System and method for monitoring and producing audio feedback based on respiratory movements

Abstract

A system and method for monitoring respiratory movements and generating real-time audio feedback reflects a user's breathing pattern. The system includes a respiratory sensor, a signal processor, an audio synthesis unit, a wireless transmitter, and an audio output device. Respiratory signals are captured, filtered, and analyzed to determine inhalation and exhalation phases. The audio synthesis unit produces corresponding sound signals that are dynamically modulated based on the user's breathing intensity and rhythm. The audio output provides a continuous, natural-sounding auditory representation of breathing, which can be used for caregiver reassurance, sleep monitoring, or relaxation purposes. Abnormal events such as breathing cessation or muscle spasms may trigger audio alerts or spoken messages. Processed physiological metrics, including respiratory rate, heart rate, and sleep state, being transmitted to a cloud platform for long-term analysis. The system may also interface with smart lighting or external alerting devices to support hearing-impaired caregivers.

Description

SYSTEM AND METHOD FOR MONITORING AND PRODUCING AUDIO FEEDBACK BASED ON RESPIRATORY MOVEMENTSCROSS-REFERENCE TO RELATED APPLICATIONS

[0001] This application claims priority to provisional application U.S. Serial No. 63 / 693,710, which was filed on September 12, 2024, which is pending, and which is hereby incorporated in its entirety for all purposes.

[0002] This application claims priority to provisional application U.S. Serial No. 63 / 784,928, which was filed on April 07, 2025, which is pending, and which is hereby incorporated in its entirety for all purposes.BACKGROUND OF THE INVENTIONFIELD OF THE INVENTION

[0003] The present invention relates to systems and methods for non-invasively monitoring respiratory movements and generating real-time audio feedback based on detected breathing patterns. It further includes mechanisms for abnormal event detection and alerting via sound or spoken messages.

[0004] Monitoring respiratory activity is essential for ensuring the health and safety of vulnerable individuals, such as newborns, patients with neurological disorders (e.g., epilepsy), or persons in assisted living environments.DISCUSSION OF THE RELATED ART

[0005] Several devices currently available on the market are capable of monitoring the respiratory movements of newborn babies and children and providing alerts in case of apnea or other abnormalities. However, they often lack the ability to produce continuous, natural-sounding audio feedback that mirrors real-time breathing. This kind of feedback could be highly reassuring for caregivers, such as parents monitoring a sleeping infant in another room. Existing systems also rarely do not provide intelligible spoken alerts describing the exact nature of a problem, relying instead on vague alarming-Page 1 of 26-4820PCT_spec_claims_abst_03b.docx 51790 7574 185sound signals or visual indicators that may go unnoticed, or often cause unnecessarily scary moments.

[0006] Furthermore, current technologies seldom include capabilities for detecting and reporting other critical conditions — such as tonic-clonic seizures — which involve abnormal, repetitive muscle movements that may go unnoticed without continuous monitoring. There is also a growing demand for solutions that can provide real-time physiological feedback in wellness- oriented applications, such as mindfulness training, guided breathing, or yoga practice. In parallel, cloud-based platforms and immersive AR / VR environments are emerging as promising interfaces for capturing, visualizing, and responding to respiratory data in both clinical and consumer contexts.

[0007] In neonatal settings, the ability to distinguish between deep and light sleep in real time enables medical staff to time interventions — such as blood draws or examinations — more sensitively, helping to avoid unnecessary disturbance and to support rest and recovery. Existing systems rarely support such fine-grained, continuous feedback or integrate seamlessly with smart environments, limiting their applicability beyond narrow diagnostic use.

[0008] Several commercial products exist for monitoring respiratory activity, particularly for infants and individuals with medical conditions. These systems typically utilize sensors such as piezoelectric films, strain gauges, or radar to detect breathing motion. Most are designed to detect and alert upon apnea or respiratory cessation events. However, such technologies tend to be reactive rather than proactive: they monitor for dangerous deviations and trigger an alarm only after a predefined threshold is crossed, lacking the ability to provide continuous or informative feedback in real time. They are not designed for home use, for healthy babies, just for parents' reassurance.

[0009] Moreover, current systems do not generate or synthesize audio that reflects the user’s actual breathing pattern. They offer no real-time auditory feedback that could help calm an anxious caregiver, support guided breathing, or enhance sleep and relaxation. Existing products are generally not capable of transmitting real-time respiratory data to external audio devices or immersive platforms such as AR / VR headsets. Nor are they optimized for-Page 2 of 26-4820PCT_spec_claims_abst_03b.docx 51790 7574 185ultra-low-power operation through efficient data transmission — for example, by sending only respiratory rate or waveform metadata instead of continuous audio streams.

[0010] Another widely used solution for infant monitoring is the baby monitor, often referred to as a “baby alarm.” These devices typically transmit ambient audio from the baby’s room to a parent unit located elsewhere in the home. However, during quiet sleep periods — when the infant is not moving or vocalizing — these systems often transmit little more than background noise or silence, potentially leaving caregivers uncertain about the baby’s well-being during prolonged stillness.

[0011] US20100241018A1 discloses a baby monitor that detects at least one of a child's vital signs — such as breathing, heartbeat, or movement — and allows remote monitoring via Internet or telephone networks. The system may alert caregivers when physiological parameters fall outside preset thresholds and may provide audible or visual reassurance (e.g., LEDs or repetitive sound indicators). While this prior art addresses general vital sign monitoring and remote alerting, its audio feedback is focused on producing discrete tones or pulses (e.g., one tick per heartbeat), rather than generating naturalistic, continuous audio that mimics the real-time respiratory pattern of the user.Such tick-based feedback could even be perceived as stressful or unnatural in some care environments. Furthermore, the cited prior art does not address integration with immersive ARA / R environments, or the collection of wellness- related metrics (such as HRV or sleep phase estimates) for long-term cloudbased analysis — features that are central to the present invention.

