Entrance guard access method and system based on multi-source data and entrance guard terminal
Through the access control method of multi-source data, using technologies such as reinforcement learning and edge computing, dynamically optimizing the verification process, and combining infrared fill light and iris verification, it solves the security and convenience issues of existing access control systems in complex scenarios, and realizes efficient and secure access control.
Patent Information
- Application Number
- CN202510684050.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-05-26
- Publication Date
- 2025-09-05
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
Existing access control systems with a single data source are difficult to meet the requirements of security and convenience in complex scenarios, especially when the access card is lost or the biometric accuracy is poor, there is a problem of insufficient security.
The access control method adopts multi-source data, collects historical attack data and environmental parameters, uses reinforcement learning models to dynamically optimize the verification process, and combines infrared fill light, iris verification, zero-knowledge proof, edge computing and dual redundant communication technologies to achieve adaptive adjustment and enhance security.
It improves the security and response speed of the access control system in different environments and attack modes, ensures the security and privacy of data, and enhances the flexibility and reliability of the system.
Smart Images
Figure CN120599730A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of access control technology, and in particular to an access control method, system and access control terminal based on multi-source data. Background Art
[0002] Access control systems primarily used simple electronic locks or magnetic card locks, with doors unlocked using keys or cards. This technology was less secure and vulnerable to hacking and theft. With the development of proximity card technology, card-based access control systems began to emerge. These systems use a card reader to read the card information carried by the user for authentication and door opening control, offering greater security.
[0003] With the improvement of security needs and technological advancement, the traditional single data source access control system of existing technology can no longer meet the security and convenience requirements in complex scenarios. If the access card is accidentally lost and the accuracy of facial biometric recognition technology is poor, relying solely on card swiping or facial recognition to unlock the door is still lacking in security and convenience. Summary of the Invention
[0004] The object of the present invention is to provide an access control method, system and access control terminal based on multi-source data to solve the problems raised in the above background technology.
[0005] In order to solve the above technical problems, the present invention provides the following technical solutions: a multi-source data-based access control method, system and access control terminal, comprising the following steps: A1: Collect historical attack data and environment parameters; A2: Utilizes a reinforcement learning model to dynamically optimize the verification process based on the historical attack data and environmental parameters; A3: Adaptively adjust the verification method based on current environment parameters and real-time detected attacks.
[0006] Preferably, the environmental parameters include light intensity and background noise decibels, and the adaptive adjustment verification method includes: automatically enabling infrared fill light to enhance facial recognition in a low-light environment; increasing the voiceprint verification weight; and temporarily adding an iris verification link when a high-frequency attack is detected.
[0007] Preferably, the specific implementation method of temporarily adding an iris verification step when a high-frequency attack is detected is: B1: Count the number of abnormal access attempts to the access control system within a set time period. When the number of abnormal access attempts exceeds the preset high-frequency attack threshold, it is determined that a high-frequency attack has been detected. B2: After determining it is a high-frequency attack, a command is sent to the access control terminal to temporarily enable the iris verification function and require the user to perform iris verification; B3: After the high-frequency attack is resolved, that is, when the number of abnormal access attempts is lower than the preset recovery threshold, the original verification method is restored.
[0008] Preferably, it includes a zero-knowledge proof module, which is used to encrypt the user's biometric hash value and store it in the alliance chain node; a smart contract module, which is used to complete feature matching through the on-chain smart contract during verification; and a local data protection module, which ensures that the original biometric data never leaves the local device.
[0009] Preferably, the access rights management module supports user-defined access rights and generates temporary visitor keys; the digital credential generation module generates a verifiable digital credential for each verification, realizing decentralized auditing and anti-tampering traceability.
[0010] The preferred architecture includes a low-latency access control response architecture based on edge computing. The architecture deploys lightweight AI inference chips at the access control terminal to realize localized calculation of core algorithms for optical flow analysis and voiceprint extraction. The architecture also includes cloud-based model training and threat intelligence synchronization modules to ensure data privacy through federated learning.
[0011] Preferably, it also includes: dual redundant communication modules, including 5G and LoRa communication modules, to ensure that basic verification can still be completed in a network disconnected environment.
