Street lamp vehicle-road cooperative control system

Abstract

The invention, which relates to the technical field of intelligent traffic and illumination control, discloses a street lamp vehicle-road cooperative control system comprising a data acquisition module, a data processing module, a control execution module and a decision management module. The data acquisition module is used for acquiring traffic signal lamp phase, vehicle position and speed, street lamp working state and environment illumination intensity data; according to the invention, through a cross-system collaborative algorithm, the traffic flow direction is predicted in combination with the phase of the traffic signal lamp, the brightness of the street lamp illumination area is adjusted in advance, the energy distribution according to needs is realized, the energy consumption is effectively reduced, and through a vehicle-lamp-cloud three-level decision-making mechanism, the computing power priority is dynamically distributed according to the traffic condition. Traffic signal lamp control and vehicle driving path planning are optimized, the road passing efficiency is improved, traffic congestion is relieved, by pushing the lighting optimization path to the vehicle, the vehicle is driven under the proper lighting condition, the emergency brake risk is reduced, and the road traffic safety level is remarkably improved.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent transportation and lighting control, and specifically provides a road lamp vehicle-road collaborative control system. Background Art

[0002] As an important part of urban infrastructure, road lamps refer to lighting devices installed in public places such as roads, streets, and squares. Their core function is to drive lighting sources through electricity to provide lighting at night or in low-light environments, ensuring the travel safety of pedestrians and vehicles and enhancing the night visibility of public spaces. In the early days, high-pressure sodium lamps were mostly used for road lamps, which had the characteristics of high luminous efficiency and long lifespan; with the development of technology, LED road lamps have gradually become popular and have become the mainstream choice for modern urban lighting due to their energy-saving, environmental protection, and good light color advantages. Vehicle-road collaborative control is one of the key technologies of intelligent transportation systems. Based on technologies such as 5G, the Internet of Things, and sensors, an information interaction network between vehicles and vehicles, vehicles and roads, vehicles and people, and vehicles and the cloud is constructed. By real-time collecting and analyzing information such as vehicle operation status, road conditions, and traffic signals, dynamic perception and intelligent regulation of traffic flow are realized, aiming to improve traffic efficiency, reduce the risk of traffic accidents, and provide support for the development of autonomous driving technology.

[0003] However, there are still certain defects in the existing technology. The existing road lamp lighting system has a low degree of intelligence, mostly adopting a fixed brightness control mode, unable to adjust in real time and dynamically according to traffic flow, ambient light intensity, road conditions, etc., resulting in serious energy waste. The traffic signal system and the road lamp lighting system are independent of each other, lacking an information interaction and collaboration mechanism. During peak traffic hours, vehicles start and stop frequently, exacerbating traffic congestion. Vehicle-road collaborative control lacks overall optimization of vehicle driving paths and lighting conditions, and is unable to effectively reduce the risk of sudden braking caused by insufficient lighting. Therefore, it is of great significance to develop a road lamp vehicle-road collaborative control system. Summary of the Invention

[0004] The purpose of the present invention is to make up for the deficiencies of the existing technology and provide a road lamp vehicle-road collaborative control system. It can, through a cross-system collaborative algorithm, combine traffic signal phases to predict the traffic flow direction, adjust the brightness of the road lamp lighting area in advance, and achieve reasonable utilization of energy. Through a vehicle-lamp-cloud three-level decision-making mechanism, dynamically allocate the computing power priority, optimize traffic flow, relieve congestion, reduce the risk of sudden braking by pushing optimized lighting paths to vehicles, and improve road safety, thereby effectively solving the deficiencies in the existing technology.

[0005] To solve the above technical problems, the present invention provides the following technical solution: A road lamp vehicle-road collaborative control system, the system includes: a data collection module, a data processing module, a control execution module, and a decision-making management module; The data acquisition module is used to collect data on traffic signal phases, vehicle positions and speeds, street lamp operating states, and environmental light intensities; The data processing module includes a traffic flow direction prediction unit and an illumination optimization calculation unit. The traffic flow direction prediction unit uses a spatio-temporal convolutional attention network algorithm. By calculating the traffic flow feature matrix , where is the activation function, is the weight matrix, and respectively represent convolutional operations in the time dimension and the space dimension, is the input data, is the data sequence length, is the bias term. The illumination optimization calculation unit generates a street lamp illumination adjustment plan and a vehicle illumination optimization path according to the traffic flow direction prediction result and in combination with a preset illumination model; The control execution module adjusts the street lamp illumination brightness according to the instruction output by the data processing module and pushes the illumination optimization path to the vehicle; The decision management module realizes three-level decision-making for the roadside, street lamps, and cloud platform, and dynamically manages the operation of the system.

