ES2967807T3 - Remote medical care system - Google Patents

Abstract

A method for operating a telehealth care system is described. The method comprises analyzing ultrasonic sound waves reflected by a client to identify, in real time, a pattern in the reflected ultrasonic sound waves corresponding to the movements of the client's chest, and in response to determine a deviation in the identified real-time pattern of an expected pattern, predicting an alert event based on the deviation. Even if an alert is predicted, an alert signal is issued. Suitably, a client can be monitored in a non-invasive and contactless manner, and events warranting an alert to a remote alarm center can be predicted at an early stage, thus facilitating faster dispatch of help to the client. (Automatic translation with Google Translate, without legal value)

Description

DESCRIPTION

Remote healthcare system

field of invention

The present disclosure relates to a remote medical care system and a method of operation thereof. In particular, the present disclosure relates to a telecare system that provides predictive care for a client through contactless detection of the client's current condition.

Background

Remote healthcare systems used in social care settings to socially monitor a client are well known. Typical systems include a variety of wearable sensors, i.e. contact sensors, that collect data about the customer and transmit that data to a remote monitoring site, such as a call center, via a central unit installed in the home. the client's. For example, one sensor could monitor the client's breathing rate, while another sensor monitors the heart rate. In the event of an emergency situation, such as a heart attack (which would likely activate both a breathing and heart rate monitor), an alarm is triggered in the call center which sends help to the appropriate customer. A telephone call is often established between the call center and the customer's mainframe (which has suitable audio inputs/outputs) so that the customer and the call center operator can converse; for example, so the operator can provide status updates and reassurance to the customer.

The alarm is activated in the call center when the sensor detects a signal that exceeds a predefined threshold indicating that an event serious enough for the customer has occurred or is occurring. A limitation of these systems is that events that do not generate a sensor signal above the threshold are not reported, even if it is an event that may be causing discomfort and distress to the client, requiring the client to manually request help (typically The portable device includes a manually operated alarm button).

Another limitation is that the sensors are only operable if the customer chooses to wear them, which may not always be the case, and the customer may even inadvertently forget to wear the sensor. It will be appreciated that an emergency situation is not detectable in such scenarios. Some systems monitor that the sensors are being used and notify the customer if they are not, but there is still a period of inoperability during that time in between.

Therefore, it is highly desirable to develop a remote healthcare system that overcomes these limitations.

In the state of the art, document US 2021/0169375 discusses a lung function analysis system that includes motion detection devices.

Summary

The exemplary embodiments have been provided with the aim of addressing at least some of the difficulties encountered in current remote healthcare systems, whether those difficulties have been specifically mentioned above or are otherwise apparent from the discussion here. In particular, the goal is to enable predictive care that can be used to alert telemedicine care professionals and the client about a change in current circumstances that could potentially be the precursor to more serious events, in order to detect a possible emergency in its early stages to seek appropriate care before such more serious events.

In addition, it seeks to eliminate the need for contact sensors to monitor breathing and heart rate. Instead, the example embodiments use a non-contact detection technique.

It is another goal to leverage machine learning to monitor sensor data and characterize and adapt over time the uniqueness of each user and therefore detect any deviation from the norm to more quickly enable predictive care and alert accordingly. necessary for the user and remote healthcare professionals.

The present invention is defined according to the independent claims. Additional features of the dependent claims and description herein would be appreciated.

In this way, the client can be monitored in a contactless, non-invasive way, and can even be directly informed of a possible emergency through an appropriate audible alert signal emitted for example by speakers inside the home. Thus, the customer can behave inside their home without worrying about whether they are wearing a suitable sensor or whether the sensor has sufficient battery charge.

Additionally, issuing an alert at an early stage of a potential emergency can save valuable time in sending help to the customer. Suitably, an alert signal issued to a remote alarm center may include data related to the alert event (e.g., pattern data, prediction data) so that an operator can make an informed decision about escalating the alert. alert, for example, to send help (i.e. allowing the operator to take an informed and human opinion on the automatic prediction).

The alert event may be any suitable event that may be desirable for a remote healthcare system operator - usually located at a remote monitoring center - to be aware of. For example, the predicted alert event may be one of difficulty breathing, heart attack, arrhythmia, and infection.

The pattern(s) that are identified (and, consequently, the expected patterns with which they are compared) may correspond to the client's respiratory rate and/or the client's heart rate. In particular, the pattern may be a waveform that is determined by measuring reflected ultrasonic sound waves received at a suitable sampling frequency. The sampling rate may be at least 100 Hz for measuring respiratory rate and at least 480 Hz for measuring heart rate. The waveform may be suitably analyzed using digital processing to extract the waveform from background noise using environmental noise recorded in real time from a plurality of audio input devices.

