Monitoring and early warning method based on intelligent wearing

Abstract

The invention discloses a monitoring and early warning method based on intelligent wearing, and aims to improve the safety of a user. The method comprises the following steps: continuously collecting original data streams from various built-in sensors through intelligent wearable equipment, and reflecting the state, posture, position and environment information of a user in real time; and carrying out real-time preprocessing, time-space synchronization and multi-modal fusion on the data to construct unified data representation. Based on this, analysis and feature extraction are performed, key indicators are calculated and compared with threshold values, and preset safety or auxiliary events are identified and confirmed in real time, the events at least including user tumble, location abnormality, external threats or active help seeking of the user. And according to the identified event type, severity and context, automatically selecting and triggering a corresponding execution module built in the equipment to execute a preset response action. According to the invention, various danger or help-seeking scenes can be accurately identified, and timely and effective automatic response is provided.

Description

Technical Field

[0001] The present invention belongs to the field of monitoring and early warning, and in particular relates to a monitoring and early warning method based on smart wearable devices. Background Art

[0002] With the rapid development of smart wearable technology, health and safety monitoring and early warning based on smart wearable devices have become an important trend. In existing technologies, some smart wearable devices already have basic activity monitoring, heart rate detection, or single-function early warning, such as simple fall detection based on acceleration thresholds, or location tracking and electronic fence alarms using GPS. However, these existing systems often have significant limitations: they usually rely only on a single type or a limited number of sensor data, lack the ability to deeply integrate and collaboratively analyze multimodal data, resulting in an inability to fully and accurately perceive the user's complex status and surrounding environmental conditions; their event detection algorithms are usually relatively simple, mainly based on preset hard thresholds, and it is difficult to effectively identify multiple types and levels of safety or auxiliary events, especially complex scenarios involving user interaction with the environment.

[0003] Furthermore, existing systems have relatively limited early warning and response mechanisms, often limited to sending simple notifications or alerts. They lack the ability to automatically trigger diverse, targeted response actions based on event type, severity, and context, such as rapid physical protection or localized warnings and guidance based on environmental awareness. This limits their effectiveness in protecting and assisting in emergencies. Therefore, there is an urgent need for a smart wearable monitoring and early warning method that can fully utilize multi-source sensor data, accurately identify multiple potential risks or needs through intelligent analysis, and implement a coordinated and diverse set of effective responses. Summary of the Invention

[0004] In response to the above technical problems, the present invention provides a monitoring and early warning method based on smart wearable devices, which is used to monitor and warn based on wearable devices and handle special situations.

[0005] A monitoring and early warning method based on smart wearable devices comprises the following steps: Step S1, continuously collecting raw data streams from multiple built-in sensors of different types through the smart wearable device, wherein the data streams reflect the user's current motion state, posture information, geographic location, surrounding environment conditions and / or physiological indicators in real time, and the raw data streams include at least acceleration data, angular velocity data and position data.

[0006] Step S2, performing real-time preprocessing, spatiotemporal synchronization, and multimodal data fusion on the raw data streams collected from the various types of sensors to construct a unified data representation; based on the unified data representation, performing multi-level data analysis and feature extraction, and applying pattern recognition algorithms and / or judgment rules to calculate a series of key indicators reflecting user status and / or environmental conditions, and comparing the key indicators with multiple preset thresholds in real time to identify and confirm the occurrence of preset safety or auxiliary events, the identification process includes calculating the key indicators and threshold judgment according to preset logic.

[0007] Step S3, based on the specific safety or auxiliary event type, severity and context information identified and confirmed in step S2, automatically select and trigger the corresponding execution module built into the smart wearable device to execute the preset corresponding safety or auxiliary response action.

[0008] Furthermore, the step S1 specifically includes: collecting acceleration data through the built-in inertial measurement unit and angular velocity The raw data of motion and posture, including Separate moments The acceleration components along the three coordinate axes, Separate moments angular velocity components along the three coordinate axes; collecting raw position data including GPS signals through the built-in positioning module; collecting raw environmental data including video images and audio signals through the built-in environmental perception module; and / or collecting raw physiological data of the user through the built-in physiological sensors.

