated. The examples used herein are intended merely to facilitate an understanding of ways in which the embodiments herein can be practiced and to further enable those skilled in the art to practice the embodiments herein. Accordingly, the examples should not be construed as limiting the scope of the embodiments herein.
As is traditional in the field, embodiments may be described and illustrated in terms of blocks that carry out a described function or functions. These blocks, which may be referred to herein as units or modules or the like, are physically implemented by analog or digital circuits such as logic gates, integrated circuits, microprocessors, microcontrollers, memory circuits, passive electronic components, active electronic components, optical components, hardwired circuits, or the like, and may optionally be driven by firmware and software. The circuits may, for example, be embodied in one or more semiconductor chips, or on substrate supports such as printed circuit boards and the like. The circuits constituting a block may be implemented by dedicated hardware, or by a processor (e.g., one or more programmed microprocessors and associated circuitry), or by a combination of dedicated hardware to perform some functions of the block and a processor to perform other functions of the block. Each block of the embodiments may be physically separated into two or more interacting and discrete blocks without departing from the scope of the invention. Likewise, the blocks of the embodiments may be physically combined into more complex blocks without departing from the scope of the invention.
The accompanying drawings are used to help easily understand various technical features and it should be understood that the embodiments presented herein are not limited by the accompanying drawings. As such, the present disclosure should be construed to extend to any alterations, equivalents, and substitutes in addition to those which are particularly set out in the accompanying drawings. Although the terms first, second, etc. may be used herein to describe various elements, these elements should not be limited by these terms. These terms are generally only used to distinguish one element from another.
In one or more embodiments, the disclosed method or system is designed to identify and mitigate electromagnetic (EM) noise in wearable devices utilized for bio-signal acquisition, as discussed throughout the disclosure (FIGS. 1 to 7). The disclosed method or system employs advanced techniques to intelligently measure ambient EM noise, utilizing collected data to improve noise cancellation and ensure the reliable capture of high-quality bio-signals, including but not limited to EEG (e.g., ear EEG), EOG (e.g., ear EOG), ECG (e.g., ear ECG), and EMG (e.g., ear EMG) readings. This capability is particularly critical in environments characterized by significant electromagnetic interference, especially when sensors are positioned in areas such as the periphery of the ear, which are susceptible to heightened EM disturbances compared to sensors inside the ear. The disclosed method or system incorporates specialized sensors, metadata, and sophisticated algorithms for detecting EM disturbances, employing a precise Machine Learning (ML) model for active EM noise reduction. By continuously monitoring and adapting to the surrounding environment, the ML model evaluates the impact of various noise sources and assigns or adjusts one or more dynamic weights for the noise cancellation process to prioritize the most significant factors while effectively mitigating those of lesser impact. This one or more dynamic weights assignment not only ensures efficient noise cancellation but also enhances the overall performance of wearable devices across diverse conditions, ultimately leading to accurate and reliable bio-signal measurements, which may be used for a variety of applications, such as medical monitoring, sports performance monitoring, and brain-computer interfaces.
The disclosed method or system addresses the distinct challenges associated with certain existing methods or solutions; for example, an ear-worn bio-signal acquisition device specifically designed for recording EEG signals. The disclosed method or system is characterized by several key differentiators:
Compact, real-time noise mitigation: The disclosed method or system incorporates sophisticated machine learning algorithms that enable the dynamic classification and cancellation of noise from various sources in real-time, as described in conjunction with FIG. 1, FIG. 2, and FIG. 4. This capability is essential for maintaining the integrity of the acquired EEG signals amidst environmental disturbances.
Targeted noise detection: The disclosed method or system employs adaptive mechanisms that effectively distinguish and mitigate noise originating from co-located components, as described in conjunction with FIG. 2, FIG. 3, and FIG. 6. This includes the ability to identify and manage interference from ambient electromagnetic noise, radio-emitting modules, and speaker-generated sounds, thereby enhancing the signal quality.
Optimized for feeble signals: Recognizing that EEG signals are among the weakest bio-signals, the disclosed method or system is specifically tailored to preserve their integrity. It ensures high fidelity in signal acquisition, even in environments characterized by significant interference, thereby facilitating accurate data collection for subsequent analysis, as described in conjunction with FIG. 2 and FIG. 6.
Hardware and software design: The approach leverages specialized sensors, adaptive scaling techniques, and advanced signal processing methodologies, as described in conjunction with FIG. 1, FIG. 2, FIG. 5, and FIG. 6. This combination allows for superior noise cancellation capabilities within a compact and wearable format, addressing the practical constraints of ear-worn devices.
Overall, the disclosed method or system uniquely integrates advanced hardware and software techniques to overcome the limitations inherent in existing large-scale electromagnetic noise cancellation methods. It presents a targeted and portable solution for the effective acquisition of bio-signals in wearable technology, thereby advancing the field of bio-signal monitoring and analysis, as discussed throughout the disclosure.
Referring now to the drawings, and more particularly to FIGS. 1 to 7, where similar reference characters denote corresponding features consistently throughout the figures, there are shown preferred embodiments.
FIG. 1 illustrates a block diagram of a wearable electronic device 100 for Electro-Magnetic (EM) noise detection and cancellation, according to an embodiment as disclosed herein. Examples of the wearable electronic device 100 include, but are not limited to a fitness tracker, a smartwatch, a health monitoring patch, and a wearable ECG monitor.
In one or more embodiments, the wearable electronic device 100 comprises a system 101. The system 101 may include a memory 110, a processor 120, and a communicator 130. In one or more embodiments, the system 101 may be implemented at one or multiple electronic devices (not shown in FIG.1).
In one or more embodiments, the memory 110 stores instructions to be executed by the processor 120 for the EM noise detection and cancellation, as discussed throughout the disclosure. The memory 110 may include non-volatile storage elements. Examples of such non-volatile storage elements may include magnetic hard discs, optical discs, floppy discs, flash memories, or forms of electrically programmable memories (EPROM) or electrically erasable and programmable (EEPROM) memories. In addition, the memory 110 may, in some examples, be considered a non-transitory storage medium. The term “non-transitory” may indicate that the storage medium is not embodied in a carrier wave or a propagated signal. However, the term “non-transitory” should not be interpreted that the memory 110 is non-movable. In some examples, the memory 110 can be configured to store larger amounts of information than the memory. In certain examples, a non-transitory storage medium may store data that can, over time, change (e.g., in Random Access Memory (RAM) or cache). The memory 110 can be an internal storage unit, or it can be an external storage unit of the wearable electronic device 100, a cloud storage, or any other type of external storage.
In one or more embodiments, the processor 120 communicates with the memory 110 and the communicator 130. The processor 120 is configured to execute instructions stored in the memory 110 and to perform various processes for the EM noise detection and cancellation, as discussed throughout the disclosure. The processor 120 may include one or a plurality of processors, maybe a general-purpose processor, such as a Central Processing Unit (CPU), an Application processor (AP), or the like, a graphics-only processing unit such as a Graphics Processing Unit (GPU), a Visual Processing Unit (VPU), and/or an Artificial intelligence (AI) dedicated processor such as a Neural Processing Unit (NPU).
In one or more embodiments, the processor 120 may include an Electro-Magnetic (EM) noise controller 121. The EM noise controller 121 is implemented by processing circuitry such as logic gates, integrated circuits, microprocessors, microcontrollers, memory circuits, passive electronic components, active electronic components, optical components, hardwired circuits, or the like, and may optionally be driven by firmware. The circuits may, for example, be embodied in one or more semiconductor chips, or on substrate supports such as printed circuit boards and the like.
In one or more embodiments, the EM noise controller 121 is configured to collect one or more bio-signals from a user of the wearable electronic device 100. The one or more bio-signals may include, but are not limited to, an Electroencephalography (EEG) signal, an Electrooculography (EOG) signal, an Electromyography (EMG) signal, and an Electrocardiography (ECG) signal. In one or more embodiments, the EM noise controller 121 is configured to receive the one or more bio-signals from one or more other electronic devices (e.g., ECG monitor, EEG devices, etc.) or server (not shown in FIG. 1).
In one or more embodiments, the EM noise controller 121 is configured to identify one or more EM noises present in a surrounding environment of the wearable electronic device 100 during the collection of the one or more bio-signals. The one or more EM noises may include, but is not limited to, an ambient electromagnetic noise (EMambient), an EM noise from speakers (EMspeaker), an EM noise from internal device components (EMradio), and an impulsive noise (EMother). To identify the one or more EM noises, the EM noise controller 121 is configured to perform multiple operations, which are given below and which may relate to FIG. 2.
The EM noise controller 121 is configured to detect, via an EM sensor, the one or more EM noises. The EM sensor may include, but is not limited to, a loop antenna, a dipole antenna, an antenna array, and a Micro-Electro-Mechanical Systems (MEMS) based Inductor-Capacitor (LC) tank circuit, which may relate to FIG. 5. The EM noise controller 121 is further configured to quantifying the one or more EM noises detected by the EM sensor to generate one or more measurable parameters indicative of one or more EM noise characteristics. The EM noise controller 121 is further configured to classify the one or more quantified EM noises, based on a predefined classification mechanism, by utilizing at least one ML model, to identify the one or more EM noises present in the surrounding environment of the wearable electronic device.
To identify the EMambient, the EM noise controller 121 may perform multiple operations. For instance, the EM noise controller 121 is configured to determine a presence of one or more electronic devices within the surrounding environment that emit EM interference. Examples of the one or more electronic devices may include, but are not limited to, a Wi-Fi module, an antenna, a mobile phone, a smartwatch, a Wi-Fi router, a wireless charger, a microwave, a speaker, a compressor, and a transformer. The EM noise controller 121 is further configured to identify the ambient electromagnetic noise (EMambient) based on the one or more determined electronic devices. The ambient electromagnetic noise (EMambient) is characterized as relatively constant over time and comprises the electromagnetic interference generated by the one or more determined electronic devices.
To identify the EMspeaker, the EM noise controller 121 may perform multiple operations. For instance, the EM noise controller 121 is configured to determine one or more characteristics of an input audio signature and EM noise metadata. The EM noise metadata is a descriptive file formatted according to a standard specification. The EM noise metadata is designed to be paired with one or more audio files intended for playback. The EM noise metadata comprises information associated with one or more ML models that weights pertinent to the EM noise emissions anticipated from a variety of speaker models and sound configurations during the playback of the one or more audio files. The EM noise controller 121 is further configured to identify the EM noise from speakers (EMspeaker) based on the one or more determined characteristics, which may relate to equation-1, as discussed hereinafter in conjunction with description of FIG. 2.
To identify the EMradio, the EM noise controller 121 may perform multiple operations. For instance, the EM noise controller 121 is configured to determine a presence of one or more electronic devices within the surrounding environment that emit EM interference. The EM noise controller 121 is further configured to determine, by utilizing the one or more ML models, one or more radio parameters comprise an operation characteristic of a radio circuit associated with the one or more determined electronic devices, data ingress, and egress within the radio circuit, one or more radio transmission principles. The EM noise controller 121 is further configured to identify the EM noise from internal device components (EMradio) based on the one or more determined radio parameters, which may relate to equation-2, as discussed hereinafter in conjunction with description of FIG. 2.
To identify the EMother, the EM noise controller 121 may perform multiple operations. For instance, the EM noise controller 121 is configured to determine at least one category of short-duration, high-amplitude bursts produced by one or more sources. The one or more sources may include, but is not limited to, one or more atmospheric events (e.g., lightning strikes) and one or more domestic appliances (e.g., microwave ovens, refrigerators, etc.).
In one or more embodiments, the EM noise controller 121 is configured to scale the one or more collected bio-signals, and the one or more identified EM noises, which may relate to FIG. 2.

