A human perception false alarm filtering system, a false alarm prevention system, an alarm cancellation system and an old person emergency rescue control system
Patent Information
- Application Number
- CN202611050920.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Priority Date
- 2026-05-04
- Filing Date
- 2026-07-15
- Publication Date
- 2026-09-15
Smart Images
Figure CN122761531A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of civilian intelligent security technology, specifically a false alarm filtering system based on human perception, a system to prevent missed alarms, an alarm cancellation system, and a control system for emergency rescue of the elderly. Background Technology
[0002] As society ages, home safety for elderly people living alone or in empty nests has become a major public concern. Elderly individuals who fall or suffer sudden medical emergencies are often unable to call for help independently, and rescuers may have difficulty quickly reaching their homes, easily missing crucial rescue time. Currently available mainstream home health monitoring and emergency alarm devices suffer from several inherent and difficult-to-resolve industry flaws: 1. The accuracy of single sensing devices is poor, and false alarms and missed alarms are prominent problems. Existing monitoring equipment mostly uses single hardware to collect data, such as cameras, millimeter-wave radar, ground pressure sensors, and wearable wristbands. Each type of equipment has its own shortcomings in recognition: millimeter-wave radar is easily interfered with by pets walking, curtains swaying, and furniture moving, resulting in false alarms. It also fails to recognize non-standard falling postures such as slow sitting, fainting, and leaning, leading to missed reports; visual cameras have significantly reduced recognition accuracy due to backlighting, darkness at night, clothing obstruction, and interference from television screens; ground pressure sensors can only recognize falls caused by heavy pressure and cannot monitor situations where the patient falls while lying in bed or in a chair; wearable devices rely on the elderly wearing them for a long time, and become completely ineffective after being dropped or when the power is cut off.
[0003] Existing equipment relies solely on raw sensor data to determine potential hazards, lacking a local voice interaction confirmation process. Suspected anomalies are directly pushed with alarm messages, leading to frequent false alarms and disturbances to residents. Furthermore, the equipment only passively monitors falling actions and cannot identify emergencies such as myocardial infarction, stroke, or indoor harassment without falls, posing a serious risk of missed reports. It also lacks a complete local autonomous filtering and fault-tolerant handling logic.
[0004] 2. The system is highly dependent on the external network and cloud backend; the entire machine will be completely paralyzed in the event of a network outage or power failure. Currently, mainstream elderly care monitoring systems rely entirely on cloud servers and home broadband networks for anomaly detection, alarm notifications, and electric lock authorization. The entire control logic cannot operate independently locally. Once the home broadband or mains power is interrupted, the functions of human body sensing, voice interaction, emergency unlocking, and alarm notifications all fail. The main control equipment, cameras, radar, human-machine interaction terminals, and other terminals lack independent power supply, and in extreme failure scenarios, they completely lose their rescue capabilities.
[0005] 3. Fragmented interaction logic and poor linkage effect. The existing products have fragmented functions. After the sensors identify a suspected danger, there is no local confirmation process. If the alarm is misjudged, it will directly push an alarm to the outside. The products rely solely on passive perception to identify falls and do not have an independent voice call channel to deal with non-fall emergencies. Once the alarm process is initiated, there is no local cancellation mechanism. If the elderly misjudge or the alarm is triggered by environmental noise, the products can only wait for remote manual handling, and the process cannot be closed locally.
[0006] 4. The emergency unlocking mode is limited, making it difficult to simultaneously address both home security and rapid on-site rescue. Traditional smart emergency electric locks only have two control logics: one is to automatically open the door after detecting an anomaly, which can lead to the door remaining open for extended periods during false alarms, posing a significant security risk to homes. This unlocking method is a one-time unlock, and the door cannot be reopened if external forces such as wind close the lock. The other relies on a temporary unlocking password sent from the cloud, which can be severely hampered by network latency or outages. Both solutions lack a comprehensive temporary access mechanism, preventing neighbors and passersby from entering to provide on-site first aid and missing a crucial window for rescue. Furthermore, there is no independent hardware backup unlocking channel, meaning the door cannot be opened after a malfunction in the electric lock's control module or a power outage. Additionally, all unlocking logic lacks a unified and adjustable delay buffer period, making it unsuitable for rapid rescue of high-risk elderly individuals or for allowing younger, healthy elderly individuals to decide whether to open the door, thus failing to meet the diverse needs of accidental unlocking. It also does not support the combined use of multiple unlocking modes, resulting in significant limitations in its application scenarios.
[0007] 5. Existing equipment cannot complete the closed-loop on-site rescue in extreme scenarios of triple offline conditions such as power outage, broadband outage, and no cellular base station signal. Most devices only rely on cellular or broadband to push alarms outward. Once the base station signal is lost or the broadband is disconnected, they cannot automatically unlock the door or call neighbors with a high-power external loudspeaker. Even if a few devices have backup batteries, they can only provide simple local alarms and lack an autonomous secondary verification mechanism. They will not actively open the door to facilitate emergency rescue for passersby. In extreme failure environments, they cannot achieve the first-time emergency rescue of "someone knows, someone enters, and someone rescues". Summary of the Invention
[0008] To address the shortcomings of existing elderly care rescue equipment, such as severe false alarms and missed alarms, heavy reliance on cloud networks, failure of rescue functions after power outages, network outages, and base station failures, and inability to trigger local neighbor assistance, this invention provides a human-sensory false alarm filtering system, a missed alarm prevention system, an alarm cancellation system, and a control system for emergency rescue of the elderly. It constructs a three-level linked human-machine interactive closed loop of "false alarm filtering - missed alarm prevention - cancellation," with the entire sensing, judgment, confirmation, and execution process performed locally offline autonomously, without the need for cloud, backend, or external communication support. The system is equipped with an independent backup battery unit to achieve self-sustaining operation even in the absence of broadband or cellular base station signals. It can still operate fully, including autonomous hazard assessment, secondary voice confirmation, automatic door unlocking, and on-site warning to neighbors via an external loudspeaker, enabling the machine to self-judge, self-confirm, and autonomously initiate local neighbor assistance even in extreme three-disruption conditions.
