A collision warning system for an intelligent vehicle

Abstract

This invention discloses a collision avoidance warning system for intelligent vehicles, comprising: an interaction module for interacting with a device in a target scene to acquire scene environment data; the scene environment data includes multiple collision nodes and collision markers corresponding to each collision node; a data acquisition module for real-time acquisition of driving data and driving environment data of the intelligent vehicle when entering the target scene; a data analysis module for obtaining the current collision node of the intelligent vehicle based on the driving data and driving environment data, combined with the scene environment data; and a collision warning module for real-time acquisition of collision avoidance warning information about the collision markers of the current collision node and issuing a warning when the intelligent vehicle is within the warning range of the current collision node. This invention solves the technical problem of limited collision avoidance warning functionality of intelligent vehicles in complex scenarios in existing technologies.

Description

Technical Field

[0001] This invention relates to the field of intelligent vehicle driver assistance technology, specifically a collision avoidance warning system for intelligent vehicles. Background Technology

[0002] With the rapid development of the automotive industry, the number of intelligent vehicles on the road continues to increase, making travel more convenient for the public. Intelligent vehicles have a significant advantage in assisted autonomous driving, using sensors installed on the vehicle to collect data and perform multiple assisted autonomous driving functions such as lane departure warning, forward collision warning, and cruise control.

[0003] Currently, most intelligent vehicle collision avoidance warning systems rely on high-precision sensor devices to collect environmental data and complex algorithms to identify and determine the existence of collision risks, making them suitable for well-lit and open roads. However, in complex scenarios with poor lighting and numerous non-standard obstacles, such as garages in older residential areas, intelligent vehicle sensors may malfunction, GPS signals may be lost, and obstacles may be difficult to identify. This often makes it difficult for intelligent vehicle collision avoidance warning functions to achieve the desired results.

[0004] This indicates that existing technologies have limitations in the collision avoidance warning function of intelligent vehicles in complex scenarios. Summary of the Invention

[0005] The purpose of this invention is to provide a collision avoidance warning system for intelligent vehicles, which solves the technical problem that the collision avoidance warning function of intelligent vehicles is limited in complex scenarios in the prior art.

[0006] This invention provides a collision avoidance warning system for intelligent vehicles, the system comprising:

[0007] An interaction module is used to interact with the device to be interacted with in the target scene and obtain scene environment data of the target scene; the scene environment data includes multiple collision nodes and collision markers corresponding to each collision node;

[0008] The data acquisition module is used to collect real-time driving data and driving environment data of the intelligent vehicle when entering the target scene;

[0009] The data analysis module is used to obtain the current collision node of the intelligent vehicle based on driving data and driving environment data, combined with scene environment data;

[0010] The collision warning module is used to acquire and issue a warning in real time about the collision warning information of the collision markers at the current collision node when the intelligent vehicle is within the warning range of the current collision node.

[0011] Furthermore, the interaction module interacts with the device to be interacted with in the target scene to obtain scene environment data of the target scene, including:

[0012] When the interaction module placed on the smart car is within the preset interaction range of the device to be interacted with in the target scene and the device to be interacted with recognizes the smart car's intention to enter the target scene, the device to be interacted with provides the interaction module with at least one preset interaction method.

[0013] The interaction module obtains scene environment data of the target scene by interacting with at least one preset interaction method.

[0014] Furthermore, the scene environment data also includes scene end nodes and their corresponding end node markers; when the data acquisition module enters the target scene, it collects the intelligent vehicle's driving data and driving environment data in real time, including:

[0015] After the interaction module acquires the scene environment data of the target scene, the data acquisition module collects the driving environment data of the intelligent vehicle in real time; the driving environment data includes vehicle radar data and vehicle camera data;

[0016] Based on vehicle radar data and vehicle camera data, information on obstacles in front of the intelligent vehicle is obtained.

[0017] Based on obstacle information in the area in front of the intelligent vehicle, identify and acquire at least one target end node marker;

[0018] Based on at least one target end node marker, confirm the time node when the intelligent vehicle enters the target scene and the target scene end node;

[0019] After confirming the time node or target scene endpoint when the intelligent vehicle enters the target scene, the data acquisition module begins to collect driving data; the driving data includes vehicle speed data, steering wheel rotation data, and acceleration data.

