Dual-mode barrier gate control system and method based on LoRa and Wi-Fi

Abstract

The invention discloses a dual-mode barrier gate control system and method based on LoRa and Wi-Fi. The method comprises the following steps: S1, system initialization; s2, dynamically selecting a communication mode by adopting a Markov decision process algorithm; s3, analyzing the historical behavior data of the visitor by combining a hypergeometric learning algorithm and a manifold learning algorithm, and generating a dynamic two-dimensional code; s4, encrypting the dynamic two-dimensional code, and transmitting the dynamic two-dimensional code to the visitor terminal through LoRa or Wi-Fi; s5, the visitor scans the two-dimensional code to request for passing, and the edge computing device locally verifies the validity of the two-dimensional code and device binding information; s6, when an abnormal behavior is detected, the system triggers a multi-factor authentication mechanism; s7, after the authentication is passed, the system controls a barrier gate execution mechanism to open a gate, and traffic information is recorded and uploaded to the cloud; and S8, the system collects visitor data for a long time, generates a behavior model by using big data, and optimizes an approval process and two-dimensional code management. According to the invention, the dynamic two-dimensional code is generated by combining LoRa and Wi-Fi dual-mode communication and an intelligent algorithm, so that the dynamic management of visitors is realized.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent access control, and in particular to a dual-mode gate access control system and method based on LoRa and Wi-Fi. Background Art

[0002] With the rapid development of Internet of Things technology, intelligent access control systems have been widely used, especially in large buildings, office parks, residential areas and other scenarios, where visitor management has become an important part of security management. Most existing access control systems rely on a single communication mode, such as Wi-Fi or wired networks, to transmit data and control access control devices. However, a single communication mode has many limitations when facing diverse and dynamically changing usage scenarios, especially in large-scale, multi-visitor, and complex environments, where its communication stability, real-time performance, and power consumption are insufficient. This situation leads to inefficient systems, poor user experience, and even security vulnerabilities.

[0003] Although traditional Wi-Fi access control systems perform well in short-range high-bandwidth data transmission, their communication distance is limited. The typical coverage range of Wi-Fi is about tens of meters to one hundred meters, which is difficult to meet the communication needs of long-distance scenarios, especially in large campuses or building complexes. The distance between visitors and access control systems often exceeds the effective coverage range of Wi-Fi. In addition, the high-bandwidth characteristics of Wi-Fi are also accompanied by high power consumption. In scenarios that require long standby or low power consumption, this high power consumption limits the application flexibility of the system. For example, for long-distance visitor devices, the Wi-Fi communication mode may cause the device to consume power quickly, which is not suitable for low-power scenarios that require long-term use.

[0004] Compared with Wi-Fi, LoRa (Low Power Wide Area Network) is a low-power, long-distance communication technology that is gradually emerging in IoT applications. LoRa has the advantages of long communication distance, low power consumption, and strong penetration. It can support communications within a range of several kilometers and performs well in scenarios that require frequent communications but small amounts of data. However, LoRa's low bandwidth and high latency characteristics limit its application in scenarios that require large amounts of data transmission and low-latency response, especially in complex, multi-device, and high-concurrency environments. LoRa's transmission speed and stability are poor, making it difficult to meet high bandwidth requirements.

[0005] Most of the current access control systems are based on static QR codes for identity authentication. Such static QR codes are usually permanently fixed after being generated and cannot be dynamically adjusted according to the visitor's real-time behavior, access frequency or environmental changes. This verification method has security risks. For example, if the QR code is copied and used after being screenshotted, the access control system cannot effectively verify the authenticity of the visitor's identity, which is prone to security vulnerabilities. On the other hand, the identity authentication process of traditional access control systems is relatively simple and often relies on fixed verification mechanisms, such as QR code scanning and password input, and lacks multi-level and multi-factor identity authentication methods. This fixed verification mechanism is difficult to deal with abnormal behaviors, such as abnormal access frequency, frequent access at night, etc., and it is impossible to take additional security verification measures for high-risk behaviors, which poses a security risk.

[0006] In addition, the access control system in the prior art usually relies on a centralized cloud computing platform for data processing and management. All visitor data and access control logic need to be transmitted and processed through the cloud. This centralized processing method is prone to excessive pressure on the cloud in high-concurrency scenarios, resulting in delays or slow system response, which reduces the real-time and stability of the system. Moreover, the processing method that relies on the cloud also increases the security risks in the data transmission process, which may cause the data to be stolen or tampered with during the transmission process.

[0007] For visitor behavior management, traditional access control systems mostly use static rules or simple access record analysis, lack in-depth analysis and learning of visitor historical behavior patterns, and cannot make intelligent predictions and decisions based on visitor historical data. Most current systems are difficult to achieve personalized management of visitor behavior, and cannot automatically make intelligent adjustments based on visitor access frequency, time period, and behavior patterns, resulting in a lack of adaptive capabilities in the access control system, which cannot dynamically adjust access control permissions and access policies based on the behavior of different visitors.

[0008] Therefore, how to provide a dual-mode gate access control system and method based on LoRa and Wi-Fi is a problem that technicians in this field urgently need to solve. Summary of the invention

[0009] One purpose of the present invention is to propose a dual-mode gate access control system and method based on LoRa and Wi-Fi. The present invention combines LoRa and Wi-Fi dual-mode communication technologies, and uses intelligent algorithms such as Markov decision algorithm, hypergeometric learning and manifold learning to dynamically analyze and predict visitors' geographic location, network status and historical behavior data, generate dynamic QR codes and optimize the approval process, thereby realizing intelligent management and rapid passage of visitors. Through edge computing and big data analysis, the system not only improves the flexibility and adaptability of communication, but also significantly improves the security of QR codes, effectively prevents the risk of abuse, and has the advantages of efficient, safe and intelligent management, which is particularly suitable for access control management needs in complex and multi-scenario situations.

