Road traffic management and control method and system based on multi-source energy coordination and hierarchical decision-making

Abstract

This invention proposes a road traffic management method and system based on multi-source energy coordination and hierarchical decision-making, belonging to the field of intelligent transportation technology. It includes: integrating solar photovoltaic power generation modules, high-entropy energy harvesting devices, and phase-change energy storage units for multi-source energy coordinated power supply; using a lightweight convolutional neural network and long short-term memory network model to perform real-time feature extraction, anomaly identification, and short-term traffic flow prediction on raw traffic data, uploading key feature data to the centralized control center; employing a two-stage robust optimization model to determine the energy storage operation boundary, and combining deep reinforcement learning algorithms to coordinate and optimize traffic signal control, electric vehicle charging scheduling, and roadside unit start-stop strategies. This invention enables intelligent real-time processing of traffic data while ensuring a self-sustaining and sustainable energy supply, thereby providing effective decision-making for road traffic management.

Description

Technical Field

[0001] This invention belongs to the field of intelligent transportation technology, and in particular relates to a road traffic control method and system based on multi-source energy coordination and hierarchical decision-making. Background Technology

[0002] The statements in this section are merely background information related to the present invention and do not necessarily constitute prior art.

[0003] Intelligent transportation systems aim to improve road traffic efficiency, ensure traffic safety, and optimize traffic management through information technology. To achieve this goal, it is necessary to widely deploy data collection, communication, and control equipment along roads and build a central system capable of processing data in real time and making intelligent decisions.

[0004] However, existing technical solutions face the following technical bottlenecks in achieving the above objectives: Energy supply is highly dependent on and unstable. Existing roadside equipment and control centers heavily rely on traditional power grids. In remote areas, newly constructed roads, or disaster scenarios, wiring costs are prohibitive or even impossible, and a power grid outage would paralyze the entire traffic monitoring and control system. Although there have been attempts to use new energy sources such as solar power, their output is severely affected by weather and day / night cycles, exhibiting significant fluctuations and intermittent characteristics, making it difficult to meet the core requirement of continuous and stable 24 / 7 operation for intelligent transportation systems.

[0005] The data processing and decision-making chain suffers from high latency and insufficient intelligence. Existing systems typically transmit massive amounts of raw data (such as video streams) directly back to the central cloud platform for processing, resulting in enormous network bandwidth pressure and severe delays in decision response, making it impossible to support real-time intervention in emergencies such as traffic accidents and congestion. Furthermore, central decision-making is often based on simple rules or single-dimensional data, lacking in-depth insight and collaborative optimization capabilities regarding the coupled relationships between multiple factors such as traffic flow, vehicle status, and energy supply and demand, making it difficult to achieve the leap from passive response to proactive prediction. Summary of the Invention

[0006] To overcome the shortcomings of the prior art, this invention provides a road traffic management method and system based on multi-source energy coordination and hierarchical decision-making. It can achieve intelligent real-time processing of traffic data while ensuring the independent and sustainable supply of energy, thereby providing effective decision-making for road traffic management.

[0007] To achieve the above objectives, one or more embodiments of the present invention provide the following technical solutions: The first aspect of this invention provides a road traffic control method based on multi-source energy coordination and hierarchical decision-making.

[0008] Road traffic management methods based on multi-source energy coordination and hierarchical decision-making include: By integrating solar photovoltaic power generation modules, high-entropy energy harvesting devices, and phase change energy storage units, multi-source energy is used for coordinated power supply. By deploying traffic data collection components on the roadside, vehicle identity information, traffic flow, vehicle speed, lane occupancy rate, abnormal events and environmental data are collected in real time as raw traffic data; Lightweight convolutional neural network and long short-term memory network models are run at roadside edge nodes to perform real-time feature extraction, anomaly identification and short-term traffic flow prediction on the obtained raw traffic data, and only the identified key feature data is uploaded to the central control center. Based on the received key feature data, the central control center uses a two-stage robust optimization model to determine the energy storage operation boundary, and combines deep reinforcement learning algorithms to coordinate the optimization of traffic signal control, electric vehicle charging scheduling, and roadside unit start-stop strategies. The optimization results are displayed on an interactive screen, and early warning information and control instructions are pushed to roadside equipment, vehicles, and management personnel.

