CN-122024365-A - Method and system for cooperatively unlocking door lock

CN122024365ACN 122024365 ACN122024365 ACN 122024365ACN-122024365-A

Abstract

The invention discloses a method and a system for cooperatively unlocking a door lock, wherein the method comprises the steps of constructing a distributed processing network for cooperative identification according to an auxiliary identification request triggered by a target door lock when a low-power or fault state is detected, synchronously collecting and sharing a detected mobile equipment MAC address list and signal intensity information thereof according to an event that any neighbor door lock in the distributed processing network captures the face of a target user for the first time, iteratively updating the binding confidence coefficient of each MAC address and the face according to the event that a subsequent door lock continuously captures the face of the target user and monitors the signal intensity change of the MAC address until reaching a preset binding confirmation threshold, and triggering the target door lock to execute unlocking according to the confirmed binding relation and the signal characteristics of portable equipment of the target user. By utilizing the embodiment of the invention, reliable identity confirmation and unlocking independent of independent operation of the main door lock can be realized when the main door lock is abnormal, and the robustness and user experience of the system are improved.

Inventors

YE FEI
ZHU ZHILING
SANG SHENGWEI
TANG JUNXIONG
DENG YEHAO
XI JUAN

Assignees

浙江德施曼科技智能股份有限公司

Dates

Publication Date: 20260512
Application Date: 20260415

Claims (10)

