Urban intelligent street lamp control method and system

Abstract

The invention discloses an urban intelligent street lamp control method and system. The method comprises the following steps: acquiring road environment parameters (PM2.5 concentration, humidity, environment illumination and SO2 concentration) and traffic state parameters (traffic flow, pedestrian flow, average lane speed and lane occupancy); constructing a brightness adjustment model and a color temperature adjustment model, performing combined weighting through an entropy weight method and an analytic hierarchy process, introducing a dynamic weight correction factor and fusing real-time wind speed data to perform diffusion compensation on SO2 concentration, and generating a brightness judgment value and a color temperature judgment value; and based on the judgment value, dividing a brightness adjustment level and a color temperature adjustment level, and dynamically adjusting the brightness and color temperature of the street lamp. According to the invention, fine adjustment of the brightness and the color temperature of the street lamp is realized, the scientificity, the accuracy, the real-time performance and the adaptability of the adjustment are improved, the street lamp illumination is ensured to meet the road illumination requirement, and the system can adapt to different traffic conditions and weather conditions.

Description

Technical Field

[0001] The present invention relates to the technical field of road traffic lighting, and particularly to an urban intelligent street lamp control method and system. Background Art

[0002] In modern urban traffic management, street lamps, as an important part of urban infrastructure, not only provide lighting for night driving and pedestrians, but also play an important role in improving road safety, energy conservation and emission reduction, and shaping the urban image. With the acceleration of the urbanization process, the urban traffic pressure is increasing day by day, and the traditional street lamp control system has been unable to meet the growing demand for intelligent and refined management. Traditional street lamp control mostly relies on timed switching or simple photosensitive control, lacking flexibility and intelligence, and unable to dynamically adjust the lighting intensity and time according to the actual traffic flow and environmental changes, resulting in energy waste and low lighting efficiency.

[0003] In recent years, with the development and popularization of the Internet of Things technology, the concept of intelligent street lamps has emerged. Intelligent street lamps can realize remote monitoring and intelligent management of street lamps by integrating sensors, wireless communication modules and intelligent control algorithms. For example, through vehicle detectors and pedestrian sensors installed on street lamps, the street lamp system can collect traffic flow data in real time and dynamically adjust the lighting intensity and on / off time according to the data, so as to achieve the purpose of energy conservation and emission reduction. In addition, intelligent street lamps can also be integrated with other urban management systems, such as traffic signal control systems, video monitoring systems, etc., to realize information sharing and collaborative work, and further improve the intelligent level of urban traffic management.

[0004] The patent with the publication number of CN118317484B discloses an intelligent street lamp lighting control method and system based on the Internet of Things technology, revealing an intelligent street lamp lighting control method and system using the Internet of Things technology. For each street lamp in the intelligent street lamp system, obtain its current lighting parameter configuration scheme, and use the Internet of Things technology to collect the street lamp lighting data corresponding to this scheme. Then, based on the collected street lamp lighting data, calculate the energy consumption generated by this scheme. If the energy consumption exceeds the pre-set energy consumption threshold range, convert each lighting parameter in the scheme into a gene string structure. Use the preset objective function to perform iterative optimization processing on the gene string, so as to obtain an optimized lighting parameter configuration scheme.

[0005] However, the existing intelligent street lamp systems still have some limitations. For example, most systems have limited capabilities in data processing and decision support, and are unable to implement lighting control strategies in complex scenarios. At the same time, due to the lack of effective data analysis and prediction models, the adaptive ability of the street lamp system is insufficient. Summary of the Invention

[0006] The object of the present invention is to overcome the deficiencies of the prior art and provide an urban intelligent street lamp control method and system, aiming to solve the problem of how to realize the fine adjustment of the brightness and color temperature of street lamps by comprehensively collecting road environmental parameters and traffic state parameters to ensure that the street lamp lighting meets the road lighting requirements.

[0007] The object of the present invention is achieved by the following technical solutions: An urban intelligent street lamp control method includes the following: S1. Collect road environmental parameters and traffic state parameters. The environmental parameters include PM2.5 concentration, humidity, ambient illuminance, and SO2 concentration, and the traffic state parameters include traffic flow, pedestrian flow, average lane speed, and lane occupancy; S2. Construct a brightness adjustment model and a color temperature adjustment model, and calculate the weights of each parameter through a combined weighting algorithm respectively: The brightness adjustment model takes traffic flow, pedestrian flow, PM2.5 concentration, humidity, and ambient illuminance as input parameters, combines the entropy weight method and the analytic hierarchy process for weighting, and introduces a dynamic weight correction factor to generate a brightness judgment value; The color temperature adjustment model takes PM2.5 concentration, SO2 concentration, and humidity as input parameters, combines the entropy weight method and the analytic hierarchy process for weighting, and compensates for the diffusion of SO2 concentration by integrating real-time wind speed data to generate a color temperature judgment value; S3. Based on the brightness judgment value and the color temperature judgment value, divide the brightness adjustment level and the color temperature adjustment level respectively, and dynamically adjust the brightness and color temperature of the street lamp.

[0008] As a preferred method, it further includes: S4. Realize the linkage of street lamps in the upstream and downstream sections through a coordinated control strategy: The upstream centralized controller calculates the adjustment time, holding time, and target brightness / color temperature value of the downstream street lamps according to the detected traffic flow, pedestrian flow, average lane speed, and lane occupancy, in combination with a dynamic traffic flow prediction model based on LSTM; Send the adjustment instruction to the downstream centralized controller through a LoRaWAN and 5G dual-mode redundant communication protocol, so that the downstream street lamps complete the pre-adjustment of brightness and color temperature before the arrival of the traffic flow.

[0009] As a preferred method, the calculation method of the dynamic weight correction factor in the combined weighting algorithm is: ; Where: Is the objective weight calculated by the entropy weight method; Is the absolute deviation between the current parameter value and the historical mean, and the historical mean is the mean value in the same time period in the past 7 days; is the maximum allowable fluctuation range of the parameter, set according to national standards; The dynamically corrected comprehensive weight is: Among them, is the comprehensive weight of the th parameter, indicating the final influence of this parameter in the brightness or color temperature adjustment model; is the weight obtained by the analytic hierarchy process, indicating the th parameter, is the total number of parameters input in the model, and the specific value depends on the adjustment target. Brightness adjustment model: (Traffic flow, pedestrian flow, PM2.5 concentration, humidity, ambient illuminance); Color temperature adjustment model: (PM2.5 concentration, SO2 concentration, humidity).

