AI intelligent home security control system

Abstract

The application relates to the field of home safety, in particular to an AI intelligent home safety control system; the AI intelligent home safety control system comprises a management center, the management center is in communication connection with a home user behavior module, a home user control module and a safety detection control module. Historical instruction data of the user to the intelligent home are acquired, the historical instruction data are analyzed, a user instruction graph is constructed, and user control data are generated in combination with the user instruction graph; a network path link is constructed in combination with the historical instruction data and the user control data; the network path link is analyzed to generate a standard path link template; real-time home path links are generated through the acquired current control behavior of the user, a home behavior control model is obtained according to the user instruction graph and the network path link; real-time instruction data collected are analyzed in combination with the home behavior control model to generate a risk mapping index table, and the intelligent home is controlled in safety through the risk mapping index table. The application can improve the safety of the intelligent home.

Description

Technical Field

[0001] This invention relates to the field of home security, specifically to an AI-powered smart home security control system. Background Technology

[0002] With the development of technologies such as the Internet of Things, artificial intelligence, and edge computing, smart home systems have been widely adopted. Users can control the operation of home devices through mobile terminals, voice assistants, and other means to achieve intelligent control of lighting, temperature control, security protection, and other functions. To enhance user experience, smart home platforms have gradually acquired intelligent capabilities such as learning user behavior preferences and controlling multiple devices in a coordinated manner. At the same time, AI technology has shown significant advantages in home behavior recognition, device anomaly detection, and automated response, injecting intelligent perception and decision-making mechanisms into smart home systems.

[0003] Chinese patent CN115309065A discloses an IoT-based smart home security control system and method, including: an information collection module, a data management center, a monitoring information analysis module, a home device monitoring module, and a security control early warning module. The information collection module collects user operation data on the smart home, the data management center stores and manages the collected data, the monitoring information analysis module identifies aging smart home devices and analyzes their degree of aging, the home device monitoring module plans the monitoring interval for aging smart home devices, sensors are installed on the devices for monitoring, and the security control early warning module analyzes user operation data to predict and warn of the risk of unauthorized control of home devices, reducing the probability of security incidents during smart home device control and the risk of unauthorized control of home devices.

[0004] In existing technologies, most smart home systems rely solely on single command execution for control permission assessment, lacking comprehensive analysis of user behavior and device interactions. This makes it difficult for the system to promptly identify risky behaviors in scenarios such as abnormal user commands, path changes, or device replacement, leading to security vulnerabilities such as unauthorized control, accidental device triggering, and path hijacking. Furthermore, existing smart home security strategies are mostly based on static rule bases, unable to generate dynamic security response strategies in real time, resulting in poor adaptability to the real-time and complex nature of control commands. These are the problems we need to solve. Summary of the Invention

[0005] The purpose of this invention is to address the problems existing in the background technology by proposing an AI smart home security control system.

[0006] The technical solution of this invention: An AI smart home security control system, comprising a management center, wherein the management center is communicatively connected to a home user behavior module, a home user control module, and a security detection and control module.

[0007] The home user behavior module is used to acquire historical command data of users for smart home devices, analyze the historical command data, construct a user command graph, and generate user control data by combining the user command graph.

[0008] The home user control module is used to construct network path links by combining historical command data and user control data; analyze the network path links to generate standard path link templates; generate real-time home path links based on the acquired user's current control behavior; and obtain a home behavior control model based on the user command graph and network path links.

[0009] The security detection and control module is used to analyze the collected real-time command data in conjunction with the home behavior control model, generate a risk mapping index table, and use the risk mapping index table to control the security of smart homes.

[0010] Preferably, the process of acquiring historical user commands for smart home devices, analyzing this historical command data, and constructing a user command map includes:

[0011] Set up a smart home control platform, which obtains historical command data from users regarding smart home devices. The historical command data includes control timestamps, control terminal identifiers, control object identifiers, control action types, and execution response status.

