Space-ground integrated dynamic traffic flow analysis and road topology reasoning method and device

Abstract

The invention discloses a space-ground integrated dynamic traffic flow analysis and road topology reasoning method and device. A cloud server generates and issues real-time traffic information through multi-level processing such as map matching, traffic flow analysis, prediction and early warning and road topology analysis and updating; specifically, firstly, the high-precision positioning and communication capability of a low-orbit satellite is utilized, data fusion of an inertial measurement unit and a wheel speed sensor is combined, and a multi-source fusion algorithm is constructed to improve the positioning precision; secondly, traffic flow dynamic analysis and prediction are carried out based on a Transform interaction analysis model, and real-time updating of a road topological structure is realized at a cloud end by using a clustering algorithm and a graph theory model; finally, the traffic flow and road topology information is sent to related vehicles and a traffic management system, and real-time monitoring, early warning and optimization of the traffic state are achieved. According to the invention, the accuracy and real-time performance of traffic information acquisition, processing and issuing can be effectively improved, and the intelligent level of traffic management and decision making is enhanced.

Description

Technical Field

[0001] The present invention relates to the fields of intelligent transportation systems and autonomous driving, and particularly to a device and method for traffic flow analysis and road topology inference and update based on low Earth orbit (LEO) satellite communication and high-precision positioning technology. Background Art

[0002] With the rapid development of intelligent transportation systems (ITS) and autonomous driving technologies, cooperation between vehicles and real-time and accurate perception of traffic information have become key technical points for improving traffic efficiency and safety. Traditional traffic information collection means, such as fixed traffic monitoring devices, information collection vehicles, or ground communication technologies, have problems of limited coverage and slow update speed, and it is difficult to meet the requirements of future highly automated traffic scenarios. Low Earth orbit satellites, with their characteristics of low latency, high bandwidth, and high-precision space-based positioning, have become an important technical foundation for supporting future intelligent transportation systems. However, current research and applications of high-precision traffic flow collection and road topology information inference strategies for cooperation between low Earth orbit satellites and vehicles are still not perfect. Summary of the Invention

[0003] Aiming at the deficiencies in the prior art, the present invention provides a method and device for dynamic traffic flow analysis and road topology inference based on low Earth orbit satellite real-time communication and high-precision positioning, which realizes accurate collection, inference, publication, and update of traffic information; combines intelligent vehicle information collection devices, low Earth orbit satellite communication, cloud data processing and distribution systems to form a highly collaborative intelligent transportation system. This method can infer, publish, and update traffic information in real time, and realize functions such as traffic flow monitoring, road topology analysis, and real-time warning.

[0004] The present invention achieves the above technical objectives through the following technical means.

[0005] An integrated space-ground dynamic traffic flow analysis and road topology inference device, comprising:

[0006] A ground monitoring station for monitoring on-orbit satellite data;

[0007] An intelligent vehicle on-board module, which obtains the real-time heading, speed, and position of the vehicle based on low Earth orbit satellite data and ground monitoring station data;

[0008] A cloud server, which receives the data transmitted by the intelligent vehicle on-board module, performs multi-level calculations and analyses, and realizes two groups of task flows on the basis of map matching, including traffic flow distribution, prediction and warning, and road topology structure analysis and update;

[0009] A ground information interaction module, which is used to receive the road topology and traffic flow information processed by the cloud server and distribute it to target vehicles or traffic management systems to realize road state and traffic information sharing.

[0010] In the above technical solution, the in-vehicle module of the intelligent vehicle includes, but is not limited to, an inertial measurement unit, a wheel speed sensor, a low-earth orbit satellite communication module, and a data processing unit.

[0011] A method for integrated space-ground dynamic traffic flow analysis and road topology inference:

[0012] Based on low-earth orbit satellite data and ground monitoring station data, the in-vehicle module of the intelligent vehicle obtains real-time vehicle motion state data and uploads it to the cloud server;

[0013] The cloud server uses the motion state data of intelligent vehicles passing through key traffic areas to perform a preliminary match of map data using a navigation-level map, and further performs a high-precision docking of vehicle trajectories with the road map to achieve an accurate match of each vehicle's position on the map;

[0014] Traffic flow distribution, prediction, and early warning: The cloud server collects all motion state data within a specified time window, calculates the average vehicle speed, traffic flow density, and vehicle spacing distribution, and uses a Transformer-based heterogeneous element interaction model for driving scenarios to analyze the dynamic relationships between vehicles and road map elements to obtain traffic flow states, congestion, and dangerous situations;

[0015] The long-term accumulated vehicle motion state data in the detection area is used to analyze and update the lane centerlines and lane topology relationship maps of the road, realizing the analysis and update of the road topology structure;

[0016] The traffic flow information and road topology are distributed to target vehicles or traffic management systems.