[0010] As a preferred method, the division method of the brightness adjustment level is: divide the brightness judgment value interval [0,1] into an adaptive Gaussian interval based on the standard deviation of historical traffic flow. The specific formula is: interval ; Among them, is the central value of the th brightness level, taking values of 25%, 50%, 75%, 100% of the base brightness; is the interval standard deviation, dynamically adjusted according to the standard deviation of traffic flow in the past 1 hour.

[0011] The division of the brightness adjustment level is dynamically adjusted based on the fluctuation of historical traffic flow. The system divides the brightness interval [0,1] into four levels (25%, 50%, 75%, 100%), and the central value of each level corresponds to the base brightness. By statistically calculating the standard deviation of traffic flow in the past 1 hour in real time, the interval range of each brightness level is dynamically calculated: if the traffic flow fluctuates greatly (large standard deviation), the interval range is automatically widened to avoid frequent adjustment; if the traffic flow is stable (small standard deviation), the interval is narrowed and the brightness response is more sensitive. This adaptive mechanism can intelligently balance energy consumption and lighting requirements according to the actual traffic flow changes on the road, ensuring driving safety and reducing energy waste.

[0012] As a preferred method, in the brightness adjustment model, the inverse correlation relationship between ambient illuminance and brightness is realized through a piecewise function: ; Among them, is the base brightness (unit: %); is the ambient illuminance (unit: lux), and the coefficients 0.03 and 0.01 implicitly have the unit of dimensionless proportional coefficient / lux, or directly take of the numerical value for substitution calculation, is the lowest brightness threshold (default 20%) to avoid negative values or excessive dimming.

[0013] As a preferred method, the color temperature adjustment levels are divided as follows: the color temperature judgment value range [0, 1] is divided into three sub-ranges, corresponding to the dynamic color temperature values based on the haze index: ; Among them, is the color temperature judgment value, and the calculation formula is: ; : PM2.5 normalization value. When the PM2.5 concentration ≤ 300 μg / m 3 , ; when the PM2.5 concentration > 300 μg / m 3 , ; : SO2 normalization value, (based on the 24-hour average limit of GB 3095-2012); : humidity normalization value, ; is the combined weight, satisfying , and the calculation formula is: ; The objective weight of the th parameter calculated by the entropy weight method (based on parameter volatility); : The subjective weight of the th parameter determined by the analytic hierarchy process (expert experience).

[0014] Calculation of the objective weight by the entropy weight method: (brightness model); (color temperature model); is the objective weight of the PM2.5 parameter in the brightness model and the color temperature model, corresponding to (brightness model) and (color temperature model).

[0015] ; : The original objective weight of PM2.5 calculated by the entropy weight method; : The absolute deviation of the current PM2.5 concentration from the historical mean; the maximum allowable fluctuation range of PM2.5, for example .

[0016] When the deviation of PM2.5 concentration is amplified by the dynamic correction factor to reflect the real-time volatility of the pollution concentration. The corrected weights need to be renormalized to ensure that the sum of all parameter weights is 1, where only the weight of PM2.5 is corrected and the weights of other parameters remain unchanged.

[0017] As a preferred method, the remote communication adopts the LRaWAN and 5G dual-mode redundant transmission protocol, with a transmission delay of less than 2 seconds and a packet loss rate of less than 0.1%.

[0018] As a preferred method, in the collaborative control strategy, the adjustment time of the downstream street lights is determined by the improved fuzzy PID control algorithm, specifically: Taking the average lane speed (unit: km / h) and the lane occupancy rate (unit: %) as inputs, a membership function of the new congestion index is added: where is the highest designed road speed (unit: km / h); The membership function of the output adjustment time (unit: seconds) adopts a dynamic asymmetric Gaussian distribution: ; where is the expected value of the adjustment time calculated from historical data, the standard deviation, and the fluctuation range of the adjustment time.

[0019] As a preferred method, in the improved fuzzy PID control algorithm, the formula for dynamically adjusting the proportional coefficient is: ; where is the initial proportional coefficient; is the vehicle speed predicted by the LSTM model (unit: km / h); is the actually detected vehicle speed (unit: km / h).

[0020] As a preferred method, after receiving the adjustment instruction, if the deviation between the actual traffic flow and the predicted value exceeds 20%, the downstream centralized controller re-predicts the target brightness / color temperature value through the Kalman filter algorithm: state equation: ; Observation equation: ; Where: is the state vector, including traffic flow (unit: vehicles per minute) and vehicle speed (unit: km / h); is the system matrix; are the process noise and the observation noise.

[0021] As an optimal method, it also includes an abnormal data processing step: when the PM2.5 concentration exceeds 300 μg / m 3 , start the haze penetration mode, and force to set , directly trigger , and lock the color temperature at 3000K; and increase the brightness to 120% of the rated value, and continue until the PM2.5 drops below 200 μg / m 3 , and execute after verifying the adjustment effect through the digital twin platform.

[0022] An urban intelligent street lamp control system includes: A data acquisition module for obtaining road environment parameters and traffic state parameters in real time; A dynamic weight calculation module, deployed in the centralized controller, includes: A brightness weight unit, integrating the exponential smoothing filtering algorithm for PM2.5 concentration, and the filtering formula is: ; Wherein, is the historical mean (unit: μg / m 3 ); is the current value, is the PM2.5 concentration value processed by the exponential smoothing filtering algorithm; A color temperature weight unit, integrating the wind speed compensation model for concentration, and the compensation formula is: Wherein, is the current wind speed; An adjustment decision module for generating a brightness level instruction and a color temperature level instruction according to the brightness judgment value and the color temperature judgment value; A collaborative control module, includes: An upstream controller, configured to run a traffic flow prediction model based on LSTM, and output the downstream adjustment time, the holding time, and the target brightness / color temperature value; A downstream controller, configured to calibrate the target brightness / color temperature value in real time through the Kalman filtering algorithm, and execute the adjustment when the prediction time arrives.

