A Method and System for Adaptive Lighting Control of Urban Roads Based on Multi-Source Urban Data

Abstract

This invention discloses an adaptive lighting control method and system for urban roads based on multi-source urban data. The method includes: acquiring fixed road parameters and multi-source real-time operational data of the target road; preprocessing the acquired multi-source real-time operational data; calculating a fixed characteristic quantity φ based on the fixed road parameters and calling a basic illuminance interval; identifying road functional attributes and regional functional attributes to obtain a road attribute-corrected illuminance interval; dynamically adjusting the road attribute-corrected illuminance interval based on the preprocessed multi-source real-time operational data to obtain a target illuminance value E and calculate a target luminous flux F; generating a lighting control command based on the target luminous flux F and sending the lighting control command to the lighting execution device. The method disclosed in this invention can effectively adapt to road lighting in different urban environments, and has advantages such as good lighting safety and energy saving.

Description

Technical Field

[0001] This invention relates to the field of urban road lighting control technology, and in particular to an adaptive lighting control method and system for urban roads based on multi-source urban data. Background Technology

[0002] Urban road lighting systems are an important component of urban infrastructure. The intensity of road lighting directly affects nighttime road traffic safety, pedestrian visibility, the quality of public space use, and urban lighting energy consumption. With the development of LED luminaires, individual lamp controllers, sensors, image acquisition equipment, IoT communication, and data processing technologies, road lighting systems now have the technological foundation to dynamically adjust according to road operating conditions and external environmental conditions.

[0003] Existing urban road lighting control methods mainly include timed on / off control, ambient light control, traffic flow control, zoned brightness control, and remote lighting control based on individual lamp controllers. Some existing solutions can collect data such as ambient brightness, vehicle flow, and pedestrian flow through illuminance sensors, cameras, radar, or traffic management platforms, and adjust street light brightness according to real-time road conditions. These solutions reduce lighting energy consumption in unoccupied or low-activity road sections to some extent and also improve the automation level of urban road lighting systems.

[0004] However, existing technologies still have the following shortcomings: First, some lighting control schemes mainly adjust based on time, ambient illuminance, or a single traffic flow variable, making it difficult to simultaneously consider fixed physical conditions of the road, such as road grade, road width, lamp installation height, lamp spacing, lamp arrangement, and road surface material. Second, some schemes mainly adjust based on the current road conditions in real time, lacking a layer-by-layer coupling of road lighting standard constraints, road functional attributes, regional functional attributes, and real-time operating status. Third, although some schemes introduce algorithm models, they lack an interpretable technical chain for road lighting engineering formulas and target luminous flux calculation processes, making it difficult to directly convert algorithm results into dimming commands that can be executed at the lamp end. Fourth, some schemes lack feedback correction mechanisms based on measured illuminance and historical operating data after execution, making it difficult to adapt to changes in lighting needs under different roads, different time periods, and different intensities of pedestrian and vehicle activity.

[0005] Therefore, there is a need for an adaptive lighting control method and system for urban roads that can, under the premise of meeting the constraints of road lighting standards, comprehensively consider fixed road parameters, road functional attributes, regional functional attributes, vehicle flow, pedestrian flow, ambient light, accident status, time information, and historical measured illuminance data, and further convert the target illuminance into target luminous flux and executable dimming commands. Summary of the Invention

[0006] This invention addresses the problem that existing urban road lighting control methods struggle to generate target illuminance and target luminous flux while simultaneously meeting road lighting standards, and incorporating fixed physical attributes, functional attributes, regional functional attributes, real-time traffic conditions, pedestrian activity, ambient light conditions, and historical feedback data. This results in insufficient adaptability, a single adjustment basis, and difficulties in coordinating lighting safety and energy-saving effects in road lighting control. Therefore, this invention proposes an adaptive urban road lighting control method based on multi-source urban data.

[0007] The technical means employed in this invention are as follows: An adaptive lighting control method for urban roads based on multi-source urban data includes the following steps: Step 1: Obtain the fixed road parameters of the target road; Step 2: Obtain multi-source real-time operational data for the target road; Step 3: Perform timestamp alignment, outlier removal, data smoothing, and normalization on the acquired multi-source real-time running data to obtain preprocessed multi-source real-time running data; Step 4: Calculate the fixed feature quantity φ based on the obtained fixed road parameters of the target road; Step 5: Based on the road grade parameters in the fixed road parameters of the target road, retrieve the corresponding basic illuminance range from the preset road lighting standard library; Step 6: Identify the road function attributes and area function attributes of the target road, and correct the upper and / or lower limits of the basic illuminance range based on the identified road function attributes and area function attributes to obtain the road attribute corrected illuminance range. Step 7: Based on the acquired preprocessed multi-source real-time operating data, dynamically adjust the road attribute correction illuminance range to obtain the target illuminance value E; Step 8: Calculate the target luminous flux F based on the obtained target illuminance value E and the fixed characteristic quantity φ; Step 9: Generate a lighting control command based on the calculated target luminous flux F, and send the lighting control command to the lighting actuator; Step 10: Collect the measured illuminance data of the target road after the lighting control is implemented, and update the dynamic adjustment weights based on the historical measured illuminance data and historical operation data.

[0008] Furthermore, the fixed road parameters include road grade, road width W, lamp installation height H, lamp spacing S, lamp arrangement N, road surface material, lamp maintenance factor k, and lamp utilization factor U.

[0009] Furthermore, the multi-source real-time operational data includes vehicle traffic flow, pedestrian traffic flow, ambient light intensity, accident status, time information, weather information, and visibility information.

[0010] Furthermore, the formula for calculating the fixed characteristic quantity φ is as follows: φ = kSW / NU(1) The fixed characteristic quantity φ is determined based on the road width W, the lamp spacing S, the lamp arrangement N, the lamp maintenance factor k, and the lamp utilization factor U.

[0011] Furthermore, the formula for calculating the target illuminance value E is as follows: E = Emin + αV + βP + γL + δA + etaT + λQ + μR (2) Where Emin is the lower limit of the road attribute correction illuminance range; V is the normalized vehicle flow; P is the normalized pedestrian flow; L is the ambient light correction term; A is the accident status correction term; T is the time correction term; Q is the weather correction term; R is the visibility correction term; and α, β, γ, δ, η, λ, and μ are the weights of the corresponding dynamic factors.

[0012] Furthermore, the formula for calculating the target luminous flux F is as follows: F = E × φ(3) Where F is the target luminous flux; E is the target illuminance value; and φ is a fixed characteristic quantity.

[0013] An adaptive lighting control system for urban roads based on multi-source urban data includes: The fixed road parameter acquisition module is used to acquire the fixed road parameters of the target road. The real-time running data acquisition module is used to acquire multi-source real-time running data of the target road; The data preprocessing module is used to perform timestamp alignment, outlier removal, data smoothing, and normalization on the acquired multi-source real-time running data to obtain preprocessed multi-source real-time running data; A fixed feature quantity calculation module is used to calculate a fixed feature quantity φ based on the acquired fixed road parameters; The standard constraint module is used to call the corresponding basic illuminance zone from the preset road lighting standard library based on the road level parameters in the fixed road parameters of the target road. The road attribute recognition module is used to identify the road function attributes and area function attributes of the target road; The road attribute correction module is used to correct the upper and / or lower limits of the basic illuminance range based on the identified road functional attributes and regional functional attributes, so as to obtain the road attribute corrected illuminance range. The dynamic illuminance adjustment module is used to dynamically adjust the illuminance range of road attribute correction based on the acquired preprocessed multi-source real-time operating data to obtain the target illuminance value E. The target luminous flux calculation module is used to calculate the target luminous flux F based on the acquired target illuminance value E and the fixed characteristic quantity φ. The control output module is used to generate lighting control commands based on the calculated target luminous flux F, and send the lighting control commands to the lighting actuator. Additionally, a feedback optimization module is used to collect measured illuminance data of the target road after the execution of lighting control, and to update the dynamic adjustment weights based on historical measured illuminance data and historical operation data.

