Lighting optimization control method and system based on environment monitoring

Abstract

The invention provides an illumination optimization control method and system based on environmental monitoring, and the method comprises the steps: collecting the illumination intensity data and human body existence state data of a target region; performing space-time alignment on the data, external astronomical data and historical energy consumption data at the cloud to obtain a fused data set; generating an illumination demand prediction curve through a time sequence model based on the fused data set; according to the illumination demand prediction curve, dynamically constructing a collaborative strategy fusing an illumination threshold, human body induction and a shiftable time table; compiling the cooperation strategy into a physical instruction, and driving a lighting loop through edge equipment; the loop state of the lighting loop is monitored in real time, the re-learning process of the lighting demand prediction curve is triggered when abnormality is detected, the lighting demand is predicted through various data, and then the collaborative strategy is generated and lighting control is performed, so that compared with single environmental parameter or cloud analysis, the method can adapt to complex environmental changes, realize accurate lighting control and improve the lighting control efficiency. Resource waste is reduced; and user experience is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of lighting optimization, and in particular to a lighting optimization control method and system based on environmental monitoring. Background Art

[0002] As a core component of building energy consumption, intelligent control of lighting systems is crucial for achieving energy conservation and emission reduction goals. With the development of IoT technology, environmental sensing-based lighting control has become a key research direction in the field of smart buildings. In particular, in industrial plants and large public facilities, achieving precise lighting control through real-time monitoring of environmental parameters such as light intensity and human activity has become a key technical challenge in reducing operating costs and improving energy efficiency.

[0003] The current mainstream lighting control solutions mainly include three categories: (1) time-sharing and timing control: lighting equipment is turned on and off according to a preset schedule, but it cannot adapt to sudden changes in weather or the flow of people; (2) single sensor control: if it relies only on light sensors to achieve automatic switching, it is easy to cause false operations due to environmental interference and cannot sense the presence of people; (3) centralized management on the cloud platform: historical energy consumption data is analyzed through the cloud to generate control strategies, but there is a response delay, which makes it difficult to meet real-time control needs.

[0004] However, timing control cannot respond to sudden changes in the environment, single-sensor solutions ignore the human-light correlation, and centralized cloud computing leads to control lags. This makes it difficult to achieve precise lighting control when responding to complex environmental changes, resulting in energy waste or a decline in user experience. Summary of the Invention

[0005] The present invention provides a lighting optimization control method and system based on environmental monitoring, which are used to solve the defects of the prior art that the lighting control method is single, resulting in waste of resources and reduced user experience.

[0006] In a first aspect, the present invention provides a lighting optimization control method based on environmental monitoring, comprising: Collect light intensity data and human presence data in the target area through a distributed sensor network; Performing spatiotemporal alignment on the cloud for the light intensity data and human presence status data, external astronomical data, and historical energy consumption data to obtain a fused data set; Based on the fused data set, generating a lighting demand prediction curve through a time series model; Based on the lighting demand forecast curve, dynamically build a collaborative strategy that integrates lighting thresholds, human body sensing, and shiftable schedules; Compiling the collaborative strategy into physical instructions to drive the lighting loop via an edge device; The circuit status of the lighting circuit is monitored in real time, and when an abnormality is detected, a re-learning process of the lighting demand prediction curve is triggered.

[0007] According to a lighting optimization control method based on environmental monitoring provided by the present invention, the method collects light intensity data and human presence status data of a target area through a distributed sensor network, including: Eliminate environmental interference through multi-sensor collaboration to obtain light intensity data; Track the movement trajectory of the human body based on the topological relationship of distributed sensors to obtain the human body's presence status data; Dynamic noise suppression is performed on the light intensity data and the human body presence status data.

[0008] According to a lighting optimization control method based on environmental monitoring provided by the present invention, generating a lighting demand prediction curve based on the fused data set through a time series model includes: Perform feature engineering on the fused dataset to extract periodic patterns of light and human activity; Based on the periodic pattern, a multi-channel model is constructed to process the spatiotemporal characteristics; The processed spatiotemporal features are weighted to generate a probabilistic demand distribution and construct a lighting demand prediction curve.

[0009] According to the present invention, a lighting optimization control method based on environmental monitoring further includes: Gridding the target area based on the distributed sensor topological relationship; Count the human activity frequency in each grid; Based on the human activity frequencies within the grids, a spatial heat map is constructed, and a multi-channel model is constructed by combining the periodic patterns of illumination and human activity.

[0010] According to a lighting optimization control method based on environmental monitoring provided by the present invention, the dynamically constructing a shiftable schedule includes: calculating a reference offset based on the external astronomical data; Adjusting the reference offset based on a deviation between an actual ambient light measurement value and a predicted value of the lighting demand prediction curve; When the deviation exceeds a threshold, the reference parameter is updated to generate a feasible schedule.

[0011] According to a lighting optimization control method based on environmental monitoring provided by the present invention, compiling the collaborative strategy into physical instructions and driving the lighting circuit through the edge device includes: caching the collaborative strategy; When the network is disconnected, an emergency instruction is generated based on the cached collaborative strategy and the light intensity data and human presence status data; When the network is restored, the offline log corresponding to the emergency instruction is uploaded and the lighting demand prediction curve is calibrated.

[0012] According to a lighting optimization control method based on environmental monitoring provided by the present invention, the circuit status of the lighting circuit is monitored in real time, and when an abnormality is detected, a relearning process of the lighting demand prediction curve is triggered, including: Analyzing the loop current waveform of the lighting circuit to identify abnormal patterns; Comparing the abnormal pattern with the abnormal pattern library, classifying and determining the fault type; According to the fault type, a response action is executed to trigger a re-learning process of the lighting demand prediction curve.

[0013] According to a lighting optimization control method based on environmental monitoring provided by the present invention, after collecting light intensity data and human presence status data of a target area through a distributed sensor network, the method further includes: When the deviation between a single sensor data and the monitoring value of the adjacent sensor cluster exceeds the dynamic threshold, it is marked as a suspicious node; Calling the historical illumination model and astronomical data of the associated area to verify the physical rationality of the data of the suspicious node; If verification fails, it automatically switches to the backup sensor data stream and triggers an equipment maintenance alarm.

