Multi-source heterogeneous data real-time monitoring method for intelligent lighting network

Abstract

The invention specifically relates to the technical field of big data analysis, and discloses a multi-source heterogeneous data real-time monitoring method for an intelligent lighting network, which comprises the following steps: S1, collecting multi-source heterogeneous data: constructing a network lighting multi-source data set; s2, preprocessing the multi-source data; s3, edge aggregation and feature extraction: constructing a network illumination multi-source data feature set; s4, multi-dimensional data analysis: calculating to obtain a comprehensive illumination evaluation index of the target intelligent illumination network, and obtaining illumination demand information; s5, generating a real-time illumination strategy: formulating the real-time illumination strategy according to the illumination demand information and the illumination state; s6, feedback control and optimization: optimizing model parameters; through the cloud server, a comprehensive illumination evaluation index is obtained, an illumination state grade is judged, illumination demand information and the illumination state grade are matched, a real-time illumination strategy is generated, lamp parameters are adjusted, and model parameters are optimized in combination with historical data, so that the data analysis precision is improved, the real-time response capability is enhanced, and the energy waste is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of big data analysis, and more specifically, to a method for real-time monitoring of multi-source heterogeneous data for an intelligent lighting network. Background Art

[0002] As the global energy crisis intensifies, the lighting system, as an important area of ​​energy consumption, is the key to energy conservation and emission reduction. With the advancement of smart city construction, the lighting system has become an important part of urban infrastructure and needs to be managed intelligently. The popularization of Internet of Things technology has connected lighting equipment to the network, generating massive multi-source heterogeneous data. The data sources of smart lighting networks are diverse and the formats are heterogeneous. Traditional centralized data processing methods are difficult to meet real-time requirements. Existing real-time monitoring methods for smart lighting network data include real-time data collection, stable data transmission, data processing and analysis, real-time monitoring and control, as well as feedback and optimization, which realize efficient processing and intelligent regulation of lighting network data and improve the real-time and reliability of lighting network data monitoring.

[0003] However, in actual use, it still has some shortcomings. First, the data collection scope is limited. The existing real-time monitoring methods for intelligent lighting network data collection are relatively limited. They usually rely on a few fixed types of sensors and do not accurately analyze the lamps in the lighting network, which reduces the accuracy of data analysis. Second, the dynamic adjustment capability is insufficient. Existing real-time monitoring methods for intelligent lighting network data mostly use preset time or threshold rules for control. This makes it difficult to make dynamic adjustments based on real-time changes in the environment and user behavior. When a fault occurs, the lighting brightness and area cannot be automatically adjusted in time, which cannot meet actual lighting needs. Third, the depth of intelligent analysis is insufficient. The existing real-time monitoring method of intelligent lighting network data only stays at the level of data collection and simple statistics. It lacks in-depth mining of time series data and spatial data, and is adjusted based on fixed time. It does not consider dynamic factors and controls single lamps. It lacks analysis of all lamps in the area, resulting in serious energy waste. Summary of the Invention