[0009] Furthermore, the high-entropy energy harvesting device is deployed below the road surface to collect the mechanical vibration energy caused by vehicle traffic and convert it into electrical energy. Specifically, the high-entropy energy harvesting device is a piezoelectric energy harvesting device, whose output power is dynamically adjusted according to traffic flow and is directly coupled to the phase change energy storage unit, serving as a supplementary power source when solar energy output is insufficient.

[0010] Furthermore, the traffic data acquisition components include: a visible light communication transceiver, an RFID reader / writer, a high-definition camera, a millimeter-wave radar, a geomagnetic sensor, and a weather station.

[0011] Furthermore, the lightweight convolutional neural network is used to identify traffic accidents, illegal parking, and road litter in real time, and the long short-term memory network is used to predict traffic flow within a preset number of minutes in the future.

[0012] Furthermore, a two-stage robust optimization model is adopted to determine the energy storage operation boundary, including: the first stage, with the goal of minimizing the energy storage boundary setting cost, determining the upper and lower limits of the charging and discharging power of each energy storage unit under extreme scenarios; the second stage, within the operation boundary determined in the first stage, with the goal of minimizing the load reduction risk and system operating cost, solving the optimal scheduling strategy under the worst wind and solar power output scenario.

[0013] Furthermore, the deep reinforcement learning algorithm is a soft actor-critic algorithm. The state space of the soft actor-critic algorithm includes traffic flow at each node of the road network, state of charge and temperature of electric vehicle batteries, solar photovoltaic output and electricity price information; the action space of the soft actor-critic algorithm is any one of the charging current of each charging pile or the transmission power of the roadside unit.

[0014] Furthermore, the road traffic management method based on multi-source energy coordination and hierarchical decision-making also includes a direct energy interaction step between vehicles and infrastructure. That is, within the target area, by deploying solar photovoltaic carports and solar receivers on the roof of vehicles, contactless solar charging is achieved during vehicle parking, and identity authentication and billing are completed through visible light communication.

[0015] The second aspect of this invention provides a road traffic control system based on multi-source energy coordination and hierarchical decision-making.

[0016] A road traffic control system based on multi-source energy coordination and hierarchical decision-making includes: The multi-source energy collaborative power supply module includes a solar photovoltaic power generation unit, a high-entropy energy harvesting device, a phase change energy storage unit, and an energy router, which is used to provide stable power for traffic monitoring and communication equipment; The multimodal data acquisition module includes a visible light communication transceiver, an RFID reader / writer, a high-definition camera, a millimeter-wave radar, a geomagnetic sensor, and a weather station, used to collect vehicle identity information, traffic flow, vehicle speed, lane occupancy rate, abnormal events, and environmental data in real time. The edge intelligent preprocessing module, deployed at roadside nodes, has a built-in embedded GPU and lightweight AI model to process raw data in real time and extract key feature data. The central hierarchical decision-making and optimization module is deployed on the central control center server. It has a built-in two-stage robust optimization model and a deep reinforcement learning decision engine to generate traffic control and energy dispatch strategies. The multimodal interaction and execution module includes an interactive large screen, a voice recognition unit, a gesture sensor, and an AR projection device, which are used to display system status, early warning information, and issue control commands.

[0017] The above one or more technical solutions have the following beneficial effects: This invention integrates solar photovoltaic power generation modules, high-entropy energy harvesting devices, and phase change energy storage units to achieve multi-source energy collaborative power supply. It innovatively combines unstable solar energy with high-entropy electrical energy generated in real time according to traffic flow, and utilizes the stabilizing and storing capabilities of phase change energy storage. This method overcomes the shortcomings of existing technologies that are heavily dependent on energy supply, and its advantages are particularly evident in remote road sections where the power grid is difficult to cover, newly built roads, or disaster emergency scenarios. Simultaneously, the multi-source complementarity and energy storage buffering mechanism can effectively overcome the large fluctuations in output from a single renewable energy source, thereby ensuring the continuous and stable operation of traffic monitoring, communication, and processing equipment 24 / 7 under various weather conditions, significantly improving the overall system's deployment flexibility and energy resilience.

