Traffic risk identification method and device based on vehicle infrastructure cooperation

Abstract

The invention discloses a traffic risk identification method and device based on vehicle-road cooperation. The method comprises the steps that vehicle position data and first moving object data are collected in real time at the current visual angle of a vehicle through vehicle-mounted virtual equipment; respectively generating first head-up display information according to the vehicle position data and the first moving object data; when second moving object data sent from any data source end is received, updating the first head-up display information according to the second moving object data, and generating second head-up display information; and risk prediction is performed according to the second head-up display information, a risk prediction result is generated, and a corresponding prompt module is triggered to perform output, so that safety risk early warning is quickly and accurately performed in combination with a vehicle-road cooperation mode.

Description

Technical Field

[0001] The present invention relates to the field of risk identification technology, and in particular to a traffic risk identification method and device based on vehicle-road collaboration. Background Art

[0002] Against the backdrop of the deep integration of digitalization and intelligence, information technology is reshaping the ecology of various industries in all aspects. Virtual Reality (VR), Augmented Reality (AR), and Mixed Reality (MR) technologies are booming, constantly expanding the boundaries of human perception and interaction. From a global perspective, R&D investment in related technologies continues to rise. In the next few years, the market size of VR / AR / MR will grow exponentially, showing great application potential in many fields such as entertainment, education, medical care, and industry.

[0003] In terms of VR technology, although the immersive virtual environment it creates has made significant progress, the high cost of equipment, poor wearing comfort and relative lack of content ecology have seriously hindered its large-scale popularization. For AR technology, although it has initially possessed the attributes of mobile phones, it is limited by physical conditions and is highly dependent on existing mobile products such as mobile phones. The location information interaction method mostly relies on devices that match the mobile terminal (such as Bluetooth or data cable connection), which inevitably leads to large interaction delays and is difficult to meet application scenarios with strict real-time requirements, such as industrial remote real-time collaboration and AR navigation. Even though AR-HUD (Augmented Reality Head - Up Display) combines AR with vehicle-mounted hardware, it has solved the portability, mobility and some hardware physical limitations of AR to a certain extent, but the stability of the overall system and the richness of information presentation still need to be improved.

[0004] In the existing technology, whether it is L2, L3 or L4 level, the single-vehicle intelligence (Autonomous Driving, AD) still relies on the network to determine the interactive method of location information during the actual implementation process. It is unable to give an effective presentation at the feedback end, and it is difficult to quickly and accurately issue safety risk warnings. Summary of the invention

[0005] The present invention provides a traffic risk identification method and device based on vehicle-road collaboration, which solves the technical problem that the existing single-vehicle intelligent technology still relies on the network to determine the interaction mode of location information during actual implementation, cannot give effective presentation at the feedback end, and is difficult to quickly and accurately provide safety risk warnings.

[0006] The present invention provides a traffic risk identification method based on vehicle-road cooperation, which is applied to an in-vehicle virtual device. The method includes:

[0007] Collect vehicle position data and first moving object data in real time from the current perspective of the vehicle;

[0008] Generate first head-up display information respectively according to the vehicle position data and the first moving object data;

[0009] When receiving second moving object data sent from any data source end, update the first head-up display information according to the second moving object data to generate second head-up display information;

[0010] Perform risk prediction according to the second head-up display information, generate a risk prediction result and trigger a corresponding prompt module to output.

[0011] Optionally, the data source end includes a wireless metropolitan area network, a GPS system or a smart city traffic system; the data source end is specifically used for:

[0012] Periodically collect initial position data corresponding to moving objects that meet the preset identification type;

[0013] Perform self-calibration on the initial position data, generate second moving object data and send it to the in-vehicle virtual device.

[0014] Optionally, the first moving object data includes a first moving object coordinate and first moving object motion data; the generating first head-up display information respectively according to the vehicle position data and the first moving object data includes:

[0015] Convert the vehicle position data and the first moving object coordinate to an in-vehicle virtual coordinate system to obtain a first virtual coordinate;

[0016] Calculate a first position vector according to the first virtual coordinate combined with the first moving object motion data;

[0017] After performing real-time correction on the first position vector, load map data corresponding to the current perspective of the vehicle to construct a virtual scene;

[0018] Render the corrected first position vector to the virtual scene to generate first head-up display information.