[0009] Furthermore, step S2 specifically includes: performing noise filtering, calibration and standardization preprocessing on the raw data of each sensor; performing time synchronization on the preprocessed multi-source data based on timestamps; performing feature-level or decision-level fusion on the synchronized data; based on the fused data, using signal processing, machine learning algorithms and / or rule-based expert systems to analyze the user's movement patterns, posture change trajectories, position sequences, and visual and auditory features of the surrounding environment, identify patterns related to preset events, and calculate and compare key indicators and thresholds according to specific decision logic.

[0010] Furthermore, in step S2, preset safety or auxiliary events are identified and confirmed, and the events include at least one or more of the following: identifying a specific motion-posture composite pattern indicating that the user has suddenly, involuntarily and rapidly fallen; identifying a position sequence pattern indicating that the user has stayed in an abnormal position for a long time, deviated from a preset safety area, or moved on an atypical path; identifying an environmental perception pattern or abnormal impact signal indicating the presence of abnormal proximity, collision or violent behavior in the surrounding environment; identifying a help or navigation request signal triggered by a user through a specific physical operation.

[0011] Furthermore, the step S2 identifies a specific motion-posture composite pattern indicating a sudden, involuntary, rapid fall of the user, specifically including analyzing the fused motion and posture data, and calculating a fall characteristic index reflecting the intensity of the user's motion and the rapid change of posture, wherein the characteristic index includes the instantaneous modulus of the combined acceleration. , the instantaneous modulus of the resulting angular velocity and the tilt angle of the posture relative to the vertical direction ; Based on the characteristic indicators, calculate the fall risk score , the calculation formula is: ; in, For the moment The combined acceleration modulus length; For the moment The modulus of the resultant angular velocity; is the peak value of the combined acceleration modulus during the event period; is the peak value of the combined angular velocity mode length; is the maximum rate of change of attitude tilt angle; The duration of time the user remains stationary or in slight motion after the event occurs; , , , is the weight coefficient of each characteristic index; , , , is a function that maps the corresponding characteristic index to the risk contribution value; when the calculated fall risk score When the preset fall determination threshold is exceeded, a fall event is determined to have occurred.

[0012] Furthermore, the step S2 identifies the environmental perception pattern or abnormal impact signal indicating the presence of abnormal approach, collision or violent behavior in the surrounding environment, specifically including: analyzing the video image and audio signal stream collected by the environmental perception module, using the pattern recognition algorithm to identify whether there is a preset abnormal visual pattern or a preset abnormal auditory pattern, the preset abnormal auditory pattern includes a collision sound, glass breaking sound, screaming or threatening voice; the preset abnormal visual pattern includes but is not limited to a stranger approaching the wearer quickly or continuously, wandering for a long time, and a specific aggressive or inducing gesture; analyzing the inertial measurement unit data , detect whether there is a sudden high-intensity acceleration change and calculate the impact severity index , the calculation formula is as follows: ; in, Indicates the duration window of the detected high-intensity acceleration change; when the abnormal visual pattern or the preset abnormal auditory pattern is identified, or the calculated impact severity index is When the preset collision determination threshold is exceeded, it is determined that there is an external safety risk or collision event.

[0013] Furthermore, the step S2 identifies a position sequence pattern indicating that the user has stayed in an abnormal position for a long time, deviated from a preset safety area, or moved on an atypical path, specifically including: based on the user's precise current position sequence continuously obtained by the positioning module; calculating the geographical distance between the original position data and a preset reference position; judging whether the geographical distance continuously exceeds the preset safety area radius; judging whether the original user position data is continuously located outside the preset safety electronic fence area for a preset total time threshold. When the distance exceeding time or the total time outside the area threshold conditions are met, or the user path deviates significantly from the historical typical path, it is determined that the user is in a lost or abnormal position state.

[0014] Furthermore, step S3 specifically includes automatically executing one or more of the following actions according to the type and severity of the event: triggering the deployment or adjustment of the physical protection mechanism; triggering a localized sound and light warning or voice broadcast of help information; sending an emergency help or alarm message containing the user's precise location, event type, timestamp and preset health information to one or more contacts in a preset emergency contact list through a communication module; starting real-time recording of environmental monitoring data, and storing the recorded data in a local storage medium and / or uploading it to a cloud platform via the network; providing the user with voice navigation guidance based on the current location to a preset destination.