[0012] The system combines control timestamps to align the corresponding historical command data with timestamps, converts the control terminal identifiers in the historical command data into unified standard device identifiers, converts the historical command data into a unified field format, and extracts key fields to obtain terminal key identifiers, object key identifiers, action field types, and response key states. Based on the control action types in the historical command data, the system extracts typical control action semantics and constructs a preset action coding table. The system then converts the control action types into numerical codes according to the preset action coding table to generate control action codes.

[0013] Based on the control timestamp sequence of historical instruction data, home behavior nodes are constructed using terminal key identifiers, object key identifiers, and control action codes. These home behavior nodes include device graph nodes, user graph nodes, and control nodes. State transition edges are constructed between corresponding home behavior nodes by responding to key states. The state transition edges are then linked to the corresponding home behavior nodes to construct the user instruction graph.

[0014] Preferably, the process of generating user control data by combining user command graphs includes:

[0015] Extract the frequency and time of occurrence of each home behavior node in the user command graph, and set the running time period; set the running time period as the column of the matrix, set the device graph node and its corresponding control node as the row of the matrix, store the frequency of occurrence of device graph node and control node in the corresponding item of the matrix, and construct the frequency distribution matrix of user control of home.

[0016] Based on the control timestamp, the time distribution of each device graph node and control node within the day and week is statistically analyzed to generate a set of behavior patterns; the state transition edges between home behavior nodes are analyzed, and the transition states of the state transition edges include normal and abnormal; a control behavior chain from the initial state to the target state is generated; and user control data is generated based on the frequency distribution matrix of each home behavior node, the set of behavior patterns, and the control behavior chain.

[0017] Preferably, the process of constructing network path links by combining historical command data and user control data, and analyzing the network path links to generate standard path link templates, is as follows:

[0018] Extract the sending device identifier, the target device identifier, and node identifier information during the transmission process from historical command data and user control data;

[0019] By combining node identification information, the network connection order and topology relationship between each corresponding home behavior node are analyzed. By combining the sending device identifier and the target device identifier with the network connection order and topology relationship of the control behavior chain, the control behavior chain is further analyzed to construct the network path chain between the home behavior nodes corresponding to the sending device and the target device.

[0020] The network path chain is segmented and marked to obtain the path start point, relay node and target node, and to record the network hop count, access protocol type and device link location index information of each node in the network path chain.

[0021] Filter out all network path chains with state transition edges that are in a normal transition state, and construct historical network path chain samples; set a frequency reference threshold, count the occurrence frequency of historical network path chain samples, and select historical network path chain samples with an occurrence frequency greater than or equal to the frequency reference threshold as standard path chain templates.

[0022] Preferably, the process of generating a real-time home path chain based on the acquired user's current control behavior is as follows:

[0023] Obtain the target device identifier and control action type corresponding to the user's current control behavior, extract the device transmission node identifier information during the user control process, construct a real-time behavior node, which includes a real-time device node, a real-time user node, and a real-time control node, obtain the network path generated by the user's current control behavior within the real-time node, and construct the real-time home path chain corresponding to the current control behavior in combination with the real-time behavior node.

[0024] Preferably, the process of obtaining a home behavior control model based on the user command graph and network path chain includes:

[0025] Based on the control behavior node chain composed of control terminal identifier, control object identifier, and control action code in the user command graph, the target device identifier and control action code of the network path chain are extracted. Node-level alignment is performed on the user command graph and network path chain using the target device identifier to construct a correspondence index between home behavior nodes and path start points, relay nodes, and target nodes. Based on the correspondence index, a joint chain structure of user control behavior and path transmission structure is generated. The path hop count sequence, device access sequence, and network topology node order corresponding to each control behavior node in the joint chain structure are extracted. A home behavior control model is constructed by combining the user command graph and the network path chain.

[0026] Preferably, the process of analyzing the collected real-time command data by combining a home behavior control model to generate a risk mapping index table, and then using the risk mapping index table to perform security control of the smart home includes:

[0027] Collect real-time command data sent by the control terminal, extract the control object, identify the control action, and add a timestamp to the real-time command data; match the control object, control action, and path information contained in the real-time command data with the home behavior control model constructed from historical command data;

[0028] Based on the path hop count differences, control action code offsets, and target device identifier changes identified during the matching process, risk labels are generated at the corresponding nodes or state transition edges where changes occur, and the corresponding risk reasons are stored in the risk labels. The risk reasons include path hop count differences, control action code offsets, and target device identifier changes.

