System for the intelligent elimination of motion artifacts in portable systems for the acquisition of biomedical signals

Abstract

A system for the intelligent removal of motion artifacts in portable biomedical signal acquisition systems, comprising: a housing configured to be worn on a person's body surface and enclosing a variety of functionally interconnected components; A biosignal acquisition unit is arranged within the housing structure and configured to capture physiological signals from the subject via at least one sensor interface; an inertial sensor unit comprising at least one accelerometer and at least one gyroscope, wherein the inertial sensor unit is configured to generate motion-related signals corresponding to the movement of the housing structure; an analog processing unit that is operationally coupled with the biosignal acquisition unit and is configured to amplify, impedance-match, and filter the acquired physiological signals using programmable amplification circuits and anti-aliasing filters; a digitization unit that is operationally connected to the analog processing unit and the inertial sensor unit, wherein the digitization unit comprises at least one analog-to-digital converter configured to convert the physiological signals and the motion-related signals into synchronized digital representations; a storage unit configured to store digitized signal data, calibration parameters, and adaptive coefficients; a computing unit comprising at least one processor operationally linked to the digitization unit and the storage unit, wherein the computing unit is configured to temporally align the physiological signals and the motion-related signals, decompose the physiological signals into multiple signal components, identify motion-correlated components based on their correlation with the motion-related signals, generate adaptive filter coefficients, and reconstruct an artifact-cleaned physiological signal; and a communication unit configured to transmit the reconstructed physiological signal to an external device.

Description

Technical field of the invention

[0001] The present invention relates to systems for acquiring and processing biomedical signals, in particular a hardware-based system and device for the real-time identification, isolation, and removal of motion-induced artifacts in physiological signals acquired by wearable sensors. The invention relates in particular to structural and computational systems for improving signal quality in electrocardiography (ECG), electromyography (EMG), photoplethysmography (PPG), and similar biosignals in dynamic environments. BACKGROUND OF THE INVENTION

[0002] Portable biomedical devices are increasingly used for continuous monitoring of physiological parameters in outpatient settings. However, the accuracy and reliability of such systems are significantly affected by motion artifacts caused by body movements, sensor displacement, and environmental influences. Conventional filtering methods, including static bandpass filters and adaptive filtering techniques, are insufficient to compensate for non-stationary noise patterns that superimpose on the frequency components of the desired biomedical signals.

[0003] Existing systems typically rely on software-based post-processing methods, which introduce latency, increased computational overhead, and limited real-time capability. Furthermore, these methods are not robust enough to handle multi-axis motion artifacts and do not dynamically adapt to changing signal conditions. Therefore, there is a need for a structurally integrated, hardware-oriented system that can intelligently remove motion artifacts in real time without compromising signal integrity.

[0004] Wearable biomedical signal acquisition systems have become an indispensable component of modern healthcare, enabling continuous, real-time monitoring of physiological parameters such as cardiac activity, muscle response, blood oxygen saturation, and respiratory patterns. These systems are widely used in areas such as remote patient monitoring, fitness tracking, rehabilitation, and early disease detection. Despite significant advances in miniaturization, sensor design, and wireless communication, motion artifacts remain a persistent and fundamental challenge to the reliability of such systems. Motion artifacts are unwanted disturbances of the acquired biosignals caused by relative movement between the sensor and skin, tissue deformation, impedance changes between the electrode and skin, and external mechanical vibrations.These artifacts often overlap spectrally and temporally with the actual physiological signals, making their identification and removal particularly complex.

[0005] Conventional wearable systems typically use analog front-end filtering techniques for noise reduction. These include bandpass filters, which preserve the relevant frequency components while attenuating interference outside the frequency band. For example, electrocardiogram signals are often filtered within a predefined frequency range to eliminate baseline wander and high-frequency interference. However, motion artifacts do not follow fixed frequency bands and often exhibit non-stationary characteristics that overlap with the desired signal spectrum. Therefore, static filtering techniques are inherently limited, as they either fail to effectively remove artifacts or unintentionally distort important signal features such as QRS complexes in ECGs or pulse peaks in photoplethysmography signals.

