A Smart Lighting Control Data Processing Method and System

Abstract

This invention discloses a smart lighting control data processing method and system, relating to the field of smart lighting control. This method and system first completes data standardization and vector loading, then uses public vectors to achieve a binary mode separation between public and private scenarios. It employs two independent processes—constraint-first and user-driven—to conduct multi-dimensional data processing, ultimately generating control commands and completing closed-loop feedback. This method distinguishes between scenario-specific decision logic to adapt to different usage needs, utilizes vector mapping for flexible parameter adjustment, and possesses good scalability. It also adds a tiered fault degradation strategy to improve operational fault tolerance and uses a weighted fusion approach to coordinate user operation and automatic control, adapting to actual usage requirements.

Description

Technical Field

[0001] This invention relates to the field of intelligent lighting control, and in particular to an intelligent lighting control data processing method and system. Background Technology

[0002] With the rapid development of IoT and AI technologies, smart lighting technology has been widely applied in various fields such as urban public lighting, commercial complexes, industrial plants, and smart homes. Existing smart lighting control technology has evolved from early timed on / off control to single-parameter automatic control based on human body and light sensing, and now to the mainstream multi-sensor data fusion control. Current mainstream technologies typically collect multi-source sensing data such as time, illuminance, and the number of people, and calculate the target brightness value through a preset weighted summation algorithm. Some solutions also incorporate scene mode switching, remote control, and energy consumption statistics, achieving a certain degree of automated control of the lighting system and reducing lighting energy consumption.

[0003] However, existing intelligent lighting control data processing technologies, on the one hand, adopt a single-track unified control architecture, where all application scenarios share the same control processing logic, failing to consider the fundamental differences in the priority of control objectives across different scenarios. For public lighting scenarios where safety, reliability, and unified service are core requirements, existing technologies either employ a constant-on mode to ensure safety, resulting in energy waste, or use sensor control for energy saving, leading to delayed lighting response and potential safety hazards. For indoor lighting scenarios where user experience and autonomous control are core requirements, the automatic control logic of existing technologies often conflicts with actual user needs, requiring frequent manual intervention and resulting in a poor intelligent experience. On the other hand, existing technologies mostly use multi-parameter fusion algorithms with fixed weights, lacking dynamic adjustment mechanisms. When key variables such as illumination and personnel reach critical thresholds, they cannot achieve single-variable-dominated control, easily leading to control logic chaos. Furthermore, existing technologies have insufficient fault tolerance for sensor failures; a single sensor failure can cause weight calculation failure, rendering the system unable to operate normally. To address these issues, this invention provides an intelligent lighting control data processing method and system to solve the aforementioned problems. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides an intelligent lighting control data processing method and system. This method solves several problems with current intelligent lighting control data processing technologies. Firstly, existing technologies employ a single-track unified control architecture, where all application scenarios share a single set of control processing logic, failing to consider the fundamental differences in the priority of control objectives across different scenarios. For public lighting scenarios where safety, reliability, and unified service are core requirements, existing technologies either use a constant-on mode to ensure safety, resulting in energy waste, or employ sensor control for energy saving, leading to delayed lighting response and potential safety hazards. Secondly, for indoor lighting scenarios where user experience and autonomous control are core requirements, the automatic control logic of existing technologies often conflicts with actual user needs, requiring frequent manual intervention and resulting in a poor intelligent experience. Thirdly, existing technologies often use fixed-weight multi-parameter fusion algorithms, lacking dynamic adjustment mechanisms. When key variables such as illumination and personnel reach critical thresholds, single-variable-dominated control cannot be achieved, easily leading to control logic chaos. Furthermore, existing technologies lack sufficient fault tolerance for sensor failures; a single sensor failure can cause weight calculation failure, preventing the system from operating normally.

[0005] To achieve the above objectives, the present invention provides the following technical solution: an intelligent lighting control data processing method and system, comprising: S1. System initialization and multi-source sensing data standardization processing: Load the five pre-configured control vectors, and collect data on time, illumination, personnel and sensor working status; S2. Binary Mode Rigid Diversion and Basic Rule Set Generation: Perform mode judgment based on the value of the commonity vector and the preset diversion threshold, load the basic rule set of the corresponding mode based on the mode judgment result, the basic rule set defines the meta-rules of all control behaviors under the mode, and initialize the basic control parameter set of the mode based on the basic rule set; S3. When it is determined to be a public mode, perform constraint-priority multi-dimensional data processing: generate a basic decision vector based on the lighting demand determination rules specific to the public mode, and sequentially perform comfort dimension constraint processing, energy-saving dimension constraint optimization, and collaborative dimension constraint enhancement to generate the final decision vector of the public mode; S4. When it is determined to be a personalized mode, perform user-led multi-level data processing: generate a basic decision vector based on the lighting demand determination rules specific to the personalized mode, and sequentially perform user integration processing of the comfort dimension, adaptive optimization of the energy-saving dimension, and scenario-based enhancement of the collaboration dimension to generate the final decision vector of the personalized mode; S5. Standardized control command generation and closed-loop feedback processing.