[0014] Furthermore, the fixed road parameters of the target road acquired by the fixed road parameter acquisition module include: road grade, road width W, lamp installation height H, lamp spacing S, lamp arrangement N, road surface material, lamp maintenance coefficient k, and lamp utilization coefficient U; The real-time operation data acquisition module obtains multi-source real-time operation data of the target road, including vehicle flow, pedestrian flow, ambient light intensity, accident status, time information, weather information, and visibility information.

[0015] Furthermore, the formula for calculating the fixed characteristic quantity φ is as follows: φ = kSW / NU(1) The fixed characteristic quantity φ is determined based on the road width W, the lamp spacing S, the lamp arrangement N, the lamp maintenance factor k, and the lamp utilization factor U. The formula for calculating the target luminous flux F is: F = E × φ(3) Where F is the target luminous flux; E is the target illuminance value; and φ is a fixed characteristic quantity.

[0016] Furthermore, the formula for calculating the target illuminance value E is as follows: E = Emin + αV + βP + γL + δA + etaT + λQ + μR (2) Where Emin is the lower limit of the road attribute correction illuminance range; V is the normalized vehicle flow; P is the normalized pedestrian flow; L is the ambient light correction term; A is the accident status correction term; T is the time correction term; Q is the weather correction term; R is the visibility correction term; and α, β, γ, δ, η, λ, and μ are the weights of the corresponding dynamic factors.

[0017] Compared with existing technologies, the urban road adaptive lighting control method based on multi-source urban data disclosed in this invention has the following beneficial effects: First, this invention determines the basic illuminance range from a preset road lighting standard library by using road grade parameters, and then performs road attribute correction and dynamic data adjustment, so that the target illuminance value is generated under the constraints of road lighting standards, reducing the risk of insufficient illuminance caused by simply adjusting the dimming based on real-time traffic flow.

[0018] Secondly, this invention uses a fixed characteristic quantity φ to characterize fixed conditions such as road width, luminaire installation height, luminaire spacing, luminaire arrangement, road surface material, maintenance factor, and utilization factor. This allows the same control method to adapt to different road conditions, improving the road adaptability of the lighting control model. In other words, this invention incorporates the differences in physical properties of different roads into the same control method, enabling roads with larger widths, larger luminaire spacing, or lower utilization factors to achieve correspondingly higher target luminous flux, avoiding inconsistent lighting effects caused by applying the same brightness level to different roads.

[0019] Third, this invention acquires vehicle traffic flow, pedestrian traffic flow, ambient light and visibility status through deep learning target detection, multi-target tracking, semantic segmentation, crowd density estimation, illuminance sensor acquisition and image brightness analysis, making the input data for road lighting control collectable and verifiable.

[0020] Fourth, this invention identifies road functional attributes and regional functional attributes through GIS overlay analysis, point of interest density identification, planning map layer identification, or street view image semantic recognition, enabling road lighting control to combine road space usage characteristics, rather than adjusting solely based on traffic flow.

[0021] Fifth, this invention generates target illuminance values ​​by combining safety constraint rules with machine learning regression models, enabling the system to dynamically respond to vehicle traffic, pedestrian traffic, ambient light, accident status, and time information while meeting the minimum lighting safety requirements.

[0022] Sixth, this invention further converts the target illuminance into a target luminous flux F, and generates brightness level instructions, dimming ratio instructions, PWM duty cycle control instructions, DALI dimming instructions, or single-lamp controller control instructions based on the target luminous flux F, so that the algorithm calculation results can be connected with the lamp control terminal, improving the feasibility of the technical solution.

[0023] Seventh, this invention establishes a feedback update mechanism through historical measured illuminance and historical operational data to update the dynamic adjustment weights of vehicle flow and pedestrian flow, enabling the system to gradually correct the dynamic adjustment model as the actual road usage changes, thereby improving the stability and interpretability of adaptive lighting control. Attached Figure Description

[0024] Figure 1This is a flowchart illustrating the adaptive lighting control method for urban roads based on multi-source urban data disclosed in this invention. Figure 2 This is a schematic diagram of the multi-source real-time running data acquisition and preprocessing process in this invention; Figure 3 This is a schematic diagram of the three-layer constraint calculation process in this invention; Figure 4 This is a schematic diagram illustrating the calculation process of the fixed characteristic quantity φ and the target luminous flux F in this invention; Figure 5 This is a schematic diagram illustrating the process of generating and executing lighting control commands in this invention; Figure 6 This is a schematic diagram of the feedback optimization process based on the least squares method or recursive least squares method in this invention; Figure 7 This is a structural block diagram of the urban road adaptive lighting control method system based on multi-source urban data disclosed in this invention; Figure 8 This is a schematic diagram showing the verification results of road parameters and illumination of the target road in an embodiment of the present invention. Detailed Implementation

[0025] Example 1 like Figure 1 , Figure 2 , Figure 3 , Figure 4 , Figure 5 and Figure 6 As shown, the urban road adaptive lighting control method based on multi-source urban data disclosed in this invention includes the following steps: Step 1: Obtain the fixed road parameters of the target road; Step 2: Obtain multi-source real-time operational data for the target road; Step 3: Perform timestamp alignment, outlier removal, data smoothing, and normalization on the acquired multi-source real-time running data to obtain preprocessed multi-source real-time running data; Step 4: Calculate the fixed feature quantity φ based on the obtained fixed road parameters of the target road; Step 5: Based on the road grade parameters in the fixed road parameters of the target road, retrieve the corresponding basic illuminance range from the preset road lighting standard library; Step 6: Identify the road function attributes and area function attributes of the target road, and correct the upper and / or lower limits of the basic illuminance range based on the identified road function attributes and area function attributes to obtain the road attribute corrected illuminance range. Step 7: Based on the acquired preprocessed multi-source real-time operating data, dynamically adjust the road attribute correction illuminance range to obtain the target illuminance value E; Step 8: Calculate the target luminous flux F based on the obtained target illuminance value E and the fixed characteristic quantity φ; Step 9: Generate a lighting control command based on the calculated target luminous flux F, and send the lighting control command to the lighting actuator; Step 10: Collect the measured illuminance data of the target road after the lighting control is implemented, and update the dynamic adjustment weights based on the historical measured illuminance data and historical operation data.

[0026] Specifically, in one optional embodiment of this application, the steps are as follows: Step 1 involves obtaining fixed road parameters for the target road, including road grade, road width W, lamp installation height H, lamp spacing S, lamp arrangement N, road surface material, lamp maintenance factor k, and lamp utilization factor U. These fixed road parameters can be obtained from a GIS spatial database, road design data, street light asset ledger, on-site surveying equipment, or mobile surveying equipment. For example, the road centerline, road boundaries, road grade, and street light locations are retrieved from the GIS spatial database; the lamp spacing S is calculated based on the spatial distance between adjacent street light locations or the road mileage; the road width W is calculated based on the vertical distance between road boundary lines or road cross-section data; and the lamp installation height H, lamp arrangement N, lamp maintenance factor k, and lamp utilization factor U are obtained from the street light asset ledger.

[0027] like Figure 2 As shown, step 2 involves acquiring multi-source real-time operational data for the target road, including vehicle traffic flow, pedestrian traffic flow, ambient light intensity, accident status, time information, weather information, and visibility information. The methods for acquiring various real-time operational data are as follows: Vehicle flow is obtained through road video image recognition. Specifically, the system identifies vehicle targets in video frames using a convolutional neural network-based target detection model, and uses a multi-target tracking algorithm to associate the same vehicle target in consecutive video frames to form vehicle trajectories. When a vehicle trajectory crosses a preset virtual detection line or enters a preset detection area, the system generates a vehicle passage event and counts the number of vehicle passage events per unit time as the vehicle flow.

[0028] Pedestrian flow is obtained through one or more of the following methods: pedestrian target detection, multi-object tracking, semantic segmentation, or crowd density estimation. Specifically, the system uses a pedestrian target detection model and a multi-object tracking algorithm to obtain pedestrian trajectories and calculates pedestrian flow based on the number of times the pedestrian trajectory crosses the detection line. When the target road is located in a commercial district, near a school, a transportation hub, or other densely populated areas, the system uses a semantic segmentation model or a crowd density estimation model to obtain the proportion of people or the crowd density integral within the detection area, and uses the proportion of people or the crowd density integral as a pedestrian activity index. That is, the system identifies pedestrian areas through a semantic segmentation model and calculates the area proportion of pedestrian areas within the detection area; or it generates a density heatmap through a crowd density estimation model and uses the integral value of the density heatmap within the detection area as a crowd activity index.