[0006] To overcome these limitations, adaptive filtering techniques have been introduced. These techniques use motion-correlated reference signals, such as accelerometer data, to estimate and subtract artifact components. Techniques like least mean squares (RLS) and recursive least squares (RLS) are employed to dynamically adjust the filter coefficients to changing signal conditions. While these methods offer improved performance compared to static filters, they are often implemented in software and require iterative computations, resulting in increased processing latency and higher power consumption. Furthermore, the effectiveness of adaptive filters is highly dependent on the quality and orientation of the reference motion signals.In practical applications with portable devices, misalignment between motion sensors and biosignal acquisition points can lead to inaccurate artifact estimation and incomplete artifact removal.

[0007] Another category of existing solutions includes signal decomposition techniques such as wavelet transformation, empirical mode decomposition (EMD), and independent component analysis (ICA). These methods decompose the acquired signal into multiple components based on frequency, scale, or statistical independence to identify and remove components associated with motion artifacts. While these methods improve the processing of non-stationary signals, they are computationally intensive and typically performed in post-processing rather than in real time. This limits their applicability in continuous monitoring systems that require immediate feedback. Furthermore, decomposition-based methods often require the manual or heuristic selection of parameters, such as...Wavelet basic functions or intrinsic mode selection criteria that may not translate well to different users or movement conditions.

[0008] Machine learning has also been explored for the detection and removal of motion artifacts. These methods train models using annotated datasets to distinguish between clean and corrupted signal segments. Techniques such as support vector machines, neural networks, and deep learning architectures have shown promising results in controlled environments. However, when applied to wearable systems, these approaches have several drawbacks. First, they require large amounts of annotated training data, which are difficult to obtain for diverse real-world motion scenarios. Second, the computational cost of such models is high, requiring powerful processors and memory resources that are typically unavailable in compact wearable devices.Thirdly, these models often lack interpretability and may not be applicable to different sensor positions, user behavior, and environmental conditions.

[0009] Hardware-based approaches to reducing motion artifacts have also been proposed, focusing on improving electrode design, sensor attachment mechanisms, and mechanical stability. For example, flexible electrodes and adhesive materials have been developed to ensure consistent skin contact during movement. While such solutions reduce the occurrence of motion artifacts to some extent, they do not eliminate the problem entirely, especially during intense activities. Furthermore, improvements in mechanical design often increase the complexity and cost of the devices, as well as their comfort, limiting their widespread adoption.

[0010] Another limitation of existing systems is the lack of synchronized multimodal sensors. Many wearables operate with independent sensor units that do not share a common temporal reference point. This leads to a misalignment of physiological signals and motion data. This temporal discrepancy impairs the effectiveness of artifact detection methods based on correlation analysis. Furthermore, existing systems often lack feedback mechanisms to adapt to changing signal conditions. Consequently, their performance deteriorates during extended monitoring scenarios where sensor characteristics and user behavior change.

[0011] Energy efficiency is another crucial aspect of portable systems. Advanced signal processing techniques, especially software-based ones, consume significant computing resources, leading to increased power consumption and reduced battery life. This creates a trade-off between signal quality and device lifespan, which is undesirable in applications requiring continuous monitoring over extended periods. Furthermore, frequent data transfers to external devices for processing introduce additional power consumption and potential latency, impacting the system's real-time capability.

[0012] In many existing solutions, the separation of sensors, data processing, and communication leads to inefficiencies in data processing. Raw signals or only minimally processed signals are often transmitted to external systems for analysis, which increases bandwidth requirements and exposes sensitive physiological data to potential security risks. The lack of integrated data processing in the wearable not only limits real-time artifact removal but also places an additional burden on the external infrastructure.

[0013] Another disadvantage of current methods is their limited ability to preserve clinically relevant features while simultaneously removing artifacts. Excessive filtering or aggressive artifact suppression can distort signal morphology, leading to misinterpretations and potential misdiagnoses. For example, in cardiac monitoring, subtle variations in waveform morphology are crucial for detecting arrhythmias and other disorders. Existing systems often fail to achieve an optimal balance between artifact removal and signal preservation.

[0014] Furthermore, scalability and adaptability remain significant challenges. Wearables are used by diverse populations with varying physiological characteristics, activity levels, and environmental conditions. Existing solutions are often tailored to specific use cases and lack the flexibility to adapt to different scenarios without extensive recalibration or reconfiguration. This limits their applicability for large-scale deployments and heterogeneous user environments.

