A target screening method for lane lines according to a self-vehicle motion state

Abstract

The application discloses a target screening method for optimizing lane lines according to a self-vehicle motion state. When screening a target, a corresponding lane line is updated by considering a self-vehicle motion trend on the basis of an existing lane line, and the corresponding lane line is used as a target screening lane line. The method utilizes the information of the lane line and comprehensively considers the motion state of the self-vehicle, thereby ensuring the accuracy of target screening, providing reliable target parameters for longitudinal control of adaptive cruise control (ACC), improving the comfort of the whole longitudinal control, avoiding unexpected deceleration braking and other actions of the longitudinal control caused by target screening, and avoiding collision in the longitudinal control by timely decelerating according to a target vehicle when deceleration and braking are needed.

Description

TECHNICAL FIELD

[0001] The present application belongs to the field of vehicle longitudinal control, and particularly relates to a target screening method for optimizing lane lines according to a motion state of a vehicle. BACKGROUND

[0002] After the adaptive cruise control (ACC) of the current vehicle is activated, a target screening algorithm is needed to select a preceding vehicle to be followed. In the target screening algorithm, whether a preceding vehicle is in the current lane needs to be determined according to the positional relationship between the lane line and the preceding target vehicle, and the closest target is selected as the following target from all the target vehicles in the current lane. In the target screening, the current common methods for processing the lane line mainly include the following two methods: 1. When there is a lane line, the lane line in which the vehicle is located is directly used as the lane line for target screening. After the vehicle changes lanes, the lane line after the lane change is used. When the vehicle turns, the trajectory line of the vehicle is calculated as the lane line according to the trajectory of the vehicle. 2. Without relying on the actual lane line, the lane line is calculated according to the motion trajectory of the vehicle, and the target is screened according to the lane line.

[0003] The current two methods for processing the lane line mainly have the following problems: In the first method, directly using the lane line for target screening may lead to inaccurate target screening in some cases. The vehicle may deviate to one side when moving in the lane, or the vehicle may deviate from the lane during lane changing or turning, which may cause some deviation between the motion of the vehicle and the lane line. Directly using the lane line in which the vehicle is located to screen the target may lead to selecting some targets that are not intended to be selected or failing to select some targets that are intended to be selected. In the second method, the lane line is directly calculated according to the motion trend of the vehicle. Since the sensor of the vehicle may have some errors, the calculation of the trajectory of the vehicle may have some deviation, which may lead to instability of the target screening. Since the lane line width and shape are different in different road sections, the lack of information of the lane line may lead to the occurrence of misselection. Therefore, the present application provides a target screening method for optimizing the lane line according to the motion state of the vehicle to solve the above problems. SUMMARY

[0004] To solve at least one problem in the background technology, the present application provides a target screening method for optimizing the lane line according to the motion state of the vehicle.

[0005] To achieve the above object, the present application provides the following technical scheme: a target screening method for optimizing the lane line according to the motion state of the vehicle, comprising the following steps:

[0006] Step 1: acquiring the sensor input information of the vehicle and calculating the motion formula of the vehicle;

[0007] Step 2: According to the input information of the camera, the lane center line is calculated according to the left and right lane lines;

[0008] Step 3: According to steps 1 and 2, the starting position, the lateral position error of the vehicle to the lane center line, and the lateral position of the vehicle at the preview point, the lane line position at the preview point, and the distance of the vehicle to the lane center line at the preview point are calculated;

[0009] Step 4: According to the information in step 3, the motion state of the vehicle is calculated using the lateral position error at the starting point and the lateral position error at the preview point: the vehicle is centered, the vehicle is stable offset, the vehicle is left lane changing without crossing the line, the vehicle is left lane changing with crossing the line, the vehicle is right lane changing without crossing the line, and the vehicle is right lane changing with crossing the line;

[0010] Step 5: According to the information in step 4, the lane center line is updated and the preview point position parameter information for lane changing is found using the updated lane center line;

[0011] Step 6: According to the information in step 5, the motion parameters of the current point and the lane changing preview point are used to fit the lane center line with a quintic polynomial;

[0012] Step 7: According to the information in steps 4 and 5, the lane center line is fitted with a cubic polynomial within the effective distance from the lane changing preview point to the lane line;

[0013] Step 8: According to the information in step 7, for positions beyond the effective length of the lane line, the lane line is extended using Taylor expansion at the vanishing point of the lane line.