[0018] This invention achieves localized real-time processing of massive amounts of raw data, such as video, by deploying an edge intelligent preprocessing module embedded with a lightweight AI model on the roadside. This allows for vehicle recognition, abnormal event detection, and short-term traffic flow prediction, while only uploading a small proportion of key feature data to the central control center. This significantly reduces network bandwidth consumption and shortens the latency from event occurrence to central response to the second level, meeting the real-time intervention requirements for emergencies such as traffic accidents and congestion. Building on this, the central control center's hierarchical decision-making and optimization module comprehensively utilizes algorithms such as two-stage robust optimization and deep reinforcement learning. It deeply integrates multi-dimensional information such as real-time traffic flow, vehicle battery status, multi-source energy output, and road network topology to achieve coordinated optimization and dynamic adjustment of traffic signal control, electric vehicle charging scheduling, and roadside unit start-stop strategies. This drives traffic management from passive response based on simple rules to proactive prediction and precise control with multiple objectives and forward-looking capabilities, thereby significantly improving road network efficiency and safety.

[0019] Advantages of additional aspects of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0020] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an improper limitation of the invention.

[0021] Figure 1 This is a flowchart of the road traffic control method based on multi-source energy coordination and hierarchical decision-making in Embodiment 1 of the present invention. Detailed Implementation

[0022] It should be noted that the following detailed descriptions are exemplary and intended to provide further illustration of the invention. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0023] It should be noted that the terminology used herein is for the purpose of describing particular implementations only and is not intended to limit the exemplary implementations of the present invention.

[0024] Where there is no conflict, the embodiments and features in the embodiments of the present invention can be combined with each other.

[0025] Example 1 This embodiment discloses a road traffic control method based on multi-source energy coordination and hierarchical decision-making.

[0026] like Figure 1 As shown, the road traffic control method based on multi-source energy coordination and hierarchical decision-making includes: Step S1: Multi-source energy is used for coordinated power supply by integrating solar photovoltaic power generation modules, high-entropy energy harvesting devices, and phase change energy storage units; Step S2: By deploying traffic data collection components on the roadside, collect vehicle identity information, traffic flow, vehicle speed, lane occupancy rate, abnormal events and environmental data in real time as raw traffic data; Step S3: Run a lightweight convolutional neural network and long short-term memory network model at the roadside edge node to perform real-time feature extraction, anomaly identification and short-term traffic flow prediction on the obtained raw traffic data, and upload only the identified key feature data to the central control center. Step S4: Based on the received key feature data, the central control center uses a two-stage robust optimization model to determine the energy storage operation boundary, and combines deep reinforcement learning algorithms to collaboratively optimize traffic signal control, electric vehicle charging scheduling, and roadside unit start-stop strategies; the optimization results are displayed on an interactive screen, and early warning information and control instructions are pushed to roadside equipment, vehicles, and management personnel.

[0027] Based on the above process, this invention can achieve intelligent real-time processing of traffic data while ensuring a self-sustaining and sustainable energy supply, thereby providing effective decision-making for road traffic management. To facilitate understanding of the technical solution of this invention, the specific implementation methods of this invention will be further explained and described below.

[0028] In step S1, multi-source energy is used for coordinated power supply by integrating solar photovoltaic power generation modules, high-entropy energy harvesting devices, and phase change energy storage units.

[0029] Multi-source energy is used for coordinated power supply by integrating solar photovoltaic (PV) power generation modules, high-entropy energy harvesting devices, and phase change energy storage units. Specifically, solar PV power generation modules are deployed on roadside or overhead scaffolds to convert solar energy into DC power. The phase change energy storage unit uses paraffin-based phase change materials to store and release heat energy through heat storage and release processes, and provides stable temperature environment regulation and partial power supplementation to the system via thermoelectric conversion. An energy router serves as the core control unit to monitor the output power of solar PV, the state of charge of the phase change energy storage unit, temperature, and system load demand in real time. Its control logic is as follows: priority is given to using the direct output of solar PV power generation; when PV output is excessive, the excess electricity is used to drive a heat pump to store heat in the phase change materials or directly stored in the backup battery; when PV output is insufficient, priority is given to releasing the heat energy in the phase change energy storage, while the high-entropy energy harvesting device is activated as a supplementary power source.