In one aspect of the invention there is provided a computer program comprising instructions that, when executed by a computer, cause the computer to substantially carry out at least part of the method described above.

In one aspect of the invention there is provided a remote medical care system arranged to implement the aforementioned techniques. A system of this type comprises a communications module, one or more audio outputs and inputs, and a calculation unit. The communications module is arranged to communicate with a remote alarm center. The audio outputs are arranged to emit ultrasonic sound waves, and the audio inputs are arranged to receive ultrasonic sound waves that are reflected from a client. The computing unit is arranged to, in real time, identify a pattern in the reflected ultrasonic sound waves that corresponds to the movements of the client's chest and, in response to determine that the pattern identified in real time deviates from an expected pattern , predict an alert event based on the deviation. The computing unit can then control either an audio output (including the possibility of multiple audio outputs), or the communications module, or even both an audio output and a communications module, to issue an alert signal.

It will be appreciated that said system can be implemented by a single device incorporating each of the main elements, each of the elements can be part of separate devices, or any logical combination thereof. For example, preferably the audio inputs and outputs are configured as one type of device, while the communications module and the computing unit are arranged together in a single device. In another example, the computing unit is provided as part of a server, therefore it may not even be hosted within the customer's home.

Brief description of the figures

For a better understanding of the present disclosure, reference will be made solely by way of example to the attached drawings, in which:

Figure 1 shows a general principle of using sound waves to measure the distance to an object;

Figure 2 shows the application of Figure 1 to measure the displacement of a client's chest;

Figure 3 illustrates an example waveform that arises when measuring chest movement;

Figure 4 shows digital processing of a received audio input to extract a waveform; Figure 5 shows examples of comparing a deviated waveform with an expected waveform; Figure 6 summarizes a method of operating a remote healthcare system;

Figure 7 shows a remote medical care system arranged in a client's home;

Figure 8 shows an example of a central unit of a remote healthcare system;

Figure 9 shows an example of an audio input/output unit of a remote healthcare system.

Detailed description

Contactless monitoring of a client

One solution to the problem of clients not using a monitor assigned to them, or overcoming the problems of clients not using or configuring devices appropriately for proper monitoring, is to move to a contactless client monitoring system . Furthermore, it is desirable that such a contactless system be operable with as little customer input or configuration as possible. Consider that typical clients of remote healthcare systems are not technology savvy, so it is desirable to reduce the need for client interaction with a monitoring device (e.g., sensor).

The approach to non-contact monitoring adopted in the present disclosure is to use sound waves (more specifically ultrasonic waves) to measure, in real time, movements in a client's chest to determine at least one of the client's respiratory rate and heart rate. customer.

Respiratory rate and heart rate provide a useful indication of a client's current status because changes in respiratory and/or heart rate may be correlated with various situations for which the client may require assistance. For example: a low respiratory rate may correspond to difficulty breathing, and therefore to panting movements; shallow breathing may correspond to an impending heart attack; an erratic heart rate may correspond to arrhythmia and indicate heart problems; an increase in the heart rate and pulse (especially when the client is at rest) may indicate that the client has an infection or potentially more serious condition; and of course, it is extremely helpful to know if a client's breathing or heart rate has stopped completely. Figure 1 shows the general principle of using sound waves to measure the position of an object. Here sound waves 10 are emitted (at a predetermined frequency) from an emitter 12 and reflected by an object 14. When used in a remote healthcare system, the object 14 would be the client at home, and can be provided more than one sender and receiver. The reflected waves 16 are captured by a receiver 18 (generally built into the same device as the emitter 12 for simplicity, as shown) and a time of flight from the emission of the emitted waves 10 and the reception of the reflected waves 16 can be used to determine the distance to the object 14.

Figure 2 shows the measurement of chest movements of client 1 (in a social care setting) based on the measurement of reflected sound waves; that is, the measurement of the displacement of the chest. Figure 2A shows the position of client 2's chest after inhaling. Figure 2B shows the position of client 2's chest after exhaling. Measuring the position of the client's chest while transitioning between inhalation and exhalation positions produces a pattern from which the client's respiratory rate can be determined (i.e., based on the time between inhalation and exhalation). More specifically, respiratory rate can be measured as a function of time to look for changes in respiratory rate.

The fluctuation in the position of chest 2, that is, the up and down movement of the chest, can be measured as long as it is sampled at a sufficient frequency; This is because the movement of a patient's chest is inherently continuous, however, analysis of reflected sound waves must be done unobtrusively. The Nyquist-Shannon theorem states that to accurately resolve a periodic signal, it must be sampled at a frequency at least twice as high as the highest frequency component of the signal waveform in question.