[0020] Furthermore, the obstacle information includes several obstacle edges, shape data of each obstacle edge, and proportional data between any two obstacle edges; based on the obstacle information in the area in front of the intelligent vehicle, at least one target end node marker is identified and acquired, including:

[0021] Obtain the feature edge data of each end node marker; the feature edge data includes the feature edge, the size data of each feature edge, the shape data, and the ratio data between any two related feature edges;

[0022] Based on the feature edge data of each end node marker, several obstacle edges in the obstacle information are filtered to obtain several valid obstacle edges;

[0023] Based on a number of valid obstacle edges, multiple combinations of edges to be matched are obtained; each combination of edges to be matched includes at least two valid obstacle edges.

[0024] Each edge combination to be matched is matched with each end node marker to obtain at least one target edge combination and the target end node marker corresponding to each target edge combination.

[0025] Furthermore, based on at least one target end-node marker, the time point at which the intelligent vehicle enters the target scene and the target scene end-node are confirmed, including:

[0026] Constructing a spatial rectangular coordinate system based on intelligent vehicles;

[0027] Based on the feature edges of the target end node markers and the effective obstacle edges of the corresponding matching target edge combinations, the coordinate data of the target end node markers in the spatial rectangular coordinate system are confirmed.

[0028] Based on the coordinate data of multiple target end node markers in a spatial rectangular coordinate system, the position data of the intelligent vehicle in the target scene is confirmed;

[0029] If the location data of the current intelligent vehicle in the target scene is within a preset range of a scene end node, the scene end node is confirmed as the target scene end node, and the time is marked as a time node.

[0030] Furthermore, the data analysis module uses driving data and driving environment data, combined with scene environment data, to obtain the current collision node where the intelligent vehicle is located, including:

[0031] Based on the size data and driving data of intelligent vehicles, a virtual bounding box of the intelligent vehicle is constructed and the driving trajectory and vehicle pose of the virtual bounding box in the target scene are obtained.

[0032] Based on driving environment data, information on obstacles located in front of the intelligent vehicle is identified and obtained.

[0033] Based on the vehicle's virtual bounding box, driving trajectory, vehicle pose, and obstacle information in the target scene, it can determine in real time whether there is at least one collision marker in the area in front of the intelligent vehicle.

[0034] If so, based on at least one collision marker, identify the collision node corresponding to the collision marker, as well as other collision markers corresponding to the collision node.

[0035] Furthermore, based on the vehicle's virtual bounding box trajectory, vehicle pose, and obstacle information in the target scene, it is determined in real time whether there is at least one collision marker in the area in front of the intelligent vehicle, including:

[0036] Based on the vehicle's driving trajectory and vehicle pose in the target scene using the vehicle's virtual bounding box, the initial position and initial orientation of the intelligent vehicle in the target scene are obtained.

[0037] Based on the initial position and initial orientation of the intelligent vehicle in the target scenario, several collision markers to be confirmed are obtained;

[0038] Based on obstacle information, several collision markers to be confirmed are verified one by one.

[0039] When at least one collision marker to be confirmed passes verification, it is determined that there is at least one collision marker in the area in front of the intelligent vehicle.

[0040] Furthermore, based on obstacle information, several collision markers to be confirmed are verified one by one, including:

[0041] Acquire the feature edge data of the collision marker to be confirmed;

[0042] Based on the feature edge data of the collision marker to be confirmed, determine whether there is a preset number of associated obstacle edges in the obstacle information;

[0043] If so, the collision marker verification must be confirmed.

[0044] Furthermore, when the intelligent vehicle is within the warning range of the current collision point, the collision warning module acquires real-time collision avoidance warning information about the collision markers at the current collision point and issues a warning, including:

[0045] Obtain the safe distances of all collision markers at the current collision node;

[0046] Based on the driving trajectory and vehicle pose of the vehicle's virtual bounding box in the target scene, determine whether there is an overlap between the vehicle's virtual bounding box and the warning range of the current collision node;

[0047] If so, the intelligent vehicle enters the warning range of the current collision point;

[0048] The actual distance between the vehicle's virtual border and each collision marker is obtained in real time, and all collision markers at the current collision node are divided into risky collision markers and non-risky collision markers; when the actual distance between the vehicle's virtual border and a collision marker is less than the safe distance of the collision marker, the collision marker is a risky collision marker;

[0049] Obtain the name, risk type, and position of the risk collision marker relative to the vehicle's virtual border;

[0050] Based on the name of the risk collision marker, the risk type, and its position relative to the vehicle's virtual border, collision avoidance warning information is obtained and warnings are issued according to preset warning methods.

