Tunnel dynamic dimming method and system based on Internet of Things

Abstract

The invention discloses a tunnel dynamic dimming method and system based on the Internet of Things, and belongs to the field of dynamic dimming, and the method comprises the steps: an edge gateway receives sensor data in a tunnel, and processes the sensor data through a multi-source heterogeneous sensing algorithm built in the edge gateway; the edge gateway sends the received sensor data and the dimming intensity demand calculated by the multi-source heterogeneous sensing algorithm to the cloud, and the adaptive dimming algorithm generates a dimming strategy according to the sensor data and the dimming intensity demand; and issuing a dimming strategy generated by the adaptive dimming algorithm to a dimming control module in the tunnel. According to the invention, illumination can be dynamically adjusted according to the real-time environment in the tunnel, and the problem of energy waste or insufficient illumination is avoided.

Description

Technical Field

[0001] The present invention relates to the field of dynamic dimming, and particularly to a tunnel dynamic dimming method and system based on the Internet of Things. Background Art

[0002] With the development of society, the number of highway tunnels is increasing continuously. The tunnel lighting system plays a crucial role in tunnel traffic safety. There are some problems in the existing tunnel lighting control systems, which cannot dynamically adjust the lighting according to the real-time environmental changes, resulting in energy waste or insufficient lighting, and cannot meet the requirements of energy consumption, safety and visual comfort at the same time, and lack of adaptability to complex environments such as haze. Summary of the Invention

[0003] One object of the present invention is to provide a tunnel dynamic dimming method based on the Internet of Things to solve the problem in the prior art that the requirements of energy consumption, safety and visual comfort cannot be met at the same time, and the lighting control lacks comprehensiveness.

[0004] The present invention is realized by the following technical solutions. A tunnel dynamic dimming method based on the Internet of Things includes the following steps: S100, an edge gateway receives sensor data in a tunnel and processes the sensor data through a multi-source heterogeneous perception algorithm built in the edge gateway; S200, the edge gateway sends the received sensor data and the dimming illuminance requirement calculated by the multi-source heterogeneous perception algorithm to the cloud, and an adaptive dimming algorithm in the cloud generates a dimming strategy according to the sensor data and the dimming illuminance requirement; the adaptive dimming algorithm includes a dynamic scene classification mechanism, an energy efficiency priority model and a safety priority model; the dynamic scene classification mechanism performs scene classification, selects different priority modules according to different scenes, and different priority models generate different dimming strategies; S300, the dimming strategy generated by the adaptive dimming algorithm is sent down to a dimming control module in the tunnel, and the dimming control module adjusts tunnel lamps according to the dimming strategy.

[0005] Further, the sensors in the tunnel include core sensors and auxiliary sensors. The core sensor is a distributed fiber optic illuminance sensor; the auxiliary sensors include: a millimeter wave radar array, a thermal imaging sensor, a temperature / humidity sensor, an illuminance sensor and an image sensor.

[0006] Furthermore, the multi-source heterogeneous perception algorithm includes: S110, extracting and integrating the input data of the sensors to obtain the measured illuminance data, the visual demand illuminance data, and the haze attenuation factor; S120, defining dynamic weight parameters according to the measured illuminance data, the visual demand illuminance data, and the haze attenuation factor. The dynamic weight parameters are a parameter group Ω=(α,β,γ) composed of multiple dynamic weights. The functions of each parameter are as follows: α determines the importance of the measured illuminance data in the final adjusted illuminance, β determines the importance of adjusting the visual demand illuminance in the final dimming, and γ is used to adjust the importance of the attenuation factor of the influence of haze on illuminance; S130, combining the measured illuminance data, the visual demand illuminance data, the haze attenuation factor, and the dynamic weight parameters into the final adjusted illuminance demand through weighted summation. The adjusted illuminance demand is expressed by the following formula:

[0007] ,

[0008] , wherein, is the measured illuminance data, is the visual demand illuminance data, is the haze attenuation factor, is the attenuation factor, used to characterize the influence of haze on illuminance, z is the haze light transmission attenuation weight coefficient.

[0009] Furthermore, the measured illuminance data is calculated according to the following formula:

[0010] , where, is the sum of all illuminance sensors, is a single illuminance sensor, is the original data of the i-th sensor, is the sensor noise reduction residual, is the geometric attenuation weight; the visual demand illuminance data is calculated according to the following formula:

[0011] , where, is the contrast sensitivity factor, is the average vehicle speed, is the road speed limit value, is the exponential weight of the relationship between speed and visual demand illuminance; the haze attenuation factor is calculated according to the following formula:

[0012] , where, is the regression coefficient obtained by training based on foggy day image data, is the gray level distribution of the thermal imaging image, is the humidity ratio.