[0014] According to the present invention, a lighting optimization control method based on environmental monitoring further includes: Calculate the actual energy saving rate and malfunction rate after executing the collaborative strategy; Based on the actual energy saving rate and the malfunction occurrence rate, a correlation matrix between strategy parameters and abnormal events is established to locate the characteristics of the failed strategy; The failure characteristics are injected into the model training set of the lighting demand prediction curve to generate a strategy blacklist and optimize the collaborative strategy generation rules.

[0015] In a second aspect, the present invention further provides a lighting optimization control system based on environmental monitoring, comprising: The acquisition module is used to collect light intensity data and human presence status data of the target area through a distributed sensor network; A fusion module is used to align the light intensity data and human presence status data with external astronomical data and historical energy consumption data in the cloud in time and space to obtain a fused data set; A prediction module, configured to generate a lighting demand prediction curve using a time series model based on the fused data set; A construction module for dynamically constructing a collaborative strategy integrating lighting thresholds, human body sensing, and a shiftable schedule based on the lighting demand prediction curve; a driver module, configured to compile the collaborative strategy into physical instructions and drive the lighting circuit via an edge device; The feedback module is used to monitor the circuit status of the lighting circuit in real time and trigger the re-learning process of the lighting demand prediction curve when an abnormality is detected.

[0016] In a third aspect, the present invention also provides an electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the lighting optimization control method based on environmental monitoring as described above is implemented.

[0017] In a fourth aspect, the present invention further provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements any of the above-described lighting optimization control methods based on environmental monitoring.

[0018] In a fifth aspect, the present invention further provides a computer program product, comprising a computer program, which, when executed by a processor, implements any of the above-described lighting optimization control methods based on environmental monitoring.

[0019] The present invention provides a lighting optimization control method and system based on environmental monitoring. The method collects light intensity data and human presence status data of the target area through a distributed sensor network; the light intensity data and human presence status data are spatiotemporally aligned with external astronomical data and historical energy consumption data in the cloud to obtain a fused data set; based on the fused data set, a lighting demand prediction curve is generated through a time series model; according to the lighting demand prediction curve, a collaborative strategy that integrates light thresholds, human body sensing and offset schedules is dynamically constructed; the collaborative strategy is compiled into physical instructions and the lighting circuit is driven by the edge device; the loop status of the lighting circuit is monitored in real time, and when an anomaly is detected, the re-learning process of the lighting demand prediction curve is triggered. Since the lighting demand is predicted by multiple data, a collaborative strategy is generated and lighting control is performed. Compared with a single environmental parameter or cloud analysis, it can adapt to complex environmental changes, achieve precise lighting control, reduce resource waste and improve user experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the technical solutions in the present invention or the prior art, a brief introduction is given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0021] Figure 1 is a flow chart of the lighting optimization control method based on environmental monitoring provided in this embodiment; Figure 2 is a schematic structural diagram of a lighting optimization control system based on environmental monitoring provided in this embodiment; Figure 3 Schematic diagram of the structure of the electronic device provided in this embodiment. DETAILED DESCRIPTION

[0022] To make the objectives, technical solutions, and advantages of the present invention more clear, the technical solutions of the present invention will be clearly and completely described below in conjunction with the accompanying drawings. Obviously, the embodiments described are only some of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts shall fall within the scope of protection of the present invention.

[0023] Figure 1 3 is a flow chart of the lighting optimization control method based on environmental monitoring provided in this embodiment.

[0024] like Figure 1 As shown, the lighting optimization control method based on environmental monitoring provided by the embodiment of the present invention mainly includes the following steps: 101. Collect light intensity data and human presence status data in the target area through a distributed sensor network.

[0025] In a specific implementation, to collect light intensity data and human presence data in a target area through a distributed sensor network, various sensor nodes can be deployed in the target area. For light intensity data collection, high-precision light sensors, such as the BH170 photosensor, are used and evenly distributed throughout the target area. These sensors are connected to ZigBee end nodes with wireless communication capabilities via GPIO ports. The sensors sense the light intensity in real time and convert the analog signal into a digital signal for transmission to the end node.

[0026] To collect human presence data, infrared sensors, such as the HC-SR501 sensor, are installed in the area and connected to the ZigBee end node. These sensors can detect human activity within a certain range and send a high-level signal to the end node upon detecting human movement.

[0027] After collecting light intensity and human presence data, the ZigBee end nodes transmit this data to the coordinator node via the ZigBee network. The coordinator node aggregates the data from each end node and transmits it via serial communication to a host computer, such as a computer, for storage, analysis, and display. Furthermore, the coordinator node can be equipped with an OLED screen for local, real-time display of the collected data.

[0028] This distributed sensor network data collection method accurately captures light intensity and human presence information at different locations within a target area. This enables comprehensive monitoring of lighting conditions across the area, preventing undetected localized lighting anomalies and providing a basis for subsequent precise dimming of lighting equipment. Based on human presence data, intelligent lighting control can be implemented, effectively reducing ineffective lighting energy consumption in unoccupied areas, improving energy efficiency, and providing users with a more comfortable and intelligent lighting environment.

[0029] 102. Align the light intensity data and human presence status data with external astronomical data and historical energy consumption data in the cloud in time and space to obtain a fused data set.

[0030] Specifically, at the data collection end, after the light intensity data and human presence status data are transmitted to the coordinator node through the ZigBee network, the coordinator node uploads the data to the cloud server through the 4G / 5G communication module or Wi-Fi.

[0031] The cloud server also retrieves external astronomical data from the astronomical data platform via an API, including sunrise and sunset times, solar altitude, and cloud cover. This data reflects natural lighting conditions at different times. It also retrieves historical energy consumption data from the historical energy consumption database, including electricity consumption data and on / off times for lighting equipment in the area over the past period.

[0032] In the cloud server, timestamps and geographic coordinates are used to align the collected light intensity data, human presence data, external astronomical data, and historical energy consumption data in time and space. Taking the time dimension as an example, various data types at the same moment are correlated and integrated. In the spatial dimension, the location information of the target area is combined to ensure that the data corresponds to the actual area, ultimately forming a fused dataset.

[0033] By aligning multiple data sets in the cloud, creating a fused dataset through temporal and spatial alignment, we can comprehensively analyze the relationship between natural light, human activity, and energy consumption. For example, by combining astronomical data with historical energy consumption data, we can predict appropriate lighting needs under different weather conditions and time periods. Lighting equipment can be adjusted in advance based on light intensity and human presence data to reduce energy waste. By analyzing the differences between historical and current energy consumption data, we can assess energy-saving effects and provide data support for optimizing lighting control strategies, further enhancing the intelligence and energy efficiency of lighting systems.