Resting breathing rates in humans are typically in the range of 12-25 breaths per minute 'rpm', with breathing rates under stress typically not exceeding 50 rpm. The Nyquist rate for sampling normal breathing is therefore around 50 Hertz 'Hz', and during stress around 100 Hz. Therefore, when sampling sound waves reflected from the client's chest at a frequency of at least 100 Hz, a pattern corresponding to the client's respiratory rate can be established. However, it will be appreciated that these sampling rates are minimal and correspond to idealized, repetitive and stable patterns. A client's breathing rate may not necessarily be repetitive and stable, and may fluctuate considerably, especially when in distress, so in general, the higher the sampling rate, the better.

Another consideration is that to properly resolve breast rise and fall, sound waves must have wavelengths of the same order (or shorter) as the rise and fall. For most adults, chest movements due to breathing are between 5 and 15 centimeters ('cm'). However, it will be appreciated that those with lung conditions or generally lower lung capacity will have shallower breaths and correspondingly shorter chest movements; possibly in the range of around 1 to 2 centimeters. Therefore, the sound waves should preferably have a frequency of at least 7 kHz (corresponding to chest movements of 5 cm), and even more preferably of at least 17 kHz to 34 kHz (corresponding to chest movements of 1 to 2 cm).

It will be appreciated, however, that the range of sound frequencies mentioned is within the audible spectrum for humans (approximately 20 Hz to 20,000 Hz). Monitoring the client using sound waves within this spectrum is strongly discouraged due to the constant hum that would be heard by client 1.

Instead, it is preferred that, for continuous operation within a customer's home, monitoring of the customer's chest movement be performed at ultrasonic frequencies (above 20 kHz), which are generally above the hearing range of humans. . To adequately ensure that the sound is above the range of human hearing, it is further preferred that monitoring be performed at 22 kHz or higher. Suitably, ultrasonic frequencies are well above Nyquist rates for sampling respiratory movement, and also allow measurement of chest movements as small as 1.5 cm (22 kHz). Advantageously, these ultrasonic frequencies can be achieved using common audio technology such as paper cone/magnet speakers, which also allows these speakers to be used for the output of normal audible frequencies. Similar to what was described above, measuring the movement of client 1's chest can provide information about the client's heart rate; again, as long as it is sampled and resolved appropriately. That is, deviations in the client's chest position due to his or her heartbeat will generate a pattern that can be analyzed to determine the client's heart rate.

Human heart rates typically range between 40 and 150 beats per minute (bpm) in adults and around 150 to 240 bpm in infants. The Nyquist rate for sampling the heart rate of an adult is approximately 300 Hz, and for infants it is up to approximately 480 Hz.

The typical range of chest motion due to heartbeat is between 4 millimeters 'mm' and 12 mm. The corresponding sound frequency range to resolve such motion is 28.5 kHz to 86 kHz. That is, using sound waves of at least 86 kHz it should be possible to detect chest movements due to heartbeats of only 4 mm. Frequencies above 28.5 kHz are already well above audible frequencies and are therefore suitable for use in a customer's home.

Therefore, it will be appreciated that monitoring a client using ultrasonic sound waves allows for simultaneous measurement of both respiratory rate and heart rate. Preferably, such monitoring should be carried out using sound waves of at least 28.5 kHz, and even more preferably at least 86 kHz. Frequencies closer to the lower end of the range (i.e. around 28.5 kHz) can be achieved advantageously using common audio equipment that can also emit audible frequency sounds, such as paper cone/magnet speakers. Frequencies at the higher end of the range (i.e. around 86 kHz) may require dedicated hardware, such as a piezoelectric transducer, although this approach is correspondingly more expensive.

Figure 3 is an idealized illustrative example of a pattern that could emerge from appropriately sampling a client's chest movement using ultrasonic waves. More specifically, Figure 3 illustrates a waveform 20 that could be expected by plotting the position (x) of the client's chest as a function of time t, based on measurements of the time of flight of ultrasonic waves. It will be appreciated that an offset of the waveform in the x-axis (i.e., distance from 0) can be used to approximate a general position of the client 1 from the sender/receiver 12/18. It will also be appreciated that other methods can be used to analyze patterns in reflected wave measurements without viewing and analyzing them specifically as a waveform.

