Coating film intelligent control system and control method for self-luminous road marking

Abstract

The application discloses a coating film intelligent control system and control method for self-luminous road marking, relates to the technical field of road engineering, constructs a road marking, environment and traffic digital twin model, runs a deep reinforcement learning algorithm on an edge side, outputs optimal luminous parameter decisions through the digital twin model based on key characteristic parameters and traffic prediction data, forms a preliminary flexible controllable luminous unit control instruction, constructs a vehicle-road cloud integrated collaborative interaction mechanism, links vehicles, traffic platforms and cross-department systems, supplements real-time interaction data for algorithm decisions, optimizes flexible controllable luminous unit control instructions and outputs them. The application realizes multi-dimensional regulation and control of luminous parameters, enhances environmental and traffic adaptability, and an adaptive algorithm driven by a digital twin considers visible distance and energy consumption optimization, so that the energy consumption is greatly reduced while ensuring that the visible distance is greater than or equal to a set threshold value; the vehicle-road cloud integrated collaborative interaction mechanism greatly reduces response delay in emergency scenarios and improves road safety.

Description

Technical Field

[0001] This invention relates to the field of road engineering technology, and in particular to an intelligent control system and control method for coatings used in self-illuminating road markings. Background Technology

[0002] With increasing demands for road traffic safety, self-illuminating road markings have gained widespread use due to their active nighttime illumination properties. However, existing technologies have several limitations:

[0003] The perception dimension is singular, mostly collecting only light or meteorological data, lacking comprehensive monitoring of road conditions and traffic flow details, resulting in insufficient basis for control decisions;

[0004] Most light-emitting units are static coating structures with fixed light intensity and mode, which cannot be dynamically adjusted according to changes in environment and traffic, resulting in poor adaptability.

[0005] The control algorithms mostly use fixed parameter logic and do not combine digital twin and other technologies to achieve scenario-based simulation optimization, making it difficult to balance visual effects and energy consumption; the vehicle-road cooperative interaction capability is weak, and it is unable to link vehicles, traffic platforms and cross-departmental systems, resulting in a lag in response to emergencies such as accidents and severe weather.

[0006] The inadequate energy management and maintenance system, insufficient energy supply stability, and lack of predictive maintenance mechanisms affect the long-term reliable operation of the system. These problems make it difficult for existing self-illuminating road markings to meet the needs of intelligent transportation development in terms of both effectiveness and operational efficiency. Summary of the Invention

[0007] To address the aforementioned technical problems, this invention provides an intelligent control system and method for self-illuminating road marking coatings. The technical solution is as follows:

[0008] A smart control method for coatings used in self-illuminating road markings includes the following steps:

[0009] Step 1: Deploy a distributed intelligent sensor network in the self-illuminating road marking coating system. The distributed intelligent sensor network collects environmental perception data, traffic flow data, and road surface condition data, and transmits them to edge nodes. The edge nodes preprocess the data using the Kalman filter algorithm and extract key characterization parameters for different types of data.

[0010] Step 2: The flexible controllable light-emitting unit adopts a three-layer composite modular structure. The bottom layer is a flexible long afterglow substrate, the middle layer is an energy storage layer, and the top layer is an intelligent excitation layer. The flexible long afterglow substrate is Z-axis gradient doped with rare earth ions and quantum dot sensitizers. The modules are connected by magnetic interfaces.

[0011] Step 3: Construct digital twin models of road markings, environment, and traffic; run deep reinforcement learning algorithms at the edge; based on the key representation parameters and traffic prediction data from Step 1, output optimal luminous parameter decisions through the digital twin models; and form preliminary flexible and controllable luminous unit control commands.

[0012] Step 4: Construct an integrated vehicle-road-cloud collaborative interaction mechanism to link vehicles, transportation platforms, and cross-departmental systems, supplement real-time interactive data for the algorithm decision-making in Step 3, optimize the control commands of the flexible and controllable light-emitting unit, and output them.

[0013] Optionally, in step 1, the distributed intelligent sensor network includes a light intensity sensor, a weather sensor, a traffic flow sensor, a visibility sensor, and a road surface condition sensor.

[0014] Optionally, in step 1, the following parameters are extracted from the environmental perception data: light change rate, visibility level, and humidity change trend; traffic flow data: traffic flow density, average vehicle speed, proportion of large vehicles, and headway; and road surface condition data: water accumulation thickness change rate and friction coefficient gradient value. These parameters are then combined with time-domain and frequency-domain features for multi-dimensional characterization.

[0015] Optionally, a full lifecycle energy management system can be built to provide a continuous and stable energy supply and efficient scheduling for the sensing, calculation, execution and interaction processes in steps 1-4.

[0016] Optionally, an AI vision-assisted predictive maintenance system can be constructed to monitor the status of the light-emitting unit and the system, ensuring the long-term reliability of the hardware carrier in step 2 and the execution of instructions in step 3.

[0017] Optionally, in step 3, the process of forming preliminary flexible and controllable light-emitting unit control commands from the digital twin models of road markings, environment, and traffic includes:

[0018] Step 31: Multimodal feature input fusion of key representation parameters from Step 1, traffic flow prediction data based on LSTM neural network, coating health status data, and real-time vehicle data transmitted via V2X.