[0034] 103. Based on the fused data set, a lighting demand prediction curve is generated through a time series model.

[0035] After acquiring the fused dataset from the cloud server, the data is first preprocessed. Outliers, such as obviously erroneous light intensity data and unusual energy consumption fluctuations, are removed. Missing data is filled using interpolation to ensure data integrity and continuity. The processed data is organized according to time series and divided into training and test sets. The training set is used for model training, while the test set is used to evaluate the model's predictive accuracy.

[0036] Select an appropriate time series model, such as a long short-term memory (LSTM) network or a gated recurrent unit (GRU). Taking LSTM as an example, input the training set data into the LSTM model, set appropriate hyperparameters such as the number of network layers and neurons, and train the model using a backpropagation algorithm. This continuously adjusts the model parameters so that the model can learn the temporal dependencies and patterns in the data and capture the inherent connections between natural light changes, human activity patterns, and lighting needs.

[0037] After training is complete, the model is evaluated using the test set. If the model prediction error is within an acceptable range, the trained model is used to predict lighting demand for a specific period of time. The prediction results are then connected chronologically to form a curve to generate a lighting demand forecast curve. This forecast curve can intuitively demonstrate the changing trends of lighting demand at different times, providing a basis for the development of subsequent lighting control strategies.

[0038] Based on the fused dataset, a lighting demand forecast curve is generated using a time series model. This allows for the prediction of target area lighting needs at different time periods in the future. For example, during periods of daylight with ample activity and low activity, the forecast curve can be used to preemptively reduce lighting brightness or turn off some lamps to avoid over-illumination. Furthermore, before peak activity arrives, lighting equipment can be adjusted to an appropriate brightness to meet lighting needs. This forecast curve provides dynamic and precise control guidance for the lighting system, further improving energy efficiency and reducing operating costs. It also enhances the intelligent management of the lighting system, creating a more comfortable and energy-efficient lighting environment for users.

[0039] 104. Based on the lighting demand forecast curve, dynamically build a collaborative strategy that integrates light thresholds, human body sensing, and offset schedules.

[0040] After obtaining the lighting demand forecast curve, the first step is to determine the lighting threshold. Based on historical data and environmental characteristics, critical light intensity values ​​are set for different time periods. For example, when the indoor light intensity falls below 300 lux during the day, auxiliary lighting is activated, and when it exceeds 800 lux, some lamps are turned off. Furthermore, combined with human sensor data, human activity zones and peak activity times are identified. When the forecast curve indicates that a certain area is about to enter peak activity and the current light intensity falls below the threshold, the lighting brightness in that area is prioritized.

[0041] For shiftable schedules, refer to the low lighting demand period in the forecast curve and arrange maintenance and inspection of lighting equipment in non-critical areas during that period; if the forecast curve shows that there is sufficient natural light and very little human activity during a certain period, the time for shutting down lighting equipment can be appropriately extended.

[0042] An algorithm dynamically integrates light thresholds, human sensor data, and a shiftable schedule. When the forecast curve indicates rising lighting demand in the future, human sensor data is prioritized to determine activity areas, and lighting brightness in those areas is adjusted based on the light threshold. During non-critical periods within the shiftable schedule, lighting energy consumption is appropriately reduced while still meeting basic lighting needs, thus forming a dynamic collaborative strategy.

[0043] The collaborative strategy dynamically constructed based on the lighting demand forecast curve realizes the intelligent linkage of multiple factors. The lighting threshold ensures that the lighting brightness meets the environmental requirements, human body sensing ensures real-time lighting in the personnel activity area, and the schedule can be offset to optimize the equipment operation arrangement. The three work together to avoid wasting lighting resources. For example, when the low peak period at night is predicted and natural light is still available, the system can automatically turn off most lamps and only retain necessary lighting; when people temporarily enter the unlit area, human body sensing is combined to timely fill the light. This strategy significantly improves the flexibility and energy-saving effect of the lighting system, reduces energy consumption and operation and maintenance costs, while ensuring the quality of lighting services.

[0044] 105. Compile the collaborative strategy into physical instructions to drive the lighting loop through edge devices.

[0045] After dynamically constructing a collaborative strategy, a compiler is used to convert it into computer-readable code. For example, policy details such as light intensity threshold adjustment, human sensor response, and device on / off timing can be compiled into control code written in Python or C. This control code is further converted into physical instructions for a specific protocol, such as Modbus protocol instructions, with specific instructions such as "Turn on group X of lights in a certain area" or "Adjust the brightness of a certain light to Y%."

[0046] The compiled physical instructions are transmitted to edge devices, such as smart lighting controllers, via a local area network or wireless network. The controller, which has a built-in microprocessor and communication module, receives the instructions, parses and verifies them, and once verified, transmits the electrical signals to the corresponding lighting circuits via hardware components such as relays and dimming modules. For example, to turn on a lamp circuit, the controller closes the circuit via a relay; to dim a lamp, the controller changes the output voltage or current via a dimming module, thereby driving the lighting devices to operate according to the coordinated strategy.

[0047] Compiling collaborative strategies into physical instructions and driving lighting circuits through edge devices achieves precise implementation from intelligent strategies to actual control. This ensures that lighting devices strictly adhere to the energy-saving and intelligent control requirements of the collaborative strategy, avoiding delays or errors from manual intervention and enabling the lighting system to respond to environmental changes in real time. Furthermore, edge devices process instructions locally, reducing latency in uploading data to the cloud, improving control timeliness, and ensuring the stability and reliability of the lighting system. This reduces network transmission pressure and data security risks, further enhancing the intelligent management level and energy efficiency of the lighting system.

[0048] 106. Monitor the circuit status of the lighting circuit in real time, and trigger the re-learning process of the lighting demand prediction curve when an abnormality is detected.

[0049] Specifically, monitoring devices such as current sensors and voltage sensors are deployed in the lighting circuit to collect current and voltage data in real time. This data is then transmitted to edge devices via wired or wireless communications. The edge devices analyze the collected data in real time and set thresholds for normal current and voltage fluctuations. For example, the normal operating current range is 0.1A-2A, and the voltage range is 200V-240V. If the monitored data exceeds the preset thresholds, or if a sudden current drop or abnormal voltage jump occurs, the lighting circuit is identified as abnormal, such as a lamp failure or a short circuit, and an abnormality alarm is generated.