[0023] The present invention has at least the following beneficial effects: By comprehensively collecting road environment parameters and traffic state parameters, the present invention realizes the refined adjustment of the brightness and color temperature of street lamps. By constructing a brightness adjustment model and a color temperature adjustment model, and using a combined weighting algorithm to calculate the weights of various parameters, the scientificity and accuracy of the adjustment basis are ensured. At the same time, by introducing a dynamic weight correction factor and real-time wind speed data for the diffusion compensation of SO2 concentration, the real-time performance and adaptability of the adjustment are further improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] In order to reveal the technical details of the embodiments of the present invention, the drawings involved in the embodiments will be briefly introduced next. It should be emphasized that these drawings only present several embodiments of the present invention and should not be regarded as a definition of the scope of the invention. For those of ordinary skill in the art, without creative labor, other related drawings can still be derived based on these drawings.

[0025] Figure 1 It is a schematic flow chart of the first embodiment; Figure 2 It is a schematic structural diagram of a city intelligent street lamp control system. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0026] The technical solution of the present invention will be further described in detail below with reference to the drawings, but the protection scope of the present invention is not limited to the following.

[0027] In the following content, the embodiments of the present disclosure will be described in detail with the help of the drawings. However, it should be clear that the present disclosure is not limited to the specific forms shown here. On the contrary, it should be understood to cover various changes, equivalent forms and / or alternative solutions of the embodiments of the present disclosure. In the process of elaborating the drawings, the same reference numerals will be used to label similar components.

[0028] It should be clear that although specific details are provided in the following description to help a comprehensive understanding of the example embodiments. Those skilled in the art should know that even without these specific details, the example embodiments can still be implemented. For example, the system can be presented in the form of a block diagram to avoid excessive details interfering with the clarity of the example. In other cases, in order to maintain the clarity of the example, unnecessary details of those well-known processes, structures and technologies may be omitted.

[0029] As Figure 1 shown, a method for controlling an intelligent street lamp in a city includes the following: S1. Collect road environmental parameters and traffic state parameters. The environmental parameters include PM2.5 concentration, humidity, ambient illuminance, and SO2 concentration, and the traffic state parameters include traffic flow, pedestrian flow, average lane speed, and lane occupancy rate. S2. Construct a brightness adjustment model and a color temperature adjustment model, and calculate the weights of each parameter through a combined weighting algorithm respectively: The brightness adjustment model takes traffic flow, pedestrian flow, PM2.5 concentration, humidity, and ambient illuminance as input parameters, combines the entropy weight method and the analytic hierarchy process for weighting, and introduces a dynamic weight correction factor to generate a brightness judgment value; The color temperature adjustment model takes PM2.5 concentration, SO2 concentration, and humidity as input parameters, combines the entropy weight method and the analytic hierarchy process for weighting, and fuses real-time wind speed data to compensate for the diffusion of SO2 concentration to generate a color temperature judgment value. S3. Based on the brightness judgment value and the color temperature judgment value, divide the brightness adjustment level and the color temperature adjustment level respectively, and dynamically adjust the brightness and color temperature of the street lamp.

[0030] In this embodiment, by real-time monitoring of environmental and traffic data, the brightness and color temperature of the street lamp are dynamically adjusted, taking into account the needs of road safety and energy conservation. Data collection and parameter integration: The system collects two types of data through sensors in real time: Environmental parameters: including PM2.5 concentration (reflecting haze), humidity (affecting visibility), ambient illuminance (natural light intensity), SO2 concentration (pollution index), and wind speed (affecting pollutant diffusion). Traffic parameters: traffic flow, pedestrian flow (determining lighting requirements), average lane speed, and lane occupancy rate (judging congestion level).

[0031] The brightness adjustment model synthesizes five parameters: traffic flow, pedestrian flow, PM2.5, humidity, and ambient illuminance, and combines the "entropy weight method + analytic hierarchy process" for weighting (taking into account data objectivity and expert experience), and introduces a dynamic correction mechanism. For example, when the PM2.5 concentration suddenly increases, the system automatically increases its weight to give priority to ensuring the fog penetration. Brightness adjustment: The brightness judgment value is divided into four levels (25%, 50%, 75%, 100%). For example, during the evening rush hour when the traffic is dense, the brightness is adjusted to 100%, and when there are few cars in the early morning, it is reduced to 25%, realizing "more cars, brighter lights; fewer cars, dimmer lights". This part of the content will be described in detail in the subsequent embodiments.

[0032] The color temperature adjustment model is based on three parameters: PM2.5, SO2, and humidity, and combines wind speed data to compensate for the SO2 concentration (the greater the wind speed, the faster the pollutant diffusion, and the lower the influence weight of SO2). For example, in windy weather, even if the SO2 concentration is high, the system still maintains a high color temperature to optimize the light effect. Color temperature adjustment: The color temperature judgment value is divided into three levels (3000K warm yellow light, 4000K neutral light, 5000K cold white light). Automatically switch to 3000K on hazy days to enhance fog penetration, and use 5000K high-efficiency cold light on clear nights to save energy.

[0033] When the environment suddenly changes (such as a sharp increase in PM2.5) or traffic fluctuates, the model quickly adjusts the weights to avoid adjustment lag. By hierarchical adjustment, the balance between energy consumption and demand is achieved. In haze days, safe fog penetration is ensured, and in low-flow periods, ineffective lighting is reduced. Wind speed data corrects the influence of SO2, avoiding excessive reduction of color temperature on windy days and improving the rationality of adjustment. Through multi-parameter collaborative decision-making, the street lights of the present invention can sense the environment and predict demands, realizing intelligent and refined management and control in complex scenarios. The specific content will be described in detail in the following embodiments.

[0034] In a preferred embodiment, it further includes: S4. Realizing the linkage of street lights on upstream and downstream sections through a collaborative control strategy: The upstream centralized controller calculates the adjustment time, holding time, and target brightness / color temperature value of the downstream street lights according to the detected traffic flow, pedestrian flow, average vehicle speed of the lane, and lane occupancy rate, in combination with the dynamic traffic flow prediction model based on LSTM; LSTM is a recurrent neural network dedicated to processing long-term dependencies in time series data. Compared with ordinary RNNs, LSTM effectively solves the problem of gradient disappearance or explosion in traditional RNNs, enabling it to remember long-term historical information and dynamically adjust the memory content. The upstream centralized controller realizes dynamic traffic flow prediction through the LSTM model, and the input data includes historical traffic flow data, time features, and environmental parameters. Historical traffic flow data: includes time series data such as traffic flow, average vehicle speed of the lane, and lane occupancy rate. Time features: time period (peak / off-peak), date (weekday / holiday). Environmental parameters: factors affecting visibility such as weather (haze, rainfall), PM2.5 concentration, etc. The LSTM network is trained with historical data to learn the spatio-temporal laws of traffic flow changes. Such as the pattern of sudden increase in traffic flow during the evening peak and sparse traffic flow in the early morning. The input data is sliced into sequences according to a time window (such as the past 1 hour), and the output is the future time period (such as the arrival time and duration of downstream traffic flow). The prediction targets include the adjustment time and the holding time. Adjustment time: the time required for the traffic flow to reach the downstream from the upstream, used to trigger the dimming of downstream street lights in advance. Holding time: the time when the traffic flow passes through the downstream area, determining the duration for which the street lights maintain high brightness. The upstream controller real-time collects the current traffic flow data (such as vehicle speed, lane occupancy rate), inputs it into the LSTM model, and predicts the arrival time of downstream traffic flow. The prediction result is transmitted to the downstream controller through LoRaWAN+5G dual-mode communication, and the downstream adjusts the brightness and color temperature of the street lights in advance.

