A scene linkage control system and method of a whole-house intelligent switch panel

Abstract

The application discloses a scene linkage control system and method of a whole-house intelligent switch panel, relates to the technical field of control analysis, and comprises the following steps: splitting a target execution node set into an instant execution queue and a risk suspension queue according to a cross judgment result, generating power-off control data of the instant execution queue; starting state polling on the risk suspension queue, generating consumption reduction sleep data in the state polling, reacquiring new working condition identification data and new current fluctuation sequence data, removing a corresponding node and generating supplementary power-off control data when both the new working condition identification data and the new current fluctuation sequence data meet safety conditions; accumulating a total waiting time length of the state polling, and when the total waiting time length exceeds an intervention threshold value and there is a stranded node in the risk suspension queue, terminating the state polling and generating deadlock warning data of the stranded node. The application has the effect of improving the scene linkage control efficiency.

Description

Technical Field

[0001] This application relates to the field of control analysis technology, and in particular to a scene linkage control system and method for a whole-house smart switch panel. Background Technology

[0002] As the centralized interaction point in a home IoT system, the smart switch panel serves as the unified control over lights, sockets, curtains, and other electrical devices. Its stability directly impacts the efficiency of whole-house scene linkage and user experience. Meanwhile, smart sockets and their connected terminal devices are no longer simply power loads. Some devices may undergo internal processes such as firmware upgrades, data writing, cache synchronization, and sleep switching during operation. Therefore, the continuity of power supply is crucial for device safety and data integrity.

[0003] In scenarios such as "whole house goodnight" or "power off when leaving home," switch panels often need to perform unified control over multiple circuits and sockets within a short period. While these scenarios appear to be merely power on / off actions, their actual impact extends to the specific loads downstream of the sockets. When a device is in an uninterruptible operating condition, the power-off action triggered by the scenario linkage may interrupt the device's internal operations, potentially leading to system anomalies, data corruption, or even hardware failure. Therefore, there is an implicit correlation between power scenario control and the real-time operating status of devices; simply executing on / off control based on scenario rules is insufficient to meet the safety requirements of complex home electrical environments.

[0004] However, existing smart home scenario linkages largely rely on pre-configured fixed rules, typically only determining whether devices are online and sockets are on, lacking comprehensive identification of real-time load conditions and electrical operating characteristics. For smart devices capable of reporting status, existing systems often remain at the single-query level, unable to handle status delays, missing feedback, or temporary task changes; for loads unable to communicate proactively, it's even more difficult to determine whether immediate power outage is appropriate. This leads to problems such as blind execution, overall blockage, or simple timeout prompts in power outage scenarios, affecting the continuity of scenario execution and failing to promptly prevent potential device damage risks. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this application provides a scene linkage control system and method for whole-house smart switch panels.

[0006] Firstly, this application provides a scene-linked control method for a whole-house smart switch panel, comprising the following steps: Receive global power outage scenario command, parse scenario feature identifier code and area coverage verification code, verify real-time network topology feature code based on area coverage verification code, and extract initial node identifier list from scenario feature identifier code after verification, perform connectivity filtering on initial node identifier list, and generate target execution node set; For the target set of execution nodes, obtain operating condition identification data and current fluctuation sequence data within the same sampling window; The operating condition identification data is compared with the safety identification database, and the current fluctuation sequence data is matched with the electrical feature fingerprint database. The comparison results and matching results are then fused to generate a cross-judgment result. Based on the cross-determination results, the target execution node set is split into an immediate execution queue and a risk suspension queue, and power outage control data for the immediate execution queue is generated. Start status polling for the risk suspension queue. During status polling, generate power consumption reduction and hibernation data, reacquire new operating condition identification data and new current fluctuation sequence data, and remove the corresponding node and generate supplementary power outage control data when both meet the safety conditions. The total waiting time of the cumulative status polling is calculated. When the total waiting time exceeds the intervention threshold and there are stranded nodes in the risk suspension queue, the status polling is terminated and deadlock warning data of the stranded nodes is generated.

[0007] Secondly, this application provides a scene linkage control system for a whole-house smart switch panel, including: The parsing module is used to receive global power outage scenario commands, parse the scenario feature identifier code and the area coverage verification code, verify the real-time network topology feature code based on the area coverage verification code, and extract the initial node identifier list from the scenario feature identifier code after the verification is passed. The initial node identifier list is then filtered for connectivity to generate the target execution node set. The acquisition module is used to acquire operating condition identification data and current fluctuation sequence data within the same sampling window for the target set of execution nodes; The comparison module is used to compare the operating condition identification data with the safety identification database, match the current fluctuation sequence data with the electrical feature fingerprint database, and fuse the comparison results and matching results to generate a cross-judgment result. The splitting module is used to split the target execution node set into an immediate execution queue and a risk suspension queue according to the cross-determination results, and generate power-off control data for the immediate execution queue; The polling module is used to start status polling of the risk suspension queue. During status polling, power consumption reduction and hibernation data are generated, new operating condition identification data and new current fluctuation sequence data are reacquired, and the corresponding node is removed and reissued power outage control data is generated when both meet the safety conditions. The output module is used to accumulate the total waiting time of status polling. When the total waiting time exceeds the intervention threshold and there are stranded nodes in the risk suspension queue, the status polling is terminated and deadlock warning data of the stranded nodes is generated.