[0013] Furthermore, the parameter group Ω = (α, β, γ) is calculated according to the following formula:

[0014] ,

[0015] where, is the weight adjustment based on the energy consumption gradient change, is the natural constant, is the adjustment rate, is the energy consumption gradient change amount, is the gradient change weight parameter based on visual requirements and environmental interference, is the gradient of the visual requirement illuminance, is the gradient of the haze attenuation factor, is the quality constraint in the form of the Sigmoid activation function, SQI is the sensor signal quality index, is the fault flag, which is 1 if the sensor fails and 0 otherwise.

[0016] Furthermore, the scene feature vector modeling is represented by the following formula:

[0017] , where, is the fog concentration change rate, which reflects the change trend of the environmental visibility; is the total kinetic energy of vehicles within 100 meters is the curvature of the longitudinal illuminance requirement along the tunnel, is the image clarity index of the image sensor.

[0018] Furthermore, the fog concentration change rate can be calculated by the following formula:

[0019] , where, is the fog concentration at time t, t is the current time, is the time interval, usually a fixed short time. During the scene determination process, if the fog concentration change rate exceeds a certain threshold, it indicates that the fog concentration is changing rapidly, and the lighting brightness needs to be quickly increased or decreased to adapt to the change in visibility.

[0020] Furthermore, the total kinetic energy of vehicles within 100 meters can be calculated by the following formula:

[0021] , where M is the total number of vehicles within 100 meters, is the estimated mass of the i-th vehicle, is the speed of the \(i\)-th vehicle. During the scene determination process, if the total kinetic energy of the vehicles within 100 meters exceeds a certain threshold, it indicates that many vehicles are traveling at high speeds. At this time, it is necessary to increase the lighting brightness to ensure the driver's clear vision and reduce the possibility of accidents. If the total kinetic energy of the vehicles within 100 meters is lower than a certain threshold, it means that the vehicle speed is relatively slow, possibly due to traffic congestion or other reasons. In this case, the lighting brightness can be appropriately reduced to save energy.

[0022] Furthermore, the longitudinal curvature of the illuminance requirement in the tunnel can be calculated by the following formula:

[0023] , where is the illuminance requirement at position \(x\), and \(x\) is the longitudinal position of the tunnel. During the scene determination process, if the change curvature of the illuminance requirement is large, it indicates that the illuminance requirements at different positions in the tunnel change rapidly. For example, at sudden tunnel exits or entrances, the light changes significantly, and the lighting brightness needs to be adjusted quickly. If the change curvature is small, it means that the illuminance requirements in the tunnel are relatively stable, and the current lighting brightness can be maintained.

[0024] Furthermore, the image sharpness index can be calculated by the following formula:

[0025] , is the Laplace transform value at pixel \((x,y)\), representing the high-frequency component of the image; is the total number of pixels in the image. During the scene determination process, if the image sharpness index is low, it indicates that the image quality captured by the camera is poor, possibly due to haze or other factors affecting the vision. In this case, it is necessary to increase the lighting brightness to improve the vision. If the image sharpness index is high, it means that the current lighting conditions are already good, and the current lighting brightness can be maintained or appropriately reduced to save energy.

[0026] Furthermore, the energy efficiency priority model is defined by the objective function and constraints. The purpose of the objective function is to minimize the energy consumption while meeting the required illuminance, which is expressed by the following formula:

[0027] , where \(u\) k is the control input, \(k\) is a single moment, \(K\) is the sum of moments \(k\), is the actual illuminance value at the \(k\)-th moment, is the required illuminance value at the \(k\)-th moment, is the weight parameter, is the energy consumption corresponding to the control input \(u\) k at the \(k\)-th moment; the constraints ensure that the system complies with a series of physical and operational limitations while meeting the objective function. The constraints include: longitudinal gradient constraint: , this constraint limits the rate of change of illuminance along the longitudinal direction of the tunnel, ensuring that the illuminance does not change too much, and the maximum allowable rate of change is 15 lux per meter; Transient jump constraint: , this constraint limits the rate of change of the control input u k to ensure that the adjustment of the lighting system is smooth and does not change suddenly by a large margin. The maximum allowable rate of change is 10% per second; No-glare constraint: , this constraint ensures that the illuminance ratio between the brightest and darkest areas of the lighting system at any position does not exceed 5 times to avoid glare problems.

