Vehicle state perception method and system in vehicle-infrastructure integrated environment

Abstract

The application discloses a vehicle state perception method and system in a vehicle-road cooperative environment, and relates to the technical field of intelligent transportation, comprising: collecting real-time traffic data of a target intersection; establishing a temporary V2X communication link; receiving real-time traffic data and K real-time vehicle state information of K driving vehicles; if the single-dimensional risk detection results of the real-time traffic data and the K real-time vehicle state information by a roadside edge computing device are all 0, compensating the K real-time vehicle state information with the real-time traffic data to obtain K compensated vehicle state information, constructing a global dynamic traffic model, and predicting trajectory conflicts to locate potential collision risks; according to a risk avoidance priority, broadcasting the potential collision risks to the K driving vehicles through the V2X communication link. The application solves the technical problems of inaccurate traffic modeling and untimely collision risk identification caused by incomplete vehicle-mounted perception information in the prior art, and achieves the technical effects of improving the accuracy of vehicle state perception and the timeliness of collision risk warning.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent transportation, in particular to a vehicle state perception method and system in a cooperative vehicle infrastructure environment. BACKGROUND

[0002] In a cooperative vehicle infrastructure environment, vehicles usually rely on vehicle-mounted perception devices to obtain their own state and surrounding environment information. However, due to the limitations of sight distance obstruction, weather interference or perception range, the information obtained by the vehicle-mounted devices has the problems of poor timeliness and insufficient integrity, which leads to incomplete vehicle state information and makes it difficult to support accurate modeling of the dynamic relationship between multiple vehicles in complex scenarios such as intersections, thereby affecting the timely identification and effective avoidance of potential collision risks. SUMMARY

[0003] The present application provides a vehicle state perception method and system in a cooperative vehicle infrastructure environment, which is used to solve the technical problems of inaccurate traffic modeling and untimely collision risk identification caused by incomplete vehicle-mounted perception information in the prior art.

[0004] In view of the above problems, the present application provides a vehicle state perception method and system in a cooperative vehicle infrastructure environment.

[0005] In a first aspect of the present application, a vehicle state perception method in a cooperative vehicle infrastructure environment is provided, which comprises:

[0006] The roadside sensing network collects real-time traffic data of a target intersection by a millimeter wave radar; a temporary V2X communication link between a vehicle-mounted OBU of a vehicle entering the target intersection and a roadside edge computing device is established based on a geographic fence trigger; the roadside edge computing device receives the real-time traffic data and K real-time vehicle state information of K driving vehicles, where K≥2 and K is a positive integer; if the single-dimensional risk detection results of the roadside edge computing device on the real-time traffic data and the K real-time vehicle state information are all 0, the K real-time vehicle state information is compensated by the real-time traffic data to obtain K compensated vehicle state information; a global dynamic traffic model is constructed based on the K compensated vehicle state information; trajectory conflict prediction is performed based on the global dynamic traffic model to locate potential collision risks; and the potential collision risks are broadcast to the K driving vehicles through the V2X communication link according to the risk avoidance priority.

[0007] In a second aspect of the present application, a vehicle state perception system in a cooperative vehicle infrastructure environment is provided, which comprises:

[0008] The traffic data collection module is used for collecting real-time traffic data of a target intersection by a roadside sensing network through a millimeter wave radar; the communication link establishment module is used for establishing a temporary V2X communication link between an on-board OBU of a vehicle entering the target intersection and a roadside edge computing device based on a geographic fence trigger; the information receiving module is used for receiving the real-time traffic data and K pieces of real-time vehicle state information of K driving vehicles by the roadside edge computing device, where K is greater than or equal to 2 and K is a positive integer; the compensation module is used for compensating the K pieces of real-time vehicle state information by the real-time traffic data to obtain K pieces of compensated vehicle state information if single-dimensional risk detection results of the real-time traffic data and the K pieces of real-time vehicle state information by the roadside edge computing device are all 0; the traffic model construction module is used for constructing a global dynamic traffic model based on the K pieces of compensated vehicle state information; the trajectory conflict prediction module is used for performing trajectory conflict prediction based on the global dynamic traffic model to locate potential collision risks; and the broadcasting module is used for broadcasting the potential collision risks to the K driving vehicles through the V2X communication link according to a risk avoidance priority.

[0009] One or more technical solutions provided in the present application have at least the following technical effects or advantages:

[0010] The roadside sensing network collects real-time traffic data of a target intersection through a millimeter wave radar; a temporary V2X communication link is established between an on-board OBU of a vehicle entering the target intersection and a roadside edge computing device based on a geographic fence trigger; the roadside edge computing device receives the real-time traffic data and K pieces of real-time vehicle state information of K driving vehicles, where K is greater than or equal to 2 and K is a positive integer; if single-dimensional risk detection results of the real-time traffic data and the K pieces of real-time vehicle state information by the roadside edge computing device are all 0, the K pieces of real-time vehicle state information are compensated by the real-time traffic data to obtain K pieces of compensated vehicle state information; a global dynamic traffic model is constructed based on the K pieces of compensated vehicle state information; trajectory conflict prediction is performed based on the global dynamic traffic model to locate potential collision risks; and the potential collision risks are broadcast to the K driving vehicles through the V2X communication link according to a risk avoidance priority. The present application solves the technical problems of inaccurate traffic modeling and untimely collision risk identification caused by incomplete vehicle perception information in the prior art, and achieves the technical effects of improving vehicle state perception accuracy and collision risk warning timeliness by introducing roadside sensing data to compensate vehicle state information and constructing a global dynamic traffic model for trajectory conflict prediction. BRIEF DESCRIPTION OF DRAWINGS