[0051] Furthermore, the risk types include frontal collisions and side collisions.

[0052] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0053] In this invention, an interaction module quickly acquires scene environment data of the target scenario, serving as a reference for the intelligent vehicle's environment after entering the target scenario. A data acquisition module obtains driving data and driving environment data of the intelligent vehicle in the target scenario, which is then used by a data analysis module to analyze the scene environment data and determine the current collision node of the intelligent vehicle. A collision warning module provides collision avoidance warnings to multiple collision markers corresponding to the current collision node. This improves the accuracy of obstacle location and identification for the intelligent vehicle in complex target scenarios and solves the technical problem of limited collision avoidance warning functionality for intelligent vehicles in complex scenarios in existing technologies. Attached Figure Description

[0054] Figure 1 This is a schematic diagram of a collision avoidance warning system for an intelligent vehicle according to the present invention. Detailed Implementation

[0055] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, 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 some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0056] like Figure 1 As shown, a collision avoidance warning system for an intelligent vehicle includes:

[0057] An interaction module is used to interact with the device to be interacted with in the target scene and obtain scene environment data of the target scene; the scene environment data includes multiple collision nodes and collision markers corresponding to each collision node;

[0058] The data acquisition module is used to collect real-time driving data and driving environment data of the intelligent vehicle when entering the target scene;

[0059] The data analysis module is used to obtain the current collision node of the intelligent vehicle based on driving data and driving environment data, combined with scene environment data;

[0060] The collision warning module is used to acquire and issue a warning in real time about the collision warning information of the collision markers at the current collision node when the intelligent vehicle is within the warning range of the current collision node.

[0061] The specific implementation process of this embodiment includes:

[0062] In this embodiment, the interaction module quickly acquires scene environment data of the target scenario, which serves as a reference for the scene environment after the intelligent vehicle enters the target scenario. The data acquisition module acquires driving data and driving environment data of the intelligent vehicle in the target scenario, which are then used by the data analysis module to analyze the scene environment data and determine the current collision node of the intelligent vehicle. The collision warning module provides collision avoidance warnings to multiple collision markers corresponding to the current collision node. This improves the accuracy of obstacle location and identification for the intelligent vehicle in complex target scenarios and solves the technical problem of limited collision avoidance warning functions of intelligent vehicles in complex scenarios in existing technologies.

[0063] In this embodiment, the interaction module interacts with the device to be interacted with in the target scene to obtain scene environment data of the target scene, including:

[0064] S11: When the interaction module placed on the smart car is within the preset interaction range of the device to be interacted with in the target scene and the device to be interacted with recognizes the intention information of the smart car to enter the target scene, the device to be interacted with provides at least one preset interaction method to the interaction module.

[0065] In this embodiment, the devices to be interacted with in the target scene include an image acquisition device and an image recognition device.

[0066] The image acquisition device is set at the entrance of the target scene to acquire images of the license plates of vehicles in front of the intelligent vehicle; the image recognition device is used to recognize the license plate images in front of the vehicle to obtain license plate information.

[0067] Image recognition and image acquisition devices work together to identify the intention of intelligent vehicles to enter target scenes.

[0068] The interactive device also includes one or more of the following: a QR code dynamic display device, a Bluetooth interactive device, and an NFC interactive device, used to provide at least one preset interactive path to interact with the interactive module on the smart car.

[0069] S12: The interaction module interacts with the target scene to obtain scene environment data by means of at least one preset interaction method.

[0070] In this embodiment, the interaction module includes one or more of a QR code scanning device, a Bluetooth interaction device, and an NFC interaction device. The interaction module interacts with the device to be interacted with to obtain scene environment data of the target scene, thereby improving the efficiency of intelligent vehicles in obtaining scene environment data of the target scene.

[0071] In this embodiment, the scene environment data of the target scene includes a pre-constructed scene plan view and a scene 3D view. Multiple collision nodes are pre-marked in the scene plan view and scene 3D view based on expert experience and collision history data, with each collision node corresponding to at least one collision marker. Collision markers include fire hydrants, pillars, ground locks, and blocks.