[0050] The edge device uploads the anomaly alarm information, along with detailed data such as the time and location of the anomaly, to the cloud server. Upon receiving the anomaly information, the cloud server automatically triggers a relearning process for the lighting demand forecast curve. First, the anomaly data and related data from the period before and after the anomaly (such as light intensity and human presence) are added to the fused dataset. Then, based on the updated fused dataset, the time series model is retrained, and model parameters are adjusted to adapt to the changing lighting demand patterns following the anomaly. This generates a new lighting demand forecast curve, providing a more accurate basis for subsequent lighting control.

[0051] Real-time monitoring of lighting circuit status and triggering relearning of prediction curves effectively responds to unexpected lighting system conditions. When a circuit anomaly causes a change in lighting performance, the abnormal data is promptly incorporated into analysis, allowing the lighting demand prediction curve to quickly adapt to the new environmental conditions and avoid lighting strategy failures caused by equipment failure. For example, if a lamp in a certain area breaks down, the system can quickly adjust the prediction curve based on the new data, replanning the brightness and switching times of other lamps to maintain the overall lighting effect. Furthermore, by continuously updating the model, the accuracy and reliability of the prediction curve are improved, the lighting system's adaptability is optimized, maintenance costs are reduced, and the long-term stable and efficient operation of the lighting system is guaranteed.

[0052] Furthermore, based on the above embodiments, in this embodiment, light intensity data and human presence status data of the target area are collected through a distributed sensor network, including: eliminating environmental interference through multi-sensor collaboration to obtain light intensity data; tracking human movement trajectory based on the distributed sensor topological relationship to obtain human presence status data; and implementing dynamic noise suppression on light intensity data and human presence status data.

[0053] Specifically, multiple light sensors and human infrared sensors are evenly deployed in the target area to build a distributed sensor network. Light sensors with ambient light compensation are selected to eliminate environmental interference through multi-sensor collaboration. At least three light sensors are formed into a sensor cluster, each collecting light data from different angles in the same area. The data is then fused using a weighted averaging algorithm. Weights are assigned based on the accuracy parameters of each sensor, with higher weights assigned to high-precision sensors. This helps offset interference factors such as reflected light and shadows in the environment, thereby obtaining accurate light intensity data.

[0054] Human movement is tracked based on the topological relationships of distributed sensors. Human infrared sensors are distributed in a grid pattern, with each sensor node carrying a unique identifier and coordinate information. When a person passes through the sensor's detection range, the sensor triggers and sends the detection information (including its own identifier) ​​to the edge device. Based on the sensor's topological relationships and triggering time sequence, the edge device uses a Kalman filter algorithm to process the discrete detection data, predicting the person's position at different times, and ultimately obtaining the person's presence status data and movement trajectory.

[0055] Dynamic noise suppression is implemented on collected light intensity and human presence data. Using a wavelet transform algorithm, the number of decomposition layers and thresholds are dynamically adjusted based on real-time data fluctuations. For light intensity data, high-frequency noise signals are decomposed and adaptive thresholds are set to filter out noise. For human presence data, abnormal signals caused by sensor mis-triggering are identified and removed, improving data accuracy and stability.

[0056] By collaborating across multiple sensors to eliminate environmental interference, the impact of complex environments on light intensity data collection can be effectively reduced, ensuring that the acquired data more closely reflects actual lighting conditions and providing a reliable basis for lighting system dimming. Tracking human movement based on distributed sensor topology allows for precise understanding of human activity patterns, enabling on-demand control of lighting equipment and avoiding ineffective lighting. Dynamic noise suppression technology improves data quality, reduces control errors caused by data errors, enhances the stability and reliability of the lighting system, improves energy efficiency, and provides a high-quality data foundation for subsequent data fusion and analysis.

[0057] Furthermore, based on the above embodiments, this embodiment generates a lighting demand prediction curve based on a fused data set through a time series model, including: performing feature engineering on the fused data set to extract periodic patterns of light and human activity; based on the periodic patterns, building a multi-channel model to process spatiotemporal features; performing feature weighting on the processed spatiotemporal features to generate a probabilistic demand distribution, and constructing a lighting demand prediction curve.

[0058] Specifically, after acquiring the fused dataset, feature engineering is first performed on it. Using algorithms such as Fourier transforms and sliding windows, the light intensity data and human presence data are analyzed to extract cyclical patterns in light changes and human activity. For example, by analyzing historical data, the cyclical variations in light intensity between weekdays and weekends, as well as peak activity times during different times of the day, can be identified.

[0059] Based on the extracted periodic patterns, a multi-channel model is constructed to process spatiotemporal features. Time series data (such as minute-by-minute and hour-by-hour light and human activity data), spatial data (sensor location information), and periodic pattern data are fed as separate channels into a deep learning model such as an LSTM or Transformer. Using a multi-layered neural network structure, the model automatically learns the spatiotemporal correlations between data from different channels, exploring the potential connections between light, human activity, and lighting needs in both temporal and spatial dimensions.

[0060] The spatiotemporal features output by the multi-channel model are weighted. Using an attention mechanism, each feature is weighted based on its importance to predicting lighting demand, highlighting the impact of key features. For example, during peak activity periods, features related to human presence are weighted more highly; at night, features related to light intensity are weighted more highly. A weighted calculation generates a probabilistic demand distribution, representing the likelihood of lighting demand at different times and in different areas in a probabilistic manner. Finally, the probabilistic demand distribution is converted into an intuitive curve to construct a lighting demand prediction curve, which is used to guide the development of subsequent lighting control strategies.

[0061] A method for generating lighting demand forecast curves based on fused datasets extracts periodic patterns through feature engineering, enabling in-depth exploration of the patterns between lighting and human activity, providing data support for predictions. A multi-channel model processes spatiotemporal features, fully accounting for both temporal and spatial dimensions, making the model more realistic. Feature weighting generates a probabilistic demand distribution, improving the accuracy and reliability of predictions. This method can accurately predict lighting demand for different time periods and regions in advance, enabling the lighting system to adjust brightness and on / off status in advance, avoiding energy waste and reducing operating costs. Furthermore, it dynamically adapts to changes in the environment and human activity, enhancing the intelligence and automation of the lighting system and providing users with a more comfortable and energy-efficient lighting environment.

