An ai intelligent independent secondary safety supervision system and method based on device state data

Abstract

The application provides an AI intelligent independent secondary safety supervision system and method based on device state data, and belongs to the technical field of artificial intelligence safety supervision, device state recognition and privacy protection. The system comprises: a primary device layer: a local data processing unit built in a controlled device, used for collecting device state data, pre-processing the device state data, and reporting standardized data through an Internet of Things protocol at a preset frequency; a secondary independent safety hub: independent of the controlled device, built-in lightweight AI inference framework and rule engine, used for receiving and processing the standardized data reported by the primary device layer without collecting audio and video, extracting multi-dimensional device state features, and adopting a double determination logic of "quantifiable hard threshold + AI confidence" for risk identification and safety control. The application can be widely applied to scenes such as home intelligent interaction, AI education, intelligent vehicles, intelligent transportation, and automatic driving assistance.

Description

Technical Field

[0001] This invention belongs to the field of artificial intelligence safety supervision, equipment status recognition and privacy protection technology, and relates to an AI intelligent independent secondary safety supervision system and method based on equipment status data. Background Technology

[0002] With the widespread adoption of smart devices, the demand for safety supervision of smart homes, assistive devices for the elderly, children's toys, drones, automobiles, and autonomous vehicles is increasing. Current technologies primarily rely on the following two methods to achieve safety supervision: One type relies on multimedia data collection devices such as cameras and microphones for behavior recognition, including fall detection, abnormal movement recognition, and driver status monitoring. This type of method has problems such as high privacy infringement risks, strong user resistance, and significant legal risks of data leakage.

[0003] Another type relies on conventional security algorithms built into the device for local control. It lacks an independent external intelligent monitoring layer, has a simple security architecture, low level of intelligence, often uses fixed threshold judgments, and lacks AI autonomous optimization capabilities. When the device is hacked or the algorithm malfunctions, it is prone to security failure.

[0004] In addition, existing technologies have limited functionality and poor adaptability to different scenarios. They only cover personal safety and lack comprehensive protection against property and life risks such as vehicle loss of control, driver abnormalities, payment risks, and equipment malfunctions. They are also difficult to work effectively in closed scenarios without networks or cameras and lack autonomous, scenario-based intelligent judgment capabilities.

[0005] Currently, there is no publicly available technology that discloses a two-layer AI intelligent security monitoring architecture that is independent of the controlled device, does not collect audio or video, and is based on device status data. It is also unable to provide underlying technical support for privacy and security for upper-layer intelligent interaction, online education, smart cars, drones, autonomous driving and other applications. Summary of the Invention

[0006] In view of this, the purpose of this invention is to provide an AI-powered intelligent independent secondary security monitoring system and method based on device status data, which can solve the problems of high privacy risks, single security architecture, poor scenario adaptability, and lack of intelligent judgment capabilities in the existing technology, and realize AI-powered intelligent secondary security monitoring independent of the controlled intelligent device.

[0007] To achieve the above objectives, the present invention provides the following technical solution: An AI-powered, independent Level 2 security monitoring system based on device status data, comprising: Level 1 Device Layer: This is the local data processing unit built into the controlled device, used to collect device status data, preprocess the device status data, and then report standardized data through the Internet of Things protocol at a preset frequency. The secondary independent security hub operates independently of the controlled devices. It has a built-in lightweight AI inference framework and rule engine to receive and process standardized data reported by the primary device layer without collecting audio and video. It extracts multi-dimensional device status features and uses a dual judgment logic of "quantifiable hard threshold + AI confidence" for risk identification and security control.

[0008] Furthermore, the secondary independent security hub is implemented based on an embedded industrial control board, and its deployment forms include home security gateways, cloud service nodes, vehicle-mounted security units, or drone ground monitoring nodes, and it has multi-channel communication capabilities.

[0009] Furthermore, the risk identification includes one or more of the following: fall detection, device malfunction, abnormal payment, fraud prevention for the elderly, smart home security, abnormal vehicle driver, vehicle malfunction, abnormal drone, abnormal robot, anti-theft of shared devices, industrial equipment security, or pet smart device security; each risk identification adopts a dual judgment logic that combines a preset numerical hard threshold with the risk confidence level output by the AI ​​model.

[0010] Furthermore, the secondary independent safety hub is also used to perform independent forced control on automobiles, drones, autonomous vehicles, industrial robotic arms, or shared equipment. Control commands include one or more of the following: deceleration, stopping, hazard lights, forced landing, emergency stop, or locking the equipment.

[0011] Furthermore, the secondary independent security hub is also used to output standardized security status codes and synchronize them to the upper-layer application in JSON format, while also outputting standardized alarm interfaces.