In one or more embodiments, the EM noise controller 121 is configured to analyze the one or more scaled bio-signals, and the one or more scaled EM noises by utilizing at least one Machine Learning (ML) model, which may relate to FIG. 4. To train the ML model, the EM noise controller 121 may perform multiple operations. For instance, the EM noise controller 121 is configured to perform one or more data processes on the one or more identified EM noises. The one or more data processes may include, but is not limited to, a data integration to consolidate various sources of data, a data cleaning to remove inaccuracies and inconsistencies, a data reduction to minimize the volume of data while retaining essential information, a data transformation to convert data into a suitable format for analysis, and a data discretization to convert continuous data into discrete values for improved processing. The EM noise controller 121 is further configured to split, after processing, the one or more identified EM noises into at least one of a training set, a validation set, and a testing set. The training set is utilized for training the at least one ML model, the validation set is used to optimize one or more hyperparameters of the at least one ML model, and the testing set is reserved for unbiased performance evaluation of at least one trained ML model.
The EM noise controller 121 is further configured to feed the split data into at least one ML model for the training process comprising at least one of a model initialization to establish one or more starting parameters, a loss function selection to define an error measurement, an optimizer selection to determine an algorithm for updating model parameters, and a hyperparameter tuning to refine model performance. The EM noise controller 121 is further configured to perform one or more evaluation processes to determine an optimal model comprising utilizing performance metrics to assess model effectiveness, conducting testing set evaluation to validate model accuracy, and executing error analysis to identify and correct prediction errors.
In one or more embodiments, the EM noise controller 121 is configured to generate a compound noise signal (EMtotal) based on a result of the analysis, which may relate to equation-3, as discussed hereinafter in conjunction with description of FIG. 2.
In one or more embodiments, the EM noise controller 121 is configured to cancel the generated compound noise signal (EMtotal) from the one or more collected bio-signals, to enhance quality of the one or more collected bio-signals, which may relate to equation-4, as discussed hereinafter in conjunction with description of FIG. 2. To cancel the generated compound noise signal (EMtotal) from the one or more collected bio-signals, the EM noise controller 121 may perform multiple operations. For instance, the EM noise controller 121 is configured to programmatically cancel the one or more EM noises from the one or more collected bio-signals in real-time. For another instance, the EM noise controller 121 is configured to cancel the one or more EM noises from the one or more collected bio-signals by utilizing one or more adaptive filtering mechanisms.
In one or more embodiments, the communicator 130 is configured for communicating internally between internal hardware components and with external devices (e.g., server) via one or more networks (e.g., radio technology). The communicator 130 includes an electronic circuit specific to a standard that enables wired or wireless communication.
In one or more embodiments, the wearable electronic device 100 may include a display module (not shown in FIG.1). The display module may accept user inputs and is made of a Liquid Crystal Display (LCD), a Light Emitting Diode (LED), an Organic Light Emitting Diode (OLED), or another type of display. The user inputs may include but are not limited to, touch, swipe, drag, gesture, and so on.
In one or more embodiments, the wearable electronic device 100 may include or be associated with, for example, but is not limited to, one or more hardware entities (as shown in FIG. 5).
In one or more embodiments, a function associated with the various components of the wearable electronic device 100 may be performed through the non-volatile memory, the volatile memory, and the processor 120. One or a plurality of processors controls the processing of the input data in accordance with a predefined operating rule or At least one ML model stored in the non-volatile memory and the volatile memory. The predefined operating rule or At least one ML model is provided through training or learning. Here, being provided through learning means that, by applying a learning algorithm to a plurality of learning data, a predefined operating rule or At least one ML model of the desired characteristic is made. The learning algorithm is a method for training a predetermined target device (for example, a robot) using a plurality of learning data to cause, allow, or control the target device to decide or predict. Examples of learning algorithms include, but are not limited to, supervised learning, unsupervised learning, semi-supervised learning, or reinforcement learning.
The At least one ML model may consist of a plurality of neural network layers. Each layer has a plurality of weight values and performs a layer operation through a calculation of a previous layer and an operation of a plurality of weights. Examples of neural networks include, but are not limited to, Large Language Models (LLMs), Transformer Based ML Models, Convolutional Neural Network (CNN), Deep Neural Network (DNN), Recurrent Neural Network (RNN).
Although FIG. 1 shows various hardware components of the wearable electronic device 100, but it is to be understood that other embodiments are not limited thereon. In other embodiments, the wearable electronic device 100 may include less or more number of components. Further, the labels or names of the components are used only for illustrative purposes and do not limit the scope of the invention. One or more components can be combined to perform the same or substantially similar functions for the EM noise detection and cancellation.
FIG. 2 is a flow diagram illustrating a method 200 for the EM noise detection and cancellation, according to an embodiment as disclosed herein. The EM noise detection and cancellation disclosed method 200 is a sophisticated framework designed to address the challenges posed by the EM noise in various environments, particularly those involving bio-signal monitoring. This process begins with the critical step of measuring and quantifying EM noise, which serves as the foundation for subsequent analysis and mitigation efforts. The disclosed method 200 may execute multiple operations for the EM noise detection and cancellation, which are given below.
At operations 201-202, in an initial phase, specialized EM sensors are deployed to capture EM disturbances across a wide frequency spectrum. These EM sensors are capable of detecting various types of noise, which are then classified by the subsequent operations of disclosed method 200. The classification process is vital for understanding the nature and sources of EM noise present in the environment. The disclosed method 200 may utilize the ML model that has been meticulously trained to categorize the detected noise into four primary classifications, as mentioned below.
In one embodiment, a first classification is “EMambient”. This classification includes EM noise that is relatively stable over time, primarily generated by fixed sources such as power lines, electrical appliances, and wireless communication devices. Examples of these fixed sources may include Wi-Fi modules, 5G antennas, mobile phones, smartwatches, Wi-Fi routers, wireless chargers, microwaves, speakers, compressors, and transformers. The data collected regarding EMambient noise serves multiple purposes, it helps assess whether the environment is conducive for bio-signal testing and alerts users to potentially hazardous levels of EM noise. This information is crucial for ensuring the safety and accuracy of bio-signal measurements.
In one embodiment, a second classification is “EMspeaker”. Conventional electromechanical speakers are known to emit EM noise while reproducing audio signals. The disclosed method 200 may predict the resultant EM noise by analyzing the input signature comprising the audio input and noise cancellation features in real-time. By leveraging EM noise metadata associated with the audio source, the disclosed method 200 may enhance the accuracy of its predictions. A machine learning-based predictive model plays a crucial role in this process, allowing for the anticipation of EM noise generated by speakers that are in close proximity to bio-signal sensors. This predictive capability is essential for optimizing the environment for bio-signal acquisition. The mathematical representation of the EMspeaker is defined by the below-mentioned equation-1.
EMspeaker = f (Input audio file, EM metadata) (1)
In addition, “EMspeaker” refers to electromagnetic noise, which can arise from the speaker magnets or equivalent diaphragm actuators within the device. This noise is generated due to the diaphragm’s movement and the associated motion of the magnets. Predictive analysis of the input signals fed to the device can be employed to anticipate such noise.
In one embodiment, a third classification is “EMradio”. This classification pertains to unpredictable EM noise that varies over time, typically generated, for example, by Bluetooth and Wi-Fi modules in devices such as mobile phones and wearables. The presence of one or more radio modules, including Bluetooth Low Energy (BLE), Wi-Fi, and Ultra-Wideband (UWB), contributes to this noise during their operation. The disclosed method 200 may capitalize on one or more operational characteristics of these radio modules to predict the EM noise they generate. By analyzing standardized input data and understanding the behavior/ principles of the radio circuits, the disclosed method 200 may derive more accurate predictions of the EM noise produced, thereby enhancing the overall detection and cancellation capabilities. The mathematical representation of the EMradio is defined by the below-mentioned equation-2.
EMradio = f (operation characteristics, inputs, radio transmission principles, ML model) (2)
Herein, the operation characteristics define the functional behavior of radio circuits. The inputs and outputs signify the signals and data that ingress and egress these circuits. The radio transmission principles define the fundamental concepts governing the propagation of radio signals. Furthermore, the ML model refers to a machine-learning framework that analyzes and predicts electromagnetic noise based on the aforementioned factors.
In one embodiment, a fourth classification is “EMother”. This classification encompasses all other sources of electromagnetic noise, including impulsive noise characterized by short, high-amplitude bursts. Such noise can be generated by phenomena like lightning strikes or by household appliances such as mixers and switching devices. The unpredictable nature of this type of noise presents significant challenges for mitigation, necessitating advanced detection and cancellation strategies.
At operation 203, once the EM noise has been classified, the disclosed method 200 may employ an adaptive system to conduct adaptive signal scaling and real-time analysis of the classified EM noise. This adaptive system is designed to ensure that both bio-signal and EM noise signals are scaled appropriately, which is critical for accurate identification and analysis of the various noise components, even when they span diverse ranges. The ability to perform real-time analysis allows the system to respond dynamically to changes in the environment, thereby enhancing its effectiveness.
At operation 204, in the next phase, the adaptive system implements machine learning-based analysis on the scaled signals. The ML model is trained on extensive datasets that include both EM noise and bio-signal data, enabling it to learn the unique characteristics of different noise sources. This integration of machine learning not only improves the precision of noise analysis but also facilitates structured data processing, which refines the recognition and classification of EM noise types. Additionally, the disclosed method 200 may feature advanced visualization tools that display real-time and historical data, providing users with critical insights into the noise environment and system performance. This feedback loop is essential for continuous adaptation and calibration of the system based on the quantified EM noise data.
