# US 2019/0167186 A1 · Method and System for Identifying Respiratory Events

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2019/0167186 A1
- Document ID: US20190167186A1
- Status: Published application
- Kind: A1
- Application number: 16/092,585
- Filed: 2017-03-23
- Published: 2019-06-06
- Earliest claimed: 2016-04-15
- Assignees: Clebre SpóLka Z O.O.
- Applicants: CLEBRE Spólka z o.o.
- Inventors: Marcel Mlynczak; Maciej Migacz; Wojciech Kukwa
- Web page: https://patentgrep.com/patent/US20190167186A1

## CPC

- A61B 5/4818: {Sleep apnoea}
- A61B 5/0004: {characterised by the type of physiological signal transmitted}
- A61B 5/0823: {Detecting or evaluating cough events}
- A61B 5/0826: {Detecting or evaluating apnoea events (sleep apnoea A61B5/4818)}
- A61B 5/113: occurring during breathing
- A61B 5/14551: {for measuring blood gases}
- A61B 5/7264: {Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems}
- A61B 5/7282: {Event detection, e.g. detecting unique waveforms indicative of a medical condition (cough events A61B5/0823; seizures A61B5/4094; sleep apnoea A61B5/4818)}
- A61B 5/7455: {characterised by tactile indication, e.g. vibration or electrical stimulation}
- A61B 7/003: {Detecting lung or respiration noise}
- A61B 2562/0204: Acoustic sensors
- G16H 50/20: for computer-aided diagnosis, e.g. based on medical expert systems
- A61B: DIAGNOSIS; SURGERY; IDENTIFICATION (usefulness limited to only animals A61D)
- G16H: HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA

## Abstract

The present disclosure relates to a method and a system for examining respiratory disorders whereby signals coming from the examined person are recorded by means of a wireless sensor equipped with a microphone and an accelerometer and then sent to a monitoring station. The monitoring station receives a digital data stream from the wireless sensor, cuts out respiratory episodes from the signal and, using a classification assembly constructed from three independent detection modules, classifies a respiratory episode as being normal or as snoring as well as determines the occurrence of apnoea.

## Claims

### Claim at position 1

1-15. (canceled)

### Claim 16

A method for identifying respiratory events whereby,
a sound signal and a motion signal, which are generated during respiration, are recorded by a wireless sensor equipped with a microphone sensor and a motion sensor;
signals from sensors are converted into a digital data stream by a microcontroller;
the digital data stream is sent to a monitoring station by a wireless transmission module;
wherein, the digital data stream from the wireless sensor is received in the monitoring station and then digital data representing sound and motion signals are pre-filtered;
in a segmentation module, the digital data representing a sound signal are divided into time windows and transformed to a frequency domain and then the sound signal is divided into segments corresponding to respiratory episodes on the basis of signal changes in the frequency domain in the time windows of the sound signal for specific sound signal frequencies;
input vectors are created containing sound signal parameters in a time domain and in the frequency domain, statistical parameters specified on the basis of historical data as well as motion signals parameters in the time domain, and then the input vector containing sound signal parameters is fed into inputs of an assembly of at least three independent and different detection modules which have been designed to generate a signal classifying a respiratory event on the basis of the sound signal,
the vector containing motion signal parameters is fed into an input of a motion signals classification module which has been designed to generate a motion/position classification signal;
data obtained at an output of the respiratory event of the classification module and the motion signals classification module are fed into an inference module at the output of which a respiratory event identification signal is output.

### Claim 17 (depends on 16)

The method of claim 16 wherein the respiratory event identification signal is generated in an independent detection module which is a multi-layer neural network whose weights have been set in such a manner that a signal, which differentiates respiratory disorders from normal respiration, is generated at the output of the multi-layer neural network and contains a relative confidence factor of identification of respiratory disorders whereby in each detection module, the weights of neural networks have been selected independently from the weights of neural networks in other detection modules.

### Claim 18 (depends on 16)

The method of claim 16 wherein a respiratory disorders signal is generated correspondingly to a reading of the detection module which generates an output signal with the highest confidence factor.

### Claim 19 (depends on 16)

The method of claim 16 wherein an input vector is created in which statistical data refers to the population.

### Claim 20 (depends on 16)

The method of claim 16 wherein an input vector is created in which statistical data refers to the historical data of an examined individual.

### Claim 21 (depends on 16)

The method of claim 16 wherein the wireless sensor is equipped with a vibratory signaling device.

### Claim 22 (depends on 16)

The method of claim 16 wherein the wireless sensor is equipped with a reflectance-based pulse oximeter.

### Claim 23

A system for identifying respiratory events during examination constructed from
a wireless sensor comprising:
a microphone and a motion sensor, which record sound signals and motion signals generated during respiration,
a microcontroller, which converts signals from the sensors into a digital data stream, and
a wireless transmission module;
a monitoring station comprising,
a wireless transmission module,
a signals pre-processing module, which has been designed to pre-filter the data stream from the wireless transmission module, divided into time windows, and transforms both the sound signal and the motion signal into a frequency domain for subsequent time windows;
wherein the monitoring station further comprises,
a segmentation module, which has been designed to divide the sound signal and the motion signal into segments corresponding to respiratory episodes on the basis of sound signal changes in the frequency domain as well as into sound signal and motion signal time windows;
a transformation module, which has been designed to create an input vector containing sound signal parameters in the time and frequency domains, historical data and statistical parameters;
a classification module, consisting of at least three independent and different detection modules which have been designed to generate a signal classifying a respiratory event on the basis of the input vector and a motion signal classification module, which has been designed to generate a position classification signal; and
an inference module, which has been designed in such a manner that a position identification output signal is output at its output.

### Claim 24 (depends on 23)

The system of claim 23 wherein the independent detection module is a multi-layer neural network whose weights have been set in such a manner that a signal, which differentiates respiratory disorders from normal respiration, is generated at the output of the neural network and contains a relative confidence factor of identification of respiratory disorders, whereby in each detection module the weights of neural networks have been selected independently from the weights of neural networks in other detection modules.

### Claim 25 (depends on 23)

The system of claim 23 wherein each detection module has a different set of weights of neural networks.

### Claim 26 (depends on 24)

The system of claim 24 wherein the inference module is adapted to generate a respiratory disorders signal based on the detection module with the highest confidence factor of the output.

### Claim 27 (depends on 24)

The system of claim 24 wherein the motion sensor is an accelerometer/gyroscope or a combination of these two sensors.

### Claim 28 (depends on 24)

The system of claim 24 wherein the sound signal is obtained from a microphone sensor or a signal coming from the motion sensor.

### Claim 29 (depends on 24)

The system of claim 24 wherein the wireless sensor is equipped with a vibratory signaling device.

### Claim 30 (depends on 24)

The system of claim 24 wherein the wireless sensor is equipped with a reflectance-based pulse oximeter.