The waveform 20 is a superposition of a first waveform 22 corresponding to the respiratory movement of the chest and a second waveform 24 corresponding to the movement of the heartbeat. By properly analyzing the superimposed waveform 20, details related to the first and second constituent waveforms 22 and 24 can be extracted, thereby determining the corresponding respiratory and heart rates. For example, in the first waveform 22 (corresponding to the largest sine wave in waveform 20), the peak-to-peak separation corresponds to a full exhalation followed by a full inhalation. Similarly, in the second waveform 24 (corresponding to the smaller sine wave following the larger one in waveform 20), the peak-to-peak separation corresponds to the time between heartbeats.

It will be appreciated that, in real world use, the superimposed waveform 20, corresponding to the chest deflections measured from the reflected ultrasonic waves 16, will be mixed with the ambient noise of the environment; for example, noise from a television, mobile phone and/or kitchen, to name just a few possible sources. Therefore, the chest movement waveform 20 will typically need to be extracted from this background noise (at least in all but the quietest environments). This is achieved by recording ambient noise using multiple audio inputs (e.g. a microphone array) and processing it digitally to remove unwanted noise or extract the first and second waveforms 22, 24. The audio input array could be the same as, or separate from, the audio inputs used to receive the reflected ultrasonic waves 16. Additionally, the audio input array may be formed as part of a single device unit (provided there is some spatial separation of the inputs inside the device), or it could be located throughout the home as part of separate devices communicating with each other.

Figure 4 shows an example of digital processing for noise cancellation. A first signal 26 is a signal received by an audio input that acts as a receiver 18 for ultrasonic sound waves. The first signal 26 comprises noise 28 and a waveform 30 corresponding to the chest movements of client 1 hidden within the noise 28. A second (noise) signal 32 is formed based on the audio recorded by one or more audio inputs. additional that is invested; That is, the second signal represents an inversion of the ambient noise signal. The two signals 28 and 32 are superimposed, resulting in the third signal 34, where it can be seen that waveform 30 can now be resolved on top of a low-level white noise signal 36. Predicting an alert event

Obtaining real-time information about a client's breathing and heart rate does not by itself provide useful information for predicting an alert event; that is, a customer-related situation that would justify activating an alarm in a call center.

The goal of predictive care is to trigger a warning signal before the warning event becomes a reality; for example, alerting an alarm center before a heart attack occurs, rather than during or after (although current techniques would also indicate such events). This way, valuable seconds can be saved when dispatching help to a customer (if needed), potentially saving lives. To determine if an alert event may occur, real-time chest movement data is compared to a normal movement expected for the client. Stated another way, the current chest movement data derived from the reflected ultrasonic waves corresponding to the client's chest movement is compared in real time with the expected chest movement data for the client. Deviations in real-time data compared to expected data are then used to predict whether an alert event is likely to occur.

Figure 5 shows examples of a deviation in a currently measured pattern from an expected pattern. More specifically, Figure 5 shows a comparison of the current and expected waveforms for a client. The various signals are measured over 1 minute (x-axis) with an amplitude corresponding to the measured range of chest movement. High amplitude signals correspond to chest movements due to breathing. Barely visible are the small amplitude signals that follow the large amplitude wave and correspond to chest movements due to the heart rate. The signals are separated on the y-axis with respect to an average number of beats per minute determined from the small amplitude signals. A reference waveform 38 represents an expected "normal" pattern for Client 1. As an example, this waveform corresponds to approximately 7.5 breaths per minute and an average heart rate of 83 bpm. Analysis of the deviation of other signals from this baseline can be used to predict the onset of an alert event.

A first waveform 40 corresponds to a first example state of client 1. Here client 1 is measured with only 5 breaths per minute and an average heart rate of 77 bpm. As the client's respiratory rate has decreased, this indicates that the client may be having difficulty breathing and is taking large breaths of air to compensate (corresponding to large chest width measurements). In such a case, it may also be appropriate to supplement the analysis with audio recording (for example, via the ultrasonic audio input, or other audio input that can be used for noise cancellation) to determine whether the client is indeed gasping for air (and possible suffocation).

A second waveform 42 corresponds to a second example state of client 1. Here client 1 is measured with a respiratory rate of 15 breaths per minute and an average heart rate of 86 beats per minute. Since the client's breathing rate and heart rate have increased moderately, this could indicate that Client 1 is fighting an infection.

In such cases, it may also be appropriate to combine pattern analysis with real-time absolute measurements of the client's position relative to the ultrasound receiver, to verify whether the client is straining while moving or is at rest. In other words, the client's locomotion can be analyzed. In this way false positives can be prevented. Measuring the client's absolute position may also be useful in other situations, not just for predicting infections.

A third waveform 44 corresponds to a third example state of client 1. Here client 1 is measured with a respiratory rate of 30 breaths per minute and an average heart rate of 92 beats per minute. The client's very shallow breathing compared to baseline 38 (indicated by the small amplitude signal), in addition to high respiratory and heart rate, may indicate a precursor to a heart attack.

