Multi-modal fusion neonatal safety auxiliary monitoring system and method for maternal and infant care institutions

Abstract

This invention discloses a multimodal fusion neonatal safety monitoring system and method for maternal and infant care institutions. The system includes a bedside monitoring terminal, a centralized management platform at the nurse station, and a tiered viewing terminal for parents. Through a distributed wireless local area network, the bedside monitoring terminal integrates millimeter-wave radar and a video monitoring module. It employs a dual-modal data spatiotemporal synchronization cross-verification mechanism, generating a valid alarm only when both types of sensors detect an anomaly within the same time window, significantly reducing the false alarm rate. The bedside terminal uses a passive audio-visual design and is physically separated from the parent display terminal, completely eliminating audio-visual pollution. The system constructs a dual-stack operation mode of "local LAN backup + optional cloud synchronization" to ensure that core monitoring remains effective even when the external network is interrupted. Combining tiered access control and differentiated alarm strategies, it balances privacy compliance, a quiet environment, and emergency response, making it suitable for commercial scenarios such as postpartum care centers and maternal and infant care institutions.

Description

Technical Field

[0001] This invention relates to the field of non-medical safety auxiliary monitoring technology for infants and young children, specifically to a centralized newborn care, safety backup monitoring, and digital operation management system based on multimodal sensor fusion and local area network distributed architecture, for B-end commercial scenarios such as postpartum care centers and maternal and infant care institutions. Background Technology

[0002] Currently, newborn monitoring systems in B-end scenarios such as postpartum care centers and maternal and infant care institutions are still in a rudimentary stage, mainly consisting of live video streaming and human supervision. The mainstream solutions fall into two categories, both of which suffer from fundamental shortcomings in adapting to different scenarios:

[0003] Category 1: Home / Security-Grade Solutions. This involves directly using home-use infant monitoring devices and ordinary security cameras to build a passive video streaming system, simply pushing the footage to the parents' room or mobile phone. This type of solution lacks a centralized management platform, proactive anomaly monitoring, and safety fallback warnings. It relies on manual monitoring to ascertain the infant's status, making it unsuitable for institutions with multiple beds, cross-regional operations, and centralized management models. Monitoring efficiency is low, and risk control capabilities are insufficient.

[0004] The second category is medical-grade retrofit solutions. These involve simplified modifications to medical-grade neonatal monitoring equipment. However, such equipment is complex to deploy, has cumbersome wiring, high procurement and maintenance costs, and requires a high level of professional expertise to operate. Furthermore, it lacks features suitable for commercial nursing institutions, such as privacy control, lightweight viewing for parents, and operational data traceability. It fails to meet the needs of non-medical, commercial, and efficient operations in postpartum care centers, making large-scale adoption difficult.

[0005] Both types of solutions have irreconcilable technical contradictions:

[0006] 1. The contradiction between accurate early warning and low false alarm: The false alarm rate of a single sensor (video only or radar only) is extremely high (such as false alarms caused by clothing obscuring the image or changes in light and shadow), which interferes with the normal operation of the institution;

[0007] 2. Conflict between safety alarms and quiet environment: Existing monitoring equipment has built-in audible and visual alarms, which can seriously disturb newborns and mothers when deployed near infants, thus disrupting the quiet environment.

[0008] 3. The conflict between centralized operation and privacy compliance: The existing solution cannot balance the parents' viewing needs with the protection of the baby's privacy, and it is highly dependent on the external network connection. Once the network is disconnected, the monitoring will be invalid, which poses a major security risk.

[0009] Therefore, there is an urgent need for a newborn safety monitoring system that can simultaneously resolve the above contradictions and is suitable for B-end business scenarios. Summary of the Invention

[0010] The purpose of this invention is to overcome the shortcomings of the prior art and provide a multimodal fusion neonatal safety auxiliary monitoring system and method for maternal and infant care institutions.

[0011] To achieve the above objectives, the present invention adopts the following technical solution:

[0012] A multimodal fusion neonatal safety monitoring system for maternal and infant care institutions, characterized in that it includes: a bedside monitoring terminal and a centralized management platform at the nurse station; the bedside monitoring terminal is deployed next to the neonatal care bed and includes:

[0013] The radar monitoring module is used to monitor the respiratory signs and body movement of newborns.

[0014] The video monitoring module is used to identify the newborn's posture and the condition of the surrounding coverings;

[0015] The bed-side monitoring terminal is connected to the centralized management platform at the nurse station via the institution's wireless local area network;

[0016] The bedside front-end monitoring terminal is configured to execute dual-modal joint judgment logic: within the same time window, it performs time-series matching and feature comparison on radar monitoring data and video monitoring data, and generates a valid alarm signal only when both types of data are determined to be in an abnormal state.

[0017] Furthermore, the radar monitoring module is configured to monitor the newborn's respiratory rate and body movement amplitude, as well as other vital signs data.

[0018] The video monitoring module is configured to identify the newborn's posture, surrounding coverings, and limb compression environment data.

[0019] In the dual-modal joint judgment logic, abnormalities in the vital signs dimension and abnormalities in the environmental dimension must match within the same judgment time window in order to be judged as valid security risks.

[0020] Furthermore, the dual-modal joint determination logic specifically includes:

[0021] The radar monitoring data is processed to extract respiratory rate and body movement amplitude features. If the features deviate from the preset safety threshold, they are marked as abnormal in the vital signs dimension.

[0022] Image recognition is performed on video frame data to extract features such as infant position, facial coverings, and limb compression. If the features match the preset risk model, they are marked as abnormal in the environmental dimension.

[0023] The timestamps of abnormalities in the physical and environmental dimensions are synchronized and calibrated. If the two overlap within the time window T, it is considered a valid alarm; if only a single module is abnormal, it is considered interference and filtered out.

[0024] Furthermore, the wireless LAN distributed networking architecture supports dual-mode operation: "local backup" and "cloud synchronization".

[0025] When the external network connection is normal, the system will synchronize alarm data and monitoring records to the cloud server;

[0026] When the external network is interrupted, the system automatically switches to pure local area network mode. The bedside front-end monitoring terminal still executes the dual-modal joint judgment logic and pushes the effective alarm signal to the nurse station centralized management platform through the local area network to ensure that the core monitoring function is uninterrupted.