[0029] Ambient light intensity is collected by illuminance sensors installed on light poles, road edges, or around the target road. The system adds a timestamp to each illuminance data point and timestamps the illuminance data with vehicle traffic flow, pedestrian traffic flow, and control commands.

[0030] Visibility information is obtained through road image brightness analysis. The system acquires images of the target road, calculates one or more of the following: the mean grayscale value, the mean luminance component value, the proportion of low-brightness pixels, or the image contrast, and determines the environmental visibility of the target road based on the calculation results. When the environmental visibility is lower than a preset visibility threshold, the system generates a low visibility status marker.

[0031] Time information is obtained through one or more of the following methods: local system clock, network time service, lighting control platform timestamp, or city IoT platform timestamp. Specifically, the system obtains time information through the local system clock, network time service, lighting control platform timestamp, or city IoT platform timestamp.

[0032] Weather information is obtained through meteorological sensors near the target road, road environment monitoring equipment, urban meteorological data interfaces, traffic management platforms, or third-party meteorological service interfaces.

[0033] Accident status is obtained through video analysis or external traffic event data. The system continuously tracks vehicle targets. When a vehicle target's speed in a non-stop area is lower than a preset speed threshold and the duration exceeds a preset time threshold, the vehicle target is marked as a suspected abnormal target. When the number of suspected abnormal targets in the same road segment exceeds a preset number, or the vehicle queue length continues to increase and exceeds a preset length threshold, the system marks the target road accident status as abnormal. The system can also read accident, construction, road closure, or traffic control data from the traffic management platform and update the accident status accordingly.

[0034] like Figure 2As shown, step 3 involves performing timestamp alignment, outlier removal, data smoothing, and normalization on the acquired multi-source real-time operational data to obtain preprocessed multi-source real-time operational data. Specifically, the system timestamps vehicle traffic flow, pedestrian traffic flow, ambient light intensity, and measured illuminance data according to a preset sampling period; outliers are removed using thresholding, interquartile range, Z-score, or sliding window mean method; time series data are smoothed using moving average filtering, exponential smoothing, or Kalman filtering; vehicle traffic flow is normalized based on the target road's design capacity; pedestrian traffic flow is normalized based on historical high quantiles or maximum observed values; and ambient light intensity is normalized based on a preset illuminance threshold.

[0035] like Figure 4 As shown, in step 4, a fixed characteristic quantity φ is calculated based on the obtained fixed road parameters of the target road. The fixed characteristic quantity φ characterizes the combined influence of the road's physical conditions and luminaire arrangement on the lighting effect. Even if different roads have the same target illuminance, they may require different luminous flux outputs due to differences in road width, luminaire spacing, luminaire arrangement, luminaire maintenance status, and road surface material. Therefore, the fixed characteristic quantity φ needs to be calculated first. Specifically, in step 4, the fixed road parameters include road width W, luminaire installation height H, luminaire spacing S, luminaire arrangement N, luminaire maintenance coefficient k, and luminaire utilization coefficient U. The fixed characteristic quantity φ is calculated based on the road width W, luminaire spacing S, luminaire arrangement N, luminaire maintenance coefficient k, and luminaire utilization coefficient U. The fixed characteristic quantity φ is used to convert the target illuminance value into a target luminous flux F and to distinguish the differences in the required luminous flux output of different roads under the same target illuminance. The formula for calculating the fixed characteristic quantity φ is: φ = kSW / NU(1) Where k is the luminaire maintenance factor, S is the luminaire spacing, W is the road width, N is the luminaire arrangement parameter corresponding to the luminaire arrangement method, and U is the luminaire utilization factor. The above formula is an optional implementation method, and the fixed characteristic quantity φ can also be equivalently transformed according to the road lighting engineering formula.

[0036] like Figure 3As shown, in step 5, based on the road level parameters in the fixed road parameters of the target road, the system retrieves the corresponding basic illuminance interval from the preset road lighting standard library. The basic illuminance interval is used to define the lighting reference range of the target road under the corresponding road level. The specific process is as follows: The preset road lighting standard library uses road level as an index field and stores the corresponding lower limit of average illuminance, recommended average illuminance interval, illuminance uniformity requirements, and glare limitation requirements. The system retrieves the basic illuminance interval from the preset road lighting standard library based on the road level of the target road. The preset road lighting standard library includes basic illuminance intervals or basic illuminance thresholds corresponding to different road levels, such as arterial roads, secondary arterial roads, and local roads. The corresponding basic illuminance interval is retrieved from the preset road lighting standard library as the basis for subsequent road attribute correction and dynamic adjustment. For example, when the road level is an arterial road, the system retrieves the basic illuminance interval corresponding to the arterial road; when the road level is a secondary arterial road, the system retrieves the basic illuminance interval corresponding to the secondary arterial road; when the road level is a local road, the system retrieves the basic illuminance interval corresponding to the local road.

[0037] In step 6, the road function attributes and regional function attributes of the target road are identified, and the upper and / or lower limits of the basic illuminance interval are corrected based on the identified road function attributes and regional function attributes to obtain the road attribute corrected illuminance interval. The specific process is as follows: The road function attributes and regional function attributes are obtained through one or more of the following methods: GIS overlay analysis, point of interest density identification, planning map layer identification, or street view image semantic recognition, and a functional attribute label for the target road is generated. The system generates a road influence buffer based on the centerline of the target road, and counts the number and density of different types of points of interest within the road influence buffer. When the density of commercial service point of interest is higher than a preset commercial threshold, the system marks the regional function attribute of the target road as a commercial area. When the proportion of residential point of interest or residential map patches is higher than a preset residential threshold, the system marks the regional function attribute of the target road as a residential area. When the density of education, medical, and public service point of interest is higher than a preset public service threshold, the system marks the regional function attribute of the target road as an educational public area or a public service area.

[0038] Road attribute correction employs either a rule engine or fuzzy inference. When using a rule engine, the system invokes preset correction rules based on road and area functional attributes to correct the upper and / or lower limits of the basic illuminance range. When using fuzzy inference, the system uses commercial service intensity, residential sensitivity, pedestrian activity intensity, and public service intensity as input variables and outputs road attribute correction coefficients; the system then corrects the upper and / or lower limits of the basic illuminance range based on these road attribute correction coefficients.

[0039] When the target road is a long-distance traffic road, the system increases the upper limit of the basic illuminance range; when the target road is a commuter road, the system increases the upper limit of the basic illuminance range; when the target road is a commercial activity road, the system increases the upper limit of the basic illuminance range and decreases the lower limit of the basic illuminance range; when the target road is a pedestrian activity road, the system decreases the lower limit of the basic illuminance range. When the target road is located in a residential area, the system decreases the lower limit of the basic illuminance range; when the target road is located in a commercial area, the system increases the upper limit of the basic illuminance range; when the target road is located in an educational public area, the system increases the upper limit of the basic illuminance range and decreases the lower limit of the basic illuminance range.

[0040] like Figure 3 As shown, in step 7, the road attribute correction illuminance range is dynamically adjusted based on the acquired preprocessed multi-source real-time operating data to obtain the target illuminance value E. The multi-source real-time operating data includes vehicle traffic flow, pedestrian traffic flow, ambient light intensity, accident status, time information, weather information, and visibility information. The specific processing procedure is as follows: In step S7, dynamic adjustment can be achieved by combining safety constraint rules and machine learning regression models. Safety constraint rules are used to ensure that the target illuminance value is not lower than the illuminance safety lower limit corresponding to the target road, and to increase the target illuminance value in accident status or low visibility status. The inputs of the machine learning regression model include normalized vehicle traffic flow, normalized pedestrian traffic flow, ambient light intensity, time information, accident status, road functional attributes, and regional functional attributes, and the output is the target illuminance correction amount.