[0012] US 10,874,332 B2 discloses a respiration monitoring system for infants that combines a garment — with a high-contrast geometric pattern visible in near-infrared wavelengths — with a camera and IR illumination system mounted above the crib. The system detects respiratory motion by tracking the rhythmic movement of the pattern across the thorax. Unlike the present invention, this prior art relies on visual pattern tracking and image-based respiratory detection using camera systems and does not produce naturalistic audio feedback that mimics breathing rhythms.-Page 3 of 26-4820PCT_spec_claims_abst_03b.docx 51790 7574 185

[0013] US patent 7,697,891 B2 describes a baby monitor system comprising a “child unit” equipped with a transducer, ADC, microprocessor, and wireless transmitter, and a “parent unit” receiving and reconstructing the audio for caregivers. Security features include channel scanning, encryption, and frequency hopping. However, this system solely transfers ambient sounds (e.g., crying) using standard radio techniques, and lacks sensor-based respiratory motion detection, continuous synthetic audio feedback, adaptive modulation, immersive application integration, intelligent condition-specific alerting, or cloud-based wellness data analysis — all key elements of the present invention.SUMMARY OF THE INVENTION

[0014] The present invention provides a system and method for monitoring respiratory movements and converting them into soothing, natural-sounding audio feedback. The generated audio can be played back in real time through a nearby speaker — such as a Bluetooth-connected audio device — to provide peace of mind to caregivers or parents, or to support self-regulation in wellness contexts. The audio output mimics the rhythm of natural breathing and functions as a real-time indicator of ongoing respiration.

[0015] The system comprises at least one respiratory sensor selected from a group including: piezoelectric sensors (e.g., PVDF or PZT), quasipiezoelectric sensors such as permanently charged polypropylene-based ferroelectrets, strain gauges, mmWave radar, optical fiber sensors, or accelerometers. Sensor modalities may also include PPG- or ECG-based sensor respiration estimators commonly used in wearable devices. In the present innovation, signals from these sensors are processed to determine breathing rate, phase, and intensity, which are then used to drive an audio synthesis unit that generates breathing-like sound.

[0016] In one embodiment, filtered white noise is modulated in real time by the respiratory signal to create a continuous audio stream that reflects the user’s current respiration. In another low-power embodiment, only breathing rate data is transmitted via Bluetooth to a receiving audio device. The receiver then generates a sine wave at the corresponding frequency, which modulates-Page 4 of 26-4820PCT_spec_claims_abst_03b.docx 51790 7574 185pre-filtered inhalation and exhalation noise samples using amplitude modulation and zero-order hold interpolation. The final signal is output via PWM and played through a speaker — efficiently recreating the rhythm of breathing with minimal data and computational overhead.

[0017] In addition to real-time playback, the system can upload analyzed physiological data — such as breathing rate, heart rate, heart rate variability (HRV), and sleep / wake state estimates — to a cloud service for long-term storage and tracking of well-being. This enables applications such as sleep diaries and retrospective pattern analysis, without the bandwidth or storage overhead of transmitting raw audio.

[0018] Optional integrations with augmented reality (AR) and virtual reality (VR) platforms are also supported, allowing the audio feedback or respiratory metrics to be used in immersive training, meditation, or therapeutic applications.

[0019] The system includes an intelligent alert mechanism that responds to critical conditions by either modifying the audio output (e.g., switching to a distinctive sound) or playing pre-recorded spoken messages that clearly describe the issue. Example messages may include: “Your baby is restless and breathing cannot be analysed”, “Bluetooth connection between the cot sensor and the speaker is disconnected”, “Cot sensor’s power is disconnected”, “Repetitive muscle spasms detected - check child’s condition”.

[0020] These alerts ensure that caregivers receive timely and understandable information about the monitored subject’s condition. In cases where the caregiver or parent is hearing-impaired, the system may additionally or alternatively trigger a connected light source to provide visual reassurance or flashing alarms.

[0021] In one embodiment, the system may also serve as or integrate with a traditional baby monitor, combining ambient audio transmission with real-time physiological feedback. Unlike prior art systems that use playback of prerecorded sounds (e.g., maternal heartbeats), the present invention generates dynamic, real-time audio based on the user's own respiratory signal — resulting in a personalized and adaptive experience.-Page 5 of 26-4820PCT_spec_claims_abst_03b.docx 51790 7574 185

[0022] The invention enables a novel combination of passive monitoring, therapeutic audio feedback, and intelligent multi-modal alerting — all within a compact, low-power system suitable for infants, wellness users, and immersive technology applications.BRIEF DESCRIPTION OF DRAWINGS

[0023] The invention is described in more detail by means of examples by referring to the following drawings:

[0024] Fig. 1 is an isometric view of a newborn baby 104 sleeping in a cot 105 on a mattress 106 in one room 103, and the mother 102 of the baby in another room 100 listening to the baby’s breathing sound via a device 101 of the innovation.

[0025] Fig. 2 is an isometric view of a sensor 200 in accordance with one or more embodiments of the present invention. In the picture are shown cable 201 with USB-C connector for power input; plastic corner case 202 covering integrated electronics; and holes 204 for fixing the sensor on the bed frame, under the mattress.

[0026] Fig. 3A is a cross-sectional view of the sensor in Fig. 2, in accordance with one or more embodiments of the present invention.

[0027] Fig. 3B is a cross-sectional view taken at cross section A-A of Fig. 3A in accordance with one or more embodiments of the present invention showing electronics printed circuit board 207 inside of the corner case 202. Aferro- electret sensor 203 is seen in between soft foams 205A and 205B, covered with for example thin sheets of PVC, 206 A and 206B.

[0028] Fig. 3C is a cross-sectional view taken at section B-B of Fig. 3A in accordance with one or more embodiments of the present invention.