[0028] Further, the safety-first model dynamically adjusts the RGB of tunnel lights according to the monitored fog concentration by real-time monitoring of environmental conditions. mix The light color ratio generates a compensation light spot through the inverse Monte Carlo ray tracing algorithm to increase the illuminance of the target area and compensate for the light loss caused by haze; and the illuminance is forced to increase to 1.5 times the standard value within 20 meters of the tunnel entrance section to ensure that the driver has sufficient illuminance transition when entering the tunnel and avoid discomfort and safety hazards caused by sudden changes in illuminance.

[0029] Further, the adjustment of the light color ratio of the tunnel lights RGB mix is determined by the following formula:

[0030] , where RGB mix is the light color mixing ratio, and C f is the fog concentration. When the fog concentration C f < 0.3, the first set of ratios is used, with 70% blue light, 20% green light, and 10% red light. When the fog concentration C f ≥ 0.3, the second set of ratios is used, with 60% blue light, 15% green light, and 25% red light; The adjustment strategy of the compensation light spot is determined by the following formula:

[0031] , where is the compensation illuminance at the position (x, y), D is the number of all samples that need to be compensated for illuminance, is the luminous intensity distribution function of the lamp, indicating the luminous intensity of the lamp in the tunnel in the direction , is the scattering coefficient of haze, is the propagation distance of light in fog.

[0032] On the other hand, the present invention provides a tunnel dynamic dimming system for the Internet of Things, which includes a processor and a memory. A computer program is stored in the memory. When the computer program is executed by the processor, the tunnel dynamic dimming method for the Internet of Things as described above is implemented.

[0033] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0034] 1. The present invention can dynamically adjust the lighting according to the real-time environment, traffic flow and lighting conditions in the tunnel, avoiding the problems of energy waste or insufficient lighting.

[0035] 2. The present invention fuses multi-dimensional information such as environment, traffic flow and lighting through a multi-source heterogeneous perception algorithm, and introduces a haze attenuation factor and dynamic weight parameters, improving the adaptability of the system to complex environments such as haze.

[0036] 3. The present invention combines an adaptive dimming algorithm, which can simultaneously meet the requirements of energy consumption, safety and visual comfort. By perceiving traffic information such as traffic flow and vehicle speed, the lighting can be optimized according to the traffic conditions, improving the safety of personnel driving in the tunnel. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] The drawings described herein are used to provide a further understanding of the embodiments of the present invention, form a part of this application, and do not constitute a limitation on the embodiments of the present invention. In the drawings:

[0038] Figure 1 is a flowchart of the overall steps provided in Embodiment 1 of the present invention.

[0039] Figure 2 is a timing diagram of the overall steps provided in Embodiment 1 of the present invention.

[0040] Figure 3 is a flowchart of the multi-source heterogeneous perception algorithm provided in Embodiment 1 of the present invention.

[0041] Figure 4 is a timing diagram of the multi-source heterogeneous perception algorithm provided in Embodiment 1 of the present invention.

[0042] Figure 5 is a timing diagram of the adaptive dimming algorithm provided in Embodiment 1 of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0043] To make the objectives, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. Usually, the components of the embodiments of the present invention described and illustrated herein can be arranged and designed in various different configurations.

[0044] Embodiment 1

[0045] The prior art cannot dynamically adjust the lighting according to the real-time environmental changes in the tunnel, resulting in energy waste or insufficient lighting, and cannot meet the requirements of energy consumption, safety, and visual comfort simultaneously. The lighting control lacks comprehensiveness and the adaptability to complex environments such as haze, and cannot adjust the lighting parameters according to environmental changes. The lighting control system lacks the ability to sense traffic information such as traffic flow and vehicle speed, and cannot optimize the lighting according to traffic conditions. Moreover, the lighting control system lacks the ability to sense traffic information such as traffic flow and vehicle speed, and cannot optimize the lighting according to traffic conditions. To solve the above problems, this embodiment discloses a tunnel dynamic dimming method based on the Internet of Things. This embodiment can dynamically adjust the lighting according to the real-time environment, traffic flow, and lighting conditions in the tunnel. The solution collects multi-dimensional information such as the environment, traffic flow, and lighting in the tunnel by arranging various types of high-precision sensors in the tunnel, and uses a multi-source heterogeneous perception algorithm to fuse and process this information to obtain the required dimming illuminance. At the same time, through the adaptive dimming algorithm deployed in the cloud, the required dimming illuminance is converted into an execution strategy that meets the requirements of energy consumption, safety, and light quality, and the execution strategy is sent to the dimming control module in the tunnel to achieve intelligent lighting control.

[0046] The method disclosed in this embodiment can dynamically adjust the tunnel lighting according to the real-time environment, traffic flow, and lighting conditions in the tunnel, which can not only meet the requirements of driving safety and visual comfort, but also save energy to the greatest extent and improve energy utilization efficiency, thus solving the problems existing in the prior art.