[0015] Furthermore, the action of triggering the deployment or adjustment of the physical protection mechanism specifically includes: immediately sending an instruction to the actuator built into the smart wearable device after identifying that the user has fallen or suffered an external impact; driving the actuator to control the physical protection structure of the device to move quickly to a protection position within a preset response time, so as to provide cushioning and protection for key parts of the user's head before contacting the ground or objects, and the response time is less than the time from the user's fall / impact to the maximum impact that the head may suffer.

[0016] Furthermore, the triggering of localized sound and light warnings or voice broadcasting of help information specifically includes: when a fall event is identified as being caused by non-external reasons, and the user fails to recover to a normal posture within a preset short time limit for self-recovery, a preset help voice message is cyclically broadcasted through the speaker of the smart wearable device; the help voice message content includes user identity information, event status and / or help content, and a preset emergency contact number; and / or when a potential external threat event is identified, a sound and light warning is issued through the smart wearable device to attract the attention of people around.

[0017] Compared with the prior art, the present invention has the following beneficial effects: (1) The present invention improves the accuracy and comprehensiveness of detecting various safety or auxiliary events by fusing multimodal sensor data and performing multi-level analysis; (2) The present invention can automatically trigger a variety of targeted responses, including rapid physical protection, based on the type and severity of the incident, effectively improving user safety. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 This is a flowchart of an exemplary step of the monitoring and early warning method of the present invention. DETAILED DESCRIPTION

[0019] like Figure 1 As shown in this embodiment, a monitoring and early warning method based on smart wearables is provided, which includes the following steps: Step S1, continuously collecting raw data streams from multiple built-in sensors of different types through the smart wearable device. The data streams reflect the user's current motion status, posture information, geographic location, surrounding environment conditions and / or physiological indicators in real time. The raw data streams include at least acceleration data, angular velocity data and position data.

[0020] In one embodiment, the smart wearable device is designed as a headwear-like product, such as a peaked cap. Its built-in sensors include, but are not limited to: an inertial measurement unit (IMU) for collecting acceleration and angular velocity data; a positioning module (e.g., GPS / Beidou) for collecting location data; an environmental perception module, including at least four high-definition cameras for capturing 360-degree video images; and an optional microphone for collecting audio signals; and optional physiological sensors for collecting raw physiological data from the user. The raw data collected by these sensors forms the basis for subsequent event detection and response. For example, inertial data is crucial for fall and collision detection, location data is crucial for trajectory monitoring and navigation, and environmental data is crucial for threat identification and event recording.

[0021] Step S2: Perform real-time preprocessing, spatiotemporal synchronization, and multimodal data fusion on the raw data streams collected from various types of sensors to construct a unified data representation; based on the unified data representation, perform multi-level data analysis and feature extraction, and apply pattern recognition algorithms and / or judgment rules to calculate a series of key indicators reflecting user status and / or environmental conditions, and compare the key indicators with multiple preset thresholds in real time to identify and confirm the occurrence of preset safety or auxiliary events. The identification process includes calculating key indicators and threshold judgment according to preset logic.

[0022] In one embodiment, the preset safety or auxiliary events identified and confirmed in step S2 include but are not limited to: a sudden, involuntary, rapid fall of the user; a position sequence pattern event in which the user stays in an abnormal position for a long time, deviates from a preset safety area, or moves on an atypical path; an environmental perception pattern or abnormal impact signal event in which there is abnormal proximity, collision, or violent behavior in the surrounding environment; and a help or navigation request signal event triggered by a specific physical operation of the user.

[0023] For example, for the identification of sudden, involuntary, rapid falls of users, the system analyzes the fused inertial data, calculates the fall characteristic indicators that reflect the intensity of the user's movement and the rapid change of posture, and calculates the fall risk score based on these characteristic indicators. When the fall risk score exceeds the preset threshold, it is determined that a fall event has occurred. This fall detection logic is applicable to a variety of situations, including but not limited to: falls caused by elderly people due to non-external impacts such as imbalance or slippery roads; sudden falls of patients with cardiovascular and cerebrovascular diseases or other latent diseases. In these scenarios, the system accurately identifies falls and provides a basis for subsequent timely rescue.