[0017] Furthermore, the data structure collected by the cloud server includes the collected vehicle motion state information and prior map information : , ; where is the time-series state vector of the th vehicle in the scenario, is the vectorized representation of the th map element in the scenario, and respectively represent the vehicles and road map elements included in the currently detected driving scenario; and Input into the Transformer-based heterogeneous scenario element interaction model, analyze the comprehensive spatio-temporal interaction relationship of scenario elements, and obtain scenario dynamic interaction feature information; use the scenario dynamic interaction feature information to analyze the traffic flow state, congestion, and dangerous situations, further analyze the traffic flow density and vehicle speed in each traffic area, and combine historical data and current interaction feature information to predict and identify congestion or collision risk areas and their risk probabilities and output them.

[0018] Furthermore, combine the lane centerline and the turning and lane-changing behaviors in the vehicle trajectory, use the graph theory model to construct a lane topology relationship graph, and define: nodes represent intersections or key positions, edges represent lane connection relationships, edge weights are passing constraints, and at the same time mark special topological structures.

[0019] Even further, the update of the road topological structure is specifically as follows: by comparing the existing high-precision road map information with the lane topology relationship graph, detect newly added, closed, or adjusted lanes, mark the changed positions, and perform incremental updates on the high-precision map; at the same time, based on the detection results, update the navigation map and send it to relevant vehicles.

[0020] Furthermore, when uploading to the cloud server, traffic anomaly information adopts an immediate transmission strategy and is given priority to be uploaded, and regular vehicle status information is transmitted at a fixed frequency.

[0021] Furthermore, when uploading to the cloud server, the time slot allocation for data transmission is dynamically adjusted according to the number of vehicles and the communication bandwidth.

[0022] Furthermore, the traffic flow analysis has the highest priority and performs real-time corresponding traffic flow prediction and early warning; the road topology analysis is used as an auxiliary and is responsible for long-term update and improvement of information sharing.

[0023] Furthermore, the distribution rule is: according to the vehicle position and driving direction, determine the target vehicles that need to receive data, and preferentially distribute the information to the target vehicles closer to the event point.

[0024] The beneficial effects of the present invention are:

[0025] (1) By combining low-earth orbit satellite high-precision positioning and vehicle status information collection, the present invention designs a multi-source data fusion strategy, which can significantly improve the accuracy of vehicle status data collection and the reliability of data transmission, and ensure the stability and adaptability of the system decision-making and control strategy in a complex traffic environment.

[0026] (2) The present invention constructs a real-time traffic flow analysis and prediction model, uses short-term trajectory data to monitor the dynamic traffic flow state, and combines neural networks to predict the future traffic flow trend, realizing real-time detection and early warning of traffic congestion and dangerous situations, and effectively improving the initiative and intelligent level of traffic management.

[0027] (3) Based on long-term vehicle trajectory data, the present invention dynamically generates a road topological structure by using clustering and graph theory methods, and combines with the existing navigation map for real-time update, which can quickly identify road changes (such as new lanes or intersection adjustments), ensure the accuracy and real-time nature of traffic information, and improve the path planning efficiency of the vehicle autonomous driving system.

[0028] (4) Through a multi-task collaborative processing architecture, the present invention combines traffic flow analysis, traffic flow prediction and road topology update, optimizes the data utilization efficiency, enhances the global perception ability of the system, and provides an efficient support solution for intelligent transportation systems and autonomous driving technologies.

[0029] (5) The method of the present invention can be seamlessly integrated with existing traffic management platforms and autonomous driving vehicle control systems, providing a technical guarantee for the comprehensive upgrade of intelligent transportation infrastructure and the wide application of autonomous driving. Brief Description of the Drawings

[0030] Figure 1 It is a schematic diagram of the system architecture of the disclosed embodiment of the present invention.

[0031] Figure 2 It is a flowchart of traffic flow analysis and road topology reasoning based on low-earth orbit satellite communication and high-precision positioning in the disclosed embodiment of the present invention.