[0047] Figure 1 The flowchart showing the overall steps in this embodiment is as follows. Figure 2 The sequence diagram showing the overall steps in this embodiment is as follows. It can be seen from the figure that this embodiment includes the following steps:

[0048] Step 1: Arrange various types of high-precision sensors in the tunnel to collect various information in the tunnel, and transmit the data collected by the sensors to the edge gateway through the wireless Internet of Things.

[0049] The edge gateway is built-in with a multi-source data processing module based on FPGA. This module processes the data collected by the sensors arranged in the tunnel through a multi-source heterogeneous perception algorithm. This multi-source heterogeneous perception algorithm realizes the perception of multi-dimensional information of the environment-traffic flow-light field through the processing of sensors and redundant networks.

[0050] It should be noted that in this embodiment, multiple types of high-precision sensors are arranged, including dividing different types of sensors into core sensors and auxiliary sensors. Among them, the core sensor is a distributed fiber optic illuminance sensor, which is installed every 2m in the tunnel; the auxiliary sensors include a millimeter wave radar array for vehicle information perception in the tunnel, a thermal imaging sensor for fog concentration perception, a temperature / humidity sensor, a light sensor, and an image sensor for environmental perception. The data collected by the sensors is first sent to the edge gateway, and the edge gateway first preprocesses the data. Since the sensors may be affected by noise and other interferences, some preprocessing needs to be performed on them before processing. It can include: using wavelet transform to denoise the data and removing errors generated by environmental noise or the sensors themselves. At the same time, in order to remove the high-frequency fluctuations in the measurement process, a low-pass filter can be used to smooth the sensor data. Finally, the denoised sensor data value is obtained.

[0051] Figure 3 The flowchart of the multi-source heterogeneous perception algorithm in this embodiment is shown. Figure 4 The timing diagram of the multi-source heterogeneous perception algorithm in this embodiment is shown. It can be seen from the figure that the multi-source heterogeneous perception algorithm includes the following steps:

[0052] 1) Extract the input data related to the lighting intensity requirement. These input data usually come from multiple sensors and environmental monitoring devices, and the input data is integrated to obtain the measured illuminance data, visual demand illuminance data, and haze attenuation factor.

[0053] Among them, the measured illuminance data L opt , which is the real-time illuminance directly measured by the sensor.

[0054] In this embodiment, it can be calculated by the following formula:

[0055] ,

[0056] Among them, is the sum of all illuminance sensors, is a single illuminance sensor, is the original data of the i-th sensor, is the sensor noise reduction residue, is the geometric attenuation weight, which can be calculated according to the position of the lighting fixtures in the tunnel. This formula obtains a comprehensive light measurement value by averaging all sensor data and considering the geometric attenuation weight.

[0057] The visual demand illuminance data L vd , which is calculated based on the illuminance for visual comfort and demand. This value can determine the required illuminance according to the passing vehicle speed, road conditions, and speed limit value in the tunnel.

[0058] In this embodiment, the visual demand illuminance data L vd can be calculated by the following formula:

[0059] ,

[0060] where is the contrast sensitivity factor, which can be obtained by looking up the table according to the lighting rules in the Lighting Guide for Tunnels and Underpasses (Standard CIE 88 - 2004); is the average vehicle speed, is the road speed limit value, is the exponential weight of the relationship between speed and visual demand illuminance, which determines how the change in the average vehicle speed relative to the safe speed affects the visual demand illuminance. When δ > 1, a higher average speed will result in a greater visual demand illuminance, that is, the faster the speed, the higher the lighting demand for drivers or pedestrians; when δ < 1, the change in speed has a smaller impact on the visual demand illuminance, and even if the speed increases, the lighting demand will not increase significantly. When δ = 1, the visual demand illuminance is linearly related to the speed.

[0061] The haze attenuation factor C fog , since haze will cause strong attenuation of light, it is necessary to model this factor. The haze concentration index can be inversely calculated through thermal imaging technology or other image processing methods to obtain a haze attenuation value calculated based on environmental conditions (such as humidity, wind speed, etc.).

[0062] In this embodiment, the haze attenuation factor C fog can be calculated by the following formula:

[0063] ,

[0064] where is the regression coefficient obtained by training based on foggy day image data, is the gray - level distribution of the thermal imaging image, is the humidity ratio, is the second - order mixed partial derivative of the gray - level value of the thermal imaging image, representing the change rate of the image in the x and y directions, is the water vapor concentration for normalization, quantifying the attenuation effect of water vapor on light.