[0029] Fig. 4 is an isometric view of an audio receiver and playback unit 101 , in accordance with one or more embodiments of the present invention. It has speaker 401 under the top part, and a volume control 402, and a power on-off button 403.

[0030] Fig. 5 shows two block diagrams, one of a sensor and the other of an audio receiver and playback unit, in accordance with one or more embodiments of the present invention.-Page 6 of 26-4820PCT_spec_claims_abst_03b.docx 51790 7574 185

[0031] Fig. 6 is a flowchart describing the signal processing flowDETAILED DESCRIPTION OF THE INVENTION

[0032] This section provides a comprehensive explanation of the system architecture, main individual components, signal processing flow, and audio synthesis techniques used to transform respiratory activity into real-time audio feedback.System Overview

[0033] Referring to the Fig. 1 , the invention comprises a sensor-based system that in a typical use case scenario, monitors the respiratory motion of a sleeping newborn baby 104 in a room 103 and generates a corresponding audio signal to another room 100 via an audio playback device 101. The system may consist of either one or two physical units: first, a sensor unit (Fig.2 - 200) with onboard signal processing and wireless transmission (Fig. 5 - 620), and secondly, a receiver and audio playback unit 101 (Fig. 1 - 101 , Fig. 4 - 101 , Fig. 5 - 630).

[0034] The sensor unit to detect breathing can be of many kinds. It can be a wearable (for example, PPG-based) sensor, or a nearby proximity sensor (for example, mmWave based), or hidden under-mattress type (for example, charged ferro-electret, or piezoelectric, or optical cable-based). Additionally, a microphone for sound recording may be integrated.

[0035] The separate receiver and audio playback unit (Fig. 1 - 101 , Fig. 4 - 101 , Fig. 5 - 630) is equipped with a speaker (Fig. 4 - 401 , Fig. 5 - 615) for audio output and a volume control (Fig. 4 - 401 , Fig. 5 - 617) for it. In the preferred embodiment, it also has cloud connectivity (Fig. 5 - 615).

[0036] Fig. 3A is a cross-sectional view of the sensor in Fig. 2, in accordance with one or more embodiments of the present invention.

[0037] Fig. 3B is a cross-sectional view taken at cross section A-A of Fig. 3A in accordance with one or more embodiments of the present invention showing electronics printed circuit board 207 inside of the corner case 202. A ferro- electret sensor 203 is seen in between soft foams 205A and 205B, covered with for example thin sheets of PVC, 206 A and 206B.-Page 7 of 26-4820PCT_spec_claims_abst_03b.docx 51790 7574 185

[0038] Fig. 3C is a cross-sectional view taken at section B-B of Fig. 3A in accordance with one or more embodiments of the present invention.

[0039] In a minimal embodiment, the system may be integrated into a single enclosure that performs both sensing and playback functions, while the generated white noise-based audio can be transmitted using a commercially available baby cry monitor system.

[0040] In the preferred embodiment, the respiratory sensor is a permanently charged ferro-electret film type, installed under a mattress (Fig 1 - 106), but it can be of type that is capable of detecting breathing motions and / or rate. Possible types include, but are not limited to, strain gauges, piezoelectric sensors (e.g., PVDF or PZT), accelerometers, mm Wave sensors (radar), optical fiber-based movement sensors, video cameras with motion analysis algorithms. Alternatively, the respiratory signal may be derived from heart rate variability, measured using photoplethysmography (PPG) or electrocardiogram (ECG)-based sensors. One example of such a method is described in the scientific publication “Deriving Respiration from the Pulse Photoplethysmographic Signal,” published in Computing in Cardiology, 2011 , volume 38, pages 713-716.

[0041] Depending on the sensor technology, certain implementations may also support simultaneous detection of cardiac signals (e.g., heart rate and heart rate variability) using bandpass filtering or waveform analysis.

[0042] The system may further include algorithms capable of detecting abnormal muscle activity — such as the rapid contractions associated with tonic-clonic seizures — using threshold-based or machine learning-enhanced signal classification. The feasibility of such capabilities, particularly with ferroelectret-based sensors, has been demonstrated in prior publications, including: “Prospective Study of the Emfit Movement Monitor,” Journal of Child Neurology, 2013, volume 28(11 ), pages 1434-1436; and “Assessment of a Quasi-Piezoelectric Mattress Monitor as a Detection System for Generalized Convulsions,” Epilepsy & Behavior, 2013, volume 28, pages 172-176.

[0043] To minimize latencies and power consumption, Aa central controller (e.g., microcontroller) orchestrates the sensor’s data acquisition (Fig. 5 --Page 8 of 26-4820PCT_spec_claims_abst_03b.docx 51790 7574 185604), processing, audio synthesis (if at the sensor), Bluetooth transmission (Fig. 5 - 606), and cloud communication (Fig. 5 - 615), if that is also integrated into the same device, while minimizing latency and power consumption.Respiratory Sensor

[0044] In the preferred embodiment, the respiratory sensor detects mechanical movements associated with breathing, such as thoracic or abdominal expansion and contraction. For infants, the sensor can be integrated into a wearable device (e.g., chest band, sock, or smart onesie), but is most preferably implemented as a durable pad placed under the mattress to ensure comfort and non-intrusive monitoring. This pad preferably incorporates an internally charged ferro-electret sensor material, selected for its passive operation, durability, and high sensitivity.

[0045] One preferred embodiment utilizes a permanently charged, laminated ferro-electret film sensor developed explicitly by the inventors. The design and manufacturing method of such a sensor are described in detail in U.S. Patent Application Ser. Nos. 10 / 996,263 and 10 / 018,413. In this invention, the sensor is preferably laminated between PVC or PU sheets, forming a flexible pad structure optimized for placement under a mattress.