[0030] Furthermore, the high-entropy energy harvesting device is deployed below the road surface to collect the mechanical vibration energy caused by vehicle traffic and convert it into electrical energy. Specifically, the high-entropy energy harvesting device is a piezoelectric energy harvesting device, whose output power is dynamically adjusted according to traffic flow and is directly coupled with a phase change energy storage unit to serve as a supplementary power source when solar energy output is insufficient.

[0031] As an optional embodiment, piezoelectric ceramic sheets or piezoelectric composite material arrays are encapsulated in a waterproof, pressure-resistant substrate and laid beneath the road asphalt layer. When a vehicle drives over them, tire pressure causes the piezoelectric material to deform and generate an electric charge, which is then converted into a stable voltage direct current via a charge collection circuit and a DC-DC converter. The output of this electrical energy is directly connected to the input of the thermoelectric conversion system of the phase change energy storage unit. The photovoltaic output is continuously monitored by an energy router. When it falls below a preset threshold (e.g., 70% of load demand) and the available power reserve of the phase change energy storage unit is insufficient, the energy output collected on the piezoelectric surface is maximized via an instruction to supplement the power shortage.

[0032] In step S2, traffic data collection components deployed on the roadside collect vehicle identity information, traffic flow, vehicle speed, lane occupancy rate, abnormal events, and environmental data in real time as raw traffic data.

[0033] Traffic data acquisition components deployed along roadsides, including visible light transceivers, RFID readers, high-definition cameras, millimeter-wave radar, geomagnetic sensors, and weather stations, are used to collect real-time vehicle identification information, traffic flow, vehicle speed, lane occupancy, abnormal events, and environmental data as raw traffic data. This can be achieved through the following methods: Visible light communication transceiver: Installed on a street light pole, it modulates the LED light source at a specific frequency to send a signal containing information such as the node ID and status, and receives light signals from a vehicle-specific transponder to obtain vehicle identity, basic status information, and security messages.

[0034] RFID readers: Deployed at the start, end, or intersection of key road sections to read RFID electronic tags affixed or embedded in passing vehicles and obtain the vehicle's unique identifier.

[0035] High-definition cameras: capture traffic video streams at a specific frame rate (e.g., 10-30 fps) for subsequent analysis.

[0036] Millimeter-wave radar: Deployed on the roadside, it emits millimeter waves and receives echoes to directly measure the distance, speed, and angle of target vehicles, generating point cloud data.

[0037] Geomagnetic sensor: Buried under the lane, it detects changes in the geomagnetic field caused by vehicles passing by, in order to count and sense the presence of vehicles.

[0038] Weather station: Deployed by the roadside to measure and output data on ambient temperature, humidity, wind speed, wind direction, rainfall, and solar irradiance.

[0039] Data from all the above components can be aggregated in real time to the edge intelligent preprocessing module of this roadside node via wired (e.g., Ethernet) or wireless (e.g., 5G CPE) methods.

[0040] In step S3, a lightweight convolutional neural network and long short-term memory network model are run at the roadside edge nodes to perform real-time feature extraction, anomaly identification, and short-term traffic flow prediction on the obtained raw traffic data. Only the identified key feature data is uploaded to the centralized control center.

[0041] This system employs a lightweight convolutional neural network to identify events such as traffic accidents, illegal parking, and road littering in real time. Specifically, a pre-trained lightweight convolutional neural network model based on the MobileNetV3-Small architecture is loaded onto the embedded GPU in the edge intelligence preprocessing module. This model takes video frames captured by a high-definition camera as input. During inference, the model performs forward propagation on each frame, outputting bounding boxes and class probabilities. By setting a confidence threshold (e.g., 0.7), when the model outputs a confidence level exceeding the threshold for "traffic accidents" (e.g., multi-vehicle collisions, vehicle rollovers), "illegal parking" (e.g., vehicles remaining stationary in no-parking zones for an extended period), or "road littering" (e.g., fallen goods, large obstacles), it is determined that the corresponding abnormal event has occurred. Simultaneously, the frame timestamp, lane, and bounding box location of the event are recorded.