It will be appreciated that these are simply illustrative examples of the current art, and not limiting. Other average respiration rates and heart rates may be determined to indicate these types of alert events (i.e., client situations). Additionally, other situations can be determined from pattern analysis. For example, an erratic heart rate waveform (such that it may not even be possible to accurately determine an average heart rate) could indicate that Client 1 is suffering from an arrhythmia. In another example, the lack of a measurable heart rate may indicate that the client's heart has stopped.

To determine what is normal for a client 1, for example what the base 38 waveform is for a particular client, some type of calibration is useful. Preferably, the reference standard is updated regularly; That is, a continuous readjustment of the baseline is carried out. This is because what is considered "normal" for a client 1 will change over time due to age, illness, physical condition, etc.

In one example, data from a measurement window at some time in the past (a subsequent calibration window) can be used to determine the reference standard. An example measurement window is a couple of days centered one week in the past; although it will be appreciated that other values can also be chosen. A subsequent calibration window provides a profile for client 1, assuming that this window is normal for the client. In this way, the new referral is made automatically, without any specific involvement of the client. Accordingly, data corresponding to alert events can be marked so that such data can be easily recognized and ignored if it subsequently falls within the subsequent calibration window.

In another example, calibration is performed by measuring the movement of the client's chest for a specific time and/or while performing a specific task. Such an approach is not generally preferred, however, since requiring a customer to "do" something for the calibration to take effect is unlikely to be done regularly, even if done correctly.

Preferably, analysis of the chest movement pattern is performed in real time in comparison with an expected pattern using a suitable Machine Learning 'ML' algorithm implemented on a suitable computer (or network of computing devices). Compared to normal software algorithms, machine learning (ML) provides significant advantages to quickly classify certain waveforms as related to a certain condition. More specifically, a classification-type machine learning algorithm can be used to classify a particular waveform (which is a deviation from the reference waveform) as being produced by a particular condition affecting the client (i.e., which would give lead to an alert event).

Along the same lines, a machine learning algorithm may be able to more easily identify certain patterns as related to a specific activity that does not correspond to an alert event. That is, in addition to classifying waveforms to predict alert events, the ML algorithm can classify waveforms that do not correspond to an alert event. Consider that the client's breathing and heart rate are likely to increase by measurable degrees when performing certain 'normal' tasks such as exercising, getting up and sitting down, and moving around the home. Additionally, events that are classified as non-alert events can be flagged and excluded from calibration if they fall within the calibration window, so 'normal' chest movement for a client can be better established (i.e. excluding data). abnormalities corresponding to the activity).

It will be appreciated, of course, that normal (non-machine learning) software algorithms could also be used, implemented on a suitable computer architecture.

In response to the determination that there is a deviation from the real-time chest movement pattern, and the correlation of that deviant pattern with an alert event (i.e., the prediction of a deteriorated condition of client 1), a warning is issued. warning sign. The alert signal may be broadcast to client 1, to an alarm center (to be received by an operator there), and/or to both.

In one example, an alert signal issued to the client may take the form of a voice communication, for example, using a suitable audio output device provided within the client's home (which may even be the ultrasonic emitter operating in audible frequency ranges). This could be specific information about the type of event being detected, or a more general instruction to contact the alarm/call center. In another example, the alert to the client may be a signal transmitted to a portable social care device (such as an alarm trigger) that causes a light to come on, periodically flash, etc., to inform the client that it has been detected. abnormal breathing or heart rate; The customer can be trained to recognize that the light means that an abnormal state has been detected and contact the alarm/call center accordingly.

In an example, an alert signal issued to the alarm center may be an alarm notification that includes information about the client and the alert event being predicted, whereby an operator is informed that a potentially dangerous condition has been detected. customer deterioration 1. In addition to the alarm notification, the alert signal may include data related to the predicted alert event; For example, the alert signal may include data about the measured waveform that gave rise to the prediction, a sound recording captured by audio inputs in the customer's home, and/or video data captured by a camera device inside the customer's home. of the client's home. In this way, the alarm center operator can be provided with additional information on which to base an appropriate course of action.

For example, in response to receiving the alert signal, an operator at the alarm center may contact the customer to determine whether the customer is, for example, in pain, distressed, feeling unwell, or otherwise impaired. The operator can also send help to the customer, such as on-site caregivers, lifeguards, paramedics, etc., either before or after contacting the customer (or even without contacting the customer at all). Sending help immediately may be appropriate in situations where a severe alert event is being predicted, so it is vital that help reaches the customer without delay.

It will be appreciated that the notification of the alert signal can be managed using suitable software and computing and communication modules.