[0065] 2) For the measured illuminance data, visual demand illuminance data, and haze attenuation factor, define the dynamic weight parameter Ω, where the dynamic weight parameter Ω=(α,β,γ) is a parameter group composed of multiple dynamic weights, and the role of each parameter is as follows:

[0066] α determines the importance of the measured illuminance data in the final adjusted illuminance. It is usually a dynamic value and may be adjusted according to energy consumption changes, lighting conditions, etc.

[0067] β determines the importance of adjusting the visual demand illuminance in the final dimming. Its value may vary according to traffic conditions, road types, or the driver's needs.

[0068] γ is used to adjust the importance of the attenuation factor of the influence of haze on lighting. Its value may depend on the signal quality of the sensor or environmental conditions such as humidity, haze concentration, etc.

[0069] These weights are dynamically changing, that is, they are adjusted in real time according to factors such as environmental conditions and sensor signal quality. Through these dynamically changing weight parameters, the flexibility and adaptability of the entire system are ensured, enabling it to operate efficiently in a complex and changing Internet of Things environment.

[0070] Specifically, in this implementation, the dynamic weight parameter Ω can be calculated according to the following formula:

[0071] ,

[0072] where, is the weight adjustment based on the energy consumption gradient change, is the natural constant, is the adjustment rate, is the energy consumption gradient change amount, is the gradient change weight parameter based on visual demand and environmental interference, is the gradient of the visual demand illuminance, used to represent the change rate of the function in each direction,

[0073] is the norm of the illuminance gradient, representing the magnitude of the gradient vector, quantifying the total change rate or change intensity in space; is the gradient of the haze attenuation factor, is the norm of the haze attenuation factor gradient, quantifying the total change rate or change intensity of the haze attenuation factor in space; is the weight parameter based on the sensor signal quality index and the fault flag, is the quality constraint in the form of the Sigmoid activation function, SQI is the sensor signal quality index, is the fault flag, which is 1 if the sensor fails and 0 otherwise.

[0074] 3) Combine the measured illuminance data, visual demand illuminance data, haze attenuation factor, and dynamic weight parameter into the final adjusted illuminance demand L through weighted summation req, the required dimming illuminance can be expressed by the following formula:

[0075] ,

[0076] ,

[0077] Wherein, is the attenuation factor, which is used to characterize the influence of haze on illuminance, z is the haze light transmission attenuation weight coefficient.

[0078] Step 2: After receiving the sensor data and the required dimming illuminance sent by the edge gateway, the cloud data center uses the adaptive dimming algorithm deployed in the cloud to convert the required dimming illuminance into a dimming execution strategy that can meet energy consumption, safety, and light quality in combination with the sensor data.

[0079] Specifically, the adaptive dimming algorithm includes a dynamic scene classification mechanism, an energy efficiency priority model, and a safety priority model. First, the dynamic scene classification mechanism is used for scene classification, different priority modules are selected according to different scenes, and different priority models generate different dimming execution strategies according to rules. Figure 5 shows the timing diagram of the adaptive dimming algorithm of this embodiment.

[0080] In this embodiment, the scene feature vector modeling is expressed by the following formula:

[0081] ,

[0082] Wherein, is the fog concentration change rate, which reflects the change trend of environmental visibility; is the total kinetic energy of vehicles within 100 meters, which reflects the comprehensive situation of traffic flow and vehicle speed; is the curvature of the longitudinal illuminance requirement along the tunnel, that is, the second derivative of the longitudinal illuminance requirement along the tunnel, which reflects the change degree of the illuminance requirement and is used to judge the drastic change of the lighting requirement in the tunnel; is the image clarity index of the image sensor, which reflects the current visibility and image quality.

[0083] In this embodiment, the fog concentration change rate can be calculated by the following formula:

[0084] ,

[0085] Wherein, is the fog concentration at time t, and t is the current time, is the time interval, usually a fixed short time. During the scene determination process, if the change rate of fog concentration exceeds a certain threshold, it indicates that the fog concentration is changing rapidly, and it is necessary to quickly increase or decrease the lighting brightness to adapt to the change in visibility.

[0086] The total kinetic energy of vehicles within 100 meters can be calculated by the following formula:

[0087] ,

[0088] where M is the total number of vehicles within 100 meters, is the estimated mass of the i-th vehicle, is the speed of the i-th vehicle. During the scene determination process, if the total kinetic energy of vehicles within 100 meters exceeds a certain threshold, it indicates that many vehicles are traveling at high speeds. At this time, it is necessary to increase the lighting brightness to ensure the driver's field of vision is clear and reduce the possibility of accidents. If the total kinetic energy of vehicles within 100 meters is lower than a certain threshold, it indicates that the vehicle speed is relatively slow, which may be due to traffic congestion or other reasons. In this case, the lighting brightness can be appropriately reduced to save energy.