At operation 205, following the analysis, the adaptive system generates the compound noise (EMtotal) signal that represents the aggregate of all detected EM noise components. This compound noise signal is pivotal for the active noise cancellation process, as the compound noise signal provides a comprehensive representation of the noise environment that needs to be mitigated. The mathematical representation of the EMtotal is defined by the below-mentioned equation-3.
EMtotal = (EMambient + EMradio + EMspeaker +EMother) (3)
At operation 206, the disclosed method 200 may implement a software-based noise cancellation strategy to remove the compound noise signal from the recorded bio-signal, also known as “EMcancel”. The mathematical representation of the EMcancel is defined by the below-mentioned equation-4.
EMcancel = - (EMambient + EMradio + EMspeaker+ EMother) (4)
This software-based noise cancellation strategy may utilize an adaptive filtering algorithm that continuously monitors both the bio-signal and EM noise signals in real-time. By programmatically canceling out the noise, the system minimizes interference, thereby enhancing the accuracy and reliability of bio-signal readings. This sophisticated approach ensures that the wearable bio-signal monitoring devices (e.g., 100) can operate effectively in diverse environments, ultimately improving their performance and user experience. Overall, the disclosed method 200 not only addresses the immediate challenges posed by EM noise but also lays the groundwork for future advancements in bio-signal monitoring technology, contributing to safer and more accurate health assessments.
In one or more embodiments, the disclosed method 200 may incorporate a descriptive metadata file in a standardized format, which is intended to accompany the audio files designated for playback, ideally provided by the music streaming or source provider alongside the audio content. This metadata file may contain essential information and machine learning model weights pertinent to the EM noise emissions produced by various speaker models and sound configurations during audio playback. By integrating this information, for example, each earphone system can predict and subsequently attenuate EM noise emissions, eliminating the need for real-time computation of EM noise generated by a specific audio file upon receipt. This approach significantly reduces processing load and enhances the effectiveness of EM noise cancellation.
In one or more embodiments, to achieve precise noise cancellation in wearable devices (e.g., 100), it is crucial to assign appropriate weights to each noise variable (e.g., W1, W2, W3, W4), including ambient electromagnetic noise, EM noise from speakers, and EM noise from the device’s internal components. Machine learning techniques are instrumental in this process, enabling the dynamic adjustment of these weights based on real-time data and contextual information. Through continuous environmental monitoring and learning, the machine learning models can assess the relevance of various noise sources, adapting the cancellation strategy to prioritize the most significant factors while minimizing the impact of less critical ones. This dynamic weight assignment not only ensures effective noise cancellation but also enhances the overall performance of wearable devices under varying conditions, ultimately leading to more accurate and reliable bio-signal measurements. A simplified representation of the real-time processing for each sampled data point within the machine learning model can be illustrated by the equation-5 mentioned below.
EMtotal = W1*EMambient + W2*EMradio + W3*EMspeaker + W4*EMother (5)
In one or more embodiments, the software-based noise cancellation methodologies may include different mechanisms to address the EM noise, for example, which are mentioned below.
EM noise cancellation by an actual playback (audio playback) mechanism;
EM noise cancellation by a simulation-based mechanism; and
EM noise cancellation by a metadata-based mechanism.
In the context of the actual playback mechanism, real-time data from speaker and radio modules is utilized to dynamically detect and mitigate EM noise. During the audio playback, an ML algorithm continuously analyzes the EM noise produced by a magnetic diaphragm of the speaker and the high-frequency emissions from radio modules such as Bluetooth or Wi-Fi. By monitoring the interaction between playback signals and detected noise, the algorithm learns how specific combinations of audio signals and operational radio frequencies contribute to overall EM interference. The ML model assesses the required compensation by identifying patterns in noise generation over repeated cycles, allowing it to adjust noise cancellation mechanisms in real time, thereby ensuring that the bio-signal acquisition system remains unaffected by playback-induced noise. This mechanism offers precise, real-time mitigation of EM interference and optimizes device performance across varying playback environments, as one of the advantages of the disclosed method 200.
In the context of the simulation-based mechanism, the simulation-based defined EM noise detection and cancellation relies on pre-recorded or simulated data to analyze and mitigate EM noise generated by speaker and radio modules. Controlled simulations are conducted to ascertain the EM noise profile of the speaker during audio playback under diverse conditions and the radio module during active data transmission. These simulations provide comprehensive insights into expected noise outputs for various operational scenarios. The simulated noise data is analyzed to identify patterns and quantify noise levels, enabling the system to understand the relationship between input audio signals and resultant EM noise, as well as variations caused by changes in signal frequency, amplitude, or hardware operations. The pre-defined noise profiles derived from these simulations are then used to programmatically cancel anticipated noise during real-world device operation, enhancing bio-signal clarity by compensating for noise before it affects signal acquisition. This mechanism facilitates offline testing and optimization of noise cancellation algorithms while establishing a robust framework for predicting and mitigating noise without the need for constant real-time learning, as one of the advantages of the disclosed method 200.
In the context of the metadata-based mechanism, the metadata-based defined EM noise detection and cancellation, leverages pre-defined metadata that accompanies audio files to anticipate and cancel EM noise. Each audio file includes metadata detailing the expected noise characteristics from the speaker based on audio signal properties (e.g., frequency, amplitude, etc.) and the anticipated interference patterns from the radio module during playback. When the audio file is played, the disclosed method 200 references this metadata to predict the EM noise likely to be generated, eliminating the need for real-time computation of noise profiles and significantly reducing processing power requirements. Using the metadata, the disclosed method 200 applies pre-determined filters and adaptive algorithms to effectively cancel the defined EM noise, ensuring high accuracy tailored to the specific audio file and device configuration. This mechanism reduces computational load by leveraging pre-defined noise profiles and ensures highly accurate cancellation tailored to the specific audio file and device configuration, as one of the advantages of the disclosed method 200.
The above-mentioned mechanisms collectively address the challenge of EM noise cancellation through distinct approaches such as the actual playback mechanism providing real-time, adaptive noise mitigation; the simulation-based mechanism offers pre-defined profiles derived from controlled testing; and the metadata-based mechanism optimizes efficiency by embedding noise definitions within audio files. Together, these mechanisms ensure comprehensive noise cancellation, enhancing the performance of ear-worn bio-signal acquisition devices in dynamic environments.
In one embodiment, the disclosed method 200 may refine the noise cancellation process by incorporating methodologies that adapt to real-time operational conditions, leverage simulation data for offline optimization, and utilize pre-defined metadata for efficient and accurate interference mitigation. This enhanced mechanism ensures superior biosignal acquisition in diverse and challenging electromagnetic environments.
In one embodiment, the disclosed method 200 may isolate speaker-generated sound within the ear canal while audio is being played through earphone speakers. This isolation is accomplished through the utilization of metadata associated with the earphones, the speakers, and the audio playback signal. The disclosed method 200 may leverage this metadata to predict the acoustic signature of the earphone speakers, subsequently removing it from the composite signal captured within the ear canal. This process enables the isolation of acoustic biosignals, such as the sound of blood flow in the veins or arteries like carotid artery, which can be utilized for medical diagnostics, including the early detection of stroke symptoms.
The disclosed method 200 may begin with metadata-driven audio prediction, where, upon audio playback through the earphones, the processing unit accesses metadata files related to the specific earphone model. Utilizing this metadata alongside the playback signal, the wearable electronic device 100 may compute the expected acoustic output of the speakers, encompassing both frequency and amplitude characteristics.
Next, signal acquisition occurs through embedded microphones within the earphones that capture a composite signal from the ear canal. This composite signal comprises the audio output generated by the speakers, background environmental noise, and the acoustic biosignal.
Following this, real-time signal subtraction is performed by the disclosed method 200, where the predicted speaker output, derived from the metadata, is subtracted from the composite signal captured by the microphones. This subtraction process isolates the residual signal, which primarily consists of the acoustic biosignal with minimal interference from other noise sources.
The isolated residual signal undergoes acoustic bio signal identification, wherein the disclosed method 200 may analyze the waveform to extract the acoustic biosignal. Machine learning techniques are applied to identify distinct characteristics of the biosignal pulse, including its frequency, amplitude, and temporal patterns.
Finally, the extracted bio signal pulse is subjected to a health assessment, evaluating it for abnormalities such as irregular timing, amplitude variations, or waveform distortions that may indicate medical conditions like stroke or arterial blockages .
The above-mentioned embodiment signifies a substantial advancement in wearable health technology by integrating bio-signal monitoring and real-time diagnostics into everyday audio devices. The disclosed method 200 ensures precise and reliable acquisition of acoustic biosignals in acoustic dynamic environments.
In one embodiment, the disclosed method 200 may enhance the accuracy of audio cancellation processes within the ear canal through the implementation of a unit test signal. The disclosed method 200 may validate an effective identification and software-based removal of speaker-generated sounds utilizing metadata. The disclosed method 200 may determine whether any deviations are detected during the unit testing, the metadata is dynamically modified and updated to accurately reflect the specific acoustic characteristics of the earphones under current usage conditions.
The disclosed method 200 may begin with a post-cancellation validation. After isolating the residual biosignal, such as the carotid artery pulse sound, by eliminating speaker-generated sounds from the composite signal, the system initiates a validation sequence.
Following this, a unit test signal playback is conducted, wherein a predefined unit test signal is played through the earphone speakers. This signal is specifically designed to encompass key frequencies and amplitudes representative of the speaker’s operational range. For instance, an example test signal may consist of a sequence of sine waves at frequencies of 100 Hz, 1 kHz, 5 kHz, and 15 kHz, each played for a fixed duration at calibrated amplitudes.