[0008] In summary, this application includes at least one of the following beneficial technical effects: 1. This application provides a scene-linked control method for a whole-house smart switch panel. After receiving a global power outage scene command, the method first parses the scene feature identifier code and the area coverage verification code, and then uses the area coverage verification code to verify the real-time network topology feature code. After the verification is passed, an initial node identifier list is extracted and connectivity filtering is performed, thus limiting the set of target execution nodes actually participating in the power outage linkage from the data source. Simultaneously, this application does not directly issue power outage actions based solely on scene rules, but synchronously acquires operating condition identifier data and current fluctuation sequence data within the same sampling window for the target execution node set, ensuring that the reported status of the device corresponds to the actual electrical operating status in time. By comparing the operating condition identifier data with a safety identifier library and performing waveform matching between the current fluctuation sequence data and an electrical feature fingerprint library, and then fusing the two types of judgment results to generate a cross-judgment result, it avoids misjudgments caused by relying solely on the device's online status, switch status, or a single current magnitude. This creates a continuous data closed loop for node selection, status identification, and risk judgment in the power outage scene, thereby improving the control accuracy and execution reliability of whole-house power outage linkage under complex load environments. 2. By splitting the target execution node set into an immediate execution queue and a risk suspension queue based on the cross-judgment results, nodes that can be safely powered off can generate power outage control data in a timely manner, while nodes with operational risks enter the subsequent state polling process, avoiding the blockage of the entire global power outage scenario by a single risky node. In the risk suspension queue, this application generates power-saving dormancy data by combining operating condition identification data, and re-acquires new operating condition identification data and new current fluctuation sequence data during the polling process. Only when both meet the safety conditions is the corresponding node removed and a supplementary power outage control data is generated, allowing the suspended nodes to continue to complete the scenario execution after the state changes to safe. At the same time, by accumulating the total waiting time of state polling, and terminating state polling and generating deadlock warning data when the intervention threshold is exceeded and there are stagnant nodes, this application establishes an orderly collaborative relationship between immediate execution, delayed supplementary data, and abnormal warning. This reduces the risk of data corruption and operational abnormalities caused by sudden power outages of critical task equipment, and avoids prolonged lack of feedback in whole-house power outage scenarios, enabling smart switch panels to have more granular, safer, and more realistic scene linkage control capabilities that conform to the actual household power consumption status. Attached Figure Description

[0009] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0010] Figure 1 This is a flowchart of the scene linkage control method for a whole-house smart switch panel according to an embodiment of this application.

[0011] Figure 2 This is a flowchart of the method for setting up the target execution node set in an embodiment of this application.

[0012] Figure 3 This is a schematic diagram of the scene linkage control system of the whole-house smart switch panel in the embodiment of this application. Detailed Implementation

[0013] The following description, in conjunction with the implementation of the present invention, is merely an example and illustration of the concept of the present invention. Those skilled in the art can make various modifications or additions to the specific embodiments described, or use similar methods to replace them, as long as they do not deviate from the concept of the invention, and all such modifications and additions should fall within the protection scope of the present invention.

[0014] Example 1: This application is applicable to power-related control scenarios using a whole-house smart switch panel as the scene entry point, especially suitable for scenarios requiring unified power-off processing of multiple power nodes, such as "Goodnight," "Leaving Home," and "Whole-House Energy Saving." The core idea of ​​this application is that after receiving a global power-off scene command, the switch panel does not directly cut off all nodes at once according to fixed scene rules. Instead, it first forms a continuous data processing link based on node affiliation, communication connectivity, operating condition identification data, and current fluctuation sequence data. Then, it cross-determines the operating condition judgment and the electrical judgment, thereby distinguishing nodes that can be immediately powered off and nodes that need to be temporarily observed. For nodes that need to be temporarily observed, status polling, power-saving sleep data, resending power-off control data, and deadlock warning data are also introduced, ensuring that the global power-off scene maintains execution integrity while reducing the risk of sudden power outages to equipment performing critical tasks.

[0015] This application discloses a scene linkage control method for a whole-house smart switch panel.

[0016] Reference Figure 1-2 A method for scene linkage control of a whole-house smart switch panel includes the following steps: Receive global power outage scenario command, parse scenario feature identifier code and area coverage verification code, verify real-time network topology feature code based on area coverage verification code, and extract initial node identifier list from scenario feature identifier code after verification, perform connectivity filtering on initial node identifier list, and generate target execution node set; Furthermore, it receives global power outage scenario commands, parses the scenario feature identifier code and area coverage verification code, and generates a set of target execution nodes, including: The data frames of the global power outage scenario command are unpacked, and the scenario feature identifier code and the area coverage verification code are extracted. Read the real-time network topology feature code from the local topology record, compare the area coverage check code with the real-time network topology feature code to obtain the topology comparison difference; When the topology comparison difference is less than the topology change threshold, the initial node identifier list is extracted by addressing in the scene mapping table based on the scene feature identifier code. Connectivity detection data is generated according to the initial node identifier list, and heartbeat response data is received from the nodes; The heartbeat response data is parsed to obtain the communication delay value and the response integrity mark. Node identifiers whose communication delay value exceeds the packet loss limit and whose response integrity mark is invalid are deleted. The remaining node identifiers are combined into a target execution node set.

[0017] In one specific implementation, the whole-house smart switch panel is equipped with a scene instruction parsing program, a topology verification program, a node sampling program, a cross-judgment program, a queue scheduling program, and a log update program. These programs can be executed by the processor within the switch panel calling instructions stored in memory. Global power outage scene instructions can originate from panel touch buttons, timed scene tasks, voice scene gateway forwarding data, or home automation rule-triggered data. For ease of explanation, this implementation uses an away-from-home power outage scene as an example. The global power outage scene instruction uses a fixed-length data frame, which includes at least a frame header, a scene feature identifier code, a region coverage verification code, a timestamp, and a verification tail code. The scene feature identifier code indicates that the current scene is a whole-house power outage scene, and the region coverage verification code identifies the room area, circuit area, and node group range that the scene should cover.

[0018] Specifically, after the switch panel receives a global power outage scenario command, the scenario command parsing program parses the command's fields to obtain the scenario feature identifier code and the area coverage verification code. Subsequently, the topology verification program reads the real-time network topology feature code from the local topology record. The local topology record is updated by the switch panel during daily operation based on node network entry / exit, room binding relationships, and loop binding relationships. The real-time network topology feature code can be represented as a binary string, where different bit segments correspond to the node presence status and loop association status in areas such as the living room, bedroom, kitchen, and bathroom. The area coverage verification code and the real-time network topology feature code are generated using the same encoding rule, therefore they can be directly compared for consistency. In actual processing, the area coverage verification code can be denoted as... The real-time network topology feature code is denoted as Both have a number of digits. Topological alignment difference Can be customized according to different number of digits and The ratio is obtained, that is ,in, This indicates the number of bits that differ in value between the two encoded strings. This represents the total number of bits in the encoded string. The topology change threshold can be determined based on the number of nodes in the household; for example, 0.1 for no more than 32 nodes and 0.05 for more than 32 nodes. If the topology comparison difference is less than the topology change threshold, it means that the coverage of the global power outage scenario command is basically consistent with the current network topology, and the scenario mapping table can continue to be used. If the topology comparison difference is not less than the topology change threshold, the current scenario command is recorded as a topology anomaly command, and the switch panel will not enter the subsequent power outage queue generation process.