[0011] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0012] Figure 1 A schematic diagram of the vehicle status perception method in a vehicle-road cooperative environment provided in the embodiments of this application;

[0013] Figure 2 This is a schematic diagram of the vehicle status perception system in a vehicle-road cooperative environment provided in an embodiment of this application.

[0014] Figure labeling: Traffic data acquisition module 11, communication link establishment module 12, information receiving module 13, compensation module 14, traffic model construction module 15, trajectory conflict prediction module 16, broadcasting module 17. Detailed Implementation

[0015] This application provides a vehicle status perception method and system in a vehicle-road cooperative environment. It addresses the technical problems of inaccurate traffic modeling and untimely collision risk identification caused by incomplete vehicle-mounted perception information in existing technologies. By introducing roadside sensor data to compensate for vehicle-mounted vehicle status information and constructing a global dynamic traffic model for trajectory conflict prediction, it achieves the technical effect of improving the accuracy of vehicle status perception and the timeliness of collision risk warning.

[0016] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0017] It should be noted that any variation of the terms "comprising" and "having" is intended to cover non-exclusive inclusion, for example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units that are explicitly listed, but may include other steps or modules that are not explicitly listed or that are inherent to such processes, methods, products, or devices.

[0018] Example 1, as Figure 1 As shown, this application provides a vehicle status perception method in a vehicle-road cooperative environment, the method comprising:

[0019] Step S100: The roadside sensor network collects real-time traffic data of the target intersection using millimeter-wave radar.

[0020] In the embodiments of the present application, the side sensing network collects real-time traffic data of the target intersection through a millimeter wave radar. The millimeter wave radar detects vehicles within the target intersection by emitting and receiving electromagnetic wave signals, obtains information such as the position, speed, and distance of the vehicles. The radar signal is reflected back to the receiver after encountering an object, and by measuring the round-trip time of the signal, the precise distance and relative speed of the object are calculated, thereby obtaining real-time traffic data of each vehicle within the target intersection.

[0021] Step S200: Based on the geographic fence triggering, the vehicle-mounted OBU of the vehicle entering the target intersection establishes a temporary V2X communication link with the roadside edge computing device.

[0022] In the embodiments of the present application, based on the geographic fence triggering, the vehicle-mounted OBU of the vehicle entering the target intersection establishes a temporary V2X communication link with the roadside edge computing device. Specifically, first, the geographic fence technology defines an area around the target intersection by setting a virtual boundary, and once the vehicle enters the area, the geographic fence system triggers an event. The vehicle-mounted OBU (On-Board Unit, vehicle-mounted unit) as an intelligent device within the vehicle, monitors the position information of the vehicle in real time. When the vehicle enters the virtual boundary, the OBU obtains the accurate position of the vehicle through the GPS positioning system, and sends the current position data to the roadside edge computing device. After receiving the vehicle position data, the roadside edge computing device starts the V2X (Vehicle-to-Everything, vehicle networking) communication link, which is a temporary data transmission connection based on wireless communication technology, and can realize fast exchange of information between the vehicle and the roadside device. Through this communication link, the vehicle-mounted OBU and the roadside edge computing device can realize bidirectional transmission of vehicle state data and real-time traffic information, ensuring efficient interaction between the vehicle and the traffic management system.

[0023] Step S300: The roadside edge computing device receives the real-time traffic data and K real-time vehicle state information of K driving vehicles, where K ≥ 2, K is a positive integer.

[0024] In the embodiments of the present application, the side edge computing device receives real-time traffic data and K real-time vehicle state information of K driving vehicles, where K≥2, K is a positive integer. Specifically, the roadside sensor network continuously monitors the traffic situation of the target intersection through millimeter wave radar and other sensing devices to obtain real-time traffic data, such as the number of vehicles, traffic flow, density, and other traffic environment information. At the same time, the vehicle-mounted OBU is responsible for monitoring and transmitting the real-time vehicle state information of the vehicle, including the position, speed, acceleration, and turn signal state. These vehicle state information are collected in real time by the vehicle-mounted device: the vehicle position is obtained by the GPS system, the speed and acceleration are provided by the vehicle-mounted sensor (such as the vehicle speed sensor, accelerometer), and the turn signal state is detected by the control system of the vehicle. When K driving vehicles enter the coverage area of the roadside sensor network, the vehicle-mounted OBU sends the vehicle state information and the current position of each vehicle to the roadside edge computing device through the V2X communication link.

[0025] Step S400: If the single-dimensional risk detection results of the real-time traffic data and K real-time vehicle state information by the roadside edge computing device are all 0, the K real-time vehicle state information is compensated by the real-time traffic data to obtain K compensated vehicle state information.