[0010] A dual-mode gate access control method based on LoRa and Wi-Fi according to the present invention comprises the following steps: S1. Initialize the system. The gate control module establishes a connection with the cloud server through the edge computing device, loads the visitor management module, starts the communication module, and selects LoRa or Wi-Fi as the initial communication mode based on the dual-mode communication architecture; S2, the gate device uses the Markov decision process algorithm to dynamically select the communication mode according to the visitor's geographical location, network status and device status; S3, the gate equipment analyzes the historical behavior data of visitors, including visit frequency, time period and visit area, by combining hypergeometric learning and manifold learning algorithms, learns and predicts future visit time, and generates dynamic QR codes; S4. The dynamic QR code is encrypted with the visitor's behavior data, timestamp and device binding information, and transmitted to the visitor's terminal via LoRa or Wi-Fi; S5. Visitors request passage by scanning a dynamic QR code. The barrier device uses edge computing technology to process and verify the QR code in real time locally, including QR code validity, device binding verification, and dynamic validity period adjustment. S6. When the edge computing device detects abnormal behavior of the visitor, the system triggers the multi-factor authentication mechanism and requires the visitor to perform face recognition or SMS verification; S7. After passing the QR code verification and multi-factor authentication, the system controls the gate actuator to open the gate to allow visitors to pass. At the same time, the system uploads the pass records to the cloud server via LoRa or Wi-Fi for archiving and analysis; S8. The system collects and analyzes visitor access data over a long period of time, generates behavioral models through big data technology, and dynamically optimizes approval processes and QR code management.

[0011] Optionally, the S2 specifically includes: S21. Define a state space S, which includes the visitor's geographic location, network status, and device status. It includes geographic location L, network status N and device status E, expressed as s=(L,N,E); The geographical location L is represented by a distance range based on the relative distance between the visitor and the barrier device. The distance range is divided into three states: far, medium, and near, which represent the visitor's long distance, medium distance, and short distance from the barrier device, respectively. The longer distance adopts the LoRa communication mode, the medium distance adopts the Wi-Fi or LoRa communication mode, and the shorter distance adopts the Wi-Fi communication mode; S22. Define the action space ,in Indicates that LoRa communication mode is selected. Indicates the selection of Wi-Fi communication mode. Each action selects the communication mode based on the current state s. S23. Define the reward function R(s,a), which is determined by the visitor's geographic location, network status and device status: ; in, , and represents the weight coefficient; It indicates the benefits of communication mode selection based on geographic location. If the visitor is far away, choosing LoRa has a higher reward; Indicates the benefits of communication mode selection based on network status. Signal quality and network load determine the choice of communication mode. Indicates the benefits of communication mode selection based on device status, remaining battery and processing power determine the priority of selecting low power mode; S24, state transition probability It represents the probability of transferring from the current state s to the state s' after executing action a. The state transition takes into account the visitor's movement and network condition changes. The model is trained through historical data to predict the possible state changes of visitors. S25. Based on the Bellman equation of the Markov decision process, calculate the long-term return value V(s) of state s: ; in, represents the discount factor; S26. In each state, the optimal communication mode is selected by iteratively calculating the Bellman equation. .

[0012] Optionally, the S23 specifically includes: S231. The influence of geographical location on communication mode is reflected in distance. A nonlinear distance attenuation model is introduced to express the relationship between distance and communication mode. d is defined as the distance between the gate and the visitor device. For LoRa communication mode : ; in, Indicates the sensitivity of the complexity of the environment to the attenuation of the distance signal, Indicates the effective communication distance of LoRa. represents the nonlinear attenuation coefficient; For Wi-Fi communication mode : ; in, Indicates the environmental sensitivity coefficient of Wi-Fi, Indicates the optimal communication distance of Wi-Fi. represents the weight of the position influence, represents the smoothing term; S232, network status by signal strength and network load It was decided to introduce a nonlinear signal strength model and a logarithmic network load impact model to reflect the impact of signal quality and network load on rewards; For LoRa communication mode : ; in, and Indicates the signal strength and network load weight coefficient of the LoRa communication mode, represents the nonlinear increase of LoRa reward due to control signal strength, represents the signal strength threshold, It represents a constant term to prevent errors caused by taking the logarithm when the network load is zero; For Wi-Fi communication mode : ; in, and Indicates the signal strength and network load weight coefficient of the Wi-Fi communication mode. represents the nonlinear effect of controlling network load on Wi-Fi rewards, Indicates the threshold value of network load; S233, device status includes device battery level and processing power , introducing a quadratic polynomial model and logarithmic correction terms to reflect the balance between power and processing capacity; For LoRa communication mode : ; in, Indicates the control of the reward growth when the battery is low. Indicates the critical point of battery power. represents the processing capacity impact coefficient, represents the smoothing term; For Wi-Fi communication mode : ; in, Indicates the reward growth when the power is high. Indicates the high battery threshold, Indicates the penalty item when the battery is low.

[0013] Optionally, the S3 specifically includes: S31. Build a visitor behavior dataset ,in Represents the behavioral feature vector of the i-th visitor, including the visitor's access frequency , Time period and access area , expressed as , Indicates the actual access result of the i-th visitor, including the visitor's access time and whether the verification is passed; S32. Use the hypergeometric learning algorithm to model the visitor's behavioral characteristics in the negative curvature space, map the visitor's historical behavior data to the hypergeometric space, and assume that the space is a hypergeometric space with negative curvature k. , where d represents the embedding dimension of the data: ; in, Represents two behavioral features in hypergeometric space and The distance between represents the Euclidean norm, Represents the inverse function of hyperbolic cosine; S33. Generate behavioral pattern clusters by clustering visitor behavior characteristics in hypergeometric space , where each cluster Represent a class of similar visitor behavior patterns, and predict the visitor's future behavior patterns, including visit time and frequency, based on the characteristics of each cluster; S34, using a manifold learning algorithm to analyze the embedded structure of visitor behavior data in a high-dimensional space, extracting local and global geometric structures of behavior features, and mapping high-dimensional data to a low-dimensional manifold space through manifold learning; S35. Generate the visitor's future visit time based on the combination of hypergeometric learning and manifold learning A prediction model that analyzes visitors' historical visit periods and frequency , generating a probability distribution for predicted access times : ; in, The average visit time of visitors. represents the standard deviation of the time distribution, Z represents the normalization constant, and exp represents the exponential function; S36. Based on the predicted visitor behavior data, the system dynamically generates a QR code and dynamically adjusts the validity period of the QR code according to the visitor's behavior pattern.

