EP4662651A1 - Dispositif et procédé de surveillance d'une personne - Google Patents
Dispositif et procédé de surveillance d'une personneInfo
- Publication number
- EP4662651A1 EP4662651A1 EP24703775.7A EP24703775A EP4662651A1 EP 4662651 A1 EP4662651 A1 EP 4662651A1 EP 24703775 A EP24703775 A EP 24703775A EP 4662651 A1 EP4662651 A1 EP 4662651A1
- Authority
- EP
- European Patent Office
- Prior art keywords
- person
- class
- during
- sounds
- interest
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G08—SIGNALLING
- G08B—SIGNALLING SYSTEMS, e.g. PERSONAL CALLING SYSTEMS; ORDER TELEGRAPHS; ALARM SYSTEMS
- G08B21/00—Alarms responsive to a single specified undesired or abnormal condition and not otherwise provided for
- G08B21/02—Alarms for ensuring the safety of persons
- G08B21/04—Alarms for ensuring the safety of persons responsive to non-activity, e.g. of elderly persons
- G08B21/0407—Alarms for ensuring the safety of persons responsive to non-activity, e.g. of elderly persons based on behaviour analysis
- G08B21/043—Alarms for ensuring the safety of persons responsive to non-activity, e.g. of elderly persons based on behaviour analysis detecting an emergency event, e.g. a fall
-
- G—PHYSICS
- G08—SIGNALLING
- G08B—SIGNALLING SYSTEMS, e.g. PERSONAL CALLING SYSTEMS; ORDER TELEGRAPHS; ALARM SYSTEMS
- G08B21/00—Alarms responsive to a single specified undesired or abnormal condition and not otherwise provided for
- G08B21/02—Alarms for ensuring the safety of persons
- G08B21/04—Alarms for ensuring the safety of persons responsive to non-activity, e.g. of elderly persons
- G08B21/0438—Sensor means for detecting
- G08B21/0446—Sensor means for detecting worn on the body to detect changes of posture, e.g. a fall, inclination, acceleration, gait
-
- G—PHYSICS
- G08—SIGNALLING
- G08B—SIGNALLING SYSTEMS, e.g. PERSONAL CALLING SYSTEMS; ORDER TELEGRAPHS; ALARM SYSTEMS
- G08B21/00—Alarms responsive to a single specified undesired or abnormal condition and not otherwise provided for
- G08B21/02—Alarms for ensuring the safety of persons
- G08B21/04—Alarms for ensuring the safety of persons responsive to non-activity, e.g. of elderly persons
- G08B21/0438—Sensor means for detecting
- G08B21/0469—Presence detectors to detect unsafe condition, e.g. infrared sensor, microphone
-
- G—PHYSICS
- G08—SIGNALLING
- G08B—SIGNALLING SYSTEMS, e.g. PERSONAL CALLING SYSTEMS; ORDER TELEGRAPHS; ALARM SYSTEMS
- G08B29/00—Checking or monitoring of signalling or alarm systems; Prevention or correction of operating errors, e.g. preventing unauthorised operation
- G08B29/18—Prevention or correction of operating errors
- G08B29/185—Signal analysis techniques for reducing or preventing false alarms or for enhancing the reliability of the system
- G08B29/186—Fuzzy logic; neural networks
-
- G—PHYSICS
- G08—SIGNALLING
- G08B—SIGNALLING SYSTEMS, e.g. PERSONAL CALLING SYSTEMS; ORDER TELEGRAPHS; ALARM SYSTEMS
- G08B31/00—Predictive alarm systems characterised by extrapolation or other computation using updated historic data
Definitions
- the technical field of the invention concerns the monitoring of a person, with the aim of anticipating the occurrence of a fall.
- the devices can detect a person's fall. These devices usually include motion sensors, such as accelerometers. Based on the accelerometer's measurements, the occurrence of a fall is detected. The devices can be configured to generate an automatic alert signal, for the attention of third parties responsible for monitoring the person or for the attention of emergency services.
- US2019/197861 describes a device configured to detect a fall of a person, based on a variation in the height at which a person is moving. The fall can be confirmed by a detection of a lying posture of the person or by sounds generated by the person.
- a first object of the invention is a method of monitoring a person, the method comprising:
- step b) comprises a frequency analysis of the sound recorded in each processing period, so as to form input data for the classification algorithm
- the classification algorithm assigns a score to each class, the score being representative of a probability of belonging to said class;
- the classification algorithm may be a supervised learning artificial intelligence algorithm.