[0027] Furthermore, it also includes a parent-child tiered monitoring terminal, which is physically separated from the bedside front-end monitoring terminal and can be started and stopped independently;

[0028] The system also includes a tiered, differentiated alarm mechanism:

[0029] When the centralized management platform of the nurse station receives a valid alarm signal, it triggers an audible and visual alert and a pop-up notification.

[0030] When the parent-child tiered viewing terminal receives a valid alarm signal, it triggers a silent vibration or a pop-up notification on the interface, without emitting an alarm sound that penetrates the room.

[0031] Furthermore, the parental control tiered viewing terminal includes a local wireless display terminal and a remote mobile viewing terminal, and the system is configured with a tiered access control module:

[0032] The local wireless display terminal is deployed in a parent-only area, allowing parents to view real-time status data on-site without authorization.

[0033] The remote mobile viewing terminal can only access data after identity binding verification and privacy authorization agreement, and remote access is automatically prohibited when the external network is interrupted.

[0034] Furthermore, it also includes an auxiliary sensing module, which includes a temperature and humidity sensor and a crying sound acquisition module;

[0035] The system is configured with a heterogeneous data distribution and view isolation module:

[0036] The data collected by the auxiliary sensing module does not participate in the dual-modal joint determination logic, but is only used for state visualization.

[0037] The centralized management platform for nurse stations hides the auxiliary sensor data by default and only displays core alarm information. It supports manual viewing to assist in handling the situation.

[0038] The parent-child tiered viewing terminal displays all auxiliary sensor data by default.

[0039] Furthermore, the system also includes a heterogeneous data distribution and view isolation module, which is configured as follows:

[0040] The collected data is divided into core alarm data and auxiliary environment data. The core alarm data is generated by the radar monitoring module and the video monitoring module and participates in the dual-modal joint determination. The auxiliary environment data is generated by the auxiliary sensing module and does not participate in the dual-modal joint determination logic.

[0041] The centralized management platform at the nurse station distributes and displays only the core alarm data by default to support nursing staff in quickly locating risks; the parent-level viewing terminal distributes and displays the full amount of core alarm data and auxiliary environment data by default to improve parents' perception of the nursing environment.

[0042] This invention also provides a non-medical newborn safety monitoring method for maternal and infant care institutions, applied to the aforementioned system, comprising the following steps:

[0043] Data acquisition steps: Simultaneously acquire radar monitoring data and video monitoring data through the bedside monitoring terminal;

[0044] Cross-validation steps: Spatiotemporal synchronization of dual-modal data is performed in the local processor. A valid alarm signal is generated only when both types of data are determined to be abnormal within the same time window. During this process, the front end of the bed remains silent and without light output.

[0045] Alarm distribution steps: Valid alarm signals are pushed to the centralized management platform of the nurse station via wireless LAN to trigger audible and visual alerts;

[0046] Data management steps: Store monitoring data locally via the local area network and selectively upload it to the cloud based on network conditions.

[0047] Furthermore, the data management steps also include network disaster recovery and dual-stack operation mode switching steps:

[0048] Network status monitoring sub-step: Real-time monitoring of external network connection status;

[0049] Cloud synchronization sub-step: When the external network connection is normal, the effective alarm signals and monitoring record data will be synchronized to the cloud server, and data access will be supported by remote mobile viewing terminals;

[0050] Local backup steps: When an external network interruption is detected, the system automatically switches to a pure LAN operating mode, blocks the access permissions of remote mobile viewing terminals, and maintains the dual-modal joint judgment logic of the bed-side front-end monitoring terminal and the alarm distribution function within the LAN to ensure that core monitoring services are not interrupted.

[0051] The beneficial effects of this invention are:

[0052] 1. More reliable safety monitoring: Through millimeter-wave radar (vital signs) + video (environment) dual-modal spatiotemporal synchronous cross-verification, the error of a single sensor is effectively filtered out, and the false alarm rate is reduced by more than 90% compared with the single sensor solution, providing accurate early warning of high-risk situations such as suffocation;

[0053] 2. Adaptable to B-end operations: The distributed wireless LAN networking supports cross-regional and centralized management, and the core monitoring does not fail when the external network is interrupted, which is in line with the large-scale operation mode of postpartum care centers;

[0054] 3. More comfortable environment: The monitoring terminal and display terminal are physically separated, and the baby terminal has a passive sound and light design, which completely eliminates the sound and light pollution of the device and creates a suitable quiet environment for mothers and babies.

[0055] 4. More efficient operation and management: Centralized control at the nurse station and full-process data traceability enable closed-loop management of monitoring work, and differentiated data display to balance nursing efficiency and parents' right to know;

[0056] 5. Enhanced privacy compliance: A tiered access control mechanism (local access without authorization, remote access requiring binding) balances parents' viewing needs with privacy protection, meeting the compliance requirements for commercial services;

[0057] 6. Enhanced service experience: Visualized display of auxiliary data such as temperature, humidity, and crying sounds improves parents' recognition and trust in the institution's services. Attached Figure Description

[0058] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the present invention will be further described below in conjunction with the accompanying drawings and embodiments. The drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort:

[0059] Figure 1 This is an overall architecture diagram of a non-medical newborn safety auxiliary monitoring system for maternal and infant care institutions according to an embodiment of the present invention;

[0060] Figure 2 This is a flowchart of a non-medical newborn safety monitoring method for maternal and infant care institutions according to an embodiment of the present invention;

[0061] Figure 3 This is a flowchart of the dual-modal joint determination logic method according to an embodiment of the present invention;

[0062] Figure 4 This is a flowchart of the network disaster recovery and dual-stack operation mode switching method according to an embodiment of the present invention. Detailed Implementation

[0063] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, a clear and complete description will be provided below in conjunction with the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the protection scope of the present invention.

[0064] This invention provides a non-medical newborn safety auxiliary monitoring system for maternal and infant care institutions, such as... Figure 1 As shown, the system includes a bedside monitoring terminal 10 and a centralized management platform at the nurse station 20. The bedside monitoring terminal 10 is deployed next to the neonatal care bed and includes: a radar monitoring module 11 for monitoring the newborn's respiratory signs and body movement; and a video monitoring module 12 for identifying the newborn's posture and the status of surrounding coverings. The bedside monitoring terminal 10 and the centralized management platform at the nurse station are connected via the institution's wireless local area network. The bedside monitoring terminal 10 is configured to execute a dual-modal joint judgment logic: within the same time window, it performs time-series matching and feature comparison on radar monitoring data and video monitoring data, and generates a valid alarm signal only when both types of data are determined to be in an abnormal state.