[0015] In summary, while numerous techniques have been developed to address motion artifacts in wearable biomedical signal acquisition systems, each method has inherent limitations. Static filtering is inadequate for non-stationary noise, adaptive filtering is computationally intensive and dependent on the quality of the reference signal, decomposition is not real-time capable, machine learning requires substantial resources and training data, and hardware-based solutions do not completely eliminate artifacts. Furthermore, issues related to synchronization, energy efficiency, data processing, and signal integrity further complicate the problem.These challenges underscore the need for a structurally integrated, hardware-oriented system that can intelligently and in real time remove motion artifacts and operate efficiently within the limitations of portable devices, while ensuring high signal quality. SUBJECT OF THE INVENTION

[0016] The present invention aims to provide a structurally integrated system and device for intelligent motion artifact suppression in portable biomedical signal acquisition systems, wherein the system comprises synchronized data acquisition with multiple sensors, hardware-level preprocessing, adaptive signal decomposition, and real-time artifact suppression using embedded computing units. SUMMARY OF THE INVENTION

[0017] The present invention discloses a system and a corresponding device comprising a housing structure enclosing a plurality of interconnected hardware components, including a biosignal acquisition unit, an inertial sensor unit, an analog front-end conditioning unit, a digitization unit, a storage unit and a computing unit with one or more processors.

[0018] The biosignal acquisition unit is configured to receive physiological electrical signals via electrodes or optical sensors on the user's body. The inertial sensor unit includes three-axis accelerometers and gyroscopes that capture motion-related parameters corresponding to the movement of the wearable device.

[0019] The analog input stage includes programmable amplifiers, impedance matching circuits, and anti-aliasing filters for conditioning the captured signals. The digitization unit consists of high-resolution analog-to-digital converters that convert the processed analog signals into digital representations.

[0020] The processing unit is configured to perform a sequence of hardware-implemented operations, including time-synchronized data alignment, multi-channel signal decomposition, motion correlation analysis, adaptive filter coefficient generation, and artifact suppression. The system also includes a memory unit for storing intermediate signal states, calibration parameters, and adaptive coefficients.

[0021] The device is structurally configured to operate in real time. Motion artifacts are identified by correlating inertial sensor data with biomedical signal disturbances, followed by selective attenuation or reconstruction of the disturbed signal components.

[0022] The present invention relates to a system and an associated wearable device for the intelligent removal of motion artifacts in biomedical signal acquisition systems. The system is designed to detect, characterize, and suppress motion-induced disturbances in real time without compromising the integrity of the underlying physiological signals. The object of the invention is high-precision signal acquisition in dynamic environments through the integration of biosignal sensor components with motion sensors and hardware-based computing units in a unified architecture.

[0023] A further objective of the invention is to provide a device for the synchronous acquisition of physiological signals and motion parameters using spatially arranged sensor units. This enables precise temporal alignment and correlation between motion events and signal distortions. The invention aims to identify motion artifacts with high precision by utilizing the spatial and temporal coherence between several sensor modalities integrated in a common housing.

[0024] A further objective of the invention is the implementation of signal conditioning and digitization mechanisms at the hardware level, which improve signal quality prior to computer-aided processing, thereby reducing noise propagation and increasing the effectiveness of subsequent artifact removal operations. The invention aims to integrate programmable gain control, impedance stabilization, and anti-aliasing functions into the analog front end to ensure optimal signal reproduction under various physiological and environmental conditions.

[0025] A further objective of the invention is to provide a computing arrangement with one or more processors that execute adaptive and transformation-based signal decomposition methods in real time. The system is capable of isolating motion-correlated components and selectively attenuating or reconstructing affected signal segments. The invention further aims to enable the dynamic generation and updating of filter parameters based on the continuous evaluation of signal characteristics and the estimation of the residual error.

[0026] A further objective of the invention is to minimize latency and power consumption by integrating artifact removal operations into the portable device itself. This eliminates the need for external processing and reduces data transmission overhead. The invention aims for energy-efficient operation suitable for continuous long-term monitoring while simultaneously ensuring computational accuracy and responsiveness.

[0027] A further objective of the invention is to preserve clinically relevant features of biomedical signals during artifact suppression, in order to ensure that critical waveform characteristics are maintained for precise diagnosis and analysis. The invention aims to preserve the morphological integrity of signals such as ECG, EMG, and PPG while effectively eliminating motion-induced distortions.