[0026] In the embodiments of the present application, first, the roadside edge computing device performs single-dimensional risk detection on the real-time vehicle state information of each vehicle to check whether there is risk in each dimension (such as vehicle position, speed, acceleration, and turn signal state). For example, the speed of the vehicle is compared with the speed limit, the acceleration is compared with the preset threshold, the turn signal state is checked for consistency with the driving direction of the vehicle, and the vehicle position is compared with the lane boundary. If the detection result of any dimension shows that there is risk, the device will set the corresponding detection result to 1, indicating that there is potential risk; if the detection results of all dimensions are 0, it means that the current state of the vehicle is not detected to be abnormal.

[0027] When all single-dimensional risk detection results are 0, the roadside edge computing device compensates the vehicle state information of K vehicles using real-time traffic data. Real-time traffic data includes traffic flow on the road, position and speed information of other vehicles, etc., which can provide a reference for the compensation process. When the vehicle state information (such as speed, position, etc.) of a vehicle is incorrect or missing, the roadside edge computing device will compensate it using real-time traffic data. For example, based on the traffic flow of the target intersection and the state of the adjacent vehicles, the device can infer the possible position or speed of a vehicle. In this way, the device can fuse the real-time traffic data with the vehicle state information of K driving vehicles to generate K compensated vehicle state information.

[0028] Further, the method provided by the embodiments of the present application further comprises:

[0029] According to the road condition information of the target intersection, a traffic-related risk feature is matched and called; the roadside edge computing device uses the traffic-related risk feature to traverse the real-time traffic data, and outputs a first risk detection result; according to the position information of the target intersection, a driving speed limit feature is called in a network; the roadside edge computing device uses the driving speed limit feature to traverse the K real-time vehicle state information, and outputs a second risk detection result; if the first risk detection result and the second risk detection result are both 0, then the K real-time vehicle state information is compensated by using the real-time traffic data, and K compensated vehicle state information is obtained.

[0030] In the embodiment of the present application, the roadside edge computing device first matches and calls the corresponding traffic-related risk feature template in the local or cloud database according to the current road condition information (such as traffic flow density, road topological structure, signal state, etc.) of the monitored target intersection. Common risk features include two key indicators: too small inter-vehicle distance and vehicle reverse. Among them, the judgment of inter-vehicle distance is based on whether the Euclidean distance between vehicles is less than a dynamic safety threshold, which is dynamically set according to vehicle speed and traffic environment; the reverse judgment is based on whether the angle between the actual driving direction of the vehicle and the direction defined by the road topological structure exceeds a set threshold (such as 180°±30°). Then the traffic data sequence returned by the radar is analyzed frame by frame using a sliding window traversal method, and the spatial relationship of each pair of vehicles is analyzed and the direction consistency is verified, and a first risk detection result is output.

[0031] After completing the traffic environment risk feature detection, the roadside edge computing device obtains the speed limit standard and driving specification library corresponding to the road segment based on the geographic coordinate information of the intersection, and calls the driving speed limit feature. In this stage, the speed field in the real-time vehicle state information of K vehicles is extracted, and a judgment method based on the speed limit upper and lower threshold interval is used for analysis. If the speed of a vehicle exceeds the set speed limit threshold (such as speed limit 60km / h, tolerance ±10%), it is determined that the vehicle has a speed limit risk. The detection method combines time stamp and position data for dynamic speed calculation, that is, by differentiating the position difference between adjacent time points and dividing by the time interval to obtain a more true instantaneous speed. After analyzing K vehicles, a second risk detection result is generated.

[0032] After receiving the first risk detection result and the second risk detection result, Boolean logic judgment is performed. If all K vehicles do not trigger an alarm in both risk dimensions (i.e., the detection results are all 0), it is considered that there is no significant traffic conflict or abnormal behavior at the current intersection, and the compensation mechanism startup condition is met.

[0033] Under the premise that the risk detection result is 0, K real-time vehicle state information is compensated by using real-time traffic data. In this process, first, the real-time position data of each vehicle is extracted from the K real-time vehicle state information. Then, the positions of these vehicles are offset compensated in combination with the real-time traffic data to correct the position deviation caused by signal interference or sensor error, to generate K compensated vehicle positions. Then, according to the compensated vehicle positions, the real-time speed and heading angle information related to each vehicle is extracted from the real-time traffic data. These detection state data are weighted and fused with the original vehicle state information to ensure that the compensated data are more accurate, and finally the compensated vehicle positions are mapped to the original vehicle state information, and K compensated vehicle state information is output.

[0034] Further, the method provided by the application embodiment further comprises:

[0035] The traffic-related risk features include vehicle safety distance constraints and vehicle driving direction constraints; the vehicle safety distance constraints and the vehicle driving direction constraints are used to traverse the real-time traffic data to locate real-time risk vehicles as the first risk detection result; if the first risk detection result is 1, the real-time V2X communication link of the real-time risk vehicle is called according to the K real-time vehicle state information; and the real-time V2X communication link is used to send a risk warning to the real-time risk vehicle.