[0046] Alternative sensor technologies may include PVDF film, PZT ceramics, strain gauges, optical fibers, mmWave radar, accelerometers, and video cameras combined with motion analysis algorithms. In neonatal or clinical environments where physical contact is undesirable (e.g., incubators), contactless approaches — such as radar-based sensors, video-based respiratory tracking, or under-mattress solutions using ferroelectret sensors — may be preferred by some users or institutions. These non-contact methods reduce infection risk and avoid the need for direct skin attachment, which may be unsuitable for preterm infants or patients with fragile or sensitive skin.Signal Processor

[0047] Referring to the Fig. 5, showing two block diagrams, in preferred embodiment utilizing a ferroelectret film sensor (600), the signal chain includes a dedicated analog front-end preamplifier (602) prior to digital-Page 9 of 26-4820PCT_spec_claims_abst_03b.docx 51790 7574 185processing. The analog front-end consists of a low-noise preamplifier stage with an extremely low high-pass cutoff frequency (approximately 0.07 Hz), allowing the system to detect even very slow breathing movements, including those interrupted by apnea events.

[0048] The preamplified signal is then split at least into two but preferably into three separate bandpass paths: one optimized for capturing respiratory movement (typically with lower gain and frequency range focused on ~0.1 - 1 Hz), and the other optimized for detecting cardiac-related micro-movements (with higher gain and a band typically centered around 5-20 Hz), and a third with small gain and about 0,1 Hz high-pass filter for detecting mentioned seizure activity related strong movements. A good low-pass cutoff for both branches is, for example, at around 80 Hz.

[0049] After analog preprocessing, the signal is digitized and analyzed by a microcontroller — for example, using a Nordic Semiconductor nRF52840 system-on-chip, integrated within a Fanstel BT840F module (Fig. 5 - 603 - 604 - 605). This module performs local signal processing and manages the Bluetooth transmission of the preprocessed breathing and, in some embodiments, also heartbeat data, to an audio receiver and playback unit.

[0050] In this preferred embodiment’s architecture, the audio receiver and playback unit also contains a BT840F module or equivalent, which handles audio synthesis and playback. This allows the computational load and audio output circuitry (including speaker, amplifier, and volume control) to be located separately from the sensor unit, improving comfort and reducing electromagnetic interference near the bed.

[0051] The audio receiver and playback unit may further include a Wi-Fi and / or LTE-M communication module (Fig. 5 - 615) to upload analyzed physiological data (e.g., respiratory rate, heart rate, HRV, and event markers) to a cloud server for long-term storage or real-time monitoring. These data transmissions are the primary sources of electromagnetic emissions, which are an essential consideration in neonatal settings, where reducing EMF exposure is often desired.-Page 10 of 26-4820PCT_spec_claims_abst_03b.docx 51790 7574 185

[0052] To enhance accessibility, the receiving unit can also provide a discrete output interface (e.g., a dry contact relay or GPIO trigger) for external alerting devices such as light-based indicators — handy for caregivers who are hearing-impaired.

[0053] In alternative embodiments, the signal processor and audio synthesis unit may be co-located within the same device — such as the sensor pad placed under the baby — allowing the generated breathing-like audio to be played directly at the sensing location. This configuration enables the use of conventional baby monitors with built-in microphones to relay the sound to caregivers, eliminating the need for a dedicated receiver unit. If desired, the same unit may also include wireless connectivity (e.g., Wi-Fi or LTE-M) for transmitting physiological data to a cloud service.

[0054] In such integrated implementations, electromagnetic field (EMF) exposure near the infant can be minimized by placing the actual electronics — including wireless modules and processors — at a short distance (e.g., 1 meter) from the passive sensor element, connected via a shielded cable. This allows the sensor pad itself to remain passive and free of emissions directly beneath the mattress. In the described, preferred configurations where a separate receiver unit is used for audio playback and cloud communication, the sensor unit may only require a low-power wireless transmitter (e.g., Bluetooth) and can be fully integrated into a sealed unit (Fig. 2 - 202) under the mattress. This unit may be powered via for example USB-C cable (Fig. 2 - 201 ) or be battery-powered, using rechargeable batteries such as NiMH cells, which are inherently safer than lithium-ion batteries and reduce the need for trailing power cables — an important safety consideration, as strangulation hazards associated with power cords have been reported in baby monitor incidents.

[0055] Optional signal analysis algorithms may be implemented in the signal processing to detect high-frequency tremors or repetitive signal spikes that are characteristic of tonic-clonic seizure activity. These algorithms may operate on the cardiac signal band (e.g., 1-20 Hz) or use additional frequency ranges as needed.-Page 11 of 26-4820PCT_spec_claims_abst_03b.docx 51790 7574 185

[0056] Upon detection of such abnormal movement patterns, the system triggers a distinct alert, which may include a change in audio output, a spoken message, cloud notification, or activation of auxiliary devices (e.g., light signals or caregiver paging systems).Audio Synthesis Unit:

[0057] Referring to the Fig. 1 - 101 , Fig. 4 - 101 , Fig. 5 - 630 and the flowchart in Fig. 6, the audio synthesis unit generates real-time audio signals that reflect the user’s breathing patterns in a natural, soothing, and perceptually synchronized manner.

[0058] Depending on implementation constraints and design goals, the unit can utilize MIDI libraries, software synthesis frameworks (e.g., Mozzi, Pure Data), or hardware audio synthesis chips such as Yamaha YMZ280B, DFPlayer Mini, or VS1053b.

[0059] The inhalation and exhalation phases may be mapped to different tones, timbres, frequencies, or filtered noise profiles to simulate the dynamic qualities of natural respiration.

[0060] In one embodiment, white noise is processed through distinct digital filters for inhalation and exhalation. The filtered signals are then modulated by the amplitude envelope of the real-time respiratory waveform — producing a breathing sound that changes fluidly with the user's own breathing depth and rhythm.