[0042] Traffic flow is predicted within a preset number of minutes in the future using a Long Short-Term Memory (LSTM) network. Specifically, an LSM network model is deployed on the same edge node. The model's input feature sequence consists of historical traffic flow data for each lane of the road segment under the node's jurisdiction over a past period (e.g., 1 hour) (which can be obtained by fusing geomagnetic sensor and radar data), along with auxiliary features such as current and prediction period meteorological data (e.g., weather condition codes) and time period type (e.g., weekday morning rush hour). After training, the model uses the aforementioned historical sequence as input and outputs the predicted traffic flow values ​​for each lane within a future time period (e.g., the next 15 minutes). The prediction period can be set to a fixed time interval as needed, such as predicting the next 15 minutes every 5 minutes.

[0043] The raw traffic data is processed in real-time for feature extraction, anomaly detection, and short-term traffic flow prediction. Only the identified key feature data is uploaded to the central control center. Specifically, the edge intelligent preprocessing module processes the aggregated raw data stream in real time. For video streams, feature information packages for anomalies are extracted and uploaded only when a lightweight CNN identifies them. These packages include: event type, timestamp, geographic location (node ​​ID and lane number), confidence level, and a small number of associated key image fragments (such as compressed images of 5 frames before and after the event). For traffic flow data, the predicted traffic flow values ​​for each lane in the future time period, obtained from LSTM prediction, along with aggregated indicators such as average vehicle speed and lane occupancy rate for the current time period, are uploaded every 5 minutes. For RFID and visible light communication data, aggregated vehicle passage records (such as a list of vehicle IDs and passage times) are uploaded. Large-scale data such as raw video streams and radar point clouds are discarded or repeatedly overwritten after processing at the edge nodes and are not uploaded to the central control center.

[0044] In step S4, the central control center uses a two-stage robust optimization model to determine the energy storage operation boundary based on the received key feature data, and combines deep reinforcement learning algorithms to coordinate the optimization of traffic signal control, electric vehicle charging scheduling, and roadside unit start-stop strategies. The optimization results are displayed on an interactive screen, and early warning information and control instructions are pushed to roadside equipment, vehicles, and management personnel.

[0045] After receiving key feature data, the control center uses a two-stage robust optimization model to determine the energy storage operating boundary. This includes: First, aiming to minimize the energy storage boundary setup cost, determining the upper and lower limits of the charging and discharging power of each energy storage unit under extreme scenarios; second, within the operating boundary determined in the first stage, aiming to minimize load shedding risk and system operating costs, solving for the optimal scheduling strategy under the worst-case wind and solar power output scenario. Specifically, this can be achieved through the following methods: Phase 1: The optimization engine in the control center treats each phase change energy storage unit at each roadside node as an object to be optimized. With a planning period of 24 hours, the inputs include the predicted load curves for each node, the predicted photovoltaic output curves (considering weather forecasts), electricity price information, and the physical parameters of the energy storage system (capacity, efficiency, cost coefficient). An optimization problem is constructed, with the decision variables being the upper and lower limits of the planned charge and discharge power for each energy storage unit in each time period. The objective function is to minimize the "setup cost" (which may be related to equipment wear or reserved capacity costs) incurred by all energy storage units within the planning period due to setting these operating boundaries, while implicitly satisfying load demands. By solving this optimization problem, a robust set of energy storage operating power boundary plans can be obtained.

[0046] Phase Two: Within the operational boundary framework defined in Phase One, the uncertainties in wind and solar power output are considered (described as a polyhedral set of uncertainties). The optimization engine aims to minimize the total system operating cost under the worst-case wind and solar scenario; this total operating cost includes electricity purchase cost, generation cost, and load shedding penalty cost. Decision variables include the actual charging and discharging power of each node in each uncertain scenario and time period, the power purchased from the grid, and the possible load shedding. By solving this min-max optimization problem, a day-ahead dispatch strategy capable of withstanding the worst renewable energy fluctuations can be obtained. This strategy ensures that energy storage charging and discharging are performed within the boundaries set in Phase One, regardless of any permissible uncertainties.