[0065] Among them, the bed-side monitoring terminal 10 is a passive audio-visual design, without its own buzzer, alarm indicator light or other audio-visual alarm components; the radar monitoring module 11 is configured to monitor the newborn's respiratory rate and body movement amplitude vital sign data; the video monitoring module 12 is configured to identify the newborn's body position, surrounding coverings, and limb compression environmental data; in the dual-modal joint judgment logic, abnormalities in the vital sign dimension and abnormalities in the environmental dimension must match within the same judgment time window in order to be judged as a valid safety risk.

[0066] Furthermore, it also includes a parent-led tiered viewing terminal 30, and the three are connected via a distributed wireless LAN network. The bed-side monitoring terminal 10 and the parent-led tiered viewing terminal 30 are physically separated and can be started and stopped independently. There is no direct wired connection between the two; they communicate only through a wireless LAN. The nurse station centralized management platform 20 is used to receive valid alarm signals from all beds in the area, enabling centralized control, alarm push, and data traceability. The parent-led tiered viewing terminal 30 is used to view the newborn's status in tiers without interfering with the dual-modal joint judgment logic.

[0067] In the system described above, the three core functional modules form an "end-pipe-cloud (end)" architecture: Bedside monitoring terminal 10: Deployed next to the neonatal bed, responsible for data collection and edge computing. Nurse station centralized management platform 20: The management hub of the B-end institution, responsible for data aggregation, alarm processing, and traceability. Parent-tiered viewing terminal 30: A user-facing viewing terminal, available in local and remote versions. All terminals achieve distributed networking connectivity through the institution's wireless local area network (WLAN), unlike traditional point-to-point connections or pure cloud connections, ensuring data flows locally and can operate without relying on the external network.

[0068] Preferably, the radar monitoring module is a millimeter-wave radar module, with a preferred operating frequency band of 60GHz, used to collect vital sign data, such as respiration and body movement.

[0069] Preferably, the passive audio-visual design means that the terminal itself does not have buzzers, alarm indicator lights or other audio-visual alarm components, or its functions are shielded / removed, thus eliminating alarm noise and light pollution from a physical perspective.

[0070] In the system described in the above embodiments, such as Figure 3 As shown, the dual-modal joint decision logic differs from simple "OR" logic or single sensor triggering. It is executed locally at the bedside monitoring terminal (edge ​​computing), and specifically includes:

[0071] Step S21: Process the radar monitoring data and extract respiratory rate and body movement amplitude features. If the features deviate from the preset safety threshold, they are marked as abnormal in the vital signs dimension.

[0072] Step S22: Perform image recognition on the video frame data to extract features such as infant position, face covering, and limb compression. If the features match the preset risk model, mark them as abnormal in the environmental dimension.

[0073] Step S23: Synchronize and calibrate the timestamps of abnormalities in the physical and environmental dimensions. If the two overlap within the time window T, it is determined to be a valid alarm; if only a single module is abnormal, it is determined to be interference and filtered.

[0074] For example, an alarm is only triggered when radar detects abnormal breathing (vital signs) and video detects facial obstruction (environment). If only radar or video reports a false alarm, the system does not generate an alarm, thus significantly reducing the false alarm rate.

[0075] The above embodiment creatively breaks down the assessment of suffocation risk into two orthogonal dimensions—"physical signs (radar)" and "physical environment (video)"—by introducing the concept of "spatiotemporal synchronization." An alarm is only triggered when both dimensions simultaneously exhibit problems. This logically simulates the thought process of a professional caregiver assessing risk (checking for both abnormal breathing and external obstructions), surpassing conventional dual-sensor combinations and addressing the industry pain point of "simple stacking of dual sensors leading to persistent false alarms" in existing technologies. The system must also possess timestamp synchronization calibration capabilities to eliminate logical misjudgments caused by sensor response delays or timing misalignments.

[0076] Scenario 1: Precisely capturing the risk of choking on milk / overheating in centralized childcare settings

[0077] Scenario Description: In a newborn care room, an infant accidentally pulls the blanket up to their face while sleeping, causing breathing obstruction. In this situation, a single sensor (such as pure video) may not be able to see the face clearly due to the blanket obscuring it, while a single radar may misinterpret the slight movement of the infant as normal.

[0078] The specific implementation steps are as follows:

[0079] 1. Data Acquisition and Dimensional Breakdown: Vital Signs Dimension (Millimeter-Wave Radar): The radar continuously monitors the infant's chest cavity micro-movements. If the respiratory rate drops sharply from 40 breaths / min to 10 breaths / min within 5 seconds (below the preset safety threshold of 30 breaths / min), it is marked as "Abnormal Vital Signs Dimension." Environmental Dimension (Video Monitoring): The camera, through AI image recognition, detects that the infant's facial area is obscured by more than 70% of the covering (blanket) for more than 3 seconds, and this is marked as "Abnormal Environmental Dimension."

[0080] 2. Judgment Time Window Matching: The system sets the "judgment time window" to 5 seconds. The system synchronously compares the timestamps of the radar and video. Detection: From time T0 to T5, the radar continuously outputs "abnormal breathing," and the video continuously outputs "facial occlusion" from time T1 to T4. Logical Judgment: The two types of anomalies highly overlap within the aforementioned 5-second time window (overlap time > 3 seconds).

[0081] 3. Valid Alarm Generation: Based on logical judgment, a valid high-risk asphyxiation event is identified, and a valid alarm signal is generated. Result: The nurse station immediately receives a strong alert, and nursing staff quickly arrive to remove the covering, preventing an accident.

[0082] Scenario 2: False Alarm Filtering for "Motion Shadows" in Mother-Baby Rooming-in Scenarios

[0083] Scenario Description: In a room where the mother and baby are in the same room, changes in lighting cause the curtains to cast shadows on the crib area, or the baby may be tossing and turning violently in their sleep. Existing single-video or single-radar solutions are highly likely to misinterpret such situations as abnormalities and trigger alarms, disturbing the mother and baby's rest.