[0061] The resulting audio can range from soft, wind-like ambient noise to other relaxing sounds such as a maternal heartbeat or womb-like rhythmic pulses. In some embodiments, users may select from different “sound themes” to suit infant or adult use cases.

[0062] The synthesized audio may be played through a local speaker at an adjustable volume. Volume control can be manual, app-based, or adaptive based on environmental noise levels.

[0063] In the event of a detected abnormality — such as apnea, irregular breathing, signal loss, or tonic-clonic seizure — the audio output is modified to draw the caregiver’s attention. This can involve changing the pitch, rhythm, volume, or timbre of the ongoing audio signal.-Page 12 of 26-4820PCT_spec_claims_abst_03b.docx 51790 7574 185

[0064] Alternatively, the system may interrupt the breathing sound and play prerecorded spoken alerts stored locally on the device. These voice messages may include, for example: “Your baby is restless and breathing cannot be analysed”, “Bluetooth connection between sensor and speaker is disconnected”, “Sensor power is disconnected”, “Repetitive muscle spasms detected - check the child’s condition”.

[0065] Spoken alerts may be available in multiple languages and may be customizable by the user or caregiver depending on the context of use.

[0066] Audio playback may also be synchronized with external devices (e.g., smart lights) to provide multimodal alerts. The previously mentioned Fanstel BT840F module, based on the Nordic nRF52840 chipset, provides both signal processing and Bluetooth communication.

[0067] In some implementations, alternative modules may be used — especially when the development team has existing firmware or hardware platforms based on them. One such example is the ESP32, which includes integrated Bluetooth and Wi-Fi connectivity. This makes it a practical choice for rapid prototyping or when unified handling of signal processing, cloud communication (via Wi-Fi), and transmission of the generated white-noise- based audio to a third-party Bluetooth speaker is desired. As it supports both Bluetooth Low Energy (BLE) and Bluetooth Classic (BR / EDR), it offers flexibility in audio streaming applications. However, its power consumption is typically higher than that of the BT840F, which may be a limiting factor in battery-powered designs, particularly for systems intended for continuous 24 / 7 operation near sleeping infants.Bluetooth Transmitter:

[0068] BLE may be used for transmitting metadata — such as breathing rate, sleep stage estimates, or alert status — while Bluetooth Classic is used for continuous, real-time streaming of synthesized audio signals.

[0069] In the preferred implementation, where the system is divided into two physical units: a sensor unit (with integrated analog front-end and signal processing) and a receiver unit (which handles audio synthesis, playback,-Page 13 of 26-4820PCT_spec_claims_abst_03b.docx 51790 7574 185cloud connectivity, and optional alert outputs) the Bluetooth is used to send preprocessed data from the sensor to the receiver.

[0070] In an alternative embodiment, all functionality — including sensing, signal processing, and audio synthesis — is integrated into a single device. This device may transmit audio directly to a standard, commercially available Bluetooth speaker, baby monitor, or headphone without requiring a dedicated receiver unit.

[0071] This configuration allows the system to function similarly to a conventional audio source (e.g., smartphone), streaming breathing-related audio as standard Bluetooth audio (e.g., A2DP) to any compatible playback device. Volume control and pairing may then be handled using the speaker's built-in controls.

[0072] The system may optionally maintain BLE connectivity in parallel to provide metadata (e.g., timestamps, breathing rate, alert codes) to a companion mobile app or cloud gateway.

[0073] Additionally, Bluetooth HID or custom GATT profiles may be used to interface with augmented reality (AR) or virtual reality (VR) systems for synchronized biofeedback in immersive environments.

[0074] The flexibility to operate with both proprietary and off-the-shelf Bluetooth audio devices allows for a wide range of use cases — from infant monitoring in the home to biofeedback applications in meditation studios or research settings.Cloud Communication Module

[0075] In embodiments where audio playback and connectivity are handled by a dedicated receiver unit, the receiver may include integrated Wi-Fi and / or LTE-M (or NB-loT) modules to enable wireless communication with cloud servers (Fig. 6 - 615).

[0076] In a preferred embodiment, the communication module transmits only preprocessed and summarized physiological metrics, such as breathing rate (in breaths per minute), heart rate (HR), heart rate variability (HRV), detected movement artifacts, and sleep / wake state estimates. The uploaded data can-Page 14 of 26-4820PCT_spec_claims_abst_03b.docx 51790 7574 185be used to generate automated sleep diaries, well-being summaries, anomaly logs, or trend reports accessible to caregivers, clinicians, or family members.

[0077] In the present invention, there is no need to transmit the sensor’s raw signal waveform to the cloud, as the intended functionality can be achieved without it. This approach conserves bandwidth and storage and reduces regulatory burden in sensitive environments such as infant or elder care. Nevertheless, if required — for example in clinical research studies or certain advanced applications — the raw waveform data can also be transmitted, and in practice we have often implemented such transfers.

[0078] The uploaded data can be used to generate automated sleep diaries, well-being summaries, anomaly logs, or trend reports accessible to caregivers, clinicians, or family members. Cloud connectivity also enables remote access to historical data, real-time alerts, and system health monitoring (e.g., battery level, sensor status, connection status).

[0079] Communication protocols may include secure REST APIs, MQTT, or HTTPS with TLS encryption. The device may authenticate with the cloud platform using certificate-based or token-based mechanisms.

[0080] In deployments without Wi-Fi coverage, the LTE-M or NB-loT module allows independent operation over mobile networks, making the system suitable for use in locations such as rural homes, hospitals, or mobile care units.

[0081] In some embodiments, the communication module may support dual connectivity (Wi-Fi and cellular) with automatic failovers for enhanced reliability.

[0082] Optionally, integration with third-party health platforms or caregiver dashboards may be enabled via secure API gateways.