[0019] Optionally, the second moving object data includes a second moving object coordinate and second moving object motion data; the updating the first head-up display information according to the second moving object data to generate second head-up display information when receiving second moving object data sent from any data source end includes:

[0020] Convert the coordinates of the second moving object to the vehicle-mounted virtual coordinate system to obtain second virtual coordinates;

[0021] Calculate a second position vector according to the second virtual coordinates in combination with the motion data of the second moving object;

[0022] Render the second position vector to the virtual scene to update the first head-up display information and generate second head-up display information.

[0023] Optionally, the risk prediction is performed according to the second head-up display information, a risk prediction result is generated, and a corresponding prompt module is triggered for output, including:

[0024] Filter risk objects in a preset condition area from the second head-up display information;

[0025] Perform risk prediction separately according to the type of the condition area to which the risk object belongs to generate a risk prediction result;

[0026] Trigger a corresponding prompt module for output according to the risk prediction result.

[0027] The present invention also provides a traffic risk identification device based on vehicle-road cooperation, which is applied to a vehicle-mounted virtual device. The device includes:

[0028] A data acquisition module, configured to acquire vehicle position data and first moving object data in real time from the current perspective of the vehicle;

[0029] A display generation module, configured to generate first head-up display information according to the vehicle position data and the first moving object data respectively;

[0030] A display update module, configured to update the first head-up display information according to the second moving object data to generate second head-up display information when receiving second moving object data sent from any data source end;

[0031] A risk prediction and prompt module, configured to perform risk prediction according to the second head-up display information, generate a risk prediction result, and trigger a corresponding prompt module for output.

[0032] Optionally, the data source end includes a wireless metropolitan area network, a GPS system or a smart city traffic system; the data source end is specifically configured to:

[0033] Periodically acquire initial position data corresponding to moving objects that meet a preset identification type;

[0034] Perform self-calibration on the initial position data, generate second moving object data, and send it to the vehicle-mounted virtual device.

[0035] Optionally, the first moving object data includes the coordinates of the first moving object and the motion data of the first moving object; specifically, the display generation module is configured to:

[0036] Convert the vehicle position data and the coordinates of the first moving object to the in-vehicle virtual coordinate system to obtain the first virtual coordinates;

[0037] Calculate the first position vector according to the first virtual coordinates in combination with the motion data of the first moving object;

[0038] After performing real-time correction on the first position vector, load the map data corresponding to the current perspective of the vehicle to construct a virtual scene;

[0039] Render the corrected first position vector to the virtual scene to generate the first head-up display information.

[0040] Optionally, the second moving object data includes the coordinates of the second moving object and the motion data of the second moving object; specifically, the display update module is configured to:

[0041] Convert the coordinates of the second moving object to the in-vehicle virtual coordinate system to obtain the second virtual coordinates;

[0042] Calculate the second position vector according to the second virtual coordinates in combination with the motion data of the second moving object;

[0043] Render the second position vector to the virtual scene to update the first head-up display information and generate the second head-up display information.

[0044] Optionally, the risk prediction and prompt module is specifically configured to:

[0045] Screen for risk objects in the preset condition area from the second head-up display information;

[0046] Perform risk prediction separately according to the type of the condition area to which the risk object belongs to generate a risk prediction result;

[0047] Trigger the corresponding prompt module for output according to the risk prediction result.

[0048] It can be seen from the above technical solutions that the present invention has the following advantages:

[0049] The present invention collects vehicle position data and first moving object data in real time at the current perspective of the vehicle through an in-vehicle virtual device; generates first head-up display information based on the vehicle position data and the first moving object data respectively; when receiving second moving object data sent from any data source end, updates the first head-up display information according to the second moving object data to generate second head-up display information; performs risk prediction according to the second head-up display information, generates a risk prediction result and triggers a corresponding prompt module for output, so as to quickly and accurately perform safety risk warning in combination with the vehicle-road cooperation method. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required to be used in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0051] Figure 1 It is a flowchart of the steps of a traffic risk identification method based on vehicle-road cooperation provided by an embodiment of the present invention;

[0052] Figure 2 It is a schematic diagram of data acquisition of a moving object in an embodiment of the present invention;

[0053] Figure 3 It is a structural block diagram of a traffic risk identification device based on vehicle-road cooperation provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0054] The embodiments of the present invention provide a traffic risk identification method and device based on vehicle-road cooperation, which are used to solve the technical problem that the existing single-vehicle intelligent technology still relies on the network to determine the interaction mode of position information in the actual implementation process, cannot give an effective presentation at the feedback end, and is difficult to quickly and accurately perform safety risk warning.

