Millimeter wave radar forward anti-collision early warning method and system based on lane recognition

Abstract

The invention relates to a millimeter wave radar forward anti-collision early warning method and system based on lane recognition, and relates to the technical field of intelligent driving. The method comprises the following steps: acquiring point cloud information sent by a millimeter wave radar and driving data corresponding to a current vehicle; generating target information based on the point cloud information, wherein the target information is used for indicating at least one target in a millimeter wave radar detection range; generating a predicted collision result corresponding to the current vehicle and the target in combination with the target information and the motion data; and generating anti-collision early warning information based on the predicted collision result. According to the point cloud clustering tracking result, the lane, the speed and the pose information of the target are screened and judged, the distance between the target and the front vehicle and the collision time are accurately estimated, proper early warning measures are taken, and the reliability and the anti-interference capability of the result are further improved.

Description

Technical Field

[0001] The present invention relates to the field of intelligent driving, and particularly to a millimeter-wave radar forward collision warning method and system based on lane recognition. Background Art

[0002] With the rapid development of intelligent driving technology, millimeter-wave radar, as a sensor with all-weather and strong anti-interference capabilities, has been widely used in the environmental perception system of vehicles, especially playing an important role in forward collision warning (FCW).

[0003] Traditional millimeter-wave radar collision avoidance systems mainly rely on information such as the relative speed and distance of targets to judge potential collision risks.

[0004] However, in the actual driving environment, there is a certain misjudgment rate in judging whether there is a collision risk based solely on radar data, especially in complex traffic scenarios such as multi-lane, high-speed, and lane-changing situations. Summary of the Invention

[0005] The present invention relates to a millimeter-wave radar forward collision warning method and system based on lane recognition, which can improve the warning effect of forward collision avoidance in lanes. The technical solution is as follows:

[0006] On the one hand, a millimeter-wave radar forward collision warning method based on lane recognition is provided. The method is applied to a computer device in a millimeter-wave radar forward collision warning system based on lane recognition. The method includes:

[0007] Obtain the point cloud information sent by the millimeter-wave radar and the driving data corresponding to the current vehicle;

[0008] Generate target information based on the point cloud information, where the target information is used to indicate at least one target within the detection range of the millimeter-wave radar;

[0009] Combine the target information and the motion data to generate a predicted collision result corresponding to the current vehicle and the target;

[0010] Generate a collision avoidance warning information based on the predicted collision result.

[0011] In an optional embodiment, the receiving the point cloud information sent by the millimeter-wave radar includes:

[0012] Preset a working cycle for receiving point cloud information;

[0013] Receive the point cloud information based on the working cycle.

[0014] In an alternative embodiment, generating the target information based on the point cloud information includes:

[0015] Performing target data extraction on the point cloud information to obtain point cloud data;

[0016] Performing clustering fitting on the point cloud data to generate the target information, where the target information includes at least one of target quantity information, target size information, and target position information.

[0017] In an alternative embodiment, combining the target information and the motion data to generate a predicted collision result corresponding to the target for the current vehicle includes:

[0018] Determining a first driving trajectory of the current vehicle based on the motion data;

[0019] Determining a second driving trajectory corresponding to the target based on the target information;

[0020] Generating a predicted collision result corresponding to the target based on the first driving trajectory and the second driving trajectory.

[0021] In an alternative embodiment, determining the first driving trajectory of the current vehicle based on the motion data includes:

[0022] Establishing a vehicle coordinate system corresponding to the current vehicle;

[0023] Determining the first driving trajectory of the current vehicle within the vehicle coordinate system based on the motion data.

[0024] In an alternative embodiment, the motion data includes vehicle speed data and vehicle direction data.

[0025] In an alternative embodiment, the point cloud information includes lane data;

[0026] Determining the second driving trajectory corresponding to the target based on the target information includes:

[0027] Determining target lane data and target motion data corresponding to the target based on the target data, where the target lane data is used to indicate the current lane of the target, and the target motion data is used to indicate the current state of the target;

[0028] Determining the second driving trajectory based on the target lane data and the target motion data.