[0009] This invention's system is lightweight and easily deployable. It requires only a human body state sensing unit and an alarm output unit to form an independent rescue terminal, eliminating the need for peripherals such as electric lock execution units and central control units. Relying solely on millimeter-wave radar for independent local computation and judgment, and equipped with an independent backup battery for self-sufficiency in the event of power outages, it can still complete the entire rescue process—including human anomaly recognition, secondary voice confirmation, and external loudspeaker broadcasting to neighbors—even in offline scenarios with no broadband or cellular base station signal. This low-cost hardware solution enables on-site emergency rescue in extreme conditions, making it suitable for lightweight applications such as older buildings, small apartments, and mass deployment for affordable elderly care. The technical solution adopted by this invention is as follows: A false alarm filtering system for human body perception includes a central processing unit, a human body state perception unit, and a human-computer interaction unit. The central processing unit communicates bidirectionally with the human body state perception unit and the human-computer interaction unit. The central processing unit has built-in judgment logic, performs independent local data processing, and autonomously completes anomaly judgment and / or process control, confirmation, and / or alarm triggering. The human body state perception unit collects one or more of the following data in real time: human posture, activity status, and vital signs of the target object, and transmits them to the central processing unit. The central processing unit analyzes data to identify suspected abnormal events. When a suspected abnormality is determined, it feeds back to the human-machine interaction unit and executes the corresponding operation: if it is a false alarm cancellation command, the false alarm is immediately terminated and the system returns to normal; if it is an emergency distress command or there is no effective feedback, it is determined to be a real event and the system triggers an alarm.
[0010] A human-sensing anti-missed alarm system includes a central processing unit and a human-machine interaction unit. The central processing unit and the human-machine interaction unit communicate bidirectionally. The central processing unit has built-in judgment logic, performs independent local data processing, and autonomously recognizes commands and triggers rescue actions. The human-machine interaction unit is equipped with a voice module and pre-stores and recognizes voice data. When the central processing unit receives valid voice data, it directly determines the actual emergency status and triggers an alarm.
[0011] A human-sensing alarm cancellation system includes a central processing unit and a human-machine interface unit (HMI). The HMI communicates bidirectionally with the central processing unit. The central processing unit has built-in judgment logic, performs independent local data processing, and autonomously recognizes commands, terminates alarms, and / or resets the system. The HMI is equipped with a voice module and pre-stores keywords for alarm cancellation. When the system triggers an alarm, if the HMI recognizes a valid voice command for alarm cancellation, the central processing unit immediately cancels the alarm, and the system returns to normal.
[0012] An emergency rescue control system for the elderly includes one or more of the aforementioned systems; when it includes all the systems of claims 1 to 3, it constitutes a three-level linkage closed-loop control system for local offline autonomous operation, including false alarm filtering, prevention of missed alarms, and alarm cancellation.
[0013] The aforementioned emergency rescue control system for the elderly includes an electric lock execution unit, which is equipped with an electric lock. After the emergency rescue process is triggered, the central processing unit opens the electric lock.
[0014] The aforementioned emergency rescue control system for the elderly also includes a warning output unit, which is equipped with a voice speaker. After the emergency rescue process is triggered, the central processing unit controls the playback of the voice message.
[0015] The aforementioned emergency rescue control system for the elderly is equipped with independent backup batteries for its central processing unit, human body status sensing unit, and human-computer interaction unit; the backup batteries automatically switch between mains power and mains power.
[0016] The aforementioned emergency rescue control system for the elderly includes an electronically controlled lock with three unlocking modes: Mode 1: Unlock directly with voice command When the emergency rescue process is not initiated, the electric lock remains in the normal locked state. After the central processing unit recognizes the unlocking voice command, it directly drives the electric lock to complete the automatic unlocking. Mode 2: Emergency unlocking via voice command When the emergency rescue process is not initiated, the electric lock remains in the normal locked state, and only pre-authorized personnel can unlock it. After the central processing unit recognizes the unlocking voice command, it removes the unlocking whitelist permission verification restriction, and any personnel can unlock it. Mode 3: Voice command, external independent power supply, manual emergency unlocking The electric lock is equipped with a manual emergency unlocking mechanism. When the emergency rescue process is not activated, the electric lock remains in the normal locked state. After the central processing unit recognizes the unlocking voice command, any person can unlock the electric lock by pressing the external manual emergency unlocking mechanism. The aforementioned emergency rescue control system for the elderly has a delay period for unlocking the electric lock. The delay threshold can be adjusted from 0 to a preset time. During the delay period, the entire emergency process can be terminated by a voice cancellation command. If the voice unlock command is recognized during the delay period or if no cancellation command is received when the delay period expires, the corresponding unlocking action will be automatically executed. At the same time as the unlocking action is executed, the voice speaker will be activated and information will be sent through the built-in cellular mobile communication module.
[0017] 1. When the delay threshold is greater than 0, the system enters the delay period: ① If the unlocking command is not recognized within the delay period, the warning output unit will not start, and the electric lock permission verification module will maintain the whitelist verification restriction; the command can be canceled by voice to terminate the emergency process, and the system will return to normal. ② During the delayed period, the voice command for unlocking is triggered, and the warning output unit is activated simultaneously to output a warning signal; Automatically unlock the electric lock, remove the whitelist restriction of the electric lock permission verification module, and open full access permission; remove the blocking restriction of the signal path corresponding to the external manual emergency unlocking trigger mechanism by the independent permission locking module, and the external manual emergency unlocking trigger mechanism becomes effective; ③ If no valid cancellation instruction is received by the end of the extension period, the warning signal will continue to be output through the warning output unit; Automatically unlocks the electric lock; maintains the whitelist verification restriction lifted and full access open; continuously removes the corresponding signal path blocking restriction, and keeps the external manual emergency unlocking trigger mechanism in effect; 2. When the delay threshold is equal to 0, there is no buffer delay period. After the emergency alarm is triggered, the warning output unit will be activated simultaneously to output a warning signal and the corresponding unlocking mode will be directly enabled: Automatically unlocks; removes whitelist verification restrictions; removes corresponding signal path blocking restrictions.