[0019] After successful verification, the scene instruction parsing program uses the scene feature identifier to address the scene in the scene mapping table. The scene mapping table can be stored in the local storage area of ​​the switch panel. Each record includes the scene feature identifier, node identifier, node location, load category identifier, and default execution priority. Taking the power outage scenario when leaving home as an example, the scene feature identifier can correspond to the living room lighting node, bedroom socket node, kitchen non-refrigerator socket node, and entertainment device socket node. The scene instruction parsing program extracts all node identifiers that match the scene feature identifier, forming an initial node identifier list. This initial node identifier list only indicates the nodes that should participate in the power outage at the rule level; it does not yet represent that all these nodes can currently be controlled. Therefore, connectivity filtering is still required.

[0020] Connectivity filtering is performed by a topology verification program and a communication management program. The switch panel generates connectivity detection data based on the initial node identifier list, including node identifiers, detection sequence numbers, and sending timestamps. Each node returns heartbeat response data, containing the corresponding node identifier, detection sequence number, receiving timestamp, returning timestamp, and a complete response flag. The communication management program parses the heartbeat response data and uses the difference between the returning and sending timestamps to obtain the communication latency value. The packet loss threshold can be set between 800ms and 1500ms and can be determined during the installation and commissioning phase based on the average round-trip time of the home wireless network. If the communication latency value exceeds the packet loss threshold and the complete response flag is invalid, the corresponding node identifier is removed from the initial node identifier list; the remaining node identifiers are combined into a target execution node set. This target execution node set inherits the scenario rules while excluding nodes that cannot currently interact reliably, preventing node reference breaks in subsequent data collection.

[0021] For the target set of execution nodes, obtain operating condition identification data and current fluctuation sequence data within the same sampling window; Furthermore, for the target execution node set, operating condition identification data and current fluctuation sequence data within the same sampling window are obtained, including: Based on the node identifier in the target execution node set, query the node attribute table to obtain the node protocol address and node sampling channel identifier; Generate status probe data according to the node protocol address, receive operating condition response data corresponding to the node identifier, and extract operating condition identifier data from the operating condition response data; Read the current sampling data in the sampling window according to the node sampling channel identifier, and sort the current sampling data according to the sampling timestamp to form current fluctuation sequence data; Align the received timestamp of the operating condition identification data with the sampling timestamp of the current fluctuation sequence data to generate a node operating condition sampling record.

[0022] In one specific implementation, after the target execution node set is generated, the node sampling program sequentially reads the node identifiers and queries the node attribute table. The node attribute table is established by the switch panel during equipment network distribution and routine maintenance. Each record includes a node identifier, node protocol address, node sampling channel identifier, load category identifier, and standby baseline data. The node protocol address is used to generate status probe data, and the node sampling channel identifier is used to read current sampling data; both are indexed by the same node identifier. The node sampling program generates status probe data according to the node protocol address. The status probe data includes the node identifier, status field request code, and sampling window start time. After the node returns operating condition response data, the node sampling program extracts operating condition identifier data from the operating condition response data. The operating condition identifier data can be expressed in the form of enumerated codes. For example, idle, standby, running, upgrading, writing, compressor protection, and cleaning heating states all have corresponding code values. The safety identifier library stores a set of code values ​​that allow direct power-off.

[0023] Current fluctuation sequence data and operating condition identification data are acquired within the same sampling window. In this embodiment, the sampling window can be 3 to 10 seconds after the global power outage scenario command is received, and the sampling frequency can be 20Hz to 100Hz. The node sampling program reads the current sampling data within the sampling window according to the node sampling channel identifier. Each current sampling data includes a node identifier, a sampling timestamp, and an instantaneous current value, with the instantaneous current value in amperes. Subsequently, the node sampling program sorts the current sampling data of the same node in ascending order of sampling timestamps to form current fluctuation sequence data. To avoid time discrepancies between the operating condition response data and the current sampling data, the node sampling program aligns the receiving timestamp of the operating condition identification data with the sampling timestamp of the current fluctuation sequence data. When the receiving timestamp of the operating condition identification data falls within the sampling window, the operating condition identification data is directly written into the node operating condition sampling record; when the receiving timestamp is near the boundary of the sampling window, the current sampling data closest to the receiving timestamp is selected as the alignment point, and a predetermined number of sampling data before and after the alignment point are written into the node operating condition sampling record. The node operating condition sampling record includes at least the node identifier, operating condition identifier data, current fluctuation sequence data, and corresponding timestamp.

[0024] The operating condition identification data is compared with the safety identification database, and the current fluctuation sequence data is matched with the electrical feature fingerprint database. The comparison results and matching results are then fused to generate a cross-judgment result. Furthermore, waveform matching is performed between the current fluctuation sequence data and the electrical feature fingerprint database, including: The current fluctuation sequence data is denoised and smoothed to obtain a smooth current sequence; The average current value, peak-to-valley difference, pulse count value, and slope sign sequence are extracted from the smoothed current sequence and combined to generate an electrical feature vector. Based on the load category identifier in the node attribute table, extract the corresponding fingerprint sample group from the electrical feature fingerprint database; The electrical feature vector is compared with the fingerprint sample group item by item to obtain the waveform matching degree; Matching results are generated based on waveform matching degree and risk matching threshold.