[0019] Step 32: The hierarchical control architecture is optimized by using edge real-time control and cloud strategy. Deep reinforcement learning algorithm is run at the edge to output LED brightness, flicker frequency and UV excitation power parameters.

[0020] Step 33: Simulate the luminous effect under different traffic scenarios using a digital twin model, verify and optimize the output parameters, set a dual objective function, and reduce energy consumption by a ratio greater than or equal to a set reduction threshold while ensuring that the visible distance is greater than or equal to a set distance threshold. Finally, convert the optimized luminous parameters into preliminary control commands.

[0021] Optionally, in step 4, the process of optimizing the control commands for the flexible controllable light-emitting unit includes:

[0022] Step 41: The autonomous vehicle passes through and sends the path planning intention to the road markings via short-range communication. The road markings adjust the illumination mode in the preliminary control command 50m-100m in advance. Ordinary vehicles adjust the local highlight parameters in the command by triggering the command through millimeter-wave radar.

[0023] Step 42: Connect to the city's smart transportation cloud platform and receive real-time event instructions, including accident instructions, construction instructions, and congestion instructions. Optimize the display parameters of the luminous color and dynamic warning pattern in the preliminary control instructions.

[0024] Step 43: Establish an API interface with the meteorological bureau and emergency management bureau, set the time in advance to receive severe weather warnings, and increase the luminous intensity parameter in the preliminary control command to the set speed of the normal value based on the specific data of the severe weather warning, so as to form the optimized final control command.

[0025] Optionally, the detailed control logic for establishing API interfaces with the Meteorological Bureau and Emergency Management Bureau in step 43 includes:

[0026] The relationship between weather type and visibility: heavy rain corresponds to visibility of 500m-1000m, dense fog corresponds to visibility of 200m-500m, and snow and ice corresponds to visibility of 100m-200m, which correspond to different light intensity adjustment levels.

[0027] The linkage formula is: ;

[0028] Where L is the adjusted luminous intensity. V represents normal luminous intensity, V represents real-time visibility, and K represents the warning level coefficient; blue warning K=1.0, yellow warning K=1.2, orange warning K=1.4, and red warning K=1.6.

[0029] Normal luminous intensity Value 500 cd / m 2 -800cd / m 2 Real-time visibility V is pushed in real time by the meteorological bureau's API, and the warning level coefficient K is dynamically matched according to the warning level issued by the emergency management bureau.

[0030] When visibility changes by 100m or the warning level is raised or lowered by one level, the system recalculates the L value and updates the control command within the set minimum time to ensure that the luminous intensity matches the environmental visibility in real time.

[0031] Optionally, after the initial control command is formed, an instruction execution scheduling stage is also included: the edge node converts the initial control command into a standardized JSON format command and sends it to the microcontroller of the flexible controllable light-emitting unit through the industrial Ethernet PROFINET protocol. The microcontroller synchronously drives the charging and discharging circuits of the RGB-LED array, UV-LED excitation module and energy storage module of the flexible controllable light-emitting unit, and at the same time collects and transmits the execution status data through current and voltage sensors to form a closed-loop feedback.

[0032] The intelligent control system for self-illuminating road marking coatings is used to realize the intelligent control method for self-illuminating road marking coatings. The intelligent control system includes a multi-dimensional sensing unit, a flexible and controllable light-emitting unit, a digital twin control unit, a vehicle-road-cloud collaborative interaction unit, and a full life cycle energy management unit.

[0033] The multi-dimensional perception unit consists of a distributed intelligent sensor network and edge nodes. It collects environmental perception data, traffic flow data, and road surface condition data. The edge nodes preprocess the data and extract key characterization parameters through the Kalman filter algorithm.

[0034] The flexible controllable light-emitting unit includes a flexible long-afterglow substrate, an energy storage layer, and an intelligent excitation layer. The flexible long-afterglow substrate is Z-axis gradient doped with rare earth ions and quantum dot sensitizers. The modules are connected by magnetic interfaces.

[0035] The digital twin control unit constructs a digital twin model of road markings, environment, and traffic. Deep reinforcement learning algorithms run at the edge and output preliminary control commands based on key representation parameters of the multi-dimensional perception subsystem and traffic prediction data. It also exchanges preliminary control commands with the flexible and controllable light-emitting unit and the vehicle-road-cloud collaborative interaction unit.

[0036] The vehicle-road-cloud collaborative interaction unit communicates with vehicles, connects to the city's smart transportation cloud platform, and establishes API interfaces with the meteorological bureau and emergency management bureau to supplement real-time interactive data and optimize initial control commands.

[0037] The lifecycle energy management unit provides a closed-loop energy supply for each unit, including energy collection, storage, scheduling, and recovery.

[0038] In summary, the present invention has at least one of the following beneficial technical effects:

[0039] This invention provides an intelligent control system and method for self-illuminating road marking coatings. By integrating technologies such as multi-dimensional perception, modular light-emitting units, digital twin control, and vehicle-road-cloud collaboration, the accuracy of control decisions is improved.

[0040] Flexible and controllable light-emitting units enable multi-dimensional adjustment of light-emitting parameters, enhancing environmental and traffic adaptability;

[0041] The adaptive algorithm driven by digital twins takes into account both line-of-sight distance and energy consumption optimization, significantly reducing energy consumption while ensuring that the line-of-sight distance is greater than or equal to a set threshold.