[0018] The aforementioned emergency rescue control system for the elderly includes a human state perception unit that is one or more of a visual acquisition device, a sensing radar device, a smart wearable device, or a sensing device.
[0019] The aforementioned emergency rescue control system for the elderly includes a privacy-blocking module in its visual acquisition device.
[0020] The aforementioned emergency rescue control system for the elderly also includes a central processing unit equipped with a local storage module and a communication module. The local storage module stores collected and processed data locally in a closed loop, and normal data is not uploaded to the cloud; the communication module is only activated when an emergency is triggered to push alarm information.
[0021] The beneficial effects of this application are: Firstly, false alarm prevention, missed alarm prevention, and alarm cancellation can all be embedded, in combination or individually, into human body sensing hardware, such as a single radar, camera, or sensor, to achieve complete rescue and reduce deployment costs; they can be sold as a complete set or sold separately.
[0022] Secondly, a three-level linkage human-computer interaction closed-loop control system of "false alarm-missed alarm-voice cancellation" is constructed. The entire judgment process can be completed locally offline without the need for manual intervention in the cloud, thereby reducing false alarms and eliminating missed alarms from the source. Third, the entire machine adopts local autonomous computing and is equipped with an independent backup power supply battery. Under extreme conditions such as power outages and wired broadband outages, the human body sensing, human-computer interaction, emergency unlocking, and alarm push functions can all operate normally. Fourth, three levels of controllable emergency unlocking modes, with corresponding delay control rules. The unlocking modes can be activated independently or in combination, depending on the scenario and the user's wishes, to balance home security and rapid on-site rescue, and adapt to various usage needs. Fifth, the system possesses triple offline functionality in the event of power outages, broadband outages, and base station outages. Each unit's independent backup battery automatically switches to mains power. Even without external power supply, wired network, or cellular communication signal, human body sensing, human-machine interaction verification, automatic unlocking of electric locks, and external high-volume warning horns all function normally, enabling complete local rescue operations without relying on any external communication channels. In the most extreme triple offline scenarios, it achieves a backup rescue capability of "recognition, entry, and rescue."
[0023] Sixth, once the emergency is confirmed locally, an external loudspeaker will be activated to broadcast a distress message in a loop, directly alerting neighbors and passersby to come to the nearest door for assistance. In conjunction with an automatic full-area unlocking mechanism, a closed loop of "identifying the emergency - confirming the risk - opening the door - providing on-site assistance to neighbors" will be established, enabling immediate on-site rescue and making up for the shortcomings of remote alarms that delay family members and emergency personnel, thus missing the golden rescue time.
[0024] Seventh, the product form is flexible and layered. A minimalist split model with only millimeter-wave radar and an external high-frequency speaker can be launched. It can realize three-way offline neighborhood rescue without the need for an electric lock, realizing "someone knows and someone rescues". The hardware cost is low and the installation is convenient. It is suitable for old houses, small apartments, and mass purchase scenarios for inclusive elderly care, covering user groups with different budgets and different home security needs. Attached Figure Description
[0025] Figure 1 This is a flowchart of an embodiment of the present invention. Detailed Implementation
[0026] The technical solution of the present invention will be described in detail below with reference to the accompanying drawings. The following embodiments are illustrative; unless otherwise specifically stated, the relative arrangement and numerical expressions of components and steps described in these embodiments should not be construed as limiting the scope of this application. The following description of exemplary embodiments is merely illustrative and is not intended to limit the scope of this application or its application or use in any way. Techniques, methods, and apparatus known to those skilled in the art may not be discussed in detail here, such as switching between mains power and battery power, the installation and use of radar cameras or various sensors for sensing the human body, etc. However, where these techniques, methods, and apparatuses are applicable, they should be considered part of this specification.
[0027] It should be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the exemplary embodiments according to this application. Unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. When the terms "comprising" and / or "including" are used in this specification, they indicate the presence of features, steps, operations, devices, components, and / or combinations thereof. Unless otherwise stated, the terms "installed," "connected," "linked," etc., in this application should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral part; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to an internal connection between two elements or an interaction between two elements. Those skilled in the art can understand the specific meaning of the above terms in this application according to the specific circumstances.
[0028] For ease of description, the use of terms such as "upper," "lower," "left," "right," "bottom," and "top" in this application does not limit the structure. It is merely for the purpose of understanding the structural principles of this application in conjunction with the accompanying drawings, and does not indicate or imply that the equipment or components referred to must have a specific orientation, or be constructed and operated in a specific orientation. It should not be construed as a limitation of this application.
[0029] This embodiment is an emergency rescue control system for the elderly, including a central processing unit, a human body status sensing unit, a human-computer interaction unit, an electric lock execution unit, and an alarm output unit.
[0030] The central processing unit is an edge computing box, smart gateway, or industrial control host, etc., and is equipped with multiple types of wired and wireless communication interfaces to realize data interaction and command issuance with each unit. The system as a whole adopts a local autonomous computing mode. The entire process of human anomaly detection (such as using the YOLOv8 human posture algorithm), voice interaction verification, and electric lock linkage command issuance does not require external network, cloud platform, or manual intervention in the background. The central processing unit has a built-in local storage module and mobile communication module. Daily monitoring and computing data are only stored locally in a closed loop. The mobile communication module is activated to push alarm information outward only when a real emergency event is identified.