[0025] Furthermore, the comparison and matching results are fused to generate cross-determination results, including: The working condition label data is compared with the safety label database to obtain the label safety result. The waveform matching degree in the matching results is compared with the risk matching threshold to obtain the waveform risk result; When the security result is identified as secure and the waveform risk result is identified as a missed risk, a security cross-judgment result is generated. When the safety result is marked as unsafe and the waveform risk result is that any of the risk items in the hit risk are met, a risk cross-judgment result is generated; The results of safe cross-judgment and risk cross-judgment are written into the node operating condition sampling record respectively, forming a node operating condition sampling record with a judgment label.

[0026] In one specific implementation, after forming node operating condition sampling records, the cross-determination program processes the operating condition identification data and current fluctuation sequence data separately. For the operating condition identification data, the cross-determination program compares it with a safety identification library for identification consistency. The safety identification library is established according to load category identification; for example, the safety identification of lighting nodes may include idle and standby, the safety identification of storage device socket nodes may include standby but not write, and the safety identification of kitchen appliance nodes may include idle but not heating and compressor protection. The output of the identification consistency comparison is the identification safety result, which can take two values: safe and unsafe.

[0027] For current fluctuation sequence data, the cross-validation procedure first performs denoising and smoothing. Denoising and smoothing can be achieved using a combination of moving median filtering and moving average, with a sliding window length of 3 to 7 sampling points. instantaneous current value For example, first take the adjacent sampling points before and after it to form a local window, and then calculate the median. Then for a consecutive number of The corresponding value in the smoothed current sequence is obtained by averaging. This process does not change the dimensions of the current value; it remains in amperes. Subsequently, the cross-determination procedure extracts the average current value, peak-to-valley difference, pulse count, and slope sign sequence from the smoothed current sequence. Average current value All within the sampling window Arithmetic mean; peak-to-valley difference To the maximum Subtract the minimum Pulse count value The number of pulse segments in a smooth current sequence that exceed the local reference current and continue to reach a predetermined number of sampling points; the slope sign sequence consists of adjacent... The difference values ​​are composed of signs, with differences greater than a small slope threshold being recorded as positive, differences less than a small slope threshold being recorded as negative, and all others being recorded as zero. The above average current value, peak-to-valley difference, pulse count value, and slope sign sequence are combined into an electrical feature vector.

[0028] The electrical fingerprint database stores fingerprint sample groups categorized by load category. These fingerprint sample groups are not generalized device models, but rather derived from historically collected safety power outage samples and risk operation samples from the same household environment. For example, load category identifiers such as TV boxes, network storage devices, washing machines, refrigerators, and routers each have their own average current range, peak-to-valley difference range, pulse count reference value, and slope sign reference sequence. The cross-validation program extracts the corresponding fingerprint sample groups from the electrical fingerprint database based on the load category identifier in the node attribute table, and then compares the electrical feature vector with each fingerprint sample group item by item. For average current values ​​and peak-to-valley differences, normalized differences can be used for comparison; for pulse count values, count differences can be used for comparison; and for slope sign sequences, the proportion of identical sign positions to the total number of positions can be compared. Waveform matching degree. Press Calculate, where, Indicates the similarity of average current values. Indicates the similarity of peak-valley differences. Indicates the similarity of pulse count values. Indicates the similarity of slope sign sequences. and These are weights, and the sum of the four is 1. Weights can be assigned at the factory according to the load category identifier, for example, increasing the weight for storage loads. The value of [value] improves the performance of electric heating loads. and The value of is determined by comparing the waveform matching degree with the risk matching threshold. The risk matching threshold can be determined by the lowest stable matching degree of historical risk operation samples in the electrical feature fingerprint database, or it can be rewritten according to the household equipment samples after installation and commissioning.

[0029] The safety and matching results are entered into the same cross-determination procedure. The cross-determination procedure compares the waveform matching degree in the matching result with the risk matching threshold to obtain the waveform risk result. When the waveform matching degree reaches the risk matching threshold, the waveform risk result is a hit risk; when the waveform matching degree does not reach the risk matching threshold, the waveform risk result is a miss risk. If the safety result is safe and the waveform risk result is a miss risk, a safety cross-determination result is generated; if either the safety result is unsafe or the waveform risk result is a hit risk, a risk cross-determination result is generated. The cross-determination procedure writes the safety cross-determination result and the risk cross-determination result into the node operating condition sampling record, forming a node operating condition sampling record with a determination label. This determination method avoids relying solely on the device's reported status and avoids judging whether a node can be de-energized based solely on the current magnitude.

[0030] Based on the cross-determination results, the target execution node set is split into an immediate execution queue and a risk suspension queue, and power outage control data for the immediate execution queue is generated. Furthermore, based on the cross-determination results, the target execution node set is split into an immediate execution queue and a risk suspension queue, including: Read the node operating condition sampling records with judgment labels, and extract the node identifier, judgment label, identification safety result, and waveform risk result; Write the node identifier that is determined to be safe into the immediate execution queue, and write the corresponding safe identifier result into the immediate execution queue record; Write the node identifier that is identified as risk into the risk suspension queue, and write the corresponding waveform risk result and working condition identifier data into the risk suspension queue record; Power outage control data is generated based on the immediate execution queue records, and suspension monitoring data is generated based on the risk suspension queue records.

[0031] In one specific implementation, the queue scheduler reads the node condition sampling records with judgment tags and extracts the node identifier, judgment tag, safety result, and waveform risk result. Node identifiers with a "safe" judgment tag are written to the immediate execution queue, and the corresponding safety result is written to the immediate execution queue record. Node identifiers with a "risk" judgment tag are written to the risk suspension queue, and the corresponding waveform risk result and condition identification data are written to the risk suspension queue record. The immediate execution queue record can also store the node execution order, power-off action type, and generation time. Based on this, the queue scheduler generates power-off control data for the immediate execution queue, which includes the node identifier, control action code, execution batch number, and verification field. The risk suspension queue record is then converted into suspension monitoring data. This suspension monitoring data does not trigger an immediate power-off but serves as input data for subsequent status polling.