[0036] In the application embodiment, the traffic-related risk features include vehicle safety distance constraints and vehicle driving direction constraints. First, the roadside edge computing device obtains the state data such as the position information, speed, acceleration and driving direction of each vehicle through real-time traffic data. The vehicle safety distance constraint is to calculate the distance between adjacent vehicles to ensure that the safety distance between vehicles is not less than a preset standard. This standard is generally set according to dynamic factors such as vehicle speed, road conditions and vehicle speed changes. If the distance between vehicles is less than the safety distance, it means that there is a potential collision risk. Similarly, the vehicle driving direction constraint is to compare the actual driving direction of the vehicle with the driving direction specified by the road. If the vehicle deviates from the correct driving lane or there is a reverse situation, the behavior of the vehicle constitutes a potential risk.

[0037] When the roadside edge computing device applies these two constraint conditions to the real-time traffic data, it traverses the relative positions between each vehicle and the surrounding vehicles, calculates the safety distance, and judges whether the driving direction of the vehicle is compliant. If the safety distance between a vehicle and the vehicle in front or around is too small, or the vehicle has a reverse or deviates from the driving route, the device will locate it as a real-time risk vehicle and output the first risk detection result. If the detection result is 1, it means that a potential risk vehicle is identified.

[0038] In the case of a first risk detection result of 1, the roadside edge computing device utilizes K real-time vehicle state information, especially real-time vehicle position, speed, acceleration, etc. data, to call the V2X communication link of the real-time risk vehicle. The V2X communication link is a data communication interface between the vehicle and other vehicles or roadside facilities. Through the V2X communication link, the roadside edge computing device can establish a communication connection with the real-time risk vehicle and send a risk warning to the vehicle in real time. This warning will remind the vehicle driver to pay attention to the possible danger or abnormal situation in front, so as to take appropriate measures to reduce the risk of potential collision.

[0039] Further, the method provided by the application embodiment further comprises the following steps after sending the risk warning to the real-time risk vehicle through the real-time V2X communication link:

[0040] presetting an alarm execution window; after the risk warning sending time delay to the real-time risk vehicle satisfies the alarm execution window, the roadside edge computing device receives updated traffic data returned by the roadside sensor network coverage; and the roadside edge computing device uses the traffic-related risk features to make a related risk persistence judgment on the updated traffic data.

[0041] In the application embodiment, the preset alarm execution window refers to a time range set after sending the risk warning to ensure that the alarm is responded in time. This time window is preset by technical experts according to the traffic conditions of the intersection, vehicle reaction time and other factors to ensure that the driver has enough time to receive the alarm and take appropriate measures. The length of the alarm execution window usually considers the driving speed of the vehicle, reaction delay and road characteristics and other variables to ensure that the alarm information is effective and can be processed by the driver in time.

[0042] After the risk warning sending time delay to the real-time risk vehicle satisfies the alarm execution window, the roadside edge computing device starts to receive updated traffic data from the roadside sensor network. The roadside sensor network includes millimeter wave radar, camera, laser radar and other sensors, which collect vehicle information and traffic condition data in the target area and transmit them to the edge computing device in real time. These updated traffic data include vehicle position information, speed, acceleration, traffic flow and road condition changes.

[0043] After receiving the updated traffic data, the roadside edge computing device will make a correlation risk persistence judgment on the data in combination with traffic correlation risk features. The traffic correlation risk features include multiple dimensions such as the safety distance between vehicles, vehicle speed, driving direction, etc. For example, the device will judge whether there is still a collision risk according to the distance between the current position of the vehicle and other vehicles; at the same time, it will evaluate whether the vehicle has taken appropriate deceleration measures according to the vehicle speed and road conditions. The device analyzes whether there is a potential risk that persists according to these risk features, if the risk persists, it continues to monitor, if the risk has been eliminated, it updates the risk state and terminates tracking.

[0044] Further, the method provided by the application embodiment further comprises:

[0045] The K real-time vehicle positions are called from the K real-time vehicle state information; the K real-time vehicle positions are offset compensated by checking the K real-time vehicle positions using the real-time traffic data to obtain K compensated vehicle positions; K sets of detection state data are called from the real-time traffic data according to the K compensated vehicle positions, wherein the detection state data includes real-time speed and vehicle heading angle of the vehicle; the K sets of detection state data are mapped and weighted fused into the K real-time vehicle state information, the K compensated vehicle positions are mapped and overlaid on the K real-time vehicle state information, and the K compensated vehicle state information is output.

[0046] In the application embodiment, first, K real-time vehicle positions are extracted from K real-time vehicle state information. The real-time vehicle state information of each vehicle has been obtained through the vehicle-mounted OBU and vehicle sensor system, including real-time position, speed, acceleration and heading angle of the vehicle and other data. In this step, these obtained real-time vehicle position data are directly accessed and extracted through the V2X communication link or other data acquisition network, to ensure that the device can extract K real-time vehicle positions from the existing data.

[0047] Then, the K real-time vehicle positions are checked and offset compensated using real-time traffic data. The real-time traffic data is usually collected by roadside sensors (such as millimeter wave radar, etc.), including position information of other vehicles, traffic flow, vehicle speed, etc. In this process, these real-time traffic data are compared with the K real-time vehicle positions to determine whether there is a position offset. If there is an offset, the device will compensate the position of each vehicle through a spatial correction algorithm (such as Kalman filtering or least squares method), to obtain K compensated vehicle positions. This compensation process ensures that the vehicle position is more accurate, reducing the influence of GPS error or other measurement error.