[0014] Preferably, the sensor input information of the vehicle includes but is not limited to data collected by the vehicle's speed sensor, acceleration sensor, and gyroscope sensor. These data are fused to calculate the motion formula of the vehicle, wherein the speed sensor is used to obtain the real-time driving speed of the vehicle, the acceleration sensor is used to obtain the acceleration size and direction of the vehicle, and the gyroscope sensor is used to obtain the angular velocity information of the vehicle.

[0015] Preferably, when calculating the lane center line according to the input information of the camera, the image collected by the camera is preprocessed, including image graying and filter noise reduction processing, to improve the recognition degree of the lane line information in the image, and then an edge detection algorithm is used to extract the edge contour of the lane line, and the extracted edge contour is filtered and fitted based on geometric features and prior knowledge of the lane line, so as to accurately calculate the lane center line.

[0016] Preferably, the step 3: according to step 1 and step 2, the starting position, the lateral position error of the vehicle to the lane center line, and the lateral position of the vehicle at the preview point, the lane line position at the preview point are calculated, and the distance of the vehicle to the lane center line at the preview point is calculated; the preview point is mainly used for calculation.

[0017] Preferably, the step 4 uses the lateral position error at the starting point and the lateral position error at the preview point to calculate the motion state of the vehicle; the threshold value for vehicle motion trend calculation is optimized according to the actual situation of the vehicle.

[0018] Preferably, the step 5: according to the information in step 4, the lane center line is updated, and the preview point position parameter information of lane changing is found using the updated lane center line, and the distance of the preview point is determined according to the actual vehicle.

[0019] Preferably, the step 6: according to the information in step 5, the motion parameters of the current point and the lane changing preview point are used to fit the lane center line with a quintic polynomial, wherein the fitting of the lane line includes but is not limited to fitting with a quintic polynomial.

[0020] Preferably, the step 7: according to the information in steps 4 and 5, the lane center line is fitted with a cubic polynomial within the effective distance of the lane line from the lane changing preview point, and the fitting of the lane line for this section includes but is not limited to fitting with a cubic polynomial.

[0021] Preferably, the step 8: according to the information in step 7, the lane line is extended using Taylor expansion at the vanishing point of the lane line beyond the effective length of the lane line; the extension of the lane line for this section includes but is not limited to Taylor expansion.

[0022] Compared with the prior art, the beneficial effects of the present application are:

[0023] 1. In urban road sections or road sections with heavy traffic or during vehicle motion, the vehicle has a deviation in the lane, and the accuracy of the target screening of the adaptive cruise control ACC is higher when using the optimized lane center line during lane changing and turning. Considering various motion states of the vehicle, the auxiliary driving function is more intelligent and more in line with the needs of the driver. The target screening method of the present application improves the accuracy of the longitudinal control target screening of the intelligent driving vehicle, can screen out real dangerous targets, and improves the stability of the longitudinal control target screening of the intelligent driving vehicle, providing reliable input for stable following of the adaptive cruise control ACC.

[0024] 2. At the same time, the target selection method of the present invention improves the comfort of intelligent driving vehicles, more effectively reflects the driver's driving intentions, and helps to increase public trust and acceptance of intelligent driving; it also improves the robustness and real-time performance of the longitudinal control algorithm. Attached Figure Description

[0025] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0026] Figure 1 A flowchart of a target selection method for optimizing lane lines based on the vehicle's motion state;

[0027] Figure 2 This is a schematic diagram of vehicle dynamics during vehicle turning in the present invention;

[0028] Figure 3 This is a schematic diagram of the state machine of the vehicle's motion state in this invention;

[0029] Figure 4 This is a diagram showing the fit of the lane lines. Detailed Implementation

[0030] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0031] First, refer to Figures 1 to 4 The present invention will specifically describe a target selection method for optimizing lane lines based on the vehicle's motion state, according to an embodiment of the present invention.

[0032] The invention will be further described in detail below with reference to the overall technical solution flowchart:

[0033] A target selection method for optimizing lane lines based on the vehicle's motion state includes the following steps:

[0034] Step 1: Obtain the vehicle's sensor input information and calculate the vehicle's motion formula;

[0035] Step 2: Based on the input information from the camera, calculate the lane center line according to the left and right lane lines;

[0036] Step 3: Based on Step 1 and Step 2, calculate the starting position, the lateral position error of the vehicle from the lane center line, the lateral position of the vehicle at the aiming point, the lane line position at the aiming point, and calculate the distance from the vehicle at the aiming point to the lane center line.