Figure 6 summarizes a method of operating a remote healthcare system consistent with the techniques disclosed above. In step 601, ultrasonic waves are emitted from a suitable output unit within the customer's home. In step 602, ultrasonic waves that have been reflected by a client and received by a suitable input unit are analyzed in real time to identify a pattern corresponding to the movements of the client's chest. In step 603, a deviation of the identified real-time pattern from an expected pattern is determined. In step 604, an alert event is predicted based on the deviant pattern. At step 605, an appropriate alert signal is issued.

Example Remote Healthcare System

Figure 7 shows an example of a remote healthcare facility (or system) 100, arranged within a client's home 1. Figure 8 illustrates an example central unit 110 compatible with system 100. Figure 9 illustrates an 'IO' audio input/output unit 120 compatible with system 100.

The system 100 comprises a communications module 112 arranged to communicate with a remote alarm center (not shown). The communications module 112 is configured to establish communication with the alarm center using a suitable communication network; for example, and without limitation, a public switched telephone network 'PSTN', Voice over Internet Protocol 'VoIP' and/or a cellular telecommunications network.

Here the communications module 112 is configured as part of the central unit 110. The central unit 110 may be configured to receive an alarm signal from a social alarm activation device that can be worn by a client 2. As will be familiar For those in the art, the social alarm activation device can transmit a radio frequency signal to the central unit 110 in response to an event such as a customer falling, and in turn, the central unit 110 can activate an alarm in the alarm center.

The system 100 also comprises one or more audio inputs 122 (i.e., ultrasonic receivers 18) and one or more audio outputs 124 (i.e., ultrasonic emitters 12); for example, one or more speakers (outputs) and one or more microphones (inputs). Here the audio inputs 122 and outputs 124 are configured as part of one or more audio input/output devices 120 (two of these devices are shown in Figure 7). The audio input/output devices 120 are preferably distributed throughout the home 1, that is, they are remote from the central unit 110, to provide audio coverage throughout the home, especially when it has separate and distinct rooms (as is usually be the case). Suitably, the audio input/output devices 120 are arranged to communicate electronically with the central unit 110 (either wired or wirelessly) to facilitate data transfer between the devices.

In another example, audio input(s) 122 and output(s) 124 may be provided as part of the central unit 110, either in addition to or instead of the audio input/output devices 120. In another example, audio inputs and outputs can be implemented as separate devices; This type of arrangement may be beneficial if you wish to move bulky speaker equipment to shelves at the edges of housing 1, but it is desirable to have microphones more centered within housing 1 (for example, on a coffee table).

The audio outputs 124 are configured to emit ultrasonic sound waves while the audio inputs 122 are configured to receive ultrasonic sound waves. More specifically, the audio inputs 122 are arranged to receive the ultrasonic sound 16 reflected from the client 2; that is, receiving and measuring ultrasonic sound at the frequencies emitted by the output(s) 124. In this example, the audio outputs can emit sound in a frequency range between 22 kHz and 28.5 kHz. That is, a frequency range that is appropriately above audible frequencies but that is possible to produce using common paper cone/magnet loudspeakers, and that may be suitable for detecting both the client's respiratory rate and heart rate 1 .

The system 100 comprises a calculation unit 114 arranged to identify a pattern in the reflected ultrasonic sound waves that correspond to the chest movements of the client of client 2. Preferably, the pattern is identified in substantially real time (i.e., with the shortest possible delay between receiving the reflected ultrasonic waves and analyzing the corresponding data to identify a pattern).

The calculation unit 114 is configured to determine whether the pattern identified in real time deviates from an expected movement pattern of the client's chest. If so, the calculation unit 114 is arranged to predict an alert event based on the detected deviation and control the communications module 112 to issue an alert signal.

In other words, the computing unit 114 is arranged to perform the contactless customer monitoring techniques described above.

In a preferred example, the computing unit 114 is configured as part of the central unit 110. In this way, connection to the communications module 112 is achieved more easily, and could even be configured to be controlled by the same processor. Beneficially, a single device can both analyze the ultrasonic data and issue an alert signal to the remote alarm center, which reduces potential points of failure in the system (thus providing robustness) and beneficially ensures that customer data does not have to be be transferred outside the device (except when optionally included as part of an alarm signal). The audio input/output devices 120 may send data corresponding to the received ultrasonic sound waves through a built-in communications module 126 to the central unit 110 through its communications module 112.

In another example, each of the audio input/output devices 120 may be provided with a suitably configured computing unit 128 (i.e., a processor) to analyze ultrasonic waves specifically received by an audio input 122 associated with that particular audio input/output device 120 (and also to control the ultrasonic output through output 124). The audio input/output devices may communicate the processed data to the central unit 110, or send a notification to the central unit 110 indicating that it should issue an alert.