[0048] After completing the vehicle position compensation, K sets of detection state data are extracted from the real-time traffic data according to the K compensated vehicle positions. The detection state data includes the real-time speed and heading angle of the vehicle, which are obtained by real-time acquisition through vehicle-mounted sensors such as vehicle speed sensors, accelerometers, gyroscopes, etc. The device will extract the real-time speed and heading angle data related to the K compensated vehicle positions through real-time traffic data to form K sets of detection state data.

[0049] Finally, the extracted K sets of detection state data are fused with the K compensated vehicle positions. The weighted fusion process is to ensure the accuracy of the data by combining the weight of each data point. The device uses data fusion techniques such as weighted average method or Kalman filter to fuse the real-time speed, acceleration, and heading angle of each vehicle with the compensated vehicle position to obtain the final K sets of compensated vehicle state information.

[0050] Step S500: constructing a global dynamic traffic model based on the K sets of compensated vehicle state information.

[0051] In the embodiments of the present application, when constructing a global dynamic traffic model based on K sets of compensated vehicle state information, first, the K sets of compensated vehicle state information are associated with the lane topology to generate global bound vehicle state information, so that the dynamic data of each vehicle matches the structure of the lane it is in. Next, using these global bound vehicle state information, multi-target trajectory tracking is performed on the K vehicles to output their respective real-time trajectory sequences. Finally, these real-time trajectory sequences are fitted into the spatial modeling of the target intersection to ultimately obtain the global dynamic traffic model.

[0052] Further, the method provided by the embodiments of the present application further comprises:

[0053] associating the K sets of compensated vehicle state information with the lane topology to obtain global bound vehicle state information; performing multi-target trajectory tracking on the K vehicles according to the global bound vehicle state information to output K real-time trajectory sequences; fitting the K real-time trajectory sequences to the spatial modeling of the target intersection to obtain the global dynamic traffic model.

[0054] In the embodiments of the present application, first, the K sets of compensated vehicle state information are associated with the lane topology. In this process, spatial mapping technology is used to match the vehicle position, speed, and other dynamic data in each compensated vehicle state information with the lane topology data of the intersection. The lane topology data includes the geometric structure, lane connection relationship, and traffic flow direction of each lane of the intersection. By mapping the vehicle state information of each vehicle to a specific lane position, the dynamic state of the vehicle in its corresponding lane is accurately reflected, thereby generating global bound vehicle state information.

[0055] Next, according to the global binding vehicle state information, multi-target trajectory tracking is performed on the K driving vehicles, and K real-time trajectory sequences are output. In this process, the multi-target trajectory tracking (MOT) algorithm is used to track the dynamic data of the K vehicles in real time, and K real-time trajectory sequences are output to represent the driving trajectory of each vehicle.

[0056] Finally, the K real-time trajectory sequences are fitted to the spatial modeling of the target intersection to obtain a global dynamic traffic model. The device combines the lane topology data of the intersection with the real-time trajectory sequence, and maps the trajectory of the vehicle to the geometric structure of the intersection through the spatial modeling algorithm. This fitting process not only considers the geometric shape of the road, but also includes traffic signals, lane switching rules, traffic flow and other factors, to ensure that the driving trajectory of the vehicle matches the layout of the intersection and the traffic rules. Finally, through this process, a global dynamic traffic model is obtained, which reflects the traffic flow and vehicle behavior of the intersection and its surrounding area.

[0057] Step S600: trajectory conflict prediction based on the global dynamic traffic model to locate potential collision risks.

[0058] In the embodiments of the present application, when trajectory conflict prediction is performed based on the global dynamic traffic model to locate potential collision risks, first, the distribution of the conflict area is predefined according to the lane connection topology of the target intersection to determine the area where the collision may occur. Then, the CTRA model is used to predict the trajectory of the K real-time trajectory sequences to generate K predicted time series trajectories. Then, the spatiotemporal overlap algorithm is used to analyze these predicted time series trajectories to extract multiple conflict trajectories falling into the predefined conflict area. According to the mapping relationship between these conflict trajectories and the predicted trajectories, multiple driving vehicles involved are located, and the corresponding vehicle IDs are called, and finally these vehicle IDs are packaged as potential collision risks output.

[0059] Further, in the method provided by the embodiments of the present application, the trajectory conflict prediction based on the global dynamic traffic model to locate potential collision risks further comprises:

[0060] The distribution of the conflict area is predefined according to the lane connection topology of the target intersection; the CTRA model is used to predict the trajectory along the K real-time trajectory sequences to obtain K predicted time series trajectories; after spatiotemporal overlap of the K predicted time series trajectories, multiple conflict trajectories falling into the conflict area distribution are extracted; according to the mapping relationship between the multiple conflict trajectories and the K predicted time series trajectories, multiple driving vehicles are located; after calling multiple vehicle IDs of the multiple driving vehicles, the multiple vehicle IDs are packaged as the potential collision risk output.