[0072] In this embodiment, the scene environment data also includes scene end nodes and corresponding end node markers; when the data acquisition module enters the target scene, it collects the intelligent vehicle's driving data and driving environment data in real time, including:

[0073] S21: After the interaction module obtains the scene environment data of the target scene, the data acquisition module collects the driving environment data of the intelligent vehicle in real time; the driving environment data includes vehicle radar data and vehicle camera data;

[0074] In this embodiment, the intelligent vehicle is equipped with vehicle radar and vehicle camera, which are used to collect vehicle radar data and vehicle camera data, respectively.

[0075] S22: Based on vehicle radar data and vehicle camera data, obtain information about obstacles in the area in front of the intelligent vehicle;

[0076] In this embodiment, the obstacle information includes several obstacle edges, shape data of each obstacle edge, and ratio data between any two obstacle edges;

[0077] S23: Based on obstacle information in the area in front of the intelligent vehicle, identify and acquire at least one target end node marker;

[0078] S231: Obtain the feature edge data of each end node marker; the feature edge data includes the feature edge, the size data of each feature edge, the shape data, and the ratio data between any two related feature edges;

[0079] In this embodiment, feature edge data of each end node marker is obtained based on scene environment data.

[0080] S232: Based on the feature edge data of each end node marker, filter several obstacle edges in the obstacle information to obtain several valid obstacle edges;

[0081] In this embodiment, due to various factors such as lighting, the vehicle-mounted radar and vehicle-mounted camera are limited in the target scene, making it difficult to ensure the integrity of the obstacle edges when collecting data on unknown obstacles.

[0082] Therefore, obstacle edges in the current obstacle information are filtered based on the feature edge data of known end node markers in the scene environment data. When the shape of an obstacle edge is similar to a feature edge, the obstacle edge is considered a valid obstacle edge.

[0083] In this embodiment, the similarity between the shape of an obstacle edge and a feature edge means that each part of the obstacle edge is proportionally reduced or enlarged as a feature edge.

[0084] S233: Based on several valid obstacle edges, obtain multiple combinations of edges to be matched; each combination of edges to be matched includes at least two valid obstacle edges;

[0085] In this embodiment, a combination of edges to be matched is obtained by randomly arranging and combining several effective obstacle edges.

[0086] It should be noted that the number of valid obstacle edges in each edge combination to be matched does not exceed 4.

[0087] S234: Match each edge combination to be matched with each end node marker to obtain at least one target edge combination and the target end node marker corresponding to each target edge combination.

[0088] In this embodiment, when matching the edge combination to be matched with each end node marker, the matching is performed in order of increasing number of effective obstacle edges.

[0089] When the preset number of edge combinations to be matched successfully match the end node markers, the matching process stops.

[0090] S24: Based on at least one target end node marker, confirm the time node when the intelligent vehicle enters the target scene and the target scene end node;

[0091] In this embodiment, the scene end node represents the first node that the intelligent vehicle may encounter after entering the target scene. In some embodiments, the scene end node includes a collision node.

[0092] S25: After confirming the time node or target scene end node when the intelligent vehicle enters the target scene, the data acquisition module begins to collect driving data; the driving data includes vehicle speed data, steering wheel rotation data and acceleration data.

[0093] In this embodiment, the data acquisition module begins collecting driving data after confirming the time point at which the intelligent vehicle enters the target scene or the target scene endpoint, ensuring the timeliness of the driving data and facilitating subsequent analysis by the data analysis module.

[0094] In this embodiment, based on at least one target end node marker, the time node at which the intelligent vehicle enters the target scene and the target scene end node are confirmed, including:

[0095] S241: Constructing a spatial rectangular coordinate system based on intelligent vehicles;

[0096] In this embodiment, the X and Y axes of the spatial index coordinate system are determined based on the plane where the chassis of the intelligent vehicle is located. The centerline along the length of the intelligent vehicle chassis is used as the Y-axis, the centerline along the width is used as the X-axis, the direction of vehicle travel is used as the positive direction of the Y-axis, and the direction to the right of the direction of vehicle travel is used as the positive direction of the X-axis. Based on the X and Y axes, the Z-axis of the spatial rectangular coordinate system is determined.