[0032] Figure 3 It is a flowchart of the data processing method in the disclosed embodiment of the present invention. Detailed Embodiment

[0033] The present invention will be further described below in conjunction with the drawings and specific embodiments, but the protection scope of the present invention is not limited thereto.

[0034] First, the specific implementation of the system architecture. The system architecture of the present invention consists of an intelligent vehicle on-board module, a cloud server, a ground monitoring station, and a ground information interaction module. Each module works collaboratively to achieve high integration and efficient operation. The specific implementation is as follows:

[0035] The in-vehicle module of an intelligent vehicle includes, but is not limited to, an inertial measurement unit (IMU), wheel speed sensors, a low-earth orbit satellite communication module, a data processing unit, etc. Taking a sedan as an example, the inertial measurement unit is installed at the center of the vehicle chassis and fixed by a rigid bracket to ensure that it can accurately sense the changes in the vehicle's acceleration and angular velocity. The wheel speed sensors are installed near the hubs of the wheels and concentric with the rotation axes of the wheels to accurately measure the rotational speeds of the wheels. After the installation of the inertial measurement unit and the wheel speed sensors, calibration operations are carried out. The IMU undergoes static calibration. By collecting IMU data in multiple postures, error parameters such as the zero bias and scale factor of the IMU are calculated and compensated. For the wheel speed sensors, dynamic calibration is performed using a standard speed source. By comparing the actual speed with the speed measured by the wheel speed sensors, the parameters of the wheel speed sensors are adjusted to control the measurement error within a very small range. The low-earth orbit satellite communication module receives communication signals from multiple satellites and data from ground monitoring stations, corrects errors, and realizes the solution of positioning and navigation data. The data collected by the inertial measurement unit, wheel speed sensors, and low-earth orbit satellite communication module are processed by the data processing unit through Kalman filtering or nonlinear least squares optimization to remove noise interference and output real-time information such as the vehicle's heading, speed, and position.

[0036] The cloud server can adopt a high-performance server cluster architecture to collect the positioning and navigation data of intelligent vehicles to meet the concurrent processing requirements of a large amount of vehicle data. For example, when processing traffic flow analysis tasks, the computing tasks are distributed to each node for parallel execution through the cluster management system, greatly improving the data processing speed and ensuring the efficient operation of the entire system, providing a solid hardware foundation for subsequent tasks such as map matching, traffic flow analysis, and road topology generation.

[0037] The ground monitoring station is used to monitor the data of on-orbit satellites, conduct two-way communication with low-earth orbit satellites, provide functions of data forwarding, signal reception and transmission, and ensure the stability and accuracy of the signals between the satellites and user equipment.

[0038] The ground information interaction module, such as using a satellite communication receiver, 4G / 5G network information receiver, etc., is used to receive the road topology and traffic flow information processed by the cloud server to realize the sharing of road conditions and traffic information.

[0039] In the second aspect, the specific implementation of the method flow is used to illustrate the data processing method and link based on low-earth orbit satellite communication and high-precision positioning, which specifically includes:

[0040] S1. Data acquisition. The low-earth orbit satellite communication module is used to receive multiple groups of satellite signals, perform real-time position, speed, and time synchronization calculations for the vehicle. At the same time, an improved method based on the differential positioning algorithm of the ground monitoring station is adopted to improve the positioning accuracy. Specifically, the low-earth orbit satellite communication module of the vehicle continuously receives multiple groups of low-earth orbit satellite data, and calculates the pseudo-range between the vehicle and the satellite by measuring the propagation time of the satellite signal. At the same time, the ground monitoring station data is used to achieve precise positioning. The ground monitoring stations are distributed around the target area and have accurately known position coordinates. After receiving the satellite signal, the ground monitoring station calculates the measurement error of the satellite signal, including the errors caused by factors such as atmospheric delay and satellite orbit error, and transmits this error information to the vehicle end through the network.