[0029] Preferably, the process of analyzing the collected real-time command data by combining a home behavior control model to generate a risk mapping index table, and then using the risk mapping index table to perform security control of the smart home includes:

[0030] Extract real-time command data and user control data to each node of the real-time home path chain and control behavior chain in the home behavior control model, respectively, and construct the current control command sequence and the historical standard command sequence. Based on the node jumps, control action mutations, and inconsistencies in the path chain node order between the current control command sequence and the historical standard command sequence, construct an offset map between the real-time home path chain and the control behavior chain. Mark the risk labels at the corresponding positions in the offset map to generate a risk mapping index table.

[0031] If a risk label exists in the risk mapping index table, the user's control over the smart home will be cut off through the smart home control platform, and the user's control permissions for the smart home will be revoked. If a risk label exists in the risk mapping index table, the user's control behavior will be stopped through the smart home control platform, and the user will be reminded to perform the operation again, until no risk mapping index table is generated.

[0032] Compared with the prior art, the above-mentioned technical solution of the present invention has the following beneficial technical effects: The home behavior control model constructed based on the user command graph and network path chain can jointly model the path structure, node sequence and control action code of user control behavior, thereby realizing fine-grained identification of control behavior structure deviation and path chain topology change, which helps to accurately identify potential risk paths and command deviation behaviors; Through the offset graph constructed between the real-time home path chain and the control behavior chain, the structure comparison between real-time control commands and historical standard behaviors in dimensions such as hop count, sequence, and action code is realized, and node-level risk labeling is carried out in combination with the risk mapping index table, thereby realizing the structured execution of dynamic judgment of control permissions and response strategy invocation, avoiding control failure or overreach caused by path detours, device replacement, etc., and improving the security of smart homes. Attached Figure Description

[0033] Figure 1 This is a schematic diagram of one embodiment of the present invention. Detailed Implementation

[0034] like Figure 1 As shown, the present invention proposes an AI smart home security control system, including a management center, which is communicatively connected to a home user behavior module, a home user control module, and a security detection and control module.

[0035] The home user behavior module is used to acquire historical command data of users for smart home devices, analyze the historical command data, construct a user command graph, and generate user control data by combining the user command graph.

[0036] The home user control module is used to construct network path links by combining historical command data and user control data; analyze the network path links to generate standard path link templates; generate real-time home path links based on the acquired user's current control behavior; and obtain a home behavior control model based on the user command graph and network path links.

[0037] The security detection and control module is used to analyze the collected real-time command data in conjunction with the home behavior control model, generate a risk mapping index table, and use the risk mapping index table to control the security of smart homes.

[0038] It should be further explained that, in the specific implementation process, the process of acquiring historical user commands for smart home devices, analyzing this historical command data to construct a user command graph, and then generating user control data based on this graph is as follows:

[0039] Set up a smart home control platform, which acquires historical command data of users for smart home devices;

[0040] The historical instruction data includes control timestamp, control terminal identifier, control object identifier, control action type, and execution response status;

[0041] Specifically, the historical command data refers to a complete set of executed control commands issued by home users to smart home devices through various control terminals (such as voice assistants, mobile apps, wall panels, etc.); the control timestamp refers to the standardized time record value of the control command issued by the user, usually marked in Coordinated Universal Time (UTC) format, accurate to the second or millisecond; the control terminal identifier refers to the original identification information of the device initiating the control command, such as MAC address, IMEI number, device number, etc., representing the source of the control interface used by the user; the control object identifier refers to the unique identification code of the controlled smart device (such as device ID, binding code, etc.); the control action type refers to the type of control behavior issued by the user to the target device, such as "on", "off", "temperature adjustment", "mode setting", etc.; the execution response status refers to the actual response result of the target device after receiving the control command, which may include status codes such as "successful execution", "execution failure", "device no response";

[0042] The historical instruction data is timestamped by combining control timestamps, and the control terminal identifiers in the historical instruction data are converted into unified standard device identifiers. The historical instruction data is also converted into a unified field format, and key fields are extracted to obtain terminal key identifiers, object key identifiers, action field types, and response key states. Based on the control action types in the historical instruction data, typical control action semantics are extracted, and a preset action coding table is constructed. Each control action type in the preset action coding table corresponds to a unique coding identifier. The control action types are converted into numerical codes according to the preset action coding table to generate control action codes.