[0061] In the embodiments of the present application, the distribution of the conflict area is predefined according to the lane connection topology of the target intersection. First, by analyzing the lane structure and traffic flow direction of the intersection, the area where the collision is likely to occur is determined. The conflict area usually includes the lane intersection, the lane changing area and other areas with dense traffic flow, which are considered as potential collision risk areas. Through the lane topology data and the intersection geometry, the spatial distribution of these conflict areas is predefined.

[0062] Then, the K real-time trajectory sequences are predicted using the CTRA (Continuous-Time Random Access) model. The CTRA model can predict the motion trajectory of each vehicle at future time under the continuous time framework, combining the current position, speed and acceleration of the vehicle. This process considers the dynamic behavior of the vehicle and generates K predicted time series trajectories, each describing the driving path of the vehicle in the future period of time.

[0063] After obtaining the K predicted time series trajectories, the trajectories are analyzed using the space-time overlap method. Through the space-time overlap algorithm, the predicted trajectory of each vehicle is compared with the predefined conflict area distribution to find out which vehicles' trajectories overlap with the conflict area. The space-time overlap refers to the spatial matching of the predicted trajectory of each vehicle and the synchronization in the time dimension, extracting the trajectories falling into the conflict area and identifying multiple conflict trajectories.

[0064] Next, according to the mapping relationship between the multiple conflict trajectories and the K predicted time series trajectories, the involved driving vehicles are located. By comparing each conflict trajectory with the actual driving trajectory of the vehicle, it is determined which vehicles will collide in the conflict area. At this time, through the real-time position, speed and driving direction of the vehicle, the vehicle ID involved in the collision risk is extracted, so as to determine these vehicles.

[0065] Finally, multiple vehicle IDs are packaged as potential collision risks and output. These vehicle IDs represent vehicles that may collide at future time.

[0066] Step S700: According to the risk avoidance priority, the potential collision risk is broadcast to the K driving vehicles through the V2X communication link.

[0067] In the embodiments of the present application, the potential collision risk is broadcast to the K driving vehicles through the V2X communication link according to the risk avoidance priority. Specifically, first, the severity and urgency of the potential collision are determined based on the collision prediction result. The risk avoidance priority is evaluated according to the spatial overlap degree, time window, and vehicle relative speed of the predicted collision risk, and the closer to the collision situation, the higher the risk priority. Then, according to the evaluated risk priority, each risk is assigned a corresponding emergency response level. For high-priority potential collision risks, immediate broadcast of warnings is selected; for low-priority risks, later or other ways of prompting may be selected.

[0068] Once the risk priority is determined, the roadside system broadcasts the relevant risk information to the K driving vehicles through the V2X communication link. Through the V2X communication link, detailed information of the potential collision risk (such as the time, location of the collision prediction, ID of the related vehicle, warning information, etc.) is sent to the on-board unit of the K driving vehicles in real time. This enables the driver to learn about the risk in a timely manner and take necessary risk avoidance measures, such as reducing speed, changing lanes, etc., thereby reducing the occurrence of potential collisions.

[0069] In the embodiments of the present application, as described above, the embodiments of the present application have at least the following technical effects:

[0070] The roadside sensing network of the present application collects real-time traffic data of the target intersection through a millimeter wave radar; a temporary V2X communication link is established between the on-board OBU of the vehicle entering the target intersection triggered by the geo-fencing and the roadside edge computing device; the roadside edge computing device receives the real-time traffic data and K real-time vehicle state information of K driving vehicles, where K≥2, K is a positive integer; if the single-dimensional risk detection results of the real-time traffic data and the K real-time vehicle state information are all set to 0, the K real-time vehicle state information is compensated using the real-time traffic data to obtain K compensated vehicle state information; a global dynamic traffic model is constructed based on the K compensated vehicle state information; trajectory conflict prediction is performed based on the global dynamic traffic model to locate potential collision risks; and the potential collision risks are broadcast to the K driving vehicles through the V2X communication link according to the risk avoidance priority. The present application solves the technical problems of incomplete vehicle sensing information leading to inaccurate traffic modeling and untimely collision risk identification in the prior art. By introducing roadside sensing data to compensate for vehicle state information and constructing a global dynamic traffic model for trajectory conflict prediction, the technical effects of improving the accuracy of vehicle state perception and the timeliness of collision risk warning are achieved.

[0071] Embodiment two, based on the same inventive concept as the vehicle state perception method in the aforementioned embodiment under the vehicle-road cooperation environment, such as Figure 2As shown, the present application provides a vehicle state perception system in a vehicle-road cooperation environment. The system and method embodiments in the present application are based on the same inventive concept. The system comprises:

[0072] a traffic data collection module 11 for collecting real-time traffic data of a target intersection by a roadside sensing network through a millimeter wave radar; a communication link establishment module 12 for establishing a temporary V2X communication link between an on-board unit (OBU) of a vehicle entering the target intersection and a roadside edge computing device based on a geo-fencing trigger; an information receiving module 13 for receiving the real-time traffic data and K pieces of real-time vehicle state information of K driving vehicles by the roadside edge computing device, where K≥2 and K is a positive integer; a compensation module 14 for compensating the K pieces of real-time vehicle state information with the real-time traffic data to obtain K pieces of compensated vehicle state information if the single-dimensional risk detection results of the real-time traffic data and the K pieces of real-time vehicle state information by the roadside edge computing device are all 0; a traffic model construction module 15 for constructing a global dynamic traffic model based on the K pieces of compensated vehicle state information; a trajectory conflict prediction module 16 for predicting a trajectory conflict and locating a potential collision risk based on the global dynamic traffic model; and a broadcasting module 17 for broadcasting the potential collision risk to the K driving vehicles through the V2X communication link according to a risk avoidance priority.