[0012] Preferably, the response delay of the lightweight AI inference chip is reduced to less than 200ms.
[0013] Preferably, it includes a user interaction interface for displaying verification information, alarm information and providing a user operation entrance; a hardware interface for connecting external devices, including a camera, a microphone, an infrared fill light, and an iris recognition sensor.
[0014] Compared with the prior art, the present invention has the following beneficial effects: First, the present invention constructs a reinforcement learning model that dynamically optimizes the verification process based on historical attack data and environmental parameters. In low-light environments, infrared fill light is automatically enabled to enhance facial recognition, while voiceprint verification weight is increased. The reinforcement learning model dynamically optimizes the verification process based on collected data. The reinforcement learning model can continuously adjust its strategy based on feedback to adapt to different attack patterns and environmental changes. The learning model uses the Q-learning reinforcement learning algorithm and is trained using historical data. The access control system records relevant information about each access attempt in real time, including access time, visitor identity, and access results. Within a specified ten-minute period, the number of abnormal access attempts within that time period is counted, including facial recognition failures and voiceprint verification failures. If five or more abnormal access attempts occur within ten minutes, it is determined to be a high-frequency attack, and iris verification is enabled. The user is prompted to perform iris verification through a user interface. The user needs to align their eyes with the iris recognition sensor to complete iris information collection and verification. The iris information verification information and collection results are stored, allowing subsequent staff or administrators to view who is using iris verification. Upon successful verification or if there are no more than two access control attempts within half an hour, high-frequency attack verification is deactivated and the original verification method is restored.
[0015] Second, the present invention, in which the zero-knowledge proof module encrypts the hash value of fingerprints and facial features and stores it in the alliance chain node to ensure the security of biometric data during transmission and storage. Zero-knowledge proof technology allows the verifier to confirm the user's identity without obtaining or storing the user's original biometric data. During the verification process, the smart contract module completes feature matching through the on-chain smart contract. The smart contract automatically executes the preset verification rules to ensure the fairness and transparency of the verification process. The local data protection module ensures that the original biometric data never leaves the local device, further enhancing the security of the data. The access permission management module supports user-defined access permissions and generates temporary The digital credential generation module generates a verifiable digital credential for each verification, realizing decentralized audit and anti-tampering tracing. The digital credential records the time, place and result of the verification, which facilitates subsequent audit and tracing. It uses zero-knowledge proof technology to encrypt the user's biometric hash value and store it in the alliance chain node. During verification, feature matching is completed through the on-chain smart contract. The original data never leaves the local device, supports user-defined access rights and temporary visitor keys, and generates a verifiable digital credential for each verification, realizing decentralized audit and anti-tampering tracing.
[0016] Third, the present invention, among other things, deploys a lightweight AI inference chip on the access control terminal based on edge computing to realize localized calculation of core algorithms such as optical flow analysis and voiceprint extraction. This architecture reduces data transmission delay and improves the response speed of the access control system. The cloud-based model training and threat intelligence synchronization module is only used for incremental model training and threat intelligence synchronization on the cloud side. Federated learning is used to ensure data privacy. Federated learning allows models to be trained on multiple data sources without sharing original data, thereby protecting user privacy. Dual redundant communication modules ensure that basic verification can still be completed in a network-disconnected environment. The 5G communication module provides high-speed data transmission, while the LoRa communication module provides long-distance, low-power communication capabilities. The combination of the two ensures the reliability of the access control system. Lightweight AI inference chips are deployed on the access control terminal to realize localized calculation of core algorithms such as optical flow analysis and voiceprint extraction. The response delay is reduced to less than 200ms. The cloud-based model training and threat intelligence synchronization module is only used for incremental model training and threat intelligence synchronization. Data privacy is ensured by federated learning. At the same time, dual redundant communication modules (5G+LoRa) are designed to ensure that basic verification can still be completed in a network-disconnected environment. BRIEF DESCRIPTION OF THE DRAWINGS
[0017] Figure 1 This is a schematic diagram of the structural method flow of the present invention; DETAILED DESCRIPTION The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.