[0014] Optionally, the S5 specifically includes: S51: After the visitor scans the QR code, the gate edge computing device receives the QR code data Q, which includes the encrypted QR code content. , Dynamic timestamp and dynamic behavior characteristics , the complete data of the QR code is represented as: ; Wherein, Enc represents a function for encryption using a key K; S52, the edge computing device encrypts the QR code content Decryption is performed using a hybrid encryption algorithm that combines asymmetric encryption and symmetric encryption. The decryption formula is: ; in, Indicates the device private key. After decryption, the original data of the QR code is obtained. ; S53. After decryption, the edge computing device extracts the dynamic timestamp in the QR code , by verifying whether the time when the visitor generated the QR code is within the time limit allowed by the system, to determine whether the QR code is expired: ; in, Indicates the current time. If , then the QR code is valid. Indicates the maximum time difference allowed; In order to prevent the timestamp from being tampered with, the system uses a hash chain verification method to verify the timestamp and the encrypted QR code content in the QR code. Perform hash binding and the verification formula is: ; Where H represents the hash function, Indicates the hash value pre-stored in the system; S54, the edge computing device verifies whether the device bound to the QR code is an authorized device, and uniquely identifies the visitor's device and For comparison, the device binding verification formula is: ; Where d represents the similarity measurement function of device features, represents the matching threshold; S55. Extract the visitor's dynamic behavior characteristics from the QR code after decryption , edge computing devices are based on visitors’ historical visit behavior data Real-time matching is performed. The behavioral characteristics include the visitor's visit frequency, visit period, and visit area. The similarity between the current behavior and the historical behavior is calculated through the behavioral pattern matching algorithm in the hypergeometric space: ; Among them, Sim represents the cosine similarity of behavioral features. If , then the authentication is successful. Indicates the set similarity threshold; S56. Based on the verification of the QR code, the system dynamically adjusts the verification according to the risk level of the visitor's current visit behavior and calculates the risk level Risk: ; in, , and represents the weight factor, if , then multi-factor authentication is triggered.

[0015] Optionally, the S8 specifically includes: S81. Every time a visitor passes through, the system collects multi-dimensional data through the gate equipment and visitor terminal equipment. The data includes the visitor's identity , access the device's hardware status information , QR code verification result and access behavior trajectory data , the data set is represented as ,After the initial preprocessing by the edge computing device, all the data is uploaded to the cloud server based on the distributed data architecture to complete the initial cleaning and standardization; S82. The system extracts behavioral features from the data set through machine learning algorithms. , the system analyzes the visitor's device usage patterns, access location shifts, and network performance changes, and constructs a multi-dimensional behavioral feature vector : ; in, Characteristic function representing the state of the device, Characteristic function representing network performance, The characteristic function representing the behavior trajectory is used to classify the behavior pattern of each visitor through the support vector machine algorithm to determine whether there is abnormal behavior. If abnormal, the If the set threshold is exceeded, the system marks it as a high-risk behavior: ; in, Represents the weight factor of each feature dimension; S83. For each visitor, the system generates an adaptive behavior model based on the behavior characteristics. , the behavior model The recursive neural network captures the temporal characteristics of visitor behavior, and the system predicts the visitor's future visit behavior to obtain the prediction vector ; S84, based on behavioral model The system dynamically optimizes the approval process based on the output of the visitor. Through the reinforcement learning algorithm, the system adjusts the approval process according to the visitor's behavioral characteristics and prediction results: ; Among them, EFF represents approval efficiency, Represents the risk indicator, The approval process is time-consuming. and They represent the parameters that weigh approval efficiency and security. When the approval efficiency EFF exceeds the threshold, the system automatically skips some approval steps. S85, system based on behavioral model Dynamically adjust the QR code management strategy and the validity period of the QR code And use conditions to adaptively update based on visitor behavior predictions: ; in, Indicates the initial validity period. represents the adjustment factor, represents the access behavior prediction function at time t; S86, the system is based on long-term collection of visitor access data and uses a greedy strategy to model visitor behavior Perform periodic updates.

[0016] According to the present invention, a dual-mode gate access control system based on LoRa and Wi-Fi includes the following modules: System initialization module: used to initialize the gate control module, establish a connection with the cloud server through the edge computing device, load the visitor management module, and start the communication module; Dynamic communication selection module: used to select dynamic communication mode based on the visitor's geographic location, network status and device status using the Markov decision process algorithm; Dynamic QR code generation module: used to combine hypergeometric learning and manifold learning algorithms to analyze visitors' historical behavior data and generate dynamic QR codes; Edge computing verification module: used to decrypt the visitor's dynamic QR code, verify the timestamp, verify the device binding, and match the behavior characteristics, and trigger multi-factor authentication as needed; Passage control and recording module: used to control the opening of the gate after the QR code verification is passed, and upload the passage record to the cloud server for archiving; Behavioral data analysis and optimization module: used to collect and analyze visitor access data, generate behavioral models, and dynamically optimize approval processes and QR code management.

[0017] The beneficial effects of the present invention are: First of all, the present invention successfully solves the limitations of a single communication mode in different scenarios in the prior art by introducing a dual-mode communication mechanism of LoRa and Wi-Fi. As a low-power, long-distance communication technology, LoRa can maintain stable communication with low power consumption when the visitor is far away from the gate, meeting the communication needs in large-scale parks, building complexes or remote monitoring scenarios. Wi-Fi is suitable for use in close-range, high-bandwidth demand scenarios, and can provide higher communication speeds and data transmission capabilities to ensure smooth information exchange during QR code generation and verification. Through the Markov decision process algorithm, the present invention can dynamically switch the LoRa or Wi-Fi communication mode according to the visitor's geographic location, network status, and device status, thereby automatically selecting the optimal communication method in different scenarios, greatly improving the adaptability and flexibility of the system.

[0018] Secondly, the innovation of the present invention in the management of QR codes overcomes the problem of insufficient security of traditional static QR codes. The generation and management method of dynamic QR codes can dynamically adjust the validity period and permissions of QR codes according to factors such as visitor's visit frequency, visit period, and device status by performing real-time analysis and prediction on visitor's behavior patterns. Compared with traditional static QR codes, the dynamic QR code technology of the present invention effectively prevents the risk of QR codes being screenshot, copied or abused, and improves the security of the system. At the same time, through the mechanism of binding the QR code to the device, even if the QR code is copied to other devices, the system can still prevent unauthorized devices from using the QR code through the device verification mechanism, further improving the security protection capability of the access control system.

[0019] The present invention achieves in-depth analysis and accurate prediction of visitor behavior patterns by combining hypergeometric learning with manifold learning algorithms. This innovation enables the system to automatically learn visitor access patterns based on historical behavior data, predict future visit behaviors, and dynamically generate more accurate QR codes. It can also optimize visitor approval processes in advance and reduce the workload of manual intervention. This behavior prediction mechanism based on intelligent analysis not only improves the efficiency of visitor access, but also enhances the system's automation level and intelligent management capabilities.