During the signal capture and analysis phase, the microphones or sensors embedded within the earphones capture the audio within the ear canal while the unit test signal is being played. The disclosed method 200 may then compare the captured signal against the expected output derived from the metadata.
Error detection is critical, as any discrepancies between the captured signal and the predicted signal indicate inaccuracies in the audio cancellation process. Such deviations may stem from various factors, including variations in speaker performance due to wear and tear, environmental influences such as temperature and humidity, or user-specific anatomical differences that affect sound propagation.
If deviations are identified, a metadata editing process is initiated by the disclosed method 200. The metadata file is dynamically updated using a feedback loop to correct the frequency response, amplitude, or distortion characteristics, thereby enhancing the accuracy of subsequent audio cancellation processes
Finally, the disclosed method 200 may undergo a revalidation process, where the unit test signal is replayed following the metadata adjustments to confirm the accuracy of the updated cancellation parameters. This iterative process ensures precise isolation of biosignals, thereby enhancing the overall effectiveness of the audio cancellation methodology.
In one embodiment, the disclosed method 200 may utilize a dynamically edited metadata, the dynamically edited metadata is stored in the user’s profile on the device or on a secure cloud server, facilitating the preservation and synchronization of personalized audio cancellation settings across multiple devices. After local modifications are made to the metadata file, the updated file is encrypted and uploaded to the user’s cloud profile. When the user connects a new device, such as an additional pair of earphones or another wearable, the disclosed method 200 may download the relevant metadata and apply it to ensure consistent audio cancellation across all devices linked to the user’s account. To maintain accuracy, a version control system tracks changes made to the metadata, ensuring that the most recent file is utilized while retaining previous versions for potential rollback if necessary. Additionally, the cloud-stored metadata may include annotations regarding environmental conditions, such as high humidity or temperature, to enhance performance in similar future scenarios.
Furthermore, the disclosed method 200 may improve the accuracy of audio isolation in wearable devices through the auto-calibration of the audio cancellation algorithm, even when the earphones are not actively worn by the user. This is achieved by simulating an operational environment using predefined test signals and analyzing the acoustic and electromagnetic feedback within the device housing. The auto-calibration process refines metadata and cancellation parameters, enhancing isolation performance before active use. An integrated module within the signal processing unit generates predefined audio test signals across a broad frequency range (e.g., 20 Hz to 20 kHz), specifically designed to probe the acoustic response of the earphones and simulate typical usage conditions. Furthermore, microphones positioned within the earphone housing capture audio signals emitted by the speakers during calibration, allowing the detection of speaker-generated audio while minimizing interference from external noise and user-specific anatomical variations.
In an embodiment, the present invention provides the method and/or system for dynamically predicting and accounting for the EM noise generated by speaker movements during Active Noise Cancellation (ANC) operations in wearable electronic devices (e.g., 100). The and/or system analyzes an anti-noise signal transmitted to the speakers in real-time, where the anti-noise signal is specifically designed to negate external acoustic disturbances. The speaker movements induced by this anti-noise signal generate transient EM noise due to the oscillation of the diaphragm and associated magnetic field interactions within the speaker assembly. Using predefined metadata that includes speaker-specific parameters such as frequency response curves, coil dynamics, harmonic distortion, and transient behavior, the system predicts the resulting EM emissions. This predicted ANC-induced EM noise is continuously monitored, quantified, and incorporated into the overall noise cancellation process, enabling precise subtraction of EM noise artifacts from biosignals. By accounting for these complex EM interactions, the and/or system ensures enhanced biosignal isolation and robustness in environments with significant active audio processing.
FIG. 3 is a flow diagram illustrating a method 300 for environmental diagnosis for the EM noise detection and cancellation, according to an embodiment as disclosed herein. The disclosed method 300 may execute multiple operations for environmental diagnosis for the EM noise detection and cancellation, which are given below.
At operation 301, the method 300 initiates by receiving at least one input from the user, which serves as a trigger to commence the recording or collection of the one or more bio-signals. At operation 302, subsequent to the receipt of the at least one input, the method 300 involves measuring the ambient electromagnetic noise within the surrounding environment. At operation 303, following this measurement, the method 300 assesses whether the measured value of the ambient electromagnetic noise exceeds a predefined electromagnetic noise threshold (designated as EMthreshold). Should the evaluation indicate that the ambient electromagnetic noise (EMambient) is below the established electromagnetic noise threshold (EMthreshold), the method proceeds to collect the specified bio-signals from the user of the wearable electronic device 100. At operation 304, the method 300 further includes collecting the one or more bio-signals from the user of the wearable electronic device 100 in response to determining that the value of the ambient electromagnetic noise (EMambient) is lower than the value of the predefined electromagnetic noise threshold (EMthreshold).
At operation 305, the method 300 further includes transmitting an alert message to the user in response to determining that the value of the ambient electromagnetic noise (EMambient) is greater than the value of the predefined electromagnetic noise threshold, where the alert message indicates that an EM noise level is too high for collecting the one or more bio-signals. At operation 306, the method includes a subsequent operation to ascertain whether the user still wishes to proceed with the collection of the bio-signals, despite the prior alert regarding elevated electromagnetic noise levels. At operation 306-304, if the user expresses a desire to continue, the method 300 may then facilitate the collection of the one or more bio-signals from the user of the wearable electronic device (e.g., 100).
For instance, consider an example scenario where the user wearing a health monitoring device (e.g., 100) activates it by pressing a button, indicating their desire to start monitoring their health. Upon activation, the health monitoring device (e.g., 100) measures the ambient electromagnetic noise (EMambient) in the surrounding area, which may be influenced by electronic devices such as smartphones, microwaves, or Wi-Fi routers. The health monitoring device (e.g., 100) then compares the measured EMambient value to the predefined electromagnetic noise threshold (EMthreshold). If the EMambient is lower than the EMthreshold, the device proceeds to collect bio-signals from the user, ensuring accurate health monitoring. Conversely, if the EMambient exceeds the EMthreshold, the device sends the alert to the user (e.g., voice message, display message, etc.), indicating that the electromagnetic noise is too high for reliable data collection. After receiving this alert, the user is prompted to decide whether to continue with the bio-signal collection despite the noise. If the user chooses to proceed, the health monitoring device (e.g., 100) attempts to collect the bio-signals, even in the noisy environment; if the user opts not to proceed, the health monitoring device (e.g., 100) remains idle until the user decides to try again later. This scenario effectively illustrates how the health monitoring device (e.g., 100) intelligently manages bio-signal collection while keeping the user informed about environmental conditions.
In one or more embodiments, the method 300 includes determining the ambient electromagnetic noise (EMambient= N) associated with the one or more EM noises based on one or more parameters, as illustrated in the equation-6. The one or more parameters may include a wavelength (?), a permeability (µ), a permittivity (?), a current (I), a voltage (V), one or more environmental factors (e.g., pressure, humidity, temperature), a distance between a noise source and an EM sensor, and a constant (k).
(6)
FIG. 4 illustrates operations for training a Machine Learning (ML) model for electromagnetic (EM) noise detection and cancellation, according to an embodiment as disclosed herein.
Initially 401, a comprehensive EM noise dataset is collected, encompassing various noise types (e.g., EMambient, EMspeaker, EMradio, EMother) with diverse signal amplitudes, frequencies, and characteristics. Next 402, the acquired data undergoes processing, including data integration 403, data cleaning 404, data reduction 405, data transformation 406, and data discretization 407, followed by labeling to annotate combined noise waveforms. At operation 408, the processed data is split into training, validation, and testing sets for effective model development. During model training 409, key steps include neural network design (e.g., CNNs for time-frequency features, RNNs for temporal dependencies), model initialization (e.g., Xavier/He methods), loss function selection (e.g., MSE/MAE), optimizer selection (e.g., SGD, Adam), and hyperparameter tuning (e.g., learning rate, batch size). Post-training, iterative evaluations 410 refine the model via architecture adjustments, error analysis, and performance validation against testing sets using metrics like MSE, MAE, and correlation coefficients.
The final trained model 411 ensures robust noise prediction and generalizability, validated on unseen data with error analysis guiding further refinements 412-414. This process supports applications such as wearable devices (e.g., 100) that monitor EEG signals in real-time, improving cognitive state analysis through enhanced signal quality and model performance.
FIG. 5 illustrates an EM noise sensing system 500 for EM signal detection (Antenna), according to an embodiment as disclosed herein.
In one or more embodiments, the EM noise sensing system 500 may utilize a Micro-Electro-Mechanical Systems (MEMS) antenna 501 as its primary receiver for EM signals. The MEMS antenna 501 operates by detecting incoming EM signals from the surrounding environment and converting them into electrical signals. Leveraging MEMS technology allows for significant miniaturization and enhanced sensitivity, enabling the detection of weak EM signals even in noise-prone environments, which is crucial for applications requiring high precision.
In one or more embodiments, the EM noise sensing system 500 may utilize a combination of loop and dipole antennas or an antenna array for magnetic and electric field detection. For instance, but not limited to, a plurality of loop antennas may be used for detecting magnetic fields. The Loop Antennas may comprise coils of wire that may produce a voltage when they encounter a changing magnetic field. Even dipole antennas may be useful for detecting electric fields, and further being straight, these antennas may be made of a conductive metal like copper.
In one or more embodiments, to filter and amplify the received signals, the EM noise sensing system 500 may incorporate an LC tank circuit 502, consisting of an inductor (L) and a capacitor (C) connected in parallel. This LC tank circuit 502 is meticulously designed to resonate at a specific frequency, as shown by the below-mentioned equation-7, allowing only those signals that align with or are near this resonant frequency to pass through efficiently, while other frequencies are attenuated.
f = 1 / (2 * p * v(L * C)) (7)
The quality factor (Q) of the circuit plays a pivotal role in determining the sharpness of the signal selection at resonance, ensuring that the system can effectively isolate target frequencies, as shown by below-mentioned equation-8.
Q = (1 / R) * v(L / C) (8)