[0024] For example, to identify environmental perception patterns or abnormal impact signal events indicating unusual proximity, collision, or violent behavior in the surrounding environment, the system analyzes the video image and audio signal streams captured by the environmental perception module and uses a pattern recognition algorithm to identify preset abnormal visual or auditory patterns. Simultaneously, it analyzes inertial measurement unit data to detect sudden, high-intensity acceleration changes and calculates an impact severity index. When the impact severity index exceeds a preset collision determination threshold, an external impact event is determined.

[0025] Step S3, based on the specific safety or auxiliary event type, severity and context information identified and confirmed in step S2, automatically select and trigger the corresponding execution module built into the smart wearable device to execute the preset corresponding safety or auxiliary response action.

[0026] In one embodiment, the safety or auxiliary response actions performed in step S3 include but are not limited to: Triggering the deployment or adjustment of the physical protection mechanism: After identifying a fall or external impact event of the user, the system immediately sends instructions to the actuator built into the smart wearable device, and quickly moves to the protection position within the preset response time to provide cushioning and protection for the user's head before contacting the ground or objects.

[0027] For example, the physical protection structure may use high-strength and high-toughness carbon fiber as the outer layer, high-strength and high-shock-resistant fiber synthetic plastic as the middle layer, and non-toxic and harmless silicone as the inner energy-absorbing layer, aiming to absorb impact energy to the greatest extent, reduce head injuries, and even resist short-term, non-lethal external physical attacks, buying time for users to escape or call for help.

[0028] Provides voice navigation guidance based on the user's current location to a preset destination: When the user triggers a navigation request signal by pressing a button, the system provides clear voice navigation guidance through the built-in speaker based on the user's current GPS location and the preset destination to assist the user. At the same time, this operation can also optionally send the user's current location information to family members.

[0029] Step S1 specifically includes: collecting acceleration data through the built-in inertial measurement unit and angular velocity The raw data of motion and posture, including Separate moments The acceleration components along the three coordinate axes, Separate moments angular velocity components along the three coordinate axes; collecting raw position data including GPS signals through the built-in positioning module; collecting raw environmental data including video images and audio signals through the built-in environmental perception module; and / or collecting raw physiological data of the user through the built-in physiological sensors.

[0030] Step S2 specifically includes: performing noise filtering, calibration and standardization preprocessing on the raw data of each sensor; performing time synchronization on the preprocessed multi-source data based on timestamps; fusing the synchronized data at the feature level or decision level; and using signal processing, machine learning algorithms and / or rule-based expert systems based on the fused data to analyze the user's motion patterns, posture change trajectories, position sequences, and visual and auditory features of the surrounding environment, identify patterns related to preset events, and calculate and compare key indicators and thresholds according to specific decision logic.

[0031] In step S2, preset safety or auxiliary events are identified and confirmed, and the events include at least one or more of the following: identifying a specific motion-posture composite pattern indicating that the user has suddenly, involuntarily and rapidly fallen; identifying a position sequence pattern indicating that the user has stayed in an abnormal position for a long time, deviated from a preset safety area, or moved on an atypical path; identifying an environmental perception pattern or abnormal impact signal indicating that there is abnormal proximity, collision or violent behavior in the surrounding environment; identifying a help or navigation request signal triggered by a specific physical operation of the user.

[0032] In step S2, a specific motion-posture composite pattern indicating a sudden, involuntary, rapid fall of the user is identified, specifically including analyzing the fused motion and posture data and calculating a fall characteristic index reflecting the intensity of the user's motion and the rapid change of posture, the characteristic index including the instantaneous modulus of the combined acceleration. , the instantaneous modulus of the resulting angular velocity and the tilt angle of the posture relative to the vertical direction ; Based on the characteristic indicators, calculate the fall risk score , the calculation formula is: ; in, For the moment The combined acceleration modulus length; For the moment The modulus of the resultant angular velocity; is the peak value of the combined acceleration modulus during the event period; is the peak value of the combined angular velocity mode length; is the maximum rate of change of attitude tilt angle; The duration of time the user remains stationary or in slight motion after the event occurs; , , , is the weight coefficient of each characteristic index; , , , is a function that maps the corresponding characteristic index to the risk contribution value; when the calculated fall risk score When the preset fall determination threshold is exceeded, a fall event is determined to have occurred.