[0028] A further objective of the invention is to provide a feedback-based calibration mechanism within the device, enabling the system to continuously adapt to user-specific movement patterns, variations in sensor placement, and environmental changes, thereby increasing robustness and reliability over extended periods of use. The invention aims to support adaptive learning capabilities through the storage and use of historical signal data and calibration parameters.

[0029] Another objective of the invention is to provide a compact and ergonomically designed portable device structure that integrates sensor, processing and communication components into a single housing, thus ensuring ease of use, mechanical stability and uniform sensor contact with the user's body.

[0030] Another objective of the invention is to enable the secure and efficient transmission of processed, artifact-free signals to external monitoring systems via an integrated communication interface, thereby supporting applications in the field of remote treatment and real-time clinical decision-making.

[0031] Overall, the invention aims to overcome the limitations of existing solutions by providing a structurally integrated, hardware-oriented system capable of intelligently, adaptively, and in real time removing motion artifacts, thereby significantly improving the reliability and applicability of portable biomedical signal acquisition systems under practical, real-world conditions. BRIEF DESCRIPTION OF THE IMAGE

[0032] These and other features, aspects and advantages of the present invention will be better understood if the following detailed description is read with reference to the accompanying drawing, in which the same symbols represent the same parts: Fig. Figure 1 shows a block diagram of a quantum-optimized system for detecting fake news.

[0033] Furthermore, those skilled in the art will recognize that the elements in the drawing are simplified and not necessarily drawn to scale. For example, the flowcharts illustrate the process by highlighting the main steps to facilitate understanding of the present disclosure. With regard to the construction of the device, one or more components may be represented in the drawing by conventional symbols. The drawing may show only those specific details relevant to understanding the embodiments of the present disclosure, so as not to clutter the drawing with details that are already apparent to those skilled in the art from the description contained herein. Detailed description of the invention

[0034] To facilitate understanding of the principles of the invention, reference is made below to the embodiment shown in the drawing, which is described using specific terms. It is understood, however, that this does not limit the scope of protection of the invention. Rather, modifications and further developments of the depicted system, as well as further applications of the inventive principles shown therein, are conceivable, insofar as they would normally occur to a person skilled in the art in the field of the invention.

[0035] It will be clear to those skilled in the art that the foregoing general description and the following detailed description are exemplary and explanatory of the invention and are not to be understood as a limitation thereof.

[0036] References to “an aspect”, “another aspect”, or similar phrases in this description mean that a particular feature, structure, or property described in connection with the embodiment is included in at least one embodiment of the present disclosure. Therefore, phrases such as “in one embodiment”, “in another embodiment”, and similar expressions in this description may, but do not necessarily, all refer to the same embodiment.

[0037] The terms "includes," "comprehensive," or similar expressions denote non-exclusive inclusion. Thus, a procedure or method containing a list of steps does not only include those steps but may also include further steps not explicitly listed or inherent in the procedure or method. Likewise, the statement "includes..." for one or more devices, subsystems, elements, structures, or components, without further limitations, does not preclude the existence of other devices, subsystems, elements, structures, or components.

[0038] Unless otherwise defined, all technical and scientific terms used herein have the same meanings generally known to those skilled in the art in the field to which this invention belongs. The systems, methods, and examples described herein serve only for illustration and are not to be understood as limiting.

[0039] Embodiments of the present disclosure are described in detail below with reference to the attached drawing.

[0040] Fig.Figure 1 shows a block diagram of a system for the intelligent removal of motion artifacts in wearable biomedical signal acquisition systems. The system 100 comprises: a housing (102) worn on the body surface of a test subject and containing several functionally interconnected components; a biosignal acquisition unit (104) located inside the housing, which acquires physiological signals from the test subject via at least one sensor interface; an inertial sensor unit (106) with at least one accelerometer and at least one gyroscope, which generates motion-related signals according to the movement of the housing;An analog signal processing unit (108) coupled to the biosignal acquisition unit, which amplifies, impedance-matches, and filters the acquired physiological signals using programmable gain circuits and anti-aliasing filters. A digitization unit (110) connected to the analog signal processing unit and the inertial sensor unit, the digitization unit comprising at least one analog-to-digital converter for converting the physiological and motion-related signals into synchronized digital representations; a storage unit (112) for storing digitized signal data, calibration parameters, and adaptive coefficients;a computing unit (114) with at least one processor, which is connected to the digitization unit and the storage unit and which temporally aligns the physiological and motion-related signals, decomposes the physiological signals into several signal components, identifies motion-correlated components based on their correlation with the motion-related signals, generates adaptive filter coefficients and reconstructs an artifact-free physiological signal; and a communication unit (116) for transmitting the reconstructed physiological signal to an external device.

