Vision-based eba deceleration request method, system and vehicle

Abstract

This invention discloses a vision-based EBA deceleration request method, comprising the following steps: S1, after detecting a longitudinal collision risk between the vehicle and the target vehicle, periodically predicting the longitudinal trajectory T of the vehicle. H Longitudinal trajectory T of the target vehicle G S2, Detect the longitudinal trajectory T of this vehicle within the current prediction period. H The longitudinal trajectory T of the target vehicle during this period G If overlap is detected, and the result is yes, the deceleration increment requested in the current prediction cycle is determined based on the relative vehicle speed at the time of collision; S3, after detecting that the driver of this vehicle intends to brake, this vehicle performs a deceleration action based on the sum of the deceleration increments requested in the current and previous prediction cycles. This invention periodically predicts the longitudinal travel trajectory T of this vehicle. H Longitudinal trajectory T of the target vehicle G This is used to achieve collision detection. When a collision risk is detected, the vehicle calculates a deceleration request based on the actual vehicle calibration and adjusts its speed to reduce the risk of a collision between the two vehicles.

Description

Technical Field

[0001] This invention belongs to the field of vehicle driver assistance technology, and more specifically, this invention relates to a vision-based EBA deceleration request method, system, and automobile. Background Technology

[0002] ADAS (Advanced Driver Assistance Systems) technology can use a front-facing camera to identify target vehicles, output the kinematic relationship between the vehicle and the target vehicle, and thus analyze the collision risk with the target vehicle. When a collision risk is detected, it will issue relevant warnings to the driver. Even if the driver presses the brake pedal, the collision risk may still exist.

[0003] Existing technology relies on a front-facing monocular camera to identify target vehicles within the lane and acquire information such as the target vehicle's longitudinal distance, speed, and acceleration. However, since the acquired target vehicle information differs from the actual target vehicle information, simply relying on a formula to calculate the vehicle's current required deceleration—that is, relative speed divided by time to collision (TTC)—will result in a deviation from the actual target deceleration required by the vehicle.

[0004] If the resulting deceleration is too large compared to the required target deceleration, it will cause a false triggering situation, where the driver could have avoided the collision risk by stepping on the brakes, but instead the driver requested excessive deceleration, resulting in a driver complaint; if the resulting deceleration is too small compared to the required target deceleration, it may cause a collision risk between the vehicle and the target vehicle in front. Summary of the Invention

[0005] This invention provides a vision-based EBA deceleration request method, which aims to improve the above-mentioned problems.

[0006] This invention is implemented as follows: a vision-based EBA deceleration request method, the method comprising the following steps:

[0007] S1. After detecting a longitudinal collision risk between the vehicle and the target vehicle, the vehicle begins to periodically predict its longitudinal trajectory T. H Longitudinal trajectory T of the target vehicle G ;

[0008] S2. Detect the longitudinal trajectory T of this vehicle within the current prediction period. H The longitudinal trajectory T of the target vehicle during this period G If there is overlap, and the detection result is yes, then the deceleration increment requested in the current prediction cycle is determined based on the relative vehicle speed at the time of the collision.

[0009] S3. Upon detecting that the driver of this vehicle intends to brake, this vehicle performs a deceleration action based on the sum of the deceleration increments requested in the current and previous prediction cycles.

[0010] Furthermore, if the longitudinal trajectory T of this vehicle within the current prediction period... H The longitudinal trajectory T of the target vehicle during this period G If there is no overlap, the deceleration increment for the next prediction period is determined based on the minimum longitudinal distance between the vehicle and the target vehicle within the current prediction period.

[0011] Furthermore, the longitudinal driving trajectory T of the vehicle within each prediction period H The longitudinal trajectory T of the target vehicle G The specific method for obtaining it is as follows:

[0012] Set the prediction step size N within the prediction period, and gradually predict the longitudinal travel distance of the vehicle and the target vehicle within each prediction step to form the longitudinal travel trajectory of the vehicle and the target vehicle.

[0013] Furthermore, the acceleration 'a' of the target vehicle at the beginning of the prediction period is... G [0] represents the acceleration a of the target vehicle during this prediction period. G To obtain the longitudinal trajectory T of the target vehicle within the prediction period. G .