[0042] The integrated vehicle-road-cloud collaborative mechanism significantly reduces response delays in emergency scenarios and improves road safety;

[0043] A full lifecycle energy management and predictive maintenance system ensures continuous and stable system operation and reduces operation and maintenance costs.

[0044] The overall solution enhances the intelligence and practical value of self-illuminating road markings, providing strong support for the construction of smart transportation. Attached Figure Description

[0045] Figure 1 This is a flowchart illustrating the intelligent control method for coatings used in self-illuminating road markings according to the present invention.

[0046] Figure 2 This is a schematic diagram of the intelligent control system architecture for self-illuminating road markings according to a specific embodiment of the present invention; Detailed Implementation

[0047] The present invention will be further described in detail below with reference to the accompanying drawings.

[0048] This invention discloses an intelligent control system and control method for coatings used in self-illuminating road markings.

[0049] Reference Figure 1 and Figure 2 Example 1, a method for intelligent control of coatings used in self-illuminating road markings, includes the following steps:

[0050] Step 1: Deploy a distributed intelligent sensor network in the self-illuminating road marking coating system. The distributed intelligent sensor network collects environmental perception data, traffic flow data, and road surface condition data, and transmits them to edge nodes. The edge nodes preprocess the data using the Kalman filter algorithm and extract key characterization parameters for different types of data.

[0051] Step 2: The flexible controllable light-emitting unit adopts a three-layer composite modular structure. The bottom layer is a flexible long afterglow substrate, the middle layer is an energy storage layer, and the top layer is an intelligent excitation layer. The flexible long afterglow substrate is Z-axis gradient doped with rare earth ions and quantum dot sensitizers. The modules are connected by magnetic interfaces.

[0052] Step 3: Construct digital twin models of road markings, environment, and traffic; run deep reinforcement learning algorithms at the edge; based on the key representation parameters and traffic prediction data from Step 1, output optimal luminous parameter decisions through the digital twin models; and form preliminary flexible and controllable luminous unit control commands.

[0053] Step 4: Construct an integrated vehicle-road-cloud collaborative interaction mechanism to link vehicles, transportation platforms, and cross-departmental systems, supplement real-time interactive data for the algorithm decision-making in Step 3, optimize the control commands of the flexible and controllable light-emitting unit, and output them.

[0054] In Example 2, step 1, the distributed intelligent sensor network includes a light intensity sensor, a weather sensor, a traffic flow sensor, a visibility sensor, and a road surface condition sensor.

[0055] In Example 3, in step 1, the following parameters are extracted from the environmental perception data: light change rate, visibility level, and humidity change trend; traffic flow data: traffic flow density, average vehicle speed, proportion of large vehicles, and headway; and road surface condition data: water accumulation thickness change rate and friction coefficient gradient value. These parameters are then combined with time-domain and frequency-domain features for multi-dimensional characterization.

[0056] Example 4: A full lifecycle energy management system is built to provide a continuous and stable energy supply and efficient scheduling for the sensing, calculation, execution and interaction processes of steps 1-4.

[0057] Example 5: Construct an AI vision-assisted predictive maintenance system to monitor the status of the light-emitting unit and the system, ensuring the long-term reliability of the hardware carrier in step 2 and the execution of instructions in step 3.

[0058] By adopting the above technical solution, multi-source data is collected across the entire domain through a distributed intelligent sensor network. Different types of sensors capture environmental, traffic flow and road surface status information respectively. After being transmitted to the edge node, the Kalman filter algorithm uses state estimation to eliminate noise interference. Then, through feature extraction, key parameters that are strongly correlated with light emission control are screened from the massive data to provide accurate input for subsequent decision-making.

[0059] The three-layer composite modular structure is divided into different functions: the bottom flexible long afterglow substrate provides basic light-emitting energy storage performance; the middle energy storage layer realizes efficient energy caching; the top intelligent excitation layer is responsible for fine adjustment of light-emitting parameters; gradient doping of rare earth ions and quantum dot sensitizers improves excitation response efficiency through energy level transitions; and the magnetic interface simplifies module replacement and maintenance process through standardized connection.

[0060] A digital twin model is constructed to achieve synchronous mapping between the physical world and virtual space. The edge deep reinforcement learning algorithm is trained by interacting with the virtual scene, and outputs the optimal combination of luminous parameters by combining perception parameters and traffic prediction data, which is then converted into preliminary control commands.

[0061] By breaking down system silos through integrated vehicle-road-cloud collaboration, real-time data from vehicles, transportation platforms, and cross-departmental systems dynamically supplement decision-making basis, and scenario-based corrections are made to the initial instructions to ensure that the output instructions meet the actual application needs.

[0062] By using light intensity sensors to monitor changes in light intensity, meteorological sensors to collect environmental parameters such as temperature and humidity, traffic flow sensors to obtain vehicle traffic density and speed, visibility sensors to determine atmospheric transparency, and road surface condition sensors to detect road surface water accumulation, icing, and friction coefficient, multiple types of sensors work together to achieve full coverage of perception dimensions and avoid decision-making biases caused by incomplete data from a single sensor.