[0029] In an alternative embodiment, generating a predicted collision result corresponding to the target based on the first driving trajectory and the second driving trajectory includes:

[0030] Determine the lane consistency between the current vehicle and the target based on the first driving trajectory and the second driving trajectory;

[0031] In response to the target and the current vehicle being in the same lane, determine the predicted collision time data between the current vehicle and the target;

[0032] In response to the predicted collision time data being less than the collision time threshold, determine that the predicted collision result is a collision risk.

[0033] In an alternative embodiment, the method further includes:

[0034] In response to the target and the current vehicle not being in the same lane, determine the existence result of the trajectory intersection point based on the first driving trajectory and the second driving trajectory;

[0035] In response to the existence result of the trajectory intersection point indicating that there is an intersection point between the first driving trajectory and the second driving trajectory, determine the arrival time difference data based on the first driving trajectory and the second driving trajectory;

[0036] In response to the arrival time difference data being less than the arrival time difference threshold, determine the arrival time data of the current vehicle reaching the intersection point;

[0037] In response to the arrival time data being less than the arrival time threshold, determine that the predicted collision result is a collision risk.

[0038] On the other hand, a millimeter-wave radar forward collision warning system based on lane recognition is provided, and the system includes a millimeter-wave radar and a computer device;

[0039] The millimeter-wave radar is communicatively connected to the computer device;

[0040] The millimeter-wave radar is used to generate point cloud information and send it to the computer device;

[0041] The computer device is used to execute any one of the above-mentioned millimeter-wave radar forward collision warning methods based on lane recognition.

[0042] The beneficial effects brought by the technical solution provided by the present invention at least include:

[0043] According to the point cloud clustering tracking result, screen and judge the lane, speed and pose information of the target, accurately estimate the distance and collision time from the vehicle ahead, and make appropriate warning measures, further improving the reliability and anti-interference ability of the result. Description of the Drawings

[0044] To more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the accompanying drawings required for the description of the embodiments. Obviously, the accompanying drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other accompanying drawings can be obtained based on these drawings.

[0045] Figure 1 The block diagram of a millimeter-wave radar forward collision avoidance warning system based on lane recognition provided by an exemplary embodiment of the present application is shown.

[0046] Figure 2 The flowchart of a millimeter-wave radar forward collision avoidance warning method based on lane recognition provided by an exemplary embodiment of the present application is shown.

[0047] Figure 3 The flowchart of another millimeter-wave radar forward collision avoidance warning method based on lane recognition provided by an exemplary embodiment of the present application is shown.

[0048] Figure 4 The schematic diagram of coordinate system establishment provided by an exemplary embodiment of the present application is shown. Detailed implementation manners

[0049] To make the objectives, technical solutions, and advantages of the present invention clearer, the following further describes the embodiments of the present invention in detail with reference to the accompanying drawings.

[0050] Figure 1 The block diagram of a millimeter-wave radar forward collision avoidance warning system based on lane recognition provided by an exemplary embodiment of the present application is shown. The system includes a millimeter-wave radar 110 and a computer device 120, and the millimeter-wave radar 110 is communicatively connected to the computer device 120.

[0051] In the embodiments of the present application, the millimeter-wave radar is a radar product attached to the vehicle, and the detection range of the millimeter-wave radar is the front of the vehicle. The millimeter-wave radar is used to generate point cloud information and send it to the computer device for the computer device to generate corresponding processing results and further generate a visualization result. It should be noted that the millimeter-wave radar selected in the embodiments of the present application is a 4D millimeter-wave radar. According to the actual application scenario, different types of millimeter-wave radars can be selected to adapt to the corresponding application environment.

[0052] In the application scenario of front collision avoidance, optionally, the millimeter-wave radar can be located at the central axis position of the vehicle. In one example, the millimeter-wave radar is located on the front bumper of the vehicle and at the central axis position of the vehicle.