[0047] Based on this, a deep reinforcement learning algorithm is used to collaboratively optimize traffic signal control, electric vehicle charging scheduling, and roadside unit start-stop strategies. The deep reinforcement learning algorithm is a soft actor-critic algorithm. The state space of this algorithm includes traffic flow at each node of the road network, the state of charge and temperature of electric vehicle batteries, solar photovoltaic output, and electricity price information. The action space of this algorithm is any one of the charging current of each charging pile or the transmission power of the roadside unit.

[0048] Specifically, the control center deploys a deep reinforcement learning (DRL) agent with a soft actor-critic algorithm. This agent interacts with a simulated environment (or directly with the actual system).

[0049] state At each decision time t, the state includes: 1) real-time traffic flow and predicted traffic flow uploaded by each node of the road network; 2) the state of charge and temperature of the electric vehicle batteries connected to the charging pile (obtained through the vehicle network); 3) the actual output of the current solar photovoltaic system; and 4) the time-of-use electricity price information for the current period.

[0050] action The actions output by the intelligent agent include: 1) suggestions for adjusting the traffic light cycle at each intersection (such as increasing or decreasing the green light ratio); 2) charging current instructions to each electric vehicle from each roadside charging pile; 3) start / stop instructions (open or close) for some non-critical roadside units (such as some information display screens).

[0051] award The environment rewards actions based on their effects. The reward function is designed as a comprehensive indicator, including: improvements in average road network speed, a trade-off between charging efficiency and safety (encouraging fast charging but penalizing battery overheating), and negative costs such as energy-saving effects.

[0052] Through extensive offline training and online fine-tuning, the intelligent agent learns the optimal policy and is able to output the optimal cooperative control action given a state.

[0053] Finally, the optimization results are displayed on an interactive large screen, and warning information and control instructions are pushed to roadside equipment, vehicles, and management personnel. Specifically, the interactive large screen in the control center receives optimization results and instructions from the central hierarchical decision-making and optimization module. The screen uses a GIS map layer to display the real-time traffic status (smooth / congested / accident), traffic light control mode, charging pile usage status, and energy storage unit power levels of the entire road network using different colors and icons. When a Level 1 warning occurs (such as an accident), a red warning icon flashes at the corresponding location on the screen, and a video feed of the incident location automatically pops up. The system pushes traffic management suggestions, charging scheduling instructions, and variable message sign content to the onboard terminals of relevant vehicles or roadside execution units via vehicle-to-infrastructure communication units (such as C-V2X) or mobile networks. Simultaneously, through an integrated SMS and application push gateway, warning information containing an event summary and location is sent to the management personnel responsible for the area.

[0054] Furthermore, the road traffic management method based on multi-source energy coordination and hierarchical decision-making also includes a direct energy interaction step between vehicles and infrastructure. Specifically, within the target area, contactless solar charging during vehicle parking is achieved by deploying solar photovoltaic carports and rooftop solar receivers on vehicles, with identity authentication and billing completed via visible light communication. Specifically, carports constructed of solar photovoltaic panels are built in target areas such as toll stations, service areas, and parking lots. A visible light communication transmitter module is integrated into the carport's supporting structure. Vehicles with receiving capabilities have a small-area photovoltaic receiver and a visible light communication receiver installed on their roofs. When a vehicle parks, the vehicle receiver and the carport transmitter establish communication via a visible light link to complete vehicle identity authentication. After successful authentication, the carport's energy management system controls DC power distribution, transmitting the DC power generated by the photovoltaic system to the receiver on the vehicle's roof via contactless methods such as electromagnetic induction or laser to charge the vehicle's auxiliary or main battery. The billing system automatically calculates the cost based on the charging amount and the current rate, and sends the bill to the vehicle owner's account via visible light communication or a mobile network.

[0055] Example 2 This embodiment discloses a road traffic control system based on multi-source energy coordination and hierarchical decision-making.