[0006] Preferably, in step S1, raw time data, raw illumination data, raw personnel data, and sensor working status signals are collected, and a normalization transformation is performed on all raw sensing data to generate four input variables whose values ​​are limited to the interval [0,1]. In step S1, the five control vectors include a commonality vector, a personalization vector, a comfort vector, an energy-saving vector, and a synergy vector, wherein: The publicness vector is used to characterize the degree of publicness of the service objects in the scenario and their requirements for unified service standards; The personalized vector is used to characterize the specificity of the scene service object and its requirements for personalized customization; The comfort vector is used to characterize the requirements for lighting control precision, response speed, and smoothness; The energy efficiency vector is used to characterize the energy efficiency requirements without breaking the basic constraints. The coordination vector is used to characterize the requirements for coordinated control of lighting in adjacent areas.

[0007] Preferably, in step S2, the rigid splitting of the binary pattern and the generation of the basic rule set specifically involve: When the value of the commonality vector is greater than the preset diversion threshold, it is determined to be in common mode, and the common mode basic rule set is loaded. The common mode basic rule set includes security constraints, minimum service standards, control priority order, and sensor fault degradation strategies. When the value of the commonality vector is less than or equal to the preset diversion threshold, it is determined to be in personalized mode, and the personalized mode basic rule set is loaded. The personalized mode basic rule set includes user permission range, intervention priority, adaptive learning rules, and sensor fault degradation strategies. The preset diversion threshold is a configurable parameter with a value in the range (0,1).

[0008] Preferably, in step S3, the common pattern constraint priority type multi-level dimension data processing specifically includes: S31. Execute public mode exclusive lighting demand determination: When the value of the standardized illumination variable is less than or equal to the public mode illumination threshold, or the value of the standardized time variable is greater than or equal to the public mode time threshold, it is determined that there is a lighting demand; when there is a lighting demand, when the value of the standardized personnel variable is greater than 0, or the value of the standardized time variable is greater than or equal to the public mode time threshold, it is determined that the lights are on, and a basic decision vector containing the on / off status flag, the minimum allowable brightness value, and the maximum allowable brightness value is generated; S32. Perform comfort dimension constraint processing: Generate a set of comfort processing parameters including control resolution parameters, adjustment granularity parameters, and smoothness parameters based on the values ​​of the comfort vector, transform the basic decision vector, apply the safety constraints defined by the basic rule set to verify the transformation result, and correct it to the constraint boundary if it exceeds the constraint range. S33. Perform energy-saving dimension constraint optimization: Generate an energy-saving optimization parameter set containing idle delay parameters, intensity attenuation parameters, and time period weight parameters based on the energy-saving vector values, optimize the comfort processing results, and verify the optimization results by applying the minimum service standard defined by the basic rule set; S34. Enhance the coordination dimension constraint: Generate a set of coordination control parameters including coordination range parameters, pre-trigger lead time parameters, and data interaction frequency parameters based on the values ​​of the coordination vector. Exchange state data and intermediate decision vectors with adjacent nodes within the coordination range to enhance the energy-saving optimization results. Apply the consistency constraints defined by the basic rule set to verify the enhanced results.

[0009] Preferably, in step S4, the personalized mode user-driven multi-dimensional data processing specifically includes: S41. Execute personalized mode exclusive lighting demand determination: When a user actively turns on the light, or the value of the standardized illumination variable is less than or equal to the personalized mode illumination threshold and the value of the standardized personnel variable is greater than 0, or the value of the standardized time variable is greater than or equal to the personalized mode time threshold and the value of the standardized personnel variable is greater than 0, it is determined that there is a lighting demand and a basic decision vector for the light-on state is generated. S42. Perform user intervention detection and weight calculation: Real-time detection of user manual intervention events, calculation of user intervention weight based on the value of the personalized vector, the user intervention weight is positively correlated with the value of the personalized vector; when user intervention is detected, a user intervention vector containing brightness adjustment value, color temperature adjustment value, and scene mode value is generated. S43. Perform user fusion processing for comfort dimension: Generate a set of comfort processing parameters based on the values ​​of the comfort vector, and generate an intermediate decision vector for automatic control comfort; when there is effective user intervention, the user intervention vector and the intermediate decision vector for automatic control comfort are weighted and fused according to the user intervention weight; S44. Perform adaptive optimization of energy efficiency dimension: Generate a set of energy efficiency optimization parameters based on the value of the energy efficiency vector, optimize the comfort processing results, record user intervention data and corresponding contextual environment data, and update the user preference model; S45. Enhance the collaborative dimension in a scenario-based manner: Generate a set of collaborative control parameters based on the values ​​of the collaborative vector. When the user selects a preset scenario mode, load the corresponding scenario's collaborative control rules to enhance the energy-saving optimization results.

[0010] Preferably, in step S2, the specific preprocessing process for the four main input variables includes: Raw data filtering: The collected raw illuminance data and raw personnel count data are subjected to noise suppression using a moving average filtering algorithm. The size of the filtering window can be configured according to the needs of the scene. For abnormal data that exceeds the physical range, the 3σ criterion is used for identification and removal. Time synchronization and calibration: All edge gateways and terminal nodes are synchronized with the cloud time server via the NTP network time protocol. For offline scenarios where the network cannot be connected, a high-precision real-time clock module is used for local timing and automatically calibrated after the network is restored. Standardization and normalization transformation: A linear normalization transformation is performed on the filtered raw data to map it to the [0,1] interval; for time data, the sunrise and sunset times of the day and place are calculated using astronomical algorithms based on the latitude and longitude coordinates of the edge gateway, and used as the reference value for normalization.