[0084] Specific implementation steps:

[0085] 1. Data Acquisition and Dimensional Segmentation: Vital Signs Dimension (Millimeter-Wave Radar): When the infant rolls over, the radar detects a significant increase in body movement (belonging to a normal active state), but the respiratory rate remains stable at 38 breaths / min, not triggering the abnormal vital signs threshold. The system determines "Vital Signs Dimension Normal." Environmental Dimension (Video Monitoring): The camera detects a dark shadow on the bed surface (curtain shadow). The image recognition algorithm misjudges this as "unknown covering" or "abnormal body position," and the system marks it as "Environmental Dimension Abnormal."

[0086] 2. Judgment Time Window Matching: The system sets the "judgment time window" to 3 seconds. Timestamps are compared: Radar data shows normal breathing (no abnormal vital signs) from time T0 to T3; video data marks environmental anomalies at time T1. Logical Judgment: Within the 3-second time window, only environmental anomalies exist; vital sign data are not synchronously abnormal.

[0087] 3. Interference Filtering and Silencing: Based on the above logic, since the two types of data were not simultaneously identified as abnormal within the same time window, the system determined it to be an interference signal. Result: The system did not generate a valid alarm signal, the bed terminal remained silent, and there were no pop-ups at the nurse station, successfully avoiding a false alarm caused by changes in light and shadow, and ensuring a quiet environment for the mother and baby in the same room.

[0088] In a further embodiment, respiratory rate and body movement amplitude features are marked as abnormal in the vital signs dimension if they deviate from a preset safety threshold; image recognition is performed on video frame data to extract infant position, facial coverings, and limb compression features, and if the features conform to a preset risk model, they are marked as abnormal in the environmental dimension; the timestamps of the two types of abnormalities are synchronously calibrated, and if the two overlap within an adjustable time window of 1-10 seconds, they are determined to be a valid safety risk; if only a single dimension is abnormal, it is determined to be interference and filtered out.

[0089] The "dual-modal joint decision logic" in the previous embodiment is decomposed into 3 serial processing steps + 1 final decision rule. The execution subject, input, processing procedure, and output of each step are clearly defined:

[0090] Step 1: Marking anomalies in vital signs (executed by the millimeter-wave radar processing unit)

[0091] Execution unit: Millimeter-wave radar processing unit built into the bed-side front-end monitoring terminal (60GHz band, sampling frequency 10 times / second)

[0092] Input: Raw point cloud data acquired by millimeter-wave radar

[0093] Processing flow:

[0094] Filter and reduce noise in point cloud data to extract features of chest cavity micro-movement caused by breathing and body movement features caused by limb activity;

[0095] Two core characteristic values ​​were calculated: respiratory rate (breaths / min) and body movement amplitude (quantitative value).

[0096] Compare with the "preset safety thresholds" that can be customized by B-end institutions (such as respiratory rate 20-60 breaths / min, body movement amplitude ≤ threshold X).

[0097] Output: If the feature value deviates from the safety threshold, mark it as "abnormal in vital sign dimension" with an abnormal timestamp; otherwise, no abnormality is marked.

[0098] Step 2: Environmental Dimension Anomaly Marking (Executed by Video Processing Unit)

[0099] Execution Entity: The video processing unit built into the bedside front-end monitoring terminal (integrated NPU edge computing module, video frame rate 5 frames / second)

[0100] Input: Raw video frame data collected by the video monitoring module

[0101] Processing flow:

[0102] AI image recognition was performed on the video frames to extract three core features: infant position (prone / supine / side-lying), percentage of face covering (%), and limb entrapment status (whether stuck in the crib rails / wrapped in bedding).

[0103] Compare with the "preset risk models" that B-end institutions can customize (such as "prone position + face coverage > 50%" and "limb compression" are identified as risk characteristics).

[0104] Output: If the feature matches the risk model, mark it as "Environmental Dimension Anomaly" with an anomaly timestamp; otherwise, no anomaly mark is given.

[0105] Step 3: Time synchronization and overlap determination (executed by the central processing unit MCU)

[0106] Execution Entity: Central Processing Unit of Bedside Monitoring Terminal

[0107] Input: The exception markers and corresponding timestamps output from steps 1 and 2.

[0108] Processing flow:

[0109] Timestamp synchronization calibration: Align the timestamps of radar and video data to control the timing deviation of multiple sensors within 100ms and solve the problem of hardware sampling asynchrony.

[0110] Adjustable time window matching: The judgment time window can be configured from 1 to 10 seconds (e.g., 5 seconds by default for centralized care scenario, 3 seconds by default for mother-infant rooming-in scenario, and can be extended to 10 seconds for premature infant scenario).

[0111] Overlap determination rules: If the time period of "abnormal physical signs" and the time period of "abnormal environmental conditions" overlap for ≥ 50% of the window duration (customizable) within the determination time window, it is determined as a valid security risk; if only a single dimension has an abnormal marker, or the overlap duration is insufficient → it is directly determined as interference, and no alarm is generated.

[0112] It should be noted that existing dual-sensor monitoring solutions typically use "OR" logic (an alarm is triggered if either sensor malfunctions) or simple "AND" logic (an alarm is triggered if both sensors output abnormalities simultaneously). This differs fundamentally from the solution presented in this embodiment.

[0113] Comparison Dimensions Existing dual-sensor solutions This plan Feature extraction Two types of sensors extract similar features (such as both monitoring respiration), resulting in overlapping dimensions. Radar extracts only the vital signs dimension, and video extracts only the environmental dimension; the features are completely orthogonal and have no redundancy. Time determination There is no concept of a time window; it is triggered whenever an exception is flagged. The exceptions must overlap within a configurable time window, and the exception must overlap for a specified duration before triggering. Scene adaptation Fixed thresholds cannot be adapted to different nursing scenarios. The time window, safety threshold, and risk model can all be customized by B-end institutions to adapt to different scenarios such as centralized care, mother-infant rooming-in, and premature infants.

[0114] Practice has proven that the technical solution adopted in the embodiments of the present invention has the following significant effects:

[0115] 1. Significantly reduced false alarm rate: The false alarm rate is reduced by more than 90% compared to single-sensor solutions, and by an additional 60% compared to ordinary dual-sensor AND logic solutions;

[0116] 2. Controllable response speed: The alarm can be triggered in as little as 1 second, meeting the response requirements of the "golden 4 minutes" for suffocation risk;

[0117] 3. Adaptable to B-end operations: Adjustable parameters allow institutions to customize according to newborn type and care level, adapting to the needs of large-scale deployment.