[0041] In one embodiment, the biosignal acquisition unit (104) comprises a plurality of electrode interfaces or optical sensor elements arranged spatially to ensure stable contact with the body surface, and further comprises shielding conductors configured to reduce electromagnetic interference during signal acquisition.

[0042] In one embodiment, the inertial sensor unit (106) is physically arranged together with the biosignal acquisition unit at a predefined distance to ensure spatial correspondence between motion measurements and physiological signal disturbances. The inertial sensor unit is configured to sample motion data along three orthogonal axes.

[0043] In one embodiment, the analog processing unit (108) comprises a cascade arrangement of low-noise amplifiers and tunable filters, wherein the gain parameters are dynamically adjusted based on signal amplitude changes detected by the computing unit.

[0044] In one embodiment, the digitization unit (110) comprises a plurality of high-resolution analog-to-digital converters configured to operate at a sampling rate sufficient to capture both the bandwidth of physiological signals and the dynamics of motion, and wherein the digitization unit includes a synchronization circuit configured to assign a timestamp to each sampled signal.

[0045] In one embodiment, the computing unit (114) is configured to perform signal decomposition using hardware-implemented transform-based operations, whereby the physiological signals are decomposed into several frequency or scale components that represent different signal characteristics.

[0046] In one embodiment, the computing unit (114) is further configured to compare each decomposed component with corresponding motion-related signals in order to determine correlation coefficients and to classify components that exceed a predefined correlation threshold as motion artifact components.

[0047] In one embodiment, the computing unit (114) is configured to generate adaptive filter coefficients by means of iterative optimization circuits based on minimizing the error between predicted artifact components and actual signal components.

[0048] In one embodiment, the computing unit (114) is configured to selectively attenuate or reconstruct the identified motion artifact components and to recombine the remaining components to generate the artifact-suppressed physiological signal.

[0049] In one embodiment, the computing unit (114) further comprises a feedback circuit configured to continuously estimate the residual error between the reconstructed physiological signal and the expected signal characteristics and to update the filter coefficients stored in memory.

[0050] The system is fully implemented in dedicated electronic and electromechanical hardware components. The biosignal acquisition unit utilizes physical electrodes or optical sensor interfaces coupled to low-noise instrumentation amplifier circuits and impedance matching stages. The inertial sensor unit consists of integrated accelerometers and gyroscopes that output analog or digital motion signals along orthogonal axes. The analog signal conditioning unit is equipped with cascaded operational amplifier networks, programmable amplifier circuits, and passive and active analog filters configured for aliasing correction prior to digitization. The digitization unit comprises hardware analog-to-digital converter circuits with clocked sampling and synchronization logic that assigns time-stamped digital values ​​to physiological and motion signals.The processing unit is implemented with fixed-functionality or reconfigurable digital logic, including dedicated arithmetic units, multiplier-accumulator arrays, correlator hardware blocks, and transformation processing circuits that decompose signal streams into frequency or scale components without software instructions. Correlation evaluation, component classification, and adaptive coefficient generation are performed using hardware-based comparison circuits, iterative weight update circuits, and feedback-controlled filter coefficient registers. The storage unit is implemented as a physical electronic storage element containing calibration parameters and adaptive filter weights, while the communication unit includes a dedicated transmission circuit for forwarding reconstructed signals.

[0051] The present invention relates to a system for the intelligent removal of motion artifacts in wearable biomedical signal acquisition systems. The system's operation is controlled by a sequence of hardware-executed signal acquisition, synchronization, decomposition, correlation, adaptive filtering, and reconstruction processes, performed by a computing unit operationally coupled with sensor, processing, and storage elements. The system is designed such that physiological signals from a subject and motion-related signals from an inertial sensor unit are temporally synchronized and processed computationally efficiently to achieve real-time artifact suppression while simultaneously preserving clinically relevant signal morphology.