[0006] Furthermore, the data acquisition module includes a roadside data acquisition unit and a street lamp data acquisition unit. The roadside data acquisition unit uses a fusion acquisition method of millimeter-wave radar and high-definition camera. Through the Kalman filtering algorithm, the vehicle speed, distance, and angle information obtained by the millimeter-wave radar and the vehicle type, position, and driving trajectory information obtained by the high-definition camera are fused and processed. The street lamp data acquisition unit integrates a light sensor and a current sensor. The light sensor monitors the environmental light intensity based on the principle of a silicon photocell, and the current sensor detects the street lamp operating current through the Hall effect. Both transmit the data to the data processing module through the data transmission channel of the 5G-MEC network.

[0007] Even further, in the spatio-temporal convolutional attention network algorithm, the attention weight matrix is calculated through , where , are the query matrix and the key matrix respectively, is the dimension of the key matrix. In the model training stage, through the backpropagation algorithm and the stochastic gradient descent method, with the goal of minimizing the traffic flow prediction error, the parameters in the attention weight matrix are iteratively optimized, and a regularization term is introduced. By adjusting the regularization coefficient, the fitting ability and generalization ability of the model are balanced.

[0008] Even further, the preset illumination model of the illumination optimization calculation unit considers the road curvature , vehicle speed and lighting distance relationship, by determine the street lamp brightness adjustment coefficient , where , are empirical coefficients determined by performing multiple linear regression analysis on road curvature, speed, lighting distance, and driver visual feedback data under different road types and traffic flow conditions.

[0009] Furthermore, the control execution module includes a street lamp control sub-module and a vehicle networking communication sub-module. The street lamp control sub-module uses PWM dimming technology to adjust the street lamp drive current by setting different duty cycles according to the instructions of the data processing module. The duty cycle adjustment step size is determined according to the street lamp response time and dimming accuracy requirements. The vehicle networking communication sub-module is based on UWB positioning technology and uses the two-way ranging TWR algorithm to achieve high-precision communication between the vehicle and the roadside unit.

[0010] Furthermore, the decision management module includes a roadside decision unit. When a traffic accident is detected, the roadside decision unit controls the street lamps around the accident to enter the high-brightness constant-brightness mode according to the preset emergency lighting strategy, and pushes warning information to surrounding vehicles through the vehicle networking communication sub-module. In the preset emergency lighting strategy, the brightness value of the high-brightness constant-brightness mode is determined according to the brightness-rescue efficiency model established based on historical accident rescue data and human eye visual characteristics, and the roadside decision unit dynamically adjusts the push range and frequency of the warning information according to the severity and impact range of the accident.

[0011] Furthermore, the decision management module also includes a cloud platform decision unit. The cloud platform decision unit is used to construct a traffic congestion assessment model and comprehensively analyze the traffic flow density and vehicle average delay time data of each road section. In the traffic congestion assessment model, the congestion index is calculated by weighted summation of each data index, and the weight of each data index is calculated by the analytic hierarchy process AHP according to the judgment matrix constructed by traffic domain experts' scoring of the importance of each index. When the congestion index of a certain road section exceeds the threshold, computing power resources are preferentially allocated to perform traffic flow direction prediction and lighting optimization calculation for that road section.

[0012] Furthermore, the data processing module also includes a data preprocessing unit. The data preprocessing unit normalizes the data transmitted by the data acquisition module, processes missing data using the multiple imputation method, generates multiple imputation values according to the data distribution characteristics and correlations and then comprehensively analyzes them, and uses the density-based local outlier factor detection algorithm LOF for abnormal data, and determines whether it is an abnormal point by calculating the local reachability density of the data point and performs correction and elimination.