[0056] A road traffic control system based on multi-source energy coordination and hierarchical decision-making includes: The multi-source energy collaborative power supply module includes a solar photovoltaic power generation unit, a high-entropy energy harvesting device, a phase change energy storage unit, and an energy router, which is used to provide stable power for traffic monitoring and communication equipment; The multimodal data acquisition module includes a visible light communication transceiver, an RFID reader / writer, a high-definition camera, a millimeter-wave radar, a geomagnetic sensor, and a weather station, used to collect vehicle identity information, traffic flow, vehicle speed, lane occupancy rate, abnormal events, and environmental data in real time. The edge intelligent preprocessing module, deployed at roadside nodes, has a built-in embedded GPU and lightweight AI model to process raw data in real time and extract key feature data. The central hierarchical decision-making and optimization module is deployed on the central control center server. It has a built-in two-stage robust optimization model and a deep reinforcement learning decision engine to generate traffic control and energy dispatch strategies. The multimodal interaction and execution module includes an interactive large screen, a voice recognition unit, a gesture sensor, and an AR projection device, which are used to display system status, early warning information, and issue control commands.

[0057] Furthermore, the high-entropy energy harvesting device is a piezoelectric ceramic array laid under the road surface, and its output is directly connected to the input of the phase change energy storage unit through a DC-DC converter to realize the real-time collection and storage of vibration energy.

[0058] Furthermore, the visible light communication transceiver employs an adjustable optical attenuator array to maintain stable vehicle-to-road and vehicle-to-vehicle communication links under strong ambient light interference.

[0059] Furthermore, the central hierarchical decision-making and optimization module is also connected to a roadside unit planning submodule, which is used to optimize the deployment location, type, daily start-stop plan, and whether to equip the roadside units with solar panels based on a mixed integer quadratic constraint programming model, given a budget and coverage requirements.

[0060] Furthermore, the road traffic control system based on multi-source energy coordination and hierarchical decision-making also includes: an electric vehicle charging management submodule, which is deployed on the roadside charging pile controller to receive charging current commands issued by the central hierarchical decision-making and optimization module, and to estimate the internal state of the battery in real time based on an extended Kalman filter to achieve safe and fast charging under multiple physical constraints.

[0061] Furthermore, the multimodal interaction and execution module supports a tiered early warning mechanism, specifically including: Level 1 early warning: for traffic accidents or major malfunctions, triggering an audible and visual alarm and automatically sending location and on-site video to traffic police and rescue departments; Level 2 early warning: for traffic congestion or severe weather, issuing detour suggestions through information boards and vehicle terminals; Level 3 early warning: for power supply fluctuations or equipment anomalies, highlighting them on the central control screen and recording logs.

[0062] Furthermore, the system also integrates a solar-powered autonomous driving toll submodule, deployed at the toll station, including: a solar-powered RFID reader and servo gate; an IoT communication unit for sending a confirmation email containing the amount, location, and time to the vehicle owner after the toll is deducted; and a license plate recognition camera as a redundant verification method for the RFID system.

[0063] Those skilled in the art will understand that the modules or steps of the present invention described above can be implemented using general-purpose computer devices. Optionally, they can be implemented using computer-executable program code, thereby allowing them to be stored in a storage device for execution by a computer device, or they can be fabricated as separate integrated circuit modules, or multiple modules or steps can be fabricated as a single integrated circuit module. The present invention is not limited to any particular combination of hardware and software.

[0064] While the specific embodiments of the present invention have been described above in conjunction with the accompanying drawings, this is not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art without creative effort based on the technical solutions of the present invention are still within the scope of protection of the present invention.

Claims

1. A road traffic control method based on multi-source energy coordination and hierarchical decision-making, characterized in that, include: By integrating solar photovoltaic power generation modules, high-entropy energy harvesting devices, and phase change energy storage units, multi-source energy is used for coordinated power supply. By deploying traffic data collection components on the roadside, vehicle identity information, traffic flow, vehicle speed, lane occupancy rate, abnormal events and environmental data are collected in real time as raw traffic data; Lightweight convolutional neural network and long short-term memory network models are run at roadside edge nodes to perform real-time feature extraction, anomaly identification and short-term traffic flow prediction on the obtained raw traffic data, and only the identified key feature data is uploaded to the central control center. Based on the received key feature data, the central control center uses a two-stage robust optimization model to determine the energy storage operation boundary, and combines deep reinforcement learning algorithms to coordinate the optimization of traffic signal control, electric vehicle charging scheduling, and roadside unit start-stop strategies. The optimization results are displayed on an interactive screen, and early warning information and control instructions are pushed to roadside equipment, vehicles, and management personnel.