[0063] Differentiated key characterization parameters are extracted for different data characteristics. The rate of change of illumination and visibility level of environmental perception data reflect the dynamic trend of environmental change. The density and proportion of large vehicles of traffic flow data reflect the road traffic load. The rate of change of water accumulation thickness of road surface condition data reflects the road safety conditions. Then, by combining time domain features (mean, variance, peak value) and frequency domain features (Fourier transform main frequency component), the data can be characterized in multiple dimensions, improving feature recognition and decision relevance.

[0064] By capturing ambient light energy and vehicle rolling mechanical energy through multi-source energy harvesting devices (such as transparent perovskite solar thin films and piezoelectric power generation units), energy storage modules achieve energy buffer storage, and intelligent scheduling algorithms dynamically allocate energy according to the real-time energy consumption requirements of each link of perception, calculation, execution and interaction. The decommissioned modules are transferred to energy storage power stations through tiered utilization, forming an energy closed loop of collection-storage-scheduling-recovery, ensuring that each step of the system operates continuously and stably with optimal energy consumption.

[0065] Built-in fiber optic grating sensors and photoelectric sensors monitor the health status data of the light-emitting unit in real time, such as strain, temperature, and afterglow intensity. The drone inspection system is equipped with a high-resolution camera and thermal imager to scan the road markings along the entire section. Combined with the YOLOv8 algorithm, it identifies appearance defects such as coating damage and peeling. Then, it analyzes the status data and predicts the remaining lifespan through the Weibull-ARIMA fusion model. The intelligent operation and maintenance platform automatically generates maintenance work orders and matches the optimal resources according to the fault level, ensuring the long-term reliability of the hardware carrier and command execution.

[0066] Example 6, step 3, the process of forming preliminary flexible and controllable light-emitting unit control commands from the digital twin models of road markings, environment, and traffic includes:

[0067] Step 31: Multimodal feature input fusion of key representation parameters from Step 1, traffic flow prediction data based on LSTM neural network, coating health status data, and real-time vehicle data transmitted via V2X.

[0068] Step 32: The hierarchical control architecture is optimized by using edge real-time control and cloud strategy. Deep reinforcement learning algorithm is run at the edge to output LED brightness, flicker frequency and UV excitation power parameters.

[0069] Step 33: Simulate the luminous effect under different traffic scenarios using a digital twin model, verify and optimize the output parameters, set a dual objective function, and reduce energy consumption by a ratio greater than or equal to a set reduction threshold while ensuring that the visible distance is greater than or equal to a set distance threshold. Finally, convert the optimized luminous parameters into preliminary control commands.

[0070] By adopting the above technical solution, a single type of data cannot fully reflect the complexity of the actual scenario. Therefore, the key characterization parameters of step 1 (environment, traffic, and road surface basic conditions), LSTM neural network traffic flow prediction data (based on historical traffic data to train models to predict short-term traffic conditions), coating health status data (afterglow decay rate, yellowing index, etc. reflecting the performance of light-emitting units) and V2X transmitted real-time vehicle data (vehicle speed, heading angle, and other dynamic vehicle information) are integrated. Through the complementarity of multi-dimensional data, a comprehensive and dynamic input basis is provided for control decisions, avoiding decision bias caused by single data.

[0071] The edge and cloud perform different functions to balance real-time performance and optimization. The edge, located close to the sensing and execution devices, runs deep reinforcement learning (DQN) algorithms to achieve low-latency decision-making. It continuously interacts with the environment to train the model and outputs parameters that can directly drive the hardware, such as LED brightness, flicker frequency, and UV excitation power. The cloud has powerful computing capabilities and updates control strategy parameters every 24 hours based on massive amounts of scene data. These updates are then sent to the edge to optimize the algorithm model, forming a highly efficient decision-making mode of real-time response and strategy iteration.

[0072] A digital twin model is constructed to create a virtual scene that maps 1:1 to the physical world. This virtual scene can simulate the luminous effects under different environmental conditions (such as heavy rain and fog) and traffic conditions (such as peak-hour congestion and off-peak sparse traffic). The parameters output from the edge are then virtually verified. A dual objective function is set: a visibility distance of ≥180m is used as a safety constraint, and an energy consumption reduction of ≥35% is used as an energy efficiency target. Through iterative adjustment of parameters in the virtual scene, the optimal combination of luminous parameters that meets both safety requirements and energy conservation is selected. Finally, this is converted into standardized preliminary control commands, laying the foundation for subsequent optimization.

[0073] In Example 7, step 4, the process of optimizing the control commands for the flexible controllable light-emitting unit includes:

[0074] Step 41: The autonomous vehicle passes through and sends the path planning intention to the road markings via short-range communication. The road markings adjust the illumination mode in the preliminary control command 50m-100m in advance. Ordinary vehicles adjust the local highlight parameters in the command by triggering the command through millimeter-wave radar.

[0075] Step 42: Connect to the city's smart transportation cloud platform and receive real-time event instructions, including accident instructions, construction instructions, and congestion instructions. Optimize the display parameters of the luminous color and dynamic warning pattern in the preliminary control instructions.

[0076] Step 43: Establish an API interface with the meteorological bureau and emergency management bureau, set the time in advance to receive severe weather warnings, and increase the luminous intensity parameter in the preliminary control command to the set speed of the normal value based on the specific data of the severe weather warning, so as to form the optimized final control command.