[0018] Example 1 like Figure 1 As shown, the present invention provides a technical solution: a door access method based on multi-source data, comprising the following steps: A1: Collect historical attack data and environment parameters; A2: Utilizes a reinforcement learning model to dynamically optimize the verification process based on the historical attack data and environmental parameters; A3: Adaptively adjust the verification method based on current environment parameters and real-time detected attacks.
[0019] The environmental parameters include light intensity and background noise decibels. The adaptive adjustment verification method includes: automatically enabling infrared fill light to enhance facial recognition in a low-light environment; increasing the voiceprint verification weight; and temporarily adding an iris verification link when a high-frequency attack is detected.
[0020] The specific implementation method of temporarily adding an iris verification step when a high-frequency attack is detected is: B1: Count the number of abnormal access attempts to the access control system within a set time period. When the number of abnormal access attempts exceeds the preset high-frequency attack threshold, it is determined that a high-frequency attack has been detected. B2: After determining it is a high-frequency attack, a command is sent to the access control terminal to temporarily enable the iris verification function and require the user to perform iris verification; B3: After the high-frequency attack is resolved, that is, when the number of abnormal access attempts is lower than the preset recovery threshold, the original verification method is restored.
[0021] Through the above technical solution, a reinforcement learning model is constructed to dynamically optimize the verification process based on historical attack data and environmental parameters. Infrared fill light is automatically enabled to enhance facial recognition in low-light environments, while voiceprint verification weight is increased. The reinforcement learning model is used to dynamically optimize the verification process based on the collected data. The reinforcement learning model can continuously adjust its strategy based on feedback to adapt to different attack patterns and environmental changes. The learning model uses the Q-learning reinforcement learning algorithm and is trained using historical data. The access control system records relevant information about each access attempt in real time, including access time, visitor identity, and access results. Within a specified ten-minute period, the number of abnormal access attempts within that time period is counted, including facial recognition failures and voiceprint verification failures. If there are more than five abnormal access attempts within ten minutes, it is determined to be a high-frequency attack and iris verification is activated. The user is prompted to perform iris verification through the user interface. The user needs to align their eyes with the iris recognition sensor to complete the collection and verification of iris information. The iris information verification information and collection results are stored so that subsequent staff or administrators can check who is using iris verification. If the verification is successful or there are no more than two access attempts within half an hour, the high-frequency attack verification is lifted and the original verification method is restored.
[0022] Example 2 like Figure 1 As shown, the present invention provides a technical solution: an access control system based on multi-source data, including a zero-knowledge proof module, which is used to encrypt the user's biometric hash value and store it in the alliance chain node; a smart contract module, which is used to complete feature matching through the on-chain smart contract during verification; and a local data protection module, which ensures that the original biometric data never leaves the local device.
[0023] The access rights management module supports users to customize access rights and generate temporary visitor keys. The digital credential generation module generates verifiable digital credentials for each verification, realizing decentralized auditing and anti-tampering traceability.
[0024] It includes a low-latency response architecture for access control based on edge computing. The architecture deploys lightweight AI inference chips at the access control terminal to realize localized calculation of core algorithms for optical flow analysis and voiceprint extraction. The architecture also includes cloud-based model training and threat intelligence synchronization modules to ensure data privacy through federated learning.
[0025] It also includes: dual redundant communication modules, including 5G and LoRa communication modules, to ensure that basic verification can still be completed in an Internet disconnection environment.
[0026] The response delay of the lightweight AI inference chip is reduced to less than 200ms.
[0027] It includes a user interaction interface for displaying verification information, alarm information and providing user operation entry; a hardware interface for connecting external devices, including cameras, microphones, infrared fill lights, and iris recognition sensors.