[0035] The adjustment instruction is sent to the downstream centralized controller through the LoRaWAN and 5G dual-mode redundant communication protocol (transmission delay < 2 seconds, packet loss rate < 0.1%), enabling the downstream street lights to complete the pre-adjustment of brightness and color temperature before the traffic flow arrives.

[0036] The upstream central controller is generally located at the road section entrance or the starting area of the traffic flow. It monitors the traffic flow, pedestrian flow, average vehicle speed of the lane, and lane occupancy in real time in this area, and collects environmental parameters such as PM2.5, humidity, and SO2 concentration through sensors. It predicts the arrival time of the downstream traffic flow and generates adjustment instructions (brightness, color temperature, adjustment time). The downstream central controller is generally located in the adjacent road section or the area where the traffic flow is about to arrive. It receives the upstream instructions, calibrates the target parameters (such as brightness, color temperature) in real time, and executes the adjustment at the specified time. It adjusts the street lamp status in advance to ensure that the lighting conditions are better or optimal when the traffic flow arrives.

[0037] The upstream controller analyzes the historical traffic flow data through the LSTM model to predict the arrival time of the downstream traffic flow. Combining the current environmental parameters (such as haze concentration), it calculates the target brightness (such as 100%) and color temperature (such as 3000K warm yellow light) required downstream. It generates instructions containing the adjustment time (when to adjust), the holding time (how long to last), and the brightness / color temperature value. It sends them to the downstream controller through LoRaWAN+5G dual-mode communication to ensure that the instructions are delivered quickly (delay < 2 seconds) and reliably (packet loss rate < 0.1%). After receiving the instructions, the downstream controller uses the Kalman filtering algorithm to eliminate sensor errors and calibrate the target brightness / color temperature.

[0038] In a preferred embodiment, the calculation method of the dynamic weight correction factor in the combined weighting algorithm is as follows: ; Where: is the objective weight calculated by the entropy weight method; is the absolute deviation between the current parameter value and the historical mean, and the historical mean is the mean value in the same time period in the past 7 days; is the maximum allowable fluctuation range of the parameter, which is set according to national standards; The dynamically corrected comprehensive weight is: Where, is the th comprehensive weight of the parameter, indicating the final influence of this parameter in the brightness or color temperature adjustment model; is the weight obtained by the analytic hierarchy process, represents the th parameter, is the total number of parameters input in the model, and the specific value depends on the adjustment target. Brightness adjustment model: (traffic flow, pedestrian flow, PM2.5 concentration, humidity, ambient illuminance); Color temperature adjustment model: (PM2.5 concentration, SO2 concentration, humidity). The dynamically adjusted weight ( ) and the subjective weight of the analytic hierarchy process ( ), combine and normalize to generate the final comprehensive weight ( ). The dynamic weight correction factor is used to adjust the influence of each environmental parameter in the brightness / color temperature adjustment model in real time, enabling the system to quickly respond to sudden environmental or traffic changes and optimize the sensitivity and stability of the adjustment.

[0039] In a preferred embodiment, the brightness adjustment levels are divided as follows: the brightness judgment value interval [0, 1] is divided into an adaptive Gaussian interval based on the standard deviation of historical traffic flow. The specific formula is: interval , where is the central value of the th brightness level, taking values of 25%, 50%, 75%, and 100% of the base brightness; is the interval standard deviation, which is dynamically adjusted according to the standard deviation of the traffic flow in the past 1 hour.

[0040] The division of the brightness adjustment levels is dynamically adjusted based on the fluctuation of historical traffic flow. The system divides the brightness interval [0, 1] into four levels (25%, 50%, 75%, 100%), and the central value of each level corresponds to the base brightness. By statistically calculating the standard deviation of the traffic flow in the past 1 hour in real time, the interval range of each brightness level is dynamically calculated: if the traffic flow fluctuates greatly (large standard deviation), the interval range is automatically widened to avoid frequent adjustment; if the traffic flow is stable (small standard deviation), the interval is narrowed and the brightness response is more sensitive. This adaptive mechanism can intelligently balance energy consumption and lighting requirements according to the actual traffic flow changes on the road, ensuring driving safety and reducing energy waste.

[0041] In a preferred embodiment, in the brightness adjustment model, the inverse correlation between the ambient illuminance and the brightness is realized through a piecewise function: ; where is the base brightness (unit: %); is the ambient illuminance (unit: lux), and the coefficients 0.03 and 0.01 implicitly have the unit of dimensionless proportional coefficient / lux, or directly substitute the value for calculation, is the lowest brightness threshold (default 20%) to avoid negative values or excessive dimming.

[0042] In this embodiment, the brightness of the street lamp is dynamically adjusted according to the ambient illuminance. The stronger the ambient light, the lower the demand for artificial lighting. The ambient illuminance is divided into two intervals, and different adjustment coefficients are used to achieve brightness control. When the ambient illuminance is below 20 lux, the brightness of the street lamp decreases rapidly as the ambient illuminance increases. At this time, for every 1 lux increase in illuminance, the basic brightness is reduced by 3%. When the ambient illuminance reaches or exceeds 20 lux, the rate of brightness decrease slows down, and it is only reduced by 1% per lux. This not only ensures a sensitive response to brightness changes in low-illuminance environments but also avoids frequent fluctuations caused by excessive adjustment in high-illuminance situations. The difference between the coefficients 0.03 and 0.01 reflects the energy-saving strategies under different lighting conditions. At night, the ambient illuminance is usually below 20 lux, and at this time, the street lamp needs to undertake the main lighting function, and the brightness is finely adjusted according to the changes of weak light sources such as moonlight and car lights. In the daytime or strong light environment, natural light can already meet the basic needs, and narrowing the rate of brightness decrease can avoid the impact on the lamp life caused by frequent start-stop.