2. The road traffic control method based on multi-source energy coordination and hierarchical decision-making as described in claim 1, characterized in that, The high-entropy energy harvesting device is deployed below the road surface to collect the mechanical vibration energy caused by vehicle traffic and convert it into electrical energy. Specifically, the high-entropy energy harvesting device is a piezoelectric energy harvesting device, whose output power is dynamically adjusted according to traffic flow and is directly coupled to a phase change energy storage unit to serve as a supplementary power source when solar energy output is insufficient.

3. The road traffic control method based on multi-source energy coordination and hierarchical decision-making as described in claim 1, characterized in that, The traffic data acquisition components include: a visible light communication transceiver, an RFID reader / writer, a high-definition camera, a millimeter-wave radar, a geomagnetic sensor, and a weather station.

4. The road traffic control method based on multi-source energy coordination and hierarchical decision-making as described in claim 1, characterized in that, The lightweight convolutional neural network is used to identify traffic accidents, illegal parking, and road debris in real time, while the long short-term memory network is used to predict traffic flow within a preset number of minutes in the future.

5. The road traffic control method based on multi-source energy coordination and hierarchical decision-making as described in claim 1, characterized in that, A two-stage robust optimization model is used to determine the energy storage operation boundary, including: the first stage, with the goal of minimizing the energy storage boundary setting cost, determining the upper and lower limits of the charging and discharging power of each energy storage unit under extreme scenarios; the second stage, within the operation boundary determined in the first stage, with the goal of minimizing the load reduction risk and system operating cost, solving the optimal scheduling strategy under the worst wind and solar power output scenario.

6. The road traffic control method based on multi-source energy coordination and hierarchical decision-making as described in claim 1, characterized in that, The deep reinforcement learning algorithm is a soft actor-critic algorithm. The state space of the soft actor-critic algorithm includes traffic flow at each node of the road network, state of charge and temperature of electric vehicle batteries, solar photovoltaic output and electricity price information; the action space of the soft actor-critic algorithm is any one of the charging current of each charging pile or the transmission power of the roadside unit.

7. The road traffic control method based on multi-source energy coordination and hierarchical decision-making as described in claim 1, characterized in that, It also includes a direct energy interaction step between vehicles and infrastructure, namely: in the target area, by deploying solar photovoltaic carports and solar receivers on the roof of vehicles, contactless solar charging is achieved during vehicle parking, and identity authentication and billing are completed through visible light communication.

8. A road traffic control system based on multi-source energy coordination and hierarchical decision-making, characterized in that, include: The multi-source energy collaborative power supply module includes a solar photovoltaic power generation unit, a high-entropy energy harvesting device, a phase change energy storage unit, and an energy router, which is used to provide stable power for traffic monitoring and communication equipment; The multimodal data acquisition module includes a visible light communication transceiver, an RFID reader / writer, a high-definition camera, a millimeter-wave radar, a geomagnetic sensor, and a weather station, used to collect vehicle identity information, traffic flow, vehicle speed, lane occupancy rate, abnormal events, and environmental data in real time. The edge intelligent preprocessing module, deployed at roadside nodes, has a built-in embedded GPU and lightweight AI model to process raw data in real time and extract key feature data. The central hierarchical decision-making and optimization module is deployed on the central control center server. It has a built-in two-stage robust optimization model and a deep reinforcement learning decision engine to generate traffic control and energy dispatch strategies. The multimodal interaction and execution module includes an interactive large screen, a voice recognition unit, a gesture sensor, and an AR projection device, which are used to display system status, early warning information, and issue control commands.

9. The road traffic control system based on multi-source energy coordination and hierarchical decision-making as described in claim 8, characterized in that, The high-entropy energy harvesting device is a piezoelectric ceramic array laid under the road surface. Its output is directly connected to the input of the phase change energy storage unit through a DC-DC converter, realizing the real-time collection and storage of vibration energy.

10. The road traffic control system based on multi-source energy coordination and hierarchical decision-making as described in claim 8, characterized in that, The visible light communication transceiver employs an adjustable optical attenuator array to maintain stable vehicle-to-road and vehicle-to-vehicle communication links under strong ambient light interference.

Images