[0014] Furthermore, the acceleration a of this vehicle H The specific method for determining it is as follows:

[0015] Calculate reaction time T R Corresponding forecast period In the first m prediction cycles, let the vehicle's acceleration be equal to the acceleration *a* at the beginning of the prediction cycle. H [0]; After m+1 prediction cycles, the vehicle's acceleration changes linearly, and the longitudinal acceleration a of the vehicle in the j+1 prediction cycle is... H,j+1 =a H,j +DecRate*T C Where DecRate is the linear rate of change of acceleration caused by braking, a H,j Let a be the longitudinal acceleration of the vehicle in the j-th prediction cycle, and detect the longitudinal acceleration a. H,j+1 If the longitudinal acceleration exceeds the acceleration threshold, and the result is yes, then the longitudinal acceleration of the vehicle is taken as the longitudinal acceleration threshold; if the result is no, then the longitudinal acceleration of the vehicle is taken as the longitudinal acceleration 'a'. H,j+1 .

[0016] Furthermore, in the longitudinal driving trajectory T G The array contains elements whose longitudinal distance is greater than the longitudinal travel trajectory T. H The corresponding vertical distance in the array indicates that there is a risk of collision between this vehicle and the target vehicle.

[0017] Furthermore, the method for determining the deceleration increment based on the relative vehicle speed at the time of collision is as follows:

[0018] The deceleration increment corresponding to the relative vehicle speed is found based on the relative vehicle speed-deceleration increment table. The specific process of forming the relative vehicle speed-deceleration increment table is as follows:

[0019] The simulation software simulates a collision between two vehicles at various relative speeds and determines the minimum deceleration increment to avoid the collision, which is then used as the initial deceleration increment at that relative speed.

[0020] Calibration and testing were performed under CNCAP standard operating conditions to optimize the initial deceleration increment at various relative vehicle speeds, ultimately resulting in a relative vehicle speed-deceleration increment table.

[0021] Furthermore, the method for determining the deceleration increment based on the minimum longitudinal distance is as follows:

[0022] The velocity increment corresponding to the current minimum longitudinal distance is found based on the minimum longitudinal distance-deceleration increment table; the specific process of forming the minimum longitudinal distance-deceleration increment table is as follows:

[0023] The simulation software simulates a collision between two vehicles at various relative longitudinal distances, and determines the minimum deceleration increment to avoid the collision, which is then used as the initial deceleration increment at that relative longitudinal distance.

[0024] Calibration and testing were performed under CNCAP standard operating conditions to optimize the initial deceleration increment for each relative longitudinal distance, ultimately resulting in a minimum longitudinal distance-deceleration increment table.

[0025] This invention is implemented as follows: a vision-based EBA deceleration request system, the system comprising:

[0026] A monocular camera located at the front of the vehicle captures images of the area in front of the vehicle and sends them to the processor of the Brake Assist (EBA) system. The processor determines the deceleration increment requested when the vehicle intends to brake based on the aforementioned vision-based EBA deceleration request method.

[0027] The present invention is implemented as follows: a car that integrates the vision-based EBA deceleration request system described above.

[0028] This invention predicts the longitudinal trajectory T of the vehicle periodically. H Longitudinal trajectory T of the target vehicle GTo achieve collision detection, when a collision risk is detected, the vehicle speed is adjusted based on the deceleration increment calibrated on the real vehicle to reduce the risk of a collision between the two vehicles. In addition, even if no collision risk is detected between the two vehicles, the vehicle speed is adjusted based on the minimum longitudinal distance between them to further reduce the risk of a collision between the two vehicles. Attached Figure Description

[0029] Figure 1 A flowchart of a vision-based EBA deceleration request method provided in an embodiment of the present invention. Detailed Implementation

[0030] The specific embodiments of the present invention will be further described in detail below with reference to the accompanying drawings, so as to help those skilled in the art to have a more complete, accurate and in-depth understanding of the inventive concept and technical solution of the present invention.