[0041] Machine learning regression models include one or more of the following: linear regression, random forest regression, gradient boosting tree regression, support vector regression, backpropagation neural network, or multilayer perceptron. The target illuminance correction output by the machine learning regression model is combined with the road attribute correction illuminance interval to generate the target illuminance value.

[0042] Dynamic adjustment can also employ time-series prediction models. Based on vehicle traffic flow, pedestrian traffic flow, ambient light intensity, and measured illuminance data from multiple past sampling periods, the system predicts the target illuminance correction for the next sampling period. Time-series prediction models include one or more of the following: Long Short-Term Memory (LSTM) networks, gated recurrent units (GNUs), temporal convolutional networks, or autoregressive models. When vehicle traffic flow increases, the system increases the target illuminance value within the road attribute correction illuminance range; when pedestrian traffic flow increases, the system increases the target illuminance value within the road attribute correction illuminance range; when ambient light intensity is below a preset illuminance threshold, the system increases the target illuminance value; when the accident status is abnormal, the system increases the target illuminance value to a preset safe illuminance level; when the time information is during a low-activity nighttime period, and both vehicle and pedestrian traffic flow are below the corresponding thresholds, the system reduces the target illuminance value without falling below the lower limit of the road lighting standard.

[0043] In step S7, the formula for calculating the target illuminance value E is: E = Emin + αV + βP + γL + δA + etaT + λQ + μR (2) Where Emin is the lower limit of the road attribute-corrected illuminance interval; V is the normalized vehicle flow rate; P is the normalized pedestrian flow rate; L is the ambient light correction term; A is the accident state correction term; T is the time correction term; Q is the weather correction term; R is the visibility correction term; and α, β, γ, δ, η, λ, and μ are the weights of the corresponding dynamic factors. The above formula is one optional implementation. The system can also use piecewise functions, regular functions, machine learning regression models, or time series prediction models to obtain the target illuminance value E.

[0044] like Figure 4 As shown, in step 8, the target luminous flux F is calculated based on the obtained target illuminance value E and the fixed characteristic quantity φ. Specifically, in step 8, the target luminous flux F is calculated based on the target illuminance value E and the fixed characteristic quantity φ. The target luminous flux F is positively correlated with the target illuminance value E and is affected by the luminaire maintenance factor, luminaire utilization factor, luminaire arrangement, luminaire spacing, and road width. In this embodiment, the formula for calculating the target luminous flux F is: F = E × φ(3) Where E is the target illuminance value and φ is a fixed characteristic quantity. The target luminous flux F is used to indicate the luminous flux that the target road light fixture or light group should output under the current operating conditions.

[0045] like Figure 5 As shown, in step S9, a lighting control command is generated based on the calculated target luminous flux F, and the lighting control command is sent to the lighting execution device. Specifically, in step S9, the system maps the target luminous flux F to the luminous dimming ratio according to a preset luminous fixture calibration curve. The luminous fixture calibration curve represents the correspondence between the luminous output luminous flux and the dimming ratio, drive current, or PWM duty cycle. The lighting control command includes one or more of the following: brightness level command, dimming ratio command, luminous fixture drive current control command, PWM duty cycle control command, DALI dimming command, 0-10V dimming command, and single-lamp controller control command. The lighting execution device includes a single-lamp controller, a luminous fixture driver, a lighting master control system, a street light centralized controller, or a lamp group controller.

[0046] like Figure 6As shown, in step S10, the measured illuminance data of the target road after the execution of lighting control is collected, and the dynamic adjustment weight is updated based on the historical measured illuminance data and historical operation data. Specifically, in step S10, the system collects the measured illuminance data after the lighting is executed through the illuminance acquisition unit, and adds a timestamp to each measured illuminance data. The system matches the measured illuminance data with vehicle flow, pedestrian flow, ambient light intensity, target illuminance value, and control commands within the same sampling period based on the timestamp to form a feedback sample.

[0047] Then, the system constructs a historical sample set based on historical vehicle flow, historical pedestrian flow, and corresponding measured illuminance; establishes a regression model based on the historical sample set; solves for the vehicle flow weight and pedestrian flow weight using the least squares method; and updates the dynamic illuminance adjustment model based on the solved weights.

[0048] The historical sample set includes vehicle traffic flow data, pedestrian traffic flow data, and measured illuminance data corresponding to multiple sampling periods. The system constructs a design matrix A, where each row includes vehicle traffic flow data and pedestrian traffic flow data corresponding to one sampling period; the system constructs a target vector Z, where each item includes measured illuminance data or illuminance deviation data for the corresponding sampling period; the system uses the least squares method to solve for a dynamically adjusted weight vector to update the influence coefficients of vehicle traffic flow and pedestrian traffic flow on the target illuminance value.

[0049] In step S10, the feedback optimization module uses recursive least squares to update the dynamic adjustment weights online. When the measured illuminance deviation exceeds the preset deviation threshold within a single sampling period but the deviation does not persist, the system does not immediately update the weights (weights α, β, γ, δ, η, λ, μ of the dynamic factors), but instead uses a sliding window to determine whether the deviation is persistent. When the deviation persists beyond the preset deviation threshold for multiple consecutive sampling periods, the system updates the dynamic adjustment weights. Specifically, the system determines whether the illuminance deviation between the measured illuminance and the target illuminance exceeds the preset deviation threshold. When the illuminance deviation does not exceed the preset deviation threshold, the system maintains the current dynamic adjustment weights and enters the next control period; when the illuminance deviation exceeds the preset deviation threshold, the system further uses a sliding window to determine whether the deviation is persistent. When the deviation persists beyond the preset deviation threshold for multiple consecutive sampling periods, the system uses least squares or recursive least squares to update the dynamic adjustment weights and inputs the updated dynamic adjustment weights into the dynamic illuminance adjustment module for target illuminance calculation in the next control period.

[0050] The adaptive lighting control method for urban roads based on multi-source urban data disclosed in this invention has the following beneficial effects: First, this invention determines the basic illuminance range from a preset road lighting standard library by using road grade parameters, and then performs road attribute correction and dynamic data adjustment, so that the target illuminance value is generated under the constraints of road lighting standards, reducing the risk of insufficient illuminance caused by simply adjusting the dimming based on real-time traffic flow.

[0051] Secondly, this invention uses a fixed characteristic quantity φ to characterize fixed conditions such as road width, luminaire installation height, luminaire spacing, luminaire arrangement, road surface material, maintenance factor, and utilization factor. This allows the same control method to adapt to different road conditions, improving the road adaptability of the lighting control model. In other words, this invention incorporates the differences in physical properties of different roads into the same control method, enabling roads with larger widths, larger luminaire spacing, or lower utilization factors to achieve correspondingly higher target luminous flux, avoiding inconsistent lighting effects caused by applying the same brightness level to different roads.

[0052] Third, this invention acquires vehicle traffic flow, pedestrian traffic flow, ambient light and visibility status through deep learning target detection, multi-target tracking, semantic segmentation, crowd density estimation, illuminance sensor acquisition and image brightness analysis, making the input data for road lighting control collectable and verifiable.

[0053] Fourth, this invention identifies road functional attributes and regional functional attributes through GIS overlay analysis, point of interest density identification, planning map layer identification, or street view image semantic recognition, enabling road lighting control to combine road space usage characteristics, rather than adjusting solely based on traffic flow.

[0054] Fifth, this invention generates target illuminance values ​​by combining safety constraint rules with machine learning regression models, enabling the system to dynamically respond to vehicle traffic, pedestrian traffic, ambient light, accident status, and time information while meeting the minimum lighting safety requirements.

[0055] Sixth, this invention further converts the target illuminance into a target luminous flux F, and generates brightness level instructions, dimming ratio instructions, PWM duty cycle control instructions, DALI dimming instructions, or single-lamp controller control instructions based on the target luminous flux F, so that the algorithm calculation results can be connected with the lamp control terminal, improving the feasibility of the technical solution.

[0056] Seventh, this invention establishes a feedback update mechanism through historical measured illuminance and historical operational data to update the dynamic adjustment weights of vehicle flow and pedestrian flow, enabling the system to gradually correct the dynamic adjustment model as the actual road usage changes, thereby improving the stability and interpretability of adaptive lighting control.