[0041] After the low-earth orbit satellite communication module receives the satellite signal, based on the high-precision positioning of the low-earth orbit satellite, combined with in-vehicle state estimation sensors such as inertial measurement units (IMUs) and wheel speed sensors, high-precision acquisition of the vehicle's motion state is achieved. The inertial measurement unit (IMU) measures the acceleration and angular velocity of the vehicle, and the speed and position changes of the vehicle are obtained through integral operations. During the positioning calculation process, the Kalman filter algorithm is used to fuse the satellite positioning data and inertial navigation data. The state vector of the Kalman filter includes the position, speed, acceleration of the vehicle, and the error state of the inertial measurement unit, etc. First, the low-earth orbit satellite provides the global positioning information of the vehicle, accurately calculating the position, speed, and heading of the vehicle; then, the IMU provides acceleration and angular velocity data for estimating the dynamic changes of the vehicle. Especially in an environment where satellite signals are unstable or blocked, the IMU makes up for the positioning gap through inertial measurement. The wheel speed sensor measures the tire rotation speed in real time, calculates the driving speed of the vehicle, and further calibrates the position information of the vehicle. By fusing the data of these sensors, the errors of a single sensor can be effectively eliminated, improving the positioning accuracy. Especially in complex terrains or signal-blocked situations, continuous and accurate acquisition of the vehicle state is ensured.

[0042] S2. Data upload. Through the low-earth orbit satellite communication module, the processed vehicle data is quickly transmitted to the cloud server, specifically including data packaging and data compression. According to the predetermined communication protocol, the processed vehicle data is compressed into small data packets to reduce the transmission delay; the data packets contain the vehicle's unique identifier (such as license plate number or device ID) so that the cloud server can distinguish the data sources.

[0043] Build a priority strategy during data transmission to sort the data by priority. Information such as traffic anomalies (sudden braking, collisions) is uploaded first, while regular status information (position, speed, heading) is transmitted at a predefined frequency. For the detection of traffic anomaly information, a judgment method based on the acceleration mutation threshold is adopted: in the vehicle's Inertial Measurement Unit (IMU), the acceleration change of the vehicle is monitored in real time; when the absolute value of the acceleration exceeds a preset threshold (such as 8 m / s²) within a short period of time (such as 2 seconds), the data from the wheel speed sensor is combined to determine whether the vehicle is in a sudden braking or collision state.

[0044] In terms of data transmission frequency and time slot allocation, for traffic anomaly information, an immediate transmission strategy is adopted. Once an anomaly is detected, the data is immediately sent to the cloud server through the low-earth orbit satellite communication module, giving priority to occupying communication resources to ensure the timeliness of information. For regular status information, it is transmitted at a predefined frequency, for example, data is transmitted every 10 seconds. In terms of time slot allocation, it is dynamically adjusted according to the number of vehicles and communication bandwidth: when there are fewer vehicles on the road and communication resources are relatively abundant, the transmission time slots for regular status information are appropriately increased to improve the real-time performance of the data; when there are more vehicles and communication is busy, the transmission time slots for regular status information are reasonably reduced to ensure the priority transmission of emergency information, avoid network congestion, and ensure the communication stability of the entire system and the high efficiency of data transmission.

[0045] During the transmission process, by combining the TCP / IP protocol and the link redundancy mechanism of low-earth orbit satellites, the integrity and reliability of the data are ensured; a retransmission mechanism is set for the packets that may be lost. When the sender sends a packet, it will wait for the acknowledgment information (ACK) from the receiver. If the ACK is not received within a certain time, the packet will be automatically retransmitted, and the maximum number of retransmissions is set to 10 times to prevent data loss caused by network congestion or temporary interruption of satellite signals.

[0046] At the same time, using the link redundancy mechanism of low-earth orbit satellites, a multi-link transmission strategy is adopted. For example, the low-earth orbit satellite communication module of the vehicle establishes connections with more than 4 low-earth orbit satellites at the same time, and sends the data to the cloud server through multiple links. At the receiving end, the cloud server compares and verifies the multiple copies of data received, and selects the complete and accurate data for processing, effectively improving the reliability of data transmission.

[0047] S3. Data processing. The cloud server performs multi-level calculations and analyzes on the uploaded data, and realizes two sets of task flows based on map matching, including inferring traffic flow distribution, prediction and warning, and road topology structure analysis and update. Specifically as follows:

[0048] S31. Map Matching: The cloud server takes the traffic areas of key concern as the smallest units, collects the motion state data (i.e., the real-time heading, speed, and position of the vehicle) sent by intelligent vehicles passing through them within a certain period, and executes multiple groups of tasks. Taking the intersection in the bustling business district of a certain city as an example of the key traffic area, the range for the cloud server to collect the trajectory state data of intelligent vehicles is set as: a circular area with a radius of 2 kilometers centered on this business district, and the time period is set to collect once every 10 minutes. The preliminary matching of map data is realized by using a navigation-level map, and further, the particle filter technology is adopted to perform high-precision docking of the vehicle trajectory with the road map to ensure the accurate matching of each vehicle in the map position, so as to remove obvious noise data and complete the missing state variables at the same time.