[0013] Furthermore, the vehicle networking communication sub-module adopts a joint optimization algorithm of channel coding and interference cancellation. When transmitting the lighting optimization path information, according to the coding rate - SNR - BER relationship model constructed based on a large amount of channel simulation data and actual test data, through the look-up table method and interpolation method, the coding rate is dynamically adjusted according to the requirements of the channel signal-to-noise ratio SNR and the bit error rate BER. The interference cancellation parameters are adjusted in real time according to the received signal characteristics by using the minimum mean square error MMSE algorithm to adjust the coefficients of the interference cancellation filter.

[0014] Compared with the prior art, the road lamp vehicle-road collaborative control system has the following beneficial effects: Through the cross-system collaborative algorithm, the present invention combines the traffic signal phase to predict the traffic flow direction, adjusts the brightness of the road lamp lighting area in advance, realizes the on-demand distribution of energy, effectively reduces energy consumption, and through the vehicle-lamp-cloud three-level decision-making mechanism, dynamically allocates the computing power priority according to the traffic conditions, optimizes the traffic signal control and vehicle driving path planning, improves the road traffic efficiency, alleviates traffic congestion, and by pushing the lighting optimization path to the vehicle, enables the vehicle to drive under suitable lighting conditions, reduces the risk of sudden braking, and significantly improves the road traffic safety level.

[0015] Other advantages, objectives and features of the present invention will be described to some extent in the subsequent specification, and to some extent, will be obvious to those skilled in the art based on the study of the following text, or can be taught from the practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments of the present invention, and those of ordinary skill in the art can obtain other drawings based on these drawings without creative efforts.

[0017] Figure 1 It is a schematic structural diagram of a road lamp vehicle-road collaborative control system; Figure 2 It is a schematic working flow diagram of a road lamp vehicle-road collaborative control system. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0018] In order to further elaborate on the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following will describe in detail the specific embodiments, structures, features and their effects of the present invention in combination with the accompanying drawings and preferred embodiments. Embodiment 1

[0019] See Figure 1 and Figure 2, on the main road connecting the commercial area and the residential area in a certain city, the daily traffic volume of this section is large, and the traffic congestion is obvious during the morning and evening rush hours. In addition, there are multiple intersections and branch road merging points along the way. The traditional street lights use a fixed brightness lighting mode, which cannot be adjusted according to the actual traffic conditions and environmental light, resulting in energy waste and difficulty in ensuring the safety of vehicle driving at night. In addition, the traffic signal system and the street light lighting system are independent of each other, lacking information interaction and coordinated control. When vehicles pass through intersections or merge into branch roads, due to insufficient lighting or mismatched traffic information, there is a risk of sudden braking and a high potential for traffic accidents. The road-light vehicle-road collaborative control system of the present invention is applied to this section, aiming to solve problems such as traffic congestion, energy waste, and potential safety hazards.

[0020] The roadside data collection unit is deployed on both sides of the road. The millimeter-wave radar is installed on a pole at a specific height from the ground to ensure that its detection range covers the entire width of the road and is not blocked by other objects. The high-definition camera is installed on the same pole as the millimeter-wave radar, and the angle is adjusted to obtain a clear road image for collecting vehicle type, position, and driving trajectory information. The light sensor is installed on the top of the street light pole to sense the environmental light intensity without obstruction; the current sensor is connected in series in the street light power supply line to accurately detect the working current of the street light. After all sensors are installed, the wiring is connected and debugged to ensure normal data collection.

[0021] Set up the hardware equipment of the data processing module in the urban traffic control center, including a high-performance server cluster for running the relevant algorithms of the traffic flow direction prediction unit, lighting optimization calculation unit, and data preprocessing unit. The server is configured with multiple high-performance CPUs, a large-capacity memory, and high-speed storage devices to meet the needs of a large amount of data processing. Install an operating system, a database management system, and a deep learning framework on the server, and deploy the trained spatio-temporal convolutional attention network model and other relevant algorithm programs to ensure the stable and efficient operation of the data processing module.

[0022] The street light control sub-module is integrated in the street light controller. The controller is installed in the control box of the street light pole and connected to the street light drive circuit to adjust the brightness of the street light through PWM dimming technology. The vehicle networking communication sub-module is installed on the vehicle, powered by the OBD interface or the vehicle power supply to ensure the normal operation of the device, and is adapted to the vehicle's communication system so that it can receive information from the roadside unit and the cloud platform and upload the vehicle information.