[0083] Cloud data may be simultaneously used for retrospective analysis, daily summaries, sleep diary generation, caregiver dashboards, or integration with third-party health platforms.Audio Output Device

[0084] Referring to the Fig. 1 - 101 , Fig. 4 - 101 , Fig. 5 - 630, in the preferred configuration, the audio output device is integrated into a dedicated receiving-Page 15 of 26-4820PCT_spec_claims_abst_03b.docx 51790 7574 185unit, which wirelessly receives respiration-derived audio signals via Bluetooth (Fig. 5 - 612) from the sensor unit (Fig. 2 - 200, Fig. 5 - 606). This receiver may include a built-in speaker (401 , 615), amplifier (616), and volume control (617), and is typically placed near the caregiver or parent to provide continuous, reassuring feedback that mimics the infant’s real-time breathing rhythm.

[0085] Alternatively, as explained earlier, a simplified “all-in-one” configuration may generate the breathing-like audio directly within the sensor unit and transmit it to a third-party Bluetooth speaker, headphones, or other playback device. In such cases, volume control is typically handled on the external audio device itself.

[0086] For added functionality, the receiver unit or system may include auxiliary output interfaces (e.g., dry contact relays or GPIO) to activate external visual alerts, such as flashing lights or color changes. This supports hearing- impaired caregivers or situations requiring non-disruptive nighttime awareness.

[0087] The system may also integrate with smart home lighting platforms, such as Philips Hue, Matter-compatible devices, or other loT ecosystems. These integrations can be used to generate synchronized light patterns that reflect the user’s breathing rhythm, or to signal alerts using changes in brightness, color, or pulsing effects.

[0088] In multi-user or multi-room environments, the system can be configured to broadcast to multiple synchronized audio and / or visual output devices, ensuring that critical alerts are perceived regardless of caregiver location.

[0089] Method for Simulating and Producing Breathing Sounds as an Audio Signal:Measurement:

[0090] Referring to Fig. 5 (600-602-603-604) and the flow chart at Fig. 6. (box 401 , 404). In the preferred embodiment, respiratory movement is detected using a ferroelectret film-based sensor placed under the user’s mattress or cushion. The sensor passively generates charge signals in response to-Page 16 of 26-4820PCT_spec_claims_abst_03b.docx 51790 7574 185mechanical deformation caused by thoracic or abdominal movements during breathing. For the high impedance signal, there is first an analogue charge preamplifier (Fig. 5 - 602) with high- and lo-pass filters and dividing the signal, in preferred embodiment to three different channels with different gains and filterings.

[0091] Alternative sensor modalities may include piezoelectric elements (e.g., PVDF or PZT), optical fiber-based strain sensors, strain gauges, accelerometers, radar (e.g., mmWave), or video-based movement analysis. Some implementations may prefer wearable solutions — such as PPG-based or ECG-based respiration estimation — due to component availability or target application. Contactless sensing (e.g., radar or camera) may be particularly suited for neonatal or critical care use cases where skin contact is undesirable.Signal Conditioning

[0092] When using an analog sensor, such as a ferroelectret or piezo film, the signal is amplified using a low-noise preamplifier and digitized via an analog- to-digital converter (ADC). The following filtering steps are then applied:

[0093] High-pass filtering attenuates low-frequency baseline drift caused by mattress sagging, environmental pressure changes, or slow posture shifts.

[0094] Low-pass filtering reduces high-frequency noise components, such as electronic interference or minor ambient vibrations.

[0095] These filters improve the signal-to-noise ratio but do not eliminate motion artifacts caused by the user’s body. Large or abrupt movements — such as limb repositioning or shifting posture — typically remain visible in the signal and may interfere with respiration detection. However, during periods of stillness, the conditioned signal is sufficiently clean to extract respiratory phase and rate information. In such cases, heartbeat-related micromovements can also be detected, enabling optional calculation of heart rate and heart rate variability (HRV).In preferred embodiments, the analog signal may be separated into multiple frequency bands — either prior to digitization (via analog filters) or after-Page 17 of 26-4820PCT_spec_claims_abst_03b.docx 51790 7574 185digitization (via digital filtering) — to allow independent analysis of respiration and cardiac components.Sound Generation:

[0096] The audio signal is synthesized using one or more methods: a. Modulated filtered white noise to simulate natural airflow sounds, b. Synthesized tones or pulses using MIDI or software oscillators, c. Pre-recorded or sampled sound segments corresponding to inhalation and exhalation, d. Womb-like, rhythmic or ambient sounds designed for infant comfort.

[0097] Separate digital filters or audio envelopes sculpt the inhalation and exhalation phases into distinct sonic profiles.

[0098] These sounds can be further shaped by attack / release parameters to simulate the dynamics of actual breathing.Modulation:

[0099] The real-time respiratory waveform modulates the amplitude of the generated sound in real time.

[0100] Typically, positive signal values control the inhalation phase, while the absolute value of negative signal values controls the exhalation phase.

[0101] This modulation results in a continuous audio stream that dynamically follows the user’s breathing rhythm and intensity.Transmission:

[0102] The modulated audio signal is sent wirelessly via Bluetooth (Classic or BLE-Audio) to a connected audio output device.

[0103] Latency is minimized to ensure perceptual synchrony between actual breathing and audible feedback.

[0104] In alternative embodiments, the audio may be played locally through a built-in speaker.Operation Flow:

[0105] The sensor captures respiratory motion and transmits the signal to the processor.

[0106] The processor identifies the breathing phase, evaluates for abnormalities (e.g., apnea, seizures), and controls the audio synthesis unit.-Page 18 of 26-4820PCT_spec_claims_abst_03b.docx 51790 7574 185

[0107] The audio unit generates a synchronized sound stream based on the real-time breathing pattern and sends it to the speaker via Bluetooth.

[0108] The system uploads summarized physiological metrics (e.g., RR, HR, HRV, sleep state) to the cloud at regular intervals.