[0118] Taking the scenario of an infant in the aforementioned centralized childcare room experiencing a risk of suffocation due to their face being covered by a blanket as an example, the specific implementation steps of this plan are as follows:

[0119] Step 1 (Radar): Respiratory rate <30 breaths / min for 5 seconds from T0 to T5 → Mark "Abnormal vital signs", timestamp T0-T5;

[0120] Step 2 (video): During times T1-T4 (3 seconds in total), the proportion of facial coverings is >70% → mark "Environmental dimension anomaly", timestamp T1-T4;

[0121] Step 3 (MCU): If the deviation after timestamp calibration is <50ms, the judgment time window is set to 5s, and the overlap time between the two is 3s (accounting for 60% ≥ 50% of the window) → it is judged as a valid security risk and an alarm signal is generated.

[0122] In a further embodiment, the system supports a dual-stack operating mode: when the external network connection is normal, the system synchronizes alarm data and monitoring records to the cloud server, supporting remote access; when the external network is interrupted, the system automatically switches to a pure local area network mode, the bed-side front-end monitoring terminal still executes the dual-modal joint judgment logic locally, and pushes valid alarm signals to the nurse station centralized management platform and the local parent viewing terminal through the local area network, ensuring uninterrupted core monitoring functions.

[0123] In the Internet of Things (IoT) field, the mainstream trend is "cloudification," with devices primarily acting as data collectors and computing power residing in the cloud. This embodiment takes the opposite approach, emphasizing "edge computing + local area network," pushing core logic (dual-modal judgment) down to the front-end terminal. The external network serves only as a channel for value-added services (remote monitoring), not a life-or-death necessity. It clearly distinguishes between "core monitoring data flow" (closed loop within the local area network) and "value-added service data flow" (synchronized over the external network). This design ensures that even if the cloud server crashes or the fiber optic cable is severed, the care services within the postpartum care center will not be affected.

[0124] Implementation scenario: External network failure at a postpartum care center at night

[0125] Fault Occurrence: At 2:00 AM, the postpartum care center's external network was suddenly interrupted due to a fault in the operator's line.

[0126] Automatic switching: When the system detects that the external network is disconnected, it automatically activates the "LAN fallback mode". Cloud synchronization stops, but the nurse station server and the front-end terminal still maintain a connection through the internal wireless LAN.

[0127] Alarm triggered: There is a risk of choking on vomiting in a certain bed, and the front-end terminal's dual-modal judgment is abnormal.

[0128] Local response: Alarm signals are instantly pushed to the nurse station (with strong sound and light alerts) and the parent's room terminal (with a silent pop-up) via the local area network, and nurses quickly arrive at the scene to handle the situation.

[0129] Result: Throughout the process, parents were unable to access the app (remote access was cut off), but the internal life safety monitoring of the institution was not affected in any way, thus avoiding a safety incident caused by network failure.

[0130] In a further embodiment, the system is configured with a differentiated alarm mechanism: when the nurse station centralized management platform receives a valid alarm signal, it triggers an audible and visual alert and a pop-up notification; when the parent-child tiered viewing terminal receives a valid alarm signal, it triggers a silent vibration or an interface pop-up notification, without emitting an external alarm sound; the bedside front-end monitoring terminal remains silent throughout the entire process of generating and transmitting a valid alarm signal.

[0131] That is, based on the aforementioned embodiments, three new core alarm rules are added to cover the entire alarm flow chain, specifically including:

[0132] 1. Nurse Station: Strong reminder configuration (matching the "treatment" role)

[0133] When the centralized management platform of the nurse station receives a valid alarm signal, it triggers an audible and visual alert and a pop-up notification.

[0134] Technical details: The alarm priority of the nurse station platform is the highest level. The audible and visual reminders must meet the requirements of "audible ≥70dB and light flashing frequency ≥2 times / second". The pop-up window must be pinned to the top and cannot be automatically closed until the nursing staff manually confirms the action.

[0135] Design logic: Nursing staff are the sole responsible party for responding to alarms. Strong reminders ensure no reports are missed, and the dual-modal, low-false-alarm characteristic avoids invalid reminders from interfering with operations.

[0136] 2. Parent's side: Mute setting (matching the "insider" role)

[0137] When a parent receives a valid alarm signal, the terminal will trigger a silent vibration or a pop-up notification on the screen, without emitting an external alarm sound.

[0138] Technical details: Parents are strictly prohibited from configuring external buzzers on the device. Alarms will only be sent via terminal vibration and a red pop-up notification on the interface. Simultaneously, a silent push notification will be sent to the remote mobile device to prevent alarm sounds from penetrating the room and disturbing the baby or mother.

[0139] Design logic: Parents only need the right to know, not the responsibility to handle the situation. The silent design satisfies the need for information while completely eliminating the interference of alarms on the quiet environment.

[0140] 3. Bedside area: Completely silent process (matching the "ward" role)

[0141] The bedside monitoring terminal remains silent throughout the entire process of generating and transmitting valid alarm signals.

[0142] Technical details: The "full process" covers three stages: ① The instant when the dual-modal system detects an anomaly and generates an alarm signal; ② The process of transmitting the alarm signal through the local area network; ③ The waiting period for the nurse station's response. No audio or visual output is allowed throughout the entire process; even if the firmware is faulty or manually modified, the software logic will block the audio and visual trigger interfaces.

[0143] Design Logic: It forms a double guarantee with the aforementioned "hardware passive sound and light design (no buzzer, no indicator light)" - there are no sound and light emission components at the hardware level, and any sound and light output logic is prohibited at the software level, completely eliminating the possibility of any interference from the front-end terminal.

[0144] Compared with existing technical solutions, this solution has the following outstanding features:

[0145] Existing alarm schemes Invention Solution Non-obviousness statement A unified audio and visual alarm is triggered across all terminals: it sounds at the front end, the nurse station, and the parent's terminal, severely disturbing the infant. Tiered by role: Strong reminder at the nurse station, mute on the parent's end, and completely silent on the front end. The conventional thinking in this field is that "the louder the alarm, the safer it is." This invention goes against the grain, based on the premise of a low false alarm rate in dual-modality mode. Only when the false alarm rate is low enough can we "silence" the alarms on the front end and the parent end. This combination of "low false alarm + layered alarm" cannot be derived from conventional technology. While the front-end sound and light effects are removed at the hardware level, no software restrictions are implemented, posing a risk of firmware bugs triggering the sound and light effects. Double hardware and software blockade: no hardware components and complete prohibition of audio and visual output throughout the software process. In this field, there is no proactive approach to implementing redundant "full-process silence" designs at the software level, unless it is for the extreme need for quiet care in the maternal and infant scenario, which is a scenario-specific innovation.