[0023] The roadside decision-making unit is implemented through edge computing devices and is deployed in cabinets near each intersection. It communicates and connects with the roadside data collection unit and the street lamp control sub-module, processes local traffic information in real time, and makes decisions. The cloud platform decision-making unit is built on the cloud server in the traffic control center, conducts data interaction with the roadside decision-making units and data processing modules of each road section, and comprehensively analyzes and makes decisions on the traffic data of the entire region by constructing a traffic congestion assessment model.

[0024] The roadside data collection unit in the data collection module and the street lamp data collection unit cooperate. The roadside data collection unit is deployed on both sides of the road and adopts a fusion collection method of millimeter-wave radar and high-definition cameras. The millimeter-wave radar continuously obtains vehicle speed, distance, and angle information, and the high-definition camera obtains vehicle type, position, and driving trajectory information through image recognition algorithms. The data collected by both are fused and processed through the Kalman filtering algorithm to eliminate data errors and redundancy.

[0025] The street lamp data collection unit integrates a light sensor and a current sensor. The light sensor monitors the ambient light intensity in real time, and the current sensor detects the working current of the street lamp. All the collected data is transmitted to the data processing module with low latency through the dedicated data transmission channel of the 5G-MEC network, and this channel uses QoS technology to ensure the priority transmission of data.

[0026] After the data processing module receives the collected data, the vehicle flow direction prediction unit and the lighting optimization calculation unit start to work. The vehicle flow direction prediction unit first preprocesses the input data, and then, by calculating the traffic flow feature matrix , where is the activation function, is the weight matrix obtained through backpropagation algorithm and optimizer. During the training process, a regularization term is introduced to prevent overfitting, and perform convolution operations in the time dimension and space dimension respectively, is the normalized input data, is the data sequence length, is the bias term.

[0027] Then, through calculate the attention weight matrix , where , are the query matrix and key matrix generated through convolution operations respectively, is the key matrix dimension. Through this mechanism, the extraction of key traffic flow features is strengthened to predict the vehicle flow direction in the future time period.

[0028] The lighting optimization calculation unit, based on the predicted result of the traffic flow direction, combines with a preset lighting model, considers the relationship between road curvature, vehicle speed and lighting distance, and through determine the street lamp brightness adjustment coefficient , where 、 are empirical coefficients determined by performing multiple linear regression analysis on the road experimental data of this section. According to the calculation results, a street lamp lighting adjustment plan and a vehicle lighting optimization path are generated.

[0029] The control execution module executes operations according to the instructions of the data processing module. The street lamp control sub-module adopts PWM dimming technology. After receiving the lighting adjustment plan, it adjusts the driving current of the street lamp through a microcontroller, and realizes dynamic brightness adjustment according to the response time and dimming accuracy requirements of the street lamp.

[0030] The vehicle networking communication sub-module, based on UWB positioning technology, adopts a bilateral two-way ranging algorithm to push the lighting optimization path information to the vehicle terminal to ensure real-time information update.

[0031] The decision-making management module realizes three-level decision-making. The roadside decision-making unit is deployed at each intersection, and real-time local traffic information is processed through edge computing devices. When a traffic accident is detected, according to a preset emergency lighting strategy, the street lamps around the accident are controlled to enter the high-brightness constant-on mode. The brightness value in this strategy is determined by analyzing historical accident rescue data and establishing a brightness-rescue efficiency model in combination with the visual characteristics of the human eye under different lighting conditions, and warning information is pushed to surrounding vehicles through the vehicle networking communication sub-module.

[0032] The cloud platform decision-making unit constructs a traffic congestion assessment model, comprehensively analyzes data such as traffic flow density and average vehicle delay time on each section. The congestion index is calculated by weighted summation, and the weights of each data index are determined by the analytic hierarchy process. When the congestion index of a certain section exceeds the threshold, computing power resources are preferentially allocated to perform traffic flow direction prediction and lighting optimization calculation on this section.