[0031] Figure 1 The flowchart of the vision-based EBA deceleration request method provided in this embodiment of the invention includes the following steps:

[0032] S1. After detecting a longitudinal collision risk between the vehicle and the target vehicle, the vehicle begins to periodically predict its longitudinal trajectory T. H Longitudinal trajectory T of the target vehicle G ;

[0033] In this embodiment of the invention, the target vehicle is the other vehicle closest to the vehicle in front in the current lane. The lane in which the vehicle is located is the current lane. When the collision time between the vehicle and the target vehicle is less than a set threshold, it is determined that there is a longitudinal collision risk between the vehicle and the target vehicle. Starting from the moment when a longitudinal collision risk between the vehicle and the target vehicle is detected, the longitudinal trajectory T of the vehicle is periodically predicted. H Longitudinal trajectory T of the target vehicle G The longitudinal trajectory T of the vehicle within each prediction period H The longitudinal trajectory T of the target vehicle G The specific method for obtaining it is as follows:

[0034] Set the prediction step size N within the prediction period, and gradually predict the longitudinal travel distance of the vehicle and the target vehicle within each prediction step to form the longitudinal travel trajectory of the vehicle and the target vehicle.

[0035] In this embodiment of the invention, the longitudinal direction refers to the vehicle's driving direction. The origin of the vehicle's coordinate system at the initial moment of the prediction period is used as the origin of the longitudinal distance traveled. The longitudinal distance traveled by the vehicle at the initial moment of the prediction period is 0. The longitudinal distance traveled by the target vehicle at the initial moment of the prediction period is the longitudinal distance between the current vehicle and the target vehicle, represented as follows:

[0036] T H [0] = 0;

[0037] T G [0] = Relative_distance;

[0038] Use T G [i] represents the longitudinal distance traveled by the target vehicle at the i-th prediction step, where i ranges from 0 to N. The acceleration a of the target vehicle at the initial moment of the prediction period is also considered. G [0] represents the acceleration of the target vehicle during that prediction period, T. G The specific formula for calculating [i] is as follows:

[0039] T G [i] = T G [i-1]+1 / 2*(v G [i]+v G [i-1])*T C ;

[0040] v G [i] = v G [i-1]+a G [i-1]*T C ;

[0041] a G [i-1] = a G [i-2]=a G [0];

[0042] Among them, T G [i-1] is the longitudinal distance traveled by the target vehicle at the (i-1)th prediction step, v G [i]、v G [i-1] represent the longitudinal vehicle speeds of the target vehicle at the i-th and (i-1)-th prediction steps, respectively, and T C To predict the duration of the cycle, a G [i-1]、a G [i-2] represent the longitudinal acceleration of the target vehicle at the (i-1)th and (i-2)th prediction steps, respectively, a G [0] represents the longitudinal acceleration of the target vehicle at the beginning of the prediction period, a G [0]、vG [0] is obtained by analyzing the images of the front of the vehicle captured by the monocular camera at the front of the vehicle.

[0043] Use T H [i] represents the longitudinal distance traveled by the vehicle at the i-th prediction step. The method for obtaining the longitudinal distance traveled by the vehicle at the i-th prediction step is as follows:

[0044] There is a driver reaction time T between detecting a longitudinal collision risk between the vehicle and the target vehicle and the driver pressing the brake pedal. R Calculate the reaction time T R Corresponding forecast period In the first m prediction cycles, let the vehicle's acceleration be equal to the acceleration *a* at the beginning of the prediction cycle. H [0]; After m+1 prediction cycles, the vehicle's acceleration changes linearly, and the longitudinal acceleration a of the vehicle in the j+1 prediction cycle is... H,j+1 =a H,j +DecRate*T C Where DecRate is the linear rate of change of acceleration caused by braking, a H,j Let a be the longitudinal acceleration of the vehicle in the j-th prediction cycle, and detect the longitudinal acceleration a. H,j+1 If the longitudinal acceleration exceeds the acceleration threshold, and the result is yes, then the longitudinal acceleration of the vehicle is taken as the longitudinal acceleration threshold; if the result is no, then the longitudinal acceleration of the vehicle is taken as the longitudinal acceleration 'a'. H,j+1 The longitudinal distance T traveled by the vehicle at the i-th predicted step. H The calculation principle of [i] is the same as above, and will not be repeated here. It involves the vehicle's acceleration a at the beginning of the prediction cycle. H [0], velocity v H [0] is read from this vehicle in real time.