[0146] Taking the aforementioned centralized care scenario as an example, when a bed triggers a suffocation risk alarm, the status of the three terminals is as follows:

[0147] 1. Bedside front end: The radar and video dual-mode system detects abnormalities at time T0, generates an alarm signal, and transmits it to the nurse station and parents' terminal via the local area network. Throughout the process, the terminal is completely silent with no flashing lights or beeping sounds.

[0148] 2. Nurse Station: A red pop-up window appears at the top of the screen at T0+0.5s, accompanied by a 70dB beep and a flashing red light. Nursing staff should respond immediately.

[0149] 3. Parents' Room: A mute pop-up window appears on the local wireless display terminal, and the parents' mobile phones receive a mute push notification with no external sound, so as not to disturb the mother's rest, but to let the parents know the progress of the treatment in real time.

[0150] In a further embodiment, the parent-controlled viewing terminal includes a local wireless display terminal and a remote mobile viewing terminal. The system is configured with a hierarchical access control module: the local wireless display terminal is deployed in a parent-only area, allowing parents to view real-time status data on-site without authorization; the remote mobile viewing terminal requires identity verification and a privacy authorization agreement before accessing the data, and remote access is automatically blocked when the external network is interrupted. This embodiment constructs a dynamic privacy protection network based on physical location trust and network status linkage, dividing parent access permissions into two dimensions and setting strict control rules. It ensures that infant data is both "visible (locally convenient)" and "protected (remotely compliant)," which is an important legal and technical guarantee for the commercialization and compliant operation of this invention.

[0151] The specific control rules are as follows:

[0152] Terminal type Deployment location Permission rules Technical Logic and Advantages Local wireless display terminal Parent-only areas (such as parents' rooms, visitation areas) View without authorization Logic: Automatic authentication via physical location (specific MAC address or IP range within the local area network). Advantages: Greatly improves viewing convenience; parents do not need to remember account passwords, allowing for instant access and enhanced service experience. Remote mobile viewing terminal Any public network location (such as a parent's mobile app) Identity binding and privacy authorization required Logic: Access is only granted after mobile phone number / ID verification and signing of an electronic privacy agreement. Advantages: Meets legal compliance requirements and prevents unauthorized disclosure of infant privacy data. Dynamic access control System Network Layer Automatically disable remote access when external network is interrupted Logic: Remote access data streams are forced to rely on the external network channel. When the system detects an external network disconnection (switching to LAN fallback mode), the remote port is automatically closed. Advantages: Physical isolation, eliminating the risk of data leakage during periods of external network instability or outages.

[0153] The innovation of this solution lies not only in its "access control" but also in its "scenario-based and dynamic" design:

[0154] Physical location equates to access rights: Existing technologies typically require passwords or facial recognition. This invention leverages the physical enclosure of maternal and infant care facilities to transform the physical fact of being "in the parents' room" into a "authorized" trust credential. This is a non-obvious design specifically for the postpartum care center scenario.

[0155] Network status-linked access control: This deeply integrates access control with the "LAN / WAN dual-mode" system. When the WAN connection is lost (entering pure LAN mode), the system automatically cuts off all remote access paths. This design is based on a security logic: in an uncontrolled state such as a network outage, data security should be prioritized over accessibility. This is not a general access control method, but a dedicated design for highly privacy-sensitive scenarios.

[0156] Differentiated service experience: Local authorization-free access ensures ease of use (service experience), while remote authorization ensures compliance (legal responsibility). The combination of the two perfectly meets the operational needs of B-end institutions.

[0157] Specific implementation scenario 1: Parents view the room (local access without authorization)

[0158] Parents can access a dedicated wireless display terminal (such as a tablet) in their own room at the postpartum care center. The terminal automatically connects to the nurse station platform via the local area network, displaying their baby's video and status data directly without requiring any account or password. The process is seamless and requires no prior authorization.

[0159] Specific implementation scenario two: Parents remotely check on their children while away from home (strong authorization)

[0160] Parents can open a mobile app while running errands. The app prompts them to enter their ID number and SMS verification code for identity verification, and then displays a "Privacy Authorization Agreement" requiring them to agree. After confirmation, the app retrieves data from the cloud via the internet to display the baby's status.

[0161] Specific implementation scenario 3: External network interruption (automatically disable remote access)

[0162] The postpartum care center's external fiber optic cable was severed, and the system automatically switched to local area network (LAN) backup mode. At this time, when parents attempted to refresh the app on their mobile phones, the system returned a "Network error, remote service suspended" message, and no data could be viewed. However, the local display terminal in the room could still view data normally (because it was connected via the LAN).

[0163] In a further embodiment, such as Figure 1As shown, the bedside monitoring terminal also integrates a temperature and humidity sensor and a crying sound acquisition module as an auxiliary sensing module 13; the data collected by the auxiliary sensing module 13 does not participate in the dual-modal joint judgment logic, but is only used for status visualization display; the nurse station centralized management platform hides the auxiliary sensing data by default and only displays the core alarm information, which can be manually enabled for viewing; the parent-level viewing terminal displays the auxiliary sensing data in full by default.

[0164] Existing maternal and infant monitoring solutions generally suffer from two types of deficiencies in data display. The design of this invention is a customized innovation for B-end scenarios and is not a conventional technical approach in this field:

[0165] Existing technical solutions This plan Non-obviousness statement All data is displayed to all roles simultaneously: the nurse station receives alarms, temperature and humidity data, and crying data at the same time, leading to information overload and affecting the efficiency of handling the situation. By role: Auxiliary data is hidden by default at the nurse station, while all data is displayed on the parent's end. The conventional approach in this field is "the more complete the data, the better." This invention takes the opposite approach, designing triage rules to address the different needs of nurses to "focus on safe procedures" and parents to "perceive the nursing environment." No existing technology provides such a scenario-based inspiration. Auxiliary sensor data is used in alarm determination: for example, excessively high temperature may also trigger an alarm, leading to an increased false alarm rate. Auxiliary data is not involved in the core alarm logic at all; it is only used for display. Those skilled in the art would typically incorporate more sensor data into alarm logic to improve "monitoring comprehensiveness." This invention proactively decouples auxiliary data from alarms, a targeted design based on the principle that "low false alarms are a core requirement for B-end operations," which can be derived from unconventional logic.