[0053] Figure 2The figure shows a schematic flowchart of a millimeter-wave radar forward collision avoidance warning method based on lane recognition provided by an exemplary embodiment of the present application. Taking the application of this method in a computer device within a system as shown in Figure 1 as an example for illustration, the method includes:

[0054] Step 201: Obtain the point cloud information sent by the millimeter-wave radar and the driving data corresponding to the current vehicle.

[0055] In the embodiment of the present application, the driving data of the vehicle is the data generated during the driving process of the current vehicle, which is used to indicate information such as the speed and pose of the vehicle.

[0056] The point cloud information is the information obtained by the millimeter-wave radar periodically or based on the data sending request of the computing device.

[0057] Step 202: Generate target information based on the point cloud information.

[0058] In the embodiment of the present application, the target information is used to indicate at least one target within the detection range of the millimeter-wave radar. Optionally, the number of targets can be at least two, which are used to indicate at least one target in the lane where the current vehicle is located and at least one target in the surrounding lanes of the current vehicle.

[0059] Step 203: Combine the target information and the motion data to generate a predicted collision result corresponding to the current vehicle and the target.

[0060] In the embodiment of the present application, the predicted collision result can be implemented in the form of Time To Collision (TTC) data, and the TTC data is implemented in numerical form.

[0061] Step 204: Generate a collision avoidance warning information based on the predicted collision result.

[0062] In the embodiment of the present application, the collision avoidance warning information can be implemented as a visual warning signal. After the computer device generates this signal, it is prompted through the sound-emitting device or display device in the current vehicle.

[0063] In summary, the method provided by the embodiment of the present application screens and judges the lane, speed, and pose information of the target according to the point cloud clustering tracking result, accurately estimates the distance and collision time with the vehicle ahead, and takes appropriate warning measures, further improving the reliability and anti-interference ability of the result.

[0064] Figure 3 The figure shows a schematic flowchart of another millimeter-wave radar forward collision avoidance warning method based on lane recognition provided by an exemplary embodiment of the present application. Taking the application of this method in a system as shown in Figure 1Taking the computer device in the system shown as an example, the method includes:

[0065] Step 301, preset the working cycle for receiving point cloud information.

[0066] Step 302, receive point cloud information based on the working cycle.

[0067] Steps 301 to 302 illustrate a way of receiving point cloud information. Optionally, in the embodiments of the present application, a way of obtaining information periodically is selected to obtain point cloud information.

[0068] Step 303, extract target data from the point cloud information to obtain point cloud data.

[0069] Step 304, perform clustering fitting on the point cloud data to generate target information.

[0070] Optionally, in the embodiments of the present application, the target information includes at least one of target quantity information, target size information, and target position information. In the embodiments of the present application, the number of targets corresponding to the target information is at least two.

[0071] Step 305, establish a vehicle coordinate system corresponding to the current vehicle.

[0072] Step 306, based on the motion data, determine the first driving trajectory of the current vehicle in the vehicle coordinate system.

[0073] Steps 305 to 306 are the establishment process of the vehicle coordinate system. Please refer to Figure 4 , combined with the position of the radar 410, an XYZ coordinate system with the radar as the origin of the coordinate axis can be established. In the coordinate system, there is a corresponding target 420.

[0074] After determining the vehicle position, the driving trajectory of the target vehicle can be determined according to the coordinate system and the transmission of multiple point cloud data.

[0075] In the embodiments of the present application, the motion data includes vehicle speed data and vehicle direction data.

[0076] Step 307, determine the lane consistency between the current vehicle and the target based on the first driving trajectory and the second driving trajectory.

[0077] In the embodiments of the present application, the target data includes target lane data corresponding to the target and target motion data, the target lane data is used to indicate the current lane of the target, and the target motion data is used to indicate the current state of the target. In this case, the lane consistency between the current vehicle and the target can be determined.