[0033] In summary, through the application of this system on this section of the road, the dynamic on-demand adjustment of the street lamp lighting brightness is realized, effectively reducing energy consumption. The vehicle-lamp-cloud three-level decision-making mechanism optimizes the traffic flow, improves the intersection passing efficiency, reduces the vehicle delay time, and the vehicle travels according to the pushed lighting optimization path, reducing the risk of sudden braking caused by insufficient lighting, significantly improving the road safety, and providing an effective solution for the construction of urban intelligent transportation. Embodiment 2

[0034] See Figure 1 and Figure 2, in a certain emerging industrial park, there are multiple freight channels connecting production workshops, warehousing centers and logistics transportation hubs. The roads in the park are distributed in a network pattern. Large freight vehicles come and go frequently at night, and the transportation routes are relatively fixed but there are concentrated traffic periods. Traditional street lights adopt a unified timing switch and fixed brightness mode. There is insufficient lighting during peak freight hours, with visual blind spots, which are prone to vehicle scratches and collisions; while high brightness lighting is still maintained during off-peak hours, resulting in waste of electric power resources. At the same time, due to the lack of traffic information coordination, freight vehicles cannot obtain road conditions and lighting status in advance at intersections, often accelerating or braking suddenly, exacerbating traffic congestion and safety hazards. The street light vehicle-road collaborative control system of the present invention is applied to this industrial park, aiming to improve freight traffic efficiency, ensure safe night driving and save energy.

[0035] Install roadside data collection units on both sides of the main roads and intersections in the park. Install millimeter-wave radars on poles with a height suitable for freight vehicles at an inclined angle to ensure that the long body contours, driving speeds and steering angles of trucks can be accurately captured; high-definition cameras are equipped with wide-angle lenses and installed in open areas, and vehicle information is collected using a special recognition algorithm for heavy trucks. In the street light data collection unit, a light sensor is equipped with a dust-proof and waterproof protective cover and installed on the top of the street light; the current sensor adopts a non-invasive design and is installed around the street light power supply line. All sensors are connected to the 5G-MEC base station through an industrial-grade fiber optic network to ensure stable data transmission.

[0036] Deploy a small data center in the park monitoring center, adopting an edge computing server cluster architecture. The servers are configured with AI acceleration chips optimized for freight scenarios, pre-installed with the TensorFlow framework and traffic data analysis software, and a local database is built to store static data such as the road topology of the park and the traffic rules of freight vehicles. At the same time, it receives and processes the dynamic data transmitted by the acquisition module in real time.

[0037] The street light control sub-module adopts an industrial-grade PLC controller, integrated in the street light distribution box, connected to the street light drive power supply through the 485 communication protocol, supporting wide voltage input and stable operation in high-temperature environments. The vehicle networking communication sub-module is pre-installed in the OBD interface device of the park's freight vehicles, equipped with a high-power antenna to enhance signal reception ability, and supports real-time communication with roadside units and cloud platforms.

[0038] The roadside decision-making unit is deployed in the intelligent control box at each intersection, with an ARM processor and a real-time operating system built in to achieve millisecond-level local decision-making responses. The cloud platform decision-making unit is built based on the park's private cloud, integrates the traffic data of the entire park through a big data analysis model, and dynamically adjusts the control strategies of each section according to the scheduling plan of freight vehicles and real-time road conditions.

[0039] The roadside data acquisition unit works continuously. The millimeter-wave radar obtains information such as the speed, distance, and azimuth angle of freight vehicles at a high sampling frequency. The high-definition camera uses the YOLO-X heavy vehicle detection algorithm to identify characteristic data such as truck types, license plates, and carriage status. After the two sets of data are processed by the weighted fusion algorithm, accurate vehicle operation status data is generated. The light sensor in the street lamp data acquisition unit detects the ambient illuminance every fixed period, and the current sensor monitors the working current of the street lamp in real time. All data is transmitted to the data processing module through the industrial-grade slice channel of the 5G-MEC network.

[0040] After receiving the data, the traffic flow direction prediction unit combines the historical passing data of freight vehicles in the park with the real-time collected information, and uses the improved spatio-temporal convolutional attention network algorithm to calculate the traffic flow feature matrix , where, considering the driving rules of freight vehicles, the window parameters of the convolution operation in the time dimension are optimized specifically to accurately capture the characteristic changes during the concentrated passing period of vehicles.

[0041] By calculating the attention weight matrix , it strengthens the extraction of key features such as the driving trajectory and turning intention of freight vehicles, and predicts the traffic flow distribution on each section in the future period.