[0031] The human body state sensing unit collects real-time data on the target object's posture, activity status, and vital signs, and transmits it to the central processing unit. The human body state sensing unit is selected from one or more of the following: a visual acquisition device camera, a human body sensing radar, a ground pressure sensor, a wearable sensor, or a fixed vital signs sensor. It is deployed in a distributed manner throughout the entire area to eliminate blind spots in indoor detection. The visual acquisition device is equipped with a privacy shielding module, supporting manual real-time shielding and timed automatic shielding modes. In the shielding state, image output and storage cease, while the background human posture and anomaly detection models continue to run. The visual acquisition device is equipped with a privacy shielding module, which can achieve: ① permanent shielding; such as permanent shielding in the bathroom, while the background continues to monitor and collect human body data (manually turned on / off); ② timed shutdown; such as in the bedroom, where the timed shielding period can be set according to the elderly person's daily habits, cycling daily. For example, if the elderly person wakes up at 8 am and rests at 9 pm, the effective shielding time is from 8 am to 9 pm, while the background continues to monitor and collect human body data (the shielding period can be set and controlled by the user).
[0032] The human-machine interface unit can be distributed throughout the room, allowing users to conveniently perform voice inquiries, emergency calls, and false alarm cancellations from any location within the room. The human-machine interface unit can be fully integrated into the central processing unit, or it can be deployed separately and integrated into sensing devices such as visual cameras, millimeter-wave radar, and ground pressure sensors. Based on the human-machine interface unit, a three-level linkage human-machine interaction closed loop is built, forming three independently operating sub-control systems. ① False Alarm Filtering: After the human body state perception unit identifies a suspected abnormality, the human-computer interaction unit initiates a voice confirmation inquiry. If a false alarm cancellation command is received (e.g., keywords such as "It's okay," "It's a mistake," etc.), the current abnormality processing is immediately terminated, and the system returns to normal. If an emergency distress command is received (e.g., keywords such as "I fell," "Call the police," "Help," etc.), then a genuine abnormality is determined, and the alarm and emergency handling procedures are triggered. If there is no effective feedback within a preset time, a genuine abnormality is determined, and the alarm and emergency handling procedures are directly triggered (if it is determined to be an elderly person's disability, coma, etc., the delay period is skipped, and the door is opened directly with the loudspeaker alarm). ② Proactive emergency response to prevent missed reports: Set up an independent voice emergency trigger channel, recognize the preset call wake-up word plus keywords (e.g., wake-up word + keywords, Xiaozhi, Xiaozhi, I fell down or Xiaozhi, Xiaozhi, I want to call the police, etc.), directly skip the human body perception detection and suspected abnormal interaction confirmation process, determine it as a real emergency event and immediately start the rescue process; ③ Cancel alarm: Alarms can be canceled via voice commands in any state. Keywords can include "turn off alarm", "turn off system", and "cancel alarm". The central processing unit can recognize a valid cancellation command (such as "turn off alarm") to cancel all alarms and immediately reset to the normal state.
[0033] The electronic lock execution unit integrates three types of independent emergency unlocking execution logic, corresponding to three unlocking modes: Mode 1: Automatic unlocking via voice command When the emergency rescue process is not initiated, the electric lock remains in normal operation. After the central processing unit recognizes the unlocking voice command, it directly drives the electric lock to complete the automatic unlocking. Mode 2: Voice command-based emergency access unlocking mode When the emergency rescue process is not initiated, the electric lock maintains its normal anti-theft status. Only pre-authorized personnel can unlock it by fingerprint, password, or face. After the central processing unit recognizes the unlocking voice command, it initiates the removal of the whitelist permission verification restriction of the electric lock, and any person can unlock it by fingerprint, password, or face verification. Mode 3: Voice command-based external independent power supply manual emergency unlocking mechanism for emergency unlocking. When the emergency rescue process is not initiated, the internal signal path is blocked and the external manual emergency unlocking mechanism fails. After the central processing unit recognizes the unlocking voice command, the path blockage is lifted, and any person can unlock the lock by pressing the external manual emergency unlocking mechanism. The external manual emergency unlocking mechanism can be used alone or as a hardware unlocking channel in the event of an electric lock malfunction or power failure.
[0034] The three unlocking modes can be used individually or in combination to adapt to different emergency scenarios. All three modes share a unified delay control rule, with the delay threshold adjustable from 0 to a custom preset duration. During the delay period, all emergency procedures can be terminated via voice cancellation. If an unlocking voice command is recognized within the delay period or if no cancellation command is received by the end of the delay, the corresponding unlocking action will be automatically executed. Simultaneously, a loudspeaker outside the door will be activated to broadcast a continuous distress call, and an alarm message will be immediately pushed to family members, the community, and emergency centers via the built-in 4G cellular mobile communication module.
[0035] The central processing unit, human status sensing unit, and human-machine interaction unit within the system are each independently equipped with a dedicated backup power supply. The backup power supply can automatically switch from the mains power supply. When the mains power is disconnected, it automatically switches to the backup power supply to ensure the continuous operation of all monitoring, interaction, alarm, and unlocking functions of the system.
[0036] Example 1: Mr. A, 75 years old, has the above-mentioned control system installed in his two-bedroom, one-living-room house.
[0037] The human body state perception unit of the Scenario 1.1 system supports flexible deployment: it can use only a single millimeter-wave radar device, only a single visual camera device, or a combination of millimeter-wave radar + visual camera + ground pressure sensor + wearable device, etc., which can be freely matched according to budget, privacy requirements and detection accuracy requirements.
[0038] The entire house is equipped with millimeter-wave radar, bedroom vision cameras, and bathroom floor pressure sensors, along with distributed voice speakers on the walls, electronic locks with three unlocking modes, external high-volume warning speakers on the doors, independent backup lithium batteries for each device, an edge gateway, local storage, and a 4G cellular communication module. The system delay is uniformly set to 30 seconds, with preset voice commands: inquiry "Have you fallen? Do you need help?"; confirmation of help "Yes, I have fallen"; wake-up word "Xiao Zhi Xiao Zhi"; door opening command "Open door for rescue"; and cancellation command "Wrong / It's okay, turn off the alarm".