[0057] This invention employs a three-layer constraint calculation process: retrieving the corresponding basic illuminance range from a preset road lighting standard library; modifying the upper and / or lower limits of the basic illuminance range based on road and regional functional attributes to obtain a road attribute-corrected illuminance range; and dynamically adjusting the road attribute-corrected illuminance range. This invention avoids insufficient illuminance caused by solely adjusting based on real-time traffic flow, and also avoids excessive energy consumption caused by fixing lighting based solely on road grade. The standard constraint layer provides a safety baseline, the road attribute correction layer provides scene adaptation, and the dynamic operation adjustment layer provides real-time response.

[0058] Preferably, this application further includes implementing different adjustment methods based on the accident status or visibility information; specifically, it first determines whether there is an accident status or whether the visibility information of the target road is in a low visibility state. When it is determined that there is an accident status or a low visibility state, a safety priority adjustment is performed, that is, the target illuminance value is increased to a preset safe illuminance level; when there is no accident status or low visibility state, the system performs conventional dynamic adjustment, that is, it calculates the target illuminance correction amount based on one or more of the following: vehicle flow, pedestrian flow, ambient light intensity, time information, weather information, and visibility information.

[0059] The specific process of the safety priority control process under abnormal conditions in this invention (which executes different adjustment methods based on the accident status or visibility information) is as follows: When the system detects a traffic accident, abnormal congestion, road construction, or other road safety incidents on the target road, the incident status is marked as abnormal. Upon receiving the abnormal status, the dynamic illuminance adjustment module pauses the energy-saving priority adjustment strategy and raises the target illuminance value to the preset safe illuminance level.

[0060] When the system detects that the ambient light intensity of the target road is lower than a preset light threshold, or the image visibility is lower than a preset visibility threshold, the dynamic illumination adjustment module increases the target illumination value. The visibility threshold can be determined based on camera image brightness, image grayscale average, haze weather data, or meteorological platform data.

[0061] When the system detects that the current time is during a low-activity nighttime period and that both vehicle and pedestrian traffic are below the corresponding thresholds, the dynamic illuminance adjustment module reduces the target illuminance value without falling below the lower limit of the road lighting standard, in order to reduce road lighting energy consumption during low-activity periods.

[0062] Through the aforementioned safety-priority control process, the present invention can perform differentiated adjustments during accident, low visibility, and low-activity nighttime periods, taking into account both road lighting safety and energy efficiency.

[0063] Example 2 like Figure 7As shown, the urban road adaptive lighting control system based on multi-source urban data disclosed in this invention includes: The fixed road parameter acquisition module is used to acquire the fixed road parameters of the target road. The real-time running data acquisition module is used to acquire multi-source real-time running data of the target road; The data preprocessing module is used to perform timestamp alignment, outlier removal, data smoothing, and normalization on the acquired multi-source real-time running data to obtain preprocessed multi-source real-time running data; A fixed feature quantity calculation module is used to calculate a fixed feature quantity φ based on the acquired fixed road parameters; The standard constraint module is used to call the corresponding basic illuminance zone from the preset road lighting standard library based on the road level parameters in the fixed road parameters of the target road. The road attribute recognition module is used to identify the road function attributes and area function attributes of the target road; The road attribute correction module is used to correct the upper and / or lower limits of the basic illuminance range based on the identified road functional attributes and regional functional attributes, so as to obtain the road attribute corrected illuminance range. The dynamic illuminance adjustment module is used to dynamically adjust the illuminance range of road attribute correction based on the acquired preprocessed multi-source real-time operating data to obtain the target illuminance value E. The target luminous flux calculation module is used to calculate the target luminous flux F based on the acquired target illuminance value E and the fixed characteristic quantity φ. The control output module is used to generate lighting control commands based on the calculated target luminous flux F, and send the lighting control commands to the lighting actuator. Additionally, a feedback optimization module is used to collect measured illuminance data of the target road after the execution of lighting control, and to update the dynamic adjustment weights based on historical measured illuminance data and historical operation data.

[0064] Specifically, in one optional embodiment of this application, the modules are as follows: The fixed road parameter acquisition module is used to acquire the fixed road parameters of the target road. These parameters include road grade, road width W, lamp installation height H, lamp spacing S, lamp arrangement N, road surface material, lamp maintenance factor k, and lamp utilization factor U. These fixed road parameters can be obtained from a GIS spatial database, road design data, street light asset ledger, on-site surveying equipment, or mobile surveying equipment, and then input into the fixed road parameter acquisition module. For example, the module reads the road centerline, road boundaries, road grade, and street light location data from the GIS spatial database; calculates the lamp spacing S based on the spatial distance between adjacent street light locations or the road mileage distance; calculates the road width W based on the vertical distance between road boundary lines or road cross-section data; and acquires the lamp installation height H, lamp arrangement N, lamp maintenance factor k, and lamp utilization factor U from the street light asset ledger.

[0065] The real-time operational data acquisition module is used to acquire multi-source real-time operational data for the target road. This data includes vehicle flow, pedestrian flow, ambient light intensity, accident status, time information, weather information, and visibility information. Real-time operational data can be acquired and input into the module via devices or platforms such as cameras, radar, light sensors, weather sensors, road event detection equipment, urban traffic management platforms, or urban lighting management platforms. Various methods for acquiring real-time operational data are as follows: Vehicle flow is obtained through road video image recognition. Specifically, the system identifies vehicle targets in video frames using a convolutional neural network-based target detection model, and uses a multi-target tracking algorithm to associate the same vehicle target in consecutive video frames to form vehicle trajectories. When a vehicle trajectory crosses a preset virtual detection line or enters a preset detection area, the system generates a vehicle passage event and counts the number of vehicle passage events per unit time as the vehicle flow.

[0066] Pedestrian flow is obtained through one or more of the following methods: pedestrian target detection, multi-object tracking, semantic segmentation, or crowd density estimation. Specifically, the system uses a pedestrian target detection model and a multi-object tracking algorithm to obtain pedestrian trajectories and calculates pedestrian flow based on the number of times the pedestrian trajectory crosses the detection line. When the target road is located in a commercial district, near a school, a transportation hub, or other densely populated areas, the system uses a semantic segmentation model or a crowd density estimation model to obtain the proportion of people or the crowd density integral within the detection area, and uses the proportion of people or the crowd density integral as a pedestrian activity index.

[0067] Ambient light intensity is collected by illuminance sensors installed on light poles, road edges, or around the target road. The system adds a timestamp to each illuminance data point and timestamps the illuminance data with vehicle traffic flow, pedestrian traffic flow, and control commands.

[0068] Visibility information is obtained through road image brightness analysis. The system acquires images of the target road, calculates one or more of the following: the mean grayscale value, the mean luminance component value, the proportion of low-brightness pixels, or the image contrast, and determines the environmental visibility of the target road based on the calculation results. When the environmental visibility is lower than a preset visibility threshold, the system generates a low visibility status marker.

[0069] Time information is obtained through one or more of the following methods: local system clock, network time service, lighting control platform timestamp, or city IoT platform timestamp. Specifically, the system obtains time information through the local system clock, network time service, lighting control platform timestamp, or city IoT platform timestamp.

[0070] Weather information is obtained through meteorological sensors near the target road, road environment monitoring equipment, urban meteorological data interfaces, traffic management platforms, or third-party meteorological service interfaces.

[0071] Accident status is obtained through video analysis or external traffic event data. The system continuously tracks vehicle targets. When a vehicle target's speed in a non-stop area is lower than a preset speed threshold and the duration exceeds a preset time threshold, the vehicle target is marked as a suspected abnormal target. When the number of suspected abnormal targets in the same road segment exceeds a preset number, or the vehicle queue length continues to increase and exceeds a preset length threshold, the system marks the target road accident status as abnormal. The system can also read accident, construction, road closure, or traffic control data from the traffic management platform and update the accident status accordingly.