[0055] In order to make the invention purpose, features, and advantages of the present invention more obvious and understandable, the following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the drawings in the embodiments of the present invention. Obviously, the embodiments described below are only some embodiments of the present invention, rather than all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the protection scope of the present invention.

[0056] Please refer to Figure 1 , Figure 1 It is a flowchart of the steps of a traffic risk identification method based on vehicle-road cooperation provided by an embodiment of the present invention.

[0057] A traffic risk identification method based on vehicle-road cooperation provided by the present invention is applied to an in-vehicle virtual device and includes:

[0058] Step 101, collect vehicle position data and first moving object data in real time from the current perspective of the vehicle;

[0059] The vehicle position data refers to the position data in the in-vehicle virtual coordinate system established with the vehicle as the center, such as the longitude and latitude coordinates, driving speed, driving direction, etc. of the vehicle. Herein, the vehicle refers to a vehicle equipped with an intelligent driving system and a virtual device such as an in-vehicle AR-HUD.

[0060] The first moving object data refers to the position data and motion data of all movable obstacles within the current perspective of the vehicle, including but not limited to other moving vehicles, pedestrians, cyclists, and roadblocks, etc.

[0061] In an embodiment of the present invention, the vehicle obtains the first moving object data within the current perspective of the vehicle through a built-in camera, and simultaneously obtains the vehicle position data, providing a data basis for the generation of subsequent head-up display information.

[0062] Step 102, generate first head-up display information according to the vehicle position data and the first moving object data respectively;

[0063] The first head-up display information refers to the information generated after data processing and visualization conversion based on the collected vehicle position data and first moving object data, and can be presented on the vehicle head-up display (HUD) device. The position, motion trend, etc. of the moving object are visualized through specific graphics and identifiers.

[0064] After obtaining the vehicle position data and the first moving object data, the processor converts information such as the longitude and latitude coordinates of the vehicle into display content adapted to the vehicle driving scenario, such as displaying the name of the current road where the vehicle is located, the distance to the destination, etc. on the AR-HUD device. For the first moving object data, the processor visualizes information such as the position and motion direction of the first moving object, and on devices such as in-vehicle AR-HUD devices or AR glasses that require the user to identify, the position and motion trend of the first moving object are displayed with specific graphics or identifiers. For example, different colored arrows are used to represent the motion direction of the moving object, and the length of the arrow represents the speed magnitude, etc.

[0065] In an example of the present invention, the first moving object data includes the first moving object coordinates and the first moving object motion data; Step 102 may include the following sub-steps:

[0066] Convert the vehicle position data and the first moving object coordinates to the in-vehicle virtual coordinate system to obtain the first virtual coordinates;

[0067] Calculate the first position vector according to the first virtual coordinate and the motion data of the first moving object;

[0068] After real-time correction of the first position vector, load the map data corresponding to the current perspective of the vehicle to construct a virtual scene;

[0069] Render the corrected first position vector to the virtual scene to generate the first head-up display information.

[0070] As Figure 2 shown, Figure 2 It shows a schematic diagram of data acquisition of a moving object in an embodiment of the present invention. It shows a schematic diagram of data acquisition of the current perspective of the vehicle and the perspectives of other devices with traffic lights as the data source end.

[0071] In an embodiment of the present invention, the collected vehicle position data and the coordinates of the first moving object are converted into a vehicle-mounted virtual coordinate system. The vehicle-mounted virtual coordinate system takes the vehicle itself as the origin, the vehicle driving direction as the positive direction of the X axis, perpendicular to the vehicle driving direction to the right as the positive direction of the Y axis, and perpendicular upward as the positive direction of the Z axis. Through a preset coordinate conversion algorithm, the first vehicle data and the coordinates of the first moving object are converted from the geographic coordinate system or the image coordinate system into coordinates in the vehicle-mounted virtual coordinate system to obtain the first virtual coordinate.

[0072] For example, for the vehicle longitude and latitude coordinates in the geographic coordinate system, first convert them into plane rectangular coordinates through projection transformation, and then map them into the vehicle-mounted virtual coordinate system according to the vehicle attitude information; for the coordinates of the first moving object in the image coordinate system, combine the internal and external parameters of the camera and convert them into the vehicle-mounted virtual coordinate system through methods such as perspective transformation.