[0077] Example 8, the detailed control logic for establishing API interfaces with the Meteorological Bureau and Emergency Management Bureau in step 43 includes:

[0078] The relationship between weather type and visibility: heavy rain corresponds to visibility of 500m-1000m, dense fog corresponds to visibility of 200m-500m, and snow and ice corresponds to visibility of 100m-200m, which correspond to different light intensity adjustment levels.

[0079] The linkage formula is: ;

[0080] Where L is the adjusted luminous intensity. V represents normal luminous intensity, V represents real-time visibility, and K represents the warning level coefficient; blue warning K=1.0, yellow warning K=1.2, orange warning K=1.4, and red warning K=1.6.

[0081] Normal luminous intensity Value 500 cd / m 2 -800cd / m 2 Real-time visibility V is pushed in real time by the meteorological bureau's API, and the warning level coefficient K is dynamically matched according to the warning level issued by the emergency management bureau.

[0082] When visibility changes by 100m or the warning level is raised or lowered by one level, the system recalculates the L value and updates the control command within the set minimum time to ensure that the luminous intensity matches the environmental visibility in real time.

[0083] By adopting the above technical solutions, differentiated instruction adjustments are achieved to meet the traffic needs of different types of vehicles. Autonomous vehicles have the ability to transmit path planning information, sending their intentions to road markings 50m-100m in advance via short-range communication, enabling the road markings to anticipate and adjust their illumination modes (such as curve guidance and lane keeping prompts), thus improving the accuracy of autonomous driving decisions. Ordinary vehicles lack active information transmission capabilities, so millimeter-wave radar detects the approach of vehicles and triggers adjustments to local highlighting parameters to ensure that the driver can clearly identify the location of the road markings. These two interaction methods cover different vehicle types, achieving full-scenario vehicle adaptation.

[0084] The city's intelligent transportation cloud platform gathers real-time event information from all road sections. Once connected to the platform, instructions regarding accidents, construction, and congestion can be obtained promptly. By mapping this event information to adjustable parameters for luminous colors (such as red warnings and yellow alerts) and dynamic warning patterns (such as text labels and arrow guidance), road markings are upgraded from simple guidance functions to traffic event warning carriers, improving road traffic safety and event handling efficiency.

[0085] Severe weather is a key factor affecting the visibility of road markings. By establishing an API interface with the meteorological bureau and emergency management bureau, weather warning information can be obtained in advance at set times. Based on the warning data, the luminous intensity parameters can be adjusted to increase the intensity in the initial instructions to a set multiple of the normal value, ensuring that the road markings still have sufficient visibility distance in low visibility conditions and avoiding safety risks caused by weather factors in advance.

[0086] A quantitative correlation mechanism is used to achieve precise matching between luminous intensity and severe weather. First, the correspondence between weather type and visibility is established based on the different impacts of various severe weather conditions on atmospheric transparency, providing a basic classification for luminous intensity adjustment. The linkage formula integrates three key factors: normal luminous intensity, real-time visibility, and warning level. Through mathematical modeling, the intensity is quantitatively calculated. Visibility V is negatively correlated with intensity (lower visibility means higher intensity), and the warning level coefficient K reflects the additive effect of the urgency of the disaster on the intensity. The range of normal luminous intensity values ​​is set by combining daily visibility needs with energy consumption balance. The dynamic adjustment mechanism monitors changes in visibility and warning levels, triggering real-time updates of intensity parameters to ensure that luminous intensity always maintains an optimal match with the current environmental visibility, meeting safety requirements while avoiding energy waste.

[0087] Example 9: After the initial control command is formed, the command execution scheduling stage is also included: the edge node converts the initial control command into a standardized JSON format command and sends it to the microcontroller of the flexible controllable light-emitting unit through the industrial Ethernet PROFINET protocol. The microcontroller synchronously drives the charging and discharging circuits of the RGB-LED array, UV-LED excitation module and energy storage module of the flexible controllable light-emitting unit, and at the same time collects and transmits the execution status data through current and voltage sensors to form a closed-loop feedback.

[0088] By adopting the above technical solution, the edge node converts the initial control commands into standardized JSON format commands. The principle is that JSON format is lightweight and easy to parse, enabling universal data interaction between different devices. The commands are then transmitted to the microcontroller via the PROFINET industrial Ethernet protocol. This protocol offers low latency and high reliability, meeting the real-time transmission requirements of control commands. The microcontroller synchronously drives the charging and discharging circuits of the RGB-LED array, UV-LED excitation module, and energy storage module. The principle is that the microcontroller, as the hardware control core, achieves coordinated action of multiple execution units through precise timing control, ensuring synchronized adjustment of parameters such as LED brightness and excitation power. Simultaneously, current and voltage sensors collect and transmit execution status data back, monitoring hardware operating parameters (such as actual brightness and module power consumption) in real time, forming a closed-loop feedback chain. This allows the edge node to promptly detect command execution deviations and correct parameters, ensuring the accuracy and stability of the control effect.

[0089] Example 10: Intelligent control system for self-illuminating road marking coatings, used to implement intelligent control method for self-illuminating road marking coatings. The intelligent control system includes a multi-dimensional sensing unit, a flexible and controllable light-emitting unit, a digital twin control unit, a vehicle-road-cloud collaborative interaction unit, and a full life cycle energy management unit.