[0004] In view of this, an embodiment of the present invention provides a real-time monitoring method for multi-source heterogeneous data in an intelligent lighting network. The method integrates sensor terminals, intelligent lamp terminals and user interaction terminals through a data acquisition terminal to collect multi-source heterogeneous data in the intelligent lighting network in real time. The edge computing gateway extracts key lighting features to construct a network lighting multi-source data feature set. The cloud server generates a real-time lighting strategy, and the visual control terminal controls the lamps for adjustment. This effectively solves the problems of limited data acquisition scope, insufficient dynamic adjustment capability and insufficient intelligent analysis depth proposed in the background technology.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a method for real-time monitoring of multi-source heterogeneous data for an intelligent lighting network, comprising a data acquisition terminal, an edge computing gateway, a cloud server, and a visual control terminal, forming a three-layer connection architecture, and the steps are as follows: S1: Multi-source heterogeneous data acquisition: The data acquisition terminal configures the sampling frequency according to the scene requirements, acquires multi-source heterogeneous data of the lighting network in real time, builds a multi-source data set of network lighting, and transmits it to the edge computing gateway; S2: Multi-source data preprocessing: The edge computing gateway performs noise filtering and data cleaning on the multi-source heterogeneous data in the network lighting multi-source dataset, removes abnormal data in the multi-source heterogeneous data, and performs data standardization; S3: Edge aggregation and feature extraction: The edge computing gateway aggregates and extracts features from pre-processed multi-source heterogeneous data, constructs a feature set of multi-source data for network lighting, transmits it to the cloud server, and performs preliminary edge diagnosis. S4: Multi-dimensional data analysis: A comprehensive lighting evaluation model is constructed through a cloud server to calculate the comprehensive lighting evaluation index of the target intelligent lighting network and obtain lighting demand information based on the network lighting multi-source data feature set; S5: Real-time lighting strategy generation: The lighting status is determined based on the comprehensive lighting evaluation index, and a real-time lighting strategy is formulated based on the lighting demand information and lighting status. The lighting control instructions are then generated to adjust the parameters of the smart lamps. S6: Feedback control and optimization: Real-time collection of adjusted multi-source heterogeneous data, calculation of the comprehensive lighting evaluation index, and the adjusted comprehensive lighting status are obtained, and the model parameters are optimized in combination with historical lighting data.

[0006] The technical effects and advantages of the present invention are as follows: 1. This invention uses a data acquisition terminal to integrate sensor terminals, smart lighting terminals, and user interaction terminals to collect multi-source heterogeneous data in the smart lighting network in real time. The user interaction terminal collects control instructions manually input by users to avoid misjudgments caused by single data. The edge computing gateway performs data cleaning, aggregation, and feature extraction, constructing a network lighting multi-source data feature set covering three dimensions: time domain, frequency domain, and statistics, thereby improving the accuracy of data analysis. 2. The present invention uses a cloud server to analyze lighting demand information based on user operations, personnel location, and natural time factors, and calculates a comprehensive lighting evaluation index based on the constructed comprehensive lighting evaluation model. The lighting status level is determined based on the comprehensive lighting evaluation index. The lighting demand information is matched with the lighting status level to generate a real-time lighting strategy. The lighting parameters are dynamically adjusted by real-time monitoring of the adjusted comprehensive lighting evaluation index to enhance the real-time response capability. 3. The present invention aggregates multi-source heterogeneous data in time dimension and integrates them in space dimension through edge computing gateway, extracts key lighting features, controls lamp switches, lamp brightness and scene modes according to lighting demand information, dynamically adjusts the lighting status level of the entire area, adds feedback data to the training set, and uses deep learning algorithms to retrain the comprehensive lighting evaluation model, optimize model parameters, and effectively reduce energy waste. BRIEF DESCRIPTION OF THE DRAWINGS

[0007] Figure 1 Schematic diagram of the method steps of the present invention.

[0008] Figure 2 It is a schematic diagram of the overall structure of the present invention.

[0009] Figure 3 Schematic diagram of the steps for constructing the network lighting multi-source data feature set in the present invention. DETAILED DESCRIPTION The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0010] As attached Figure 1 A real-time monitoring method for multi-source heterogeneous data in an intelligent lighting network is shown, which includes a data acquisition terminal, an edge computing gateway, a cloud server, and a visualization control terminal, forming a three-layer connection architecture.

[0011] In a more specific application of the present invention, it should be specifically explained that the data acquisition terminal is used to collect multi-source heterogeneous data of the smart lighting network in real time, including sensor terminals, smart lamp terminals and user interaction terminals, wherein the sensor terminals include environmental sensors, energy consumption sensors and position and motion sensors, and the smart lamp terminals refer to LED lamps with integrated sensors, which can provide real-time feedback on their own working status, and the user interaction terminals include mobile devices and smart panels, wherein users manually input scene requirements through mobile devices, and the smart panels are used to collect local control instructions, and the data acquisition terminal is connected to the edge computing gateway through an industrial protocol.