[0078] In one example, if the target meets any of the following conditions, the target is in the vehicle's lane:

[0079] 1) The target centroid x is within the current lane;

[0080] 2) The target straddles the current lane (the x - coordinate of the right border of the target x_max >= the center line of the lane and the x - coordinate of the left border of the target x_min <= the center line of the lane);

[0081] 3) The centroid x is in the left lane, but the right border of the target exceeds the left - lane line by a certain distance (such as 1 meter) and falls within the current lane;

[0082] 4) The centroid x is in the right lane, but the left border of the target exceeds the right - lane line by a certain distance (such as 1 meter) and falls within the current lane;

[0083] 5) The previous frame of the target was in the current lane and the left (or right) border of this frame is in the current lane.

[0084] Step 308, in response to the target and the current vehicle being in the same lane, determine the predicted time - to - collision data between the current vehicle and the target.

[0085] In the embodiments of the present application, after determining that the target and the vehicle are in the same lane, the TTC data is determined based on the relative distance and relative speed.

[0086] Step 309, in response to the predicted time - to - collision data being less than the collision time threshold, determine that the predicted collision result is a collision risk.

[0087] In the embodiments of the present application, this threshold can be determined according to the actual distance from the current vehicle.

[0088] Step 310, in response to the target and the current vehicle not being in the same lane, determine the existence result of the trajectory intersection based on the first driving trajectory and the second driving trajectory.

[0089] Step 311, in response to the trajectory intersection existence result indicating that there is an intersection between the first driving trajectory and the second driving trajectory, determine the time - difference - to - arrival data based on the first driving trajectory and the second driving trajectory.

[0090] Step 312, in response to the time - difference - to - arrival data being less than the time - difference - to - arrival threshold, determine the arrival - time data of the current vehicle at the intersection.

[0091] Step 313, in response to the arrival - time data being less than the arrival - time threshold, determine that the predicted collision result is a collision risk.

[0092] Optionally, when the target and the current vehicle are not in the same lane, the driving trajectory of the target will be calculated. The driving trajectory of the target is converted into the XYZ coordinate system as shown in Formula 1 below:

[0093] Formula 1:

[0094] Where (x0, y0) are the initial coordinates, vec is the speed, and AngleHeading is the heading angle. Optionally, the straight-line equation can be rewritten as: y = y0 + (x - x0)·tan(AngleHeading).

[0095] In this case, the road curvature coefficient corresponding to the current vehicle is shown in Formula 2 below:

[0096] Formula 2:

[0097] Where ObjIntr_facA2 is the road curvature coefficient, w is the total speed, and v is the straight-line speed.

[0098] The formal trajectory of this lane is shown in Formula 3 below:

[0099] Formula 3:

[0100] Combining Formula 1 and Formula 3 to solve whether there are solutions for x and y. If there are solutions and x in the solutions is greater than 0, it means that the two trajectories have an intersection point, that is, there is a possibility of collision. In this case, further determine the time difference to reach the intersection point. If the time difference between the current vehicle and the target vehicle reaching the intersection point is less than the time difference threshold, it means that there may be a collision event. At this time, it is necessary to further determine the time for the current vehicle to reach the intersection point. If it is determined that the time for the current vehicle to reach this intersection point is short, it means that there is a situation where the driver cannot react to the collision event in time. At this time, it is necessary to display a warning message.

[0101] It should be noted that the above collision prediction process will detect all targets generated from the point cloud information.

[0102] Step 313, generate a collision avoidance warning message based on the predicted collision result.

[0103] In summary, the method provided by the embodiments of the present application fully combines lane line information and radar target information to implement a more accurate and robust collision warning strategy at the computer processing level. This method constructs a spatial geometric constraint based on lane recognition, intelligently judges whether the target is in the same lane as the vehicle itself, and further combines dynamic elements such as relative speed and acceleration to dynamically calculate the collision time, thereby effectively improving the warning accuracy rate, reducing the false alarm rate, and enhancing the practicability and reliability of the system.