[0037] Step 4: According to the information in step 3, the motion state of the vehicle is calculated using the lateral position error at the starting point and the lateral position error at the preview point: the vehicle is driving in the center, the vehicle is stable deviation, the vehicle is left lane changing without crossing the line, the vehicle is left lane changing with crossing the line, the vehicle is right lane changing without crossing the line, and the vehicle is right lane changing with crossing the line;

[0038] Step 5: According to the information in step 4, update the lane center line and find the lane changing preview point position parameter information using the updated lane center line;

[0039] Step 6: According to the information in step 5, use the motion parameters of the current point and the lane changing preview point to fit the lane center line with a quintic polynomial;

[0040] Step 7: According to the information in steps 4 and 5, fit the lane center line with a cubic polynomial within the effective distance from the lane changing preview point to the lane line;

[0041] Step 8: According to the information in step 7, for positions beyond the effective length of the lane line, use Taylor expansion at the lane line disappearance point to extend the lane line.

[0042] The main goal of target screening is to provide reliable following target for the adaptive cruise ACC system. When the cruise speed set by the driver is greater than the speed of the target vehicle, the adaptive cruise ACC controls the vehicle to follow the target vehicle at the following distance set by the driver, and controls the speed of the vehicle to be consistent with the speed of the target vehicle.

[0043] The main principle of the target screening algorithm is as follows:

[0044] When there is a lane line and the lane line quality is valid, it is determined whether the front vehicle is in the lane line in which the vehicle is driving.

[0045] When the target is in the lane line in which the vehicle is driving, the vehicle closest to the vehicle is calculated as the target vehicle;

[0046] In this algorithm, the motion trend of the vehicle is considered. When the lane line does not exist, the motion trajectory of the vehicle is used to calculate the lane line, which is used as the lane line for target screening. When the lane line exists, the motion trend of the vehicle is used to optimize the lane line, so that the lane line is the driving area of the vehicle and the motion trend of the vehicle. The algorithm of the motion trend of the vehicle is as follows:

[0047] The state parameter diagram of the vehicle turning is shown in Figure 2 According to the motion parameters of the current state of the vehicle, the longitudinal speed v x , the yaw rate ω, the wheelbase + front suspension length L, the vehicle speed V, the angle θ between the front axle center and the rear axle center at the turning point, and the radius of the turning circle R fcR is the distance from the center of the front guard to the center of rotation. r R is the distance from the rear wheel to the center of rotation. f Let be the distance from the front wheel to the center of rotation. Then, the lateral velocity of the front bumper is:

[0048]

[0049] Centripetal acceleration a:

[0050]

[0051] v y Let be the initial lateral velocity of the vehicle, 'a' be the acceleration of the vehicle during lateral motion, and 'a' be the uniform longitudinal velocity. Then the motion parameters of the vehicle's front bumper are:

[0052]

[0053] Where: K is curvature

[0054] C1=L*K

[0055]

[0056] Then: y = C1*x + C2*x 2 (4)

[0057] In this algorithm, the distance ey between the current position and the lane line is calculated based on the vehicle's current position. current ,ey current The calculation method is as follows:

[0058] Based on the left and right lane lines of the vehicle, the lane center line is calculated. The formula for the calculated lane center line is:

[0059] y = a0 + a1x + a2x 2 +a3*x 3 (5)

[0060] At the starting position, the longitudinal position x is taken as 0. Substituting this into equation (4), the lateral position y of the vehicle is then calculated. current =0;

[0061] At the starting position, the longitudinal position x is taken as 0. Substituting this into equation (5), the lateral position of the lane centerline is y. current_lane =a0;

[0062] Find the first derivative of the lane lines, and the expression after the derivation is:

[0063] dy = a1 + 2a2x + a3x 2 (6)

[0064] Then x is taken as 0, and substituted into equation (6) to obtain the position error dy current_lane = a1

[0065] Then the distance between the vehicle and the lane line at the starting position is:

[0066]

[0067] According to the speed of the ego vehicle, the distance between the vehicle and the lane line at the preview point is calculated, and the specific calculation method is as follows:

[0068] The preview point can be calculated according to the speed multiplied by the preview time 3s, or according to the speed and the road radius to obtain x preview = v x * 3, or x preview is obtained according to the table. Since the motion state of the vehicle needs to be judged, the preview point is placed a distance in front of the vehicle, so the value of x preview is generally limited by the maximum and minimum values;

[0069] The motion position of the vehicle at the preview point and the position of the lane line are calculated respectively, and then the distance ey preview between the vehicle and the lane line at the preview point can be calculated.

[0070] First, according to equation (4), the lateral position ey preview of the vehicle at the preview point is calculated.