[0062] Furthermore, based on the above embodiment, this embodiment also includes: gridding the target area based on the distributed sensor topology relationship; counting the human activity frequency in each grid; constructing a spatial heat map based on the human activity frequency in each grid, and constructing a multi-channel model by combining the periodic pattern of light and human activity.

[0063] Specifically, based on the distributed sensor topology, the coordinates of each sensor within the target area are obtained, dividing the target area into evenly sized grids, each with a unique number. After the sensor collects human presence data, the edge device determines the grid to which it belongs based on the sensor's coordinates. Within a set statistical period (e.g., 15 minutes), the number of times human activity triggers the sensor within each grid is counted to calculate the human activity frequency within each grid.

[0064] Using geographic information system (GIS) technology, the human activity frequency data in each grid is mapped to a visualization interface, and different color gradients are set to represent different activity frequency ranges, such as red for high-frequency activity areas and blue for low-frequency activity areas, thereby constructing a spatial heat map to intuitively present the spatial distribution of human activities in the target area.

[0065] Combining the cyclical patterns of light and human activity extracted through feature engineering, spatial heat map data is used as a new channel and fed into deep learning models such as LSTM or Transformer, along with time series data (light intensity and activity duration) and cyclical pattern data, to build a multi-channel model. By learning the spatiotemporal correlations between data from different channels, the model explores the potential connections between spatial human activity distribution, temporal light and human activity patterns, and lighting needs, providing more comprehensive data support for subsequent lighting demand forecasts.

[0066] By gridding the target area and counting the frequency of human activity to construct a spatial heat map, the distribution of human activity in different areas can be clearly and intuitively displayed, accurately locating areas with dense populations and high-frequency activity. Combining the periodic patterns of light and human activity to construct a multi-channel model allows the model to comprehensively consider multi-dimensional factors such as space, time, and cycle, and more comprehensively capture the key features that affect lighting needs. This helps to predict the lighting needs of different areas at different times in advance and achieve precise control of lighting equipment. For example, it can increase the lighting brightness in areas with high human activity and reduce energy consumption in areas with low activity, thereby effectively improving energy utilization efficiency and reducing the operating costs of the lighting system, while providing users with a comfortable lighting environment that better meets their actual needs.

[0067] Furthermore, the dynamic construction of the shiftable schedule in this embodiment includes: calculating a baseline offset based on external astronomical data; adjusting the baseline offset based on the deviation between the actual ambient light value and the predicted value of the lighting demand prediction curve; updating the baseline parameter when the deviation exceeds a threshold, and generating a shiftable schedule.

[0068] Specifically, the system first obtains information such as sunrise and sunset times and solar altitude from an external astronomical data platform. Based on an astronomical calendar model, it calculates theoretical natural light values ​​for the target area at different dates and times. Based on these theoretical natural light values ​​and historical lighting equipment operation data, it sets an initial baseline offset. This offset is used to measure the fundamental relationship between natural light and artificial lighting demand. For example, for a certain period after sunrise, the time to delay the start of artificial lighting is determined based on the solar altitude and historical energy consumption data, which serves as the baseline offset.

[0069] Real-time ambient light measurements are collected, along with the predicted values ​​from the lighting demand forecast curve, to calculate the deviation between the two. A sliding window algorithm is used to statistically analyze the deviation over a specific time period (e.g., 30 minutes). If the measured values ​​consistently exceed the predicted values, indicating sufficient natural light, the baseline offset can be appropriately increased, meaning the lighting device's off time is extended or its on time is delayed. Conversely, if the measured values ​​are lower than the predicted values, the baseline offset can be reduced.

[0070] A deviation threshold is set (e.g., when the measured value deviates from the predicted value by more than 20%). When the statistical deviation exceeds this threshold, the baseline parameter update process is triggered. The adjusted baseline offset, measured light data, and predicted curve data are incorporated into historical data to recalculate the baseline parameters for each time period and generate a new offset schedule. This offset schedule specifies the adjustable range of lighting equipment during different time periods. For example, during daytime hours with ample daylight and low activity, lighting equipment can be turned off or dimmed.

[0071] By calculating a benchmark offset based on external astronomical data and dynamically adjusting it based on measured and predicted deviations, it can accurately match natural and artificial lighting needs. By updating benchmark parameters to generate an offsettable schedule, this provides a flexible basis for adjusting lighting equipment operations. For example, during clear weather with ample natural light, the lighting off time can be automatically extended to reduce energy consumption. During cloudy days with insufficient light, the lighting strategy can be adjusted to meet demand. This method effectively improves the lighting system's ability to adapt to environmental changes, avoids over- or under-lighting, reduces energy loss and operating costs, and enhances the intelligence and automation of the lighting system, achieving more efficient and energy-saving lighting management.

[0072] Furthermore, in this embodiment, the collaborative strategy is compiled into physical instructions and the lighting circuit is driven by the edge device, including: caching the collaborative strategy; when the network is disconnected, generating emergency instructions based on the cached collaborative strategy and light intensity data and human presence status data; when the network is restored, uploading the offline log corresponding to the emergency instruction and calibrating the lighting demand prediction curve.

[0073] Specifically, before compiling the collaborative strategy into physical instructions, the edge device's built-in storage module caches the collaborative strategy, using non-volatile memory chips to ensure data is not lost in the event of a power outage or network disconnection. The edge device continuously monitors the network connection status. When a network interruption is detected, it immediately stops requesting data from the cloud and instead runs an emergency command generation algorithm based on the cached collaborative strategy, combined with local real-time light intensity data and human presence data collected. The built-in microprocessor runs the emergency command generation algorithm. For example, based on the preset light thresholds and human presence sensing rules in the collaborative strategy, if a person is detected in a certain area and the light intensity is below the threshold, an emergency command to "turn on the lighting in that area" is directly generated, and hardware components such as relays and dimming modules drive the lighting circuit for execution.