[0052] During operation, the biosignal acquisition unit records analog physiological signals from the subject via sensor interfaces, which may include conductive electrodes or optical transmitters and receivers. These signals, which can represent electrical cardiac activity, muscle activation, or changes in blood volume, are inherently susceptible to distortion due to motion-induced changes in contact impedance and mechanical displacements. Simultaneously, the inertial sensor unit generates motion-related signals corresponding to translational and rotational movements of the housing. These motion signals represent external disturbances that contribute to artifact formation in the physiological signals.

[0053] The analog signal processing unit receives the physiological signals and optimizes them through amplification, impedance stabilization, and frequency-selective filtering. The amplification stage utilizes a low-noise circuit with dynamically adjustable gain parameters to compensate for fluctuations in signal amplitude. The filter stage attenuates interference signals outside the frequency band without significantly altering the spectrum of the physiological signal. The processed signals, along with the motion signals, are then forwarded to the digitization unit. There, high-resolution analog-to-digital converters capture the signals at a predefined sampling rate sufficient to record both the physiological and motion dynamics. The digitization unit assigns a synchronized timestamp to each sample, thus ensuring the temporal consistency of the physiological and motion data.

[0054] After digitization, the processing unit retrieves the synchronized data and performs time alignment to compensate for any remaining latency differences between the signal channels. This alignment ensures the precise assignment of corresponding physiological and motion data for further processing. The aligned physiological signal then undergoes decomposition, which is implemented using hardware-based transformation operations. This process breaks the signal down into multiple components, each representing a specific frequency band or intrinsic oscillation mode. Decomposition enables the isolation of signal components that may contain motion-related interference.

[0055] Simultaneously, the motion-related signals are analyzed to extract characteristic features such as amplitude variations, changes in direction, and temporal patterns of movement. The processing unit performs a correlation analysis between each decomposed physiological signal component and the motion-related signals. This analysis includes the calculation of correlation measures that quantify the degree of similarity between the temporal variations of the physiological components and the motion signals. Components with high correlation values ​​are identified as being influenced by motion artifacts, while components with low correlation values ​​are considered to be expressions of genuine physiological activity.

[0056] After identifying motion-related components, the processing unit generates adaptive filter coefficients using a hardware-implemented optimization circuit. Coefficient generation is based on iterative error minimization, continuously evaluating the difference between the estimated artifact component and the actual signal component. The optimization circuit adjusts the coefficients to minimize this error, thus achieving a precise estimate of the artifact present in each affected component. Compared to software-based implementations, this hardware-based optimization ensures faster convergence and lower computational latency.

[0057] The identified artifact components are then attenuated or corrected using the generated filter coefficients. In some cases, the system suppresses the amplitude of the artifact components; in others, it reconstructs the components by subtracting the estimated artifact contribution from the original signal. The corrected components are then combined with the unaffected components via an inverse transformation to reconstruct a largely motion-artifact-free physiological overall signal. This reconstruction process is carefully controlled to preserve essential signal features such as waveform morphology, peak amplitudes, and time intervals, which are crucial for accurate clinical interpretation.

[0058] The processing unit also features a feedback mechanism that continuously monitors the quality of the reconstructed signal. This mechanism assesses the residual error by comparing the reconstructed signal with the expected physiological patterns stored in memory. If deviations are detected, the system updates the adaptive filter coefficients and decomposition parameters to improve artifact suppression in subsequent processing cycles. This feedback-driven adaptation allows the system to account for changes in user behavior, sensor placement, and environmental conditions over time.

[0059] The storage unit plays a crucial role in supporting technical operations by storing intermediate signal results, decomposition parameters, correlation thresholds, and adaptive coefficients. It also stores historical signal data and calibration profiles, which the processing unit uses to optimize processing parameters based on previous observations. This allows the system to learn user-specific motion characteristics and optimize its performance over time.

[0060] The entire processing sequence is executed in a pipeline architecture within the computing unit, with multiple processing stages operating simultaneously on successive signal segments. This parallelism ensures that signal acquisition, decomposition, correlation analysis, filtering, and reconstruction occur continuously and overlapping, thus achieving real-time operation with minimal latency. The high-speed data connection between the digitization unit and the computing unit enables uninterrupted data flow, eliminates buffer delays, and ensures consistent throughput.