[0020] In terms of data processing, the present invention adopts edge computing technology to effectively reduce the load on cloud servers. Especially in high-concurrency scenarios, edge computing can complete QR code verification and visitor behavior feature matching locally, greatly shortening the response time and improving the real-time and stability of the system. Compared with traditional systems that rely on centralized cloud processing, the present invention reduces the delay of data transmission by distributing part of the data processing tasks to edge computing nodes, and improves the system's fault recovery capability, ensuring the continuity and security of access control. In addition, the introduction of edge computing also reduces the system's dependence on network conditions. When the network is unstable or the cloud load is high, the system can still maintain a high service quality and response speed.

[0021] In addition, the system of the present invention uses big data analysis technology to generate personalized behavior models by collecting visitor access data over a long period of time, and dynamically adjusts visitor approval processes and access strategies based on these models. This management method based on big data and intelligent analysis enables the system to continuously optimize over time, adapting to the behavioral characteristics of each visitor more and more accurately, improving the system's automated management efficiency and reducing the workload of manual approval. This not only improves the system's operating efficiency, but also enhances the system's adaptability in complex scenarios, enabling it to cope with changes in visitor traffic of different scales. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings: Figure 1 A flow chart of a dual-mode gate access control method based on LoRa and Wi-Fi proposed by the present invention; Figure 2 This is a schematic diagram of the structure of a dual-mode gate access control system based on LoRa and Wi-Fi proposed by the present invention. DETAILED DESCRIPTION

[0023] The present invention will now be described in further detail with reference to the accompanying drawings. These drawings are simplified schematic diagrams, which only illustrate the basic structure of the present invention in a schematic manner, and therefore only show the components related to the present invention.

[0024] refer to Figure 1 , a dual-mode gate access control method based on LoRa and Wi-Fi, comprising the following steps: S1. Initialize the system. The gate control module establishes a connection with the cloud server through the edge computing device, loads the visitor management module, starts the communication module, and selects LoRa or Wi-Fi as the initial communication mode based on the dual-mode communication architecture; S2, the gate device uses the Markov decision process algorithm to dynamically select the communication mode according to the visitor's geographical location, network status and device status; S3, the gate equipment analyzes the historical behavior data of visitors, including visit frequency, time period and visit area, by combining hypergeometric learning and manifold learning algorithms, learns and predicts future visit time, and generates dynamic QR codes; S4. The dynamic QR code is encrypted with the visitor's behavior data, timestamp and device binding information, and transmitted to the visitor's terminal via LoRa or Wi-Fi; S5. Visitors request passage by scanning a dynamic QR code. The barrier device uses edge computing technology to process and verify the QR code in real time locally, including QR code validity, device binding verification, and dynamic validity period adjustment. S6. When the edge computing device detects abnormal behavior of the visitor, the system triggers the multi-factor authentication mechanism and requires the visitor to perform face recognition or SMS verification; S7. After passing the QR code verification and multi-factor authentication, the system controls the gate actuator to open the gate to allow visitors to pass. At the same time, the system uploads the pass records to the cloud server via LoRa or Wi-Fi for archiving and analysis; S8. The system collects and analyzes visitor access data over a long period of time, generates behavioral models through big data technology, and dynamically optimizes approval processes and QR code management.

[0025] In this implementation, S2 specifically includes: S21. Define a state space S, which includes the visitor's geographic location, network status, and device status. It includes geographic location L, network status N and device status E, expressed as s=(L,N,E); The geographical location L is represented by a distance range based on the relative distance between the visitor and the barrier device. The distance range is divided into three states: far, medium, and near, which represent the visitor's long distance, medium distance, and short distance from the barrier device, respectively. The longer distance adopts the LoRa communication mode, the medium distance adopts the Wi-Fi or LoRa communication mode, and the shorter distance adopts the Wi-Fi communication mode; S22. Define the action space ,in Indicates that LoRa communication mode is selected. Indicates the selection of Wi-Fi communication mode. Each action selects the communication mode based on the current state s. S23. Define the reward function R(s,a), which is determined by the visitor's geographic location, network status and device status: ; in, , and represents the weight coefficient; It indicates the benefits of communication mode selection based on geographical location. If the visitor is far away, choosing LoRa has a higher reward; Indicates the benefits of communication mode selection based on network status. Signal quality and network load determine the choice of communication mode. Indicates the benefits of communication mode selection based on device status, remaining battery and processing power determine the priority of selecting low power mode; S24, state transition probability It represents the probability of transferring from the current state s to the state s' after executing action a. The state transition takes into account the visitor's movement and network condition changes. The model is trained through historical data to predict the possible state changes of visitors. S25. Based on the Bellman equation of the Markov decision process, calculate the long-term return value V(s) of state s: ; in, represents the discount factor; S26. In each state, the optimal communication mode is selected by iteratively calculating the Bellman equation. .

[0026] In this implementation manner, the S23 specifically includes: S231. The influence of geographical location on communication mode is reflected in distance. A nonlinear distance attenuation model is introduced to express the relationship between distance and communication mode. d is defined as the distance between the gate and the visitor device. For LoRa communication mode : ; in, Indicates the sensitivity of the complexity of the environment to the attenuation of the distance signal. Indicates the effective communication distance of LoRa. represents the nonlinear attenuation coefficient; For Wi-Fi communication mode : ; in, Indicates the environmental sensitivity coefficient of Wi-Fi, Indicates the optimal communication distance of Wi-Fi. represents the weight of the position influence, represents the smoothing term; S232, network status by signal strength and network load It was decided to introduce a nonlinear signal strength model and a logarithmic network load impact model to reflect the impact of signal quality and network load on rewards; For LoRa communication mode : ; in, and Indicates the signal strength and network load weight coefficient of the LoRa communication mode, represents the nonlinear increase of LoRa reward due to control signal strength, The threshold value of signal strength. It represents a constant term to prevent errors caused by taking the logarithm when the network load is zero; For Wi-Fi communication mode : ; in, and Indicates the signal strength and network load weight coefficient of the Wi-Fi communication mode. represents the nonlinear effect of controlling network load on Wi-Fi rewards, Indicates the threshold value of network load; S233, device status includes device battery level and processing power , introducing a quadratic polynomial model and logarithmic correction terms to reflect the balance between power and processing capacity; For LoRa communication mode : ; in, Indicates the control of the reward growth when the battery is low. Indicates the critical point of battery power. represents the processing capacity impact coefficient, represents the smoothing term; For Wi-Fi communication mode : ; in, Indicates the increase in rewards when the battery level is high. Indicates the high battery threshold, Indicates the penalty term when the battery is low.