[0012] The edge computing gateway is used to realize data aggregation at the edge of the network, uniformly collect sensor and lamp data of different protocols and formats, perform data preprocessing, complete protocol conversion, and set up local preliminary fault diagnosis to reduce cloud computing pressure. The edge computing gateway integrates multi-protocol conversion modules and edge computing chips, and is connected to the cloud server through the 5G network.

[0013] The cloud server is used to store a large amount of multi-source heterogeneous data from the smart lighting network, run data analysis algorithms, mine data value, and generate smart lighting strategies. The cloud server is composed of a high-performance server cluster, including a cloud database, a big data analysis engine, and a machine learning and deep learning algorithm library. The cloud server is connected to the visual control terminal via HTTP / HTTPS.

[0014] The visual control terminal is used to provide users with an intuitive interactive interface and administrators with a management background to view the status of lighting network equipment, energy consumption data and fault information in real time, and to display the real-time operating status of lighting equipment in the form of dynamic charts and maps. The visual terminal consists of a management background and a large visual screen.

[0015] For the connection methods of the above data acquisition terminal, edge computing gateway, cloud server and visual control terminal, see Figure 2 .

[0016] The specific embodiment of the present invention comprises the following steps: S1: Multi-source heterogeneous data acquisition: The data acquisition terminal configures the sampling frequency according to the scene requirements, acquires multi-source heterogeneous data of the lighting network in real time, builds a multi-source data set of network lighting, and transmits it to the edge computing gateway; Furthermore, the configuration of sampling frequency requires dividing the lighting scenes according to the scene real-time performance, data change rate, and energy consumption cost, and configuring the corresponding sampling frequency according to the lighting scene; In this embodiment, it should be specifically noted that lighting scenes are divided into high real-time scenes and low real-time scenes according to the real-time requirements of the scenes. High real-time scenes, such as conference rooms and corridors, require rapid response to personnel activities, and the sampling frequency is greater than 10Hz. Low real-time scenes, such as underground parking lots at night, have a sampling frequency less than 1Hz; lighting scenes are divided into dynamically changing scenes and static scenes according to the data change rate. Dynamically changing scenes, such as traffic intersections where lighting is affected by vehicles, have a sampling frequency greater than 50Hz, and static scenes, such as fixed lighting in warehouses, have a sampling frequency of 1Hz; lighting scenes are divided into battery-powered devices and wired-powered devices according to energy consumption cost requirements. Battery-powered devices, such as outdoor street light environmental monitoring, have a sampling frequency less than 0.1Hz, and wired-powered devices, such as indoor smart lamps, support high-frequency sampling of more than 100Hz.

[0017] The sensor terminal in the data acquisition terminal collects intelligent perception data in real time, the intelligent lighting terminal collects equipment status data in real time, and the user interaction terminal collects human-computer interaction data in real time; Intelligent perception data includes environmental perception data, human behavior data, environmental status data and equipment operation data; equipment status data includes lighting status data, operation parameter data and equipment energy consumption data; human-computer interaction data includes control command data and location identity data.

[0018] In this embodiment, it should be specifically explained that the environmental perception data specifically includes light intensity, color temperature and ultraviolet intensity, the human behavior data specifically includes human presence, movement trajectory and gestures, the environmental status data specifically includes temperature, humidity, carbon dioxide concentration and PM2.5, the equipment operation data specifically includes lamp voltage, operating temperature and fault code; the lamp status data specifically includes power, brightness, color temperature and switch status, the operating parameter data specifically includes cumulative working hours, light decay rate and LED junction temperature, the equipment energy consumption data specifically includes real-time power consumption and historical energy consumption curves; the control instruction data specifically includes brightness adjustment instructions, color temperature setting instructions and scene mode selection instructions, and the location identity data specifically includes user ID and current location information.