[0089] The curvature of the longitudinal illuminance demand along the tunnel can be calculated by the following formula:

[0090] ,

[0091] where, is the illuminance demand at position x, and x is the longitudinal position of the tunnel. During the scene determination process, if the change curvature of the illuminance demand is large, it indicates that the illuminance demand at different positions in the tunnel changes rapidly. For example, at the sudden tunnel exit or entrance, the light change is significant, and the lighting brightness needs to be adjusted quickly. If the change curvature is small, it indicates that the illuminance demand in the tunnel is relatively stable, and the current lighting brightness can be maintained.

[0092] The image sharpness index can be calculated by the following formula:

[0093] ,

[0094] is the Laplace transform value at pixel (x, y), representing the high-frequency component of the image; is the total number of pixels in the image. During the scene determination process, if the image sharpness index is low, it indicates that the image quality captured by the camera is poor, which may be affected by haze or other factors affecting the field of vision. In this case, it is necessary to increase the lighting brightness to improve the field of vision. If the image sharpness index is high, it indicates that the current lighting conditions are already good, and the current lighting brightness can be maintained or appropriately reduced to save energy.

[0095] It should be noted that the rule determined by vector features can be:

[0096] High fog concentration change + high total vehicle kinetic energy + high illumination demand change + low image clarity:

[0097] It is necessary to rapidly increase the illumination brightness to ensure that the driver can clearly see ahead.

[0098] Low fog concentration change + low total vehicle kinetic energy + low illumination demand change + high image clarity:

[0099] The illumination brightness can be maintained at the current level or appropriately reduced to save energy.

[0100] Medium fog concentration change + medium total vehicle kinetic energy + medium illumination demand change + medium image clarity:

[0101] Depending on the specific situation, small adjustments may be required to ensure that the lighting conditions can meet the current traffic and environmental requirements.

[0102] In this embodiment, the energy efficiency priority model can be defined by an objective function and constraint conditions.

[0103] The purpose of the objective function is to minimize energy consumption on the premise of meeting the required illumination, which can be expressed by the following formula:

[0104] ,

[0105] where u k is the control input, which can be the settings of the lighting system (such as switch state, brightness adjustment, etc.), k is a single moment, K is the sum of moments k, is the actual illumination value at the k-th moment, is the required illumination value at the k-th moment, is the weight parameter used to balance the weight between tracking accuracy and energy consumption. A larger value means more emphasis on energy consumption, and a smaller value means more emphasis on control accuracy, is the energy consumption corresponding to the control input u k at the k-th moment, This term is the square of the difference between the actual illumination value and the required illumination value, representing control accuracy. The square form makes the deviation larger, and the penalty is greater, thus prompting the system to make the actual illumination value as close as possible to the required illumination value.

[0106] The constraint conditions ensure that the system complies with a series of physical and operational limitations while meeting the objective function. In this embodiment, the constraint conditions include:

[0107] Longitudinal gradient constraint:

[0108] , this constraint limits the rate of change of illuminance along the longitudinal direction of the tunnel, ensuring that the illuminance does not change too much, avoiding discomfort for the driver. The maximum allowable rate of change is 15 lux per meter.

[0109] Transient jump constraint:

[0110] , this constraint limits the rate of change of the control input u k . It ensures that the adjustment of the lighting system is smooth and does not change suddenly by a large margin, avoiding causing dizziness or discomfort to the driver. The maximum allowable rate of change is 10% per second.

[0111] No-glare constraint:

[0112] , this constraint ensures that the illuminance ratio between the brightest and darkest areas of the lighting system at any position does not exceed 5 times, avoiding glare problems and ensuring the comfort and safety of the driver's field of vision.

[0113] In this embodiment, the safety - priority model dynamically adjusts the light - wavelength ratio by real - time monitoring of environmental conditions. According to the monitored fog concentration C f dynamically adjusts the light - color ratio of the RGB mix of the tunnel lamps; at the same time, it generates compensation light spots through reverse Monte Carlo ray tracing to increase the illuminance in specific areas and compensate for the light loss caused by haze; and it forcibly increases the illuminance to 1.5 times the standard value within 20 meters of the tunnel entrance section to ensure that the driver has sufficient illuminance transition when entering the tunnel, avoiding discomfort and safety hazards caused by sudden changes in illuminance.