[0074] When the network is restored, the edge device compiles the emergency commands generated and executed during the outage, along with their corresponding timestamps, execution results, and other information, into an offline log and uploads it to the cloud server via 4G / 5G or Wi-Fi. After receiving the offline log, the cloud server integrates the data with the current fused dataset and retrains the updated dataset using a time series model. By comparing changes in light intensity, human activity, and other data before and after the execution of the emergency commands, the cloud server analyzes their impact on actual lighting demand and calibrates the lighting demand forecast curve to better align it with actual application scenarios, improving the accuracy of subsequent forecasts.

[0075] Edge devices cache collaborative strategies and generate emergency commands during network outages, ensuring that the lighting system maintains basic intelligent lighting functions based on local data and preset strategies during network outages, preventing lighting control failures due to network issues and ensuring the continuity and stability of lighting services. After network restoration, offline logs are uploaded to calibrate the lighting demand forecast curve. This incorporates actual operating data from the outage into the analysis, correcting deviations in the forecast model caused by network outages and enabling the forecast curve to promptly adapt to environmental changes and actual needs. This further enhances the lighting system's intelligence and energy efficiency, while mitigating the negative impact of network fluctuations on lighting system management.

[0076] Furthermore, in this embodiment, the loop status of the lighting circuit is monitored in real time, and when an abnormality is detected, the re-learning process of the lighting demand prediction curve is triggered, including: analyzing the loop current waveform of the lighting circuit to identify the abnormal pattern; comparing the abnormal pattern with the abnormal pattern library, and classifying and determining the fault type; and executing a response action according to the fault type to trigger the re-learning process of the lighting demand prediction curve.

[0077] Specifically, high-precision current sensors are deployed in the lighting circuits to collect circuit current signals in real time at a fixed sampling frequency. These analog signals are converted into digital signals and transmitted to edge devices. The edge devices have built-in digital signal processing modules that use fast Fourier transform (FFT) algorithms to perform frequency domain analysis on the collected current waveforms, extracting characteristic parameters such as amplitude, frequency, and harmonics. These parameters are then compared with the current waveform characteristics during normal operation to identify abnormal patterns, such as sudden increases in current amplitude or the presence of unusual harmonic components.

[0078] An abnormal pattern library is established, storing abnormal current waveform patterns and fault types corresponding to various known faults. For example, a short circuit fault corresponds to a sudden and significant increase in current amplitude and severe harmonic distortion, and a poor contact fault corresponds to intermittent fluctuations in the current waveform. The identified abnormal patterns are compared with the data in the abnormal pattern library, and a pattern matching algorithm (such as the dynamic time warping algorithm) is used to calculate similarity. The fault type is determined based on the pattern with the highest similarity.

[0079] Once the fault type is determined, the edge device executes the corresponding response action. If it is a short circuit, a relay immediately disconnects the power supply to the corresponding lighting circuit to ensure power safety. If it is a minor fault such as poor contact, a fault warning message is sent to the management terminal. Simultaneously, the edge device uploads information such as the fault type, fault occurrence time, and current waveform data before and after the fault to the cloud server, triggering the relearning process of the lighting demand prediction curve. The cloud server adds the new fault data to the fused dataset, retrains the time series model, adjusts the model parameters, and generates a new lighting demand prediction curve to adapt to the changes in the lighting environment after the fault.

[0080] By monitoring the lighting circuit status in real time and performing fault handling based on abnormal pattern recognition, lighting circuit faults can be quickly and accurately located, allowing timely implementation of safety precautions to prevent safety incidents or lighting effects caused by the fault. Triggering the relearning process of the lighting demand prediction curve incorporates the actual operating data of the lighting system after the fault, allowing the prediction curve to promptly adapt to changes in the lighting environment caused by the fault, avoiding prediction deviations caused by the fault. For example, when the number of lamp failures in a certain area decreases, the system can quickly adjust the prediction curve and replan the lighting strategy to maintain the efficient operation of the lighting system, reduce energy waste, and improve the intelligent management level and reliability of the lighting system.

[0081] Furthermore, based on the above embodiment, after collecting the light intensity data and human presence status data of the target area through a distributed sensor network, this embodiment also includes: when the deviation between the data of a single sensor and the monitoring value of the adjacent sensor cluster exceeds a dynamic threshold, it is marked as a suspicious node; calling the historical lighting model and astronomical data of the associated area to verify the physical rationality of the data of the suspicious node; if the verification fails, automatically switching to the backup sensor data stream and triggering an equipment maintenance alarm.

[0082] Specifically, in a distributed sensor network, each sensor node sends collected light intensity data and human presence data to an edge device in real time. Based on the sensor topology, the edge device groups three to five adjacent sensors into a cluster. Using a sliding window algorithm, the average and standard deviation of each sensor's data relative to the cluster's monitored values ​​are calculated, and a dynamic deviation threshold (e.g., exceeding the cluster average ±3 standard deviations) is set. If the deviation between a single sensor's data and the values ​​monitored by the adjacent sensor cluster exceeds the dynamic threshold, the sensor is marked as a suspicious node.

[0083] For flagged suspicious nodes, the edge device retrieves historical illumination model data for the associated area. This model, built based on historical light intensity, time, weather, and other information, reflects the normal pattern of illumination changes in the area. It also obtains external astronomical data, such as solar altitude and cloud cover. The suspicious node data is substituted into the historical illumination model and, combined with the astronomical data, a calculation is performed to determine whether the data is physically reasonable in the current environment. For example, if a sensor indicates extremely low light intensity at noon on a sunny day, verification fails.

[0084] If verification fails, the edge device automatically switches to a backup sensor data stream, which can be another functioning sensor in the same area. Simultaneously, the edge device generates a maintenance alert containing the suspicious node ID, abnormal data, and verification results. This alert is sent wirelessly to the operation and maintenance management platform and to the mobile terminals of relevant personnel, prompting them to inspect and maintain the abnormal sensor.

[0085] This solution can quickly identify and address sensor data anomalies, preventing erroneous data from affecting lighting system control due to single sensor failures. Data rationality is verified by incorporating historical illumination models and astronomical data, ensuring that the data conforms to the actual physical environment and improving data reliability. Automatic switching to backup data streams ensures the continuity of monitoring data and maintains the normal operation of the lighting system. Timely triggering of equipment maintenance alerts allows operators to quickly locate faulty sensors, shortening repair time and reducing the risk of lighting control errors caused by sensor failures, thereby improving the stability and reliability of the entire environmental monitoring and lighting control system.