[0061] The reconstructed, artifact-corrected physiological signal is then transmitted via the communication unit to external monitoring systems or storage media. Before transmission, the signal can be encoded and secured using encryption circuits to protect sensitive medical data. The system is designed for continuous operation under changing movement conditions. The processing unit dynamically adjusts the processing parameters to the changes in movement intensity and direction detected by the inertial sensors.

[0062] Overall, the technology implemented in the system offers a robust and efficient mechanism for the intelligent removal of motion artifacts. This is achieved through the integration of synchronized multi-sensor data acquisition, transformation-based signal decomposition, correlation-driven artifact identification, adaptive filtering, and feedback-based optimization within a hardware-oriented framework. This integrated approach enables the precise and reliable extraction of physiological signals in real-world wearable applications where motion artifacts are common and conventional processing techniques are inadequate.

[0063] In one embodiment, the system comprises a compact, portable device in a biocompatible housing that ensures stable skin contact. The housing integrates electrode or optical sensor interfaces for capturing physiological signals with minimal background noise. The biosignal acquisition unit features conductive traces and shielding structures that minimize electromagnetic interference and ensure signal integrity.

[0064] The inertial sensor unit is spatially connected to the biosignal acquisition unit to ensure spatial correlation between motion data and physiological signal disturbances. The inertial sensor unit generates multi-axis acceleration and angular velocity signals, which are continuously sampled and transmitted to the digitization unit.

[0065] The analog input signal processing unit consists of a cascade of low-noise amplifiers and tunable filters, whose gain parameters are dynamically adjusted based on the signal amplitude and noise characteristics. The processed signals are then fed to a high-resolution digitization unit comprising sigma-delta analog-to-digital converters configured for sampling rates compatible with both the bandwidth of physiological signals and the dynamics of motion.

[0066] The processing unit consists of a dedicated hardware processor that performs signal processing operations at the hardware level. The processor is connected to the digitization unit and memory via high-speed data buses. Using timestamp alignment circuits, the processor synchronizes the biosignal data streams and the data from the inertial sensors.

[0067] The processor then performs signal decomposition using hardware-implemented, transformation-based methods, including discrete wavelet transform (DWT) or empirical mode decomposition (EMD). The input signal is decomposed into intrinsic components representing different frequency bands. Simultaneously, motion signals from the inertial sensors are analyzed to identify frequency components corresponding to motion artifacts.

[0068] The processor then performs a correlation analysis between the decomposed biosignal components and the motion signals. Components with a high correlation to the motion data are identified as artifact-dominant components. Adaptive filter coefficients are then generated using hardware-implemented optimization circuits based on the least squares method (LMS) or recursive least squares (RLS).

[0069] The identified artifact components are attenuated or reconstructed using inverse transformation techniques. This restores the clean signal by combining the unaffected and corrected components. The system ensures minimal distortion of physiological features such as QRS complexes in ECGs or pulse waveforms in PPG signals.

[0070] In another embodiment, the system features a feedback calibration mechanism in which the processor continuously updates the filter parameters based on the residual error estimate. The memory unit stores historical signal patterns and calibration profiles, thus enabling adaptive learning of user-specific movement characteristics.

[0071] The device also includes an energy management unit to regulate energy consumption and enable continuous operation in portable applications. Thermal management structures are integrated into the housing to dissipate the heat generated by the processing unit.

[0072] The system remains configured to output the processed, artifact-free signal to external devices via a communication interface with a wireless transmission circuit. This communication interface ensures secure and low-latency transmission of the processed data to monitoring systems or mobile devices.

[0073] During operation, the system continuously acquires physiological signals and motion data, compares the data streams, decomposes the signals into multiple components, identifies motion-related artifacts, applies adaptive filtering, and reconstructs a noise-free signal in real time. The structural integration of sensors, processing, and filtering enables efficient and precise artifact removal without external processing systems.

[0074] The present invention provides a structurally integrated, hardware-based solution for the intelligent removal of motion artifacts, enabling real-time processing with reduced latency. The system improves the signal quality and reliability of wearable biomedical devices in dynamic environments. The integration of inertial sensors and adaptive filtering mechanisms allows for the precise identification and suppression of motion artifacts without compromising critical physiological information. The device is compact, energy-efficient, and suitable for continuous health monitoring applications.