[0114] It should be noted that the safety - priority model adjusts the light - color mixing ratio according to the concentration of fog, realizes the self - adaptive adjustment of light wavelength to reduce scattering and improve visibility. It uses reverse Monte Carlo ray tracing to calculate the compensation illuminance to overcome the scattering and absorption of light by haze. Finally, it forcibly increases the illuminance in key areas (such as the tunnel entrance) and achieves a fast response through a high - speed communication bus. Combining these methods together ensures that sufficient lighting and safety protection can be provided even in harsh environments.

[0115] The specific strategy for adjusting the light - color ratio of tunnel lamps can be determined by the following formula:

[0116] ,

[0117] where RGB mix is the light - color mixing ratio, and C f is the fog concentration. When the fog concentration C fWhen C < 0.3, the first set of ratios (0.2, 0.7, 0.1) is used, that is, blue light accounts for 70%, green light accounts for 20%, and red light accounts for 10%. This combination helps to provide better visibility in the case of low-concentration fog. When the fog concentration C f ≥ 0.3, the second set of ratios (0.15, 0.6, 0.25) is used, that is, blue light accounts for 60%, green light accounts for 15%, and red light accounts for 25%. This combination provides better visibility in the case of high-concentration fog.

[0118] The specific compensation spot adjustment strategy can be determined by the following formula:

[0119] ,

[0120] where, is the compensation illuminance at the position (x, y), D is the number of all samples that need to be compensated for illuminance, is the luminous intensity distribution function of the lamp, indicating the luminous intensity of the lamp in the tunnel in the direction of, is the scattering coefficient of the haze, is the propagation distance of the light in the fog.

[0121] Step 3: The cloud data center sends the dimming execution strategy generated by the adaptive dimming algorithm to the dimming control module in the tunnel, and the dimming control module adjusts the brightness of the tunnel lamps according to the dimming execution strategy.

[0122] The above-described specific embodiments have further elaborated on the purpose, technical solutions, and beneficial effects of the present invention. It should be understood that the above is only the specific embodiments of the present invention and is not used to limit the protection scope of the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A tunnel dynamic dimming method based on the Internet of Things, characterized in that, The described dynamic dimming method includes: S100. The edge gateway receives sensor data in the tunnel and processes the sensor data through a multi-source heterogeneous perception algorithm built into the edge gateway. S200. The edge gateway sends the received sensor data and the dimming illumination requirement calculated by the multi-source heterogeneous perception algorithm to the cloud. The adaptive dimming algorithm in the cloud generates a dimming strategy based on the sensor data and the dimming illumination requirement. The adaptive dimming algorithm includes a dynamic scene classification mechanism, an energy efficiency priority model, and a safety priority model. The dynamic scene classification mechanism performs scene classification, selects different priority modules according to different scenes, and different priority models generate different dimming strategies. S300. Send the dimming strategy generated by the adaptive dimming algorithm to the dimming control module in the tunnel. The dimming control module adjusts the tunnel lamps according to the dimming strategy.

2. The tunnel dynamic dimming method based on the Internet of Things according to claim 1, wherein The sensors in the tunnel include core sensors and auxiliary sensors. The core sensor is a distributed fiber optic illuminance sensor. The auxiliary sensors include: a millimeter wave radar array, a thermal imaging sensor, a temperature / humidity sensor, a light sensor, and an image sensor.

3. The tunnel dynamic dimming method based on the Internet of Things according to claim 1, wherein The multi-source heterogeneous perception algorithm includes: S110. Extract and integrate the input data of the sensors to obtain measured illuminance data, visual demand illuminance data, and haze attenuation factor. S120. Define dynamic weight parameters according to the measured illuminance data, visual demand illuminance data, and haze attenuation factor. The dynamic weight parameter is a parameter group Ω=(α,β,γ) composed of multiple dynamic weights, and the function of each parameter is as follows: α determines the importance of the measured illuminance data in the final dimming illuminance. β is used to adjust the importance of the visual demand illuminance in the final dimming. γ is used to adjust the importance of the attenuation factor of the haze's influence on light. S130. Combine the measured illuminance data, visual demand illuminance data, haze attenuation factor, and dynamic weight parameters into the final dimming illuminance requirement by weighted summation. The dimming illuminance requirement is expressed by the following formula: ， ， Among them, is the measured illuminance data, is the visual demand illuminance data, is the haze attenuation factor, is the attenuation factor, used to characterize the influence of haze on illuminance, z is the haze light transmission attenuation weight coefficient.