[0109] In the case of a critical event, the system triggers an alert — either as a distinct audio tone, spoken message, or external signal.

[0110] Optional features include playback of calming prerecorded sounds such as maternal heartbeat, ocean waves, or white noise.

[0111] In some versions, smart lighting adjusts brightness or color in synchrony with the breathing pattern or alert condition.Alternative Embodiments:

[0112] The system may be integrated into a variety of form factors, such as hospital beds, neonatal incubators, elderly care mattresses, meditation cushions, or mobile sleep monitors.

[0113] In clinical or research environments, the audio synthesis may be temporarily disabled with the volume knob while retaining silent alerts and data logging.

[0114] For mobile or adult users, the system may offer customizable sound themes, intensity settings, and connectivity to wearable devices or smartphone apps.

[0115] Alert mechanisms may include audio, visual, or haptic feedback (e.g., vibration or flashing lights), supporting caregivers who are hearing-impaired or in noisy environments.

[0116] In AR / VR applications, the breathing data may be rendered as synchronized audio or visual feedback to guide breath control, relaxation training, or immersive therapeutic experiences.

[0117] These embodiments are provided as illustrative examples and are not intended to limit the scope of the invention. Many variations, modifications, and improvements may be made by those skilled in the art without departing from the spirit and scope defined by the following claims.Breathing Sound Synthesis from White Noise-Page 19 of 26-4820PCT_spec_claims_abst_03b.docx 51790 7574 185

[0118] Breathing sounds can be synthetically generated by shaping white noise to reproduce the characteristic spectral features of inhalation and exhalation. White noise is a suitable basis because it contains all frequency components uniformly, allowing the spectral content to be adjusted with filters. In natural respiration, breathing sounds arise mainly from turbulent airflow in the respiratory tract, which produces a noise-like signal. The objective of this work is to emulate these turbulence characteristics through signal processing.

[0119] Referring to the Fig. 6, steps 401 and 404, with our innovation, for the breathing sound synthesis, we have developed two input options. Both can be used in the same audio playback device, and depending on the use case or user preferences, either can be chosen without any component changes or switches. Simply by enabling and disabling features in the firmware via user interference, such as an app in a smartphone controlling the firmware via Bluetooth connection.

[0120] The input source option 1 (401 ) uses the measured breathing signal (sensor’s band pass filtered breathing waveform signal as an input. After the input, the first step is offset removal (step 402). In it the DC offset is removed from the measured breathing signal to center the waveform around zero (403).

[0121] The input source option 2 uses the measured (calculated) breathing rate as an input. In this option next step is 405, which is sinusoidal waveform generation. In it a synthetic sinusoidal waveform with frequency matching is generated from the breathing rate in breaths per minute (BPM).

[0122] In step 407 is input selection. In it, the breathing waveform signal input is selected between sources from steps 403 and 406. Selection can depend on, but is not limited to, a device configuration, quality analysis, hardware or software implementation. Only one input source is used in further steps.

[0123] In step 408 is upsampling the low sampling frequency breathing waveform signal to an audio frequency sample rate, such as but not limited to, 8 kHz to match the target sampling rate used in audio synthesis (step 409). The upsampling approach may be, but not limited to, Zero-Order Hold (ZOH), where each sample value is held constant until the next sample.-Page 20 of 26-4820PCT_spec_claims_abst_03b.docx 51790 7574 185

[0124] In step 410 is a half-cycle separation, meaning the waveform is split into positive and negative half-cycles — representing inhalation (step 411 ) and exhalation (step 412), respectively.

[0125] In step 413 is white noise generation. We found it is pleasant to listen white noise as a basis for breathing sound synthesis because it contains all frequency components uniformly, allowing the spectral content to be adjusted with filters.

[0126] In step 414 is exhalation filtering - namely, applying a lower band-pass Filter. Further, we generate and achieve an exhalation noise signal in step 415 by filtering white noise (step 413) into a lower frequency pass-band of 100 Hz - 200 Hz. First, a second-order Butterworth low-pass filter (cutoff 200 Hz, slope 12 dB / octave) emphasizes low-frequency components while strongly reducing mid- and high-frequency content. Secondly, a first-order Butterworth high-pass filter (cutoff 100 Hz, slope 6 dB / octave) suppresses the very lowest components, leaving a defined low-frequency band (100-200 Hz) that produces a soft, humming exhalation character.

[0127] In step 416 is inhalation filtering - namely applying a higher band-pass filter and generating an inhalation noise signal (step 417) by filtering white noise (step 413) into a higher frequency pass-band of 250 Hz - 1 kHz. First, a second- order Butterworth low-pass filter (cutoff 1 kHz, slope 12 dB / octave) removes the highest frequency components while retaining a broad frequency band, yielding a hiss-like sound. Secondly, applying a first-order Butterworth low-pass filter (cutoff 4 kHz, slope 6 dB / octave) brings an additional gentle roll-off to prevent sharp or artificial timbre.

[0128] Third, a first-order Butterworth high-pass filter (cutoff 250 Hz, slope 6 dB / octave) attenuates well any very low-frequency rumble, preserving midrange and higher components for an airy sound quality.

[0129] Steps 418 and 419 are amplitude Modulation. We use the up-sampled half-cycles (steps 411 and 412) to amplitude-modulate the pre-filtered inhalation (step 417) and exhalation noise signals (step 419). The pre-filtered inhalation noise was amplitude-modulated with the positive half-cycle, while the pre-filtered exhalation noise was amplitude-modulated with the negative half-cycle.-Page 21 of 26-4820PCT_spec_claims_abst_03b.docx 51790 7574 185

[0130] Step 422 is about the addition of modulated Inhalation and exhalation noise signals. We added two modulated signals (steps 420 and 421 ) to generate a complete breathing cycle sound signal (step 423).