[0075] The described system and device for intelligent motion artifact suppression represent a significant advancement in wearable biomedical signal acquisition technology. It combines multi-sensor integration, hardware-level signal processing, and adaptive artifact suppression within a unified structural framework. The invention ensures the acquisition of high-quality physiological signals under real-world conditions involving movement and environmental influences.

[0076] The present invention relates to systems for acquiring and processing biomedical signals, in particular a system and a portable device for the intelligent removal of motion-induced artifacts in physiological signals. The invention specifically relates to hardware-implemented arrangements for synchronized data acquisition with multiple sensors, signal conditioning, adaptive signal decomposition, correlation-based artifact identification, and real-time reconstruction of artifact-cleaned signals in portable biomedical monitoring environments. REFERENCES 100 A quantum-assisted system for detecting fake news. 102 Housing structure 104 Biosignal detection unit 106 Inertial sensor unit 108 analog conditioning units 110 digitization units 112 storage units 114 computing unit 116 Communication unit

Claims

A system for the intelligent removal of motion artifacts in wearable biomedical signal acquisition systems, comprising: a housing configured to be worn on the body surface of a person and enclosing a variety of functionally interconnected components; a biosignal acquisition unit arranged within the housing structure and configured to acquire physiological signals from the subject via at least one sensor interface; an inertial sensor unit comprising at least one accelerometer and at least one gyroscope, wherein the inertial sensor unit is configured to generate motion-related signals corresponding to the movement of the housing structure;an analog processing unit that is operationally coupled to the biosignal acquisition unit and configured to amplify, impedance-match, and filter the acquired physiological signals using programmable gain circuits and anti-aliasing filters; a digitization unit that is operationally connected to the analog processing unit and the inertial sensor unit, wherein the digitization unit comprises at least one analog-to-digital converter configured to convert the physiological signals and the motion-related signals into synchronized digital representations; a storage unit configured to store digitized signal data, calibration parameters, and adaptive coefficients;a computing unit comprising at least one processor operationally linked to the digitization unit and the storage unit, wherein the computing unit is configured to temporally align the physiological signals and the motion-related signals, decompose the physiological signals into multiple signal components, identify motion-correlated components based on their correlation with the motion-related signals, generate adaptive filter coefficients, and reconstruct an artifact-cleaned physiological signal; and a communication unit configured to transmit the reconstructed physiological signal to an external device. System according to claim 1, wherein the biosignal acquisition unit comprises a plurality of electrode interfaces or optical sensor elements arranged spatially to ensure stable contact with the body surface, and further comprises shielding conductors configured to reduce electromagnetic interference during signal acquisition. System according to claim 1, wherein the inertial sensor unit is physically arranged together with the biosignal acquisition unit within a predefined distance to ensure spatial correspondence between motion measurements and physiological signal disturbances, and wherein the inertial sensor unit is configured to sample motion data along three orthogonal axes. System according to claim 1, wherein the analog processing unit comprises a cascade of low-noise amplifiers and tunable filters, wherein the gain parameters can be dynamically adjusted based on signal amplitude changes detected by the computing unit. System according to claim 1, wherein the digitization unit comprises a plurality of high-resolution analog-to-digital converters configured to operate at a sampling rate sufficient to capture both the bandwidth of physiological signals and the dynamics of motion, and wherein the digitization unit comprises a synchronization circuit configured to assign a timestamp to each sampled signal. System according to claim 1, wherein the computing unit is configured to perform signal decomposition by means of hardware-implemented transform-based operations, wherein the physiological signals are decomposed into multiple frequency or scale components representing different signal characteristics. System according to claim 6, wherein the computing unit is further configured to compare each decomposed component with corresponding motion-related signals in order to determine correlation coefficients and to classify components that exceed a predefined correlation threshold as motion artifact components. System according to claim 7, wherein the computing unit is configured to generate adaptive filter coefficients by means of iterative optimization circuits based on error minimization between predicted artifact components and actual signal components. System according to claim 8, wherein the computing unit is configured to selectively attenuate or reconstruct the identified motion artifact components and recombine the remaining components to generate the artifact-suppressed physiological signal. System according to claim 1, wherein the computing unit further comprises a feedback circuit configured to continuously estimate the residual error between the reconstructed physiological signal and the expected signal characteristics and to update the filter coefficients stored in memory.

Images