[0043] Based on the control timestamp order of historical instruction data, home behavior nodes are constructed using terminal key identifiers, object key identifiers, and control action codes. These home behavior nodes include device graph nodes, user graph nodes, and control nodes, each storing terminal key identifiers, object key identifiers, and control action codes. State transition edges are constructed between corresponding home behavior nodes by responding to key states. These state transition edges are then linked to the corresponding home behavior nodes to construct a user instruction graph.

[0044] Specifically, the timestamp alignment refers to sorting the control timestamps in all historical instruction data to ensure the orderliness of control behavior in the time dimension; the unified standard device identifier refers to mapping the original control terminal identifiers (such as MAC, IMEI, device nickname, etc.) collected in the historical instruction data to a set unified standard code; the unified field format refers to converting the historical instruction data into a data table structure with a unified structured format.

[0045] Extract the frequency and time of occurrence of each home behavior node in the user instruction graph, set the running time period, which is divided into days; set the running time period as the column of the matrix, set the device graph node and its corresponding control node as the row of the matrix, store the frequency of occurrence of device graph node and control node in the corresponding item of the matrix, and construct the frequency distribution matrix of user control of home.

[0046] Based on the control timestamp, the time distribution of each device graph node and control node within the day and week is statistically analyzed to generate a set of behavior patterns; the state transition edges between home behavior nodes are analyzed, and the transition states of the state transition edges include normal and abnormal; a control behavior chain from the initial state to the target state is generated.

[0047] User control data is generated based on the frequency distribution matrix, behavior pattern set, and control behavior chain of each home behavior node. The user control data is used to represent the periodicity, preference, and variability of user control behavior.

[0048] It needs further explanation that, in the specific implementation process, the following steps are taken: First, a network path link is constructed by combining historical command data and user control data. Then, the network path link is analyzed to generate a standard path link template. Finally, a real-time home path link is generated based on the acquired user's current control behavior. The process of obtaining the home behavior control model based on the user command graph and the network path link is as follows:

[0049] Extract the sending device identifier, the target device identifier, and node identifier information during the transmission process from historical command data and user control data;

[0050] By combining node identification information, the network connection order and topology relationship between each corresponding home behavior node are analyzed. By combining the sending device identifier and the target device identifier with the network connection order and topology relationship of the control behavior chain, the control behavior chain is further analyzed to construct the network path chain between the home behavior nodes corresponding to the sending device and the target device.

[0051] The network path chain is segmented and marked to obtain the path start point, relay node and target node, and to record the network hop count, access protocol type and device link location index information of each node in the network path chain.

[0052] Specifically, the segmentation marker refers to classifying nodes in the network path chain into three categories based on their network roles and transmission order: originating nodes, relay nodes, and destination nodes. The originating node is the device node that sends user control commands; the relay node is the forwarding node in the process of transmitting control commands from the sending device to the destination device; and the destination node is the device node that receives control commands and executes corresponding actions. The network hop count, access protocol type, and device link location index information are obtained through gateway logs and device metadata parsing.

[0053] Filter out all network path chains with state transition edges that are all in a normal transition state, and construct historical network path chain samples; set a frequency reference threshold, count the occurrence frequency of historical network path chain samples, and select historical network path chain samples with an occurrence frequency greater than or equal to the frequency reference threshold as standard path chain templates.

[0054] Obtain the target device identifier and control action type corresponding to the user's current control behavior, extract the device transmission node identifier information during the user control process, construct a real-time behavior node, which includes a real-time device node, a real-time user node, and a real-time control node, obtain the network path generated by the user's current control behavior within the real-time node, and construct the real-time home path chain corresponding to the current control behavior in combination with the real-time behavior node.