[0011] Preferably, steps S3 and S4 further include a full-process sensor fault degradation processing step, specifically: In public mode, when any sensor malfunction is detected, the input data of that sensor is automatically blocked, and the data of the remaining normal sensors is used for control; when two or more sensor malfunctions are detected, it automatically degrades to pure time control mode, and only performs on / off control based on the standardized time variable; when all sensor malfunctions are detected, it is forced to enter the nighttime constant-on mode. In personalized mode, when any sensor malfunction is detected, the input data of that sensor is automatically blocked, and predictive control is performed using the user preference model; when two or more sensor malfunctions are detected, it automatically degrades to a pure user manual control mode, and control is only executed when the user manually operates it.

[0012] Preferably, the processing parameters for each dimension in steps S3 and S4 are dynamically generated through the corresponding vector mapping function. The vector mapping function is a monotonic function defined in the interval [0,1], whose input is the value of the corresponding control vector and whose output is the set of processing parameters for that dimension. Different scenarios can achieve differentiated control by configuring different vector mapping functions without modifying the core processing logic.

[0013] Preferably, the specific steps of step S5 are as follows: The elements of the final decision vector are mapped to the numerical range supported by the lighting actuator, encoded into standardized control commands according to the preset communication protocol, and sent to the corresponding lighting actuator. The operating status data and actual output parameters of the actuator are collected, and a feedback data vector is generated to update the system status parameters for the next control cycle.

[0014] Preferred, including: The initialization and data processing module is used to load pre-configured control vectors, collect and standardize multi-source sensing data, and generate standardized input variables and sensor state sets. The pattern splitting and rule loading module is used to perform binary pattern judgment based on the commonity vector, load the basic rule set of the corresponding pattern, and initialize the basic control parameters. The common pattern processing module is used to perform constraint-priority multi-dimensional data processing and generate the final decision vector of the common pattern. The personalized pattern processing module is used to perform user-driven multi-level and multi-dimensional data processing and generate the final decision vector of the personalized pattern. The instruction output and feedback module is used to convert the final decision vector into standardized control instructions and send them to the lighting actuator, and to collect execution feedback data to update the system status.

[0015] The technical effects and advantages of this invention are as follows: This intelligent lighting control data processing method and system, through a binary rigid splitting and dual-track parallel data processing architecture, isolates the control logic according to the core needs of the scenario. For public lighting scenarios where safety and reliability are paramount, a constraint-priority processing logic is adopted to achieve graded energy saving while ensuring basic nighttime lighting and minimum safe brightness in densely populated areas, thus balancing public safety and energy consumption control. For indoor lighting scenarios where user experience is paramount, a user-driven processing logic is adopted, allowing users to control the lighting beyond ambient light and time limitations, while retaining automatic control as an auxiliary supplement, significantly improving the user experience. This intelligent lighting control data processing method and system adopts a vector parameterized dimensional control mechanism. All control parameters are dynamically generated through vector mapping functions, eliminating the need to hard-code independent control rules for different scenarios. By adjusting the values ​​of the five major control vectors, it can quickly adapt to any application scenario, significantly shortening the system development cycle and reducing later maintenance costs. At the same time, the modular dimensional processing architecture allows for the addition of new control functions simply by adding the corresponding processing module, without modifying the core logic, significantly improving the system's scalability. This intelligent lighting control data processing method and system establishes a full-process hierarchical sensor fault degradation handling mechanism, adopting differentiated degradation strategies for different fault levels and scenario types. In public scenarios, when a sensor fails, it automatically degrades to pure time control or forced constant-on mode to ensure that basic lighting functions are not affected. In personalized scenarios, when a sensor fails, it automatically degrades to user preference predictive control or pure manual control mode, avoiding the problem of system crash due to a single sensor failure and significantly improving the system's operational stability. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art 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.

[0017] Figure 1 This is a flowchart illustrating the overall architecture of the present invention; Figure 2 This is a flowchart illustrating the three-level constraints of the common mode of this invention. Figure 3 This is a diagram for determining the personalized lighting requirements of this invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] To better understand the above technical solutions, the following will provide a detailed explanation of the technical solutions in conjunction with the accompanying drawings and specific implementation methods.

[0020] This invention discloses an intelligent lighting control data processing method and system, according to the appendix. Figure 1 As shown, it includes: S1. System initialization and multi-source sensing data standardization processing: Load the five pre-configured control vectors, and collect data on time, illumination, personnel and sensor working status; S2. Binary Mode Rigid Diversion and Basic Rule Set Generation: Perform mode judgment based on the value of the commonity vector and the preset diversion threshold, load the basic rule set of the corresponding mode based on the mode judgment result, define the meta-rules of all control behaviors under this mode, and initialize the basic control parameter set of this mode based on the basic rule set. S3. When it is determined to be a public mode, perform constraint-priority multi-dimensional data processing: generate a basic decision vector based on the lighting demand determination rules specific to the public mode, and sequentially perform comfort dimension constraint processing, energy-saving dimension constraint optimization, and collaborative dimension constraint enhancement to generate the final decision vector of the public mode; S4. When it is determined to be a personalized mode, perform user-led multi-level data processing: generate a basic decision vector based on the lighting demand determination rules specific to the personalized mode, and sequentially perform user integration processing of the comfort dimension, adaptive optimization of the energy-saving dimension, and scenario-based enhancement of the collaboration dimension to generate the final decision vector of the personalized mode; S5. Standardized control command generation and closed-loop feedback processing: Map each element of the final decision vector to the numerical range supported by the lighting actuator, encode it into standardized control commands according to the preset communication protocol and send it to the corresponding lighting actuator, collect the actuator's operating status data and actual output parameters, and generate a feedback data vector to update the system status parameters for the next control cycle.