[0028] Through the above technical solution, the zero-knowledge proof module encrypts the hash value of fingerprints and facial features and stores it in the alliance chain node to ensure the security of biometric data during transmission and storage. Zero-knowledge proof technology allows the verifier to confirm the user's identity without obtaining or storing the user's original biometric data. During the verification process, the smart contract module completes feature matching through the on-chain smart contract. The smart contract automatically executes the preset verification rules to ensure the fairness and transparency of the verification process. The local data protection module ensures that the original biometric data never leaves the local device, further enhancing the security of the data. The access permission management module supports user-defined access permissions and generates Temporary visitor keys allow users to flexibly set access rights for different users as needed, improving the flexibility and security of the access control system. The digital credential generation module generates a verifiable digital credential for each verification, realizing decentralized auditing and anti-tampering tracing. The digital credential records the time, location, and results of the verification, providing convenience for subsequent auditing and tracing. Using zero-knowledge proof technology, the user's biometric hash value is encrypted and stored in the alliance chain node. During verification, feature matching is completed through the on-chain smart contract. The original data never leaves the local device. It supports user-defined access rights and temporary visitor keys, and generates a verifiable digital credential for each verification, realizing decentralized auditing and anti-tampering tracing. The access control low-latency response architecture based on edge computing deploys lightweight AI inference chips at the access control terminal to localize the computation of core algorithms such as optical flow analysis and voiceprint extraction. This architecture reduces data transmission latency and improves the access control system's response speed. The cloud-based model training and threat intelligence synchronization module is only used for incremental model training and threat intelligence synchronization. Federated learning ensures data privacy, allowing models to be trained on multiple data sources without sharing original data, thereby protecting user privacy. Dual-redundant communication modules ensure that basic verification can be completed even in disconnected environments. The 5G communication module provides high-speed data transmission, while the LoRa communication module provides long-distance, low-power communication capabilities. The combination of the two ensures the reliability of the access control system. Lightweight AI inference chips are deployed at the access control terminal to localize the computation of core algorithms such as optical flow analysis and voiceprint extraction, reducing response latency to less than 200ms. The cloud-based model training and threat intelligence synchronization module is only used for incremental model training and threat intelligence synchronization. Federated learning ensures data privacy, and dual-redundant communication modules (5G+LoRa) are designed to ensure basic verification can be completed even in disconnected environments.
[0029] During use, a reinforcement learning model is constructed to dynamically optimize the verification process based on historical attack data and environmental parameters. Infrared fill light is automatically enabled to enhance facial recognition in low-light environments, while voiceprint verification weight is increased. The reinforcement learning model is used to dynamically optimize the verification process based on the collected data. The reinforcement learning model can continuously adjust its strategy based on feedback to adapt to different attack patterns and environmental changes. The learning model uses the Q-learning reinforcement learning algorithm and is trained using historical data. The access control system records relevant information about each access attempt in real time, including access time, visitor identity, and access results. Within a specified ten-minute period, the number of abnormal access attempts within that time period is counted, including facial recognition failures and voiceprint verification failures. If there are more than five abnormal access attempts within ten minutes, it is determined to be a high-frequency attack and iris verification is activated. The user is prompted to perform iris verification through the user interface. The user needs to align their eyes with the iris recognition sensor to complete the collection and verification of iris information. The verification information and collection results of the iris information are stored so that subsequent staff or administrators can check who is using iris verification. If the verification is successful or there are no more than two access control access attempts within half an hour, the high-frequency attack verification is lifted and the original verification method is restored. Among them, the zero-knowledge proof module encrypts the hash value of fingerprints and facial features and stores it in the alliance chain node to ensure the security of biometric data during transmission and storage. Zero-knowledge proof technology allows the verifier to confirm the user's identity without obtaining or storing the user's original biometric data. During the verification process, the smart contract module completes feature matching through the on-chain smart contract. The smart contract automatically executes the preset verification rules to ensure the fairness and transparency of the verification process. The local data protection module ensures that the original biometric data never leaves the local device, further enhancing the security of the data. The access rights management module supports users to customize access rights and generate temporary visitors. Key, users can flexibly set the access rights of different users according to their needs, improving the flexibility and security of the access control system. The digital credential generation module generates a verifiable digital credential for each verification, realizing decentralized auditing and anti-tampering tracing. The digital credential records the time, place and result of the verification, which provides convenience for subsequent auditing and tracing. It uses zero-knowledge proof technology to encrypt the user's biometric hash value and store it in the alliance chain node. During verification, feature matching is completed through the on-chain smart contract. The original data never leaves the local device, supports user-defined access rights and temporary visitor keys, and generates a verifiable digital credential for each verification, realizing decentralized auditing and anti-tampering tracing.The access control low-latency response architecture based on edge computing deploys lightweight AI inference chips at the access control terminal to localize the computation of core algorithms such as optical flow analysis and voiceprint extraction. This architecture reduces data transmission latency and improves the access control system's response speed. The cloud-based model training and threat intelligence synchronization module is only used for incremental model training and threat intelligence synchronization. Federated learning ensures data privacy, allowing models to be trained on multiple data sources without sharing original data, thereby protecting user privacy. Dual-redundant communication modules ensure that basic verification can be completed even in disconnected environments. The 5G communication module provides high-speed data transmission, while the LoRa communication module provides long-distance, low-power communication capabilities. The combination of the two ensures the reliability of the access control system. Lightweight AI inference chips are deployed at the access control terminal to localize the computation of core algorithms such as optical flow analysis and voiceprint extraction, reducing response latency to less than 200ms. The cloud-based model training and threat intelligence synchronization module is only used for incremental model training and threat intelligence synchronization. Federated learning ensures data privacy, and dual-redundant communication modules (5G+LoRa) are designed to ensure basic verification can be completed even in disconnected environments.