[0032] Start status polling for the risk suspension queue. During status polling, generate power consumption reduction and hibernation data, reacquire new operating condition identification data and new current fluctuation sequence data, and remove the corresponding node and generate supplementary power outage control data when both meet the safety conditions. Furthermore, initiate status polling for the risk suspension queue, including: Read the node identifier, waveform risk result, and operating condition identifier data from the risk suspension queue record, and generate polling task data associated with the node identifier; Based on the operating condition identification data, generate power consumption reduction and hibernation data, and write the generation time of the power consumption reduction and hibernation data into the polling task data; Reacquire new operating condition identification data and new current fluctuation sequence data of the node according to the detection interval in the polling task data; Write the new operating condition identification data and the new current fluctuation sequence data back to the risk pending queue record, overwriting the previous round of operating condition identification data and the previous round of current fluctuation sequence data of the corresponding node; After each round of state polling, the detection interval is updated according to the interval step size, and the polling waiting time of the node is accumulated.

[0033] Furthermore, when both safety conditions are met, the corresponding node is removed and supplementary power outage control data is generated, including: The new operating condition label data is compared with the safety label database to obtain the safety results of the new labels; A new average current value and a new peak-valley difference are extracted from the new current fluctuation sequence data and compared with the standby baseline data to obtain new electrical safety results; A single-round safety record is generated when both the new identification safety result and the new electrical safety result are safe. When the number of single-round security records for the same node reaches a predetermined number, the node identifier is removed from the risk suspension queue and written to the reissue execution queue. Based on the node identifier in the reissue execution queue, reissue power outage control data is generated, and the reissue power outage control data is associated and stored with the corresponding single-round safety record.

[0034] In one specific implementation, for the risk suspension queue, the queue scheduler initiates status polling. Status polling is not a simple, repetitive query of all risk nodes at fixed intervals; instead, it reads the node identifier, waveform risk results, and operating condition identifier data from the risk suspension queue records to generate polling task data associated with the node identifier. The polling task data includes the node identifier, current round, detection interval, polling wait time, the most recent operating condition identifier data, and the most recent current fluctuation sequence data. The initial detection interval can be 5s to 20s, and the interval step size can be 5s to 10s, with the specific value determined by the load category identifier. The queue scheduler generates power-saving sleep data based on the operating condition identifier data. For example, when the operating condition identifier data indicates that the device is running but in a power-saving state, the power-saving sleep data includes a sleep request code and a low-power wait field; when the operating condition identifier data indicates that the device is in an uninterruptible state such as writing or upgrading, the power-saving sleep data includes a task completion query field. The generation time of the power-saving sleep data is written into the polling task data to calculate the subsequent polling wait time.

[0035] In each round of state polling, the node sampling program re-acquires the node's new operating condition identification data and new current fluctuation sequence data according to the detection interval in the polling task data. The new operating condition identification data is still triggered by the state probe data, and the new current fluctuation sequence data is still read by the node sampling channel identifier within a new sampling window. The new operating condition identification data and new current fluctuation sequence data are written back to the risk suspension queue record, overwriting the previous round's operating condition identification data and previous round's current fluctuation sequence data for the corresponding node. Overwriting does not delete historical logs, but rather maintains the latest round's data in the risk suspension queue record, ensuring that subsequent safety condition judgments always consider the current node state. After each round of state polling, the queue scheduler updates the detection interval according to the interval step size and accumulates the node's polling waiting time. For example, if a node's initial detection interval is 10s and the interval step size is 5s, then the first three rounds of detection intervals are 10s, 15s, and 20s respectively, and the corresponding polling waiting time is the sum of the actual waiting times for each round. This incremental processing can avoid excessive communication load when the node cannot be powered off for a long time.

[0036] When both the new operating condition identification data and the new current fluctuation sequence data meet the safety conditions, the queue scheduling program removes the corresponding node from the risk suspension queue. The safety conditions consist of two parts. The first part compares the new operating condition identification data with the safety identification database to obtain a new identification safety result. The second part extracts the new average current value and new peak-valley difference from the new current fluctuation sequence data and compares them with the standby baseline data to obtain a new electrical safety result. The standby baseline data is stored in the node attribute table and can include the baseline average current and baseline peak-valley difference of the node in a stable standby state. If the new average current value is not higher than the upper limit of the baseline average current, and the new peak-valley difference is not higher than the upper limit of the baseline peak-valley difference, then the new electrical safety result is safe. The upper limit can be determined as a percentage of the standby baseline data; for example, the upper limit of the baseline average current is 20%, and the upper limit of the baseline peak-valley difference is 30%. When both the new identification safety result and the new electrical safety result are safe, a single-round safety record is generated. To avoid misjudgments caused by transient current drops, this implementation requires that the number of single-round safety records for the same node reach a predetermined number, such as forming single-round safety records for two or three consecutive rounds, before the node identifier is removed from the risk suspension queue and written to the re-issuance execution queue. The queue scheduler generates re-issuance power-off control data based on the node identifier in the re-issuance execution queue, and associates and stores the re-issuance power-off control data with the corresponding single-round safety record, so that the execution log can reflect the complete process of the node changing from risk suspension to re-issuance execution.

[0037] The total waiting time of the cumulative status polling is calculated. When the total waiting time exceeds the intervention threshold and there are stranded nodes in the risk suspension queue, the status polling is terminated and deadlock warning data of the stranded nodes is generated.

[0038] Furthermore, when the total waiting time exceeds the intervention threshold and there are stalled nodes in the risk suspension queue, the status polling is terminated and deadlock warning data for the stalled nodes is generated, including: Summarize the polling wait times of each node in the risk suspension queue to obtain the total wait time; When the total waiting time exceeds the intervention threshold, the node identifiers that have not been removed from the risk suspension queue are extracted as the lingering node identifiers. Read the delayed execution flag corresponding to the stalled node identifier from the scenario rule data; When the delayed execution flag is valid, the delayed power outage task data is generated based on the stalled node identifier and the last round of new operating condition identifier data. When the delayed execution is marked as invalid, deadlock warning data is generated based on the stagnant node identifier, the last round of new operating condition identifier data, and the last round of new current fluctuation sequence data, and the power outage control data corresponding to the stagnant node is frozen.