[0043] In a preferred embodiment, the division method of the color temperature adjustment level is as follows: the color temperature judgment value interval [0, 1] is divided into three sub-intervals, which respectively correspond to the dynamic color temperature values based on the haze index: ; Among them, is the color temperature judgment value, and the calculation formula is: ; : The normalized value of PM2.5. When the PM2.5 concentration ≤ 300 μg / m 3 , ; When the PM2.5 concentration > 300 μg / m 3 , ; : The normalized value of SO2, ; : The normalized value of humidity, ; is the combined weight, satisfying , and the calculation formula is: ; The objective weight of the th parameter calculated by the entropy weight method (based on parameter volatility); : The subjective weight of the th parameter determined by the analytic hierarchy process (expert experience).

[0044] Calculation of the objective weight by the entropy weight method: (brightness model), The entropy value of the th environmental parameter related to illumination brightness ( ), corresponding to 5 parameters of traffic flow, pedestrian flow, humidity, PM2.5, and ambient illuminance.

[0045] (Color temperature model), The entropy value of the th environmental parameter related to illumination color temperature ( ), corresponding to 3 parameters of PM2.5, SO2, and humidity.

[0046] The brightness model and the color temperature model are basically the same, except for the number of parameters and the meanings of the parameters. The present invention also introduces a dynamic weight correction factor. Taking PM2.5 as an example only, the same applies to other parameters.

[0047] are the objective weights of the PM2.5 parameter in the brightness model and the color temperature model, corresponding to (Brightness model) and (Color temperature model).

[0048] ; : The original objective weight of PM2.5 calculated by the entropy weight method; : The absolute deviation of the current PM2.5 concentration from the historical mean; , the maximum allowable fluctuation range of PM2.5, such as .

[0049] When the PM2.5 concentration deviation , is amplified by the dynamic correction factor and updated to reflect the real-time volatility of the pollution concentration. The corrected weights need to be renormalized to ensure that the sum of all parameter weights is 1, , where only the weight of PM2.5 is corrected, and the weights of other parameters remain unchanged.

[0050] In a preferred embodiment, the remote communication adopts the LoRaWAN and 5G dual-mode redundant transmission protocol, with a transmission delay of less than 2 seconds and a packet loss rate of less than 0.1%.

[0051] In a preferred embodiment, in the collaborative control strategy, the adjustment time of the downstream street lights is determined by an improved fuzzy PID control algorithm, specifically: Based on the average lane speed (unit: km / h) and the lane occupancy rate (% by unit) is taken as the input, and the congestion index is added Membership function of wherein is the maximum designed speed of the road (unit: km / h); Output adjustment time The membership function of adopts a dynamic asymmetric Gaussian distribution: ; wherein is the expected value of the adjustment time calculated according to historical data, Standard deviation, the fluctuation range of the adjustment time.

[0052] The determination of the downstream street lamp adjustment time is based on an improved fuzzy PID algorithm, and precise control is realized by dynamically perceiving the traffic state. Input parameter collection: The average vehicle speed (v) and lane occupancy rate (p) of the lane are obtained in real time. The vehicle speed reflects the moving efficiency of the traffic flow, and the occupancy rate reflects the congestion degree of the road. When the vehicle speed is 30 km / h and the occupancy rate is 80%, it indicates that the section is seriously congested. Congestion index calculation: The congestion index is generated by combining the vehicle speed and the occupancy rate , and this index comprehensively quantifies the congestion degree.

[0053] The fuzzy PID dynamic adjustment includes fuzzy control and PID correction. Fuzzy control: The vehicle speed and the occupancy rate are mapped into fuzzy rules (such as slow vehicle speed + high occupancy rate corresponding to an extended adjustment time), and the preliminary adjustment time is generated. PID correction: The deviation between the predicted vehicle speed (output of the LSTM model) and the actual vehicle speed is introduced, and the proportional coefficient is dynamically adjusted . The greater the deviation, The higher the weight, the faster the adjustment time response. The initial value is generated by the fuzzy rule, and the dynamic Gaussian distribution adjusts the fluctuation range according to the congestion degree, The correction of further reduces the error, forming a closed-loop control logic of fuzzy decision - PID correction - dynamic distribution. Fuzzy control main body: Processes non-linear inputs (vehicle speed v, lane occupancy rate p, congestion index ), and generates the preliminary adjustment time through fuzzy rules. For example, in the case of high congestion (such as > 0.6), the fuzzy rule directly extends to cope with the stagnation of the traffic flow. If the vehicle speed is low (such as 30 km / h) and the lane occupancy rate is high (such as 80%), then the output = 100 seconds (high congestion scenario). Fuzzy rule base: Specifies the downstream street lamp adjustment time The fuzzy rule base is based on three input variables: the average lane speed v, which is divided into 5 fuzzy sets (slow, relatively slow, medium, relatively fast, fast); the lane occupancy p, which is divided into 5 fuzzy sets (low, relatively low, medium, relatively high, high); and the congestion index , which is divided into 3 fuzzy sets (low, medium, high). By combining the fuzzy sets of the three input variables, the rule base is extended to 36 IF-THEN rules, and the specific rules are optimized and generated through expert experience and historical data. For example, for the input: v = slow, p = high, = high, the output is: = long (corresponding to the expected value of the adjustment time = 180 s , σ = 0.1 ). The fuzzy rule base is pre-determined, like a database, and is called by the program. The present invention only retains the dynamic adjustment of the proportional term , and the integral and differential actions are replaced by fuzzy rules. When the prediction error is large (such as sudden congestion or sensor noise), increases, enhancing the proportional action and quickly correcting the adjustment time of the fuzzy output , reducing the delay. It can be expressed as , where is the preliminary output according to the fuzzy rules or the fuzzy controller, is the error or prediction error, and is the preliminary adjustment time of the fuzzy controller output and the (historical) actual required time . Among them, represents the th adjustment operation (historical operation). The system statistically calculates the error mean of the last adjustment operations through the sliding window method: , which is set according to the actual requirements (e.g., can be an integer between 5 and 20, such as 10) to balance real-time performance and stability. The final value of the adjustment time is output through a dynamic asymmetric Gaussian distribution, and its mean is calculated from historical data. Through the above analysis, the membership function of the output adjustment time using a dynamic asymmetric Gaussian distribution can be changed to: ; is the preliminary value generated by the fuzzy control; is adjusted according to the prediction error to generate ; Gaussian distribution constraint: with as the mean and the standard deviation dynamically adjusted to output the final , 。