- the classification algorithm may include a convolutional neural network coupled with a multilayer perceptron.
- step c) involves emitting the alert signal based on a number of sounds classified in the second class of interest.
- Step c) can implement a classification model, fed by the number of sounds classified in the first class of interest.
- a microphone configured to pick up sound generated by the person during a recording period
- a third object of the invention is a device for detecting a person falling, comprising:
- a sensor configured to detect movement or sound from a person
- the device being characterized in that the fall detection device is connected to a monitoring device (1) according to the second object of the invention, the detection unit being programmed so that the detection threshold is adjusted according to an alert signal emitted by the processing unit of the monitoring device.
- the sensor may be an actimetry sensor, in which case the signal resulting from the sensor is a signal representative of a movement of the person.
- the sensor may be a microphone, in which case the signal resulting from the sensor is a signal representative of a sound emitted by the person.
- Figure 1 shows schematically the main components of a device according to the invention.
- Figure 2 shows examples of recorded frequency spectra corresponding to sounds emitted by a person.
- FIG. 3 shows the main steps of implementing the invention.
- Figure 4 shows a use of a device according to the invention coupled with a device for detecting a person falling.
- FIG 1 shows an example of a monitoring device 1 according to the invention.
- the device comprises a microphone 10, arranged to detect sounds generated by a person.
- the microphone can be placed in a fixed position, for example in a room in which the person moves. It can for example be a bedroom.
- the sounds detected by the microphone are transmitted to a processing unit 11.
- the processing unit 11 is connected to the microphone 10 by a wireless or wired connection.
- the processing unit comprises a microprocessor, programmed to execute instructions so as to implement the method, the main steps of which are summarized in connection with Figure 3.
- the processing unit is connected to a memory 12 in which said instructions are stored.
- the inventors recorded sounds generated by different people during consecutive nights, and focused the analysis of these sounds on the days and nights preceding the occurrence of a fall. They found that there was a statistical relationship between the occurrence of falls, on the one hand, and, on the other hand, the observation, on the recordings made, of the appearance of sounds suggesting the presence of respiratory anomalies in the recorded person, or of the increase in the frequency of such sounds over time, or of changes over time in the person's breathing pattern.
- respiratory anomalies we mean breathing marked by an acceleration of its frequency, by variations in its regularity, by modifications of the temporal characteristics of inspiration and expiration, by modifications of their sonorities, by the appearance of particular sounds punctuated by breathing (for example rales, moans, whistling, etc.), by the appearance of sounds added to breathing (coughing, sniffing, blowing the nose, sighs, etc.).
- These anomalies can occur in isolation or in combination. They can be analyzed one by one as many distinct entities. They can also be interpreted in a global and qualitative way to define, with human listening, a feeling of shortness of breath (dyspnea), difficulty breathing, respiratory pain, or even respiratory distress.
- the inventors analyzed the sounds generated by different people during consecutive nights, and more particularly consecutive nights preceding the occurrence of a fall.
- Figure 2 shows a set of consecutive frequency spectra of a person with wheezing-type respiratory abnormalities.
- the y-axis corresponds to the frequency.
- the x-axis corresponds to a number assigned to each elementary window.
- This network comprised 5 convolutional layers.
- a description of the ResNet network is given in Chen F. et al «Assessing the effects of convolutional neural network architectural factors on model performance for remote sensing image classification: An in-depth investigation».
- the convolutional layers allow feature extraction from the frequency spectra forming the input data.
- the extracted features are processed by a multilayer perceptron.
- the output of the multilayer perceptron represents a probability of belonging to several classes.
- Each class is representative of sounds emitted by the person: sounds of domestic origin (e.g. noise of sheets in the bed, noise of footsteps, noise of water in the toilet%), sounds of physiological origin (coughing, snoring, wheezing, respiratory rales, ).
- the sounds corresponding to each processing period T of 9.6 seconds were manually annotated.
- the different sounds recorded, corresponding to a sequence of 9.6 seconds were manually categorized into the different available classes.
- the annotations were performed by different operators. For each class, and for each processing period T, the fraction of operators who classified each sound, corresponding to the sequence, in said class, constituted the target of the neural network. The number of operators was between 3 and 7.
- the neural network was trained using sounds recorded on 559 microphones placed in individual rooms in a nursing home for dependent elderly people.