[0039] After the system detects a suspected anomaly and A confirms the danger via voice, a 30-second delay period is initiated. During this period, the electric lock's execution unit, the external loudspeaker, and the 4G communication module are all in standby mode. The electric lock maintains its normal anti-theft status, the loudspeaker does not broadcast any alarms, and the communication module does not push any alarms. During the delay period, A can actively issue an opening command to immediately trigger the corresponding unlocking mode. If A neither responds to the opening command nor issues a cancellation command by the end of the 30-second delay, the system automatically executes the corresponding unlocking mode. Simultaneously, the external loudspeaker is activated to continuously broadcast a distress call, and alarm information is immediately pushed to family members, community management, and the nearest emergency medical services via the built-in 4G cellular mobile communication module. Each unlocking mode can be activated independently, or two or three modes can be combined and used in conjunction to adapt to different emergency scenarios.
[0040] Scenario 1.2 Automatic unlocking via voice command Person A slipped and fell in the living room. Millimeter-wave radar detected a significant fall, classifying it as a possible anomaly. The nearest voice assistant in the living room automatically played a voice prompt: "Did you fall? Do you need help?" Person A was conscious and clearly replied, "Yes, I fell down." The system then entered a 30-second delay period. During this period, the electric lock remained locked, and the loudspeaker outside the door and the 4G alarm push were both in standby mode, without making any external noise or pushing any information.
[0041] There are three trigger states: Firstly, proactive voice triggering: 8 seconds into the delay period, A utters the command to open the door: "Open the door for rescue." The system immediately activates the electric lock, continuously unlocking at 5-second intervals until the door is open. Simultaneously, a loudspeaker outside the door plays a loop of emergency messages, such as, "An elderly person in this household has fallen. Please have neighbors or passersby enter the house immediately to help!" Upon hearing the warning, neighbors enter the house directly to check on A's condition and contact family members. At the same time, the 4G cellular module is immediately activated, sending alarm alerts containing the house address to family members' mobile phones, community grid workers, and the nearest community health center.
[0042] Secondly, automatic triggering upon timeout: If the 30-second delay period expires and A is still unable to speak, and neither issues an opening command nor a cancellation command, the system will automatically execute the electric lock unlocking action, and simultaneously activate the loudspeaker outside the door to repeatedly broadcast a distress message such as "An elderly person in this household has fallen. Please ask neighbors and passersby to come in and help immediately!" After hearing the warning sound, neighbors will directly enter the house, check A's condition immediately, and contact family members. The 4G cellular module will simultaneously push multiple alarm messages.
[0043] Third, cancel the alarm: If A does not want to disturb the neighbors and wants to notify his family or medical institution himself, A shouts "Turn off the alarm" during the delay period. The central processing unit deactivates the electric lock and external loudspeaker from their standby state, and the system returns to normal sensing and detection state.
[0044] Scenario 1.3: Global Emergency Response Mode When A got up in the bedroom, he felt dizzy and fell. He was too weak to speak. The millimeter-wave radar in the bedroom detected his fall and initiated a voice inquiry. A weakly replied to confirm the fall, and the system entered a 30-second delay period. During the delay period, the electric lock maintained its normal anti-theft status with whitelist verification, and the loudspeaker outside the door and 4G alarm push notifications were on standby but not activated.
[0045] There are three trigger states: Firstly, proactive voice triggering: During the delay period, when A utters the command "Open the door for rescue," the system immediately removes the whitelist permission verification restriction of the electric lock. It is no longer limited to family members' fingerprints, faces, and passwords. Neighbors, passersby, and emergency personnel can all open the door through fingerprint, password, or face verification. Simultaneously, a loudspeaker outside the door plays a loop of the distress message: "An elderly person in this household has fallen. Please come in and help immediately! The electric lock can be opened by anyone using fingerprint, password, or face." After hearing the warning sound, the neighbor immediately enters the house through fingerprint to check A's condition and contact family members. The 4G cellular module simultaneously pushes alarm information to multiple parties.
[0046] Secondly, automatic triggering upon timeout: If the 30-second delay period expires and A does not provide any effective voice response or cancel the alarm, the system automatically removes the whitelist restriction on the electric lock. It will no longer be limited to family members' fingerprints, faces, and passwords; neighbors, passersby, and emergency personnel can all open the door using fingerprint, password, or face verification. Simultaneously, a loudspeaker outside the door will continuously play a distress message: "An elderly person in this household has fallen. Please, neighbors and passersby, enter the house immediately to help! The electric lock can be unlocked by fingerprint, password, or face." Upon hearing the warning, neighbors will immediately enter the house using their fingerprints to check A's condition and contact family members. The 4G cellular module will simultaneously push alarm information to multiple parties.
[0047] Third, cancel the alarm: If Mr. A does not want to disturb the neighbors and wants to notify his family or medical institution himself, the elderly man shouts "Turn off the alarm" within the extended period. The central processing unit deactivates the electric lock and the external loudspeaker from their standby state, and the system returns to normal sensing and detection state.
[0048] Scenario 1.4: Manual Emergency Unlock Mode A fell in the bathroom. After the bathroom radar detected the anomaly and confirmed the danger, the electronic lock's mainboard suddenly malfunctioned, causing the electronic unlocking function to fail and the system to enter a 30-second delay period. During this delay, the internal and external manual mechanisms maintained signal blockage, and the loudspeaker outside the door and 4G alarm push notifications were all in a standby silent state.