[0090] The multi-dimensional perception unit consists of a distributed intelligent sensor network and edge nodes. It collects environmental perception data, traffic flow data, and road surface condition data. The edge nodes preprocess the data and extract key characterization parameters through the Kalman filter algorithm.

[0091] The flexible controllable light-emitting unit includes a flexible long-afterglow substrate, an energy storage layer, and an intelligent excitation layer. The flexible long-afterglow substrate is Z-axis gradient doped with rare earth ions and quantum dot sensitizers. The modules are connected by magnetic interfaces.

[0092] The digital twin control unit constructs a digital twin model of road markings, environment, and traffic. Deep reinforcement learning algorithms run at the edge and output preliminary control commands based on key representation parameters of the multi-dimensional perception subsystem and traffic prediction data. It also exchanges preliminary control commands with the flexible and controllable light-emitting unit and the vehicle-road-cloud collaborative interaction unit.

[0093] The vehicle-road-cloud collaborative interaction unit communicates with vehicles, connects to the city's smart transportation cloud platform, and establishes API interfaces with the meteorological bureau and emergency management bureau to supplement real-time interactive data and optimize initial control commands.

[0094] The lifecycle energy management unit provides a closed-loop energy supply for each unit, including energy collection, storage, scheduling, and recovery.

[0095] The following specific embodiments illustrate the implementation principle of the present invention:

[0096] Step 1: Deploy a self-illuminating road marking coating system on a section of an urban expressway. A distributed intelligent sensor network is set up with sensing nodes spaced 50 meters apart. Light intensity sensors are installed on lampposts on both sides of the road, 6 meters above the ground. Weather sensors are deployed at the highest points along the road. Traffic flow sensors are embedded in the inner edges of lane lines. Visibility sensors are installed on gantries at the road entrances and exits. Road surface condition sensors are embedded in the road surface. Each sensor collects data every 100 milliseconds and transmits it to the edge nodes along the road via a LoRa gateway. The edge nodes use a Kalman filter algorithm to estimate the state of the raw data, eliminating noise from vehicle vibration and electromagnetic interference. Then, for the environmental sensing data, extract the rate of change of illumination, visibility level, and humidity trend; for the traffic flow data, extract traffic flow density, average vehicle speed, proportion of large vehicles, and headway; and for the road surface condition data, extract the rate of change of water accumulation thickness and the gradient value of the friction coefficient. Simultaneously, multi-dimensional characterization is achieved by combining the mean, variance, and peak values ​​from the time-domain features and the Fourier transform main frequency component from the frequency-domain features.

[0097] Step 2: Flexible, controllable light-emitting units are continuously laid along the road markings, employing a three-layer composite modular structure. The bottom flexible long-afterglow substrate is made of silicate-based materials, with europium ions and perovskite quantum dot sensitizers doped in a Z-axis gradient from the surface to the inner layer at concentrations of 0.1%, 0.3%, and 0.5%. The middle energy storage layer uses a solid-state lithium battery pack with a capacity of 5000mAh. The top intelligent excitation layer integrates an RGB-LED array and a UV-LED excitation module. The modules are connected via standard magnetic interfaces, with positioning pins at the interfaces to ensure installation accuracy. Each module measures 100cm × 15cm × 2cm and weighs no more than 3kg.

[0098] Step 3: Construct a digital twin model of road markings, environment, and traffic. Based on BIM technology, establish a 3D model of the road segment, mapping real-time data collected by sensors to the virtual scene at a frequency of 20Hz. Deploy a deep reinforcement learning algorithm at the edge, using DQN as the core network. The input layer receives the key representation parameters from Step 1 and the 15-minute short-term traffic flow prediction data output by the LSTM neural network. The hidden layer consists of three fully connected layers, and the output layer outputs LED brightness (0-1000 cd / m²). 2 The parameters include flicker frequency (0-10Hz) and UV excitation power (0-50W). Using a digital twin model, 200 typical scenarios such as heavy rain, dense fog, and morning / evening rush hour were simulated to verify and optimize the output parameters. A dual objective function was set to reduce energy consumption by at least 35% while ensuring a visibility distance of at least 180m. Finally, the optimized luminescence parameters were converted into preliminary control commands.

[0099] Step 4: Establish an integrated vehicle-road-cloud collaborative interaction mechanism. Autonomous vehicles send their path planning intentions to road markings via C-V2X short-range communication. 80 meters in advance, the road markings adjust their illumination mode in the initial control command. When a vehicle is detected entering a curve, the corresponding road marking LED array switches to an arrow directional pattern. Ordinary vehicles adjust local high-brightness parameters in the trigger command via millimeter-wave radar. The radar detection distance is set to 50 meters, and when a vehicle approaches, the brightness of the LEDs in the local area increases by 50%. Simultaneously, the system connects to the city's smart transportation cloud platform to receive real-time event commands such as accidents, construction, and congestion. When an accident command is received at K2+300, the illumination color of road markings within 500 meters before and after that location is adjusted to red, and a dynamic detour warning pattern is displayed.