[0012] This invention also provides an AI-powered intelligent independent secondary safety monitoring method based on device status data, the method comprising the following steps: S1. The device status data is collected by the first-level device layer built into the controlled device. After the data is preprocessed, standardized structured data is reported through the Internet of Things protocol at a preset frequency. S2. The standardized structured data is received and verified by a secondary security center independent of the controlled device to ensure its integrity and legality, and multi-dimensional quantifiable device status features are extracted. S3. Risk is determined by a combination of "quantifiable hard threshold and AI scenario-based fusion judgment", and graded response and control are implemented according to risk level without collecting audio or video. S4. Generate standardized security logs and analyze historical logs using AI models to uncover patterns of potential security risks.

[0013] Furthermore, the method is applicable to one or more of the following: automobiles, drones, elevators, robotic arms, high-speed trains, ships, industrial equipment, medical equipment, agricultural machinery, smart homes, assistive devices for the elderly, or children's toys.

[0014] Furthermore, the primary device layer and the secondary independent security hub conduct end-to-end data transmission via the TLS encryption protocol, the standardized data are reported via the MQTT protocol, and the device status data includes one or more of acceleration, angular velocity, position, velocity, or pressure data.

[0015] Furthermore, the multi-dimensional device state features include one or more of mechanical data, motion data, state data, vehicle data, or AI intelligent optimization features; the dual judgment logic is executed collaboratively by the local rule engine and the AI ​​model, and the AI ​​model includes the XGBoost classification model or the reinforcement learning model.

[0016] The beneficial effects of this invention are as follows: 1) Zero privacy dependence: Completely eliminates cameras and microphones, and achieves security monitoring solely through device status data, fundamentally eliminating privacy risks; 2) Dual-layer intelligent redundancy protection: The independent secondary security hub and the device local layer form dual redundancy in hardware and algorithm. Even if the device local algorithm fails, the hub can still force intervention, which significantly improves security and reliability. 3) Quantifiable AI fusion algorithm: Risk assessment adopts a dual logic of "hard threshold + AI confidence", which results in accurate and reproducible judgment with a low false alarm rate; 4) Comprehensive protection across all scenarios: It covers both personal safety and property safety, and is compatible with multiple scenarios such as home, industry, shared equipment, automobiles, drones, and autonomous driving. 5) High security and compatibility: It adopts TLS encrypted communication and independent hardware deployment, supporting the access of various smart devices with sensors and communication modules without the need to modify the device hardware; 6) The end-to-end response time in core risk scenarios is short, meeting the needs of real-time interaction. The AI ​​model is continuously optimized online and can autonomously adapt to the usage habits and device characteristics of different scenarios.

[0017] Other advantages, objectives, and features of the invention will be set forth in part in the description which follows, and in part will be apparent to those skilled in the art from the following examination, or may be learned from practice of the invention. The objectives and other advantages of the invention can be realized and obtained through the following description. Attached Figure Description

[0018] To make the objectives, technical solutions, and advantages of the present invention clearer, the preferred embodiments of the present invention will be described in detail below with reference to the accompanying drawings, wherein: Figure 1 This is a block diagram of the overall two-layer architecture of the system of the present invention; Figure 2 This is a flowchart of the method of the present invention. Detailed Implementation

[0019] The following specific examples illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and various details in this specification can be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of the present invention. Unless otherwise specified, the following embodiments and features can be combined with each other.

[0020] The accompanying drawings are for illustrative purposes only and are schematic diagrams, not actual pictures. They should not be construed as limiting the invention. To better illustrate the embodiments of the invention, some parts in the drawings may be omitted, enlarged, or reduced, and do not represent the actual product dimensions. It is understandable to those skilled in the art that some well-known structures and their descriptions may be omitted in the drawings.

[0021] In the accompanying drawings of the embodiments of the present invention, the same or similar reference numerals correspond to the same or similar components. In the description of the present invention, it should be understood that if terms such as "upper," "lower," "left," "right," "front," and "rear" indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, they are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, the terms used to describe positional relationships in the drawings are only for illustrative purposes and should not be construed as limiting the present invention. For those skilled in the art, the specific meaning of the above terms can be understood according to the specific circumstances.

[0022] Example 1: System Architecture and Hardware Deployment Figure 1 This is a block diagram of the overall two-layer architecture of the system of the present invention. This embodiment provides an AI intelligent two-level security monitoring system, including a first-level device layer and a second-level independent security hub, as detailed below: The first-level device layer is located within the controlled device and is implemented using a low-power MCU (such as the STM32L4 series). This layer includes sensor arrays such as a triaxial accelerometer (such as the ADXL345), an angular velocity sensor (such as the MPU6050), and a GPS / BeiDou positioning module (such as the NEO-M8N). The MCU acquires raw data at a frequency of 10Hz, performs a 5Hz Butterworth low-pass filter on the acceleration and angular velocity data to remove high-frequency noise, and performs min-max normalization on the position and velocity data, unifying the units to m / s and m, with data precision retained to two decimal places. The preprocessed data is encapsulated in JSON format at a fixed frequency of 1Hz and reported to the second-level security hub via the MQTT v3.1.1 protocol through a Wi-Fi 6 or Bluetooth 5.2 communication link. The data reporting is set to timeout retransmission three times, with a retransmission interval of 500ms.