[0021] In this embodiment, the method adopts a distributed cloud-edge collaborative architecture, consisting of three layers: a cloud-based big data platform, a regional edge gateway, and a terminal sensing execution node. All data processing logic follows the principle of "real-time edge processing and global cloud optimization." The system adopts a fixed-cycle pipeline-style data processing mode, and the data processing flow within each control cycle is as follows: The cloud-based big data platform synchronizes the pre-configured control vector set and vector mapping function parameters to all regional edge gateways through a batch parameter distribution interface. After receiving the parameters, the edge gateways store them in their local non-volatile memory and automatically load them after the system restarts. Terminal sensing nodes collect raw sensing data at a preset sampling frequency and upload the data to the edge gateway of their respective area via a low-power wireless communication protocol; the edge gateway performs timestamp alignment and data integrity verification on the received data and discards invalid data packets; The edge gateway performs multi-source sensing data standardization processing, converting raw sensing data of different formats and ranges into standardized input variables with values ​​limited to the range of [0,1], and generating the corresponding sensor state set; The edge gateway performs rigid traffic splitting in binary mode and loads the basic rule set. It determines the mode of the current scene based on the value of the commonity vector, loads the basic rule set of the corresponding mode, and initializes the basic control parameters. Based on the pattern determination result, the edge gateway executes the corresponding multi-level and multi-dimensional data processing flow to generate the final decision vector. During the processing, all intermediate data is stored in the local cache of the edge gateway for subsequent troubleshooting and system optimization. The edge gateway encodes the final decision vector into standardized control commands and sends them to the corresponding terminal execution nodes via the bus; after receiving the commands, the terminal execution nodes execute the corresponding control actions and feed back the execution results to the edge gateway. The edge gateway periodically uploads operational data, control logs, and energy consumption data to the cloud-based big data platform. The cloud platform mines and analyzes the massive amounts of data, generates system optimization suggestions, and can dynamically adjust the control vector and mapping function parameters based on the analysis results.

[0022] Furthermore, in step S1, raw time data, raw illumination data, raw personnel data, and sensor operating status signals are collected, and a normalization transformation is performed on all raw sensing data to generate four input variables whose values ​​are limited to the interval [0,1]. In step S1, the five control vectors include the commonality vector, the individualization vector, the comfort vector, the energy efficiency vector, and the synergy vector, among which: The publicness vector is used to characterize the degree of publicness of the service objects in a scenario and their requirements for unified service standards. Personalized vectors are used to characterize the specificity of the service objects in a scenario and their requirements for personalized customization; The comfort vector is used to characterize the requirements for lighting control precision, response speed, and smoothness. The energy efficiency vector is used to characterize energy-saving requirements without exceeding the basic constraints. The coordination vector is used to characterize the requirements for coordinated control of lighting in adjacent areas.

[0023] In this embodiment, the values ​​of the five control vectors are all in the range of [0,1]. The common vector and the personalized vector are binary basic vectors, which serve as the highest-level meta-rules to constrain the scope and strength of the three secondary functional vectors. The specific mechanism is as follows: The constraining effect of the commonality vector: The commonality vector serves as the safety baseline for all control behaviors, and its value is negatively correlated with the maximum permissible influence strength of the three secondary functional vectors. The larger the value of the commonality vector, the stronger the constraint on the secondary functional vectors, ensuring a unified service standard and security reliability in common scenarios; The amplifying effect of personalized vectors: As a dimension for enhancing user experience, the value of personalized vectors is positively correlated with the actual strength of the three secondary functional vectors. The larger the value of the personalized vector, the more significant the impact of the secondary functional vectors on the control effect, thus meeting the personalized needs of private scenarios; Independence of secondary functional vectors: The three secondary functional vectors of comfort, energy saving and synergy act on three independent dimensions of lighting control, and there is no direct constraint relationship between them. The value of any one vector can be adjusted individually according to needs without affecting the control effect of the other two dimensions. Vector priority ranking: During control logic execution, vector priorities are ranked from highest to lowest as follows: common vector > personalized vector > comfort vector > energy-saving vector > collaborative vector. When control requirements of different vectors conflict, the requirements of the higher-priority vector are satisfied first. The generation of the four input variables is as follows: The standardized time variable is calculated from the raw time data and the local astronomical sunrise and sunset times of the day. A value of 0 corresponds to the sunrise time, a value of 1 corresponds to the sunset time, a value greater than 1 corresponds to the nighttime period after sunset, and a value less than 0 corresponds to the early morning period before sunrise. Standardized illumination variables are calculated using raw illuminance data and preset illuminance baseline values. A value of 0 corresponds to a completely dark environment, and a value of 1 corresponds to a bright environment where the illuminance reaches or exceeds the baseline value. The standardized personnel variable is calculated by using the original personnel count data and the maximum capacity of the preset area. A value of 0 corresponds to no one in the area, and a value of 1 corresponds to the number of people in the area reaching or exceeding the maximum capacity. The sensor status set contains three binary status variables, which correspond to the working status of the time system, light sensor, and personnel sensor, respectively. A value of 0 indicates that the corresponding device is working normally, and a value of 1 indicates that the corresponding device is faulty or has missing data.