[0049] There are three trigger states: Firstly, proactive voice triggering: During the delay period, when A utters the command "Open the door for rescue," the system immediately releases the signal blockage of the external manual emergency unlocking mechanism, and the pressing mechanism enters the operable state. Simultaneously, the loudspeaker outside the door continuously broadcasts a distress message: "An elderly person in this household has fallen. Please have neighbors or passersby enter the door to help immediately! The emergency button outside the door is activated and can be pressed manually to open the door." After hearing the warning sound, neighbors immediately press the emergency button to unlock the door and enter the house, check A's condition immediately, and contact family members. The 4G cellular module simultaneously pushes multiple alarm information.
[0050] Secondly, automatic triggering upon timeout: If the 30-second delay period expires and A does not respond effectively or cancel the alarm, the system automatically releases the signal path blockage of the external manual mechanism, the pressing mechanism enters the operable state, and the loudspeaker outside the door simultaneously starts to broadcast a help message in a loop: "An elderly person in this household has fallen. Please come in and help as soon as possible if you are a neighbor or passerby! The emergency button outside the door is activated and can be pressed manually to open the door." After hearing the warning sound, the neighbor immediately presses the emergency button to unlock the door and enter the house, checks A's condition and contacts the family members. The 4G cellular module simultaneously pushes multiple alarms.
[0051] Third, cancel the alarm: If A does not want to disturb the neighbors and wants to notify his family or medical institution himself, A shouts "Turn off the alarm" during the delay period. The central processing unit deactivates the electric lock and external loudspeaker from their standby state, and the system returns to normal sensing and detection state.
[0052] Example 2: Active voice distress call to bypass detection and prevent missed detections in sensing blind spots / device obstruction. Mr. B suffered a sudden heart attack while sitting on the living room sofa. He did not fall to the ground. The living room curtains blocked the camera, and neither the visual nor radar equipment detected any abnormality in the human body. This is a typical scenario of missed detection.
[0053] B experienced severe chest pain and uttered a wake-up call followed by a distress command: "Xiao Zhi, Xiao Zhi, I've fallen down."
[0054] The system recognizes independent, proactive distress commands, directly skipping the millimeter-wave and camera-based human detection steps, as well as the pre-confirmation voice inquiry and 30-second delay, initiating emergency response without any buffer: The unlocking mode can be activated in any way, or in combination of two or three modes; the unlocking action is executed synchronously, the loudspeaker outside the door continuously broadcasts help requests, and the 4G cellular network immediately pushes alarm information to multiple parties. Even if the home broadband is interrupted at this moment, the cellular module is not affected, and the alarm is delivered to the family and the community in a normal manner.
[0055] The entire process requires no cloud server or human back-end verification; it is executed automatically and locally.
[0056] Example 3: Suspected fall triggers alarm, accidentally shouts out alarm, voice cancellation and reset within the delay period: When Mr. C was tidying up the clutter on the balcony, his large bending motion was detected by millimeter-wave radar as a possible fall. The nearest human-computer interaction unit then initiated an inquiry: "Mr. C, have you fallen? Do you need help?" In a moment of panic, C accidentally uttered the keyword "call for help." The system then entered a 30-second unlocking buffer period. During this period, the electric lock remained locked, the loudspeaker outside the door and the 4G module were all muted, and no external announcements or alarms were sent.
[0057] Ten seconds later, C realized what was happening and immediately gave the cancellation command: "Wrong, turn off the alarm."
[0058] The central processing unit recognizes the valid voice command to cancel, instantly terminates all preparatory states, and the system directly returns to the normal human body monitoring mode. Throughout the process, the electric lock is not triggered to unlock, the loud announcement is not activated, no alarm information is sent to family members, and no disturbance records are stored in the cloud.
[0059] Example 4: Emergency Rescue for Triple Offline Scenario of Whole House Power Outage, Broadband Outage, and Cellular Base Station Disconnection In the early hours of the morning, a power distribution failure in the residential area caused a complete power outage for the entire house. Simultaneously, a fault in the operator's line disconnected the home's broadband, and a nearby base station malfunction resulted in no cellular signal, leaving the device completely disconnected from external power and network connections. Mr. D tripped over his slippers while trying to get up in the middle of the night and was unable to get up.
[0060] Each unit of the system automatically switches to independent backup battery power. The central processing unit, human status sensing units such as millimeter-wave radar sensing units, human-machine interaction units such as distributed voice interaction units, electric locks, and loudspeakers outside the door all operate independently without relying on any external network, cloud, or base station signals.
[0061] The millimeter-wave radar in the bedroom continuously collects human posture data using a backup battery. Local edge computing identifies suspected abnormalities such as falls, and the nearest human-computer interaction unit automatically initiates a voice confirmation question, "D, have you fallen? Do you need help?" to complete the system's autonomous secondary confirmation process.
[0062] Scenario 4.1: The elderly person is conscious and confirms the emergency with a voice reply. The system then enters a customizable delay period. During the delay period, the elderly person issues a door-opening rescue command. The system automatically executes the unlocking action and simultaneously activates a high-powered external loudspeaker outside the door to continuously broadcast a distress call, alerting nearby neighbors and passersby to enter the door and provide immediate assistance. This achieves a closed-loop rescue system in the triple offline extreme scenario of "someone aware, someone enters, and someone rescues." Due to the base station being down, the 4G communication module cannot push remote messages, but the local door-opening + loudspeaker neighbor rescue closed loop remains unaffected, relying on nearby personnel to complete the first-time emergency rescue.
[0063] Scenario 4.2: An elderly person loses the ability to move and cannot speak. After the delay period expires, the system automatically determines the actual danger, automatically unlocks the door, and continuously activates the external loudspeaker to repeatedly warn nearby residents. After hearing the alarm, neighbors can directly open the door to enter and carry out first aid. The entire process of danger assessment, risk confirmation, door opening, and on-site warning is completed locally and autonomously by the device, without the involvement of external networks, power supply, or base stations.