[0019] S2: Multi-source data preprocessing: The edge computing gateway performs noise filtering and data cleaning on the multi-source heterogeneous data in the network lighting multi-source dataset, removes abnormal data in the multi-source heterogeneous data, and performs data standardization; Furthermore, noise filtering requires selecting appropriate filtering algorithms based on the noise type and building a real-time filtering mechanism on the edge computing gateway; In this embodiment, it should be specifically explained that the noise existing in the network lighting multi-source dataset includes environmental noise, equipment noise and communication noise. The filtering algorithms include sliding average filtering, median filtering, Kalman filtering and wavelet transform filtering. The edge computing gateway uses sliding average filtering to preliminarily smooth the high-frequency noise in the data in the network lighting multi-source dataset transmitted by the data acquisition terminal, and uses the median filtering algorithm for key indicators to eliminate sudden outliers, and establishes a normal threshold range in combination with historical data, and marks data exceeding the normal threshold range as suspicious noise.

[0020] The data after noise filtering is processed for missing values, outliers are corrected, and the format is unified to build a cloud-edge collaborative cleaning mechanism.

[0021] In this embodiment, it should be specifically explained that format unification is to unify the heterogeneous data output by sensors with different protocols into the same format. The cloud-edge collaborative cleaning mechanism refers to lightweight cleaning at the edge computing gateway, real-time filtering of high-frequency noise and processing of simple missing values, reducing the amount of data uploaded to the cloud server, and performing deep cleaning on the cloud server, batch analysis of historical network lighting data, and identification of complex abnormal data that is difficult for the edge computing gateway to identify.

[0022] S3: Edge aggregation and feature extraction: The edge computing gateway aggregates and extracts features from pre-processed multi-source heterogeneous data, constructs a feature set of multi-source data for network lighting, transmits it to the cloud server, and performs preliminary edge diagnosis. Further, such as Figure 3 As shown in Figure 2, the steps for constructing the network lighting multi-source data feature set are as follows: A1: The edge computing gateway aggregates pre-processed multi-source heterogeneous data, including time dimension aggregation, spatial dimension aggregation, and feature dimension aggregation. A2: Extract the time domain features, frequency domain features, statistical features, and state features of the aggregated data to obtain the key lighting features in the network lighting multi-source data feature set; In this embodiment, it should be specifically explained that the key lighting characteristics include the physical characteristics of the lighting environment, the operating status characteristics of the lamps and the user interaction demand characteristics. The physical characteristics of the lighting environment specifically include light intensity, color temperature, color rendering index and light distribution uniformity. The operating status characteristics of the lamps specifically include active power, power factor, temperature, continuous working time, brightness adjustment percentage and color parameters. The user interaction demand characteristics specifically include user operation instructions, personnel location and natural time.

[0023] A3: Design a feature set architecture to structure the storage of key lighting features and dynamically update the network lighting multi-source data feature set based on real-time monitoring of multi-source heterogeneous data.

[0024] In this embodiment, it should be specifically explained that the feature set architecture divides the key lighting features according to the type of data acquisition terminal, including a subset of physical features of the lighting environment, a subset of features of the lamp operating status, and a subset of features of user interaction needs. The data is structured and stored according to the time dimension, space dimension, and feature dimension, and the network lighting multi-source data feature set is updated in real time to ensure that the features are synchronized with the physical status.

[0025] Edge preliminary diagnosis refers to comparing the data in the multi-source data feature set of network lighting with the preset threshold to determine whether the lamps in the lighting network are faulty. If a fault occurs, fault information is generated to remind relevant personnel to conduct inspection and maintenance.

[0026] S4: Multi-dimensional data analysis: A comprehensive lighting evaluation model is constructed through a cloud server to calculate the comprehensive lighting evaluation index of the target intelligent lighting network and obtain lighting demand information based on the network lighting multi-source data feature set; Furthermore, the calculation steps of the comprehensive lighting evaluation index are as follows: B1: The cloud server obtains the network lighting multi-source data feature set transmitted by the edge computing gateway, sets a time window, obtains the data in the network lighting multi-source data feature set of the target monitoring area within the time window, and performs normalization processing to obtain the illuminance L, color temperature T, lamp active power P, lamp temperature T1, and lamp continuous operating time t in the interval [0,1]. B2: Using the formula through illumination: , Calculate the illumination evaluation index I L , using the illuminance evaluation index and color temperature formula: , Calculate the lighting environment assessment index I e , where a1 and a2 represent the weight coefficients of illumination and color temperature respectively, and the formula is used based on the active power, temperature and continuous working time of the lamp: , Calculate the lamp status assessment index I d , where k represents the attenuation coefficient of the lamp, t0 represents the rated life of the lamp, b1, b2 and b3 represent the weight coefficients of the lamp active power, lamp temperature and continuous working time respectively; B3: Obtain the lighting environment evaluation index I of the target monitoring area e and the lamp status evaluation index I corresponding to n lamps in the target monitoring area di , using the formula The comprehensive lighting evaluation index I of the target monitoring area is calculated, where ω1 and ω2 represent the weight coefficients of the lighting environment evaluation index and the lighting status evaluation index, respectively.