[0073] After obtaining the first virtual coordinate, determine the direction and length of the vector starting from the coordinates of the first moving object in the vehicle-mounted virtual coordinate system and combining the motion direction and speed of the first moving object. For example, if the coordinates of the first moving object in the vehicle-mounted virtual coordinate system are (X1, Y1, Z1), the motion direction is θ, and the speed is v, then the direction of the first position vector is θ, and the length is the displacement of v in unit time. The specific numerical representation of the first position vector is obtained through vector operations.

[0074] Considering that during the actual operation, the calculation result of the position vector may be affected by sensor errors, vehicle vibrations, environmental interferences, etc., after obtaining the first position vector, the Kalman filter algorithm can be used to filter the first position vector. By using the state equation and observation equation of the system, the predicted value and observed value of the position vector are fused to obtain a more accurate estimated value of the position vector. At the same time, attitude compensation is performed on the position vector in combination with the real-time attitude information of the vehicle (such as pitch angle, roll angle, heading angle) to ensure the accuracy of the position vector. Meanwhile, the map data corresponding to the current perspective of the vehicle is loaded. The map data is stored in the storage module on the vehicle side and includes road information, landmark information, etc. The processor extracts the map information within the current perspective range of the vehicle from the map data according to the current position and driving direction of the vehicle, and constructs a virtual scene in combination with the in-vehicle virtual coordinate system. The virtual scene shows the road environment around the vehicle and the approximate positional relationship of the first moving object in the form of three-dimensional graphics.

[0075] After the virtual scene is constructed, the corrected first position vector is rendered into the virtual scene. Specifically, the position vector can be converted into a visual graph or identifier through a graphics rendering engine. For example, an arrow is used to represent the movement direction and speed of the moving object, and icons of different shapes and colors are used to represent different types of moving objects, etc. At the same time, relevant information of the vehicle itself, such as vehicle speed, driving direction, etc., is superimposed on the virtual scene, and finally the first head-up display information is generated and displayed on the AR-HUD device, enabling the driver to intuitively understand the positions and motion states of the moving objects around the vehicle.

[0076] Step 103, when receiving the second moving object data sent from any data source end, update the first head-up display information according to the second moving object data to generate the second head-up display information;

[0077] The data source end refers to the data end connected to this vehicle through vehicle networking communication or other data connection methods, including but not limited to wireless metropolitan area network, GPS system, or smart city traffic system, etc.

[0078] The second head-up display information refers to the head-up display information that covers the visual dead angle or blind area after supplementing and correcting the first head-up display information.

[0079] When this vehicle receives the second moving object data sent from any data source end, the display content of the second moving object in the first head-up display information is corrected and supplemented according to the second moving object data, such as more accurate position, speed and other information, to generate the second head-up display information.

[0080] In an example of the present invention, the second moving object data includes the second moving object coordinates and the second moving object motion data; Step 103 may include the following sub-steps:

[0081] Convert the coordinates of the second moving object to the vehicle-mounted virtual coordinate system to obtain the second virtual coordinates;

[0082] Calculate the second position vector according to the second virtual coordinates combined with the motion data of the second moving object;

[0083] Render the second position vector to the virtual scene to update the first head-up display information and generate the second head-up display information.

[0084] The generation processes of the second virtual coordinates and the second position vector can refer to the generation processes of the foregoing first virtual coordinates and the first position vector, which will not be elaborated here.

[0085] In the embodiment of the present invention, after obtaining the second position vector, since the virtual scene has been constructed in the foregoing steps, in order to improve the position accuracy and information richness and avoid visual blind spots, the second position vector is rendered to the virtual scene, and the position and motion state of the second moving object are presented in specific graphics, identifiers, etc. For example, different types of moving objects are distinguished by graphics of different colors and sizes, and their motion directions and speed changes are shown by dynamic arrows.

[0086] In addition, if there is information that is repeated or in conflict with the first moving object, it is corrected based on the data of the second moving object. Finally, the second head-up display information is generated and displayed on the AR-HUD device, enabling the driver to obtain more accurate and comprehensive information about the moving objects around the vehicle. The global urban coordinate system and the vehicle local coordinate system are non-contact calibrated through roadside beacons (non-V2X devices) to improve the interactive presentation of the position information of the in-vehicle AR-HUD.