[0033] Step S2 identifies environmental perception patterns or abnormal impact signals indicating abnormal approach, collision or violent behavior in the surrounding environment, specifically including: analyzing the video image and audio signal stream collected by the environmental perception module, using a pattern recognition algorithm to identify whether there is a preset abnormal visual pattern or a preset abnormal auditory pattern, wherein the preset abnormal auditory pattern includes a crash, glass breaking, screaming or threatening voice; the preset abnormal visual pattern includes but is not limited to a stranger approaching the wearer quickly or continuously, wandering for a long time, and a specific aggressive or inducing gesture; analyzing the inertial measurement unit data , detect whether there is a sudden high-intensity acceleration change and calculate the impact severity index , the calculation formula is as follows: ; in, Indicates the duration window of the detected high-intensity acceleration change; when the abnormal visual pattern or the preset abnormal auditory pattern is identified, or the calculated impact severity index is When the preset collision determination threshold is exceeded, it is determined that there is an external safety risk or collision event.

[0034] In step S2, a position sequence pattern indicating that the user has stayed in an abnormal position for a long time, deviated from a preset safety area, or moved on an atypical path is identified, specifically including: the user's precise current position sequence continuously obtained based on the positioning module; calculating the geographical distance between the original position data and the preset reference position; judging whether the geographical distance continuously exceeds the preset safety area radius; judging whether the original user position data is continuously located outside the preset safety electronic fence area for a preset total time threshold. When the distance exceeding the limit time or the total time outside the area threshold conditions are met, or the user path deviates significantly from the historical typical path, it is determined that the user is in a lost or abnormal position state.

[0035] Step S3 specifically includes automatically executing one or more of the following actions based on the type and severity of the event: triggering the deployment or adjustment of the physical protection mechanism; triggering a localized sound and light warning or voice broadcast of help information; sending an emergency help or alarm message containing the user's precise location, event type, timestamp and preset health information to one or more contacts in the preset emergency contact list through the communication module; starting real-time recording of environmental monitoring data, and storing the recorded data in a local storage medium and / or uploading it to a cloud platform via the network; providing the user with voice navigation guidance based on the current location to the preset destination.

[0036] Triggering the deployment or adjustment of the physical protection mechanism specifically includes: immediately sending instructions to the built-in actuator of the smart wearable device after identifying that the user has fallen or suffered an external impact; driving the actuator to control the physical protection structure of the device to move quickly to the protection position within a preset response time, so as to provide cushioning and protection for the key parts of the user's head before contacting the ground or objects. The response time is less than the time from the user's fall / impact to the maximum impact that the head may suffer.

[0037] Triggering localized audio and visual warnings or voice broadcasts of help messages, specifically including: when a fall event is identified as not caused by external reasons, and the user fails to recover to a normal posture within a preset short time limit for self-recovery, a preset help voice message is broadcast in a loop through the speaker of the smart wearable device; the help voice message includes user identity information, event status and / or help content, and a preset emergency contact number; and / or when a potential external threat event is identified, an audio and visual warning is issued through the smart wearable device to attract the attention of people around.

[0038] The smart wearable device's exterior can be designed to resemble a standard peaked cap, while its internal structure rationally distributes components such as the sensors, actuators, battery, circuit board, CPU, storage module, charging port, manual switch, and automatic start module. A carbon fiber layer, a drop-resistant layer, and an energy-absorbing layer form the product's protective outer shell and inner lining. Key areas such as eye and face protection, the back of the head, the forehead, and both sides of the brain are protected by appropriate structures.