[0027] In this embodiment, it is necessary to specifically explain that the data in the network lighting multi-source data feature set within the target monitoring area are standardized, linear normalization is used for features with clear upper and lower limits, and z-score is used for standardization for features without fixed boundaries; high illumination in the lighting environment can easily cause waste of resources, low illumination can affect human vision, too high or too low color temperature can affect human rhythm, and too high temperature in the lamp state can accelerate lamp aging. The above weights should be set based on historical data and expert experience, and the time window should be set according to the maximum sampling frequency within the target monitoring area.

[0028] Furthermore, the acquisition of lighting demand information requires the cloud server to classify and semantically analyze user operation instructions in the network lighting multi-source data feature set, build a personnel location-demand mapping model, and combine it with natural time to obtain lighting demand information.

[0029] In this embodiment, it should be specifically explained that user operation instructions specifically include brightness adjustment instructions, color temperature adjustment instructions, scene mode instructions and timing control instructions. Natural language processing technology is used to extract keywords in user operation instructions, map them to the lighting parameter space, establish an instruction-demand comparison table, set the priority of real-time instructions to be higher than the preset instructions, and obtain lighting demand information based on personnel location and natural time, providing a data basis for the subsequent generation of real-time lighting strategies.

[0030] S5: Real-time lighting strategy generation: The lighting status is determined based on the comprehensive lighting evaluation index, and a real-time lighting strategy is formulated based on the lighting demand information and lighting status. The lighting control instructions are then generated to adjust the parameters of the smart lamps. Furthermore, the judgment of the lighting status requires the construction of a comprehensive lighting status evaluation standard, dividing the lighting status levels, including the first lighting status level, the second lighting status level, the third lighting status level and the fourth lighting status level, matching the comprehensive lighting evaluation index in the target monitoring area with the lighting status evaluation standard to obtain the lighting status level.

[0031] In this embodiment, it should be specifically explained that when the comprehensive lighting evaluation index is greater than 0.8 and less than 1, the target monitoring area is in the first lighting status level, indicating that the lighting status is excellent and the lighting quality is excellent; when the comprehensive lighting evaluation index is greater than 0.6 and less than 0.8, the target monitoring area is in the second lighting status level, indicating that the lighting status is good, the lighting quality meets the basic requirements, and needs to be optimized; when the comprehensive lighting evaluation index is greater than 0.4 and less than 0.6, the target monitoring area is in the third lighting status level, indicating that the lighting status is general, there are local problems with the lighting quality, and improvement is needed; when the comprehensive lighting evaluation index is less than 0.4, the target monitoring area is in the fourth lighting status level, indicating that the lighting status is poor and requires urgent adjustment.

[0032] Furthermore, the generation of real-time lighting strategies requires mapping the matching relationship between the lighting demand information and the lighting status level through a rule engine according to the lighting demand information and the lighting status level, and generating lighting control instructions.

[0033] In this embodiment, it should be specifically explained that mapping the matching relationship between lighting demand and lighting status level requires the design of matching relationship principles, including the user operation instruction priority principle, the personnel position and scene correlation principle, and the time and natural light linkage principle. The user operation instructions in the lighting demand information are quantified into switching lights, brightness adjustment and scene mode, the personnel position is spatialized, and the time period code and date attributes of natural time are obtained, including weekdays, weekends and holidays, and the lighting demand score is quantified. The user demand score is mapped and matched with the lighting status level, and the lighting control instruction is generated according to the lighting demand to control the lamps through the visual control terminal to adjust the lighting status level to the corresponding state.