[0039] In one specific implementation, the status polling also includes a termination branch. The queue scheduler summarizes the polling waiting time of each node in the risk suspension queue to obtain the total waiting time. The total waiting time can be the maximum value of the polling waiting time of each node, or it can be the cumulative time that the risk suspension queue has not been cleared since the scenario started. This implementation prefers to use the cumulative time that the risk suspension queue has not been cleared since the scenario started, so that it is consistent with the waiting time of the global power outage scenario perceived by the user. The intervention threshold can be configured according to the scenario rule data, for example, 180 seconds for the power outage scenario when leaving home and 300 seconds for the power outage scenario when staying up late. When the total waiting time exceeds the intervention threshold, the queue scheduler extracts the node identifiers that have not been removed from the risk suspension queue as the lingering node identifiers. Then, it reads the postponed execution flag corresponding to the lingering node identifier in the scenario rule data. The postponed execution flag is used to distinguish whether the node is allowed to continue waiting after the current global power outage scenario ends. If the delayed execution flag is valid, delayed power-off task data is generated based on the stalled node identifier and the last round of new operating condition identifier data. This delayed power-off task data is written to the scene task cache and its status is polled again during the next scheduled wake-up. If the delayed execution flag is invalid, deadlock warning data is generated based on the stalled node identifier, the last round of new operating condition identifier data, and the last round of new current fluctuation sequence data. The power-off control data corresponding to the stalled node is then frozen. Freezing does not disconnect the node; rather, it prevents the power-off control data for that stalled node from being issued again in this scene batch. Simultaneously, the deadlock warning data is written to the panel's local prompt queue and the scene execution log data.

[0040] Furthermore, it also includes performing a log update step, which includes: After the global power outage scenario instruction processing is completed, the target execution node set, operating condition identification data, current fluctuation sequence data, cross-judgment results, queue change records and control data generation records are summarized to form scenario execution log data. Extract the node records that have generated supplementary power outage control data from the scenario execution log data, and calculate the standby baseline data of the corresponding node before it is removed from the risk suspension queue; Write the standby baseline data into the node attribute table and establish an index relationship with the corresponding node identifier; Extract the node records that generate deadlock warning data from the scenario execution log data, extract the corresponding last round of new current fluctuation sequence data, and generate risk fingerprint candidate data; When the difference between the risk fingerprint candidate data and the existing fingerprint samples in the electrical feature fingerprint database is greater than the fingerprint update threshold, the risk fingerprint candidate data is written into the electrical feature fingerprint database.

[0041] In one specific implementation, after the complete process described above is completed, the log update program executes the log update step. The completion of the global power outage scenario instruction processing can be due to both the immediate execution queue and the reissue execution queue having generated corresponding control data, and the risk suspension queue being empty; or it can be due to the status polling terminating because the total waiting time exceeds the intervention threshold. The log update program summarizes the target execution node set, operating condition identification data, current fluctuation sequence data, cross-judgment results, queue change records, and control data generation records to form scenario execution log data. In the scenario execution log data, each node has at least one master record, which includes the node identifier, the time of entry into the target execution node set, operating condition identification data, current fluctuation sequence data, cross-judgment results, the change trajectory of the queue to which it belongs, and the actual generation records from power outage control data, reissue power outage control data, delayed power outage task data, and deadlock warning data.

[0042] The log update program extracts node records from the scenario execution log data that have generated supplementary power outage control data, and calculates the standby baseline data of the corresponding node before it was removed from the risk suspension queue. The standby baseline data preferentially takes the new current fluctuation sequence data during the period when the node reaches a predetermined number of single-round safety records. If three consecutive rounds of single-round safety records are formed, the average of the three rounds of new average current values ​​can be taken as the baseline average current, and the average of the three rounds of new peak-valley differences can be taken as the baseline peak-valley difference. The standby baseline data obtained in this way originates from the period when the node actually exits the risk state, and is more closely aligned with the load of a residential site than the factory-fixed value. Subsequently, the log update program writes the standby baseline data into the node attribute table and establishes an index relationship with the corresponding node identifier. When the next global power outage scenario command is triggered, this standby baseline data can directly participate in the generation of new electrical safety results.

[0043] For node records that generate deadlock warning data, the log update program extracts the corresponding last round of new current fluctuation sequence data from the scenario execution log data to generate risk fingerprint candidate data. The risk fingerprint candidate data includes the new average current value, new peak-to-valley difference, new pulse count value, and new slope sign sequence extracted from the last round of new current fluctuation sequence data. The log update program then compares the risk fingerprint candidate data with existing fingerprint samples in the electrical feature fingerprint database. The difference values ​​are then used to determine the differences. It can be calculated as a weighted sum of the normalized differences of each item, for example ,in This represents the normalized difference between the new average current value and the existing sample average current value. This represents the normalized difference between the new peak-to-valley difference and the existing sample peak-to-valley differences. This represents the normalized difference between the new pulse count value and the existing sample pulse count values. This indicates the proportion of difference between the new slope sign sequence and the existing sample slope sign sequences. and The sum of these values ​​is 1. The fingerprint update threshold can be determined by the dispersion of historical samples of the same type of load in the electrical feature fingerprint database, for example, by taking the average difference value of the same type of samples plus one standard deviation. When the difference between the risk fingerprint candidate data and the existing fingerprint samples in the electrical feature fingerprint database is greater than the fingerprint update threshold, the log update program writes the risk fingerprint candidate data into the electrical feature fingerprint database. Thus, the electrical feature fingerprint database can be gradually updated as home field devices change, and the updates are based on node records that have already generated deadlock warning data during the execution of this scenario, ensuring continuity in data source and subsequent usage path.