[0071]

[0072] According to equation (5), the position y preview_lane of the lane center line at the preview point of the vehicle is calculated.

[0073]

[0074] According to equation (6), the first derivative dy preview_lane of the lane center line at the preview point is calculated.

[0075]

[0076] Then the distance ey preview between the vehicle and the lane line at the preview point can be calculated according to the following formula:

[0077]

[0078] According to the position error ey current at the starting point of the vehicle and the position error ey preview at the preview point calculated above,The motion trend of the vehicle is determined according to different situations, and the lane center line is offset according to different motion trends. The motion of the vehicle is divided into: the vehicle is centered, the vehicle is stably offset, the vehicle is left lane changing without crossing the line, the vehicle is left lane changing with crossing the line, the vehicle is right lane changing without crossing the line, and the vehicle is right lane changing with crossing the line. The determination of the motion state of the vehicle is discussed as follows:

[0079] The vehicle is left lane changing without crossing the line: ey current <0&&ey preview -ey current <-0.7, it is determined that the vehicle is left lane changing without crossing the line.

[0080] The vehicle is right lane changing without crossing the line: ey current >0&&ey preview -ey current >0.7, it is determined that the vehicle is right lane changing without crossing the line.

[0081] The vehicle is left lane changing with crossing the line: when the vehicle is in the left lane changing without crossing the line in the last period, a0 of the lane center line at this moment and a 0last of the last moment are used to determine whether the vehicle has crossed the line, specifically a0-a 0last <-0.6*W, W is the width of the lane.

[0082] The vehicle is right lane changing with crossing the line: when the vehicle is in the right lane changing without crossing the line in the last period, a0 of the lane center line at this moment and a 0last of the last moment are used to determine whether the vehicle has crossed the line, specifically a0-a 0last >0.6*W, W is the width of the lane.

[0083] The vehicle is stably offset: when the vehicle is not in the above four states, when the conditions (ey current <0&&ey preview <-0.5) || (ey current >0&&ey preview >0.5) are met.

[0084] The vehicle is centered: when the above five conditions are not met, the vehicle is centered.

[0085] The state machine diagram of the vehicle motion state is shown in Figure 3 ;

[0086] The calculated vehicle motion state is used to optimize the lane center line, and the specific algorithm is as follows:

[0087] The lane center line equation is optimized according to various conditions by formula (5);

[0088] I. Vehicle left lane change without crossing the line: the lane center line a0 adopts the a0 of the left lane line.

[0089] II. Vehicle right lane change without crossing the line: the lane center line a0 adopts the a0 of the right lane line.

[0090] III. Vehicle left lane change has crossed the line: the lane center line a0 remains unchanged, and the a0 of the lane center line is adopted.

[0091] IV. Vehicle right lane change has crossed the line: the lane center line a0 remains unchanged, and the a0 of the lane center line is adopted.

[0092] V. Vehicle stable offset remains: the a0 of the lane center line is offset to the a0 of the current lane line plus the lateral offset value y at the preview point preview .

[0093] VI. Vehicle driving in the center of the lane: the center line of the lane remains unchanged.

[0094] In the above cases, the a1, a2 and a3 values of the lane center line all adopt the lane center line parameters, and the parameters at the lane change preview point are calculated using the updated lane center line formula (5). The lane change preview distance can be x preview = v x *4.5, that is, the preview is 4.5s; using the current motion parameters of the vehicle and the parameters at the lane change preview point, a quintic polynomial is used to fit the lane line from the starting point after the lane change to the effective distance of the lane line, and a cubic polynomial is used to fit the part exceeding the distance of the lane line, and the part is estimated using the lane line development at the vanishing point. The specific implementation is shown in Figure 4 .

[0095] Using the updated lane center line, the lane width adopts the width of the lane line of the ego vehicle, and the updated lane center line is used for target screening of adaptive cruise ACC.

[0096] The present application considers the motion trend of the ego vehicle, and uses different lane center lines according to different ego vehicle motion states, so that the target screening is more accurate, and the screened target reflects the driving intention of the vehicle, which can make the auxiliary driving function more intelligent.

[0097] Although embodiments of the present application have been shown and described, those of ordinary skill in the art can understand that various changes, modifications, replacements and variations can be made to these embodiments without departing from the principles and purposes of the present application, and the scope of the present application is defined by the claims and their equivalents.