[0034] S6: Feedback control and optimization: Real-time collection of adjusted multi-source heterogeneous data, calculation of the comprehensive lighting evaluation index, and the adjusted comprehensive lighting status are obtained, and the model parameters are optimized in combination with historical lighting data.

[0035] Furthermore, the optimization of model parameters requires real-time monitoring of multi-source heterogeneous data, real-time processing by edge computing gateways and cloud processors to obtain real-time lighting demand information and adjusted lighting status levels, recording the matching information of historical lighting demand information and lighting status levels, and optimizing the weight coefficients in the comprehensive lighting evaluation model through machine learning.

[0036] In this embodiment, it is necessary to specifically explain that the multi-source heterogeneous data, lighting demand information and lighting status level monitored in real time are displayed in real time through the visual control terminal, and a query function is provided to obtain historical data. Multi-source heterogeneous data are monitored in real time, for example, the data information in the multi-source heterogeneous data is updated every 5 minutes, the lighting demand score and lighting status level are recalculated, the lighting control instructions are corrected, and the feedback data is added to the training set, the comprehensive lighting evaluation model is retrained, and the optimized weight coefficient is obtained to improve the accuracy of the lighting status judgment.

[0037] Secondly: The drawings of the embodiments disclosed in the present invention only involve structures related to the embodiments disclosed in the present invention. Other structures may refer to conventional designs. The same embodiment and different embodiments of the present invention may be combined with each other without conflict. Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A real-time monitoring method for multi-source heterogeneous data in intelligent lighting networks, characterized in that: It includes data acquisition terminal, edge computing gateway, cloud server and visual control terminal, forming a three-layer connection architecture. The steps are as follows: S1: Multi-source heterogeneous data acquisition: The data acquisition terminal configures the sampling frequency according to the scene requirements, acquires multi-source heterogeneous data of the lighting network in real time, builds a multi-source data set of network lighting, and transmits it to the edge computing gateway; S2: Multi-source data preprocessing: The edge computing gateway performs noise filtering and data cleaning on the multi-source heterogeneous data in the network lighting multi-source dataset, removes abnormal data in the multi-source heterogeneous data, and performs data standardization; S3: Edge aggregation and feature extraction: The edge computing gateway aggregates and extracts features from pre-processed multi-source heterogeneous data, constructs a feature set of multi-source data for network lighting, transmits it to the cloud server, and performs preliminary edge diagnosis. S4: Multi-dimensional data analysis: A comprehensive lighting evaluation model is constructed through a cloud server to calculate the comprehensive lighting evaluation index of the target intelligent lighting network and obtain lighting demand information based on the network lighting multi-source data feature set; S5: Real-time lighting strategy generation: The lighting status is determined based on the comprehensive lighting evaluation index, and a real-time lighting strategy is formulated based on the lighting demand information and lighting status. The lighting control instructions are then generated to adjust the parameters of the smart lamps. S6: Feedback control and optimization: Real-time collection of adjusted multi-source heterogeneous data, calculation of the comprehensive lighting evaluation index, and the adjusted comprehensive lighting status are obtained, and the model parameters are optimized in combination with historical lighting data.

2. The method for real-time monitoring of multi-source heterogeneous data for intelligent lighting networks according to claim 1, characterized in that: The configuration of the sampling frequency requires dividing the lighting scenes according to the scene real-time performance, data change rate and energy consumption cost, and configuring the corresponding sampling frequency according to the lighting scene; The sensor terminal in the data acquisition terminal collects intelligent perception data in real time, the intelligent lighting terminal collects equipment status data in real time, and the user interaction terminal collects human-computer interaction data in real time; Intelligent perception data includes environmental perception data, human behavior data, environmental status data and equipment operation data; equipment status data includes lighting status data, operation parameter data and equipment energy consumption data; human-computer interaction data includes control command data and location identity data.