[0054] Generally speaking, the present invention includes a fuzzy control stage and a PID correction stage. Fuzzy control stage: The membership function is used to fuzzify the vehicle speed v and the lane occupancy p into linguistic variables (such as "slow + high occupancy"). Combining with fuzzy rules (such as "slow vehicle speed and high occupancy are converted into a long adjustment time"), a preliminary fuzzy output is generated. PID correction stage: (proportional coefficient) is dynamically adjusted through the prediction error. When the prediction error is large (such as sudden congestion), increases, forcing the shortening of fuzzy output to accelerate the response. Membership function: Determines the range of the fuzzy rule output (such as fluctuation range). Adjustment: Based on the error, correct the deviation of the fuzzy output to ensure that quickly converges to the actual required value. The membership function defines the basis of the fuzzy rule, provides dynamic correction, and the two together achieve closed-loop control of coarse adjustment (fuzzy) + fine adjustment (PID).

[0055] In a preferred embodiment, in the improved fuzzy PID control algorithm, the formula for dynamically adjusting the proportional coefficient is: ; Wherein, is the initial proportional coefficient; is the vehicle speed predicted by the LSTM model (unit: km / h); is the actually detected vehicle speed (unit: km / h). Let , the PID output , ; , where is initial value, calculated based on historical traffic flow data. Here, is equivalent to in the foregoing embodiment, such as the distance d1 from the upstream central controller to the downstream central controller, and the average speed of the traffic flow passing through this section in the past period of time (such as 1 hour), .

[0056] In a preferred embodiment, after receiving the adjustment instruction, if the deviation between the actual traffic flow and the predicted value exceeds 20%, the downstream central controller re-predicts the target brightness / color temperature value through the Kalman filter algorithm: State equation: ; Observation equation: ; where is the predicted state vector of the system at time instant, is the state vector, including traffic flow and vehicle speed; , is the traffic flow, is the average vehicle speed; is the system matrix, such as , is the sampling period, is the coefficient of traffic flow varying with vehicle speed (empirical value); , where is the influence of brightness adjustment on traffic flow, is the influence of color temperature adjustment on vehicle speed; it means that brightness adjustment only inhibits vehicle speed. , is the observation matrix. is the control input, , and represent the brightness adjustment amount and the color temperature adjustment amount respectively. are the process noise and the observation noise. The process noise describes the unmodeled dynamic disturbances in the state equation (such as sudden congestion, weather changes, etc.), which follows a zero-mean Gaussian distribution with the covariance matrix , , : the process noise of traffic flow, and the variance is characterized by in the . The process noise of vehicle speed, and the variance is characterized by in the . The observation noise describes the sensor measurement errors (such as traffic flow counting deviation, vehicle speed estimation noise), which follows a zero-mean Gaussian distribution with the covariance matrix , , The observation noise of traffic flow, and the variance is characterized by in the . The observation noise of vehicle speed, and the variance is characterized by in the . is the actual observation vector at time instant.

[0057] When the deviation between the actual traffic flow and the predicted value exceeds 20%, the downstream central controller will start the Kalman filtering algorithm to recalibrate the target brightness and color temperature. The core of this algorithm is to continuously optimize the adjustment parameters by dynamically balancing the predicted and measured data. The system first predicts the future state based on the historical traffic flow model, such as estimating the road load in the next time period according to the current traffic flow and vehicle speed. This prediction process takes into account the natural fluctuations of traffic changes (such as the regularity of morning and evening rush hours) and random interference factors (such as sudden accidents or weather impacts).

[0058] In a preferred embodiment, it further includes an abnormal data processing step: when the PM2.5 concentration exceeds 300 μg / m 3 , start the haze penetration mode and force the setting of , directly trigger , lock the color temperature at 3000K, or force the color temperature to be adjusted to: ; And increase the brightness to 120% of the rated value and continue until the PM2.5 drops below 200 μg / m 3 , and execute after verifying the adjustment effect through the digital twin platform.

[0059] An urban intelligent street lamp control system (see Figure 2 ), includes: A data acquisition module for real-time acquisition of road environment parameters and traffic status parameters; A dynamic weight calculation module, deployed in the upstream central controller, includes: A brightness weight unit, integrating the exponential smoothing filtering algorithm for PM2.5 concentration, and the filtering formula is: ; Among them, is the historical mean value (unit: μg / m 3 ); is the current value, is the PM2.5 concentration value processed by the exponential smoothing filtering algorithm; Calculate the objective weights of traffic flow, pedestrian flow, humidity, PM2.5, and ambient illuminance through the entropy weight method; calculate the subjective weights of the above parameters through the analytic hierarchy process; generate a combined weight by combining weights. In addition, a dynamic correction factor is introduced to adjust the weights and perform normalization processing.

[0060] A color temperature weight unit, integrating the wind speed compensation model of the concentration, and the compensation formula is: Among them, is the current wind speed; Calculate the objective weights of PM2.5, SO2, and humidity by the entropy weight method; calculate the subjective weights of the above parameters by the analytic hierarchy process; combine the weights to generate the comprehensive weight. It should be noted that the combined weighting algorithm, dynamic correction factor, parameter normalization, etc. are all in the dynamic weight calculation module. For specific details, please refer to the description of the method embodiments. This part of the method and system corresponds to each other, and only a brief introduction is given here.

[0061] The adjustment decision module is used to generate a brightness level instruction and a color temperature level instruction according to the brightness judgment value and the color temperature judgment value; normalize parameters such as traffic flow and humidity, and generate the brightness judgment value and the color temperature judgment value through weighting, and divide the brightness levels (25%, 50%, 75%, 100%) and color temperature levels (3000K, 4000K, 5000K). For flexible division and level adjustment, please refer to the method embodiments, and this part of the content will not be elaborated here.

[0062] The collaborative control module includes: The upstream controller is configured to run a traffic flow prediction model based on LSTM and output the downstream adjustment time, holding time, and target brightness / color temperature value; The downstream controller is configured to calibrate the target brightness / color temperature value in real time through the Kalman filter algorithm and perform adjustment when the prediction time arrives.