[0097] S242: Based on the feature edges of the target end node markers and the effective obstacle edges of the corresponding matching target edge combinations, confirm the coordinate data of the target end node markers in the spatial rectangular coordinate system;

[0098] In this embodiment, each end node marker surface is provided with several feature points, and each feature point is located on the corresponding feature edge; the position of the feature point on the effective obstacle edge is obtained according to the feature edge, and the spatial coordinates of the position in the camera coordinate system or radar coordinate system are calculated;

[0099] Then, based on the transformation relationship between the camera coordinate system or radar coordinate system and the spatial rectangular coordinate system of the intelligent vehicle, the coordinate data of the feature points in the spatial rectangular coordinate system of the intelligent vehicle are obtained.

[0100] The coordinate data of the target end node marker in the spatial rectangular coordinate system includes the coordinate data of at least one feature point.

[0101] S243: Based on the coordinate data of multiple target end node markers in a spatial rectangular coordinate system, confirm the position data of the intelligent vehicle in the target scene;

[0102] In this embodiment, each feature point is pre-set with corresponding BeiDou coordinates;

[0103] Based on pre-set BeiDou coordinates for multiple feature points and their coordinates in the intelligent vehicle's Cartesian coordinate system, coordinate transformation is used to obtain the BeiDou coordinates of the origin of the intelligent vehicle's Cartesian coordinate system within the target scene. This serves as the intelligent vehicle's position data within the target scene, preventing signal issues from causing the intelligent vehicle to be unable to locate itself.

[0104] S244: If the location data of the current intelligent vehicle in the target scene is within a preset range of a scene end node, the scene end node is confirmed as the target scene end node, and the time is marked as a time node.

[0105] In this embodiment, the data analysis module obtains the current collision node of the intelligent vehicle based on driving data and driving environment data, combined with scene environment data, including:

[0106] S31: Based on the size data and driving data of the intelligent vehicle, construct the virtual bounding box of the intelligent vehicle and obtain the driving trajectory and vehicle pose of the virtual bounding box in the target scene;

[0107] In this embodiment, the size data of the intelligent vehicle includes the length and width of the intelligent vehicle projected onto the horizontal plane, and the center of the vehicle's virtual frame coincides with the origin of the intelligent vehicle's spatial rectangular coordinate system.

[0108] After confirming the target scene endpoints, the location data of the intelligent vehicle is used as the starting point of the driving trajectory, and the vehicle's pose at the starting point is obtained. After the starting point of the driving trajectory, the driving trajectory within the target scene is obtained in real time based on the intelligent vehicle's speed data, steering wheel rotation data, and acceleration data. The driving trajectory acquisition process does not require the participation of a positioning module and can be used even when the positioning signal in the target scene is poor. The vehicle pose is obtained in real time based on the steering wheel rotation data.

[0109] S32: Based on driving environment data, identify and obtain information about obstacles in the area in front of the intelligent vehicle;

[0110] S33: Based on the vehicle's virtual bounding box's driving trajectory, vehicle pose, and obstacle information in the target scene, determine in real time whether there is at least one collision marker in the area in front of the intelligent vehicle; specifically as follows:

[0111] S331: Based on the vehicle's virtual bounding box trajectory and vehicle pose in the target scene, obtain the initial position and initial orientation of the intelligent vehicle in the target scene;

[0112] S332: Based on the initial position and initial orientation of the intelligent vehicle in the target scene, obtain several collision markers to be confirmed;

[0113] S333: Verify each of the collision markers to be confirmed based on obstacle information;

[0114] S334: When at least one collision marker to be confirmed passes verification, it is determined that there is at least one collision marker in the area in front of the intelligent vehicle.

[0115] S34: If so, based on at least one collision marker, identify the collision node corresponding to the collision marker, and other collision markers corresponding to the collision node.

[0116] In this embodiment, several collision markers to be confirmed are verified one by one based on obstacle information, including:

[0117] S3331: Obtain the feature edge data of the collision marker to be confirmed;

[0118] S3332: Based on the feature edge data of the collision marker to be confirmed, determine whether there is a preset number of associated obstacle edges in the obstacle information;

[0119] S3333: If so, the collision marker verification is pending confirmation.