[0104] That is, the method provided by the embodiments of the present application:

[0105] ① Since the sensor used is a millimeter-wave radar, compared with other sensors, it has very obvious external anti-interference advantages and is suitable for traffic scenarios with complex road environments;

[0106] ②Quickly derive the point cloud occupancy probability according to the radar equation, which has the advantages of low hardware requirements, short running time, and high result reliability;

[0107] ③Screen and judge the lane, speed and pose information of the target according to the point cloud clustering tracking results, accurately estimate the distance and collision time with the vehicle ahead, and take appropriate warning measures to further improve the reliability and anti-interference ability of the results.

[0108] The above are only optional embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A millimeter-wave radar forward collision avoidance warning method based on lane recognition, characterized in that: The method is applied to a millimeter-wave radar forward collision warning system based on lane recognition, wherein the system comprises a computer device and a millimeter-wave radar, wherein the millimeter-wave radar is communicatively connected with the computer device; The method comprises: Acquire the point cloud information sent by the millimeter-wave radar and the driving data corresponding to the current vehicle; Generate target information based on the point cloud information, where the target information is used to indicate at least one target within the detection range of the millimeter wave radar; Combining the target information and the motion data, generating a predicted collision result corresponding to the current vehicle and the target; Anti-collision warning information is generated based on the predicted collision result.

2. The method according to claim 1, characterized in that The receiving of point cloud information sent by the millimeter wave radar includes: Preset the working cycle for receiving point cloud information; The point cloud information is received based on the duty cycle.

3. The method according to claim 1, characterized in that The generating target information based on the point cloud information comprises: Extracting target data from the point cloud information to obtain point cloud data; Cluster fitting is performed on the point cloud data to generate the target information, where the target information includes at least one of target quantity information, target size information, and target position information.

4. The method according to claim 1, characterized in that: The step of combining the target information and the motion data to generate a predicted collision result between the current vehicle and the target includes: Determine a first driving trajectory of the current vehicle based on the motion data; Based on the target information, determining a second driving trajectory corresponding to the target; A predicted collision result corresponding to the target is generated based on the first driving trajectory and the second driving trajectory.

5. The method according to claim 4, characterized in that The determining a first driving trajectory of the current vehicle based on the motion data includes: Establishing a vehicle coordinate system corresponding to the current vehicle; Based on the motion data, the first driving trajectory of the current vehicle is determined in the vehicle coordinate system.

6. The method according to claim 5, characterized in that The motion data includes vehicle speed data and vehicle direction data.

7. The method according to claim 5, characterized in that The point cloud information includes lane data; The determining, based on the target information, a second driving trajectory corresponding to the target includes: Based on the target data, determine target lane data and target motion data corresponding to the target, the target lane data is used to indicate a current lane of the target, and the target motion data is used to indicate a current state of the target; The second driving trajectory is determined based on the target lane data and the target motion data.

8. The method according to claim 7, characterized in that The generating a predicted collision result corresponding to the target based on the first driving trajectory and the second driving trajectory includes: Determining lane consistency between the current vehicle and the target based on the first driving trajectory and the second driving trajectory; In response to the target and the current vehicle being located in the same lane, determining estimated time-to-collision data between the current vehicle and the target; In response to the estimated collision time data being less than a collision time threshold, the predicted collision result is determined to be a collision risk.

9. The method according to claim 8, characterized in that The method further comprises: In response to the target and the current vehicle not being located in the same lane, determining a trajectory intersection existence result based on the first driving trajectory and the second driving trajectory; In response to the trajectory intersection existence result indicating that the first driving trajectory and the second driving trajectory have an intersection, determining arrival time difference data based on the first driving trajectory and the second driving trajectory; In response to the arrival time difference data being less than an arrival time difference threshold, determining arrival time data of the current vehicle arriving at the intersection; In response to the arrival time data being less than an arrival time threshold, determining that the predicted collision result is that there is a collision risk.

10. A millimeter-wave radar forward collision warning system based on lane recognition, characterized in that: The system includes a millimeter wave radar and a computer device; The millimeter wave radar is communicatively connected with the computer device; The millimeter wave radar is used to generate point cloud information and send it to the computer device; The computer device is used to execute the millimeter-wave radar forward collision avoidance warning method based on lane recognition as described in any one of claims 1 to 9.

Images