4. The tunnel dynamic dimming method based on the Internet of Things according to claim 3, wherein, The measured illuminance data is calculated according to the following formula: ， wherein, is the sum of all illuminance sensors, is a single illuminance sensor, is the original data of the i-th sensor, is the residual amount of sensor noise reduction, is the geometric attenuation weight; The visual demand illuminance data is calculated according to the following formula: ， Among them, is the contrast sensitivity factor, is the average vehicle speed, is the road speed limit value, is the exponential weight of the relationship between speed and visual demand illuminance; The haze attenuation factor is calculated according to the following formula: ， Among them, is the regression coefficient obtained by training based on foggy weather image data, is the gray scale distribution of the thermal imaging image, is the humidity ratio.

5. The tunnel dynamic dimming method based on the Internet of Things according to claim 3, characterized in that The parameter group Ω=(α,β,γ) is calculated according to the following formula: ， Among them, is the weight adjustment based on the change of energy consumption gradient, is the natural constant, is the adjustment rate, is the change amount of energy consumption gradient, is the gradient change weight parameter based on visual demand and environmental interference, is the gradient of visual demand illuminance, is the gradient of haze attenuation factor, is the quality constraint in the form of Sigmoid activation function, SQI is the sensor signal quality index, is the fault flag, which is 1 if the sensor fails, otherwise 0; t is the current time.

6. The tunnel dynamic dimming method based on the Internet of Things according to claim 1, characterized in that The scene feature vector modeling is expressed by the following formula: ， Among them, is the fog concentration change rate, which reflects the change trend of environmental visibility; is the total kinetic energy of vehicles within 100 meters is the curvature of the longitudinal illumination demand along the tunnel, is the required illumination value is the image clarity index of the image sensor.

7. The tunnel dynamic dimming method based on the Internet of Things according to claim 1, characterized in that The energy efficiency priority model is defined by an objective function and constraint conditions. The purpose of the objective function is to minimize energy consumption on the premise of meeting the required illuminance, and it is expressed by the following formula: ， where u k is the control input, k is a single moment, and K is the sum at moment k, is the actual illuminance value at the k-th moment, is the required illuminance value at the k-th moment, is the regression coefficient obtained by training based on foggy weather image data, is the control input u at the k-th moment k corresponding to the energy consumption, is the actual illuminance value; The constraint conditions ensure that the system complies with a series of physical and operation limitations while meeting the objective function. The constraint conditions include: Longitudinal gradient constraint: ， x is the longitudinal position of the tunnel; this constraint limits the rate of change of illuminance along the longitudinal direction of the tunnel, ensuring that the illuminance does not change too much, and the maximum allowable rate of change is 15 lux per meter; Transient jump constraint: ， This constraint limits the rate of change of the control input u k to ensure that the adjustment of the lighting system is smooth and does not change suddenly by a large margin. The maximum allowable rate of change is 10% per second; No glare constraint: ， is the maximum illuminance value, is the minimum illuminance value; this constraint ensures that the illuminance ratio between the brightest and darkest areas at any position in the lighting system does not exceed 5 times to avoid glare problems.

8. The tunnel dynamic dimming method based on the Internet of Things according to claim 1, characterized in that The safety - priority model dynamically adjusts the light - color ratio of the RGB tunnel lights according to the monitored fog concentration by monitoring the environmental conditions in real - time. mix of the light color; Generate a compensation spot through the reverse Monte Carlo ray tracing algorithm to improve the illuminance of the target area and compensate for the light loss caused by haze. And forcibly increase the illuminance to 1.5 times the standard value within 20 meters of the tunnel entrance section to ensure that drivers have sufficient illuminance transition when entering the tunnel and avoid discomfort and safety hazards caused by sudden changes in illuminance.

9. The tunnel dynamic dimming method based on the Internet of Things according to claim 8, wherein The RGB light color ratio of the adjustable tunnel lamp mix is determined by the following formula: ， Among them, RGB mix is the color mixing ratio of light, and C f is the fog concentration; when the fog concentration C f < 0.3, the first set of ratios is used, with blue light accounting for 70%, green light accounting for 20%, and red light accounting for 10%. When the fog concentration C f ≥ 0.3, the second set of ratios is used, with blue light accounting for 60%, green light accounting for 15%, and red light accounting for 25%; The adjustment strategy of the compensation spot is determined by the following formula: ， Among them, is the compensated illuminance at position (x, y), D is the number of all samples that need compensated illuminance, is the luminous intensity distribution function of the lamp, indicating the luminous intensity of the lamp in the tunnel in the direction . is the scattering coefficient of the haze, is the propagation distance of the light in the fog.

10. An Internet of Things-based tunnel dynamic dimming system, characterized in that, The dynamic dimming system includes: A processor; A memory stores a computer program which, when executed by a processor, implements the Internet-of-Things-based tunnel dynamic dimming method according to any one of claims 1 to 9.

Images