Additionally, the tunability of the LC tank circuit 502 enables adaptability to various EM noise frequencies, making it suitable for diverse operational environments.
In one or more embodiments, a dynamic resonance frequency control 503 is implemented to adjust the resonance frequency of the LC circuit for targeted detection. By fine-tuning the values of the inductor or capacitor, the system can match the expected EM noise frequency, ensuring precise detection even amidst varying external conditions or multiple sources of EM interference. This adaptability is essential for maintaining signal integrity and relevance. The signal processing unit 504 is tasked with extracting meaningful information from the filtered signals. The signal processing unit 504 processes these signals to identify patterns, characteristics, and noise components, employing sophisticated algorithms for signal enhancement. In addition, the signal processing unit 504 prepares the data for subsequent steps, including noise cancellation and scaling, ensuring that only the most relevant information is retained.
In one or more embodiments, an adaptive scaling mechanism 505 further enhances the system's performance by dynamically adjusting the signal amplitude to optimize it for noise cancellation. This adaptive scaling mechanism 505 analyzes the intensity of signals from the signal processing unit 504 and either amplifies or attenuates them to match the ideal levels required for effective noise cancellation. This step is critical in ensuring that the signals remain within detectable and processable ranges for the subsequent noise cancellation unit.
In one or more embodiments, a noise cancellation processing 506 employs advanced software algorithms or hardware filters to eliminate unwanted electromagnetic noise from the processed signals while preserving the integrity of the original signal. This step is vital for enhancing the accuracy of the final output, particularly in sensitive applications such as biosignal acquisition or diagnostics. Finally, the processed data is made available for interpretation or further use through display, storage, or transmission mechanisms 507. The clean, noise-free signal can be displayed in real-time, for example, on a monitor, stored for later analysis, or transmitted wirelessly for remote monitoring. This functionality ensures that the data remains usable across various applications, including health diagnostics and real-time monitoring systems, thereby maximizing the utility of the EM noise sensing system.
In one or more embodiments, the EM noise sensing system 500 may include, for example, but is not limited to, one or more hardware entities (not shown in FIG. 5), such as an EM sensing antenna, a broad-frequency front-end amplifier, a plurality of analog filters, an Analog-to-Digital Converter (ADC), a microcontroller or Digital Signal Processor (DSP), a power management circuitry, a communication interface and a user interface.
In one or more embodiments, the EM sensing antenna may be employed which may involve the innovative use of the interconnecting wires between earphone earpieces as the EM antennas to detect the EM noise. These EM antenna wires may be technically engineered to serve a dual purpose, not only transmitting audio signals but also actively capturing EM disturbances. This functionality may extend across a comprehensive frequency range, effectively detecting both low-frequency emissions typical of the power line interference and the high-frequency signals that might emanate from advanced wireless communication devices. The EM-sensing antenna wire may be integrated into various structural components of the earphones. In a wired earphone configuration, the connecting wires between the earpieces may act as the EM antenna. This integration may be achieved using materials and design techniques that may enhance the EM noise detection capabilities of the wires while maintaining their primary function of audio signal transmission.
Alternatively, in a wireless earphone system, the EM antenna may be integrated into the band or frame connecting the two earpieces in a wireless band model of Bluetooth earphones, ensuring that the EM noise detection capability may be retained. The EM antenna in this setup may be strategically placed and constructed to optimize EM noise capture without compromising the aesthetic or functional aspects of the earphones. Furthermore, the EM antenna wire may be configured to work in conjunction with the additional EM sensors within the earphone system, if present. This multi-sensor approach may allow for a more comprehensive and accurate detection of the EM noise, covering a wider range of frequencies and intensities.
In one or more embodiments, the broad-frequency front-end amplifier may be used to receive a signal from an antenna array of an EM noise sensor to amplify a tiny voltage induced in the antenna due to the EMF. The front-end amplifier must have a broad frequency response to handle various types of EMF sources.
In one or more embodiments, the plurality of analog filters may be used in the EM noise sensor for frequency filtering of the amplified signal. The analog filter may include a band-pass and low-pass filter for signal conditioning. The band-pass filter may filter out frequencies that may be outside the range of interest, enhancing the sensor's specificity. Whereas, the low-pass filter may help in eliminating high-frequency noise, which may interfere with measurements.
In one or more embodiments, the analog signal from the antenna obtained after filtering may be further converted to digital format by the ADC of the EM noise sensor wherein the analog signal may be processed by the microcontroller or DSP. The microcontroller or DSP may be adapted to process digital data from the ADC. The processing may involve tasks such as, but not limited to, Fourier transform to determine the frequencies present and to measure their intensities. The microcontroller or DSP may also run algorithms to differentiate between an ambient EMI and an internal device EMI.
In one or more embodiments, a reference signal source may aid in calibration, while an internal memory (e.g., 110) may store critical data for display and transmission. The internal memory may store calibration data, historical readings, and a potential firmware for the microcontroller or the DSP. The reference signal source may produce a known EMF signal that may be used for calibration purposes. The reference signal source may also ensure the accuracy of sensor readings.
In one or more embodiments, the power management circuitry that may ensure efficiency and a communication interface may allow the exchange of data. The power management circuitry may also incorporate features like sleep mode or low-power mode. Optionally, the user interface may be provided to the EM noise sensor to get immediate feedback, all within a shielded housing for minimal interference. Periodic calibration and non-conductive casing may also be provided as key features for accuracy and sensor protection. Additionally, the communication interface may be provided that may allow the EM sensor to communicate with other parts of the device or external devices. The communication interface may be wired (e.g., Inter-Integrated Circuit (I2C), Serial Peripheral Interface (SPI), or Universal Asynchronous Receiver-Transmitter (UART)) or wireless (e.g., Bluetooth or Wi-Fi).
In one or more embodiments, the user interface may also be optionally provided in the EM noise sensor. If the sensor may be designed to provide immediate feedback to a user, it may incorporate LEDs, a small display, a notification in their smartphones or computing device or even a buzzer. Further to shield the sensor's electronics to prevent interference from its internal electronic operations, shielding may be provided to the EM noise sensor. Shielding may be made from materials like Mu-metal, which is highly permeable and absorbs magnetic fields. The EM noise sensor may comprise a calibration mechanism that may allow the EM sensor to be periodically calibrated against known EMF sources to ensure accuracy. The entire components of the EM sensor may be housed in a housing that may be a non-conductive casing to prevent interference with measurements. The housing may also provide openings or windows to allow ambient EMF to reach the antennas.
FIG. 6 is a flow diagram illustrating a method 600 for an adaptive sensor resolution adjusting for the EM noise detection and cancellation, according to an embodiment as disclosed herein. The disclosed method 600 may execute multiple operations for adaptive sensor resolution adjusting for the EM noise detection and cancellation, which are given below.
At operation 601, initially, a systematic protocol for monitoring ambient EM noise is to be developed. This systematic protocol may define the methodologies and technologies employed to capture EM noise across various frequencies and environments. At operation 602, subsequently, the intensity and characteristics of the EM noise may be quantitatively measured and systematically recorded. This operation aims to generate a robust and comprehensive dataset that encompasses a wide range of electromagnetic noise conditions, facilitating advanced analysis. At operation 603, following the data acquisition phase, the ML algorithms may be deployed to analyze the recorded EM noise data. These ML algorithms are designed to identify patterns, correlations, and trends within the dataset, enabling the extraction of meaningful insights regarding the characteristics of the ambient electromagnetic environment.
At operation 604, based on the insights gleaned from the ML algorithms, the adaptive method 600 may evaluate and determine optimal adjustments to the sensor resolution. This operation is critical for ensuring that the sensors operate at an appropriate sensitivity level tailored to the identified EM noise patterns. At operation 605, finally, based on the analysis, the method 600 may implement dynamic modifications to the sensor settings. These adjustments may be aimed at enhancing both the resolution and accuracy of the measurements, thereby improving the overall reliability of the monitoring system in capturing ambient electromagnetic noise. For instance, during high-stress situations, the wearable electronic device (e.g., 100) may dynamically increase its sampling rate to provide more detailed data. This approach enables a deeper understanding of brain activity in everyday environments, ultimately contributing to advancements in mental health monitoring and cognitive research.
In one embodiment, the disclosed method comprises an integrated EM noise detection mechanism within a Printed Circuit Board (PCB) of the wearable electronic device 100. This wearable electronic device 100 features an EM noise sensor specifically configured to monitor electromagnetic emissions produced by the device’s internal components, including Integrated Circuits (ICs) and various electronic modules.
During operation, the wearable electronic device 100 generates predictable electromagnetic emissions, including those from Bluetooth and Wi-Fi modules, based on established operational parameters and input signals. The EM noise sensor is configured to measure these emissions in real-time and to compare the acquired signals against expected characteristics. Any discrepancies between the anticipated and actual signals are analyzed to ascertain the presence and characteristics of ambient electromagnetic noise in the surrounding environment.
In one embodiment, the disclosed method facilitates real-time identification and characterization of external electromagnetic interference, enabling the differentiation between internal device noise and external EM noise. By leveraging these measurements, the wearable electronic device 100 may dynamically optimize noise cancellation algorithms, thereby enhancing performance, particularly in applications that demand high-fidelity biosignal acquisition or other critical operations.
In one embodiment, method 200 includes an electromagnetic noise cancellation mechanism designed to address both conventional electromagnetic interference and noise induced by ionizing radiation. Wearable electronic device 100 may incorporate radiation-hardened sensors to detect and quantify ionizing radiation, such as gamma rays and X-rays, which induce transient noise in electronic circuits in industrial or medical environments. These sensors assess the impact of high-energy emissions on the device's components. The at least one integrated ML model analyzes and classifies ionizing radiation-induced noise alongside ambient electromagnetic interference. Device 100 may dynamically adjusts noise cancellation algorithms to mitigate radiation-induced noise, enhancing reliability in high-energy environments. This dual-function mechanism ensures comprehensive noise mitigation, crucial for precision in applications like biosignal monitoring systems