[0086] Furthermore, based on the above embodiment, this embodiment also includes: statistically analyzing the actual energy saving rate and false operation rate after executing the collaborative strategy; establishing a correlation matrix between strategy parameters and abnormal events based on the actual energy saving rate and false operation rate, and locating the failure strategy characteristics; injecting the failure characteristics into the model training set of the lighting demand prediction curve, generating a strategy blacklist, and optimizing the collaborative strategy generation rules.

[0087] Specifically, while the lighting system is executing a collaborative strategy, smart meters collect real-time, minute-by-minute data on lighting equipment power usage and store it on edge devices. Simultaneously, the edge devices compare data from human infrared sensors and light sensors with lighting equipment action commands to identify misoperation scenarios, such as when no one is on or when high brightness lighting is used under strong sunlight. These data then record the time and frequency of misoperation. After the strategy is executed, the actual energy savings and misoperation rate are calculated using a formula based on historical power usage data from the same period when the strategy was not executed.

[0088] Parameters in the collaborative strategy, such as light threshold, human sensor sensitivity, and offset schedule, are correlated with the actual energy savings rate and false alarm rate. The Pearson correlation coefficient algorithm is used to calculate the correlation between each parameter and the two indicators. For example, if the correlation coefficient for false alarm rate reaches 0.7 under a certain light threshold setting, it indicates a high correlation between that parameter and false alarms. The calculated correlation data is used to construct a two-dimensional correlation matrix, visually displaying the degree of correlation between each parameter and abnormal events (low energy savings, high false alarm rates).

[0089] By screening the correlation matrix for parameter combinations that are strongly correlated (with an absolute value of the correlation coefficient greater than 0.6) with low energy savings (below a set threshold, such as 15%) and high malfunction rates (above a set threshold, such as 10%), we identify failure strategy signatures. For example, if the energy savings significantly decrease and malfunctions occur frequently when the light threshold is below 200 lux and the human sensor sensitivity is above 80%, we include this parameter combination in our failure signature.

[0090] Failure signatures are added to the training set of the lighting demand forecast curve in the form of data samples, and the sample attributes are labeled. Using a time series model such as LSTM or Transformer, retraining is performed based on the updated training set. Model parameters are adjusted through backpropagation, allowing the model to learn to avoid failure signatures during training. After training is complete, a policy blacklist is generated based on the failure signatures, clearly specifying prohibited parameter combinations. When subsequently generating collaborative strategies, the system automatically retrieves the blacklist and, based on the optimized generation rules, prioritizes parameter combinations that have been validated through training, achieving dynamic strategy optimization.

[0091] This implementation provides an objective basis for evaluating the effectiveness of collaborative strategies by quantifying actual energy savings and false operation rates. A correlation matrix and failure feature location mechanism accurately identify key parameters that lead to strategy failure, avoiding subjective bias. Failure features are injected into the training set and used to generate a strategy blacklist, effectively optimizing the lighting demand prediction model and strategy generation rules, enabling subsequent strategies to automatically avoid historical failure modes. In practical applications, this can reduce the frequency of false operation in lighting systems, improve energy efficiency, minimize energy waste and equipment loss, and significantly enhance the intelligent management and economic benefits of lighting systems.

[0092] Based on the same general inventive concept, the present invention also protects a lighting optimization control system based on environmental monitoring. The lighting optimization control system based on environmental monitoring described below and the lighting optimization control method based on environmental monitoring described above can refer to each other.

[0093] Figure 2 Schematic diagram of the lighting optimization control system based on environmental monitoring provided in this embodiment.

[0094] like Figure 2 As shown, this embodiment provides a lighting optimization control system based on environmental monitoring, including: The acquisition module 201 is used to collect light intensity data and human presence status data of the target area through a distributed sensor network; Fusion module 202, for performing spatiotemporal alignment of light intensity data and human presence status data with external astronomical data and historical energy consumption data in the cloud to obtain a fused data set; A prediction module 203 is configured to generate a lighting demand prediction curve based on the fused data set and a time series model; A construction module 204 is used to dynamically construct a collaborative strategy integrating lighting thresholds, human body sensing, and a shiftable schedule based on a lighting demand prediction curve; A driver module 205 is used to compile the collaborative strategy into physical instructions to drive the lighting circuit through the edge device; The feedback module 206 is used to monitor the circuit status of the lighting circuit in real time, and trigger the re-learning process of the lighting demand prediction curve when an abnormality is detected.

[0095] Figure 3 Schematic diagram of the structure of the electronic device provided in this embodiment.

[0096] like Figure 3As shown, the electronic device may include: a processor 310, a communications interface 320, a memory 330, and a communications bus 340, wherein the processor 310, the communications interface 320, and the memory 330 communicate with each other via the communications bus 340. The processor 310 may call logic instructions in the memory 330 to execute a lighting optimization control method based on environmental monitoring, the method comprising: collecting light intensity data and human presence status data of a target area through a distributed sensor network; performing spatiotemporal alignment of the light intensity data and human presence status data with external astronomical data and historical energy consumption data in the cloud to obtain a fused data set; generating a lighting demand prediction curve based on the fused data set using a time series model; dynamically constructing a collaborative strategy that integrates light thresholds, human presence sensing, and a shiftable schedule based on the lighting demand prediction curve; compiling the collaborative strategy into physical instructions to drive a lighting circuit through an edge device; and monitoring the circuit status of the lighting circuit in real time, triggering a relearning process for the lighting demand prediction curve when an anomaly is detected.

[0097] Furthermore, the logic instructions in the aforementioned memory 330 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the portion that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product, stored in a storage medium, includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as a USB flash drive, a mobile hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0098] On the other hand, the present invention also provides a computer program product, which includes a computer program, which can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the lighting optimization control method based on environmental monitoring provided by the above-mentioned methods, the method including: collecting light intensity data and human presence status data of the target area through a distributed sensor network; aligning the light intensity data and human presence status data with external astronomical data and historical energy consumption data in the cloud in time and space to obtain a fused data set; based on the fused data set, generating a lighting demand prediction curve through a time series model; according to the lighting demand prediction curve, dynamically constructing a collaborative strategy that integrates lighting thresholds, human body sensing and offset schedules; compiling the collaborative strategy into physical instructions to drive the lighting circuit through edge devices; monitoring the circuit status of the lighting circuit in real time, and triggering the re-learning process of the lighting demand prediction curve when an abnormality is detected.