[0063] The urban intelligent street lamp control system realizes intelligent adjustment through the collaborative work of multiple modules. The system first obtains road environment parameters and traffic status data in real time by the data acquisition module, including information such as PM2.5 concentration, wind speed, and traffic flow. The dynamic weight calculation module optimizes the original data. For example, the exponential smoothing algorithm is used to eliminate the instantaneous fluctuation of the PM2.5 concentration monitoring value, making it closer to the real pollution level; for the sulfur dioxide concentration monitoring value, the system introduces a wind speed compensation mechanism to dynamically correct the influence of gas diffusion on the detection accuracy through the wind speed.

[0064] The adjustment decision module compares the processed environmental parameters with the preset standard values and generates corresponding brightness and color temperature adjustment instructions. For example, the brightness is automatically increased in haze weather to enhance the penetration, or the color temperature is increased during the peak traffic period to enhance the lighting effect. The system is specially designed with a collaborative control mechanism. The upstream controller predicts the traffic flow trend within the next 15 minutes through a deep learning model and calculates in advance the lighting parameter adjustment time node and duration required for the downstream section.

[0065] After receiving the prediction instruction, the downstream controller combines the real-time traffic flow speed change and uses the dynamic calibration algorithm to eliminate the prediction error. When the vehicle is about to enter the controlled area, the street lamp will complete the brightness and color temperature adjustment 3 - 5 seconds in advance to ensure that the lighting parameters are in the best state when the traffic flow passes, and complete the closed-loop control of prediction - calibration - execution.

[0066] In a preferred embodiment, a fuzzy PID hybrid controller is integrated in the collaborative control module, and its fuzzy rule base is extended to 36 IF-THEN rules, and the congestion index is added as a third input variable.

[0067] The congestion index can be calculated by collecting real-time data such as vehicle flow and pedestrian flow through sensors, and reflects the degree of road congestion. The fuzzy rule base contains 36 IF-THEN logic rules, which define control strategies under different input combinations. For example, when a large brightness error and a high congestion index are detected, the system will trigger the rule of "rapidly increasing brightness"; if the error is small and the congestion index is low, "smooth fine-tuning" will be executed to reduce energy consumption. During the actual control process, the controller continuously receives environmental data, calculates the PID parameters through fuzzy inference, and then combines the PID algorithm to output the final brightness or color temperature adjustment instruction. This hybrid control mode not only retains the stability of PID, but also uses fuzzy logic to handle complex and changeable road scenarios.

[0068] In a preferred embodiment, the communication unit adopts a multi-hop verification mechanism based on blockchain, specifically: a hash value is generated before the instruction transmission and recorded in the distributed ledger; after the downstream controller receives it, it verifies the hash consistency, and if it is inconsistent, it triggers a retransmission. The communication unit of the intelligent street lamp system adopts a multi-hop verification mechanism based on blockchain to ensure the reliability and security of instruction transmission. When the upstream controller generates an adjustment instruction (such as brightness and color temperature parameters), the system will generate a unique digital fingerprint (hash value) for the instruction and synchronously record the fingerprint in the distributed ledger of the blockchain network. The distributed ledger is jointly maintained by multiple nodes, and each node stores a complete hash record to ensure that the data cannot be tampered with and is transparent across the network. After receiving the instruction, the downstream controller will immediately recalculate the hash value of the instruction content and compare it with the original hash stored in the blockchain ledger. If the two are consistent, it means that the instruction has not been tampered with during the transmission process, and the downstream controller executes the adjustment operation; if they are inconsistent, it indicates that data damage or malicious attack may have occurred during the transmission process, and the system automatically triggers the retransmission mechanism, requiring the upstream controller to resend the instruction until the verification passes. The entire process does not require manual intervention. Through the distributed verification feature of blockchain, it not only avoids the risk of single-point failure, but also realizes an efficient error self-repairing ability.

[0069] In a preferred embodiment, it further includes: an adaptive compensation module. When the environmental parameters exceed the safety threshold, it automatically switches to the reinforcement learning model to generate adjustment instructions, and after virtual verification through the digital twin platform, it is executed. The adaptive compensation module is a safety mechanism of the intelligent street lamp system, which can ensure the reliability and safety of lighting control under extreme environmental conditions. When environmental parameters such as PM2.5 concentration and wind speed exceed the preset safety threshold, the system automatically triggers an exception handling process. The module switches to the reinforcement learning model, which dynamically generates brightness and color temperature adjustment instructions based on historical environmental data and adjustment effects. If the PM2.5 concentration suddenly soars to 300 μg / m 3 , the model will first increase the brightness to 90% and reduce the color temperature to 3000K to enhance the fog penetration. The instructions are sent to the digital twin platform for virtual verification. The platform simulates the light distribution, energy consumption, and impact on traffic after adjustment by constructing a virtual mirror of the road lighting system, avoiding the risks that may be caused by directly operating physical devices. After the verification passes, the instructions are sent to the physical street lamps for execution, and the adjustment effects are recorded to optimize subsequent decisions.

[0070] The urban intelligent street lamp control system described in the present invention realizes intelligent regulation through the cooperation of multiple modules. The system obtains environmental parameters (PM2.5, humidity, SO2 concentration, wind speed) and traffic parameters (traffic flow, vehicle speed, lane occupancy rate) in real time through the data acquisition module. The dynamic weight calculation module filters and compensates the data (such as PM2.5 exponential smoothing filtering and SO2 wind speed compensation) to generate denoised parameter values. The adjustment decision module dynamically assigns the weights of each parameter based on the combined weighting algorithm (entropy weight method + analytic hierarchy process), outputs the brightness and color temperature judgment values, and divides the adaptive adjustment levels according to the historical standard deviation (such as four levels for brightness and three levels for color temperature). The cooperative control module anticipates the downstream adjustment requirements through the LSTM traffic flow prediction model of the upstream centralized controller, generates the target brightness / color temperature values and time parameters, and combines the Kalman filter downstream to calibrate the error in real time to achieve a closed loop of prediction - instruction - execution; the communication unit uses blockchain multi-hop verification to ensure the security of instruction transmission, and the adaptive compensation module switches to the reinforcement learning model when the environment exceeds the limit, and executes the emergency strategy after virtual verification through the digital twin.

[0071] Although the preferred embodiments of the present invention have been described in detail, those skilled in the art may still make further adjustments and changes to these embodiments after mastering the basic innovative concepts. Therefore, the appended claims are intended to cover the preferred embodiments as well as all changes and modifications that fall within the scope of the present invention. The above content is only a description of the preferred embodiments of the present invention and is not intended to limit the scope of the present invention. It should be clear that any modification, equivalent replacement, or improvement made within the spirit and principle framework of the present invention should be regarded as being included within the protection scope of the present invention.

