Imec patent portfolio for connected health solutions

EEG/ECG Motion Artifact Removal (US 14252611)

This patent discloses a method for improving biosignal quality by reducing contribution of artefacts. Although the patent discusses motion artefacts, the method can be generalized to data contaminated by other artefacts. The method consists of two stages. The first one classifies the artefacts based on the measured biosignal and auxiliary signals, and extracts meta information describing type and severity of the artefact. The second stage uses the output of the first stage to select the most appropriate artefact reduction method and its optimal parameters and applies it. The ability to evaluate the artefact reduction effectiveness is also part of the method, and it is used to optimize the performance of both stages.

Physiological Signal Normalization (US 14471158)

A device for monitoring the physical activity of a living being is disclosed. In one aspect, there is a data input module configured to receive information about the living being's heart beat rate value, motion intensity and anthropometric characteristics. Further, there is an activity recognition and storage module configured to detect, from information received about the living being's motion intensity, the living being's activity and to store information about the living being's heart beat rate value and the motion intensity associated with that detected activity. Further, there is a heartbeat rate analysis module configured to determine, from a plurality of heart beat rate values associated with each detected activity, statistics of the distribution of heart beat rate values for each activity or a subset of activities. Further, there is a fitness estimation module configured to calculate, using the information from the heart beat rate analysis module and the anthropometric characteristics, a cardiorespiratory fitness level of the living being.

Blood pressure Measurement (US 14969300)

This patent describes a method to estimate systolic, diastolic and continuous blood pressure using photoplethysmogram (PPG) and accelerometer data.  Blood pressures are estimated based on the raw PPG signal, the accelerometer data and PPG features in combination with anthropometric characteristics, Optionally, the electrocardiogram signal is added for the extraction of the features.  The systolic and diastolic blood pressure are estimated using (a combination) of different machine learning techniques, including multiple linear regression, random forest regression model, a Bayesian model and or a general machine learning regression model. The Hilbert-Huang transform is used to calculate continuous mean blood pressure.

Personalized Stress level Estimation (US 15468645)

This invention discloses an system and methodology for determining a subject's stress condition. It consists of a multi-component system for reading physiological data during relaxation and stress task, extracting normalization values per subject, extracting relevant features for determining stress models, building/updating/storing several stress models linked to a subject and based on normalization parameters and features to predict an individual's stress level. The system include both a storage space and a reinforcement learning module able to update normalization parameters, features and models. The reinforcement learning module include also one feedback system that allow the subject to judge the quality of the estimated stress value and a system that request to perform a predefined stress task at convenient time. Both this component are used to facilitating the stress models, features and normalization parameters update. 

Wideband bio-impedance measurement (US 15247172)

This patent describes an impedance spectroscopy system for bio-impedance measurement. It includes a signal generator configured to generate a signal with a broadband frequency spectrum and to generate a stimulation current from the signal. The impedance spectroscopy system also includes an amplifier configured to measure a voltage response with respect to the stimulation current, and to simultaneously measure other biopotential signals. Further, it includes a processor configured to analyze the voltage response to derive a bio-impedance spectrum as well to derive further information from the biopotential signals.

EOG and camera-based eye tracking (US 15984537)

Image based methods are considered the most accurate non-contact methods for eye tracking, leading to an absolute position of the pupil within the eye, but they are resource hungry and hence throughput limited. On the other hand biopotential based methods (based on electro-oculography or EOG) provide a much faster update rate, but because they are differential (relative), suffer from drifts of the signal over time. This patent proposes a method to combine image-based eye tracking with EOG acquisition, to provide a fast and accurate eye tracking, benefitting from the positive aspects of both methods.

Voltage reference generator (US 15890336)

This patent discloses an integrated circuit technique for generating programmable voltage references with nA power level. To do so, conventional resistive ladder is not practical due to large Silicon area that this approach would require at such low powers. Switched capacitor is a known alternative. However, the switching frequency is normally chosen as a fixed value, despite the variability of temperature and fabrication process. This leads to excessive power consumption, because the switching frequency needs to be chosen large enough to let the system operate correctly in all conditions. To solve this problem, the disclosed technique consists in monitoring the ripple of the voltage references and controlling dynamically the switching frequency in order to keep the ripple under a certain threshold. In this way, the frequency will vary with temperature and process, maintaining the voltage reference ripple constant. This technique can be applied to all integrated circuits operating with very limited power budget, or having a low-power mode in which one or multiple voltage references need to be generated.

Radar-based Multi-parameter Characterization Methodology (US 16026465)

The proposed idea aims at creating a smart sensing electromagnetic environment enabling multi-parameter characterization (vital signs, tag-less in-door localization, fall detection, etc.) under realistic conditions, allowing therefore the operations of multi-targets tracking (detection of their speeds and locations) and vital signs monitoring.

Random Body Movement Rejection Methodology for Radar-based Vital Signs Monitoring (US 16023352)
The proposed methodology allows random body movement rejection in radar-based vital signs monitoring application. Therefore, instead of simply discarding windows of signal containing the artifact(s), this methodology tries to attenuate the artifact effect(s) in order to estimate the vital signs.

Stretchable ExG Patch (DE 202013011290.5)

The patent is related to a patch embodiment, more specifically a biopotential acquisition setup without traditional wires. The main claim made in this patent protects the idea of having an embodiment consisting of two (or more) electrodes, connected by an elastic/flexible/stretchable structure. The structure can be realized by straight, curved or meandering structures, consisting of garment or other conductive material.
It is filed as a Utility Model. It is filed only in Germany, jointly with TNO.

Total Body Water (TBW) Estimation (DE 20 2016 104065.)

This patent provides a system for estimating Total Body Water (TBW) content of a human user with more user friendliness and/or precision in estimation. TBW content is defined as a parameter reflecting the water content of a human body which is contained in the tissues, blood, bones, and elsewhere. It can be expressed as a relative value, e.g. a percentage of the total water content in relation to the body volume, or as a volume, e.g. 3.5 liters. The proposed system includes a body impedance measurement blocks which can be worn on a body part of the user. It contains a set of skin electrodes, a measurement circuit, and a processing device to estimate TBW based on body impedance and user specific parameters.

BP estimation from PAT in chair (NL 2019604)

This patent discloses a method for estimating pulse arrival time / pulse transit time, which can be used to estimate blood pressure, from a seat. This is done through the combination of at least one capacitive sensor, which measures the electrical activity of the subject’s heart, with at least one RF (radar) sensor which measures the passage of blood pulse waves through an artery in the body of the subject. In this way a fully unobtrusive system for estimation of blood pressure can be obtained.

Note that the patent list above is only a selection and only related to connected health solutions. If you want to know more about our patents, don’t hesitate to reach out.