In an embodiment, the present invention includes an auxiliary electrode probe, associated with the wearable electronic device 100 (not shown in FIGs.), positioned right next to the main active electrode probes, configured to detect ambient EM noise as an additional verification mechanism for the internal EM noise detection system. The electrode probe captures real-time EM noise signals from the surrounding environment that influence the active electrode probes as closely as possible, including interference sources such as Bluetooth, Wi-Fi, and other electronic devices. These signals are processed by a dedicated signal analysis module and compared against the EM noise detected within the ear canal by the internal EM sensing system. The auxiliary probe serves as a validation check enabling the feedback loop, ensuring consistency and accuracy in the EM noise characterization and cancellation process. Discrepancies between external and internal EM noise profiles are analyzed to identify potential anomalies, such as leakage from uncanceled EM noise or hardware-level inconsistencies. The system dynamically adjusts its noise cancellation parameters based on this cross-referenced data, thereby improving biosignal isolation and maintaining high fidelity in diverse and noisy electromagnetic environments.
In an embodiment, to identify the EM noise associated with the wearable electronic device 100, the disclosed method commences with the generation of the acoustic metadata file pertinent to the device's integrated speaker system (not shown in FIGs.). This acoustic metadata file encompasses several critical components such as, first, it includes a frequency response profile that delineates the output characteristics of the speaker system across a specified frequency range. Additionally, it captures harmonic distortion data at various operating frequencies and amplitudes, providing insight into the fidelity of audio reproduction. The acoustic metadata file also records one or more transient response times, which indicate the stabilization delays of the speaker system following the application of input signals, thereby reflecting the system’s responsiveness.
The acoustic metadata file contains phase response characteristics that correlate input and output signals across key frequencies, providing insights into speaker performance. It also includes speaker-specific profiles for different playback modes, ensuring optimal audio delivery. This file is associated with audio input signals and active Acoustic Noise Cancellation (ANC) inputs to predict the speaker's acoustic output.
The disclosed method may detect a composite sound signal within the ear canal, which includes predicted acoustic output, ambient noise, and biosignals. Using the metadata file, the predicted acoustic output is isolated, improving biosignal extraction accuracy. The file is dynamically updated in real-time to account for deviations, ensuring precise noise isolation and adaptive audio cancellation in varying environments.
FIG. 7 is a flow diagram illustrating a method 700 for the EM noise detection and cancellation, according to another embodiment as disclosed herein. The disclosed method 700 may execute multiple operations for the EM noise detection and cancellation, which are given below.
At operation 701, the method 700 includes collecting the one or more bio-signals from the user of the wearable electronic device 100. At operation 702, the method 700 includes identifying the one or more EM noises present in the surrounding environment of the wearable electronic device during the collection of the one or more bio-signals. At operation 703, the method 700 includes scaling the one or more collected bio-signals and the one or more identified EM noises. At operation 704, the method 700 includes analyzing the one or more scaled bio-signals and the one or more scaled EM noises by utilizing the at least one ML model. At operation 705, the method 700 includes generating the compound noise signal (EMtotal) based on the result of the analysis. At operation 706, the method 700 includes cancelling the generated compound noise signal (EMtotal) from the one or more collected bio-signals, to enhance quality of the one or more collected bio-signals. Further, a detailed description related to the various steps of FIG. 7 is covered in the description related to FIG. 1 to FIG. 6 and is omitted herein for the sake of brevity.
The various actions, acts, blocks, steps, or the like in the flow diagrams may be performed in the order presented, in a different order, or simultaneously. Further, in some embodiments, some of the actions, acts, blocks, steps, or the like may be omitted, added, modified, skipped, or the like without departing from the scope of the invention.
Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one ordinary skilled in the art to which this invention belongs. The system, methods, and examples provided herein are illustrative only and not intended to be limiting.
While specific language has been used to describe the present subject matter, any limitations arising on account thereto, are not intended. As would be apparent to a person in the art, various working modifications may be made to the method to implement the inventive concept as taught herein. The drawings and the forgoing description give examples of embodiments. Those skilled in the art will appreciate that one or more of the described elements may well be combined into a single functional element. Alternatively, certain elements may be split into multiple functional elements. Elements from one embodiment may be added to another embodiment.
The embodiments disclosed herein can be implemented using at least one hardware device and performing network management functions to control the elements.
The foregoing description of the specific embodiments will so fully reveal the general nature of the embodiments herein that others can, by applying current knowledge, readily modify and/or adapt for various applications such specific embodiments without departing from the generic concept, and, therefore, such adaptations and modifications should and are intended to be comprehended within the meaning and range of equivalents of the disclosed embodiments. It is to be understood that the phraseology or terminology employed herein is for the purpose of description and not of limitation. Therefore, while the embodiments herein have been described in terms of preferred embodiments, those skilled in the art will recognize that the embodiments herein can be practiced with modification within the scope of the embodiments as described herein. ,CLAIMS:WE CLAIM:

1. A method (700) comprising:
i. collecting (701) one or more bio-signals from a user of a wearable electronic device (100);
ii. identifying (702) one or more Electro-Magnetic (EM) noises present in a surrounding environment of the wearable electronic device (100) during the collection of the one or more bio-signals;
iii. scaling (703) the one or more collected bio-signals, and the one or more identified EM noises;
iv. analyzing (704) the one or more scaled bio-signals, and the one or more scaled EM noises by utilizing at least one Machine Learning (ML) model;
v. generating (705) a compound noise signal (EMtotal) based on a result of the analysis; and
vi. cancelling (706) the generated compound noise signal (EMtotal) from the one or more collected bio-signals, to enhance quality of the one or more collected bio-signals.

2. The method (700) as claimed in claim 1,
i. wherein the one or more bio-signals comprise an Electroencephalography (EEG) signal, an Electrooculography (EOG) signal, an Electromyography (EMG) signal, and an Electrocardiography (ECG) signal; and
ii. wherein the one or more EM noises comprise at least one of an ambient electromagnetic noise (EMambient), an EM noise from speakers (EMspeaker), an EM noise from internal device components (EMradio), and an impulsive noise (EMother).

3. The method (700) as claimed in claim 1, wherein identifying the one or more EM noises present in the surrounding environment of the wearable electronic device (100) comprises:
i. detecting, via an EM sensor, the one or more EM noises, wherein the EM sensor comprises at least one of a loop antenna, a dipole antenna, an antenna array, and a Micro-Electro-Mechanical Systems (MEMS) based Inductor-Capacitor (LC) tank circuit;
ii. quantifying the one or more EM noises detected by the EM sensor to generate one or more measurable parameters indicative of one or more EM noise characteristics; and
iii. classifying the one or more quantified EM noises, based on a predefined classification mechanism, by utilizing at least one ML model, to identify the one or more EM noises present in the surrounding environment of the wearable electronic device (100).

4. The method (700) as claimed in claim 1, further includes integrating an auxiliary electrode probe placed in close proximity to one or more primary active electrodes to sample the electromagnetic noise impacting the one or more primary active electrodes.
i. capturing, via the auxiliary electrode probe, real-time electromagnetic noise signals associated with interference sources that affect the primary active electrodes, wherein the captured real-time electromagnetic noise signals represent as external electromagnetic noise signals;
ii. comparing the captured real-time electromagnetic noise signals with electromagnetic noise signals identified by an internal electromagnetic noise detection system within the wearable electronic device (100);
iii. analyzing discrepancies between the external and internal electromagnetic noise signals to identify potential anomalies, including leakage of uncanceled noise or inconsistencies in hardware performance; and
iv. dynamically adjusting one or more noise cancellation parameters of the wearable electronic device (100) based on the results of the comparison, to improve accuracy in biosignal isolation and enhance the overall quality of the one or more collected bio-signals.
5. The method (700) as claimed in claim 1, wherein, when the one or more EM noises comprise an ambient electromagnetic noise (EMambient), identifying the one or more EM noises comprises:
i. determining a presence of one or more electronic devices within the surrounding environment that emit EM interference; and
ii. identifying the ambient electromagnetic noise (EMambient) based on the one or more determined electronic devices,
wherein the ambient electromagnetic noise (EMambient) is characterized as relatively constant over time and comprises the electromagnetic interference generated by the one or more determined electronic devices.

6. The method (700) as claimed in claim 1, wherein, when the one or more EM noises comprise an EM noise from speakers (EMspeaker), identifying the one or more EM noises comprises:
i. determining one or more characteristics of an input audio signature and EM noise metadata,
wherein the EM noise metadata is a descriptive file formatted according to a standard specification,
wherein the EM noise metadata is designed to be paired with one or more audio files intended for playback,
wherein the EM noise metadata comprises information associated with one or more ML models that weights pertinent to the EM noise emissions anticipated from a variety of speaker models and sound configurations during the playback of the one or more audio files; and
ii. identifying the EM noise from speakers (EMspeaker) based on the one or more determined characteristics.

7. The method (700) as claimed in claim 6, further comprising:
a. predicting EM noise generated by speaker movements during active Acoustic Noise Cancellation (ANC) operations in the wearable electronic device (100);
b. analyzing, in real-time, an ANC input signal transmitted to the speaker, wherein the ANC signal includes an anti-noise waveform generated to cancel external acoustic noise;
c. modeling the ANC-induced EM noise caused by rapid diaphragm oscillations and magnetic field fluctuations, using predefined speaker metadata and operational parameters, including input signal characteristics, driver efficiency, and coil dynamics; and
d. incorporating a predicted ANC-induced EM noise based on the modeling into the EMspeaker classification, ensuring that a total electromagnetic noise profile accurately accounts for all speaker-related emissions, to improve biosignal isolation and enhance the overall quality of the one or more collected bio-signals.

8. The method (700) as claimed in claim 1, wherein, when the one or more EM noises comprise an EM noise from internal device components (EMradio), identifying the one or more EM noises comprises:
i. determining a presence of one or more electronic devices within the surrounding environment that emit EM interference,
ii. determining, by utilizing the one or more ML models, one or more radio parameters comprise an operation characteristic of a radio circuit associated with the one or more determined electronic devices, data ingress and egress within the radio circuit, one or more radio transmission principles; and
iii. identifying the EM noise from internal device components (EMradio) based on the one or more determined radio parameters.