[0030] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and alterations may be made to the embodiments without departing from the principles and spirit thereof, and that the scope of the invention is defined by the appended claims and their equivalents.
Claims
1. A door access control method based on multi-source data, characterized by: The following steps are involved: A1: Collect historical attack data and environment parameters; A2: Utilizes a reinforcement learning model to dynamically optimize the verification process based on the historical attack data and environmental parameters; A3: Adaptively adjust the verification method based on current environment parameters and real-time detected attacks.
2. The access control method based on multi-source data according to claim 1, characterized in that: The environmental parameters include light intensity and background noise decibels. The adaptive adjustment verification method includes: automatically enabling infrared fill light to enhance facial recognition in a low-light environment; increasing the voiceprint verification weight; and temporarily adding an iris verification link when a high-frequency attack is detected.
3. The access control method based on multi-source data according to claim 2, characterized in that: The specific implementation method of temporarily adding an iris verification step when a high-frequency attack is detected is: B1: Count the number of abnormal access attempts to the access control system within a set time period. When the number of abnormal access attempts exceeds the preset high-frequency attack threshold, it is determined that a high-frequency attack has been detected. B2: After determining it is a high-frequency attack, a command is sent to the access control terminal to temporarily enable the iris verification function and require the user to perform iris verification; B3: After the high-frequency attack is resolved, that is, when the number of abnormal access attempts is lower than the preset recovery threshold, the original verification method is restored.
4. A multi-source data-based access control system, based on the multi-source data-based access control method according to any one of claims 1 to 3, characterized in that: It includes a zero-knowledge proof module, which is used to encrypt the user's biometric hash value and store it in the alliance chain node; a smart contract module, which is used to complete feature matching through on-chain smart contracts during verification; and a local data protection module, which ensures that the original biometric data never leaves the local device.
5. The multi-source data-based access control system according to claim 4, characterized in that: Also includes: Access rights management module, which supports user-defined access rights and generates temporary visitor keys; The digital certificate generation module generates a verifiable digital certificate for each verification, realizing decentralized auditing and anti-tampering traceability.
6. The multi-source data-based access control system according to claim 4, characterized in that: It includes a low-latency response architecture for access control based on edge computing. The architecture deploys lightweight AI inference chips at the access control terminal to realize localized calculation of core algorithms for optical flow analysis and voiceprint extraction. The architecture also includes cloud-based model training and threat intelligence synchronization modules to ensure data privacy through federated learning.
7. The multi-source data-based access control system according to claim 6, characterized in that: Also includes: Dual redundant communication modules, including 5G and LoRa communication modules, ensure that basic verification can be completed even in a network-disconnected environment.
8. The access control system based on multi-source data according to claim 6, characterized in that: The response delay of the lightweight AI inference chip is reduced to less than 200ms.
9. An access control terminal based on multi-source data, based on the access control method and system based on multi-source data according to any one of claims 1 to 8, characterized in that: It includes a user interaction interface for displaying verification information, alarm information and providing user operation entry; a hardware interface for connecting external devices, including cameras, microphones, infrared fill lights, and iris recognition sensors.