[0087] In an example of the present invention, the data source end includes a wireless metropolitan area network, a GPS system, or a smart city transportation system; the data source end is specifically used for:

[0088] Periodically collect the initial position data corresponding to the moving objects that meet the preset recognition types;

[0089] Perform self-calibration on the initial position data to generate the second moving object data and send it to the vehicle-mounted virtual device.

[0090] In this embodiment, the data source end periodically collects the initial position data corresponding to the moving objects that meet the preset recognition types from their respective perspectives. The preset recognition types can be vehicles, pedestrians, cyclists, etc. After obtaining the initial position data, various calibration algorithms can be used to perform self-calibration on the initial position data to eliminate data errors, and then generate the second moving object data and send it to the vehicle-mounted virtual device.

[0091] Taking a wireless metropolitan area network as an example, cameras deployed by the roadside use computer vision algorithms to identify targets such as vehicles and pedestrians in the picture, obtain their position information in the image coordinate system, and combine the position and attitude parameters of the cameras to convert them into initial position data in the geographic coordinate system; the GPS system directly obtains the longitude and latitude coordinates of moving objects through satellite positioning as the initial position data; intelligent traffic cameras and radar devices in the smart city transportation system determine the positions of moving objects by detecting the reflection signals or image features of moving objects and generate initial position data. The initial position data collected by GPS is calibrated by using the method of fusing the inertial navigation system (INS) and GPS data. The high-precision positioning information of INS in a short time is used to make up for the errors caused by factors such as GPS signal occlusion and multipath effect; for the data collected by cameras in the wireless metropolitan area network and the smart city transportation system, an image calibration algorithm based on machine learning is adopted. The model is trained through a large number of sample data to perform geometric correction and photometric calibration on the images collected by the cameras, and improve the accuracy of the position data of moving objects in the images. At the same time, the data source end also calibrates the timestamps of the collected data to ensure the time consistency of the data and avoid position data deviation caused by time errors. After calibration, the calibrated initial position data is integrated with the corresponding motion data of the moving object (such as speed, acceleration, motion direction, etc., which can be calculated from the continuously collected position data) to generate the second moving object data. The wireless metropolitan area network uses its network communication protocol to package the data and transmit it to the in-vehicle virtual device; the GPS system sends the processed data to the in-vehicle virtual device through a dedicated data transmission channel; the smart city transportation system directly transmits the data to the in-vehicle virtual device through V2X (vehicle-to-everything) communication technology.

[0092] In addition, the sending of V2X can be avoided through the standardized access protocol of Internet of Things data, and at the same time, centimeter-level virtual-real space alignment can be achieved based on the hybrid positioning architecture of urban digital twin and vehicle-mounted SLAM.

[0093] Step 104, perform risk prediction according to the second head-up display information, generate a risk prediction result, and trigger the corresponding prompt module for output.

[0094] In this embodiment, according to the data such as the position relationship, relative speed, and motion direction between the vehicle and each moving object in the second head-up display information, combined with a preset risk assessment model for calculation and analysis. When the risk prediction result is generated, the processor triggers the corresponding prompt module for output.

[0095] For example, if it is calculated that the predicted driving trajectories of a vehicle and a moving object intersect within a certain period in the future and the approaching speed exceeds a preset threshold, it is determined that there is a collision risk, and a corresponding risk prediction result is generated according to the degree of risk. After the risk prediction result is generated, the processor triggers the corresponding prompt module for output. If the risk level is low, the processor triggers the sound alarm to emit a relatively soft prompt sound, and at the same time highlights the moving object with risk on the AR-HUD device in yellow; if the risk level is high, the processor not only triggers the sound alarm to emit a rapid alarm sound, but also activates the vibration device to make the vehicle seat vibrate, and at the same time displays the risk moving object on the AR-HUD device with a red flashing mark to attract the driver's high attention.

[0096] In an example of the present invention, step 104 may include the following sub-steps:

[0097] Screen for risk objects in the preset condition area from the second head-up display information;

[0098] Perform risk prediction separately according to the type of the condition area to which the risk object belongs, and generate a risk prediction result;

[0099] Trigger the corresponding prompt module for output according to the risk prediction result.