[0073] Further, the system is also used to implement the following functions:

[0074] According to the road condition information of the target intersection, a traffic-related risk feature is matched and called; the roadside edge computing device traverses the real-time traffic data using the traffic-related risk feature to output a first risk detection result; according to the location information of the target intersection, a driving speed limit feature is called online; the roadside edge computing device traverses the K pieces of real-time vehicle state information using the driving speed limit feature to output a second risk detection result; if the first risk detection result and the second risk detection result are both 0, the K pieces of real-time vehicle state information are compensated with the real-time traffic data to obtain K pieces of compensated vehicle state information.

[0075] Further, the system is also used to implement the following functions:

[0076] The traffic-related risk feature includes a vehicle safety distance constraint and a vehicle driving direction constraint; the real-time traffic data is traversed using the vehicle safety distance constraint and the vehicle driving direction constraint to locate a real-time risk vehicle as the first risk detection result; if the first risk detection result is 1, a real-time V2X communication link of the real-time risk vehicle is called according to the K pieces of real-time vehicle state information; and a risk warning is sent to the real-time risk vehicle using the real-time V2X communication link.

[0077] Further, the system is also used to implement the following functions:

[0078] A preset alarm execution window; after the delay of sending the risk alarm to the real-time risk vehicle meets the alarm execution window, the roadside edge computing device receives the updated traffic data returned by the roadside sensing network coverage; the roadside edge computing device uses the traffic-related risk features to make a related risk persistence judgment on the updated traffic data.

[0079] Further, the system is also used to implement the following functions:

[0080] K real-time vehicle positions are called from the K real-time vehicle state information; the K real-time vehicle positions are offset compensated by using the real-time traffic data to obtain K compensated vehicle positions; K sets of detection state data are called from the real-time traffic data according to the K compensated vehicle positions, wherein the detection state data includes real-time vehicle speed and vehicle heading angle; the K sets of detection state data are mapped and weighted fused to the K real-time vehicle state information, the K compensated vehicle positions are mapped and covered to the K real-time vehicle state information, and the K compensated vehicle state information is output.

[0081] Further, the system is also used to implement the following functions:

[0082] The K compensated vehicle state information is associated with the lane topology to obtain global bound vehicle state information; multi-target trajectory tracking is performed on the K driving vehicles according to the global bound vehicle state information, and K real-time trajectory sequences are output; the K real-time trajectory sequences are fitted to the spatial modeling of the target intersection to obtain the global dynamic traffic model.

[0083] Further, the system is also used to implement the following functions:

[0084] The conflict area distribution is predefined according to the lane connection topology of the target intersection; trajectory prediction is performed along the K real-time trajectory sequences by using the CTRA model to obtain K predicted time sequence trajectories; after spatiotemporal overlapping of the K predicted time sequence trajectories, a plurality of conflict trajectories falling into the conflict area distribution are extracted; a plurality of driving vehicles are located according to the mapping relationship of the plurality of conflict trajectories and the K predicted time sequence trajectories; after calling a plurality of vehicle IDs of the plurality of driving vehicles, the plurality of vehicle IDs are packaged as the potential collision risk output.

[0085] It should be noted that the above-mentioned embodiment sequences of the present application are merely for description only, but not for representing the advantages and disadvantages of the embodiments. And the above-mentioned embodiments of the present specification have been described. The processes depicted in the accompanying drawings do not necessarily require the specific order and continuous order shown to achieve the desired results. In some embodiments, multi-task processing and parallel processing are also possible or can be advantageous.

[0086] The above only describes the preferred embodiments of the present application, and does not limit the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

[0087] The specification and drawings are merely exemplary of the present application, and any and all modifications, variations, combinations or equivalents that are within the scope of the present application should be included. Obviously, those skilled in the art can make various modifications and variations to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application belong to the scope of the present application and its equivalents, the present application is intended to include these modifications and variations.