[0049] Adopt the particle filter technology to perform high-precision docking of the vehicle trajectory with the road map, specifically: ; where is the posterior probability of the state variable at the moment to be inferred; is the observation model, that is, obtained from the data uploaded by the vehicle; is the prior distribution, that is, the state variable speculation made through the vehicle kinematic model and historical state input; is the normalization factor of the observation and is an empirical scalar selected.

[0050] S32. Real-time Traffic Flow Analysis, Prediction, and Warning. This task flow will analyze the current traffic flow state in the current traffic area based on the vehicle state data collected in the short term, and identify congestion or dangerous situations in a timely manner. The cloud server calculates macroscopic indicators such as average vehicle speed, traffic flow density, and vehicle spacing distribution by collecting all trajectory data within a specified time window, and uses the Transformer-based heterogeneous element interaction model for driving scenarios to analyze the dynamic relationship between vehicles and road map elements to obtain the traffic flow state, congestion, and dangerous situations.

[0051] The data structure collected by the cloud server includes the collected vehicle motion state information and the prior map information : , ; where is the time-series state vector of the th vehicle in the scenario, is the vectorized representation map element of the th in the scenario, and respectively represent the vehicles and road map elements included in the currently detected driving scenario.

[0052] In the data screening and preprocessing stage, a statistics-based outlier detection method is adopted, that is, reasonable threshold ranges are set according to the statistical distribution laws of parameters such as vehicle speed and acceleration under normal driving conditions. For example, for speed data, calculate the mean and standard deviation of the speeds of all vehicles in this area, and regard the speed values exceeding the range of mean ± 3 times the standard deviation as outliers for elimination. For acceleration data, a similar statistical method is also used for outlier processing to remove abnormal data caused by sensor failures or special driving conditions (such as rapid start after sudden braking).

[0053] Then and are input into the Transformer-based heterogeneous scenario element interaction model to analyze the comprehensive spatio-temporal interaction relationships of scenario elements: ; where C are the macroscopic index characteristics such as the average vehicle speed, traffic flow density, and vehicle spacing distribution obtained by preprocessing, and H is the scenario dynamic interaction feature information; multiple decoder output heads will use the interaction feature information to analyze and warn of traffic flow status, congestion, and dangerous situations: ; The decoder predicts and identifies congestion or collision risk areas and their risk probabilities by analyzing the traffic flow density and vehicle speed in each area, combining historical data and current spatio-temporal interaction characteristics. Among them, R has multiple groups of sequences, and each group of sequences includes the area positions with congestion or collision risks and the corresponding risk values.

[0054] The present invention realizes scenario risk awareness and prediction in cloud-based traffic flow analysis, and further marks and sends congestion or danger information to other relevant vehicles to provide risk warnings; at the same time, it outputs a visualized traffic flow status map for real-time monitoring and decision-making in the cloud or traffic management center. At the same time, it continuously collects new real-time data to update the prediction model (i.e., the Transformer-based heterogeneous scenario element interaction model) to improve long-term accuracy. In terms of the generation and display of the visualized traffic flow status map, the geographic information system (GIS) technology is used to present the traffic flow data in an intuitive graphical way. On the cloud platform, different colored lines are used to represent the vehicle driving speeds on different lanes. Green represents the unobstructed lanes with a vehicle speed greater than 40 km / h, yellow represents the slow lanes with a vehicle speed between 20 - 40 km / h, and red represents the congested lanes with a vehicle speed less than 20 km / h. At the same time, the occurrence positions of traffic abnormal events (such as accidents, sudden brakes, etc.) are marked with flashing icons, enabling traffic management personnel to clearly understand the traffic conditions in the entire area, providing strong support for traffic decision-making, and also facilitating vehicle drivers to obtain real-time traffic flow information through in-vehicle displays or mobile phone applications to optimize travel routes.