[0027] In this implementation, S3 specifically includes: S31. Build a visitor behavior dataset ,in Represents the behavioral feature vector of the i-th visitor, including the visitor's access frequency , Time period and access area , expressed as , Indicates the actual access result of the i-th visitor, including the visitor's access time and whether the verification is passed; S32. Use the hypergeometric learning algorithm to model the visitor's behavioral characteristics in the negative curvature space, map the visitor's historical behavior data to the hypergeometric space, and assume that the space is a hypergeometric space with negative curvature k. , where d represents the embedding dimension of the data: ; in, Represents two behavioral features in hypergeometric space and The distance between represents the Euclidean norm, Represents the inverse function of hyperbolic cosine; S33. Generate behavioral pattern clusters by clustering visitor behavior characteristics in hypergeometric space , where each cluster Represent a class of similar visitor behavior patterns, and predict the visitor's future behavior patterns, including visit time and frequency, based on the characteristics of each cluster; S34, using a manifold learning algorithm to analyze the embedded structure of visitor behavior data in a high-dimensional space, extracting local and global geometric structures of behavior features, and mapping high-dimensional data to a low-dimensional manifold space through manifold learning; S35. Generate the visitor's future visit time based on the combination of hypergeometric learning and manifold learning A prediction model that analyzes visitors' historical visit periods and frequency , generate a probability distribution for predicting access time : ; in, The average visit time of visitors. represents the standard deviation of the time distribution, Z represents the normalization constant, and exp represents the exponential function; S36. Based on the predicted visitor behavior data, the system dynamically generates a QR code and dynamically adjusts the validity period of the QR code according to the visitor's behavior pattern.

[0028] In this implementation manner, S5 specifically includes: S51: After the visitor scans the QR code, the gate edge computing device receives the QR code data Q, which includes the encrypted QR code content. , Dynamic timestamp and dynamic behavior characteristics , the complete data representation of the QR code is: ; Wherein, Enc represents a function for encryption using a key K; S52: Edge computing device encrypts the QR code content Decryption is performed using a hybrid encryption algorithm that combines asymmetric encryption and symmetric encryption. The decryption formula is: ; in, Indicates the device private key. After decryption, the original data of the QR code is obtained. ; S53. After decryption, the edge computing device extracts the dynamic timestamp in the QR code , by verifying whether the time when the visitor generated the QR code is within the time limit allowed by the system, to determine whether the QR code is expired: ; in, Indicates the current time. If , then the QR code is valid. Indicates the maximum time difference allowed; In order to prevent the timestamp from being tampered with, the system uses a hash chain verification method to verify the timestamp and the encrypted QR code content in the QR code. Perform hash binding and the verification formula is: ; Where H represents the hash function, Indicates the hash value pre-stored in the system; S54, the edge computing device verifies whether the device bound to the QR code is an authorized device, and uniquely identifies the visitor's device and For comparison, the device binding verification formula is: ; Where d represents the similarity measurement function of device features, represents the matching threshold; S55. Extract the visitor's dynamic behavior characteristics from the QR code after decryption , edge computing devices are based on visitors’ historical visit behavior data Real-time matching is performed. The behavioral characteristics include the visitor's visit frequency, visit period, and visit area. The similarity between the current behavior and the historical behavior is calculated through the behavioral pattern matching algorithm in the hypergeometric space: ; Among them, Sim represents the cosine similarity of behavioral features. If , then the authentication is successful. Indicates the set similarity threshold; S56. Based on the verification of the QR code, the system dynamically adjusts the verification according to the risk level of the visitor's current visit behavior and calculates the risk level Risk: ; in, , and represents the weight factor, if , then multi-factor authentication is triggered.

[0029] In this implementation manner, S8 specifically includes: S81. Every time a visitor passes through, the system collects multi-dimensional data through the gate equipment and visitor terminal equipment. The data includes the visitor's identity , access the device's hardware status information , QR code verification result and access behavior trajectory data , the data set is represented as ,After the initial preprocessing by the edge computing device, all the data is uploaded to the cloud server based on the distributed data architecture to complete the initial cleaning and standardization; S82. The system extracts behavioral features from the data set through machine learning algorithms. , the system analyzes the visitor's device usage patterns, access location shifts, and network performance changes, and constructs a multi-dimensional behavioral feature vector : ; in, Characteristic function representing the state of the device, Characteristic function representing network performance, The characteristic function representing the behavior trajectory is used to classify the behavior pattern of each visitor through the support vector machine algorithm to determine whether there is abnormal behavior. If abnormal, the If the set threshold is exceeded, the system marks it as a high-risk behavior: ; in, Represents the weight factor of each feature dimension; S83. For each visitor, the system generates an adaptive behavior model based on the behavior characteristics. , the behavior model The recursive neural network captures the temporal characteristics of visitor behavior, and the system predicts the visitor's future visit behavior to obtain the prediction vector ; S84, based on behavioral model The system dynamically optimizes the approval process based on the output of the visitor. Through the reinforcement learning algorithm, the system adjusts the approval process according to the visitor's behavioral characteristics and prediction results: ; Among them, EFF represents approval efficiency, Represents the risk indicator, The approval process is time-consuming. and They represent the parameters that weigh approval efficiency and security. When the approval efficiency EFF exceeds the threshold, the system automatically skips some approval steps. S85, system based on behavioral model Dynamically adjust the QR code management strategy and the validity period of the QR code And use conditions to adaptively update based on visitor behavior predictions: ; in, Indicates the initial validity period. represents the adjustment factor, represents the access behavior prediction function at time t; S86, the system is based on long-term collection of visitor access data and uses a greedy strategy to model visitor behavior Perform periodic updates.

[0030] refer to Figure 2 , a dual-mode gate access control system based on LoRa and Wi-Fi, including the following modules: System initialization module: used to initialize the gate control module, establish a connection with the cloud server through the edge computing device, load the visitor management module, and start the communication module; Dynamic communication selection module: used to select dynamic communication mode based on the visitor's geographic location, network status and device status using the Markov decision process algorithm; Dynamic QR code generation module: used to combine hypergeometric learning and manifold learning algorithms to analyze visitors' historical behavior data and generate dynamic QR codes; Edge computing verification module: used to decrypt the visitor's dynamic QR code, verify the timestamp, verify the device binding, and match the behavior characteristics, and trigger multi-factor authentication as needed; Passage control and recording module: used to control the opening of the gate after the QR code verification is passed, and upload the passage record to the cloud server for archiving; Behavioral data analysis and optimization module: used to collect and analyze visitor access data, generate behavioral models, and dynamically optimize approval processes and QR code management.