[0045] S2. Detect the longitudinal trajectory T of this vehicle within the current prediction period. H The longitudinal trajectory T of the target vehicle during this period G If there is overlap, and the detection result is yes, then the deceleration increment requested in the current prediction cycle is determined based on the relative vehicle speed at the time of the collision.

[0046] In this embodiment of the invention, if the longitudinal travel trajectory T of the vehicle within the current prediction period... H The longitudinal trajectory T of the target vehicle during this period G There is overlap, i.e., the longitudinal driving trajectory T G The array contains elements whose longitudinal distance is greater than the longitudinal travel trajectory T. HThe corresponding longitudinal distance in the array indicates a collision risk between the vehicle and the target vehicle. The relative speeds between the vehicle and the target vehicle at the moment of trajectory overlap are calculated. Based on the relative speed-deceleration increment table, the deceleration increment corresponding to the relative speed is found. The specific process of forming the relative speed-deceleration increment table is as follows:

[0047] The simulation software simulates a collision between two vehicles at various relative speeds and determines the minimum deceleration increment to avoid the collision, which is then used as the initial deceleration increment at that relative speed.

[0048] Calibration and testing were performed under CNCAP standard operating conditions to optimize the initial deceleration increment at various relative vehicle speeds, ultimately resulting in a relative vehicle speed-deceleration increment table.

[0049] S3. Upon detecting that the driver of this vehicle intends to brake, this vehicle performs a deceleration action based on the sum of the deceleration increments requested in the current and previous prediction cycles.

[0050] In this embodiment of the invention, if the longitudinal travel trajectory T of the vehicle within the current prediction period... H The longitudinal trajectory T of the target vehicle during this period G If there is no overlap, the deceleration request increment for the current period is determined based on the minimum longitudinal distance between the current vehicle and the target vehicle within the current prediction period.

[0051] In this embodiment of the invention, the velocity increment corresponding to the current minimum longitudinal distance is found based on the minimum longitudinal distance-deceleration increment table. The specific process of forming the minimum longitudinal distance-deceleration increment table is as follows:

[0052] The simulation software simulates a collision between two vehicles at various relative longitudinal distances, and determines the minimum deceleration increment to avoid the collision, which is then used as the initial deceleration increment at that relative longitudinal distance.

[0053] Calibration and testing were performed under CNCAP standard operating conditions to optimize the initial deceleration increment for each relative longitudinal distance, ultimately resulting in a minimum longitudinal distance-deceleration increment table.

[0054] This invention predicts the longitudinal trajectory T of the vehicle periodically. H Longitudinal trajectory T of the target vehicle G This system enables collision risk detection. When a collision risk is detected, the system calculates a deceleration request based on the actual vehicle calibration and adjusts the vehicle's speed accordingly to reduce the risk of a collision between the two vehicles. Furthermore, even if no collision risk is detected between the two vehicles, the system adjusts the deceleration request based on the minimum longitudinal distance between them to avoid excessive braking distance.

[0055] The present invention also provides a vision-based EBA deceleration request system, the system comprising:

[0056] A monocular camera located at the front of the vehicle captures images of the area in front of the vehicle and sends them to the processor of the Brake Assist (EBA) system. The processor determines the deceleration increment requested when the vehicle intends to brake based on the aforementioned vision-based EBA deceleration request method.

[0057] The present invention also provides an automobile that integrates the above-mentioned vision-based EBA deceleration request system, wherein a monocular camera is integrated into the windshield.

[0058] The present invention has been described by way of example. Obviously, the specific implementation of the present invention is not limited to the above-described manner. Any non-substantial improvements made using the inventive concept and technical solution of the present invention, or the direct application of the inventive concept and technical solution of the present invention to other occasions without modification, are all within the protection scope of the present invention.