[0039] In one specific embodiment, to further enhance the ability to respond to specific medical emergencies, the smart wearable device, such as a hat, is designed with a small sealed storage compartment on the left or right side. This compact compartment can be used to store emergency items pre-placed by the user based on their health status. For example, for users with hypoglycemia, sugar cubes can be pre-placed; for heart patients, emergency medicines such as quick-acting heart-saving pills can be stored. When the system analyzes physiological data and identifies that the user may have fallen or is in a dangerous state due to hypoglycemia or a sudden heart attack, it triggers an alarm, notifies emergency contacts, and sends a help message that includes a reminder of the emergency items in the storage compartment. This design allows rescuers to immediately access and utilize these first aid items, providing timely physical assistance to the user, greatly enhancing the device's practical value and auxiliary effectiveness in real-world emergency scenarios.

[0040] The above contents are merely examples and explanations of the present invention. Those skilled in the art may make various modifications or additions to the described specific embodiments or replace them in similar ways. As long as they do not deviate from the scope defined by the invention, they should all fall within the scope of protection of the present invention.

Claims

1. A monitoring and early warning method based on smart wearable devices, characterized in that: The following steps are involved: Step S1: continuously collecting raw data streams from multiple different types of built-in sensors through the smart wearable device, wherein the data streams reflect the user's current motion state, posture information, geographic location, surrounding environment conditions and / or physiological indicators in real time, and the raw data streams include at least acceleration data, angular velocity data and position data; Step S2, performing real-time preprocessing, spatiotemporal synchronization, and multimodal data fusion on the raw data streams collected from the various types of sensors to construct a unified data representation; Based on the unified data representation, multi-level data analysis and feature extraction are performed, and pattern recognition algorithms and / or determination rules are applied to calculate a series of key indicators reflecting user status and / or environmental conditions, and the key indicators are compared with multiple preset thresholds in real time to identify and confirm the occurrence of preset safety or auxiliary events. The identification process includes calculating the key indicators and determining the thresholds according to preset logic; Step S3, based on the specific safety or auxiliary event type, severity and context information identified and confirmed in step S2, automatically select and trigger the corresponding execution module built into the smart wearable device to execute the preset corresponding safety or auxiliary response action.

2. A monitoring and early warning method based on smart wearable according to claim 1, characterized in that: The step S1 specifically includes: collecting acceleration data through the built-in inertial measurement unit and angular velocity The raw data of motion and posture, including Separate moments The acceleration components along the three coordinate axes, Separate moments angular velocity components along the three coordinate axes; collecting raw position data including GPS signals through the built-in positioning module; collecting raw environmental data including video images and audio signals through the built-in environmental perception module; and / or collecting raw physiological data of the user through the built-in physiological sensors.

3. The monitoring and early warning method based on smart wearable according to claim 1, characterized in that: The step S2 specifically includes performing noise filtering, calibration and standardization preprocessing on the raw data of each sensor; The pre-processed multi-source data is time-synchronized based on timestamps; the synchronized data is fused at the feature level or decision level; based on the fused data, signal processing, machine learning algorithms and / or rule-based expert systems are used to analyze the user's motion patterns, posture change trajectories, position sequences, and visual and auditory features of the surrounding environment, identify patterns related to preset events, and calculate and compare key indicators and thresholds based on specific decision logic.

4. The monitoring and early warning method based on smart wearable according to claim 1, characterized in that: In step S2, preset safety or auxiliary events are identified and confirmed, and the events include at least one or more of the following: identifying a specific movement-posture composite pattern indicating that the user has suddenly, involuntarily and rapidly fallen; identifying a position sequence pattern indicating that the user has stayed in an abnormal position for a long time, deviated from a preset safety area, or moved on an atypical path; identifying an environmental perception pattern or abnormal impact signal indicating the presence of abnormal proximity, collision or violent behavior in the surrounding environment; identifying a help or navigation request signal triggered by a specific physical operation of the user.