[0042] According to the traffic flow direction prediction result, the lighting optimization calculation unit combines the geometric parameters of the park roads and the driving characteristics of freight vehicles, and through determines the street lamp brightness adjustment coefficient , where the empirical coefficients 、 are obtained through ridge regression analysis of the special experimental data of the night driving of freight vehicles in the park, ensuring that the street lamp brightness adjustment can not only meet the lighting needs of freight vehicles but also achieve the energy-saving goal, and then generating the brightness adjustment plan for each section of the street lamp and the vehicle lighting optimization path.

[0043] After receiving the lighting adjustment plan, the street lamp control sub-module adjusts the duty cycle of the street lamp drive power supply through the PWM dimming technology according to the preset dimming curve to achieve a smooth gradual change in the street lamp brightness. The vehicle networking communication sub-module pushes the lighting optimization path and road condition information to the in-vehicle terminal of the freight vehicle in real time through UWB positioning and V2X communication technologies, and the in-vehicle terminal guides the driver to drive through voice prompts and dashboard displays.

[0044] The roadside decision-making unit monitors the traffic conditions at intersections in real time. When it detects that the length of the truck queue exceeds the threshold or an emergency such as cargo spillage occurs, the emergency lighting plan is immediately triggered, the brightness of the street lights in the accident area is increased to the maximum, and passing vehicles are alerted through the sound and light alarm device. The cloud platform decision-making unit dynamically adjusts the computing power allocation priority of each road section based on the park's freight scheduling plan and real-time traffic data. For example, during periods of concentrated cargo transportation, priority is given to traffic flow prediction and lighting optimization on sections around logistics hubs. At the same time, staggered driving suggestions are pushed to freight vehicles through the Internet of Vehicles platform to balance traffic flow within the park.

[0045] To sum up, after the application of this system in the industrial park, the energy consumption of street lamps has been reduced. Through dynamic dimming to accurately match the traffic needs of freight vehicles, the average driving speed of freight vehicles has been improved, the number of emergency brakes has been reduced, and the traffic efficiency at intersections has been improved, effectively alleviating traffic congestion in the park. At the same time, with the help of lighting optimization path guidance and real-time road condition warning, the incidence of traffic accidents has been significantly reduced, providing reliable guarantees for efficient logistics transportation and safe production in the industrial park.

[0046] The above description is only a preferred embodiment of the present invention and does not limit the present invention in any form. Although the present invention has been disclosed as a preferred embodiment as above, it is not used to limit the present invention. Any technical personnel in this field can make some changes or modify the technical contents disclosed above into equivalent embodiments without departing from the scope of the technical solution of the present invention. However, any brief modifications, equivalent changes and modifications made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solution of the present invention are still within the scope of the technical solution of the present invention.

Claims

1. A road lamp vehicle-road collaborative control system, characterized in that, The system includes: a data acquisition module, a data processing module, a control execution module, and a decision-making management module; The data acquisition module is used to collect traffic signal phases, vehicle positions and speeds, street lamp working states, and ambient light intensity data; The data processing module includes a traffic flow direction prediction unit and an illumination optimization calculation unit. The traffic flow direction prediction unit uses a spatio-temporal convolutional attention network algorithm. By calculating the traffic flow feature matrix , where is an activation function, is a weight matrix, and respectively represent convolutional operations in the time dimension and the space dimension, is the input data, is the data sequence length, is the bias term. The illumination optimization calculation unit generates a street lamp illumination adjustment plan and a vehicle illumination optimization path according to the traffic flow direction prediction result and in combination with a preset illumination model; The control execution module adjusts the street lamp illumination brightness and pushes an illumination optimization path to the vehicle according to the instruction output by the data processing module; The decision-making management module realizes three-level decision-making of roadside, street lamps, and cloud platform, and dynamically manages the system operation.

2. The vehicle-road collaborative control system for street lamps according to claim 1, wherein The data acquisition module includes a roadside data acquisition unit and a street lamp data acquisition unit. The roadside data acquisition unit adopts a fusion acquisition method of millimeter-wave radar and high-definition camera. The vehicle speed, distance, and angle information obtained by the millimeter-wave radar and the vehicle type, position, and driving trajectory information obtained by the high-definition camera are fused and processed through the Kalman filtering algorithm. The street lamp data acquisition unit integrates a light sensor and a current sensor. The light sensor monitors the ambient light intensity based on the silicon photovoltaic cell principle, and the current sensor detects the street lamp working current through the Hall effect. The two transmit the data to the data processing module through the data transmission channel of the 5G-MEC network.