[0064] Scenario 4.3: The emergency is triggered by mistake. The elderly person gives the cancellation command within the delay period. The system automatically terminates all emergency procedures, the electric lock remains locked, the loudspeaker does not start, and normal monitoring is directly restored.
[0065] Once mains power, broadband, and base station signals are restored, the device automatically switches back to mains power supply mode. After cellular communication returns to normal, subsequent emergencies can be simultaneously pushed to family members and the community for remote alarm.
[0066] Example 5: E, a single person living in a small apartment, installed millimeter-wave radar only in the bathroom.
[0067] Install an integrated millimeter-wave radar in the bathroom. The device has built-in sound pickup, broadcasting, independent backup battery, local storage, and a built-in 4G module. It has no central processing unit, no electric lock, and no external loudspeaker. False alarm filtering, missed alarm active SOS, voice cancellation, and a three-level human-machine interaction system are all built-in. All judgments are completed locally, without relying on broadband or cloud services.
[0068] ① False alarm filtering: When E bends down to wipe the floor, the radar detects a possible fall and asks a voice inquiry; E replies "I'm fine, it was a mistake", the abnormality is terminated directly on a standby basis, the record is kept locally, and no alarm is reported; ②Actively calling for help: E felt dizzy while using the toilet but did not fall down. The radar did not recognize him. E called out, "Xiao Zhi, Xiao Zhi, I fell down." The device skipped the inquiry and directly issued a loud alarm and pushed a message to his children via 4G. ③ Voice Cancellation: When E was watching TV at home, the TV voice accidentally triggered the alarm. The device entered a short delay period. E said "Turn off the alarm", and the whole machine returned to normal monitoring without any external alarm push.
[0069] The aforementioned equipment can be sold as a standalone product, without the need for peripheral devices such as a central control unit or electric locks. Even in the event of a power or network outage, it can operate a complete three-level interactive closed loop—false alarm filtering, active distress call, and voice cancellation—relying on its own backup battery. This effectively addresses the industry pain points of existing similar devices, such as failure upon power or network outages, high false alarm rates, and high missed alarm rates. Device E can autonomously control alarm triggering and cancellation, requiring no external network connection, cloud platform, or manual intervention. The entire process—sensing and data collection, logical judgment, interactive confirmation, alarm activation, and alarm deactivation—is completed independently by the device.
[0070] Example 6: Deploying a millimeter-wave radar unit, in a scenario where the network, power, and base station are disconnected, and there is offline neighborhood voice alert for rescue. This embodiment uses a single integrated main control millimeter-wave radar as an independent human status sensing unit. The device has a built-in audio pickup module, local storage module, and independent backup lithium battery, and is equipped with a high-power external loudspeaker. The entire device is an independent alarm unit and does not include an electric lock. False alarm filtering, active distress calls to prevent missed alarms, and voice alarm cancellation are all performed locally, without the need for central control, external network, or cloud intervention. When the device is offline due to three conditions: mains power failure, home broadband disconnection, and no nearby cellular base stations, it automatically operates using the backup battery and only uses the external loudspeaker to trigger an alarm.
[0071] This system has the resident's address, the elderly person's name F, and the family's contact number pre-stored locally. The external loudspeaker has a preset loop to broadcast the voice message: "Grandpa XX from apartment XX in XX community has fallen at home. Neighbors and passersby, please call the family member XX immediately." After the base station restores power, the family member can be contacted by phone. If the household has an electric lock, the rescuer can be informed of the temporary password for the electric lock to enter the house and carry out first aid.
[0072] Scenario 6.1: Radar recognizes falling posture and automatically triggers alarm. F fell while using the restroom at night. The millimeter-wave radar detected the abnormal posture of the person falling and the device issued a voice inquiry locally: "Have you fallen? Do you need help?" Mr. F was conscious and verbally confirmed the emergency: The system entered a preset delay period. If no instruction to cancel the alarm was received from the elderly man by the end of the delay period, the external loudspeaker continuously broadcast the pre-stored help request message, fully announcing the address, the elderly man's name, and family contact number. Simultaneously, neighbors were prompted to contact family members to obtain the electric lock password for entry and rescue. Due to the "triple outage" environment of power outage, internet outage, and no base station, the device could not remotely push alarm messages and relied solely on the high-decibel loudspeaker to transmit help information, waiting for neighbors or passersby to provide on-site assistance. If Mr. F uttered keywords such as "turn off the alarm," "it's okay," or "I made a mistake" during the delay period to cancel the alarm, the external loudspeaker was immediately turned off, and normal human body monitoring resumed.
[0073] If F falls and loses consciousness and is unable to provide voice feedback: If no valid human voice is detected within the delay period, the device determines that it is a real event and continuously drives the external loudspeaker to broadcast the help request in a loop, so that passers-by, neighbors, and rescuers can enter the house to provide first aid.
[0074] Scenario 6.2 No sudden collapse, proactive voice call for help scenario Mr. F suddenly experienced chest tightness and dizziness, but did not fall to the ground. The millimeter-wave radar could not detect any abnormality. The elderly man called out the preset wake-up call, "Xiao Zhi, Xiao Zhi, I've fallen." The device skipped the human posture detection and pre-emptive voice inquiry process, without any delay, and immediately started the external loudspeaker to continuously broadcast resident information, family contact numbers, and smart lock password prompts. Even in a triple offline environment of power outage, network outage, and no base station, the loudspeaker broadcast function worked normally, relying on neighbors and passersby to complete the rescue.
[0075] Scenario 6.3 Environmental interference triggering and voice cancellation scenario If environmental disturbances such as swaying curtains or falling clothes are misinterpreted by the radar as a suspected fall, the device will automatically initiate a voice inquiry. If the elderly person responds normally with "It's okay, it was a mistake," the device will directly terminate the abnormal process without activating the external loudspeaker or triggering a disturbance alarm. If the device has already entered the delay phase, it will immediately stop emitting sound after recognizing a valid withdrawal of the voice, and the device will return to normal human body monitoring.