[0055] When training and validating the heterogeneous scenario element interaction model using historical traffic data, taking the traffic data of the intersection to be monitored in this city in the past year as an example, 80% of it is used as the training set and 20% as the validation set. During the training process, an adaptive learning rate adjustment strategy is adopted, and the learning rate is dynamically adjusted according to the loss value of the model on the validation set. In terms of the online update mechanism of the model, every 1 hour, the newly collected real-time traffic data is used to perform incremental learning and update on the model to adapt to the dynamic changes in traffic conditions, ensuring that the prediction performance of the model always remains at a high level and providing continuous and reliable support for traffic management.

[0056] S33. Road topology structure analysis and update. The long-term accumulated vehicle motion state data in the detection area can be used to analyze and update the lane centerlines and topology structure of the road, ensuring the real-time and accuracy of road information. Specifically as follows:

[0057] Trajectory data accumulation: Collect the trajectory data of all vehicles in the detection area within a long time window (such as several hours to several days), including historical positions, driving directions, and speeds; in this embodiment, the time window for long-term trajectory data accumulation is set to several days, specifically 3 days. In terms of data processing methods, a coordinate conversion algorithm based on spatial projection is adopted to convert the longitude and latitude coordinates of vehicles into plane coordinates based on the road, forming a set of road trajectory points.

[0058] Lane centerline extraction: Use clustering algorithms (DBSCAN or MeanShift) to cluster the trajectory points into lane clusters; and use polynomial regression or B-spline curve fitting for the data points of each cluster to generate the lane centerline. When using the clustering algorithm, optimize the clustering parameters according to the actual width and traffic flow of the road: for the main road with a large traffic flow and a lane width of 3.5 meters, set the neighborhood radius to 2 meters and the minimum number of sample points to 50 to ensure that the trajectory points can be accurately clustered into lane clusters; for the branch road with a small traffic flow and a lane width of 3 meters, set the neighborhood radius to 1.5 meters and the minimum number of sample points to 30. In terms of the application of curve fitting methods, for straight lanes, use second-order polynomial regression for fitting; for curves with a curvature radius less than 100 meters, use third-order B-spline curve fitting to ensure that the centerline can accurately describe the shape of the lane and control the average error of the lane centerline within 0.3 meters.

[0059] Construction of the lane topology relationship graph: Combine the lane centerlines and the turning and lane-changing behaviors of vehicle trajectories, and use a graph theory model to construct a road topology relationship graph, defining: nodes represent intersections or key positions, edges represent lane connection relationships, the edge weights are traffic constraints (such as whether there are traffic restrictions, speed limits, flow limits, etc.), and at the same time mark special topological structures (such as ramps, roundabouts, or confluence lanes).

[0060] Combined with the actual road changes, such as the construction of a new ramp in the current area, by analyzing the vehicle trajectory data, the turning behavior and speed change characteristics of vehicles in this area are identified, the starting point, ending point and connection relationship of the ramp are determined, and they are accurately added to the road topology relationship map. In terms of navigation map comparison and update, on a daily basis, the generated road topology relationship map is compared with the lane information in the existing navigation map to detect newly added, closed or adjusted lanes, mark the changed positions, and perform incremental updates on the high-precision map; at the same time, based on the detection results, the navigation map is updated and sent to relevant vehicles to ensure the accuracy of the information. For example: when it is detected that a certain road is closed due to construction, through the statistical analysis of vehicle trajectory data, the position and impact range of the closed lane are determined, and the change information is updated to the navigation map in a timely manner to ensure that the navigation map can accurately reflect the actual situation of the road, provide a reliable basis for vehicle navigation and path planning, and improve the overall operation efficiency of the intelligent transportation system.

[0061] S34. Comprehensive strategy and task coordination. To efficiently execute these tasks, the present invention designs task coordination and priority management under cloud control. Real-time traffic flow analysis has the highest priority and performs real-time corresponding traffic flow prediction and early warning; road topology analysis is used as an auxiliary, mainly responsible for long-term update and improvement of information sharing: the output results of each task can be shared through the cloud server to each in-vehicle terminal and traffic management center, using low-earth orbit satellite communication to ensure the timeliness and efficiency of information transmission. The priority dynamic adjustment mechanism among tasks such as real-time traffic flow analysis, traffic flow prediction and early warning, and road topology analysis is as follows: Under normal traffic conditions, the priority of real-time traffic flow analysis is set to the highest, and it is executed once every 5 minutes; the traffic flow prediction and early warning task is triggered according to the results of real-time traffic flow analysis. When there are obvious changes in traffic flow (such as sudden changes in vehicle speed, increase in traffic density), the prediction and early warning task is immediately executed; road topology analysis is performed for long-term update once a day to assist other tasks; when a traffic anomaly event (such as an accident, bad weather, etc.) occurs, the priority of real-time traffic flow analysis is further increased, and the execution cycle is shortened to once every 2 minutes. At the same time, the traffic flow prediction and early warning task also speeds up the execution frequency to more timely respond to traffic changes and ensure traffic safety.