1. A method for cooperatively unlocking a door lock, the method comprising: constructing an auxiliary recognition network, namely according to an auxiliary recognition request triggered by a target door lock when the target door lock detects a low-power or fault state, gradually spreading face feature vectors of a target user along left and right neighbor door lock chains of the target user, and constructing a distributed processing network for collaborative recognition; The accompanying relation detection and data synchronization is carried out, namely, a multicast tracking task is triggered to be established according to the event that the face of a target user is captured for the first time by any neighbor door lock in the distributed processing network, and the detected MAC address list of the mobile equipment and the signal intensity information thereof are synchronously collected and shared; Confirming a confidence coefficient accumulation and binding relation, namely continuously capturing the face of a target user and monitoring the event of the signal intensity change of the MAC address according to the subsequent door lock, and iteratively updating the binding confidence coefficient of each MAC address and the face through signal curve fitting and intersection calculation until a preset binding confirmation threshold value is reached; And executing collaborative decision and unlocking, namely initiating a door lock decision at the target door lock or by a task according to the confirmed binding relation and the signal characteristics of the portable equipment of the target user, and triggering the target door lock to execute unlocking action.
2. The method of claim 1, wherein the auxiliary identification network construction comprises: Generating low power detection and request, namely generating auxiliary identification request information comprising a target user face feature vector, a door lock identification and a request direction according to the fact that the target door lock detects that the self power is lower than a preset threshold value; constructing a neighbor relation chain, namely unicast transmitting the message to a left neighbor door lock and a right neighbor door lock prestored in the target door lock according to the direction mark in the request message; Step-by-step transmission of the request, namely analyzing and locally storing a face feature vector of a target user according to a neighbor door lock receiving the request, adding self identification to a neighbor relation linked list, and continuously forwarding updated information to the next neighbor door lock in the same direction until a link tail; And the distributed network is ready, wherein processing chain tables comprising a plurality of neighbor door locks are formed on two sides of the target door lock according to the propagation process, and each chain table node door lock enters an auxiliary identification mode to complete the construction of the distributed processing network.
3. The method of claim 2, wherein the companion relationship detection is synchronized with data, comprising: The first detection trigger and task creation are that according to any neighbor door lock entering an auxiliary recognition mode, a target user face is successfully matched and captured, the door lock is used as a task initiator, a unique tracking recognition task ID is created, and a new multicast group is applied to be added; The multicast network is built, according to the neighbor relation linked list, all door locks in the task initiation direction linked list unicast a command for joining the multicast group, and after each door lock responds and joins, a tracking task group based on multicast communication is formed; Acquiring and synchronizing initial data, namely acquiring MAC addresses and real-time signal strengths of all surrounding mobile devices by a task initiator through a WiFi sniffing module in real time according to a face capturing event to generate an initial MAC address set table; And the data multicast notification, namely according to the successful multicast group construction state, the task initiator packages the initial MAC address collection table, the face feature vector ID and the capturing time into a multicast message, and sends the multicast message to the whole task group to start a collaborative computing flow.
4. The method of claim 3, wherein the confidence accumulating and binding relationship validation comprises: According to the multicast message of the initial MAC address collection table, each door lock in the task group starts to periodically sample the signal intensity of the MAC address in the table, and an exponential weighted moving average algorithm is adopted to fit a sampling curve so as to smooth noise points and identify trends; The method comprises the following steps of carrying out intersection calculation, basic confidence degree lifting, peak value detection and reinforcement lifting, wherein the intersection calculation is carried out according to the event that a target user face is captured by a subsequent door lock, the basic confidence degree lifting is carried out by carrying out first value lifting on binding confidence degrees of all MAC addresses in the intersection, and the peak value detection and reinforcement lifting is carried out by detecting whether each MAC address fitting curve has a peak value or not in an active time window before and after the face capturing moment; Confirming the binding relation, namely confirming that the MAC address and the face of the target user form the binding relation according to the fact that the accumulated confidence coefficient of any MAC address exceeds a preset binding threshold; And the dynamic time window optimization comprises the steps of sharing the time difference information through multicast according to the time difference from the face capturing moment recorded by the door lock to the occurrence of the signal peak value detected by the door lock, and calculating an average value and adding a redundancy value by the subsequent door lock according to the received time difference records to serve as an initial reference value of the local active time window of the door lock.
5. The method of claim 4, wherein the collaborative decision-making and unlocking is performed, comprising: Judging the door opening condition, namely confirming the moving state of a target user according to the binding relation, namely processing the following two situations that the target door lock is online, namely continuously monitoring the signal strength of the bound MAC address if the target door lock is in a tracking task multicast group, judging that the user arrives and unlocking is executed when the signal strength is monitored to exceed the specific proportion of the historical peak value average value and to be stable continuously, and taking off the target door lock, namely calculating the time for the user to arrive at the target door lock by a task initiator according to the sequence and the average passing time of the user reported by a neighbor door lock if the target door lock is in a power-saving multicast group, and unicast sending a door opening instruction to the target door lock at the moment; unlocking execution and notification, namely executing unlocking action by a target door lock according to the condition of the door opening or receiving a remote instruction, and playing a voice prompt; task release, namely according to the successful unlocking event, the task initiating transmits a task release message to the multicast group, and each member door lock releases the task related resources and exits the multicast group.
6. The method of claim 5, further comprising the anomaly and optimization processing step of: The method comprises the steps of determining a folding direction, namely determining that a user folds back according to a binding relation, and inquiring a signal change trend of a bound MAC address from a door lock capturing a face again to an adjacent door lock of the user; Long-time stagnation processing, namely according to a total overtime threshold set when a tracking and identifying task is created, if no new user capturing or MAC signal significant change message is received within the threshold time, automatically releasing the task; Binding failure processing, namely judging that the collaborative identification fails and automatically dismissing the task according to the fact that the candidate MAC address set becomes an empty set due to intersection acquisition in the task advancing process; and the missing identification fault tolerance is that according to the condition that part of door locks are missing to capture human faces, the final binding relation confirmation and door opening flow is not affected as long as the number of successfully captured door locks is enough to enable the confidence of the correct MAC address to be accumulated to a binding threshold.
7. A system for cooperatively unlocking a door lock, the system comprising: the construction module is used for constructing an auxiliary recognition network, namely, according to an auxiliary recognition request triggered by a target door lock when the low electric quantity or the fault state is detected, the face feature vector of a target user is propagated step by step along left and right neighbor door lock chains of the target user, and a distributed processing network for collaborative recognition is constructed; the synchronization module is used for synchronizing the detection and the data along with the relationship, triggering the establishment of a multicast tracking task according to the event that the face of a target user is captured by any neighbor door lock in the distributed processing network for the first time, and synchronously collecting and sharing the detected MAC address list of the mobile equipment and the signal intensity information thereof; The determining module is used for confirming the confidence coefficient accumulation and binding relation, namely continuously capturing the face of the target user and monitoring the event of the signal intensity change of the MAC address according to the subsequent door lock, and iteratively updating the binding confidence coefficient of each MAC address and the face through signal curve fitting and intersection calculation until a preset binding confirmation threshold value is reached; and the execution module is used for executing the cooperative decision and unlocking, namely triggering the target door lock to execute the unlocking action at the target door lock or by initiating the door lock decision by a task according to the confirmed binding relation and the signal characteristics of the portable equipment of the target user.
8. The system according to claim 7, characterized in that said construction module is in particular adapted to: Generating low power detection and request, namely generating auxiliary identification request information comprising a target user face feature vector, a door lock identification and a request direction according to the fact that the target door lock detects that the self power is lower than a preset threshold value; constructing a neighbor relation chain, namely unicast transmitting the message to a left neighbor door lock and a right neighbor door lock prestored in the target door lock according to the direction mark in the request message; Step-by-step transmission of the request, namely analyzing and locally storing a face feature vector of a target user according to a neighbor door lock receiving the request, adding self identification to a neighbor relation linked list, and continuously forwarding updated information to the next neighbor door lock in the same direction until a link tail; And the distributed network is ready, wherein processing chain tables comprising a plurality of neighbor door locks are formed on two sides of the target door lock according to the propagation process, and each chain table node door lock enters an auxiliary identification mode to complete the construction of the distributed processing network.
9. A storage medium having a computer program stored therein, wherein the computer program is arranged to perform the method of any of claims 1-6 when run.
10. An electronic device comprising a memory and a processor, wherein the memory has stored therein a computer program, the processor being arranged to run the computer program to perform the method of any of claims 1-6.