[0131] The final step is 424, the Pulse-Wide-Modulation (PWM) sequence generation. Following it we convert the final waveform into a PWM sequence (step 425) suitable for driving a speaker directly. A PWM sequence was generated from the modulated noise and fed into the loudspeaker amplifier, taking polarity requirements into account.

[0132] As the result, the filtering stage successfully produced two distinct base sounds: Inhalation: hiss-like, airy sound with energy concentrated in mid and partly high frequencies; exhalation: deeper, humming sound dominated by low- frequency components.

[0133] When modulated with either measured breathing signals or synthetic sinusoidal cycles, the resulting output captured the perceptual differences between inhalation and exhalation. The PWM output method enabled direct loudspeaker driving with polarity control.

[0134] The system demonstrates that realistic breathing sounds can be generated from white noise using our filtering and modulation techniques. The two approaches — measured signal vs. sinusoidal cycle — offer flexibility depending on the availability of physiological data. The chosen sampling frequency (8 kHz) is sufficient to cover the spectral range of interest, while measured breathing signals within 25-200 Hz provide accurate amplitude modulation cues.-Page 22 of 26-4820PCT_spec_claims_abst_03b.docx 51790 7574 185

Claims

CLAIMSWhat is claimed is:1 . A system for generating real-time audio feedback based on respiratory movements of a user, the system comprising:- at least one respiratory sensor configured to detect respiratory motion of the user;- a signal processor configured to analyze the sensor signal and determine respiratory phases and / or respiratory rate;- an audio synthesis unit configured to generate an audio signal that reflects the user’s breathing activity; and- an audio output module, comprising either:(a) a wireless transmitter configured to send either (i) the generated audio signal or (ii) the extracted respiratory rate to an external audio playback device, or(b) a built-in speaker configured to play the generated audio signal locally, wherein the audio signal is dynamically modulated to represent the user’s breathing pattern, and wherein, in an alternative embodiment, the respiratory rate is used to generate a synthetic audio signal at the audio output device by modulating prerecorded sound samples using zero-order hold interpolation or equivalent signal generation methods.

2. The system of claim 1 , wherein the respiratory sensor comprises at least one of:- a ferroelectret film sensor,- a piezoelectric sensor (e.g., PVDF or PZT),- a strain gauge,- an accelerometer,- an mmWave radar unit,- an optical fiber-based motion sensor,- a camera with motion analysis algorithms, or-Page 23 of 26-4820PCT_spec_claims_abst_03b.docx 51790 7574 185- a photoplethysmographic (PPG) or electrocardiographic (ECG) sensor deriving respiration from heart rate variability.

3. The system of claim 1 , wherein the audio synthesis unit applies amplitude modulation to filtered noise samples representing inhalation and exhalation, based on the respiratory phase.

4. The system of claim 1 , wherein the wireless transmitter is a Bluetooth LowEnergy (BLE) or Bluetooth Classic (BR / EDR) module, and the external audio playback device is a Bluetooth speaker, soundbar, smart baby monitor, or AR / VR headset.

5. The system of claim 1 , wherein the audio output module includes a local speaker located on the same device as the sensor, and the emitted sound is intended to be captured by an external baby monitor or similar audio relay system.

6. The system of claim 1 , further comprising a cloud communication module configured to upload processed physiological data — including breathing rate, heart rate, heart rate variability, movement events, and sleep / wake estimates — to a remote server for monitoring or analysis.

7. A method for generating breathing-like audio feedback based on respiratory sensor data, the method comprising:- detecting respiratory motion using a sensor placed near or on the user;- calculating the breathing rate and / or determining the current respiratory phase;- generating an audio signal that reflects the breathing pattern; and- either (a) transmitting the audio signal or breathing rate to an external audio output device, or(b) playing the generated audio directly on a built-in speaker located near the sensor.

8. The method of claim 7, wherein generating the audio signal comprises:- selecting separate prerecorded or filtered noise samples representing inhalation and exhalation,- creating a sinusoidal or phase-synchronized control signal based on the respiratory rate, and-Page 24 of 26-4820PCT_spec_claims_abst_03b.docx 51790 7574 185- modulating the noise samples using amplitude modulation and zero-order hold interpolation.

9. The method of claim 7, further comprising filtering the sensor signal with high- pass and low-pass filters to reduce baseline drift and high-frequency noise, and optionally separating respiratory and cardiac components prior to audio synthesis.

10. The method of claim 7, wherein the generated audio signal is played locally near the user and is intended to be captured by an external microphone of a third-party monitoring system (e.g., baby monitor), allowing indirect relay to a caregiver without wireless transmission from the sensing unit.11 . The system of claim 1 , further comprising an alert mechanism configured to detect abnormal respiratory patterns or muscular activity, and to trigger a change in the audio output upon detection of such events.

12. The system of claim 11 , wherein the audio output is modified to include spoken messages selected from a predefined set of phrases describing the detected abnormality.

13. The system of claim 1 , wherein the audio synthesis unit is further configured to optionally play prerecorded audio tracks designed to promote relaxation, sleep, or caregiver reassurance.

14. The system of claim 1 , wherein the signal processor includes an analog frontend comprising a high-pass filter with a cutoff frequency below 0.1 Hz and a low-pass filter with a cutoff frequency below 100 Hz, to extract respiratory signals from sensor noise.

15. The system of claim 1 , wherein the system is further configured to control visual output devices — such as LEDs or smart lighting systems including Philips Hue or Matter-compatible devices — in synchronization with the user’s respiratory pattern or alert conditions.

16. The system of claim 1 , wherein the audio and visual outputs are transmitted or mirrored to multiple synchronized output devices across different rooms or locations to ensure continuous feedback or alerts are perceivable by caregivers regardless of position.-Page 25 of 26-4820PCT_spec_claims_abst_03b.docx 51790 7574 185

Images