[0023] The secondary independent security hub operates independently of the controlled devices, implemented on an embedded industrial control board (such as the RK3568), equipped with a quad-core Cortex-A55 processor and integrated 1 TOPS NPU computing power. This hub can be deployed as a home security gateway, vehicle-mounted security unit, or drone ground monitoring node. The hub receives data reported by the primary devices via an MQTT Broker, employing the TLS 1.3 encryption protocol for end-to-end data transmission, with a 256-bit AES encryption key.

[0024] Example 2: Risk Identification and Decision Control Figure 2 The flowchart of the method of the present invention is shown below. This embodiment takes a fall recognition scenario as an example to illustrate the risk judgment and decision control process.

[0025] When the secondary safety center receives status data from the elderly walking aid device, it extracts mechanical features based on a sliding time window (window length 5s, step size 1s). If a vertical acceleration greater than 2.5g is detected for a duration of less than 0.3s, and the horizontal velocity drops abruptly from greater than 0.5m / s to 0 within 0.5s, and the device's attitude angle is greater than 60°, then a hard threshold condition is triggered.

[0026] The central processing unit further input the above features into a lightweight XGBoost classification model (TensorFlow Lite quantization deployment, 4.8M parameters). The model output a fall risk confidence score of 0.95, exceeding the preset threshold of 0.9. Based on the combined hard threshold and AI confidence score, the central processing unit confirmed it as a high-risk case.

[0027] According to the hierarchical response rules, the central system executes a high-risk response: immediately sends a stop command to the first-level device. The command is sent via the MQTT protocol, with a response timeout of 200ms. If the timeout occurs, it will retry once. At the same time, a high-risk alarm is pushed to the user, and the security status code "2" is synchronized to the upper-layer application in JSON format.

[0028] Example 3: AI Self-Optimization and Log Traceability This embodiment illustrates the system's AI self-optimization capability and log tracing function.

[0029] The secondary security hub incorporates a lightweight XGBoost classification model, employing online learning. The model is fine-tuned every 1000 sets of valid feature data collected, optimizing feature weights and extraction rules to improve the accuracy of risk assessment. The optimization results are updated in real-time to the local SQLite feature extraction rule library.

[0030] Simultaneously, the central system generates standardized structured security logs in both CSV and JSON formats. Log fields include timestamp, device ID, risk type, judgment criteria, processing result, AI optimization parameters, and instruction execution status. The logs are stored locally on 8GB eMMC flash memory for at least 180 days.

[0031] The central processing unit incorporates a lightweight association rule mining algorithm (Apriori), which performs offline analysis of historical logs at 1:00 AM daily and online analysis every hour to uncover patterns of potential security risks, such as peak times for equipment anomalies and abnormal vehicle routes. The mining results are updated to the risk assessment rule base in the form of rules to achieve early warning.

[0032] Example 4: Multi-scenario Adaptation and Control Execution This embodiment illustrates the system's adaptability and control execution capabilities in different scenarios.

[0033] For new energy vehicles, the secondary safety hub is connected to the vehicle control system via the CAN bus. When the system detects that the driver has not operated for more than 15 seconds and the vehicle speed is greater than 30 km / h, and the AI ​​model outputs an anomaly confidence level of ≥0.9, the hub determines that the driver is abnormal and independently executes the mandatory control command of "deceleration → hazard lights → pull over". The command is sent to the vehicle execution unit via the CAN bus, without relying on the original vehicle control system.

[0034] For drones, the central control unit connects via a dedicated 2.4G / 5.8G communication protocol. When an attitude angle continuously >45° for >1 second is detected, and the AI ​​model outputs an abnormal confidence level ≥0.9, the central control unit immediately executes the "de-altitude → hover → forced landing" command to prevent the drone from crashing.

[0035] For smart home devices, the central hub connects via Wi-Fi. When it detects that the current of an appliance exceeds 150% of its rated value, the temperature exceeds 80°C, or it has been running continuously for more than 24 hours, and the AI ​​model outputs a safety risk confidence level ≥0.85, the central hub issues a power-off command.

[0036] Example 5: Standardized Interfaces and Ecosystem Collaboration This embodiment illustrates the system's standardized interface and ecosystem collaboration capabilities.