[0100] The preset condition area can be divided according to the vehicle driving scenario or safety requirements, and may include but not be limited to visual blind spots, visual dead corners, or other risk areas such as the area in front of the vehicle at close range, the side blind spot area, the rear close area, etc.

[0101] In this embodiment, while generating the second head-up display information, the position coordinates of each moving object in the second head-up display information (based on the vehicle-mounted virtual coordinate system) can be parsed to determine whether it is within a preset condition area. If the coordinates of the moving object satisfy the coordinate range of a certain preset condition area, the moving object is screened as a risk object. Then, corresponding risk prediction rules are respectively matched according to the type of the risk object condition area for risk prediction to generate a risk prediction result. Different prompting strategies are called according to the risk prediction result for risk prompting output. For example, when the risk level is high risk, the AR-HUD module is triggered to prominently display the risk object with a flashing red graphic, and text prompts such as "High risk, avoid immediately" are marked beside the graphic; the seat prompting module is triggered to vibrate at a high frequency and high intensity; meanwhile, the sound alarm emits a sharp and hasty alarm sound, reminding the driver of the danger from multiple dimensions of vision, touch, and hearing. When the risk level is medium risk, the AR-HUD module displays the risk object with a yellow graphic and marks the prompt text; the seat prompting module vibrates at a lower frequency and intensity; the sound alarm emits a relatively gentle prompt sound. When the risk level is low risk, the AR-HUD module only slightly flashes and displays the risk object with a light blue graphic, and the sound alarm emits a short prompt sound, reminding the driver to pay attention to the risk object, providing more guarantees for the driving of disabled persons, and improving driving safety through vehicle-road cooperation at the same time.

[0102] Specifically, if the moving object to which the second position vector risk object belongs is in the visual blind area, probability modeling is performed on the moving object of the risk object. Combining historical data, surrounding environment information, and the characteristics of the moving object, the corresponding obstacle types (such as vehicles, pedestrians, large objects, etc.) and risk probabilities are determined and used as the risk prediction result. Then, according to the obstacle type and risk probability, they are matched to the corresponding risk level, and the AR-HUD module is triggered to output semi-transparent prompt information. The prompt information is superimposed on the driver's field of vision in the form of graphics, text, etc., reminding the driver to pay attention to the potential risks existing in the visual blind area; if the moving object to which the second position vector of the risk object belongs is in the visual dead angle area, the seat prompting module is triggered to vibrate according to the position and real-time motion state of the moving object of the risk object. Through vibrations of different frequencies and intensities, the urgency of the risk is transmitted to the driver, enabling the driver to perceive the risks existing in the visual dead angle area through touch; if the moving object to which the second position vector of the risk object does not belong to the visual dead angle area or the visual blind area, a predicted trajectory band is generated by combining the above vehicle trajectory prediction, obstacle trajectory prediction, and collision risk prediction results based on Newtonian mechanics, and the AR-HUD module is triggered to display the trajectory. The predicted trajectory band is presented in the AR-HUD module in different colors and line styles, intuitively showing the motion trend of the risk object and helping the driver make driving decisions in advance.

[0103] In addition, the phase holographic waveguide display technology can be used to improve the field of view limitation of traditional HUD. Synchronize the position calibration with the position change in the physical environment, for example, presenting virtual characters similar to those described by AI on the display screen (AR-HUD).

[0104] In this embodiment, the prediction of the obstacle trajectory and collision risk can be achieved in the following ways:

[0105] 1) The vehicle trajectory can be measured by describing the vehicle motion based on Newtonian mechanics. For example:

[0106] x = v cos(a + b)

[0107] y = v sin(a + b)

[0108] z = v / L

[0109] a and b are slip angles, the front wheel steering angle, L is the wheelbase, v is the vehicle speed. By combining the above equations with the current speed, angle and other parameters of the vehicle, the possible position of the vehicle in the future period of time can be calculated, providing a basis for subsequent risk assessment.

[0110] 2) Assume that the obstacle motion mode is uniform motion, uniformly accelerated motion or random motion, and use the Kalman filter or particle filter algorithm to predict its future trajectory. At the same time, introduce a deep learning model (such as LSTM or Transformer) to perform more accurate obstacle trajectory prediction for blind spots (detected by millimeter wave radar or lidar). Specifically, use millimeter wave radar or lidar to obtain the distance, speed and other data of the obstacles in the blind spot area, input these data into the deep learning model for training and prediction, obtain the more accurate future motion trajectory of the obstacles, and determine the warning signal node in advance.