- the neural network was tested on 141,139 annotated sequences that were not used during the neural network training.
- the correlation coefficient reached 0.79 +/- 0.02 in the "breathing" class, representative of sounds indicating abnormal breathing.
- the neural network based on a combination of convolution layers and a multilayer perceptron, makes it possible to classify a sequence of frequency spectra, representative of the activity or state of the person during a processing period T of 9.6 seconds.
- Sound recordings were available from 2,398 nights. These nights included: a set of 2,000 nights during which no falls occurred by the person being monitored; a set of 398 nights during which the person being monitored had fallen.
- the set of recordings associated with each night included recordings from a total of 5 consecutive nights.
- a training set comprising 80% of the nights
- a test set comprising the remaining 20% of nights
- the recording set was segmented into 9.6-second treatment periods T.
- frequency spectra S(At) were formed on 67-ms time windows offset from each other by 23 ms.
- the S(At) spectra acquired during each treatment period T form a sequence of spectra S, used as input data to the neural network described above.
- the score corresponding to all classes was noted.
- the scores corresponding to each class of all sequences formed from the same set of 5 nights were cumulated in order to form input data for a learning model.
- each input data of the learning model was a vector, each term of which was associated with a class, and was formed by the sum of the scores, of said class, resulting from the neural network, for all sequences forming the recording set.
- the classification model was trained to predict an output (1 if it falls, 0 if it does not) from the input data.
- the learning model selected was a gradient boosting learning model programmed in the open source XGboost (extreme Gradient Boosting) algorithm in the Python programming environment.
- the average purity gain associated with each class was evaluated.
- the purity gain reflects the influence of a class on the prediction of the result provided by the classification model. A relationship was thus studied between the cumulative score of each class, during the 4 nights preceding the night considered.
- the inventors noted that two classes stood out: a first class of interest, corresponding to the "breathing” class, and, to a lesser extent, a second class of interest, corresponding to the "sheet noises" class.
- the breathing class reflects the presence of respiratory anomalies, suggestive of "breathing difficulties", as previously indicated.
- the "sheet noises” class reflects probable agitation of the person during sleep.
- the emission, by the person being monitored, of nocturnal sounds, reflecting breathing difficulties and, to a lesser extent, sleep disturbance inducing noises from sheets, can be considered as announcing a risk of the person falling during the following night.
- This finding, resulting from the use of the classification model, is an important element of the invention. This makes it possible, on the basis of a simple sound recording, to anticipate the occurrence of a fall.
- a remarkable aspect is that the sounds used are not sounds directly correlated to the person's walking. By sound directly correlated to the person's walking, we mean a sound produced by the person's footsteps on the ground.
- the sounds used are sounds of respiratory origin or sounds generated by the person's motor activity while they are in bed.
- the device can generate an alert signal, for the attention of third parties in charge of monitoring the person. This involves increasing the degree of vigilance with regard to the person being monitored, so as to prevent a fall from occurring.
- FIG 3 summarizes the main steps of an example of implementation of the invention.
- the recording of the person is carried out during the night.
- the invention can be implemented to carry out daytime or nighttime recordings.
- Step 100 Recording sounds in the vicinity of a person being monitored. This may involve recording in a room occupied by the person, for example in their bedroom. The sounds are recorded over a recording period, for example over one night or several nights.
- Step 110 segmentation of the recorded sound into processing periods T, for example lasting 9.6 seconds.
- Step 120 frequency analysis of the sound recorded during each processing period. As previously described, the frequency analysis is carried out on elementary windows At, for example of 67 ms. We thus have as many frequency spectra S(At) as elementary windows. Step 120 can be implemented by the processing unit 11.
- Step 130 from the frequency analysis performed during each processing period, formation of an input data S of a supervised learning artificial intelligence algorithm.
- the latter may be a neural network, comprising an association of convolution layers and a multilayer perceptron, as previously described.
- the output of the algorithm is a score, associated with different classes, each class being representative of a type of sound produced by the person being monitored.
- a first class is representative of abnormal breathing.
- a second class is representative of a sheet noise, or more generally a noise linked to movements of the person in their bed.
- the input data S may be a set of frequency spectra established during step 120, with or without overlapping of the elementary windows At. Step 130 is implemented by the processing unit 11.
- Step 140 analysis of the output of the algorithm implemented in step 130.
- a first cumulative score Cl is established, corresponding to the cumulative score, in the “breathing” class, in several time periods. The time periods can extend over one or more successive nights.