[0024] Furthermore, in step S2, the rigid splitting of the binary pattern and the generation of the basic rule set are specifically as follows: When the value of the commonality vector is greater than the preset diversion threshold, it is determined to be in common mode, and the basic rule set of common mode is loaded. The basic rule set of common mode includes security constraints, minimum service standards, control priority order, and sensor fault degradation strategy. When the value of the commonality vector is less than or equal to the preset diversion threshold, it is determined to be in personalized mode, and the basic rule set of personalized mode is loaded. The basic rule set of personalized mode includes user permission range, intervention priority, adaptive learning rules, and sensor fault degradation strategy. The preset diversion threshold is a configurable parameter with a value in the range (0,1).

[0025] In this embodiment, a dynamically updatable basic rule set management mechanism is adopted to achieve flexible configuration and rapid switching of control rules in different scenarios. The specific implementation is as follows: Structured storage of basic rule sets: Both public and personalized rule sets are stored in a structured data format, including fields such as rule number, rule conditions, rule actions, priority, and effective time; the rule sets can be visually edited and configured through a cloud platform; Hot update mechanism for rule sets: The cloud platform can send rule set update instructions to the edge gateway. After receiving the instructions, the edge gateway can complete the hot update of the rule set without interrupting system operation. A dual-caching mechanism is used during the update process to ensure that the rule set can be automatically rolled back to the old version if the update fails. Smooth transition of mode switching: When the value of the commonity vector crosses the splitting threshold and causes mode switching, the system uses a gradual transition to adjust the control parameters to avoid sudden changes in the light state; the transition time can be configured according to the scene requirements. Rule set version control: The cloud platform stores and manages all historical versions of the basic rule set, supports version rollback and comparative analysis of the rule set; when an anomaly occurs during system operation, it can be quickly rolled back to the previous stable version.

[0026] According to the appendix Figure 2 As shown, specifically disclosed in step S3, the processing of multi-level dimensional data with common pattern constraint priority is as follows: S31. Execute public mode exclusive lighting demand determination: When the value of the standardized illumination variable is less than or equal to the public mode illumination threshold, or the value of the standardized time variable is greater than or equal to the public mode time threshold, it is determined that there is a lighting demand; when there is a lighting demand, when the value of the standardized personnel variable is greater than 0, or the value of the standardized time variable is greater than or equal to the public mode time threshold, it is determined that the lights are on, and a basic decision vector containing the on / off status flag, the minimum allowable brightness value, and the maximum allowable brightness value is generated; S32. Perform comfort dimension constraint processing: Generate a set of comfort processing parameters including control resolution parameters, adjustment granularity parameters, and smoothness parameters based on the values ​​of the comfort vector, transform the basic decision vector, apply the safety constraints defined by the basic rule set to verify the transformation result, and correct it to the constraint boundary if it exceeds the constraint range. S33. Perform energy-saving dimension constraint optimization: Generate an energy-saving optimization parameter set containing idle delay parameters, intensity attenuation parameters, and time period weight parameters based on the energy-saving vector values, optimize the comfort processing results, and verify the optimization results by applying the minimum service standard defined by the basic rule set; S34. Enhance the coordination dimension constraint: Generate a set of coordination control parameters including coordination range parameters, pre-trigger lead time parameters, and data interaction frequency parameters based on the values ​​of the coordination vector. Exchange state data and intermediate decision vectors with adjacent nodes within the coordination range to enhance the energy-saving optimization results. Apply the consistency constraints defined by the basic rule set to verify the enhanced results.

[0027] In this embodiment, the common mode constraint priority data processing flow includes a full-process multi-level constraint verification mechanism to ensure that all control behaviors do not exceed the security boundaries and service standards defined by the basic rule set. The specific implementation is as follows: Formal description of constraints: All constraints in the basic rule set are described using a formal language and converted into computer-executable logical expressions; each constraint includes constraint type, constraint scope, and priority attributes. Hierarchical constraint verification: After processing at each dimension, the corresponding constraint verification is performed immediately. During the verification process, the processing result is compared with the constraint conditions. If it exceeds the constraint range, the processing result is corrected to the nearest constraint boundary. Constraint conflict resolution strategy: When multiple constraints conflict, they are processed in descending order of priority. High-priority constraints must be satisfied, while low-priority constraints can be adjusted appropriately without affecting the high-priority constraints. Constraint violation logging: All constraint violation events are logged in detail, including the violation time, the violation constraint, the original processing result, and the corrected result. Log data is periodically uploaded to a cloud platform for analyzing the rationality of constraints and optimizing control rules.