[0062] The specific way of information sharing adopts a publish / subscribe model based on the cloud platform, and the data format adopts JSON (JavaScript Object Notation) format to facilitate parsing and interaction between different systems and devices. For example, road topology information is published in the form of a JSON object containing fields such as road nodes, lane connection relationships, and lane attributes (such as width, direction, speed limit); traffic flow information is published in the form of a JSON array containing fields such as area ID, average vehicle speed, traffic flow density, and congestion status. In this way, vehicles and traffic management systems can quickly and accurately obtain the required information and achieve efficient information sharing.

[0063] S4. Information distribution, distributing the traffic flow and road topology information processed by the cloud server to relevant vehicles or traffic management systems. The specific steps include:

[0064] Distribution rule: When determining the target vehicle according to the vehicle's position and driving direction, a method combining geofencing technology and vehicle driving trajectory prediction is adopted. First, use the Geographic Information System (GIS) to construct a geofence for the target area on the cloud server side, divide it into multiple sub-areas, and each sub-area has a unique identifier. When the vehicle uploads its position information, the cloud server quickly filters out the set of vehicles that may need to receive information by judging the geofence sub-area where the vehicle is located.

[0065] At the same time, combining the vehicle's historical driving trajectory data and real-time speed and direction information, use the linear regression model in machine learning to make a short-term prediction of the vehicle's driving trajectory. For example, by analyzing the driving trajectory points of the vehicle in the past 10 minutes, determine its driving direction and speed change trend, and predict the area range that the vehicle may reach in the next 20 minutes. Then, match the prediction result with the geofence sub-area to further accurately determine the target vehicle that needs to receive specific information.

[0066] For preferentially distributing information to vehicles closer to the event point, a distance threshold of 20 kilometers is set. When a traffic event (such as congestion, accident, etc.) is detected, the cloud server immediately calculates the vehicles within 20 kilometers around the event point and preferentially pushes the relevant traffic flow and road topology information to these vehicles. In terms of the timeliness of information push, it is required that the time delay from the event occurrence to the information being pushed to the target vehicle does not exceed 5 seconds to ensure that the vehicle can obtain the latest traffic information in time and make decisions.

[0067] Data content: Road topology information: including lane division updated in real time, section passing capacity, road closure or congestion status; Traffic flow information: such as average vehicle speed, traffic flow density, and predicted driving path.

[0068] In the road topology information, the lane division information is stored in vector data format. Each lane consists of a series of ordered coordinate points with a coordinate accuracy of up to 0.2 meters, which can accurately describe the shape and position of the lane. The road section capacity is estimated based on the road design standards (such as the number of lanes and road grade) and historical traffic flow data, in units of the number of vehicles passing through per hour, and the data update frequency is once every half hour to reflect the real-time changes in the road capacity.

[0069] Transmission mechanism: Low Earth Orbit (LEO) satellite communication is used to achieve real-time data distribution; for nearby vehicles, information can be directly shared through V2V (vehicle-to-vehicle communication).

[0070] The described embodiments are the preferred embodiments of the present invention, but the present invention is not limited to the above embodiments. Without departing from the essential content of the present invention, any obvious improvements, substitutions, or modifications that can be made by those skilled in the art all fall within the protection scope of the present invention.

Claims

1. An integrated space-earth dynamic traffic flow analysis and road topology inference device, characterized in that Including: A ground monitoring station for monitoring on-orbit satellite data; An in-vehicle module of an intelligent vehicle, which obtains the real-time heading, speed and position of the vehicle based on low-orbit satellite data and ground monitoring station data; A cloud server, which receives the data transmitted by the in-vehicle module of the intelligent vehicle, performs multi-level calculations and analyses, and realizes two sets of task flows on the basis of map matching, including traffic flow distribution, prediction and early warning, and road topology structure analysis and update; A ground information interaction module, which is used to receive the road topology and traffic flow information processed by the cloud server and distribute it to target vehicles or traffic management systems to realize the sharing of road conditions and traffic information.