[0031] Example: In order to verify the feasibility of the present invention in implementation, the present invention is applied to a large-scale smart office park. Traditional access control systems often lead to unstable communication, inefficient visitor management, and insufficient access security due to problems such as single communication mode and static generation of QR codes. In order to cope with this situation, the dual-mode gate access control system based on LoRa and Wi-Fi of the present invention is applied to the park, realizing accurate management of visitor identities and efficient access control.

[0032] The smart park covers an area of ​​10 square kilometers, with relatively scattered buildings inside, and the access control management system needs to handle a large number of visitor access requests. In the past, the park access control system's Wi-Fi-based communication method had insufficient coverage in long-distance scenarios, making it difficult for some long-distance visitors to be effectively authenticated through the wireless network. At the same time, the traditional system uses static QR codes for identity authentication. The QR codes have a fixed validity period and are not bound to the device. They are easy to be copied by screenshots, and multiple security vulnerabilities have occurred during the system's high concurrency period. These problems have brought many security risks and management difficulties to the daily operations of the park.

[0033] The present invention optimizes and upgrades the access control system by introducing LoRa and Wi-Fi dual-mode communication technology and combining it with intelligent algorithms. A total of 100 gate control devices are deployed in the park, and edge computing devices are also installed in each gate to process real-time QR code verification and visitor behavior analysis. The visitor's geographic location, device status, and network status are dynamically sensed by the system. According to the distance between the visitor and the gate and the current network status, the system can flexibly switch the communication mode between LoRa and Wi-Fi, thereby ensuring the communication stability between visitors and the system at different distances and in different environments.

[0034] In actual applications, when visitors are far away from the gate, the system uses the LoRa mode to achieve low-power, long-distance data transmission. Especially in long-distance scenarios where the distance from the gate is more than 500 meters, the stability and low-power characteristics of LoRa communication are fully utilized. Through actual testing, the battery life of visitor terminal devices in LoRa mode has been increased by more than 30%, and stable data transmission can be maintained at a distance of 800 meters from the gate. When visitors approach the gate, the system automatically switches to Wi-Fi mode to provide high-speed, low-latency data transmission. In this scenario, the Wi-Fi mode not only improves the efficiency of QR code transmission, but also ensures efficient response during access control verification. The average waiting time for visitors to pass is shortened by about 50%.

[0035] The system also combines intelligent behavior prediction algorithms and uses hypergeometric learning and manifold learning techniques to analyze visitors' historical behaviors. By collecting data on visitor frequency, time period, and region in the past month, the system generates a personalized behavior model. When a visitor applies for a visit, the system automatically predicts his or her possible visit time and passage requirements based on historical data, and generates a QR code with dynamic adjustment functions in advance. On this basis, the QR code can not only be updated in real time as visitor behavior changes, but also prevent the QR code from being copied by screenshots through binding technology with visitor devices. Based on the dual mechanisms of device binding and timestamp verification, the system significantly improves the security of QR codes. In actual scenarios, within the first month after the system was deployed, the once frequent QR code copying and abuse incidents were reduced by about 80%, and the passage efficiency of the entire park was improved by about 25%.

[0036] In addition, the edge computing device of the present invention significantly improves the response speed of the system. Traditional access control systems often have high delays due to excessive pressure on cloud servers during high-concurrency access, especially during peak hours in the morning and evening (such as 8:00 to 9:30 every morning), when the cloud processing load is large, and the average waiting time for visitors at the access control exceeds 45 seconds. After the introduction of edge computing, the system directly completes core operations such as QR code decryption, timestamp verification, device binding verification, and visitor behavior feature matching locally on the gate device, avoiding the delay problem of uploading all data to the cloud for processing. After three months of testing during peak hours in the morning and evening, the average waiting time for visitors at the access control was reduced to less than 15 seconds, and the response speed was increased by about 66%.

[0037] At the same time, the behavioral data analysis and approval process optimization functions of the present invention further enhance the intelligence level of the system. The system automatically generates personalized behavior models by collecting and analyzing visitor access data over a long period of time. For those visitors with high visit frequency and stable behavior patterns, the system can automatically optimize the approval process based on their historical behavior, reduce unnecessary security verification steps, and improve traffic efficiency. In the past two months, 40% of high-frequency visitors in the visitor approval process have had their approval steps automatically optimized, saving visitors' waiting time when passing through, and improving the overall traffic efficiency of visitors by about 20%.

[0038] Table 1 Performance comparison between traditional access control system and the system of the present invention ; Based on Table 1 above, the LoRa and Wi-Fi dual-mode gate access control system shows significant advantages in many aspects compared to the traditional single Wi-Fi access control system.

[0039] First of all, the diversity of communication modes is a highlight of the present invention. Traditional systems rely solely on Wi-Fi for communication, but the communication distance of Wi-Fi is greatly limited, usually unable to cover larger scenes within 100 meters. In contrast, the present invention combines LoRa and Wi-Fi dual-mode communication technology, and can intelligently select the appropriate communication mode according to the visitor's geographic location, device status and network conditions, realizing seamless switching between using high-bandwidth Wi-Fi at close range and using low-power LoRa communication at long distances. Test data shows that LoRa's communication distance can reach 1,000 meters, solving the problem of difficult communication for long-distance visitors in scenarios such as large-scale parks.

[0040] In terms of the average waiting time of visitors, the present invention uses edge computing technology to process the decryption and verification of the QR code locally, thereby reducing the delay caused by relying on cloud computing. The average waiting time of the traditional system is 45 seconds, while the system of the present invention shortens this to 15 seconds, reducing the waiting time by about 66%. Especially in high-concurrency access scenarios, edge computing improves the real-time response capability of the system and ensures that visitors can pass quickly.

[0041] In addition, the present invention greatly improves the security of QR codes. In traditional systems, static QR codes are easily screenshotted and maliciously copied, with an average of 20 QR code abuse incidents per month. The present invention uses dynamic QR code technology, which effectively avoids QR code abuse by binding with visitor devices and timestamp verification, reducing QR code abuse incidents to only 4 times per month, and improving security by 80%.

[0042] In terms of device energy consumption, the introduction of the LoRa communication mode reduces power consumption during long-distance communication. Traditional systems rely only on Wi-Fi, resulting in high battery consumption in long-distance scenarios. The system of the present invention extends the battery life of the device by 30% by intelligently switching to low-power LoRa communication. This is of great practical significance in scenarios that require long-term operation and long communication distances.