[0028] According to the appendix Figure 3 As shown, the specifically disclosed personalized, user-driven, multi-dimensional data processing involves the following: S41. Execute personalized mode exclusive lighting demand determination: When a user actively turns on the light, or the value of the standardized illumination variable is less than or equal to the personalized mode illumination threshold and the value of the standardized personnel variable is greater than 0, or the value of the standardized time variable is greater than or equal to the personalized mode time threshold and the value of the standardized personnel variable is greater than 0, it is determined that there is a lighting demand and a basic decision vector for the light-on state is generated. S42. Perform user intervention detection and weight calculation: Real-time detection of user manual intervention events, calculation of user intervention weight based on the value of the personalized vector, the user intervention weight is positively correlated with the value of the personalized vector; when user intervention is detected, a user intervention vector containing brightness adjustment value, color temperature adjustment value, and scene mode value is generated. S43. Perform user fusion processing for comfort dimension: Generate a set of comfort processing parameters based on the values ​​of the comfort vector, and generate an intermediate decision vector for automatic control comfort; when there is effective user intervention, the user intervention vector and the intermediate decision vector for automatic control comfort are weighted and fused according to the user intervention weight; S44. Perform adaptive optimization of energy efficiency dimension: Generate a set of energy efficiency optimization parameters based on the value of the energy efficiency vector, optimize the comfort processing results, record user intervention data and corresponding contextual environment data, and update the user preference model; S45. Enhance the collaborative dimension in a scenario-based manner: Generate a set of collaborative control parameters based on the values ​​of the collaborative vector. When the user selects a preset scenario mode, load the corresponding scenario's collaborative control rules to enhance the energy-saving optimization results.

[0029] In this embodiment, the personalized mode employs a user preference modeling technique based on incremental learning to achieve adaptive learning and prediction of user lighting habits, specifically implemented as follows: User behavior data collection: The system collects all manual intervention behaviors of users in real time, including turning lights on / off, adjusting brightness, adjusting color temperature, and switching scenes, and records the contextual environmental data when each behavior occurs, such as time, illuminance, number of people, day of the week, and holidays. User preference model construction: A linear regression model is used to construct the user preference model, with contextual environmental data as input features and user-set lighting parameters as output labels; the model parameters are trained using the stochastic gradient descent algorithm. Incremental model updates: The system incrementally updates the user preference model every day at midnight, using user behavior data collected the previous day to fine-tune the model without retraining the entire model. The incremental learning process is completed locally at the edge gateway, protecting user privacy. Preference prediction and active control: When there is no manual user intervention, the system uses a pre-trained user preference model to predict the lighting parameters that the user is most likely to need based on the current context data, and automatically executes the corresponding control actions.

[0030] It is particularly important to emphasize that the preprocessing of the four main input variables in step S2 includes: Raw data filtering: The collected raw illuminance data and raw personnel count data are subjected to noise suppression using a moving average filtering algorithm. The size of the filtering window can be configured according to the needs of the scene. For abnormal data that exceeds the physical range, the 3σ criterion is used for identification and removal. Time synchronization and calibration: All edge gateways and terminal nodes are synchronized with the cloud time server via the NTP network time protocol. For offline scenarios where the network cannot be connected, a high-precision real-time clock module is used for local timing and automatically calibrated after the network is restored. Standardization and normalization transformation: A linear normalization transformation is performed on the filtered raw data to map it to the [0,1] interval; for time data, the sunrise and sunset times of the day and place are calculated using astronomical algorithms based on the latitude and longitude coordinates of the edge gateway, and used as the reference value for normalization.

[0031] It is particularly important to emphasize that steps S3 and S4 also include a full-process sensor fault degradation handling step, specifically: In public mode, when any sensor malfunction is detected, the input data of that sensor is automatically blocked, and the data of the remaining normal sensors is used for control; when two or more sensor malfunctions are detected, it automatically degrades to pure time control mode, and only performs on / off control based on the standardized time variable; when all sensor malfunctions are detected, it is forced to enter the nighttime constant-on mode. In personalized mode, when any sensor malfunction is detected, the input data of that sensor is automatically blocked, and predictive control is performed using the user preference model; when two or more sensor malfunctions are detected, it automatically degrades to a pure user manual control mode, and control is only executed when the user manually operates it. In steps S3 and S4, the processing parameters for each dimension are dynamically generated through the corresponding vector mapping function. The vector mapping function is a monotonic function defined on the interval [0,1]. Its input is the value of the corresponding control vector, and its output is the set of processing parameters for that dimension. Different scenarios can achieve differentiated control by configuring different vector mapping functions without modifying the core processing logic.

[0032] In this embodiment, a machine learning-based sensor fault detection and self-healing mechanism is employed to achieve early detection of sensor faults and automatic system degradation. The specific implementation is as follows: Sensor anomaly detection model: An isolated forest algorithm is used to construct a sensor anomaly detection model. The model is trained on historical sensor output data to learn the data distribution characteristics when the sensor is operating normally. During model runtime, it continuously detects whether the sensor output data deviates from the normal distribution. Fault classification and categorization: Based on the anomaly detection results, sensor faults are classified into three levels: minor fault, moderate fault, and severe fault. In the case of a minor fault, the system continues to use the sensor data but issues a warning. In the case of a moderate fault, the sensor data is blocked, and the remaining sensor data is used for control. In the case of a severe fault, the system is degraded. Fault self-healing process: When a sensor fault is detected, the system automatically executes the corresponding degradation strategy to ensure that the basic lighting function is not affected; at the same time, the system will attempt to self-heal by restarting the sensor and recalibrating; if self-healing fails, the fault information will be reported to the cloud platform to notify the operation and maintenance personnel to handle it. Fault data mining and analysis: The cloud platform summarizes and analyzes all sensor fault data, identifies sensor types and application scenarios with high failure rates, and provides data support for subsequent product improvement and operation and maintenance optimization; The processing parameters for each dimension are dynamically generated through a vector mapping function, which is a monotonic function defined on the interval [0,1], and its general form is: ; Where V is the numerical value of the control vector, P is the processing parameter of the corresponding dimension, and a and b are configurable function coefficients; different control can be achieved by adjusting the values ​​of a and b in different scenarios without modifying the core processing logic.