[0100] An API interface has been established with the Meteorological Bureau and the Emergency Management Bureau to receive severe weather warnings two hours in advance. Based on the correlation between weather type and visibility, different luminous intensity adjustment levels are corresponding to heavy rain (visibility 500m-1000m), dense fog (visibility 200m-500m), and snow / ice (visibility 100m-200m), respectively. This is combined with the linkage formula L=L0×2-0.001×V×K (L0 is taken as 650cd / m²). 2 V is pushed in real time by the meteorological bureau's API, and K is dynamically matched according to the warning level. The adjusted luminous intensity is calculated to form the optimized final control command.

[0101] Step 5: The full lifecycle energy management unit installs transparent perovskite solar films (50W per film) on the tops of the light poles on both sides of the road, embeds piezoelectric generator units (10W maximum output per unit) at the speed bump locations, and uses lithium iron phosphate battery packs (total capacity 100kWh) for energy storage. The intelligent scheduling algorithm dynamically allocates energy every 5 minutes based on the real-time energy consumption needs of each stage of perception, calculation, execution, and interaction. When sunlight is sufficient, the electricity generated by the solar film is prioritized for the light-emitting units, and the remaining electricity is stored in the energy storage module. When sunlight is insufficient, the energy storage module discharges to supply power, while the piezoelectric generator units supplement the energy. Retired energy storage modules are reused in surrounding community energy storage power stations, forming an energy closed loop.

[0102] Step 6: The AI ​​vision-assisted predictive maintenance system incorporates fiber optic grating sensors and photoelectric sensors to monitor in real time the strain (measurement range 0-5000με), temperature (measurement range -40℃-85℃), and afterglow intensity (measurement range 0-500mcd / m²) of the light-emitting unit. 2A weekly drone inspection system, equipped with a 20-megapixel high-resolution camera and thermal imager, scans the entire road markings. Using the YOLOv8 algorithm, it identifies surface defects such as coating damage and peeling, achieving an accuracy of ≥95%. The Weibull-ARIMA fusion model analyzes the status data to predict the remaining lifespan of the light-emitting units, with a prediction error of ≤10%. The intelligent operation and maintenance platform automatically generates maintenance work orders based on the fault level (minor, moderate, severe). Minor faults are handled by personnel within 24 hours, while severe faults are responded to within 2 hours.

[0103] Step 7, in the instruction execution scheduling stage, the edge node converts the initial control instructions into standardized JSON format instructions and sends them to the microcontroller (model STM32H743) of the flexible controllable light-emitting unit via the PROFINET industrial Ethernet protocol (transmission rate 100Mbps, latency less than or equal to 1ms). The microcontroller synchronously drives the charging and discharging circuits of the RGB-LED array, UV-LED excitation module, and energy storage module of the flexible controllable light-emitting unit, with a driving accuracy of less than or equal to ±2%. At the same time, the execution status data is collected by current and voltage sensors (measurement range 0-5A, 0-36V) and sent back to the edge node, forming a closed-loop feedback. When the actual brightness deviates from the commanded brightness by more than 10%, the edge node immediately corrects the control parameters.

[0104] The above are all preferred embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Therefore, all equivalent changes made in accordance with the structure, shape and principle of the present invention should be covered within the scope of protection of the present invention.

Claims

1. A method for intelligent control of coatings used in self-illuminating road markings, characterized in that, Includes the following steps: Step 1: Deploy a distributed intelligent sensor network in the self-illuminating road marking coating system. The distributed intelligent sensor network collects environmental perception data, traffic flow data, and road surface condition data, and transmits them to edge nodes. The edge nodes preprocess the data using the Kalman filter algorithm and extract key characterization parameters for different types of data. Step 2: The flexible controllable light-emitting unit adopts a three-layer composite modular structure. The bottom layer is a flexible long afterglow substrate, the middle layer is an energy storage layer, and the top layer is an intelligent excitation layer. The flexible long afterglow substrate is Z-axis gradient doped with rare earth ions and quantum dot sensitizers. The modules are connected by magnetic interfaces. Step 3: Construct digital twin models of road markings, environment, and traffic; run deep reinforcement learning algorithms at the edge; based on the key representation parameters and traffic prediction data from Step 1, output optimal luminous parameter decisions through the digital twin models; and form preliminary flexible and controllable luminous unit control commands. Step 3, the process of forming preliminary flexible and controllable light-emitting unit control commands from the digital twin models of road markings, environment, and traffic, includes: Step 31: Multimodal feature input fusion of key representation parameters from Step 1, traffic flow prediction data based on LSTM neural network, coating health status data, and real-time vehicle data transmitted via V2X. Step 32: The hierarchical control architecture is optimized by using edge real-time control and cloud strategy. Deep reinforcement learning algorithm is run at the edge to output LED brightness, flicker frequency and UV excitation power parameters. Step 33: Simulate the luminous effect under different traffic scenarios using a digital twin model, verify and optimize the output parameters, set a dual objective function, and reduce energy consumption by a ratio greater than or equal to a set reduction threshold while ensuring that the visible distance is greater than or equal to a set distance threshold. Finally, convert the optimized luminous parameters into preliminary control commands. Step 4: Construct a vehicle-road-cloud integrated collaborative interaction mechanism to link vehicles, transportation platforms and cross-departmental systems, supplement real-time interactive data for the algorithm decision-making in Step 3, optimize the control commands of the flexible and controllable light-emitting unit and output them. Step 4, the process of optimizing the control commands for the flexible controllable light-emitting unit, includes: Step 41: The autonomous vehicle passes through and sends the path planning intention to the road markings via short-range communication. The road markings adjust the illumination mode in the preliminary control command 50m-100m in advance. Ordinary vehicles adjust the local highlight parameters in the command by triggering the command through millimeter-wave radar. Step 42: Connect to the city's smart transportation cloud platform and receive real-time event instructions, including accident instructions, construction instructions, and congestion instructions. Optimize the display parameters of the luminous color and dynamic warning pattern in the preliminary control instructions. Step 43: Establish an API interface with the meteorological bureau and emergency management bureau, set the time in advance to receive severe weather warnings, and increase the luminous intensity parameter in the preliminary control command to the set speed of the normal value based on the specific data of the severe weather warning, so as to form the optimized final control command.