2. The integrated space-ground dynamic traffic flow analysis and road topology inference device according to claim 1, wherein The in-vehicle module of the intelligent vehicle includes, but is not limited to, an inertial measurement unit, a wheel speed sensor, a low-orbit satellite communication module and a data processing unit.

3. A method for an integrated space-ground dynamic traffic flow analysis and road topology inference device according to claim 2, characterized in that: The in-vehicle module of the intelligent vehicle obtains the real-time motion state data of the vehicle based on low-orbit satellite data and ground monitoring station data, and uploads it to the cloud server; The cloud server uses the motion state data of intelligent vehicles passing through key traffic areas to perform preliminary matching of map data using a navigation-level map, and further performs high-precision docking of vehicle trajectories with road maps to achieve precise matching of each vehicle in the map position; Traffic flow distribution, prediction and early warning: The cloud server calculates the average vehicle speed, traffic flow density, and vehicle spacing distribution by collecting all motion state data within a specified time window, and uses a heterogeneous element interaction model for driving scenarios based on Transformer to analyze the dynamic relationship between vehicles and road map elements to obtain traffic flow states, congestion and dangerous situations; The long-term accumulated vehicle motion state data in the detection area is used to analyze and update the lane center line and lane topology relationship diagram of the road to achieve the analysis and update of the road topology structure; The traffic flow information and road topology are distributed to target vehicles or traffic management systems.

4. The integrated space-ground dynamic traffic flow analysis and road topology inference method according to claim 3, characterized in that The data structure collected by the cloud server includes the collected vehicle motion state information and the prior map information : , ; where is the time-series state vector of the th vehicle in the scenario, is the vectorized representation of the th map element in the scenario, and respectively represent the vehicle and road map elements included in the currently detected driving scenario; input and into the heterogeneous scenario element interaction model based on Transformer to analyze the comprehensive spatio-temporal interaction relationship of scenario elements and obtain the scenario dynamic interaction feature information; use the scenario dynamic interaction feature information to analyze the traffic flow state, congestion, and dangerous situations, further analyze the traffic flow density and vehicle speed in each traffic area, and combine historical data and current interaction feature information to predict and identify congestion or collision risk areas and their risk probabilities and output them.

5. The method for analyzing the integrated space-earth dynamic traffic flow and inferring the road topology according to claim 3, wherein Combining the lane center line and the turning and lane-changing behaviors in the vehicle trajectory, a lane topology relationship diagram is constructed using a graph theory model, and it is defined that: nodes represent intersections or key positions, edges represent lane connection relationships, the edge weights are traffic constraints, and special topology structures are marked.

6. The method for analyzing the integrated space-earth dynamic traffic flow and inferring the road topology according to claim 4, wherein The update of the road topology structure is specifically as follows: By comparing the existing high-precision road map information with the lane topology relationship diagram, newly added, closed or adjusted lanes are detected, the changed positions are marked, and the high-precision map is incrementally updated; at the same time, based on the detection results, the navigation map is updated and sent to relevant vehicles.

7. The method for analyzing the integrated space-earth dynamic traffic flow and inferring the road topology according to claim 3, characterized in that When uploading to the cloud server, traffic anomaly information adopts an immediate transmission strategy and is given priority to be uploaded, and regular vehicle status information is transmitted at a predetermined frequency.

8. The integrated space-ground dynamic traffic flow analysis and road topology inference method according to claim 3, characterized in that When uploading to the cloud server, the time slot allocation for data transmission is dynamically adjusted according to the number of vehicles and communication bandwidth.

9. The method for analyzing the integrated space-earth dynamic traffic flow and inferring the road topology according to claim 3, wherein The traffic flow analysis has the highest priority and performs real-time corresponding traffic flow prediction and early warning; the road topology analysis is used as an auxiliary and is responsible for long-term update and improvement of information sharing.

10. The integrated space-ground dynamic traffic flow analysis and road topology inference method according to claim 3, characterized in that The distribution rule is: According to the vehicle position and driving direction, the target vehicles that need to receive data are determined, and the information is preferentially distributed to the target vehicles closer to the event point.

Images