[0055] Based on the control behavior node chain composed of control terminal identifier, control object identifier, and control action code in the user command graph, the target device identifier and control action code of the control path chain are extracted. Node-level alignment is performed on the user command graph and network path chain using the target device identifier to construct a correspondence index between home behavior nodes and path start points, relay nodes, and target nodes. Based on the correspondence index, a joint chain structure of user control behavior and path transmission structure is generated. The path hop count sequence, device access sequence, and network topology node order corresponding to each control behavior node in the joint chain structure are extracted. A home behavior control model is constructed by combining the user command graph and network path chain.

[0056] Specifically, the node-level alignment operation refers to aligning the device graph nodes, user graph nodes, and control nodes of the user instruction graph with the path start point, relay node, and target node of the network path chain.

[0057] It should be further explained that, in the specific implementation process, the collected real-time command data is analyzed in conjunction with the home behavior control model to generate a risk mapping index table. The process of using the risk mapping index table to perform security control of smart homes is as follows:

[0058] Collect real-time command data sent by the control terminal, extract the control object, identify the control action, and add a timestamp to the real-time command data; match the control object, control action, and path information contained in the real-time command data with the home behavior control model constructed from historical command data;

[0059] Based on the path hop count differences, control action code offsets, and target device identifier changes identified during the matching process, risk labels are generated at the corresponding nodes or state transition edges where changes occur, and the corresponding risk reasons are stored in the risk labels. The risk reasons include path hop count differences, control action code offsets, and target device identifier changes.

[0060] Extract real-time command data and user control data to each node of the real-time home path chain and control behavior chain in the home behavior control model, respectively, and construct the current control command sequence and the historical standard command sequence. Based on the node jumps, control action mutations, and inconsistencies in the path chain node order between the current control command sequence and the historical standard command sequence, construct an offset map between the real-time home path chain and the control behavior chain. Mark the risk labels at the corresponding positions in the offset map to generate a risk mapping index table.

[0061] If a risk label exists in the risk mapping index table, the user's control over the smart home will be cut off through the smart home control platform, and the user's control permissions for the smart home will be revoked. If a risk label exists in the risk mapping index table, the user's control behavior will be stopped through the smart home control platform, and the user will be reminded to perform the operation again, until no risk mapping index table is generated.

[0062] The embodiments of the present invention have been described in detail above with reference to the accompanying drawings. However, the present invention is not limited thereto. Various changes can be made within the scope of knowledge possessed by those skilled in the art without departing from the spirit of the present invention.

Claims

1. An AI-powered smart home security control system, comprising a management center, characterized in that, The management center is connected to a home user behavior module, a home user control module, and a security detection and control module. The home user behavior module is used to acquire historical command data of users for smart home devices, analyze the historical command data, construct a user command graph, and generate user control data by combining the user command graph. The home user control module is used to construct network path links by combining historical command data and user control data; analyze the network path links to generate standard path link templates; generate real-time home path links based on the acquired user's current control behavior; and obtain a home behavior control model based on the user command map and network path links. The security detection and control module is used to analyze the collected real-time command data in conjunction with the home behavior control model to generate a risk mapping index table, and use the risk mapping index table to perform security control of the smart home. The process of analyzing the collected real-time command data using a home behavior control model to generate a risk mapping index table, and then using this risk mapping index table to implement security control for smart homes, includes: Extract real-time command data and user control data to each node of the real-time home path chain and control behavior chain in the home behavior control model, respectively, and construct the current control command sequence and the historical standard command sequence. Based on the node jumps, control action mutations, and inconsistencies in the path chain node order between the current control command sequence and the historical standard command sequence, construct an offset map between the real-time home path chain and the control behavior chain. Mark the risk labels at the corresponding positions in the offset map to generate a risk mapping index table. If a risk label exists in the risk mapping index table, the user's control over the smart home will be cut off through the smart home control platform, and the user's control permissions for the smart home will be revoked. If a risk label exists in the risk mapping index table, the user's control behavior will be stopped through the smart home control platform, and the user will be reminded to perform the operation again, until no risk mapping index table is generated.