[0044] In the actual business process, before leaving home, the user triggers the power-off entry on the whole-house smart switch panel, and the switch panel receives the global power-off scenario command. Assume the initial node identifier list given by the scenario mapping table includes living room lighting nodes, TV socket nodes, router socket nodes, study room storage device socket nodes, and kitchen general socket nodes. After verification using the area coverage check code and real-time network topology feature code, and connectivity filtering, the target execution node set removes the kitchen general socket node that did not return complete heartbeat response data, and the remaining nodes enter the sampling. The living room lighting node's operating condition identifier data is standby; the average current value and peak-valley difference of the current fluctuation sequence data both fall within the safe range, and the cross-judgment result is a safe cross-judgment result, therefore it is written into the immediate execution queue and power-off control data is generated. The study room storage device socket node's operating condition identifier data is write; although its current value is not high, the identifier result is unsafe, and the cross-judgment result is a risk cross-judgment result, therefore it is written into the risk suspension queue. The TV socket node's operating condition identification data is set to standby, but its current fluctuation sequence data matches the risk operation samples in the electrical feature fingerprint database to a risk matching threshold, thus entering the risk suspension queue. The queue scheduler initiates status polling of the aforementioned risk suspension queue, first generating power-saving sleep data, and then re-collecting new operating condition identification data and new current fluctuation sequence data. When the TV socket node's new operating condition identification data is both set to standby for two consecutive rounds, and the new average current value and new peak-valley difference both meet the standby baseline data requirements, the node is removed from the risk suspension queue, written into the re-execution queue, and re-execution power-off control data is generated. If the study room storage device socket node remains in the write state within the intervention threshold, delayed power-off task data is generated based on the delayed execution flag; if the node is not allowed to delay execution, deadlock warning data is generated and the corresponding power-off control data is frozen. Throughout the process, the target execution node set, operating condition identification data, current fluctuation sequence data, cross-judgment results, queue change records, and control data generation records are ultimately merged into the scenario execution log data for use in updating standby baseline data and risk fingerprint candidate data.

[0045] Example 2: This application also discloses a scene linkage control system for a whole-house smart switch panel.

[0046] Reference Figure 3 A scene-linked control system for a whole-house smart switch panel, comprising: The parsing module is used to receive global power outage scenario commands, parse the scenario feature identifier code and the area coverage verification code, verify the real-time network topology feature code based on the area coverage verification code, and extract the initial node identifier list from the scenario feature identifier code after the verification is passed. The initial node identifier list is then filtered for connectivity to generate the target execution node set. The acquisition module is used to acquire operating condition identification data and current fluctuation sequence data within the same sampling window for the target set of execution nodes; The comparison module is used to compare the operating condition identification data with the safety identification database, match the current fluctuation sequence data with the electrical feature fingerprint database, and fuse the comparison results and matching results to generate a cross-judgment result. The splitting module is used to split the target execution node set into an immediate execution queue and a risk suspension queue according to the cross-determination results, and generate power-off control data for the immediate execution queue; The polling module is used to start status polling of the risk suspension queue. During status polling, power consumption reduction and hibernation data are generated, new operating condition identification data and new current fluctuation sequence data are reacquired, and the corresponding node is removed and reissued power outage control data is generated when both meet the safety conditions. The output module is used to accumulate the total waiting time of status polling. When the total waiting time exceeds the intervention threshold and there are stranded nodes in the risk suspension queue, the status polling is terminated and deadlock warning data of the stranded nodes is generated.

[0047] The above content is merely an example and illustration of the concept of the present invention. Those skilled in the art can make various modifications or additions to the specific embodiments described or use similar methods to replace them, as long as they do not deviate from the concept of the invention, they should all fall within the protection scope of the present invention.

[0048] In the description of this specification, references to terms such as "an embodiment," "example," "specific example," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0049] The preferred embodiments of the present invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to any specific implementation. Clearly, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention.

Claims

1. A scene-linked control method for a whole-house smart switch panel, characterized in that, Includes the following steps: Receive global power outage scenario command, parse scenario feature identifier code and area coverage verification code, verify real-time network topology feature code based on area coverage verification code, and extract initial node identifier list from scenario feature identifier code after verification, perform connectivity filtering on initial node identifier list, and generate target execution node set; For the target set of execution nodes, obtain operating condition identification data and current fluctuation sequence data within the same sampling window; The operating condition identification data is compared with the safety identification database, and the current fluctuation sequence data is matched with the electrical feature fingerprint database. The comparison results and matching results are then fused to generate a cross-judgment result. Based on the cross-determination results, the target execution node set is split into an immediate execution queue and a risk suspension queue, and power outage control data for the immediate execution queue is generated. Start status polling for the risk suspension queue. During status polling, generate power consumption reduction and hibernation data, reacquire new operating condition identification data and new current fluctuation sequence data, and remove the corresponding node and generate supplementary power outage control data when both meet the safety conditions. The total waiting time of the cumulative status polling is calculated. When the total waiting time exceeds the intervention threshold and there are stranded nodes in the risk suspension queue, the status polling is terminated and deadlock warning data of the stranded nodes is generated.

2. The scene linkage control method for a whole-house smart switch panel according to claim 1, characterized in that, Receive a global power outage scenario command, parse the scenario feature identifier code and area coverage verification code, and generate a target execution node set, including: The data frames of the global power outage scenario command are unpacked, and the scenario feature identifier code and the area coverage verification code are extracted. Read the real-time network topology feature code from the local topology record, compare the area coverage check code with the real-time network topology feature code to obtain the topology comparison difference; When the topology comparison difference is less than the topology change threshold, the initial node identifier list is extracted by addressing in the scene mapping table based on the scene feature identifier code. Connectivity detection data is generated according to the initial node identifier list, and heartbeat response data is received from the nodes; The heartbeat response data is parsed to obtain the communication delay value and the response integrity mark. Node identifiers whose communication delay value exceeds the packet loss limit and whose response integrity mark is invalid are deleted. The remaining node identifiers are combined into a target execution node set.

3. The scene linkage control method for a whole-house smart switch panel according to claim 1, characterized in that, For the target set of execution nodes, obtain operating condition identification data and current fluctuation sequence data within the same sampling window, including: Based on the node identifier in the target execution node set, query the node attribute table to obtain the node protocol address and node sampling channel identifier; Generate status probe data according to the node protocol address, receive operating condition response data corresponding to the node identifier, and extract operating condition identifier data from the operating condition response data; Read the current sampling data in the sampling window according to the node sampling channel identifier, and sort the current sampling data according to the sampling timestamp to form current fluctuation sequence data; Align the received timestamp of the operating condition identification data with the sampling timestamp of the current fluctuation sequence data to generate a node operating condition sampling record.