Claims

1. A method for optimizing target screening of lane lines according to a self-vehicle motion state, characterized in that, The method comprises the following steps: Step 1: obtaining sensor input information of the vehicle, and calculating a motion formula of the vehicle; Step 2: calculating a lane center line according to input information of a camera and left and right lane lines; Step 3: calculating a starting position, a lateral position error of the vehicle to the lane center line, a lateral position of the vehicle at a preview point, a lane line position at the preview point, and a lateral position error of the vehicle to the lane center line at the preview point according to step 1 and step 2; Step 4: calculating a motion state of the vehicle according to the lateral position error at the starting position and the lateral position error at the preview point in step 3, wherein the motion state comprises the following states: the vehicle is driving in the center of the lane, the vehicle is stably deviating, the vehicle is left lane changing without crossing the lane, the vehicle is left lane changing with crossing the lane, the vehicle is right lane changing without crossing the lane, and the vehicle is right lane changing with crossing the lane; Step 5: updating the lane center line according to the information in step 4, wherein if the vehicle is left lane changing without crossing the lane, the lane center line adopts the left lane line; if the vehicle is right lane changing without crossing the lane, the lane center line adopts the right lane line; if the vehicle is left / right lane changing with crossing the lane or the vehicle is driving in the center of the lane, the lane center line is unchanged; if the vehicle is stably deviating, the deviation of the lane center line is the current lane line plus the lateral deviation value at the preview point; finding the lane changing preview point position parameter information by using the updated lane center line; Step 6: fitting the lane center line by using the motion parameters of the current point and the lane changing preview point according to the information in step 5; Step 7: fitting the lane center line by using a cubic polynomial within an effective distance from the lane changing preview point to the lane line according to the information in step 4 and step 5; Step 8: extending the lane line by using Taylor expansion at a lane line disappearance point for a position beyond the effective length of the lane line, using the updated lane center line, using the width of the lane line of the vehicle as the lane width, and using the updated lane center line to select a target for adaptive cruise control (ACC).

2. The method according to claim 1, wherein, The sensor input information of the vehicle comprises but is not limited to data collected by a speed sensor, an acceleration sensor and a gyroscope sensor of the vehicle, and the motion formula of the vehicle is calculated by fusing the data, wherein the speed sensor is used to obtain the real-time driving speed of the vehicle, the acceleration sensor is used to obtain the acceleration size and direction of the vehicle, and the gyroscope sensor is used to obtain the angular velocity information of the vehicle.

3. The method of claim 1, wherein the target filtering method is optimized according to the motion state of the vehicle. When the lane center line is calculated according to the input information of the camera, the image collected by the camera is preprocessed, including image graying and filter denoising, so as to improve the recognition degree of the lane line information in the image, and then an edge detection algorithm is used to extract the edge contour of the lane line, and the extracted edge contour is screened and fitted based on geometric features and prior knowledge of the lane line, so as to accurately calculate the lane center line.

4. The method of claim 1, wherein the target filtering method is optimized according to the motion state of the vehicle. Step 3: calculating a starting position, a lateral position error of the vehicle to the lane center line, a lateral position of the vehicle at a preview point, a lane line position at the preview point, and a lateral position error of the vehicle to the lane center line at the preview point according to step 1 and step 2; mainly using the preview point for calculation.

5. The method of claim 1, wherein the target filtering method is optimized according to the motion state of the vehicle. The step 4 uses the lateral position error at the starting point and the lateral position error at the preview point to calculate the motion state of the vehicle; the threshold for the vehicle motion trend calculation is optimized according to the actual situation of the vehicle.

6. The method of claim 1, wherein the target filtering method is optimized according to the motion state of the vehicle. The step 5 updates the lane center line according to the information in the step 4, and finds the lane-changing preview point position parameter information using the updated lane center line, and the distance of the preview point is determined according to the actual vehicle.

7. The method of claim 1, wherein the target filtering method is optimized according to the motion state of the vehicle. The step 6 uses the motion parameters of the current point and the lane-changing preview point to fit the lane center line using a quintic polynomial according to the information in the step 5, and the fitting of the lane line includes but is not limited to fitting using a quintic polynomial.

8. The method of claim 1, wherein the target filtering method is optimized according to the motion state of the vehicle. The step 7 uses a cubic polynomial to fit the lane center line within the effective distance from the lane-changing preview point to the lane line according to the information in the steps 4 and 5, and the fitting of the lane line in this section includes but is not limited to fitting using a cubic polynomial.

9. The method of claim 1, wherein the target filtering method is optimized according to the motion state of the vehicle. The step 8 uses Taylor expansion at the lane line disappearance point to extend the lane line beyond the effective length of the lane line according to the information in the step 7, and the extension of the lane line in this section includes but is not limited to Taylor expansion.

Images