[0072] The data preprocessing module is used to perform timestamp alignment, outlier removal, data smoothing, and normalization on the acquired multi-source real-time operational data to obtain preprocessed multi-source real-time operational data. Specifically, the data preprocessing module performs timestamp alignment, outlier removal, data smoothing, and normalization on the acquired multi-source real-time operational data as follows: The data preprocessing module timestamps vehicle flow, pedestrian flow, ambient light intensity, and measured illuminance data according to a preset sampling period; outliers are removed using thresholding, interquartile range, Z-score, or sliding window mean method; the time series data are smoothed using moving average filtering, exponential smoothing, or Kalman filtering; vehicle flow is normalized according to the target road design capacity; pedestrian flow is normalized according to historical high quantiles or maximum observed values; and ambient light intensity is normalized according to a preset light threshold.

[0073] The fixed feature quantity calculation module is used to calculate the fixed feature quantity φ based on the acquired fixed road parameters. Specifically, the fixed road parameters acquired by the fixed feature quantity calculation module include road width W, lamp installation height H, lamp spacing S, lamp arrangement N, lamp maintenance factor k, and lamp utilization factor U. The formula for calculating the fixed feature quantity φ by the fixed feature quantity calculation module is as follows: φ = kSW / NU(1) Where k is the luminaire maintenance factor, S is the luminaire spacing, W is the road width, N is the luminaire arrangement parameter corresponding to the luminaire arrangement method, and U is the luminaire utilization factor. The above formula is an optional implementation method, and the fixed characteristic quantity φ can also be equivalently transformed according to the road lighting engineering formula.

[0074] The standard constraint module is used to retrieve the corresponding basic illuminance zone from a preset road lighting standard library based on the road level parameters in the acquired fixed road parameters of the target road. Specifically, the standard constraint module retrieves the corresponding basic illuminance interval from the preset road lighting standard library based on the road level parameters in the acquired fixed road parameters of the target road. The basic illuminance interval is used to limit the lighting reference range of the target road under the corresponding road level. The specific process is as follows: The preset road lighting standard library uses the road level as an index field and stores the corresponding lower limit of average illuminance, recommended range of average illuminance, illuminance uniformity requirements, and glare limitation requirements. The standard constraint module searches the preset road lighting standard library according to the road level of the target road to obtain the basic illuminance interval. The preset road lighting standard library includes basic illuminance intervals or basic illuminance thresholds corresponding to different road levels such as main roads, secondary roads, and local roads. The corresponding basic illuminance interval is retrieved from the preset road lighting standard library as the basis for subsequent road attribute correction and dynamic adjustment.

[0075] The road attribute identification module is used to identify the road function attributes and regional function attributes of the target road. Specifically, the road attribute identification module can identify the road function attributes and regional function attributes of the target road through one or more methods, including GIS overlay analysis, point of interest density identification, planning map layer identification, or street view image semantic recognition. A road influence buffer is generated based on the centerline of the target road, and the number and density of different types of points of interest are counted within the road influence buffer. When the density of commercial service-type points of interest is higher than a preset commercial threshold, the road attribute identification module marks the regional function attribute of the target road as a commercial area. When the proportion of residential-type points of interest or residential map patches is higher than a preset residential threshold, the road attribute identification module marks the regional function attribute of the target road as a residential area. When the density of education, medical, and public service-type points of interest is higher than a preset public service threshold, the road attribute identification module marks the regional function attribute of the target road as an educational public area or a public service area.

[0076] The road attribute correction module is used to correct the upper and / or lower limits of the basic illuminance interval based on the identified road function attributes and regional function attributes, thereby obtaining a road attribute corrected illuminance interval. Specifically, the road attribute correction module can use a rule engine or fuzzy inference method to correct the upper and / or lower limits of the basic illuminance interval based on the identified road function attributes and regional function attributes. When using a rule engine, the road attribute correction module calls preset correction rules based on the road function attributes and regional function attributes to correct the upper and / or lower limits of the basic illuminance interval. When using a fuzzy inference method, the system uses commercial service intensity, residential sensitivity, walking activity intensity, and public service intensity as input variables and outputs road attribute correction coefficients; the road attribute correction module corrects the upper and / or lower limits of the basic illuminance interval based on the road attribute correction coefficients.

[0077] When the target road is a long-distance traffic road, the road attribute correction module increases the upper limit of the basic illuminance range; when the target road is a commuter road, the module increases the upper limit of the basic illuminance range; when the target road is a commercial activity road, the module increases the upper limit of the basic illuminance range and decreases the lower limit; when the target road is a pedestrian activity road, the module decreases the lower limit. When the target road is located in a residential area, the module decreases the lower limit of the basic illuminance range; when the target road is located in a commercial area, the module increases the upper limit of the basic illuminance range; when the target road is located in an educational public area, the module increases the upper limit of the basic illuminance range and decreases the lower limit.

[0078] The dynamic illuminance adjustment module is used to dynamically adjust the illuminance range for road attribute correction based on the acquired preprocessed multi-source real-time operating data to obtain the target illuminance value E. Specifically, the dynamic illuminance adjustment module can use a combination of safety constraint rules and machine learning regression models to dynamically adjust the illuminance range for road attribute correction. The safety constraint rules are used to ensure that the target illuminance value is not lower than the illuminance safety lower limit corresponding to the target road, and to increase the target illuminance value in accident states or low visibility states. The inputs of the machine learning regression model include normalized vehicle flow, normalized pedestrian flow, ambient light intensity, time information, accident state, road functional attributes, and regional functional attributes, and the output is the target illuminance correction amount. Preferably, the dynamic illuminance adjustment module is also used to execute different adjustment methods based on the accident state or visibility information. Specifically, the dynamic illuminance adjustment module first determines whether the target road is in an accident state or whether the visibility information is in a low visibility state. When an accident or low visibility condition is detected, a safety-priority adjustment is performed, which increases the target illuminance value to the preset safe illuminance level. When no accident or low visibility condition exists, the system performs a regular dynamic adjustment, which calculates the target illuminance correction based on one or more of the following: vehicle flow, pedestrian flow, ambient light intensity, time information, weather information, and visibility information.

[0079] Machine learning regression models include one or more of the following: linear regression, random forest regression, gradient boosting tree regression, support vector regression, backpropagation neural network, or multilayer perceptron. The target illuminance correction output by the machine learning regression model is combined with the road attribute correction illuminance interval to generate the target illuminance value.

[0080] Dynamic adjustment can also employ time-series prediction models. The dynamic illuminance adjustment module predicts the target illuminance correction for the next sampling period based on vehicle traffic flow, pedestrian traffic flow, ambient light intensity, and measured illuminance data from multiple past sampling periods. The time-series prediction model includes one or more of the following: Long Short-Term Memory (LSTM) networks, gated recurrent units (GRUs), temporal convolutional networks, or autoregressive models. When vehicle traffic flow increases, the dynamic illuminance adjustment module increases the target illuminance value within the road attribute correction illuminance range; when pedestrian traffic flow increases, the module increases the target illuminance value within the same range; when ambient light intensity is below a preset illuminance threshold, the module increases the target illuminance value; when the accident status is abnormal, the module raises the target illuminance value to a preset safe illuminance level; when the time information is during a low-activity nighttime period, and both vehicle and pedestrian traffic flow are below the corresponding thresholds, the module lowers the target illuminance value without falling below the lower limit of the road lighting standard. The formula for calculating the target illuminance value E is: E = Emin + αV + βP + γL + δA + etaT + λQ + μR (2) Where Emin is the lower limit of the road attribute-corrected illuminance range; V is the normalized vehicle flow rate; P is the normalized pedestrian flow rate; L is the ambient light correction term; A is the accident state correction term; T is the time correction term; Q is the weather correction term; R is the visibility correction term; and α, β, γ, δ, η, λ, and μ are the weights of the corresponding dynamic factors. The above formula is one optional implementation. The dynamic illuminance adjustment module can also use piecewise functions, regular functions, machine learning regression models, or time series prediction models to obtain the target illuminance value E. The dynamic illuminance adjustment module prioritizes ensuring that the target illuminance is not lower than the illuminance safety lower limit corresponding to the target road, and then performs energy-saving control based on the real-time operating status.