- Step 140 may include a determination of a second cumulative score C2 in the “sheet noise” class, representative of the movement of the person being monitored in their bed.
- Step 150 Depending on the cumulative score, or each cumulative score, resulting from step 140, generation of an alert signal.
- the alert signal is generated depending on the value taken by the cumulative score, or by each cumulative score, resulting from step 140. For example, the alert signal is emitted when the cumulative score, or when each cumulative score, exceeds an alert threshold.
- Step 150 implements the classification model described above.
- the first cumulative score and the second cumulative score form the input data of the classification model.
- the output of the classification model is a scalar, preferably between 0 and 1, which represents a probability of the person falling.
- the emission of the alert signal is intended to alert a third party, in charge of the person being monitored, of an increased risk of falling during the following night.
- the third party may be a family contact, or care staff when the person being monitored is in a medicalized home, or remote monitoring staff, for example the staff of home monitoring companies.
- the monitoring device 1 previously described can be coupled with a device 2 for detecting a person falling.
- the coupling is illustrated in FIG. 4.
- This type of device is intended to be worn by a person. It usually comprises a sensor 20, for example an actimetry sensor, configured to measure a movement of the person.
- the actimetry sensor can for example comprise an accelerometer or a gyrometer or a magnetometer.
- the sensor 20 can be a sound sensor.
- all or part of the monitoring is based on a detection of sounds representative of a fall. This can be a detection of a noise produced by the fall itself, or a sound generated by the person who has fallen.
- the fall detection device 2 comprises a detection unit 21, programmed to detect a fall of the person, as a function of a characteristic of a signal resulting from the actimetry sensor.
- the detection unit 21 is configured to indicate a fall when the characteristic exceeds a detection threshold. For example, when the sensor is a accelerometer, a fall is detected when the measured acceleration, along the vertical, exceeds a certain threshold.
- the detection threshold may be a characteristic of the detected sound, for example an amplitude or a spectral power in a predetermined frequency band, said frequency band corresponding to the sound produced by a person falling.
- the sound may be due to the impact of the body on the ground, or to a sound produced by a person, for example a cry, during or following the fall.
- the monitoring device can advantageously be associated with the fall detection device.
- the detection threshold of the detection unit can be adjusted, so as to increase the sensitivity of fall detection, due to an increased risk of the monitored person falling.
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Business, Economics & Management (AREA)
- Emergency Management (AREA)
- General Health & Medical Sciences (AREA)
- Gerontology & Geriatric Medicine (AREA)
- Evolutionary Computation (AREA)
- Computing Systems (AREA)
- Artificial Intelligence (AREA)
- Fuzzy Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Computer Security & Cryptography (AREA)
- Automation & Control Theory (AREA)
- Social Psychology (AREA)
- Psychiatry (AREA)
- Psychology (AREA)
- Human Computer Interaction (AREA)
- Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)
- Alarm Systems (AREA)
Abstract
Description
Claims
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| FR2301193A FR3145639A1 (fr) | 2023-02-08 | 2023-02-08 | Dispositif et procédé de surveillance d’une personne. |
| PCT/EP2024/052957 WO2024165586A1 (fr) | 2023-02-08 | 2024-02-07 | Dispositif et procédé de surveillance d'une personne. |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| EP4662651A1 true EP4662651A1 (fr) | 2025-12-17 |
Family
ID=85937218
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| EP24703775.7A Pending EP4662651A1 (fr) | 2023-02-08 | 2024-02-07 | Dispositif et procédé de surveillance d'une personne |
Country Status (4)
| Country | Link |
|---|---|
| US (1) | US20260237288A1 (fr) |
| EP (1) | EP4662651A1 (fr) |
| FR (1) | FR3145639A1 (fr) |