[0099] On the other hand, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to execute the lighting optimization control method based on environmental monitoring provided by the above-mentioned methods, the method comprising: collecting light intensity data and human presence status data of the target area through a distributed sensor network; performing spatiotemporal alignment of the light intensity data and human presence status data with external astronomical data and historical energy consumption data in the cloud to obtain a fused data set; based on the fused data set, generating a lighting demand prediction curve through a time series model; dynamically constructing a collaborative strategy that integrates lighting thresholds, human body sensing and offset schedules according to the lighting demand prediction curve; compiling the collaborative strategy into physical instructions to drive the lighting circuit through an edge device; monitoring the circuit status of the lighting circuit in real time, and triggering the re-learning process of the lighting demand prediction curve when an abnormality is detected.

[0100] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, i.e., they may be located in one location or distributed across multiple network units. Some or all of the modules may be selected based on actual needs to achieve the objectives of the present embodiment. Persons of ordinary skill in the art will be able to understand and implement the present invention without inventive effort.

[0101] Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment can be implemented using software plus a necessary general-purpose hardware platform, or of course, hardware. Based on this understanding, the essence of the above technical solution, or the portion that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a magnetic disk, or an optical disk, and includes a number of instructions for causing a computer device (such as a personal computer, server, or network device) to execute the methods described in each embodiment or certain portions of the embodiments.

[0102] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A lighting optimization control method based on environmental monitoring, characterized in that: include: Collect light intensity data and human presence data in the target area through a distributed sensor network; Performing spatiotemporal alignment on the cloud for the light intensity data and human presence status data, external astronomical data, and historical energy consumption data to obtain a fused data set; Based on the fused data set, generating a lighting demand prediction curve through a time series model; Based on the lighting demand forecast curve, dynamically build a collaborative strategy that integrates lighting thresholds, human body sensing, and shiftable schedules; Compiling the collaborative strategy into physical instructions to drive the lighting loop via an edge device; The circuit status of the lighting circuit is monitored in real time, and when an abnormality is detected, a re-learning process of the lighting demand prediction curve is triggered.

2. The lighting optimization control method based on environmental monitoring according to claim 1, characterized in that: The method of collecting light intensity data and human presence status data of the target area through a distributed sensor network includes: Eliminate environmental interference through multi-sensor collaboration to obtain light intensity data; Track the movement trajectory of the human body based on the topological relationship of distributed sensors to obtain the human body's presence status data; Dynamic noise suppression is performed on the light intensity data and the human body presence status data.

3. The lighting optimization control method based on environmental monitoring according to claim 2, characterized in that: Generating a lighting demand prediction curve based on the fused data set through a time series model includes: Perform feature engineering on the fused dataset to extract periodic patterns of light and human activity; Based on the periodic pattern, a multi-channel model is constructed to process the spatiotemporal characteristics; The processed spatiotemporal features are weighted to generate a probabilistic demand distribution and construct a lighting demand prediction curve.

4. The lighting optimization control method based on environmental monitoring according to claim 3, characterized in that: Also includes: Gridding the target area based on the distributed sensor topological relationship; Count the frequency of human activities in each grid; Based on the human activity frequencies within the grids, a spatial heat map is constructed, and a multi-channel model is constructed by combining the periodic patterns of illumination and human activity.

5. The lighting optimization control method based on environmental monitoring according to claim 1, characterized in that: The dynamically constructed shiftable schedule includes: calculating a reference offset based on the external astronomical data; Adjusting the reference offset based on a deviation between an actual ambient light measurement value and a predicted value of the lighting demand prediction curve; When the deviation exceeds a threshold, the reference parameter is updated to generate a feasible schedule.

6. The lighting optimization control method based on environmental monitoring according to claim 1, characterized in that: Compiling the collaborative strategy into physical instructions to drive the lighting loop through the edge device includes: caching the collaborative strategy; When the network is disconnected, an emergency instruction is generated based on the cached collaborative strategy and the light intensity data and human presence status data; When the network is restored, the offline log corresponding to the emergency instruction is uploaded and the lighting demand prediction curve is calibrated.

7. The lighting optimization control method based on environmental monitoring according to claim 1, characterized in that: The real-time monitoring of the circuit status of the lighting circuit and triggering the re-learning process of the lighting demand prediction curve when an abnormality is detected include: Analyzing the loop current waveform of the lighting circuit to identify abnormal patterns; Comparing the abnormal pattern with the abnormal pattern library, classifying and determining the fault type; According to the fault type, a response action is executed to trigger a re-learning process of the lighting demand prediction curve.

8. The lighting optimization control method based on environmental monitoring according to any one of claims 1 to 7, characterized in that: After collecting the light intensity data and human presence status data of the target area through the distributed sensor network, the method further includes: When the deviation between a single sensor data and the monitoring value of the adjacent sensor cluster exceeds the dynamic threshold, it is marked as a suspicious node; Calling the historical illumination model and astronomical data of the associated area to verify the physical rationality of the data of the suspicious node; If verification fails, it automatically switches to the backup sensor data stream and triggers an equipment maintenance alarm.

9. The lighting optimization control method based on environmental monitoring according to any one of claims 1 to 7, characterized in that: Also includes: Calculate the actual energy saving rate and malfunction rate after executing the collaborative strategy; Based on the actual energy saving rate and the malfunction occurrence rate, a correlation matrix between strategy parameters and abnormal events is established to locate the characteristics of the failed strategy; The failure characteristics are injected into the model training set of the lighting demand prediction curve to generate a strategy blacklist and optimize the collaborative strategy generation rules.

10. A lighting optimization control system based on environmental monitoring, characterized in that: include: The acquisition module is used to collect light intensity data and human presence status data of the target area through a distributed sensor network; A fusion module is used to align the light intensity data and human presence status data with external astronomical data and historical energy consumption data in the cloud in time and space to obtain a fused data set; A prediction module, configured to generate a lighting demand prediction curve using a time series model based on the fused data set; A construction module for dynamically constructing a collaborative strategy integrating lighting thresholds, human body sensing, and a shiftable schedule based on the lighting demand prediction curve; a driver module, configured to compile the collaborative strategy into physical instructions and drive the lighting circuit via an edge device; The feedback module is used to monitor the circuit status of the lighting circuit in real time and trigger the re-learning process of the lighting demand prediction curve when an abnormality is detected.

Images