3. The method for real-time monitoring of multi-source heterogeneous data for an intelligent lighting network according to claim 1, characterized in that: The noise filtering requires selecting an appropriate filtering algorithm according to the noise type and building a real-time filtering mechanism on the edge computing gateway; The data after noise filtering is processed for missing values, outliers are corrected, and the format is unified to build a cloud-edge collaborative cleaning mechanism.

4. The method for real-time monitoring of multi-source heterogeneous data for an intelligent lighting network according to claim 1, characterized in that: The steps for constructing the network lighting multi-source data feature set are as follows: A1: The edge computing gateway aggregates pre-processed multi-source heterogeneous data, including time dimension aggregation, spatial dimension aggregation, and feature dimension aggregation. A2: Extract the time domain features, frequency domain features, statistical features, and state features of the aggregated data to obtain the key lighting features in the network lighting multi-source data feature set; A3: Design a feature set architecture to structure the storage of key lighting features and dynamically update the network lighting multi-source data feature set based on real-time monitoring of multi-source heterogeneous data.

5. The method for real-time monitoring of multi-source heterogeneous data for intelligent lighting networks according to claim 1, characterized in that: The calculation steps of the comprehensive lighting evaluation index are as follows: B1: The cloud server obtains the network lighting multi-source data feature set transmitted by the edge computing gateway, sets a time window, obtains the data in the network lighting multi-source data feature set of the target monitoring area within the time window, and performs normalization processing to obtain the illuminance L, color temperature T, lamp active power P, lamp temperature T1, and lamp continuous operating time t in the interval [0,1]. B2: Using the formula through illumination: ， Calculate the illumination evaluation index I L , using the illuminance evaluation index and color temperature formula: ， Calculate the lighting environment assessment index I e , where a1 and a2 represent the weight coefficients of illumination and color temperature respectively, and the formula is used based on the active power, temperature and continuous working time of the lamp: ， Calculate the lamp status assessment index I d , where k represents the attenuation coefficient of the lamp, t0 represents the rated life of the lamp, b1, b2 and b3 represent the weight coefficients of the lamp active power, lamp temperature and continuous working time respectively; B3: Obtain the lighting environment evaluation index I of the target monitoring area e and the lamp status evaluation index I corresponding to n lamps in the target monitoring area di , using the formula The comprehensive lighting evaluation index I of the target monitoring area is calculated, where ω1 and ω2 represent the weight coefficients of the lighting environment evaluation index and the lighting status evaluation index, respectively.

6. The method for real-time monitoring of multi-source heterogeneous data for an intelligent lighting network according to claim 1, characterized in that: The acquisition of the lighting demand information requires that the cloud server classify and semantically analyze the user operation instructions in the network lighting multi-source data feature set, build a personnel location-demand mapping model, and obtain the lighting demand information in combination with natural time.

7. The method for real-time monitoring of multi-source heterogeneous data for an intelligent lighting network according to claim 1, characterized in that: The judgment of the lighting state requires the construction of a comprehensive lighting state evaluation standard, dividing the lighting state levels, including the first lighting state level, the second lighting state level, the third lighting state level and the fourth lighting state level, matching the comprehensive lighting evaluation index in the target monitoring area with the lighting state evaluation standard to obtain the lighting state level.

8. The method for real-time monitoring of multi-source heterogeneous data for an intelligent lighting network according to claim 1, characterized in that: The generation of the real-time lighting strategy requires mapping the matching relationship between the lighting demand information and the lighting status level through a rule engine according to the lighting demand information and the lighting status level, and generating a lighting control instruction.

9. The method for real-time monitoring of multi-source heterogeneous data for intelligent lighting networks according to claim 1, characterized in that: The optimization of the model parameters requires real-time monitoring of multi-source heterogeneous data, real-time processing by edge computing gateways and cloud processors to obtain real-time lighting demand information and adjusted lighting status levels, recording the matching information of historical lighting demand information and lighting status levels, and optimizing the weight coefficients in the comprehensive lighting evaluation model through machine learning.

Images