In another example, the computing unit 114 may be provided as part of a remote server that is configured to receive and process ultrasonic wave measurements from the audio inputs of the system 122. The server may receive data directly from the audio inputs. if the input/output devices 120 have been provided with a suitable network connection (e.g., internet), or can receive data from the central unit 110 (e.g., via communications module 112) if the input data Audio signals are first transferred to the central unit 110 from the input/output devices 120.

The warning signal can take various forms. In one example, the alert signal may be an audio alert emitted through the same audio outputs 124 that act as ultrasonic emitters. For example, the alert signal may be a recorded voice warning that is broadcast through the input/output devices 120 and is intended to alert the client 2 about a possible medical situation. In another example, the alert signal may be an alarm notification issued to a remote alarm center operator via communications module 112.

In summary, examples of improved embodiments of a remote healthcare system that contactlessly monitors the well-being of a client have been described. In particular, exemplary embodiments monitor heart rate and respiratory rate to identify events that warrant an alert to a remote alarm center, sometimes referred to as emergency events, such as a heart attack, at an early stage of the event, i.e. , before existing techniques for monitoring a customer register a problem and raise an alarm. Suitably, valuable seconds and even minutes can be saved by activating an alarm and sending help to a customer.

The system and its components can be manufactured industrially. An industrial application of the examples described will be evident from the discussion herein.

As will be appreciated by one skilled in the art, current techniques may be incorporated as a system, method or computer program product. Accordingly, aspects of the techniques described herein may take the form of an all-hardware embodiment, an all-software embodiment, or an embodiment that combines software and hardware aspects.

Furthermore, current techniques may take the form of a computer program product incorporated into a computer-readable medium having computer-readable program code incorporated therein. The computer-readable medium may be a computer-readable signal medium or a computer-readable storage medium. A computer-readable medium may be, for example, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared or semiconductor system, apparatus or device, or any suitable combination of the above.

The computer program code for carrying out the operations of current techniques may be written in any combination of one or more programming languages, including object-oriented programming languages and conventional procedural programming languages. Code components may be embodied as procedures, methods, or the like, and may include subcomponents that may take the form of instructions or sequences of instructions at any level of abstraction, from the direct machine instructions of a native instruction set to constructs. high-level compiled or interpreted language.

Embodiments of the current techniques also provide a non-transitory data carrier that carries code that, when implemented in a processor, causes the processor to perform any of the methods described herein.

The techniques further provide processor control code to implement the methods described above, for example, in a general purpose computer system or in a digital signal processor (DSP). The techniques also provide a carrier carrying processor control code to, when executed, implement any of the above methods, in particular on a non-transitory data carrier. The code may be provided on a medium such as a disk, a microprocessor, a CD or DVD-ROM, programmed memory such as non-volatile memory (for example, Flash) or read-only memory (microprogram), or on a medium data as an optical or electrical signal carrier. The code (and/or data) to implement the embodiments of the techniques described herein may include source, object or executable code in a conventional (interpreted or compiled) programming language such as Python, C or assembly code, code for configure or control an ASIC (Application Specific Integrated Circuit) or FPGA (Field Programmable Gate Array), or code for a hardware description language such as Verilog (RTM) or VHDL (High Speed Integrated Circuit Hardware Description Language Speed). As the expert will appreciate, said code and/or data can be distributed among a plurality of components coupled in communication with each other. The techniques may comprise a controller that includes a microprocessor, working memory, and program memory coupled to one or more of the system components.

It will also be clear to a person skilled in the art that all or part of a logical method according to current art embodiments can be suitably incorporated into a logical apparatus comprising logical elements to carry out the steps of the methods described above, and that said logic elements may include components such as logic gates in, for example, a programmable logic array or an application-specific integrated circuit. Such a logical arrangement may further be incorporated into enabling elements to temporarily or permanently establish logical structures in such an assembly or circuit using, for example, a virtual hardware description language, which may be stored and transmitted using fixed or transmissible carrier media.

In one embodiment, the current techniques can be embodied in the form of a data carrier containing functional data, said functional data comprising functional computer data structures that, when loaded into a computer system or a network and operated on, allow that said computer system performs all the steps of the methods described herein.