[0043] Concurrent processing capability is also a significant advantage of the system of the present invention. The average concurrent volume of the traditional system during the daily peak period (8:00-9:30) is 200 times, while the present invention combines edge computing and dual-mode communication to increase the concurrent processing capability to 300 times, an increase of 50%. This improvement ensures that the system can maintain stable operation under high concurrency conditions and avoids system crashes or access delays caused by excessive load.

[0044] In addition, the present invention also demonstrates higher accuracy and efficiency in device verification. Due to the imperfect device binding mechanism, the failure rate of device verification in traditional systems is as high as 10%. By adopting the dual mechanism of device binding and behavioral feature verification, the present invention reduces the device verification failure rate to 1%, improving the reliability of visitor verification.

[0045] In general, the LoRa and Wi-Fi dual-mode gate access control system of the present invention not only improves the stability and efficiency in long-distance and high-concurrency scenarios through dual-mode communication, but also performs well in visitor management security and equipment energy consumption. Whether it is security, energy consumption optimization, or concurrent processing capability and response speed, the present invention is significantly superior to traditional access control systems.

[0046] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A dual-mode gate access control method based on LoRa and Wi-Fi, characterized in that: The steps include: S1. Initialize the system. The gate control module establishes a connection with the cloud server through the edge computing device, loads the visitor management module, starts the communication module, and selects LoRa or Wi-Fi as the initial communication mode based on the dual-mode communication architecture; S2, the gate device uses the Markov decision process algorithm to dynamically select the communication mode according to the visitor's geographical location, network status and device status; S3, the gate equipment analyzes the historical behavior data of visitors, including the visit frequency, time period and visit area, by combining hypergeometric learning and manifold learning algorithms, learns and predicts future visit time, and generates dynamic QR codes; S4. The dynamic QR code is encrypted with the visitor's behavior data, timestamp and device binding information, and transmitted to the visitor's terminal via LoRa or Wi-Fi; S5. Visitors request passage by scanning a dynamic QR code. The barrier device uses edge computing technology to process and verify the QR code in real time locally, including QR code validity, device binding verification, and dynamic validity period adjustment. S6. When the edge computing device detects abnormal behavior of the visitor, the system triggers the multi-factor authentication mechanism and requires the visitor to perform face recognition or SMS verification; S7. After passing the QR code verification and multi-factor authentication, the system controls the gate actuator to open the gate to allow visitors to pass. At the same time, the system uploads the pass records to the cloud server via LoRa or Wi-Fi for archiving and analysis; S8. The system collects and analyzes visitor access data over a long period of time, generates behavioral models through big data technology, and dynamically optimizes approval processes and QR code management.

2. According to claim 1, a dual-mode gate access control method based on LoRa and Wi-Fi is characterized in that: The S2 specifically includes: S21. Define a state space S, which includes the visitor's geographic location, network status, and device status. It includes geographic location L, network status N and device status E, expressed as s=(L,N,E); The geographical location L is represented by a distance range based on the relative distance between the visitor and the barrier device. The distance range is divided into three states: far, medium, and near, which represent the visitor's long distance, medium distance, and short distance from the barrier device, respectively. The longer distance adopts the LoRa communication mode, the medium distance adopts the Wi-Fi or LoRa communication mode, and the shorter distance adopts the Wi-Fi communication mode; S22. Define the action space ,in Indicates that LoRa communication mode is selected. Indicates the selection of Wi-Fi communication mode. Each action selects the communication mode based on the current state s. S23. Define the reward function R(s,a), which is determined by the visitor's geographic location, network status and device status: ; in, , and represents the weight coefficient; It indicates the benefits of communication mode selection based on geographical location. If the visitor is far away, choosing LoRa has a higher reward; Indicates the benefits of communication mode selection based on network status. Signal quality and network load determine the choice of communication mode. Indicates the benefits of communication mode selection based on device status, remaining battery and processing power determine the priority of selecting low power mode; S24, state transition probability It represents the probability of transferring from the current state s to the state s' after executing action a. The state transition takes into account the visitor's movement and network condition changes. The model is trained through historical data to predict the possible state changes of visitors. S25. Based on the Bellman equation of the Markov decision process, calculate the long-term return value V(s) of state s: ; in, represents the discount factor; S26. In each state, the optimal communication mode is selected by iteratively calculating the Bellman equation. .

3. A dual-mode gate access control method based on LoRa and Wi-Fi according to claim 2, characterized in that: The S23 specifically includes: S231. The influence of geographical location on communication mode is reflected in distance. A nonlinear distance attenuation model is introduced to express the relationship between distance and communication mode. d is defined as the distance between the gate and the visitor device. For LoRa communication mode : ; in, Indicates the sensitivity of the complexity of the environment to the attenuation of the distance signal. Indicates the effective communication distance of LoRa. represents the nonlinear attenuation coefficient; For Wi-Fi communication mode : ; in, Indicates the environmental sensitivity coefficient of Wi-Fi, Indicates the optimal communication distance of Wi-Fi. represents the weight of the position influence, represents the smoothness term; S232, network status by signal strength and network load It was decided to introduce a nonlinear signal strength model and a logarithmic network load impact model to reflect the impact of signal quality and network load on rewards; For LoRa communication mode : ; in, and Indicates the signal strength and network load weight coefficient of the LoRa communication mode, represents the nonlinear increase of LoRa reward due to control signal strength, represents the signal strength threshold, It represents a constant term to prevent errors caused by taking the logarithm when the network load is zero; For Wi-Fi communication mode : ; in, and Indicates the signal strength and network load weight coefficient of the Wi-Fi communication mode. represents the nonlinear effect of controlling network load on Wi-Fi rewards, Indicates the threshold value of network load; S233, device status includes device battery level and processing power , introducing a quadratic polynomial model and logarithmic correction terms to reflect the balance between power and processing capacity; For LoRa communication mode : ; in, Indicates the control of the reward growth when the battery is low. Indicates the critical point of battery power. represents the processing capacity impact coefficient, represents the smoothness term; For Wi-Fi communication mode : ; in, Indicates the reward growth when the power is high. Indicates the high battery threshold, Indicates the penalty item when the battery is low.