[0120] In this embodiment, when the intelligent vehicle is within the warning range of the current collision node, the collision warning module acquires and issues a warning in real time regarding the collision avoidance warning information of the collision markers at the current collision node, including:

[0121] S41: Obtain the safe distance of all collision markers at the current collision node;

[0122] In this embodiment, the current collision node of the intelligent vehicle is identified by at least one collision marker, and all collision markers corresponding to the collision node are obtained, avoiding the situation where traditional sensor recognition methods cannot identify all obstacles for various reasons.

[0123] S42: Based on the driving trajectory and vehicle pose of the vehicle's virtual bounding box in the target scene, determine whether there is an overlap between the vehicle's virtual bounding box and the warning range of the current collision node;

[0124] S43: If so, the intelligent vehicle enters the warning range of the current collision node;

[0125] S44: Real-time acquisition of the actual distance between the vehicle's virtual border and each collision marker, classifying all collision markers at the current collision node into risky collision markers and non-risky collision markers; when the actual distance between the vehicle's virtual border and a collision marker is less than the safe distance of the collision marker, the collision marker is a risky collision marker;

[0126] S45: Obtain the name, risk type, and position of the risk collision marker relative to the vehicle's virtual border;

[0127] S46: Based on the name of the risk collision marker, the risk type, and its position relative to the vehicle's virtual border, obtain collision avoidance warning information and issue a warning according to a preset warning method.

[0128] In this embodiment, the risk types include frontal collisions and side collisions. The preset warning method includes voice warnings.

[0129] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0130] The above embodiments are only used to illustrate the technical methods of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical methods of the present invention without departing from the spirit and scope of the technical methods of the present invention.

Claims

1. A collision avoidance warning system for intelligent vehicles, characterized in that: The system includes: An interaction module is used to interact with the device to be interacted with in the target scene and obtain scene environment data of the target scene; the scene environment data includes multiple collision nodes and collision markers corresponding to each collision node; The data acquisition module is used to collect real-time driving data and driving environment data of the intelligent vehicle when entering the target scene; The data analysis module is used to obtain the current collision node of the intelligent vehicle based on driving data and driving environment data, combined with scene environment data; The collision warning module is used to acquire and issue a warning in real time about the collision warning information of the collision markers at the current collision node when the intelligent vehicle is within the warning range of the current collision node.

2. The collision avoidance warning system for intelligent vehicles as described in claim 1, characterized in that: The interaction module interacts with the device to be interacted with in the target scene to obtain scene environment data of the target scene, including: When the interaction module placed on the smart car is within the preset interaction range of the device to be interacted with in the target scene and the device to be interacted with recognizes the smart car's intention to enter the target scene, the device to be interacted with provides the interaction module with at least one preset interaction method. The interaction module obtains scene environment data of the target scene by interacting with at least one preset interaction method.

3. The collision avoidance warning system for intelligent vehicles as described in claim 1, characterized in that: Scene environment data also includes scene end nodes and the corresponding end node markers; When the data acquisition module enters the target scene, it collects real-time driving data and driving environment data of the intelligent vehicle, including: After the interaction module acquires the scene environment data of the target scene, the data acquisition module collects the driving environment data of the intelligent vehicle in real time; the driving environment data includes vehicle radar data and vehicle camera data; Based on vehicle radar data and vehicle camera data, information on obstacles in front of the intelligent vehicle is obtained. Based on obstacle information in the area in front of the intelligent vehicle, identify and acquire at least one target end node marker; Based on at least one target end node marker, confirm the time node when the intelligent vehicle enters the target scene and the target scene end node; After confirming the time node or target scene endpoint when the intelligent vehicle enters the target scene, the data acquisition module begins to collect driving data; the driving data includes vehicle speed data, steering wheel rotation data, and acceleration data.

4. The collision avoidance warning system for intelligent vehicles as described in claim 3, characterized in that: Obstacle information includes several obstacle edges, shape data of each obstacle edge, and ratio data between any two obstacle edges; Based on obstacle information in the area in front of the intelligent vehicle, at least one target end node marker is identified and acquired, including: Obtain the feature edge data of each end node marker; Feature edge data includes the feature edges, the size data of each feature edge, the shape data, and the ratio data between any two related feature edges; Based on the feature edge data of each end node marker, several obstacle edges in the obstacle information are filtered to obtain several valid obstacle edges; Based on a number of valid obstacle edges, multiple combinations of edges to be matched are obtained; each combination of edges to be matched includes at least two valid obstacle edges. Each edge combination to be matched is matched with each end node marker to obtain at least one target edge combination and the target end node marker corresponding to each target edge combination.