The methods described herein can be carried out in whole or in part on an apparatus, that is, an electronic device, using a machine learning or artificial intelligence model. A function associated with AI can be performed through non-volatile memory, volatile memory and the processor. The AI model can be processed by a dedicated AI processor designed on a hardware structure specified for processing AI models. The processor may include one or more processors. At this time, one or more processors may be a general-purpose processor, such as a central processing unit (CPU), an application processor (AP), or similarly, a graphics-only processing unit, such as a graphics processing unit (GPU), a visual processing unit (VPU), and/or a processor dedicated to artificial intelligence, such as a neural processing unit (NPU). The one or a plurality of processors control the processing of input data according to a predefined operating rule or an artificial intelligence (AI) model stored in the non-volatile memory and the volatile memory.

The predefined operating rule or artificial intelligence model is provided through training or learning. Here, being provided through learning means means that by applying a learning algorithm to a plurality of learning data, a predefined operating rule or AI model of a desired characteristic is created. Learning may be carried out on the device itself where AI, according to one embodiment, is carried out, and/or may be implemented through a separate server/system. The learning algorithm is a method of training a predetermined target device (for example, a robot) using a plurality of learning data to cause the target device to make a determination or prediction. Examples of learning algorithms include, but are not limited to, supervised learning, unsupervised learning, semi-supervised learning, or reinforcement learning.

The artificial intelligence model may include a plurality of neural network layers. Each of the multiple layers of the neural network includes a plurality of weight values and performs neural network calculations by calculating between the calculation result of a previous layer and the plurality of weight values. Examples of neural networks include, but are not limited to, convolutional neural network (CNN), deep neural network (DNN), recurrent neural network (RNN), restricted Boltzmann machine (RBM), deep belief network (DBN), Bidirectional Recurrent Deep Neural Network (BRDNN), Generative Adversarial Networks (GAN), and Deep Q-Nets.

Although preferred embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that changes can be made without departing from the scope of the invention as defined in the claims.

Claims (14)

1. A method of operating a remote medical care system (100), comprising: analyze ultrasonic sound waves (16) reflected from a client (1) to identify, in real time, a pattern in the reflected ultrasonic sound waves (16) that corresponds to the chest movements of the client (1), and in response to determining a deviation in the identified pattern in real time (40, 42, 44) with respect to an expected pattern (38), predict an alert event based on the deviation and issue an alert signal, wherein the prediction of the alert event is further based on environmental noise recorded substantially simultaneously with the reflected ultrasonic sound waves (16) being analyzed. 2. The method of claim 1, wherein the real-time identified pattern (40, 42, 44) and the expected pattern (38) correspond to chest movements arising from the client's respiratory rate (1). 3. The method of claim 1 or 2, wherein the identified real-time pattern (40, 42, 44) and the expected pattern (38) correspond to chest movements arising from the client's heart rate (1) . 4. The method of any preceding claim, wherein the pattern is a waveform. 5. The method of claim 4, wherein the analysis of ultrasonic sound waves includes digital processing to extract the background noise waveform using environmental noise recorded in real time from a plurality of audio input devices. . 6. The method of any preceding claim, wherein predicting an alert event is further based on measuring the customer's locomotion (1). 7. The method of any preceding claim, wherein the expected pattern (38) is determined from historical data of ultrasonic sound waves (16) reflected from the client (1). 8. The method of any preceding claim, wherein predicting an alert event based on the deviation includes checking whether the deviation corresponds to a non-alert event, such as a normal activity performed by the client (1). 9. The method of any preceding claim, wherein the predicted alert event includes at least one of difficulty breathing, heart attack, arrhythmia and infection. 10. The method of any preceding claim, wherein issuing the alert signal includes issuing an audible warning to the customer (1). 11. The method of any preceding claim, wherein issuing the alert signal includes transmitting an alarm notification to an alarm center along with data corresponding to the identified pattern on which the predicted alert event is based. 12. The method of claim 11 further comprises, in response to receiving the alarm notification from the alarm center, sending help to the client (1). 13. A computer program comprising instructions that, when executed on a computer, cause the computer to carry out the method of any of claims 1 to 11. 14. A remote medical care system (100), which includes: a communications module (112) arranged to communicate with a remote alarm center. one or more audio outputs (120) arranged to emit ultrasonic sound waves (10), one or more audio inputs (120) arranged to receive ultrasonic sound waves (16) reflected from a client (1), and one or more more audio inputs arranged to record ambient audio substantially simultaneously with the reflected ultrasonic sound waves (16), and a calculation unit (114) arranged to, in real time, identify a pattern in the reflected ultrasonic sound waves (16) corresponding to the movements of the client's chest (1) and, in response to determine that the identified pattern in time real (40, 42, 44), deviate an expected pattern (38), predict an alert event based on the deviation and environmental noise recorded substantially simultaneously with the reflected ultrasonic sound waves (16), and control at least one of an audio output and the communications module (112) to emit an alert signal.