[0166] Scenario 1: Centralized Hosting Scenario

[0167] An infant triggered a suffocation risk alarm. After receiving the core alarm "Bed 3 Face Covering", the nurse station manually activated the auxiliary data for that bed and found that the room temperature was 26°C and the humidity was 55%. This ruled out respiratory abnormalities caused by overheating and the covering on the infant's face was quickly removed. The parents could always see the temperature, humidity, and crying data for that bed, and knew that the infant was in a comfortable environment. They expressed their approval of the institution's handling efficiency.

[0168] Scenario 2: Rooming-in with mother

[0169] Parents can check on their baby's condition at night via a local tablet and find that the crying frequency has increased. At the same time, they see that the room temperature is 25°C and the humidity is 50%, which is within the normal range. They determine that the baby is not uncomfortable due to environmental issues, but is crying because of hunger. They can directly call the nursing staff to feed the baby, reducing unnecessary nurses' rounds and improving nursing efficiency.

[0170] In a further embodiment, the system also includes a heterogeneous data distribution and view isolation module. This module is configured to: divide the collected data into core alarm data and auxiliary environment data, whereby the core alarm data is generated by the millimeter-wave radar module and video monitoring module and participates in dual-modal joint judgment; the auxiliary environment data is generated by the auxiliary sensing module 13 and does not participate in the dual-modal joint judgment logic; by default, only the core alarm data is distributed and displayed to the centralized management platform at the nurse station to support nursing staff in quickly locating risks; and by default, all core alarm data and auxiliary environment data are distributed and displayed to the parent-level viewing terminals to improve parents' awareness of the nursing environment. In other words, the "core data used for judgment (radar + video)" and the "auxiliary data used for display (temperature, humidity, crying)" are logically isolated, and differentiated displays are made according to the different needs of the nurse station (efficiency priority) and the parent terminal (right to know priority).

[0171] Existing technologies typically present all data (video, temperature, and crying sounds) together to everyone. This invention stipulates that only radar and video are responsible for "life-saving" (judgment), while temperature, humidity, and crying sounds are responsible for "experience" (display). It clearly defines that data from the "auxiliary sensing module" does not participate in bimodal judgment, but emphasizes "bimodal accuracy, auxiliary data for display."

[0172] In this embodiment of the invention, the system is only used for daily safety monitoring and risk warning of newborns, and does not involve medical-related functions such as disease diagnosis, clinical monitoring of physiological indicators, or medical treatment. It is a non-medical safety auxiliary monitoring product. To achieve its non-medical attributes, the system design proactively omits medical-grade high-precision sensors (such as medical-grade pulse oximeters) and complex pathological algorithms. For example, millimeter-wave radar is only used to monitor the presence or absence of breathing (safety dimension), rather than calculating precise respiratory gas concentration (medical dimension). This "simplification" design for B2B commercial scenarios is something that those skilled in the art would not easily conceive of when dealing with medical-grade devices.

[0173] In another embodiment of the present invention, a non-medical newborn safety monitoring method for maternal and infant care institutions is also provided, applicable to the system of any of the foregoing embodiments, such as... Figure 2 As shown, it includes the following steps:

[0174] Step S10: Data acquisition, synchronously acquiring radar vital signs data and video monitoring status and environmental data through the bed-side front-end monitoring terminal;

[0175] Step S20: Local cross-validation. On the bed-side front-end monitoring terminal, cross-validation is performed on the vital signs dimension data and status and environment dimension data within the same judgment time window. Only when both types of data are judged to have safety risks, a valid alarm signal is generated, and the bed-side front-end monitoring terminal has no audible or visual prompts throughout the process.

[0176] Step S30: Alarm distribution, push valid alarm signals to the nurse station centralized management platform and the parent-child tiered viewing terminal via wireless LAN;

[0177] Step S40: Data management step, where monitoring data is stored locally and selectively synchronized to the cloud based on network status to enable data traceability and remote viewing.

[0178] The hardware environment for implementing this method must include the aforementioned system, namely: a three-terminal architecture: a bedside front-end monitoring terminal, a nurse station centralized management platform, and a parent-level viewing terminal; a networking method: the three are connected in a distributed manner through the institution's wireless local area network; front-end terminal features: integrated millimeter-wave radar + video dual-modal sensing, passive audio-visual design, and physical separation from the parent terminal.

[0179] Preferably, such as Figure 4 As shown, the above data management steps also include network disaster recovery and dual-stack operation mode switching steps:

[0180] Network status monitoring sub-step S41: Real-time monitoring of external network connection status;

[0181] Cloud synchronization sub-step S42: When the external network connection is normal, the effective alarm signals and monitoring record data will be synchronized to the cloud server, and data access will be supported by remote mobile viewing terminals;

[0182] Local backup step S43: When an external network interruption is detected, the system automatically switches to a pure LAN operating mode, blocks the access permissions of the remote mobile viewing terminal, and maintains the dual-modal joint judgment logic of the bed-side front-end monitoring terminal and the alarm distribution function within the LAN to ensure that the core monitoring service is not interrupted.

[0183] Existing infant monitoring methods typically rely heavily on cloud processing (AI identification and alerts are handled in the cloud). This method emphasizes a layered processing approach of "local edge computing (dual-modal judgment) + LAN transmission + cloud backup," clearly defining the logic of "switching upon network disconnection and blocking remote access upon network disconnection," reflecting a non-obvious design tailored to the security needs of B-end institutions.

[0184] In summary, the invention addresses the shortcomings of existing B2B maternal and infant care institutions that rely on passive video viewing for newborn monitoring and lack adequate security safeguards. It achieves proactive and accurate monitoring and early warning of abnormal newborn conditions through dual-modal cross-validation; avoids the high false alarm rate of single-sensor monitoring, ensuring uninterrupted daily operations; eliminates noise and light pollution from monitoring equipment, creating a undisturbed environment suitable for newborns and mothers; constructs a tiered access control mechanism to balance parental monitoring needs with compliance requirements for infant privacy management; enables collaborative monitoring between institutions and parents, and full-process data retention and traceability, improving the digital operation and management system of maternal and infant care institutions; and establishes a local area network-based monitoring architecture to ensure uninterrupted core monitoring functions even during external network outages. It should be noted that the system of this invention is explicitly defined as a non-medical product and does not involve medical functions such as disease diagnosis or clinical monitoring of physiological indicators.