Description

Method and system for cooperatively unlocking door lock Technical Field The invention belongs to the technical field of door locks, and particularly relates to a method and a system for cooperatively unlocking a door lock. Background With the popularization of intelligent hotel and apartment management, intelligent door locks have become important devices for improving the experience and operation efficiency of residents. Most current smart door lock systems still rely on a single door lock terminal for user identification and door opening decisions. When the door lock is insufficient in electric quantity or fails, the identification function is often disabled, a user cannot normally pass, and the use experience and the system reliability are seriously affected. In addition, the prior art generally lacks a mechanism for effectively utilizing surrounding environment information and user portable equipment to carry out collaborative identity authentication under the condition of abnormal equipment or weak signals. Although some schemes provide a localized standby verification mode, the method has poor adaptability and insufficient security, and cannot realize dynamic and continuous user state tracking and decision sharing in a door lock network. Disclosure of Invention The invention aims to provide a method and a system for cooperatively unlocking a door lock, which are used for solving the defects in the prior art, realizing reliable identity confirmation and unlocking independent of independent operation of a main door lock when the main door lock is abnormal, and improving the robustness and user experience of the system. One embodiment of the application provides a method for cooperatively unlocking a door lock, which comprises the following steps: constructing an auxiliary recognition network, namely according to an auxiliary recognition request triggered by a target door lock when the target door lock detects a low-power or fault state, gradually spreading face feature vectors of a target user along left and right neighbor door lock chains of the target user, and constructing a distributed processing network for collaborative recognition; The accompanying relation detection and data synchronization is carried out, namely, a multicast tracking task is triggered to be established according to the event that the face of a target user is captured for the first time by any neighbor door lock in the distributed processing network, and the detected MAC address list of the mobile equipment and the signal intensity information thereof are synchronously collected and shared; Confirming a confidence coefficient accumulation and binding relation, namely continuously capturing the face of a target user and monitoring the event of the signal intensity change of the MAC address according to the subsequent door lock, and iteratively updating the binding confidence coefficient of each MAC address and the face through signal curve fitting and intersection calculation until a preset binding confirmation threshold value is reached; And executing collaborative decision and unlocking, namely initiating a door lock decision at the target door lock or by a task according to the confirmed binding relation and the signal characteristics of the portable equipment of the target user, and triggering the target door lock to execute unlocking action. Optionally, the auxiliary identification network construction includes: Generating low power detection and request, namely generating auxiliary identification request information comprising a target user face feature vector, a door lock identification and a request direction according to the fact that the target door lock detects that the self power is lower than a preset threshold value; constructing a neighbor relation chain, namely unicast transmitting the message to a left neighbor door lock and a right neighbor door lock prestored in the target door lock according to the direction mark in the request message; Step-by-step transmission of the request, namely analyzing and locally storing a face feature vector of a target user according to a neighbor door lock receiving the request, adding self identification to a neighbor relation linked list, and continuously forwarding updated information to the next neighbor door lock in the same direction until a link tail; And the distributed network is ready, wherein processing chain tables comprising a plurality of neighbor door locks are formed on two sides of the target door lock according to the propagation process, and each chain table node door lock enters an auxiliary identification mode to complete the construction of the distributed processing network. Optionally, the companion relation detection is synchronous with the data, including: The first detection trigger and task creation are that according to any neighbor door lock entering an auxiliary recognition mode, a target user face is successfully matched and captured, the door lock