5. The monitoring and early warning method based on smart wearable according to claim 1, characterized in that: The step S2 identifies a specific motion-posture composite pattern indicating a sudden, involuntary, rapid fall of the user, specifically including analyzing the fused motion and posture data, and calculating a fall characteristic index reflecting the intensity of the user's motion and the rapid change of posture, wherein the characteristic index includes the instantaneous modulus of the combined acceleration. , the instantaneous modulus of the resulting angular velocity and the tilt angle of the posture relative to the vertical direction ; Based on the characteristic indicators, calculate the fall risk score , the calculation formula is: ； in, For the moment The combined acceleration modulus length; For the moment The modulus of the resultant angular velocity; is the peak value of the combined acceleration modulus during the event period; is the peak value of the combined angular velocity mode length; is the maximum rate of change of attitude tilt angle; The duration of time the user remains stationary or in slight motion after the event occurs; , , , is the weight coefficient of each characteristic index; , , , is a function that maps the corresponding characteristic index to the risk contribution value; when the calculated fall risk score When the preset fall determination threshold is exceeded, a fall event is determined to have occurred.

6. The monitoring and early warning method based on smart wearable devices according to claim 1, characterized in that: The step S2 identifies the environmental perception pattern or abnormal impact signal indicating the presence of abnormal approach, collision or violent behavior in the surrounding environment, specifically including: analyzing the video image and audio signal stream collected by the environmental perception module, using the pattern recognition algorithm to identify whether there is a preset abnormal visual pattern or a preset abnormal auditory pattern, wherein the preset abnormal auditory pattern includes a crash, glass breaking, screaming or threatening voice; the preset abnormal visual pattern includes but is not limited to a stranger approaching the wearer quickly or continuously, wandering for a long time, and a specific aggressive or inducing gesture; analyzing the inertial measurement unit data , detect whether there is a sudden high-intensity acceleration change and calculate the impact severity index , the calculation formula is as follows: ； in, Indicates the duration window of the detected high-intensity acceleration change; when the abnormal visual pattern or the preset abnormal auditory pattern is identified, or the calculated impact severity index is When the preset collision determination threshold is exceeded, it is determined that there is an external safety risk or collision event.

7. The monitoring and early warning method based on smart wearable devices according to claim 1, characterized in that: The step S2 identifies a location sequence pattern indicating that the user has stayed in an abnormal location for a long time, deviated from a preset safety area, or moved on an atypical path, specifically including: based on the user's precise current location sequence continuously obtained by the positioning module; calculating the geographical distance between the original location data and a preset reference location; determining whether the geographical distance continuously exceeds the preset safety area radius; and determining whether the original location data of the user is continuously located outside the preset safety electronic fence area for a preset total time threshold. When the distance exceeding time or the total time outside the area threshold conditions are met, or the user path deviates significantly from the historical typical path, it is determined that the user is in a lost or abnormal location state.

8. The monitoring and early warning method based on smart wearable devices according to claim 1, characterized in that: The step S3 specifically includes automatically executing one or more of the following actions based on the type and severity of the incident: triggering the deployment or adjustment of a physical protection mechanism; triggering a localized sound and light warning or voice broadcast of a help message; sending an emergency help or alert message containing the user's precise location, incident type, timestamp, and preset health information to one or more contacts in a preset emergency contact list via a communication module; Start real-time recording of environmental monitoring data and store the recorded data in local storage media and / or upload it to the cloud platform via the network; Provides users with voice navigation guidance based on their current location to a preset destination.

9. The monitoring and early warning method based on smart wearable devices according to claim 8, characterized in that: The triggering of the deployment or adjustment action of the physical protection mechanism specifically includes: immediately sending an instruction to the actuator built into the smart wearable device after identifying that the user has fallen or suffered an external impact event; driving the actuator to control the physical protection structure of the device to quickly move to a protection position within a preset response time, so as to provide cushioning and protection for the key parts of the user's head before contacting the ground or objects. The response time is less than the time from the user's fall / impact to the maximum impact that the head may suffer.

10. The monitoring and early warning method based on smart wearable devices according to claim 8, characterized in that: The triggering of the localized sound and light warning or voice broadcast of help information specifically includes: when a fall event is identified as not caused by external reasons and the user fails to recover to a normal posture within a preset short time of self-recovery determination, a preset help voice message is cyclically broadcasted through the speaker of the smart wearable device; the help voice message content includes user identity information, event status and / or help content, and a preset emergency contact number; and / or when a potential external threat event is identified, an sound and light warning is issued through the smart wearable device to attract the attention of people around.

Images