9. The method (700) as claimed in claim 1,
i. wherein an impulsive noise (EMother) associated with the one or more EM noises comprises at least one category of short-duration, high-amplitude bursts produced by one or more sources; and
ii. wherein the one or more sources comprise at least one of one or more atmospheric events and one or more domestic appliances.

10. The method (700) as claimed in claim 1, wherein the at least one ML model is trained by:
i. performing one or more data processes on the identified electromagnetic (EM) noises, wherein the data processes comprise data integration to consolidate various data sources, data cleaning to remove inaccuracies and inconsistencies, data reduction to minimize data volume while retaining essential information, data transformation to convert data into suitable formats for analysis, and data discretization to convert continuous data into discrete values for improved processing;
ii. splitting the processed EM noises into at least one of a training set, validation set, and testing set, wherein the training set is used to train the ML model, the validation set is utilized for optimizing hyperparameters, and the testing set is reserved for unbiased performance evaluation of the trained ML model;
iii. feeding the split data into the ML model for training, wherein the training process comprises model initialization to establish starting parameters, loss function selection to define error measurements, optimizer selection to determine algorithms for parameter updates, and hyperparameter tuning to refine model performance; and
iv. performing evaluation processes to determine the optimal ML model by assessing performance metrics, conducting testing set evaluations, and executing error analysis to identify and correct prediction errors.

11. The method (700) as claimed in claim 1, further comprising:
i. implementing the at least one ML model, comprising an edge computing module within the wearable electronic device (100), to enable real-time processing and analysis of bio-signals and identified EM noises;
ii. Utilizing the at least one ML model comprising one or more transformer models to classify and predict noise patterns from the scaled bio-signals and EM noises; and
iii. employing one or more advanced ML models comprising Large Language Models (LLMs), to analyze metadata and dynamically optimize noise cancellation parameters based on real-time environmental conditions and historical data; and
iv. defining the compound noise signal (EMtotal) as a sum of all EM noises.

12. The method (700) as claimed in claim 1, wherein enhancing the quality of the one or more collected bio-signals comprises:
i. assigning a dynamic weight to each identified EM noise, wherein the weight is continuously adjusted by the at least one ML model based on real-time data and contextual information;
ii. continuously monitoring the surrounding environment to enable the at least one ML model to discern the significance of various identified EM noises; and
iii. prioritizing one or more most impactful EM noise while effectively mitigating one or more less impactful EM noise, for efficient noise cancellation.

13. The method (700) as claimed in claim 1, wherein cancelling, by the wearable electronic device (100), the generated compound noise signal (EMtotal) from the one or more collected bio-signals comprising:
i. performing at least one of :
a) programmatically canceling the one or more EM noises from the one or more collected bio-signals in real-time based on at least one mechanism comprising an actual playback mechanism, a simulation-based mechanism, and a metadata-based mechanism; or
b) canceling the one or more EM noises from the one or more collected bio-signals by utilizing one or more adaptive filtering mechanisms.

14. The method (700) as claimed in claim 13, wherein programmatically canceling the one or more EM noises from the one or more collected bio-signals in real-time based on the metadata-based mechanism comprising:
i. associating an audio file with a metadata file that specifies one or more expected electromagnetic noise characteristics generated by one or more speaker and radio modules during a playback,
a) wherein the metadata file describes the noise profile based on one or more audio signal properties comprising a frequency, an amplitude, and one or more anticipated interference patterns from one or more operations associated with the one or more speaker and radio modules; and
ii. referencing, during the playback, the metadata file to predict one or more generated EM noises, to eliminate a real-time computation, and to reduce processing load by avoiding real-time calculations;
a) wherein the metadata file enables an application of one or more pre-configured filters and adaptive algorithms to efficiently cancel one or more pre-defined EM noises, to ensure highly accurate noise cancellation tailored to a specific audio file and device configuration.

15. The method (700) as claimed in claim 13, wherein programmatically canceling the one or more EM noises from the one or more collected bio-signals in real-time based on the metadata-based mechanism comprising:
i. employing a unit test signal to validate identification and software-based removal of sounds, generated by a speaker, based on a metadata file;
ii. detecting one or more deviations during the employment of the unit test signal to assess an accuracy of an audio cancellation; and
iii. dynamically editing and updating the metadata file in response to detecting the one or more deviations, reflecting predefined acoustic characteristics under current usage conditions, for canceling the one or more EM noises from the one or more collected bio-signals in real-time.

16. The method (700) as claimed in claim 15, further comprising:
i. storing the dynamically edited and updated metadata file in a user profile on a cloud server, for ensuring preservation and synchronization across one or more electronic devices; and
ii. downloading and applying, upon synchronization across the one or more electronic devices, the saved metadata file, to ensure consistent audio cancellation across the one or more electronic devices associated with the user profile,
a) wherein each metadata file is associated with a version, and each version indicates one or more changes made to the metadata file, ensuring the use of the most recent and accurate file while preserving previous versions for potential rollback,
b) wherein each metadata file is stored in the cloud server along with one or more annotations related to one or more environmental conditions.

17. The method (700) as claimed in claim 1, wherein identifying one or more EM noises associated with the wearable electronic device (100) comprises:
i. generating an acoustic metadata file associated with a speaker system within the wearable electronic device (100), wherein the acoustic metadata file includes:
a) a frequency response profile defining output characteristics of the speaker system across a predefined frequency range;
b) harmonic distortion data at one or more operating frequencies and amplitudes;
c) one or more transient response times of the speaker system indicating stabilization delays after input signals;
d) one or more phase response characteristics correlating input and output signals across key frequencies; and
e) one or more speaker-specific operational profiles tailored for different playback modes;
ii. associating the acoustic metadata file with audio input signals intended for playback as well as the inputs for Active acoustic Noise cancellation (ANC), to predict the speaker's acoustic output;
iii. detecting a composite sound signal from within an ear canal, wherein the composite sound signal includes the predicted acoustic output, ambient noise, and one or more acoustic biosignals comprising a sound from the blood vessels in the vicinity of ear canal;
iv. isolating, via the acoustic metadata file, the predicted acoustic output of the speaker system from the composite sound signal to enhance the accuracy of biosignal extraction; and
v. dynamically updating the acoustic metadata file in response to detected deviations during real-time operation, ensuring accurate noise isolation and adaptive audio cancellation in varying conditions.

18. The method (700) as claimed in claim 1, comprising:
i. determining whether a value of an ambient electromagnetic noise (EMambient) associated with the one or more EM noises is greater than a value of an electromagnetic noise threshold (EMthreshold); and
ii. performing one of:
a) collecting the one or more bio-signals from the user of the wearable electronic device (100) in response to determining that the value of the ambient electromagnetic noise (EMambient) is lower than the value of an electromagnetic noise threshold (EMthreshold); and
b) transmitting an alert message to the user in response to determining that the value of the ambient electromagnetic noise (EMambient) is greater than the value of an electromagnetic noise threshold (EMthreshold), wherein the alert message indicates that an EM noise level is too high for collecting the one or more bio-signals and if the user wants to continue.
c) Warning the user if the detected EM noise exceeds safe limits for human exposure, indicating a potential health risk.

19. A system (101) comprising:
i. a memory (110);
ii. a processor (120);
iii. an Electro-Magnetic (EM) noise controller (121), operably connected to the memory (110) and the processor (120), configured to:
a) collect one or more bio-signals from a user of the wearable electronic device (100);
b) identify one or more Electro-Magnetic (EM) noises present in a surrounding environment of a wearable electronic device (100) during the collection of the one or more bio-signals;
c) scale the one or more collected bio-signals, and the one or more identified EM noises;
d) analyze the one or more scaled bio-signals and the one or more scaled EM noises by utilizing at least one Machine Learning (ML) model;
e) generate a compound noise signal (EMtotal) based on a result of the analysis; and
f) cancel the generated compound noise signal (EMtotal) from the one or more collected bio-signals, to enhance quality of the one or more collected bio-signals.

20. The system (101) as claimed in claim 19, wherein canceling the one or more EM noises from the one or more collected bio-signals in real-time based on a metadata-based mechanism comprising:
i. associating an audio file with a metadata file that specifies one or more expected electromagnetic noise characteristics generated by one or more speaker and radio modules during a playback,
a) wherein the metadata file describes the noise profile based on one or more audio signal properties comprising a frequency, an amplitude, and one or more anticipated interference patterns from one or more operations associated with the one or more speaker and radio modules; and
ii. referencing, during the playback, the metadata file to predict one or more generated EM noises, to eliminate a real-time computation, and to reduce processing load by avoiding real-time calculations;
a) wherein the metadata file enables an application of one or more pre-configured filters and adaptive algorithms to efficiently cancel one or more pre-defined EM noises, to ensure highly accurate noise cancellation tailored to a specific audio file and device configuration.