[0081] The target luminous flux calculation module is used to calculate the target luminous flux F based on the acquired target illuminance value E and the fixed characteristic quantity φ; specifically, the formula for calculating the target luminous flux F by the target luminous flux calculation module is: F = E × φ(3) Where E is the target illuminance value and φ is a fixed characteristic quantity. The target luminous flux F is used to indicate the luminous flux that the target road light fixture or light group should output under the current operating conditions.

[0082] The control output module generates lighting control commands based on the calculated target luminous flux F and sends these commands to the lighting actuators. Specifically, the control output module maps the target luminous flux F to a luminaire dimming ratio according to a preset luminaire calibration curve. The luminaire calibration curve represents the correspondence between the luminaire's output luminous flux and the dimming ratio, drive current, or PWM duty cycle. The lighting control commands include one or more of the following: brightness level commands, dimming ratio commands, luminaire drive current control commands, PWM duty cycle control commands, DALI dimming commands, 0-10V dimming commands, and single-lamp controller control commands. The lighting actuators include single-lamp controllers, luminaire drivers, a central lighting control system, a street light centralized controller, or a lamp group controller.

[0083] In other words, the control output module pre-stores the luminaire calibration curve. This calibration curve represents the correspondence between the luminaire's output luminous flux and the dimming ratio, drive current, PWM duty cycle, or DALI dimming level. Once the target luminous flux F is determined, the control output module queries the dimming ratio corresponding to the target luminous flux F based on the luminaire calibration curve. If the lighting actuator uses PWM dimming, the system converts the dimming ratio into a PWM duty cycle; if the lighting actuator uses DALI dimming, the system converts the dimming ratio into a DALI dimming command; if the lighting actuator uses 0-10V dimming, the system converts the dimming ratio into a corresponding voltage control signal; if the lighting actuator uses a single-lamp controller, the system converts the dimming ratio into a control command recognizable by the single-lamp controller. After receiving the lighting control command, the lighting actuator adjusts the luminaire drive current or dimming level to ensure the target road luminaire operates according to the output state corresponding to the target luminous flux F.

[0084] The feedback optimization module collects measured illuminance data of the target road after lighting control is implemented, and updates the dynamic adjustment weights based on historical measured illuminance data and historical operational data. Specifically, the feedback optimization module collects measured illuminance data after lighting implementation through the illuminance acquisition unit and adds a timestamp to each measured illuminance data point. Based on the timestamps, the feedback optimization module matches the measured illuminance data with vehicle traffic flow, pedestrian traffic flow, ambient light intensity, target illuminance value, and control commands within the same sampling period to form feedback samples.

[0085] Then, the feedback optimization module constructs a historical sample set based on historical vehicle flow, historical pedestrian flow, and corresponding measured illuminance; establishes a regression model based on the historical sample set; solves for the vehicle flow weight and pedestrian flow weight using the least squares method; and updates the dynamic illuminance adjustment model based on the solved weights.

[0086] The historical sample set includes vehicle traffic flow data, pedestrian traffic flow data, and measured illuminance data corresponding to multiple sampling periods. The feedback optimization module constructs a design matrix A, where each row includes vehicle traffic flow data and pedestrian traffic flow data corresponding to one sampling period. The feedback optimization module also constructs a target vector Z, where each element includes measured illuminance data or illuminance deviation data for the corresponding sampling period. The feedback optimization module uses the least squares method to solve for a dynamically adjusted weight vector to update the influence coefficients of vehicle traffic flow and pedestrian traffic flow on the target illuminance value.

[0087] The feedback optimization module uses recursive least squares to update the dynamic adjustment weights online. When the measured illuminance deviation exceeds the preset deviation threshold within a single sampling period but the deviation does not persist, the system does not immediately update the weights (weights of dynamic factors α, β, γ, δ, η, λ, μ). Instead, it uses a sliding window to determine whether the deviation is persistent. When the deviation persists beyond the preset deviation threshold for multiple consecutive sampling periods, the system updates the dynamic adjustment weights. Specifically, the feedback optimization module determines whether the illuminance deviation between the measured illuminance and the target illuminance exceeds the preset deviation threshold. When the illuminance deviation does not exceed the preset deviation threshold, the feedback optimization module maintains the current dynamic adjustment weights and proceeds to the next control period. When the illuminance deviation exceeds the preset deviation threshold, the feedback optimization module further uses a sliding window to determine whether the deviation is persistent. When the deviation persists beyond the preset deviation threshold for multiple consecutive sampling periods, the feedback optimization module updates the dynamic adjustment weights using least squares or recursive least squares, and inputs the updated dynamic adjustment weights into the dynamic illuminance adjustment module for target illuminance calculation in the next control period.

[0088] Specifically, this embodiment illustrates the feedback optimization process based on the least squares method or recursive least squares method as follows: The feedback optimization module collects vehicle flow, pedestrian flow, and measured illuminance data from multiple historical sampling periods. Vehicle flow and pedestrian flow are used as input variables, while measured illuminance or illuminance deviation is used as the target variable. The system constructs a design matrix A and a target vector Z.

[0089] Design matrix A is represented as follows: A = [ [V1, P1], [V2, P2], …, [Vn, Pn] ] Where Vn represents the vehicle traffic flow in the nth historical sampling period, and Pn represents the pedestrian traffic flow in the nth historical sampling period.

[0090] The target vector Z is represented as: Z = [Z1, Z2, …, Zn]^T Wherein, Zn represents the measured illuminance value or illuminance deviation value corresponding to the nth historical sampling period.

[0091] The feedback optimization module solves for the weight vector B using the least squares method: B = (A^TA)^(-1) A^TZ Here, B includes vehicle flow weights and pedestrian flow weights. The updated vehicle flow weights and pedestrian flow weights are used by the dynamic illuminance adjustment module to calculate the target illuminance value in subsequent control cycles.

[0092] In another implementation, the feedback optimization module uses recursive least squares for online updates. Recursive least squares incrementally updates the weight vector based on newly collected feedback samples, without needing to recalculate all historical samples each time. The system can also use a sliding window to determine whether the measured illuminance deviation is persistent. When the deviation occurs only within a single sampling period, the system does not immediately update the weights; when the deviation exceeds a preset deviation threshold over multiple consecutive sampling periods, the system dynamically adjusts the weights.

[0093] Through the feedback optimization process, the system can adjust the impact of vehicle and pedestrian traffic on lighting demand based on historical operating data of the target road, reduce errors caused by manual experience-based weighting, and improve the stability of adaptive lighting control.

[0094] Preferably, this application further includes implementing different adjustment methods based on the accident status or visibility information; specifically, it first determines whether there is an accident status or whether the visibility information of the target road is in a low visibility state. When it is determined that there is an accident status or a low visibility state, a safety priority adjustment is performed, that is, the target illuminance value is increased to a preset safe illuminance level; when there is no accident status or low visibility state, the system performs conventional dynamic adjustment, that is, it calculates the target illuminance correction amount based on one or more of the following: vehicle flow, pedestrian flow, ambient light intensity, time information, weather information, and visibility information.

[0095] The specific process of the safety priority control process under abnormal conditions in this invention (which executes different adjustment methods based on the accident status or visibility information) is as follows: When the system detects a traffic accident, abnormal congestion, road construction, or other road safety incidents on the target road, the incident status is marked as abnormal. Upon receiving the abnormal status, the dynamic illuminance adjustment module pauses the energy-saving priority adjustment strategy and raises the target illuminance value to the preset safe illuminance level.

[0096] When the system detects that the ambient light intensity of the target road is lower than a preset light threshold, or the image visibility is lower than a preset visibility threshold, the dynamic illumination adjustment module increases the target illumination value. The visibility threshold can be determined based on camera image brightness, image grayscale average, haze weather data, or meteorological platform data.

[0097] When the system detects that the current time is during a low-activity nighttime period and that both vehicle and pedestrian traffic are below the corresponding thresholds, the dynamic illuminance adjustment module reduces the target illuminance value without falling below the lower limit of the road lighting standard, in order to reduce road lighting energy consumption during low-activity periods.

[0098] Through the aforementioned safety-priority control process, the present invention can perform differentiated adjustments during accident, low visibility, and low-activity nighttime periods, taking into account both road lighting safety and energy efficiency.