4. The scene linkage control method for a whole-house smart switch panel according to claim 4, characterized in that, Waveform matching is performed between current fluctuation sequence data and an electrical feature fingerprint database, including: The current fluctuation sequence data is denoised and smoothed to obtain a smooth current sequence; The average current value, peak-to-valley difference, pulse count value, and slope sign sequence are extracted from the smoothed current sequence and combined to generate an electrical feature vector. Based on the load category identifier in the node attribute table, extract the corresponding fingerprint sample group from the electrical feature fingerprint database; The electrical feature vector is compared with the fingerprint sample group item by item to obtain the waveform matching degree; Matching results are generated based on waveform matching degree and risk matching threshold.

5. The scene linkage control method for a whole-house smart switch panel according to claim 4, characterized in that, The fusion of comparison and matching results generates a cross-determination result, including: The working condition label data is compared with the safety label database to obtain the label safety result. The waveform matching degree in the matching results is compared with the risk matching threshold to obtain the waveform risk result; When the security result is identified as secure and the waveform risk result is identified as a missed risk, a security cross-judgment result is generated. When the safety result is marked as unsafe and the waveform risk result is that any of the risk items in the hit risk are met, a risk cross-judgment result is generated; The results of safe cross-judgment and risk cross-judgment are written into the node operating condition sampling record respectively, forming a node operating condition sampling record with a judgment label.

6. The scene linkage control method for a whole-house smart switch panel according to claim 5, characterized in that, Based on the cross-determination results, the target execution node set is split into an immediate execution queue and a risk-suspended queue, including: Read the node operating condition sampling records with judgment labels, and extract the node identifier, judgment label, identification safety result, and waveform risk result; Write the node identifier that is determined to be safe into the immediate execution queue, and write the corresponding safe identifier result into the immediate execution queue record; Write the node identifier that is identified as risk into the risk suspension queue, and write the corresponding waveform risk result and working condition identifier data into the risk suspension queue record; Power outage control data is generated based on the immediate execution queue records, and suspension monitoring data is generated based on the risk suspension queue records.

7. The scene linkage control method for a whole-house smart switch panel according to claim 6, characterized in that, Initiate status polling for the risk suspension queue, including: Read the node identifier, waveform risk result, and operating condition identifier data from the risk suspension queue record, and generate polling task data associated with the node identifier; Based on the operating condition identification data, generate power consumption reduction and hibernation data, and write the generation time of the power consumption reduction and hibernation data into the polling task data; Reacquire new operating condition identification data and new current fluctuation sequence data of the node according to the detection interval in the polling task data; Write the new operating condition identification data and the new current fluctuation sequence data back to the risk pending queue record, overwriting the previous round of operating condition identification data and the previous round of current fluctuation sequence data of the corresponding node; After each round of state polling, the detection interval is updated according to the interval step size, and the polling waiting time of the node is accumulated.

8. The scene linkage control method for a whole-house smart switch panel according to claim 7, characterized in that, When both safety conditions are met, the corresponding node is removed and supplementary power outage control data is generated, including: The new operating condition label data is compared with the safety label database to obtain the safety results of the new labels; A new average current value and a new peak-valley difference are extracted from the new current fluctuation sequence data and compared with the standby baseline data to obtain new electrical safety results; A single-round safety record is generated when both the new identification safety result and the new electrical safety result are safe. When the number of single-round security records for the same node reaches a predetermined number, the node identifier is removed from the risk suspension queue and written to the reissue execution queue. Based on the node identifier in the reissue execution queue, reissue power outage control data is generated, and the reissue power outage control data is associated and stored with the corresponding single-round safety record.

9. A scene linkage control method for a whole-house smart switch panel according to claim 7, characterized in that, When the total waiting time exceeds the intervention threshold and there are stranded nodes in the risk suspension queue, the status polling is terminated and deadlock warning data for the stranded nodes is generated, including: Summarize the polling wait times of each node in the risk suspension queue to obtain the total wait time; When the total waiting time exceeds the intervention threshold, the node identifiers that have not been removed from the risk suspension queue are extracted as the lingering node identifiers. Read the delayed execution flag corresponding to the stalled node identifier from the scenario rule data; When the delayed execution flag is valid, the delayed power outage task data is generated based on the stalled node identifier and the last round of new operating condition identifier data. When the delayed execution is marked as invalid, deadlock warning data is generated based on the stuck node identifier, the last round of new operating condition identifier data and the last round of new current fluctuation sequence data, and the power outage control data corresponding to the stuck node is frozen.

10. A scene linkage control system for a whole-house smart switch panel, applied to the scene linkage control method for a whole-house smart switch panel as described in any one of claims 1-9, characterized in that, include: The parsing module is used to receive global power outage scenario commands, parse the scenario feature identifier code and the area coverage verification code, verify the real-time network topology feature code based on the area coverage verification code, and extract the initial node identifier list from the scenario feature identifier code after the verification is passed. The initial node identifier list is then filtered for connectivity to generate the target execution node set. The acquisition module is used to acquire operating condition identification data and current fluctuation sequence data within the same sampling window for the target set of execution nodes; The comparison module is used to compare the operating condition identification data with the safety identification database, match the current fluctuation sequence data with the electrical feature fingerprint database, and fuse the comparison results and matching results to generate a cross-judgment result. The splitting module is used to split the target execution node set into an immediate execution queue and a risk suspension queue according to the cross-determination results, and generate power-off control data for the immediate execution queue; The polling module is used to start status polling of the risk suspension queue. During status polling, power consumption reduction and hibernation data are generated, new operating condition identification data and new current fluctuation sequence data are reacquired, and the corresponding node is removed and reissued power outage control data is generated when both meet the safety conditions. The output module is used to accumulate the total waiting time of status polling. When the total waiting time exceeds the intervention threshold and there are stranded nodes in the risk suspension queue, the status polling is terminated and deadlock warning data of the stranded nodes is generated.

Images