| WO (1) | WO2024165586A1 (fr) |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US11158179B2 (en) * | 2017-07-27 | 2021-10-26 | NXT-ID, Inc. | Method and system to improve accuracy of fall detection using multi-sensor fusion |
| US11615687B2 (en) * | 2018-01-26 | 2023-03-28 | University Of Cincinnati | Automated identification and creation of personalized kinetic state models of an individual |
| KR20220053913A (ko) * | 2020-10-23 | 2022-05-02 | 주식회사 내프터 | 레이더 센서 및 포토센서를 이용한 낙상 예측 시스템 및 낙상 예측 방법 |
-
2023
- 2023-02-08 FR FR2301193A patent/FR3145639A1/fr active Pending
-
2024
- 2024-02-07 US US19/154,883 patent/US20260237288A1/en active Pending
- 2024-02-07 WO PCT/EP2024/052957 patent/WO2024165586A1/fr not_active Ceased
- 2024-02-07 EP EP24703775.7A patent/EP4662651A1/fr active Pending
Also Published As
| Publication number | Publication date |
|---|---|
| US20260237288A1 (en) | 2026-08-13 |
| WO2024165586A1 (fr) | 2024-08-15 |
| FR3145639A1 (fr) | 2024-08-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US12586453B2 (en) | System and method for monitoring life signs of a person | |
| EP2229665B1 (fr) | Procede et equipement de detection de situation critique d'un sujet | |
| US20190192110A1 (en) | Classifier ensemble for detection of abnormal heart sounds | |
| EP2122596B1 (fr) | Dispositif, systeme et procede de televigilance ambulatoire comportant un dispositif de debruitage du pouls | |
| US20140142460A1 (en) | Method for detecting potential falls and a fall detector | |
| FR2981561A1 (fr) | Procede de detection d'activite a capteur de mouvements, dispositif et programme d'ordinateur correspondants | |
| US20180040222A1 (en) | Activity monitor | |
| WO2017178205A1 (fr) | Procédé et système d'identification d'évènements respiratoires | |
| US20180064404A1 (en) | System and Method for Correcting Sleep Aberrations | |
| Peruzzi et al. | A novel methodology to remotely and early diagnose sleep bruxism by leveraging on audio signals and embedded machine learning | |
| CN111771240A (zh) | 频谱图用于监视饮食活动的系统和方法 | |
| FR3028343A1 (fr) | Procede de detection de chute d'un sujet humain et dispositif actimetrique correspondant | |
| CA3087705A1 (fr) | Systeme et procede de surveillance des signes vitaux d'une personne | |
| Liang et al. | Pre-impact alarm system for fall detection using MEMS sensors and HMM-based SVM classifier | |
| WO2024165586A1 (fr) | Dispositif et procédé de surveillance d'une personne. | |
| AU2022206692A1 (en) | Automated detection of head affecting impact events in data collected via instrumented mouthguard devices | |
| Härmä et al. | Bed exit prediction based on movement and posture data | |
| Khawandi et al. | Applying machine learning algorithm in fall detection monitoring system | |
| US20240206745A1 (en) | System and Method for Monitoring a Person for Signs of Sickness | |
| CN111048089A (zh) | 提高智能穿戴设备语音唤醒成功率的方法、电子设备、计算机可读存储介质 | |
| US20240237923A1 (en) | Mobility Analysis | |
| KR102819949B1 (ko) | 성대 신호 처리 방법 및 디바이스 | |
| NL2016471B1 (en) | Teeth grinding detection. | |
| FR2828317A1 (fr) | Procede autonome de detection de situations anormales encourues par des individus | |
| FR3059454A1 (fr) | Procede et dispositif de surveillance et de detection de chutes |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| STAA | Information on the status of an ep patent application or granted ep patent |
Free format text: STATUS: UNKNOWN |
|
| STAA | Information on the status of an ep patent application or granted ep patent |
Free format text: STATUS: THE INTERNATIONAL PUBLICATION HAS BEEN MADE |
|
| PUAI | Public reference made under article 153(3) epc to a published international application that has entered the european phase |
Free format text: ORIGINAL CODE: 0009012 |
|
| STAA | Information on the status of an ep patent application or granted ep patent |
Free format text: STATUS: REQUEST FOR EXAMINATION WAS MADE |
|
| 17P | Request for examination filed |
Effective date: 20250807 |
|
| AK | Designated contracting states |
Kind code of ref document: A1 Designated state(s): AL AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HR HU IE IS IT LI LT LU LV MC ME MK MT NL NO PL PT RO RS SE SI SK SM TR |
|
| DAV | Request for validation of the european patent (deleted) | ||
| DAX | Request for extension of the european patent (deleted) | ||
| RAP3 | Party data changed (applicant data changed or rights of an application transferred) |
Owner name: ORIKIO Owner name: SORBONNE UNIVERSITE Owner name: INSTITUT NATIONAL DE LA SANTE ET DE LA RECHERCHEMEDICALE Owner name: ASSISTANCE PUBLIQUE-HOPITAUX DE PARIS |