[0037] The secondary security hub standardizes all security statuses using "0 = Normal / Low Risk, 1 = Medium Risk, 2 = High Risk" encoding, and synchronizes them to upper-layer applications in JSON format with a synchronization latency of <50ms. Upper-layer applications, including home smart interaction systems and AI education platforms, obtain security status data in real time through standardized interfaces without requiring additional adaptation.

[0038] When used in conjunction with the floating AI interaction system submitted on the same day, the secondary security hub synchronizes the security status code to the interaction system in real time through the TLS 1.3 encryption protocol, providing it with underlying security data support and forming a technical collaboration link of "underlying security protection - mid-level interaction entry - upper-level business scenario".

[0039] In summary, this invention can be widely applied to scenarios such as home intelligent interaction, AI education, intelligent transportation, intelligent transportation, and autonomous driving assistance. As a foundational privacy and security base, it can also achieve ecological synergy with related patent architectures.

[0040] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. An AI-powered intelligent independent secondary safety monitoring system based on device status data, characterized in that, The system includes: Level 1 Device Layer: This is the local data processing unit built into the controlled device, used to collect device status data, preprocess the device status data, and then report standardized data through the Internet of Things protocol at a preset frequency. The secondary independent security hub operates independently of the controlled devices. It has a built-in lightweight AI inference framework and rule engine to receive and process standardized data reported by the primary device layer without collecting audio and video. It extracts multi-dimensional device status features and uses a dual judgment logic of "quantifiable hard threshold + AI confidence" for risk identification and security control.

2. The AI-powered intelligent independent secondary safety monitoring system based on device status data according to claim 1, characterized in that, The secondary independent security hub is implemented based on an embedded industrial control board and can be deployed as a home security gateway, cloud service node, vehicle-mounted security unit or drone ground monitoring node, and has multi-channel communication capabilities.

3. The AI-powered intelligent independent secondary safety monitoring system based on device status data according to claim 1, characterized in that, The risk identification includes one or more of the following: fall detection, device malfunction, abnormal payment, fraud prevention for the elderly, smart home security, abnormal vehicle driver, vehicle malfunction, abnormal drone, abnormal robot, anti-theft of shared devices, industrial equipment security, or smart pet device security; each risk identification uses a dual judgment logic that combines a preset numerical hard threshold with the risk confidence level output by the AI ​​model.

4. The AI-powered intelligent independent secondary safety monitoring system based on device status data according to claim 1, characterized in that, The secondary independent safety hub is also used to perform independent mandatory control over automobiles, drones, autonomous vehicles, industrial robotic arms, or shared equipment. Control commands include one or more of the following: deceleration, stopping, hazard lights, forced landing, emergency stop, or locking the equipment.

5. The AI-powered intelligent independent secondary safety monitoring system based on device status data according to claim 1, characterized in that, The secondary independent security hub is also used to output standardized security status codes and synchronize them to upper-layer applications in JSON format, while also outputting standardized alarm interfaces.

6. An AI-powered intelligent independent secondary safety monitoring method based on device status data, characterized in that, The method includes the following steps: S1. The device status data is collected by the first-level device layer built into the controlled device. After the data is preprocessed, standardized structured data is reported through the Internet of Things protocol at a preset frequency. S2. The standardized structured data is received and verified by a secondary security center independent of the controlled device to ensure its integrity and legality, and multi-dimensional quantifiable device status features are extracted. S3. Risk is determined by a combination of "quantifiable hard threshold and AI scenario-based fusion judgment", and graded response and control are implemented according to risk level without collecting audio and video. S4. Generate standardized security logs and analyze historical logs using AI models to uncover patterns of potential security risks.

7. The AI-powered intelligent independent secondary safety monitoring method based on device status data according to claim 6, characterized in that, The method is applicable to one or more of the following: automobiles, drones, elevators, robotic arms, high-speed trains, ships, industrial equipment, medical equipment, agricultural machinery, smart homes, assistive devices for the elderly, or children's toys.

8. The AI-powered intelligent independent secondary safety monitoring method based on device status data according to claim 6, characterized in that, The primary device layer and the secondary independent security hub transmit data end-to-end via the TLS encryption protocol. The standardized data is reported via the MQTT protocol. The device status data includes one or more of the following: acceleration, angular velocity, position, velocity, or pressure data.

9. The AI-powered intelligent independent secondary safety monitoring method based on device status data according to claim 6, characterized in that, The multi-dimensional device status features include one or more of mechanical data, motion data, status data, vehicle data, or AI intelligent optimization features; the dual judgment logic is executed collaboratively by the local rule engine and the AI ​​model, and the AI ​​model includes the XGBoost classification model or the reinforcement learning model.

Images