Claims

1. A vehicle state perception method in a vehicle infrastructure integration environment, characterized in that, The method comprises: The roadside sensor network collects real-time traffic data of the target intersection through a millimeter wave radar; Based on the geographic fence, a temporary V2X communication link between the OBU of the vehicle entering the target intersection and the roadside edge computing device is established; The roadside edge computing device receives the real-time traffic data and K real-time vehicle state information of K driving vehicles, wherein K≥2, K is a positive integer; If the single-dimensional risk detection result of the roadside edge computing device on the real-time traffic data and K real-time vehicle state information is all 0, the K real-time vehicle state information is compensated by the real-time traffic data to obtain K compensated vehicle state information; Based on the K compensated vehicle state information, a global dynamic traffic model is constructed; Based on the global dynamic traffic model, trajectory conflict prediction is performed to locate potential collision risks; According to the risk avoidance priority, the potential collision risks are broadcast to the K driving vehicles through the V2X communication link; Based on the K compensated vehicle state information, a global dynamic traffic model is constructed, the method comprising: The K compensated vehicle state information is associated with the lane topology to obtain global bound vehicle state information; According to the global bound vehicle state information, multi-target trajectory tracking is performed on the K driving vehicles to output K real-time trajectory sequences; The K real-time trajectory sequences are fitted to the spatial modeling of the target intersection to obtain the global dynamic traffic model. 2.The vehicle state perception method in a vehicle infrastructure integration environment according to claim 1, wherein, The method further comprises: According to the road condition information of the target intersection, traffic-related risk features are matched and called; The roadside edge computing device uses the traffic-related risk features to traverse the real-time traffic data to output a first risk detection result; According to the location information of the target intersection, driving speed features are called online; The roadside edge computing device uses the driving speed features to traverse the K real-time vehicle state information to output a second risk detection result; If the first risk detection result and the second risk detection result are all 0, the K real-time vehicle state information is compensated by the real-time traffic data to obtain K compensated vehicle state information. 3.The vehicle state perception method in a vehicle infrastructure integration environment according to claim 2, wherein, The method further comprises: The traffic-related risk features include vehicle safety distance constraints and vehicle driving direction constraints; The vehicle safety distance constraints and the vehicle driving direction constraints are used to traverse the real-time traffic data to locate real-time risk vehicles as the first risk detection result; If the first risk detection result is 1, real-time V2X communication links of the real-time risk vehicles are called according to the K real-time vehicle state information; The real-time V2X communication links are used to send risk warnings to the real-time risk vehicles. 4.The vehicle state perception method in a vehicle infrastructure integration environment according to claim 3, wherein, After sending risk warnings to the real-time risk vehicles through the real-time V2X communication link, the method comprises: A preset warning execution window is set; After the time delay of sending the risk warning to the real-time risk vehicle satisfies the warning execution window, the roadside edge computing device receives the updated traffic data returned by the roadside sensor network coverage; The roadside edge computing device uses the traffic-related risk features to make associated risk survival judgments on the updated traffic data. 5.The vehicle state perception method in a vehicle infrastructure integrated environment according to claim 2, wherein, The K real-time vehicle state information is compensated by using the real-time traffic data, and K compensated vehicle state information is obtained, and the method comprises: K real-time vehicle positions are called from the K real-time vehicle state information; The K real-time vehicle positions are offset compensated by using the real-time traffic data for checking, and K compensated vehicle positions are obtained; K sets of detection state data are called from the real-time traffic data according to the K compensated vehicle positions, wherein the detection state data comprises real-time vehicle speed and vehicle heading angle; The K sets of detection state data are mapped and weighted fused to the K real-time vehicle state information, the K real-time vehicle state information is mapped and covered by using the K compensated vehicle positions, and the K compensated vehicle state information is output. 6.The vehicle state perception method in a vehicle infrastructure integration environment according to claim 1, wherein, Trajectory conflict prediction is performed based on the global dynamic traffic model, and potential collision risks are located, and the method comprises: A conflict area distribution is predefined according to a lane connection topology of the target intersection; Trajectory prediction is performed along the K real-time trajectory sequences by using a CTRA model, and K predicted time sequence trajectories are obtained; After the K predicted time sequence trajectories are spatiotemporally overlapped, a plurality of conflict trajectories falling into the conflict area distribution are extracted; A plurality of running vehicles are located according to a mapping relationship of the plurality of conflict trajectories and the K predicted time sequence trajectories; After a plurality of vehicle IDs of the plurality of running vehicles are called, the plurality of vehicle IDs are packaged as the potential collision risks and output.

7. A vehicle state perception system in a cooperative vehicle infrastructure environment, characterized by The system is used for performing the vehicle state perception method in the vehicle infrastructure cooperation environment as claimed in any one of claims 1-6, and the system comprises: A traffic data acquisition module is used for acquiring real-time traffic data of a target intersection by a roadside sensing network through a millimeter wave radar; A communication link establishment module is used for establishing a temporary V2X communication link between a vehicle-mounted OBU of a vehicle entering the target intersection and a roadside edge computing device based on a geographic fence trigger; An information receiving module is used for receiving, by the roadside edge computing device, the real-time traffic data and K real-time vehicle state information of K running vehicles, wherein K≥2, and K is a positive integer; A compensation module is used for compensating the K real-time vehicle state information by using the real-time traffic data if single-dimensional risk detection results of the real-time traffic data and the K real-time vehicle state information are all 0, and K compensated vehicle state information is obtained; A traffic model construction module is used for constructing a global dynamic traffic model based on the K compensated vehicle state information; A trajectory conflict prediction module is used for performing trajectory conflict prediction based on the global dynamic traffic model, and locating potential collision risks; A broadcasting module is used for broadcasting the potential collision risks to the K running vehicles through a V2X communication link according to risk avoidance priorities.

Images