Claims

1. A method for controlling intelligent street lamps in a city, characterized in that, It includes the following: S1. Collect road environmental parameters and traffic state parameters. The environmental parameters include PM2.5 concentration, humidity, ambient illuminance, and SO2 concentration. The traffic state parameters include traffic flow, pedestrian flow, average lane speed, and lane occupancy. S2. Construct a brightness adjustment model and a color temperature adjustment model, and calculate the weights of each parameter through a combined weighting algorithm respectively: The brightness adjustment model takes traffic flow, pedestrian flow, PM2.5 concentration, humidity, and ambient illuminance as input parameters, combines the entropy weight method and the analytic hierarchy process for weighting, and introduces a dynamic weight correction factor to generate a brightness judgment value; The color temperature adjustment model takes PM2.5 concentration, SO2 concentration, and humidity as input parameters, combines the entropy weight method and the analytic hierarchy process for weighting, and fuses real-time wind speed data to perform diffusion compensation on the SO2 concentration to generate a color temperature judgment value. S3. Based on the brightness judgment value and the color temperature judgment value, divide the brightness adjustment level and the color temperature adjustment level respectively, and dynamically adjust the brightness and color temperature of the street lamp.

2. The urban intelligent street lamp control method according to claim 1, characterized in that, It also includes: S4. Realize the linkage of street lamps on upstream and downstream sections through a collaborative control strategy: The upstream centralized controller calculates the adjustment time, holding time, and target brightness / color temperature value of the downstream street lamp according to the detected traffic flow, pedestrian flow, average lane speed, and lane occupancy, in combination with the dynamic traffic flow prediction model based on LSTM. Send the adjustment instruction to the downstream centralized controller through the LoRaWAN and 5G dual-mode redundant communication protocol, so that the downstream street lamp completes the pre-adjustment of brightness and color temperature before the traffic flow arrives.

3. A method for controlling an intelligent street lamp in a city according to claim 1, characterized in that, The calculation method of the dynamic weight correction factor in the combined weighting algorithm is: ； Where: is the objective weight calculated by the entropy weight method; is the absolute deviation between the current parameter value and the historical mean, where the historical mean is the mean value for the same time period in the past 7 days; is the maximum allowable fluctuation range of the parameter; The dynamically corrected comprehensive weight is as follows: Among them, is the comprehensive weight of the -th parameter, indicating the final influence of this parameter in the brightness or color temperature adjustment model; is the weight obtained by the analytic hierarchy process, indicating the -th parameter, and is the total number of parameters input in the model.

4. A method for controlling an intelligent street lamp in a city according to claim 1, characterized in that, The method for dividing the brightness adjustment levels is as follows: The brightness judgment value interval [0, 1] is divided into an adaptive Gaussian interval based on the standard deviation of historical traffic flow. The specific formula is: interval , where is the central value of the th brightness level, and the values are 25%, 50%, 75%, and 100% of the base brightness; is the standard deviation of the interval, which is dynamically adjusted according to the standard deviation of the traffic flow within the past 1 hour.

5. A method for controlling an intelligent street lamp in a city according to claim 1, characterized in that, In the brightness adjustment model, the inverse correlation relationship between ambient illuminance and brightness is realized through a piecewise function: ； Among them, is the base brightness; is the ambient illuminance, is the lowest brightness threshold.

6. The control method of an urban intelligent street lamp according to claim 1, characterized in that, The division method of the color temperature adjustment level is: Divide the color temperature judgment value interval [0,1] into three sub-intervals, corresponding to the dynamic color temperature values based on the haze index respectively: ； Among them, is the color temperature judgment value, and the calculation formula is: ； : PM2.5 normalization value. When the PM2.5 concentration ≤ 300 μg / m 3 , ; When the PM2.5 concentration > 300 μg / m 3 , ; : SO2 normalization value, ; : Humidity normalization value, ; is the combined weight, satisfying , and the calculation formula is: ； is the objective weight of the th parameter calculated by the entropy weight method; is the subjective weight of the th parameter determined by the analytic hierarchy process.

7. A method for controlling an intelligent street lamp in a city according to claim 2, wherein, In the collaborative control strategy, the adjustment time of the downstream street lamp is determined by an improved fuzzy PID control algorithm, specifically: Using the average lane speed and the lane occupancy rate as inputs, a membership function for the newly added congestion index is as follows: Among them, is the maximum designed road speed; Output adjustment time The membership function of ； Among them, is the expected value of the adjustment time calculated based on historical data, is the standard deviation.

8. A method for controlling an urban intelligent street lamp according to claim 7, characterized in that, In the improved fuzzy PID control algorithm, the proportional coefficient is dynamically adjusted The formula is as follows: ； Among them, is the initial proportionality coefficient; is the vehicle speed predicted by the LSTM model; is the actually detected vehicle speed.

9. A method for controlling an urban intelligent street lamp according to claim 1, characterized in that, It also includes an abnormal data processing step: when the PM2.5 concentration exceeds 300 μg / m 3 , the haze penetration mode is activated, and is forcibly set, directly triggering . The color temperature is locked at 3000K, and the brightness is increased to 120% of the rated value and lasts until the PM2.5 drops below 200 μg / m 3 .

10. An urban intelligent street lamp control system, characterized in that, It includes: A data acquisition module for obtaining road environmental parameters and traffic state parameters in real time. A dynamic weight calculation module deployed in the centralized controller, including: A brightness weight unit integrating the exponential smoothing filtering algorithm of PM2.5 concentration, and the filtering formula is: ； Among them, is the historical mean; is the current value, is the PM2.5 concentration value processed by the exponential smoothing filtering algorithm; Color temperature weighting unit, integrated Wind speed compensation model for concentration, and the compensation formula is: Among them, is the current wind speed; An adjustment decision module for generating a brightness level instruction and a color temperature level instruction according to the brightness judgment value and the color temperature judgment value. A collaborative control module, including: An upstream controller configured to run a traffic flow prediction model based on LSTM and output the downstream adjustment time, holding time, and target brightness / color temperature value. A downstream controller configured to calibrate the target brightness / color temperature value in real time through the Kalman filtering algorithm and perform the adjustment when the prediction time arrives.

Images