2. The AI ​​smart home security control system according to claim 1, characterized in that, The process of acquiring historical user commands for smart home devices, analyzing this data, and constructing a user command map includes: Set up a smart home control platform, which obtains historical command data from users regarding smart home devices. The historical command data includes control timestamps, control terminal identifiers, control object identifiers, control action types, and execution response status. The system combines control timestamps to align the corresponding historical command data with timestamps, converts the control terminal identifiers in the historical command data into unified standard device identifiers, converts the historical command data into a unified field format, and extracts key fields to obtain terminal key identifiers, object key identifiers, action field types, and response key states. Based on the control action types in the historical command data, the system extracts typical control action semantics and constructs a preset action coding table. The system then converts the control action types into numerical codes according to the preset action coding table to generate control action codes. Based on the control timestamp sequence of historical instruction data, home behavior nodes are constructed using terminal key identifiers, object key identifiers, and control action codes. These home behavior nodes include device graph nodes, user graph nodes, and control nodes. State transition edges are constructed between corresponding home behavior nodes by responding to key states. The state transition edges are then linked to the corresponding home behavior nodes to construct the user instruction graph.

3. The AI ​​smart home security control system according to claim 2, characterized in that, The process of generating user control data by combining user command graphs includes: Extract the frequency and time of occurrence of each home behavior node in the user command graph, and set the running time period; set the running time period as the column of the matrix, set the device graph node and its corresponding control node as the row of the matrix, store the frequency of occurrence of device graph node and control node in the corresponding item of the matrix, and construct the frequency distribution matrix of user control of home. Based on the control timestamp, the time distribution of each device graph node and control node within the day and week is statistically analyzed to generate a set of behavior patterns; the state transition edges between home behavior nodes are analyzed, and the transition states of the state transition edges include normal and abnormal; a control behavior chain from the initial state to the target state is generated; and user control data is generated based on the frequency distribution matrix of each home behavior node, the set of behavior patterns, and the control behavior chain.

4. The AI ​​smart home security control system according to claim 3, characterized in that, The process of constructing network path links by combining historical command data and user control data, and analyzing these network path links to generate standard path link templates, is as follows: Extract the sending device identifier, the target device identifier, and node identifier information during the transmission process from historical command data and user control data; By combining node identification information, the network connection order and topology relationship between each corresponding home behavior node are analyzed. By combining the sending device identifier and the target device identifier with the network connection order and topology relationship of the control behavior chain, the control behavior chain is further analyzed to construct the network path chain between the home behavior nodes corresponding to the sending device and the target device. The network path chain is segmented and marked to obtain the path start point, relay node and target node, and to record the network hop count, access protocol type and device link location index information of each node in the network path chain. Filter out all network path chains with state transition edges that are in a normal transition state, and construct historical network path chain samples; set a frequency reference threshold, count the occurrence frequency of historical network path chain samples, and select historical network path chain samples with an occurrence frequency greater than or equal to the frequency reference threshold as standard path chain templates.

5. The AI ​​smart home security control system according to claim 4, characterized in that, The process of generating a real-time home path chain based on the user's current control behavior is as follows: Obtain the target device identifier and control action type corresponding to the user's current control behavior, extract the device transmission node identifier information during the user control process, construct a real-time behavior node, which includes a real-time device node, a real-time user node, and a real-time control node, obtain the network path generated by the user's current control behavior within the real-time node, and construct the real-time home path chain corresponding to the current control behavior in combination with the real-time behavior node.

6. The AI ​​smart home security control system according to claim 5, characterized in that, The process of obtaining a home behavior control model based on user command graphs and network path chains includes: Based on the control behavior node chain composed of control terminal identifier, control object identifier, and control action code in the user command graph, the target device identifier and control action code of the network path chain are extracted. Node-level alignment is performed on the user command graph and network path chain using the target device identifier to construct a correspondence index between home behavior nodes and path start points, relay nodes, and target nodes. Based on the correspondence index, a joint chain structure of user control behavior and path transmission structure is generated. The path hop count sequence, device access sequence, and network topology node order corresponding to each control behavior node in the joint chain structure are extracted. A home behavior control model is constructed by combining the user command graph and the network path chain.

Images