[0076] Once mains power, broadband, and cellular base station signals are restored, the device will automatically switch to mains power. After external communication is restored, any subsequent emergencies can be simultaneously reported to family members via the built-in communication module. In extreme offline scenarios, the device relies solely on external loudspeakers for neighborhood broadcasts as a local backup rescue method, serving the purpose of identifying emergencies and providing on-site voice broadcasts for assistance.
[0077] The above are merely preferred embodiments of this application. It should be noted that, for those skilled in the art, several structural changes or improvements can be made without departing from the principles of this application, and these changes or improvements should also be considered within the scope of protection of this application. For example, the three-level linkage function of human-computer interaction, including false alarm filtering, active emergency triggering, and voice alarm cancellation, can be fully integrated into the central processing unit; or it can be separately built into the visual camera, millimeter-wave radar, and sensing device; both arrangements fall under the protection of this invention. For example, the three unlocking modes of the electric lock can be activated individually or in multiple synchronized ways; each unlocking mode supports active voice triggering and automatic triggering after a delay period. During the delay period, the electric lock, loudspeaker, and communication module are all on standby, and unlocking simultaneously broadcasts a request for help and pushes multiple alarms via 4G.
Claims
1. A human perceptible false positive filtering system, characterized by: It includes a central processing unit, a human body state perception unit, and a human-computer interaction unit. The central processing unit communicates bidirectionally with the human body state perception unit and the human-computer interaction unit. The central processing unit has built-in judgment logic, performs independent local data processing, and autonomously completes anomaly judgment and / or process control, confirmation and / or alarm triggering. The human body state perception unit collects one or more of the target object's human posture, activity status and vital signs data in real time and transmits them to the central processing unit. The central processing unit analyzes data to identify suspected abnormal events. When a suspected abnormality is determined, it feeds back to the human-machine interaction unit and executes the corresponding operation: if it is a false alarm cancellation command, the false alarm is immediately terminated and the system returns to normal; if it is an emergency distress command or there is no effective feedback, it is determined to be a real event and the system triggers an alarm.
2. A human-aware leak detection reporting system, characterized by: It includes a central processing unit and a human-machine interaction unit, which communicate bidirectionally. The central processing unit has built-in judgment logic, performs independent local data processing, and autonomously recognizes and triggers rescue operations. The human-machine interaction unit is equipped with a voice module and pre-stores and recognizes voice. When the central processing unit receives valid voice, it directly determines the actual emergency status and triggers an alarm.
3. A human perception based alarm cancellation system, characterized by: The system includes a central processing unit and a human-machine interface unit (HMI). The HMI communicates bidirectionally with the central processing unit. The central processing unit has built-in judgment logic, performs independent local data processing, and autonomously recognizes commands, terminates alarms, and / or resets the system. The HMI is equipped with a voice module and pre-stores keywords for canceling alarms. When the system triggers an alarm, if the HMI recognizes a valid voice message for canceling the alarm, the central processing unit immediately cancels the alarm, and the system returns to normal.
4. A control system for emergency rescue of the elderly, characterized in that: It includes one or more systems as described in claims 1 to 3; when it includes the systems described in claims 1 to 3, it constitutes a three-level linkage closed-loop control system for local offline autonomous operation of false alarm filtering, false alarm prevention, and alarm cancellation.
5. The control system for emergency rescue of the elderly according to claim 4, characterized in that: It includes a door lock execution unit, which is equipped with an electric lock. After the emergency rescue process is triggered, the central processing unit opens the door lock.
6. The emergency rescue control system for the elderly according to claim 4, characterized in that: It also includes a warning output unit, which is equipped with a voice speaker. After the emergency rescue process is triggered, the central processing unit controls the playback of the voice.
7. The emergency rescue control system for the elderly according to claim 4, characterized in that: The central processing unit, human body state perception unit, and human-computer interaction unit are all equipped with independent backup batteries; the backup batteries and mains power are automatically switched.
8. The emergency rescue control system for the elderly according to claim 5, characterized in that: The electronic lock has three unlocking modes: Mode 1: Unlock directly with voice command When the emergency rescue process is not initiated, the electric lock remains in the normal locked state. After the central processing unit recognizes the unlocking voice command, it directly drives the electric lock to complete the automatic unlocking. Mode 2: Emergency unlocking via voice command When the emergency rescue process is not initiated, the electric lock remains in the normal locked state, and only pre-authorized personnel can unlock it. After the central processing unit recognizes the unlocking voice command, it removes the unlocking whitelist permission verification restriction, and any personnel can unlock it. Mode 3: Voice command, external independent power supply, manual emergency unlocking The electric lock is equipped with a manual emergency unlocking mechanism. When the emergency rescue process is not initiated, the electric lock remains in the normal locked state. After the central processing unit recognizes the unlocking voice command, any person can unlock the electric lock by pressing the external manual emergency unlocking mechanism.
9. The emergency rescue control system for the elderly according to claim 8, characterized in that: The electric lock has an unlocking delay period, and the delay threshold can be adjusted from 0 to a preset time. During the delay period, the entire emergency process can be terminated by a voice cancellation command. If the unlocking voice command is recognized during the delay period or if no cancellation command is received when the delay period expires, the corresponding unlocking action will be executed automatically. At the same time as the unlocking action is executed, the voice speaker will be activated and information will be sent through the built-in cellular mobile communication module.
10. The emergency rescue control system for the elderly according to claim 4, characterized in that: The human body state perception unit is one or more of the following: a visual acquisition device, a sensing radar device, a smart wearable device, and a sensing device.
11. The emergency rescue control system for the elderly according to claim 10, characterized in that: The visual acquisition device is equipped with a privacy shielding module.
12. The emergency rescue control system for the elderly according to claim 4, characterized in that: The central processing unit is also equipped with a local storage module and a communication module.