[0033] It is important to emphasize that the system includes: The initialization and data processing module is used to load pre-configured control vectors, collect and standardize multi-source sensing data, and generate standardized input variables and sensor state sets. The pattern splitting and rule loading module is used to perform binary pattern judgment based on the commonity vector, load the basic rule set of the corresponding pattern, and initialize the basic control parameters. The common pattern processing module is used to perform constraint-priority multi-dimensional data processing and generate the final decision vector of the common pattern. The personalized pattern processing module is used to perform user-driven multi-level and multi-dimensional data processing and generate the final decision vector of the personalized pattern. The instruction output and feedback module is used to convert the final decision vector into standardized control instructions and send them to the lighting actuator, and to collect execution feedback data to update the system status.

[0034] Finally, it should be noted that the above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for intelligent lighting control data processing, characterized in that, include: S1. System initialization and multi-source sensing data standardization processing: Load the five pre-configured control vectors, and collect data on time, illumination, personnel and sensor working status; S2. Binary Mode Rigid Diversion and Basic Rule Set Generation: Perform mode judgment based on the value of the commonity vector and the preset diversion threshold, load the basic rule set of the corresponding mode based on the mode judgment result, the basic rule set defines the meta-rules of all control behaviors under the mode, and initialize the basic control parameter set of the mode based on the basic rule set; S3. When it is determined to be a public mode, perform constraint-priority multi-dimensional data processing: generate a basic decision vector based on the lighting demand determination rules specific to the public mode, and sequentially perform comfort dimension constraint processing, energy-saving dimension constraint optimization, and collaborative dimension constraint enhancement to generate the final decision vector of the public mode; S4. When it is determined to be a personalized mode, perform user-led multi-level data processing: generate a basic decision vector based on the lighting demand determination rules specific to the personalized mode, and sequentially perform user integration processing of the comfort dimension, adaptive optimization of the energy-saving dimension, and scenario-based enhancement of the collaboration dimension to generate the final decision vector of the personalized mode; S5. Standardized control command generation and closed-loop feedback processing.

2. The intelligent lighting control data processing method according to claim 1, characterized in that, In step S1, raw time data, raw illumination data, raw personnel data, and sensor operating status signals are collected. Normalization transformation is performed on all raw sensing data to generate four input variables whose values ​​are limited to the interval [0,1]. In step S1, the five control vectors include a commonality vector, a personalization vector, a comfort vector, an energy-saving vector, and a synergy vector, wherein: The publicness vector is used to characterize the degree of publicness of the service objects in the scenario and their requirements for unified service standards; The personalized vector is used to characterize the specificity of the scene service object and its requirements for personalized customization; The comfort vector is used to characterize the requirements for lighting control precision, response speed, and smoothness; The energy efficiency vector is used to characterize the energy efficiency requirements without breaking the basic constraints. The coordination vector is used to characterize the requirements for coordinated control of lighting in adjacent areas.

3. The intelligent lighting control data processing method according to claim 1, characterized in that, In step S2, the rigid splitting of the binary pattern and the generation of the basic rule set specifically involve: When the value of the commonality vector is greater than the preset diversion threshold, it is determined to be a common mode, and the common mode basic rule set is loaded. The common mode basic rule set includes security constraints, minimum service standards, control priority order and sensor fault degradation strategy. When the value of the commonality vector is less than or equal to the preset diversion threshold, it is determined to be a personalized mode, and the basic rule set of the personalized mode is loaded. The basic rule set of the personalized mode includes user permission range, intervention priority, adaptive learning rules and sensor fault degradation strategy. The preset diversion threshold is a configurable parameter with a value in the range (0,1).

4. The intelligent lighting control data processing method according to claim 1, characterized in that, In step S3, the common pattern constraint priority type multi-level dimension data processing specifically includes: S31. Execute public mode exclusive lighting demand determination: When the value of the standardized illumination variable is less than or equal to the public mode illumination threshold, or the value of the standardized time variable is greater than or equal to the public mode time threshold, it is determined that there is a lighting demand; when there is a lighting demand, when the value of the standardized personnel variable is greater than 0, or the value of the standardized time variable is greater than or equal to the public mode time threshold, it is determined that the lights are on, and a basic decision vector containing the on / off status flag, the minimum allowable brightness value, and the maximum allowable brightness value is generated; S32. Perform comfort dimension constraint processing: Generate a set of comfort processing parameters including control resolution parameters, adjustment granularity parameters, and smoothness parameters based on the values ​​of the comfort vector, transform the basic decision vector, apply the safety constraints defined by the basic rule set to verify the transformation result, and correct it to the constraint boundary if it exceeds the constraint range. S33. Perform energy-saving dimension constraint optimization: Generate an energy-saving optimization parameter set containing idle delay parameters, intensity attenuation parameters, and time period weight parameters based on the energy-saving vector values, optimize the comfort processing results, and verify the optimization results by applying the minimum service standard defined by the basic rule set; S34. Enhance the coordination dimension constraint: Generate a set of coordination control parameters including coordination range parameters, pre-trigger lead time parameters, and data interaction frequency parameters based on the values ​​of the coordination vector. Exchange state data and intermediate decision vectors with adjacent nodes within the coordination range to enhance the energy-saving optimization results. Apply the consistency constraints defined by the basic rule set to verify the enhanced results.