[0099] The present invention also provides an electronic device, including a processor, a memory, and a computer program stored in the memory and capable of running on the processor. When the processor executes the computer program, it implements the steps of any of the above-described adaptive lighting control methods for urban roads based on multi-source urban data.

[0100] The present invention also provides a computer-readable storage medium storing a computer program. When executed by a processor, the computer program implements the steps of any of the above-described adaptive lighting control methods for urban roads based on multi-source urban data.

[0101] Example 3 like Figure 8 As shown, this embodiment uses a main urban road as the target road for illustration. The streetlights on this target road are installed at a height of 8m, the road width is 10m, the streetlights are arranged in a staggered pattern, the spacing between the lights is 25m to 30m, and the road surface material is asphalt.

[0102] The system first acquires fixed road parameters for the target road and calculates a fixed feature quantity φ. Then, based on the road grade, the system retrieves the baseline illuminance interval corresponding to the main road. Finally, the system adjusts the baseline illuminance interval according to the commercial activity attributes of the area where the target road is located.

[0103] At any given moment, the system acquires a real-time vehicle flow of 500, a real-time pedestrian flow of 1.2, and an accident status of no accidents. Based on the vehicle flow, pedestrian flow, accident status, and time information, the system dynamically adjusts the road attribute correction illuminance range to obtain the target illuminance value.

[0104] The system calculates the target luminous flux F based on the target illuminance value and a fixed characteristic quantity φ, and converts the target luminous flux F into a lighting control command. After executing the lighting control, the system collects measured illuminance data of the target road.

[0105] In this embodiment, the target illuminance value calculated by the system is 34.8 lux, while the measured illuminance value of the target road is within the range of 34.0 lux to 37.0 lux. This result indicates that the target illuminance value calculated by the system in this embodiment is close to the measured illuminance value, which can meet the lighting requirements of the target road.

[0106] In a further implementation, the system uses vehicle flow, pedestrian flow, and measured illuminance data from multiple time periods of the target road as a historical sample set, and updates the vehicle flow weights and pedestrian flow weights using the least squares method or recursive least squares method. The updated weights are then used to calculate the target illuminance for subsequent control cycles of the target road.

[0107] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A method for adaptive lighting control of urban roads based on multi-source urban data, characterized in that, Includes the following steps: Step 1: Obtain the fixed road parameters of the target road; Step 2: Obtain multi-source real-time operational data for the target road; Step 3: Perform timestamp alignment, outlier removal, data smoothing, and normalization on the acquired multi-source real-time running data to obtain preprocessed multi-source real-time running data; Step 4: Calculate the fixed feature quantity φ based on the obtained fixed road parameters of the target road; Step 5: Based on the road grade parameters in the fixed road parameters of the target road, retrieve the corresponding basic illuminance range from the preset road lighting standard library; Step 6: Identify the road function attributes and area function attributes of the target road, and correct the upper and / or lower limits of the basic illuminance range based on the identified road function attributes and area function attributes to obtain the road attribute corrected illuminance range. Step 7: Based on the acquired preprocessed multi-source real-time operating data, dynamically adjust the road attribute correction illuminance range to obtain the target illuminance value E; Step 8: Calculate the target luminous flux F based on the obtained target illuminance value E and the fixed characteristic quantity φ; Step 9: Generate a lighting control command based on the calculated target luminous flux F, and send the lighting control command to the lighting actuator; Step 10: Collect the measured illuminance data of the target road after the lighting control is implemented, and update the dynamic adjustment weights based on the historical measured illuminance data and historical operation data.

2. The urban road adaptive lighting control method based on multi-source urban data according to claim 1, characterized in that: The fixed road parameters include road grade, road width W, lamp installation height H, lamp spacing S, lamp arrangement N, road surface material, lamp maintenance factor k, and lamp utilization factor U.

3. The urban road adaptive lighting control method based on multi-source urban data according to claim 2, characterized in that: The multi-source real-time operational data includes vehicle traffic, pedestrian traffic, ambient light intensity, accident status, time information, weather information, and visibility information.

4. The urban road adaptive lighting control method based on multi-source urban data according to claim 3, characterized in that: The formula for calculating the fixed characteristic quantity φ is: φ = kSW / NU(1) The fixed characteristic quantity φ is determined based on the road width W, the lamp spacing S, the lamp arrangement N, the lamp maintenance factor k, and the lamp utilization factor U.

5. The urban road adaptive lighting control method based on multi-source urban data according to claim 4, characterized in that: The formula for calculating the target illuminance value E is: E = Emin + αV + βP + γL + δA + etaT + λQ + μR (2) Where Emin is the lower limit of the road attribute correction illuminance range; V is the normalized vehicle flow; P is the normalized pedestrian flow; L is the ambient light correction term; A is the accident status correction term; T is the time correction term; Q is the weather correction term; R is the visibility correction term; and α, β, γ, δ, η, λ, and μ are the weights of the corresponding dynamic factors.

6. The urban road adaptive lighting control method based on multi-source urban data according to claim 5, characterized in that: The formula for calculating the target luminous flux F is: F = E × φ(3) Where F is the target luminous flux; E is the target illuminance value; and φ is a fixed characteristic quantity.

7. A system for implementing the urban road adaptive lighting control method based on multi-source urban data as described in any one of claims 1 to 6, characterized in that, include: The fixed road parameter acquisition module is used to acquire the fixed road parameters of the target road. The real-time running data acquisition module is used to acquire multi-source real-time running data of the target road; The data preprocessing module is used to perform timestamp alignment, outlier removal, data smoothing, and normalization on the acquired multi-source real-time running data to obtain preprocessed multi-source real-time running data; A fixed feature quantity calculation module is used to calculate a fixed feature quantity φ based on the acquired fixed road parameters; The standard constraint module is used to call the corresponding basic illuminance zone from the preset road lighting standard library based on the road level parameters in the fixed road parameters of the target road. The road attribute recognition module is used to identify the road function attributes and area function attributes of the target road; The road attribute correction module is used to correct the upper and / or lower limits of the basic illuminance range based on the identified road functional attributes and regional functional attributes, so as to obtain the road attribute corrected illuminance range. The dynamic illuminance adjustment module is used to dynamically adjust the illuminance range of road attribute correction based on the acquired preprocessed multi-source real-time operating data to obtain the target illuminance value E. The target luminous flux calculation module is used to calculate the target luminous flux F based on the acquired target illuminance value E and the fixed characteristic quantity φ. The control output module is used to generate lighting control commands based on the calculated target luminous flux F, and send the lighting control commands to the lighting actuator. Additionally, a feedback optimization module is used to collect measured illuminance data of the target road after the execution of lighting control, and to update the dynamic adjustment weights based on historical measured illuminance data and historical operation data.

8. The system according to claim 7, characterized in that: The fixed road parameters of the target road acquired by the fixed road parameter acquisition module include: road grade, road width W, lamp installation height H, lamp spacing S, lamp arrangement N, road surface material, lamp maintenance coefficient k, and lamp utilization coefficient U. The real-time operation data acquisition module obtains multi-source real-time operation data of the target road, including vehicle flow, pedestrian flow, ambient light intensity, accident status, time information, weather information, and visibility information.

9. The system according to claim 8, characterized in that: The formula for calculating the fixed characteristic quantity φ is: φ = kSW / NU(1) The fixed characteristic quantity φ is determined based on the road width W, the lamp spacing S, the lamp arrangement N, the lamp maintenance factor k, and the lamp utilization factor U. The formula for calculating the target luminous flux F is: F = E × φ(3) Where F is the target luminous flux; E is the target illuminance value; and φ is a fixed characteristic quantity.

10. The system according to claim 9, characterized in that: The formula for calculating the target illuminance value E is: E = Emin + αV + βP + γL + δA + etaT + λQ + μR (2) Where Emin is the lower limit of the road attribute correction illuminance range; V is the normalized vehicle flow; P is the normalized pedestrian flow; L is the ambient light correction term; A is the accident status correction term; T is the time correction term; Q is the weather correction term; R is the visibility correction term; and α, β, γ, δ, η, λ, and μ are the weights of the corresponding dynamic factors.

Images