4. According to a dual-mode gate access control method based on LoRa and Wi-Fi according to claim 1, it is characterized in that: The S3 specifically includes: S31. Build a visitor behavior dataset ,in Represents the behavioral feature vector of the i-th visitor, including the visitor's access frequency , Time period and access area , expressed as , Indicates the actual access result of the i-th visitor, including the visitor's passing time and whether the verification is passed; S32. Use the hypergeometric learning algorithm to model the visitor's behavioral characteristics in the negative curvature space, map the visitor's historical behavior data to the hypergeometric space, and assume that the space is a hypergeometric space with negative curvature k. , where d represents the embedding dimension of the data: ; in, Represents two behavioral features in hypergeometric space and The distance between represents the Euclidean norm, Represents the inverse function of hyperbolic cosine; S33. Generate behavioral pattern clusters by clustering visitor behavior characteristics in hypergeometric space , where each cluster Represent a class of similar visitor behavior patterns, and predict the visitor's future behavior patterns, including visit time and frequency, based on the characteristics of each cluster; S34, using a manifold learning algorithm to analyze the embedded structure of visitor behavior data in a high-dimensional space, extracting local and global geometric structures of behavior features, and mapping high-dimensional data to a low-dimensional manifold space through manifold learning; S35. Generate the visitor's future visit time based on the combination of hypergeometric learning and manifold learning A prediction model that analyzes visitors' historical visit periods and frequency , generate a probability distribution for predicting access time : ; in, The average visit time of visitors. represents the standard deviation of the time distribution, Z represents the normalization constant, and exp represents the exponential function; S36. Based on the predicted visitor behavior data, the system dynamically generates a QR code and dynamically adjusts the validity period of the QR code according to the visitor's behavior pattern.

5. A dual-mode gate access control method based on LoRa and Wi-Fi according to claim 1, characterized in that: The S5 specifically includes: S51: After the visitor scans the QR code, the gate edge computing device receives the QR code data Q, which includes the encrypted QR code content. , Dynamic timestamp and dynamic behavior characteristics , the complete data representation of the QR code is: ; Wherein, Enc represents a function for encryption using a key K; S52, the edge computing device encrypts the QR code content Decryption is performed using a hybrid encryption algorithm that combines asymmetric encryption and symmetric encryption. The decryption formula is: ; in, Indicates the device private key. After decryption, the original data of the QR code is obtained. ; S53. After decryption, the edge computing device extracts the dynamic timestamp in the QR code , by verifying whether the time when the visitor generated the QR code is within the time limit allowed by the system, to determine whether the QR code is expired: ; in, Indicates the current time. If , then the QR code is valid. Indicates the maximum time difference allowed; In order to prevent the timestamp from being tampered with, the system uses a hash chain verification method to verify the timestamp and the encrypted QR code content in the QR code. Perform hash binding and the verification formula is: ; Where H represents the hash function, Indicates the hash value pre-stored in the system; S54, the edge computing device verifies whether the device bound to the QR code is an authorized device, and uniquely identifies the visitor's device and For comparison, the device binding verification formula is: ; Where d represents the similarity measurement function of device features, represents the matching threshold; S55. Extract the visitor's dynamic behavior characteristics from the QR code after decryption , edge computing devices are based on visitors’ historical visit behavior data Real-time matching is performed. The behavioral characteristics include the visitor's visit frequency, visit period, and visit area. The similarity between the current behavior and the historical behavior is calculated through the behavioral pattern matching algorithm in the hypergeometric space: ; Among them, Sim represents the cosine similarity of behavioral features. If , then the authentication is successful. Indicates the set similarity threshold; S56. Based on the verification of the QR code, the system dynamically adjusts the verification according to the risk level of the visitor's current visit behavior and calculates the risk level Risk: ; in, , and represents the weight factor, if , then multi-factor authentication is triggered.

6. A dual-mode gate access control method based on LoRa and Wi-Fi according to claim 1, characterized in that: The S8 specifically includes: S81. Every time a visitor passes through, the system collects multi-dimensional data through the gate equipment and visitor terminal equipment. The data includes the visitor's identity , access the device's hardware status information , QR code verification result and access behavior trajectory data , the data set is represented as ,After the initial preprocessing by the edge computing device, all the data is uploaded to the cloud server based on the distributed data architecture to complete the initial cleaning and standardization; S82. The system extracts behavioral features from the data set through machine learning algorithms. , the system analyzes the visitor's device usage patterns, access location shifts, and network performance changes, and constructs a multi-dimensional behavioral feature vector : ; in, Characteristic function representing the state of the device, Characteristic function representing network performance, The characteristic function representing the behavior trajectory is used to classify the behavior pattern of each visitor through the support vector machine algorithm to determine whether there is abnormal behavior. If abnormal, the If the set threshold is exceeded, the system marks it as a high-risk behavior: ; in, Represents the weight factor of each feature dimension; S83. For each visitor, the system generates an adaptive behavior model based on the behavior characteristics. , the behavior model The recursive neural network captures the temporal characteristics of visitor behavior, and the system predicts the visitor's future visit behavior to obtain the prediction vector ; S84, based on behavioral model The system dynamically optimizes the approval process based on the output of the visitor. Through the reinforcement learning algorithm, the system adjusts the approval process according to the visitor's behavioral characteristics and prediction results: ; Among them, EFF represents approval efficiency, Represents the risk indicator, The approval process is time-consuming. and They represent the parameters that weigh approval efficiency and security. When the approval efficiency EFF exceeds the threshold, the system automatically skips some approval steps. S85, system based on behavioral model Dynamically adjust the QR code management strategy and the validity period of the QR code And use conditions to adaptively update based on visitor behavior predictions: ; in, Indicates the initial validity period. represents the adjustment factor, represents the access behavior prediction function at time t; S86, the system is based on long-term collection of visitor access data and uses a greedy strategy to model visitor behavior Perform periodic updates.

7. A dual-mode gate access control system based on LoRa and Wi-Fi, executing the dual-mode gate access control method based on LoRa and Wi-Fi as described in any one of claims 1 to 6, characterized in that: Includes the following modules: System initialization module: used to initialize the gate control module, establish a connection with the cloud server through the edge computing device, load the visitor management module, and start the communication module; Dynamic communication selection module: used to select dynamic communication mode based on the visitor's geographic location, network status and device status using the Markov decision process algorithm; Dynamic QR code generation module: used to combine hypergeometric learning and manifold learning algorithms to analyze visitors' historical behavior data and generate dynamic QR codes; Edge computing verification module: used to decrypt the visitor's dynamic QR code, verify the timestamp, verify the device binding, and match the behavior characteristics, and trigger multi-factor authentication as needed; Passage control and recording module: used to control the opening of the gate after the QR code verification is passed, and upload the passage record to the cloud server for archiving; Behavioral data analysis and optimization module: used to collect and analyze visitor access data, generate behavioral models, and dynamically optimize approval processes and QR code management.

Images