3. The road lamp vehicle-road collaborative control system according to claim 1, characterized in that, In the spatio-temporal convolutional attention network algorithm, by calculating the attention weight matrix , where and are the query matrix and the key matrix respectively, is the dimension of the key matrix. In the model training stage, with the goal of minimizing the traffic flow prediction error, the parameters in the attention weight matrix are iteratively optimized through the backpropagation algorithm and the stochastic gradient descent method, and a regularization term is introduced to balance the fitting ability and generalization ability of the model by adjusting the regularization coefficient.

4. The vehicle-road collaborative control system for street lights according to claim 1, wherein, The preset lighting model of the lighting optimization calculation unit takes into account the road curvature , vehicle speed and lighting distance relationship, and determines the street lamp brightness adjustment coefficient through , where , are empirical coefficients determined by performing multiple linear regression analysis on road curvature, speed, lighting distance, and driver visual feedback data under different road types and traffic flow conditions.

5. The vehicle-road collaborative control system of a street lamp according to claim 1, wherein, The control execution module includes a street lamp control sub-module and a vehicle networking communication sub-module. The street lamp control sub-module adopts PWM dimming technology. According to the instruction of the data processing module, it adjusts the street lamp drive current by setting different duty cycles. The duty cycle adjustment step is determined according to the street lamp response time and dimming accuracy requirements. The vehicle networking communication sub-module is based on UWB positioning technology and uses the two-way ranging TWR algorithm to achieve high-precision communication between the vehicle and the roadside unit.

6. The vehicle-road collaborative control system for street lamps according to claim 5, wherein The decision-making management module includes a roadside decision-making unit. When a traffic accident is detected, the roadside decision-making unit controls the street lamps around the accident to enter the high-brightness constant-on mode according to the preset emergency lighting strategy, and pushes warning information to the surrounding vehicles through the vehicle networking communication sub-module. In the preset emergency lighting strategy, the brightness value of the high-brightness constant-on mode is determined according to the brightness-rescue efficiency model established based on historical accident rescue data and human eye visual characteristics, and the roadside decision-making unit dynamically adjusts the push range and frequency of the warning information according to the severity and impact range of the accident.

7. The vehicle-road collaborative control system for street lamps according to claim 6, characterized in that, The decision-making management module also includes a cloud platform decision-making unit. The cloud platform decision-making unit is used to construct a traffic congestion assessment model and comprehensively analyze the traffic flow density and vehicle average delay time data of each section. In the traffic congestion assessment model, the congestion index is calculated by weighted summation of each data index. The weights of each data index are calculated through the analytic hierarchy process AHP according to the judgment matrix constructed by the scores of traffic domain experts on the importance of each index. When the congestion index of a certain section exceeds the threshold, computing power resources are preferentially allocated to perform traffic flow direction prediction and lighting optimization calculation for that section.

8. A vehicle-road collaborative control system for street lights according to claim 1, characterized in that, The data processing module further includes a data preprocessing unit. The data preprocessing unit performs normalization processing on the data transmitted by the data acquisition module. For missing data, the multiple imputation method is used for processing. After generating multiple imputation values based on the data distribution characteristics and correlations, comprehensive analysis is carried out. For abnormal data, the density-based local outlier factor (LOF) detection algorithm is used. Whether a data point is an outlier is judged by calculating the local reachability density of the data point, and corrections and eliminations are performed.

9. The vehicle-road collaborative control system for street lamps according to claim 5, wherein The vehicle networking communication sub-module adopts a joint optimization algorithm of channel coding and interference cancellation. When transmitting the lighting optimization path information, according to the coding rate - SNR - BER relationship model constructed based on a large amount of channel simulation data and actual test data, through the look-up table method and the interpolation method, the coding rate is dynamically adjusted according to the channel signal-to-noise ratio (SNR) and the bit error rate (BER) requirements. The interference cancellation parameters are adjusted in real time according to the received signal characteristics by using the minimum mean square error (MMSE) algorithm to adjust the coefficients of the interference cancellation filter.

Images