Claims

1. A vision-based EBA deceleration request method, characterized in that, The method comprises the following steps: S1, after detecting that there is a longitudinal collision risk between the host vehicle and the target vehicle, start periodically predicting a longitudinal driving trajectory T of the host vehicle H the longitudinal driving trajectory T of the target vehicle G ; S2, detecting a longitudinal travel trajectory T of the subject vehicle in a current prediction period H a longitudinal travel trajectory T of the target vehicle in the current period G whether there is an overlap, and if the detection result is yes, determining a deceleration increment requested in the current prediction period based on a relative vehicle speed at the time of collision; S3, after detecting that the driver of the host vehicle has braking intention, the host vehicle performs deceleration action based on the sum of the requested deceleration increments in the current and previous prediction periods; The method for determining the deceleration increment based on the relative vehicle speed at the time of collision is as follows: The relative vehicle speed-deceleration increment table is formed as follows: The simulation software is used to simulate the collision of the two vehicles under each relative vehicle speed, and the minimum deceleration increment to avoid the collision of the two vehicles is determined as the initial deceleration increment under the relative vehicle speed. The calibration and test are performed under the CNCAP standard working condition, and the initial deceleration increment under each relative vehicle speed is optimized to finally form the relative vehicle speed-deceleration increment table.

2. The vision-based EBA deceleration request method of claim 1, wherein, If the longitudinal travel trajectory T H of the target vehicle in the current prediction period G there is no overlap, the deceleration increment of the current period is determined based on the minimum longitudinal distance of the host vehicle relative to the target vehicle in the current prediction period.

3. The vision-based EBA deceleration request method according to claim 1 or 2, characterized in that the longitudinal travel trajectory T of the host vehicle in each prediction cycle H the longitudinal travel trajectory T of the target vehicle G The acquisition method of the longitudinal travel trajectory T is as follows: The prediction step N in the prediction period is set, and the longitudinal driving distance of the host vehicle and the target vehicle in each prediction step is predicted step by step to form the longitudinal driving trajectory of the host vehicle and the target vehicle.

4. The vision-based EBA deceleration request method of claim 3, wherein, The acceleration a of the target vehicle at the initial time of the prediction period is obtained G [0] The acceleration a of the target vehicle within the prediction period is obtained G , to obtain a longitudinal travel trajectory T of the target vehicle within the prediction period G .

5. The vision-based EBA deceleration request method of claim 3, wherein, Acceleration a of the vehicle H The determination method is as follows: Computing the reaction time T R The corresponding prediction period In the first m prediction periods, let the vehicle acceleration equal the acceleration a of the vehicle at the initial time of the prediction period H [0]; after m+1 prediction periods, the vehicle acceleration changes linearly, and the longitudinal acceleration a of the vehicle at the j+1 prediction period is H,j+1 = a H,j + DecRate*T C , wherein DecRate is the linear change rate of the acceleration caused by braking, a H,j is the longitudinal acceleration of the vehicle at the j prediction period, and whether the longitudinal acceleration a H,j+1 is greater than the acceleration threshold value is detected, if the detection result is yes, the longitudinal acceleration of the vehicle takes the longitudinal acceleration threshold value, and if the detection result is no, the longitudinal acceleration of the vehicle takes the longitudinal acceleration a H,j+1 .

6. The vision-based EBA deceleration request method of claim 3, wherein, In the longitudinal driving track T G The longitudinal distance in the array is greater than the longitudinal driving track T H The corresponding longitudinal distance in the array, it is determined that the vehicle and the target vehicle exist collision risk.

7. The vision-based EBA deceleration request method of claim 1, wherein, The method for determining the deceleration increment based on the minimum longitudinal distance is as follows: The minimum longitudinal distance-deceleration increment table is used to find the speed increment corresponding to the current minimum longitudinal distance; the formation process of the minimum longitudinal distance-deceleration increment table is as follows: The simulation software is used to simulate the collision of the two vehicles under each relative longitudinal distance, and the minimum deceleration increment to avoid the collision of the two vehicles is determined as the initial deceleration increment under the relative longitudinal distance. The calibration and test are performed under the CNCAP standard working condition, and the initial deceleration increment under each relative longitudinal distance is optimized to finally form the minimum longitudinal distance-deceleration increment table.

8. A vision-based EBA deceleration request system characterized by, The system comprises: The monocular camera arranged at the front end of the vehicle collects the image in front of the current vehicle and sends it to the processor of the EBA system, and the processor determines the requested deceleration increment when the host vehicle has braking intention based on the visual-based EBA deceleration request method according to any one of claims 1 to 7.

9. An automobile characterized by comprising: The automobile is integrated with the visual-based EBA deceleration request system according to claim 8.

Images