[0111] 3) Adopt the collision risk prediction TTC (Time to Collision) algorithm to calculate the collision time when the two vehicles maintain the current motion state. The calculation formula is:

[0112] TTC = Δy / Δx, Δx ≠ 0

[0113] By calculating the TTC value, evaluate the risk degree of collision between the vehicle and the obstacle. The higher the TTC value, the higher the collision risk.

[0114] In an embodiment of the present invention, vehicle position data and first moving object data are collected in real time at the current perspective of the vehicle through an in-vehicle virtual device; according to the vehicle position data and the first moving object data, first head-up display information is respectively generated; when second moving object data sent from any data source end is received, the first head-up display information is updated according to the second moving object data to generate second head-up display information; risk prediction is performed according to the second head-up display information to generate a risk prediction result and trigger a corresponding prompt module to output, so as to quickly and accurately perform safety risk warning in a vehicle-road collaborative manner.

[0115] Please refer to Figure 3 , Figure 3 which shows a structural block diagram of a traffic risk identification device provided by an embodiment of the present invention based on vehicle-road collaboration.

[0116] The present invention also provides a traffic risk identification device based on vehicle-road collaboration, which is applied to an in-vehicle virtual device. The device includes:

[0117] A data acquisition module 301, configured to collect vehicle position data and first moving object data in real time at the current perspective of the vehicle;

[0118] A display generation module 302, configured to respectively generate first head-up display information according to the vehicle position data and the first moving object data;

[0119] A display update module 303, configured to update the first head-up display information according to the second moving object data to generate second head-up display information when second moving object data sent from any data source end is received;

[0120] A risk prediction and prompt module 304, configured to perform risk prediction according to the second head-up display information, generate a risk prediction result and trigger a corresponding prompt module to output.

[0121] Optionally, the data source end includes a wireless metropolitan area network, a GPS system or a smart city traffic system; the data source end is specifically configured to:

[0122] Periodically collect initial position data corresponding to moving objects that meet a preset identification type;

[0123] Perform self-calibration on the initial position data, generate second moving object data and send it to the in-vehicle virtual device.

[0124] Optionally, the first moving object data includes first moving object coordinates and first moving object motion data; the display generation module 302 is specifically configured to:

[0125] Convert the vehicle position data and the first moving object coordinates to an in-vehicle virtual coordinate system to obtain first virtual coordinates;

[0126] Calculate the first position vector according to the first virtual coordinate and the motion data of the first moving object;

[0127] After real-time correction of the first position vector, load the map data corresponding to the current vehicle view and construct a virtual scene;

[0128] Render the corrected first position vector to the virtual scene to generate the first head-up display information.

[0129] Optionally, the second moving object data includes the second moving object coordinate and the second moving object motion data; the display update module 303 is specifically configured to:

[0130] Convert the second moving object coordinate to the in-vehicle virtual coordinate system to obtain the second virtual coordinate;

[0131] Calculate the second position vector according to the second virtual coordinate and the second moving object motion data;

[0132] Render the second position vector to the virtual scene to update the first head-up display information and generate the second head-up display information.

[0133] Optionally, the risk prediction and prompt module 304 is specifically configured to:

[0134] Screen out risk objects in the preset condition area from the second head-up display information;

[0135] Perform risk prediction respectively according to the type of the condition area to which the risk object belongs to generate a risk prediction result;

[0136] Trigger the corresponding prompt module to output according to the risk prediction result.

[0137] Those skilled in the art can clearly understand that for the convenience and simplicity of description, the specific working processes of the above-described devices and modules can refer to the corresponding processes in the foregoing method embodiments and will not be elaborated herein.

[0138] In several embodiments provided by the present invention, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the modules is only a logical function division, and there may be other division methods in actual implementation. For example, multiple modules or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point, the displayed or discussed mutual coupling or direct coupling or communication connection can be through some interfaces, and the indirect coupling or communication connection of the devices or modules can be in an electrical, mechanical or other form.