[0185] It should be understood that those skilled in the art can make improvements or modifications based on the above description, and all such improvements and modifications should fall within the protection scope of the appended claims.

Claims

1. A multimodal fusion neonatal safety monitoring system for maternal and infant care institutions, characterized in that, include: Bedside monitoring terminals and centralized management platform for nurse stations; The bedside monitoring terminal is deployed next to the neonatal care bed and includes: The radar monitoring module is used to monitor the respiratory signs and body movement of newborns. The video monitoring module is used to identify the newborn's posture and the condition of the surrounding coverings; The bed-side monitoring terminal is connected to the centralized management platform at the nurse station via the institution's wireless local area network; The bedside front-end monitoring terminal is configured to execute dual-modal joint judgment logic: within the same time window, it performs time-series matching and feature comparison on radar monitoring data and video monitoring data, and generates a valid alarm signal only when both types of data are determined to be in an abnormal state. The bedside monitoring terminal adopts a passive audio-visual design, and the system is configured to push effective alarm signals to a remote nurse station centralized management platform via a local area network for notification.

2. The system according to claim 1, characterized in that, The radar monitoring module is configured to monitor the respiratory rate and body movement amplitude of newborns. The video monitoring module is configured to identify the newborn's posture, surrounding coverings, and limb compression environment data. In the dual-modal joint judgment logic, abnormalities in the vital signs dimension and abnormalities in the environmental dimension must match within the same judgment time window in order to be judged as valid security risks.

3. The system according to claim 2, characterized in that, The dual-modal joint determination logic specifically includes: The radar monitoring data is processed to extract respiratory rate and body movement amplitude features. If the features deviate from the preset safety threshold, they are marked as abnormal in the vital signs dimension. Image recognition is performed on video frame data to extract features such as infant position, facial coverings, and limb compression. If the features match the preset risk model, they are marked as abnormal in the environmental dimension. The timestamps of abnormalities in the physical and environmental dimensions are synchronized and calibrated. If the two overlap within the time window T, it is considered a valid alarm; if only a single module is abnormal, it is considered interference and filtered out.

4. The system according to claim 1, characterized in that, The wireless LAN distributed networking architecture supports dual-mode operation: "local backup" and "cloud synchronization". When the external network connection is normal, the system will synchronize alarm data and monitoring records to the cloud server; When the external network is interrupted, the system automatically switches to pure local area network mode. The bedside front-end monitoring terminal still executes the dual-modal joint judgment logic and pushes the effective alarm signal to the nurse station centralized management platform through the local area network to ensure that the core monitoring function is uninterrupted.

5. The system according to claim 1, characterized in that, It also includes a parent-child tiered monitoring terminal, which is physically separated from the bed-side front-end monitoring terminal and can be started and stopped independently; The system also includes a tiered, differentiated alarm mechanism: When the centralized management platform of the nurse station receives a valid alarm signal, it triggers an audible and visual alert and a pop-up notification. When the parent-child tiered viewing terminal receives a valid alarm signal, it triggers a silent vibration or a pop-up notification on the interface, without emitting an alarm sound that penetrates the room.

6. The system according to claim 5, characterized in that, The parental tiered viewing terminal includes a local wireless display terminal and a remote mobile viewing terminal, and the system is configured with a tiered access control module. The local wireless display terminal is deployed in a parent-only area, allowing parents to view real-time status data on-site without authorization. The remote mobile viewing terminal can only access data after identity binding verification and privacy authorization agreement, and remote access is automatically prohibited when the external network is interrupted.

7. The system according to claim 5, characterized in that, It also includes an auxiliary sensing module, which includes a temperature and humidity sensor and a crying sound acquisition module; The system is configured with a heterogeneous data distribution and view isolation module: The data collected by the auxiliary sensing module does not participate in the dual-modal joint determination logic, but is only used for state visualization. The centralized management platform for nurse stations hides the auxiliary sensor data by default and only displays core alarm information. It supports manual viewing to assist in handling the situation. The parent-child tiered viewing terminal displays all auxiliary sensor data by default.

8. The system according to claim 5, characterized in that, The system also includes a heterogeneous data distribution and view isolation module, which is configured as follows: The collected data is divided into core alarm data and auxiliary environment data. The core alarm data is generated by the radar monitoring module and the video monitoring module and participates in the dual-modal joint determination. The auxiliary environment data is generated by the auxiliary sensing module and does not participate in the dual-modal joint determination logic. The centralized management platform at the nurse station distributes and displays only the core alarm data by default to support nursing staff in quickly locating risks; the parent-level viewing terminal distributes and displays the full amount of core alarm data and auxiliary environment data by default to improve parents' perception of the nursing environment.

9. A non-medical newborn safety monitoring method for maternal and infant care institutions, applied to the system described in claim 1, characterized in that, Includes the following steps: Data acquisition steps: Simultaneously acquire radar monitoring data and video monitoring data through the bedside monitoring terminal; Cross-validation steps: Spatiotemporal synchronization of dual-modal data is performed in the local processor. A valid alarm signal is generated only when both types of data are determined to be abnormal within the same time window. During this process, the front end of the bed remains silent and without light output. Alarm distribution steps: Valid alarm signals are pushed to the centralized management platform of the nurse station via wireless LAN to trigger audible and visual alerts; Data management steps: Store monitoring data locally via the local area network and selectively upload it to the cloud based on network conditions.

10. The neonatal safety monitoring method according to claim 9, characterized in that, The data management steps also include network disaster recovery and dual-stack operation mode switching steps: Network status monitoring sub-step: Real-time monitoring of external network connection status; Cloud synchronization sub-step: When the external network connection is normal, the effective alarm signals and monitoring record data will be synchronized to the cloud server, and data access will be supported by remote mobile viewing terminals; Local backup steps: When an external network interruption is detected, the system automatically switches to a pure LAN operating mode, blocks the access permissions of remote mobile viewing terminals, and maintains the dual-modal joint judgment logic of the bed-side front-end monitoring terminal and the alarm distribution function within the LAN to ensure that core monitoring services are not interrupted.

Images