5. The collision avoidance warning system for intelligent vehicles as described in claim 4, characterized in that: Based on at least one target endpoint marker, the time point at which the intelligent vehicle enters the target scene and the target scene endpoints are confirmed, including: Constructing a spatial rectangular coordinate system based on intelligent vehicles; Based on the feature edges of the target end node markers and the effective obstacle edges of the corresponding matching target edge combinations, the coordinate data of the target end node markers in the spatial rectangular coordinate system are confirmed. Based on the coordinate data of multiple target end node markers in a spatial rectangular coordinate system, the position data of the intelligent vehicle in the target scene is confirmed; If the location data of the current intelligent vehicle in the target scene is within a preset range of a scene end node, the scene end node is confirmed as the target scene end node, and the time is marked as a time node.

6. The collision avoidance warning system for an intelligent vehicle as described in claim 4, characterized in that: The data analysis module uses driving data and driving environment data, combined with scene environment data, to obtain the current collision node of the intelligent vehicle, including: Based on the size data and driving data of intelligent vehicles, a virtual bounding box of the intelligent vehicle is constructed and the driving trajectory and vehicle pose of the virtual bounding box in the target scene are obtained. Based on driving environment data, information on obstacles located in front of the intelligent vehicle is identified and obtained. Based on the vehicle's virtual bounding box, driving trajectory, vehicle pose, and obstacle information in the target scene, it can determine in real time whether there is at least one collision marker in the area in front of the intelligent vehicle. If so, based on at least one collision marker, identify the collision node corresponding to the collision marker, as well as other collision markers corresponding to the collision node.

7. The collision avoidance warning system for intelligent vehicles as described in claim 6, characterized in that: Based on the vehicle's virtual bounding box trajectory, vehicle pose, and obstacle information in the target scene, the system determines in real time whether there is at least one collision marker in the area in front of the intelligent vehicle, including: Based on the vehicle's driving trajectory and vehicle pose in the target scene using the vehicle's virtual bounding box, the initial position and initial orientation of the intelligent vehicle in the target scene are obtained. Based on the initial position and initial orientation of the intelligent vehicle in the target scenario, several collision markers to be confirmed are obtained; Based on obstacle information, several collision markers to be confirmed are verified one by one. When at least one collision marker to be confirmed passes verification, it is determined that there is at least one collision marker in the area in front of the intelligent vehicle.

8. The collision avoidance warning system for intelligent vehicles as described in claim 7, characterized in that: Based on obstacle information, several collision markers to be confirmed are verified one by one, including: Acquire the feature edge data of the collision marker to be confirmed; Based on the feature edge data of the collision marker to be confirmed, determine whether there is a preset number of associated obstacle edges in the obstacle information; If so, the collision marker verification must be confirmed.

9. The collision avoidance warning system for an intelligent vehicle as described in claim 6, characterized in that: When the intelligent vehicle is within the warning range of the current collision point, the collision warning module acquires real-time collision avoidance warning information about the collision markers at the current collision point and issues a warning, including: Obtain the safe distances of all collision markers at the current collision node; Based on the driving trajectory and vehicle pose of the vehicle's virtual bounding box in the target scene, determine whether there is an overlap between the vehicle's virtual bounding box and the warning range of the current collision node; If so, the intelligent vehicle enters the warning range of the current collision point; The actual distance between the vehicle's virtual border and each collision marker is obtained in real time, and all collision markers at the current collision node are divided into risky collision markers and non-risky collision markers; when the actual distance between the vehicle's virtual border and a collision marker is less than the safe distance of the collision marker, the collision marker is a risky collision marker; Obtain the name, risk type, and position of the risk collision marker relative to the vehicle's virtual border; Based on the name of the risk collision marker, the risk type, and its position relative to the vehicle's virtual border, collision avoidance warning information is obtained and warnings are issued according to preset warning methods.

10. The collision avoidance warning system for an intelligent vehicle as described in claim 9, characterized in that: The risk types include frontal collisions and side collisions.

Images