2. The intelligent control method for coatings used in self-illuminating road markings according to claim 1, characterized in that, In step 1, the distributed intelligent sensor network includes light intensity sensors, weather sensors, traffic flow sensors, visibility sensors, and road condition sensors.

3. The intelligent control method for coatings used in self-illuminating road markings according to claim 2, characterized in that, In step 1, the following parameters are extracted from the environmental perception data: light change rate, visibility level, and humidity change trend; traffic flow data: traffic flow density, average vehicle speed, proportion of large vehicles, and headway; and road surface condition data: water accumulation thickness change rate and friction coefficient gradient value. These parameters are then combined with time-domain and frequency-domain features for multi-dimensional characterization.

4. The intelligent control method for coatings used in self-illuminating road markings according to claim 3, characterized in that, Establish a full lifecycle energy management system to provide continuous and stable energy supply and efficient scheduling for the sensing, calculation, execution and interaction processes in steps 1-4.

5. The intelligent control method for coatings used in self-illuminating road markings according to claim 4, characterized in that, Construct an AI vision-assisted predictive maintenance system to monitor the status of the light-emitting unit and the system, ensuring the long-term reliability of the hardware carrier in step 2 and the execution of instructions in step 3.

6. The intelligent control method for coatings used in self-illuminating road markings according to claim 5, characterized in that, The detailed control logic for establishing API interfaces with the Meteorological Bureau and Emergency Management Bureau in step 43 includes: The relationship between weather type and visibility: heavy rain corresponds to visibility of 500m-1000m, dense fog corresponds to visibility of 200m-500m, and snow and ice corresponds to visibility of 100m-200m, which correspond to different light intensity adjustment levels. The linkage formula is: ; Where L is the adjusted luminous intensity. V represents normal luminous intensity, V represents real-time visibility, and K represents the warning level coefficient; blue warning K=1.0, yellow warning K=1.2, orange warning K=1.4, and red warning K=1.

6. Normal luminous intensity The value ranges from 500 cd / m² to 800 cd / m². Real-time visibility V is pushed in real time by the meteorological bureau's API, and the warning level coefficient K is dynamically matched according to the warning level issued by the emergency management bureau. When visibility changes by 100m or the warning level is raised or lowered by one level, the system recalculates the L value and updates the control command within the set minimum time to ensure that the luminous intensity matches the environmental visibility in real time.

7. The intelligent control method for coatings used in self-illuminating road markings according to claim 6, characterized in that, After the initial control command is formed, the command execution scheduling stage is also included: the edge node converts the initial control command into a standardized JSON format command and sends it to the microcontroller of the flexible controllable light-emitting unit through the industrial Ethernet PROFINET protocol. The microcontroller synchronously drives the charging and discharging circuits of the RGB-LED array, UV-LED excitation module and energy storage module of the flexible controllable light-emitting unit, and at the same time collects and transmits the execution status data through current and voltage sensors to form a closed-loop feedback.

8. An intelligent control system for coatings used in self-illuminating road markings, characterized in that: The intelligent control system for implementing the coating intelligent control method for self-illuminating road markings according to any one of claims 1-7 includes a multi-dimensional sensing unit, a flexible and controllable light-emitting unit, a digital twin control unit, a vehicle-road-cloud collaborative interaction unit, and a full life cycle energy management unit. The multi-dimensional perception unit consists of a distributed intelligent sensor network and edge nodes. It collects environmental perception data, traffic flow data, and road surface condition data. The edge nodes preprocess the data and extract key characterization parameters through the Kalman filter algorithm. The flexible controllable light-emitting unit includes a flexible long-afterglow substrate, an energy storage layer, and an intelligent excitation layer. The flexible long-afterglow substrate is Z-axis gradient doped with rare earth ions and quantum dot sensitizers. The modules are connected by magnetic interfaces. The digital twin control unit constructs a digital twin model of road markings, environment, and traffic. Deep reinforcement learning algorithms run at the edge and output preliminary control commands based on key representation parameters of the multi-dimensional perception subsystem and traffic prediction data. It also exchanges preliminary control commands with the flexible and controllable light-emitting unit and the vehicle-road-cloud collaborative interaction unit. The vehicle-road-cloud collaborative interaction unit communicates with vehicles, connects to the city's smart transportation cloud platform, and establishes API interfaces with the meteorological bureau and emergency management bureau to supplement real-time interactive data and optimize initial control commands. The lifecycle energy management unit provides a closed-loop energy supply for each unit, including energy collection, storage, scheduling, and recovery.

Images