5. The intelligent lighting control data processing method according to claim 1, characterized in that, In step S4, the personalized mode user-driven multi-dimensional data processing specifically includes: S41. Execute personalized mode exclusive lighting demand determination: When a user actively turns on the light, or the value of the standardized illumination variable is less than or equal to the personalized mode illumination threshold and the value of the standardized personnel variable is greater than 0, or the value of the standardized time variable is greater than or equal to the personalized mode time threshold and the value of the standardized personnel variable is greater than 0, it is determined that there is a lighting demand and a basic decision vector for the light-on state is generated. S42. Perform user intervention detection and weight calculation: Real-time detection of user manual intervention events, calculation of user intervention weight based on the value of the personalized vector, the user intervention weight is positively correlated with the value of the personalized vector; when user intervention is detected, a user intervention vector containing brightness adjustment value, color temperature adjustment value, and scene mode value is generated. S43. Perform user fusion processing for comfort dimension: Generate a set of comfort processing parameters based on the values ​​of the comfort vector, and generate an intermediate decision vector for automatic control comfort; when there is effective user intervention, the user intervention vector and the intermediate decision vector for automatic control comfort are weighted and fused according to the user intervention weight; S44. Perform adaptive optimization of energy efficiency dimension: Generate a set of energy efficiency optimization parameters based on the value of the energy efficiency vector, optimize the comfort processing results, record user intervention data and corresponding contextual environment data, and update the user preference model; S45. Enhance the collaborative dimension in a scenario-based manner: Generate a set of collaborative control parameters based on the values ​​of the collaborative vector. When the user selects a preset scenario mode, load the corresponding scenario's collaborative control rules to enhance the energy-saving optimization results.

6. The intelligent lighting control data processing method according to claim 1, characterized in that, In step S2, the specific preprocessing process for the four main input variables includes: Raw data filtering: The collected raw illuminance data and raw personnel count data are subjected to noise suppression using a moving average filtering algorithm. The size of the filtering window can be configured according to the needs of the scene. For abnormal data that exceeds the physical range, the 3σ criterion is used for identification and removal. Time synchronization and calibration: All edge gateways and terminal nodes are synchronized with the cloud time server via the NTP network time protocol. For offline scenarios where the network cannot be connected, a high-precision real-time clock module is used for local timing and automatically calibrated after the network is restored. Standardization and normalization transformation: A linear normalization transformation is performed on the filtered raw data to map it to the [0,1] interval; for time data, the sunrise and sunset times of the day and place are calculated using astronomical algorithms based on the latitude and longitude coordinates of the edge gateway, and used as the reference value for normalization.

7. The intelligent lighting control data processing method according to claim 1, characterized in that, Steps S3 and S4 also include a full-process sensor fault degradation processing step, specifically: In public mode, when any sensor malfunction is detected, the input data of that sensor is automatically blocked, and the data of the remaining normal sensors is used for control; when two or more sensor malfunctions are detected, it automatically degenerates into pure time control mode, and only performs on / off control based on the standardized time variable. When all sensor malfunctions are detected, force the system to enter constant nighttime illumination mode. In personalized mode, when any sensor malfunction is detected, the input data of that sensor is automatically blocked, and predictive control is performed using the user preference model; when two or more sensor malfunctions are detected, it automatically degrades to a pure user manual control mode, and control is only executed when the user manually operates it.

8. The intelligent lighting control data processing method according to claim 1, characterized in that, In steps S3 and S4, the processing parameters for each dimension are dynamically generated through the corresponding vector mapping function. The vector mapping function is a monotonic function defined in the interval [0,1]. Its input is the value of the corresponding control vector, and its output is the set of processing parameters for that dimension. Different scenarios can achieve differentiated control by configuring different vector mapping functions without modifying the core processing logic.

9. The intelligent lighting control data processing method according to claim 1, characterized in that, The specific steps of step S5 are as follows: The elements of the final decision vector are mapped to the numerical range supported by the lighting actuator, encoded into standardized control commands according to the preset communication protocol, and sent to the corresponding lighting actuator. The operating status data and actual output parameters of the actuator are collected, and a feedback data vector is generated to update the system status parameters for the next control cycle.

10. An intelligent lighting control data processing system, used to implement the intelligent lighting control data processing method according to any one of claims 1-9, characterized in that, include: The initialization and data processing module is used to load pre-configured control vectors, collect and standardize multi-source sensing data, and generate standardized input variables and sensor state sets. The pattern splitting and rule loading module is used to perform binary pattern judgment based on the commonity vector, load the basic rule set of the corresponding pattern, and initialize the basic control parameters. The common pattern processing module is used to perform constraint-priority multi-dimensional data processing and generate the final decision vector of the common pattern. The personalized pattern processing module is used to perform user-driven multi-level and multi-dimensional data processing and generate the final decision vector of the personalized pattern. The instruction output and feedback module is used to convert the final decision vector into standardized control instructions and send them to the lighting actuator, and to collect execution feedback data to update the system status.

Images