[0139] The module described as a separation component may or may not be physically separated. The component shown as a module may or may not be a physical module, that is, it may be located in one place or may be distributed to multiple network modules. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0140] As described above, the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements for some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A traffic risk identification method based on vehicle-road cooperation, characterized in that, Applied to an in-vehicle virtual device, the method includes: Collecting vehicle position data and first moving object data in real time from the current perspective of the vehicle; Generating first head-up display information respectively according to the vehicle position data and the first moving object data; When receiving second moving object data sent from any data source end, updating the first head-up display information according to the second moving object data to generate second head-up display information; Performing risk prediction according to the second head-up display information, generating a risk prediction result and triggering a corresponding prompt module for output.

2. The method according to claim 1, wherein The data source end includes a wireless metropolitan area network, a GPS system or a smart city traffic system; specifically, the data source end is used for: Periodically collecting initial position data corresponding to moving objects that meet a preset recognition type; Performing self-calibration on the initial position data, generating second moving object data and sending it to the in-vehicle virtual device.

3. The method according to claim 1, wherein The first moving object data includes first moving object coordinates and first moving object motion data; the generating first head-up display information respectively according to the vehicle position data and the first moving object data includes: Converting the vehicle position data and the first moving object coordinates to an in-vehicle virtual coordinate system to obtain first virtual coordinates; Calculating a first position vector according to the first virtual coordinates in combination with the first moving object motion data; After performing real-time correction on the first position vector, loading map data corresponding to the current perspective of the vehicle to construct a virtual scene; Rendering the corrected first position vector to the virtual scene to generate first head-up display information.

4. The method according to claim 3, wherein The second moving object data includes second moving object coordinates and second moving object motion data; the updating the first head-up display information according to the second moving object data to generate second head-up display information when receiving second moving object data sent from any data source end includes: Converting the second moving object coordinates to the in-vehicle virtual coordinate system to obtain second virtual coordinates; Calculating a second position vector according to the second virtual coordinates in combination with the second moving object motion data; Rendering the second position vector to the virtual scene to update the first head-up display information and generate second head-up display information.

5. The method according to claim 1, wherein The performing risk prediction according to the second head-up display information, generating a risk prediction result and triggering a corresponding prompt module for output includes: Filtering risk objects in a preset condition area from the second head-up display information; Performing risk prediction respectively according to the type of the condition area to which the risk objects belong to generate a risk prediction result; Triggering a corresponding prompt module for output according to the risk prediction result.

6. A traffic risk identification device based on vehicle-road cooperation, characterized in that Applied to an in-vehicle virtual device, the device includes: A data collection module, configured to collect vehicle position data and first moving object data in real time from the current perspective of the vehicle; A display generation module, configured to generate first head-up display information respectively according to the vehicle position data and the first moving object data; A display update module, configured to update the first head-up display information according to the second moving object data when receiving the second moving object data sent from any data source end, and generate second head-up display information; A risk prediction and prompt module, configured to perform risk prediction according to the second head-up display information, generate a risk prediction result, and trigger a corresponding prompt module to output; 7. The device according to claim 6, characterized in that, The data source end includes a wireless metropolitan area network, a GPS system, or a smart city traffic system; specifically, the data source end is used for: Periodically collecting initial position data corresponding to moving objects that meet a preset recognition type; Performing self-calibration on the initial position data, generating second moving object data, and sending it to the in-vehicle virtual device; 8. The device according to claim 6, characterized in that The first moving object data includes a first moving object coordinate and first moving object motion data; specifically, the display generation module is used for: Converting the vehicle position data and the first moving object coordinate to an in-vehicle virtual coordinate system to obtain a first virtual coordinate; Calculating a first position vector according to the first virtual coordinate in combination with the first moving object motion data; After performing real-time correction on the first position vector, loading map data corresponding to the current perspective of the vehicle to construct a virtual scene; Rendering the corrected first position vector to the virtual scene to generate first head-up display information; 9. The device according to claim 7, characterized in that, The second moving object data includes a second moving object coordinate and second moving object motion data; specifically, the display update module is used for: Converting the second moving object coordinate to the in-vehicle virtual coordinate system to obtain a second virtual coordinate; Calculating a second position vector according to the second virtual coordinate in combination with the second moving object motion data; Rendering the second position vector to the virtual scene to update the first head-up display information and generate second head-up display information; 10. The device according to claim 6, characterized in that, Specifically, the risk prediction and prompt module is used for: Screening risk objects in a preset condition area from the second head-up display information; Performing risk prediction respectively according to the type of the condition area to which the risk object belongs to generate a risk prediction result; Triggering a corresponding prompt module to output according to the risk prediction result;

Images