Lane line prediction method, lane line prediction system and vehicle

Abstract

The invention discloses a lane line prediction method, a lane line prediction system and a vehicle. The lane line prediction method comprises the following steps: predicting a longitudinal movement distance of a vehicle according to motion state data of the vehicle and a lane line polynomial equation; according to the longitudinal movement distance, performing time sequence recursion processing on lane line coefficients in the lane line polynomial equation to obtain a coefficient prediction sequence; smoothing the predicted values of the lane line coefficients in the coefficient prediction sequence in the plurality of prediction periods to obtain target values of the lane line coefficients; and according to the target value of the lane line coefficient, updating the lane line polynomial equation to obtain a predicted lane line. According to the method, the change of the lane line coefficient is more continuous and smoother, the conditions of control jitter, instruction jump and the like caused by the fact that the operation period of planning control is smaller than the updating period of sensing fitting are avoided, the smoothness of the vehicle control process is guaranteed, and the vehicle control effect is effectively improved.

Description

Technical Field

[0001] This application relates to the field of vehicle technology, specifically to a lane prediction method, a lane prediction system, and a vehicle. Background Technology

[0002] With the advancement of intelligent driving technology, modern vehicles have widely adopted advanced driver assistance functions such as path planning and lateral control. The realization of these functions heavily relies on real-time perception and accurate modeling of lane line geometry. Current mainstream solutions involve sampling road images using the vehicle's front-facing camera, then fitting these images into cubic or quintic polynomial equations using a perception fusion algorithm to represent the lane lines. However, the update cycle for the perceived and fitted lane lines is typically long (e.g., 100 milliseconds), while the computation cycle for planning and control is usually short (e.g., 20 milliseconds). This difference in cycles leads to an unsmooth vehicle control process, affecting the vehicle's control performance. Summary of the Invention

[0003] This application provides a lane prediction method, a lane prediction system, and a vehicle, which makes the changes in lane coefficients more continuous and smooth, avoiding control jitter and command jumps caused by the planning control's calculation cycle being shorter than the perception fitting's update cycle, ensuring the smoothness of the vehicle control process, and effectively improving the vehicle control effect, thereby at least partially solving the above-mentioned technical problems.

[0004] In a first aspect, an embodiment of this application provides a lane line prediction method, comprising: predicting the longitudinal movement distance of a vehicle based on vehicle motion state data and a lane line polynomial equation; performing time-series recursive processing on the lane line coefficients in the lane line polynomial equation based on the longitudinal movement distance to obtain a coefficient prediction sequence, the coefficient prediction sequence including predicted values ​​of the lane line coefficients over multiple prediction periods; smoothing the predicted values ​​of the lane line coefficients over the multiple prediction periods to obtain a target value of the lane line coefficients; and updating the lane line polynomial equation based on the target value of the lane line coefficients to obtain a predicted lane line.

[0005] Optionally, the step of performing time-series recursive processing on the lane line coefficients in the lane line polynomial equation based on the longitudinal movement distance to obtain a coefficient prediction sequence includes: substituting the longitudinal movement distance into the lane line polynomial equation in the k-th prediction period to obtain the predicted value of the constant in the lane line polynomial equation in the (k+1)-th prediction period; determining the predicted value of the polynomial coefficients in the lane line polynomial equation in the (k+1)-th prediction period based on the longitudinal movement distance and the derivative of the lane line polynomial equation in the k-th prediction period; wherein the lane line coefficients include the constant and the polynomial coefficients; and k is an integer greater than or equal to 0.

[0006] Optionally, smoothing the predicted values ​​of the lane line coefficient over multiple prediction periods to obtain a target value for the lane line coefficient includes: determining a smoothing coefficient sequence based on the vehicle's yaw angle, the smoothing coefficient sequence including correction values ​​of the smoothing coefficient over multiple prediction periods; performing statistical processing on the correction values ​​of the smoothing coefficient over multiple prediction periods to obtain a cumulative value of the smoothing coefficient; performing a weighted summation on the predicted values ​​of the lane line coefficient over multiple prediction periods based on the correction values ​​of the smoothing coefficient over multiple prediction periods to obtain a fused value of the lane line coefficient; and determining the target value of the lane line coefficient based on the fused value of the lane line coefficient and the cumulative value of the smoothing coefficient.

[0007] Optionally, predicting the longitudinal movement distance of the vehicle based on the vehicle's motion state data and the lane line polynomial equation includes: predicting the yaw angle of the vehicle based on the yaw rate in the motion state data; predicting the travel distance of the vehicle based on the vehicle speed in the motion state data; if the yaw angle is zero, then determining that the longitudinal movement distance is equal to the travel distance; if the yaw angle is non-zero, then determining the longitudinal movement distance based on the yaw angle, the travel distance, and the lane line polynomial equation.

[0008] Optionally, determining the longitudinal movement distance based on the yaw angle, the travel distance, and the lane line polynomial equation includes: determining an initial value for the longitudinal movement distance based on the yaw angle, the travel distance, and a constant in the lane line polynomial equation; substituting the initial value of the longitudinal movement distance into the lane line polynomial equation to obtain an initial value for the lateral movement distance; and iteratively updating the longitudinal movement distance based on the yaw angle, the lane line polynomial equation, the initial value of the longitudinal movement distance, and the initial value of the lateral movement distance.

[0009] Optionally, the step of iteratively updating the longitudinal movement distance based on the yaw angle, the lane polynomial equation, the initial value of the longitudinal movement distance, and the initial value of the lateral movement distance includes: determining the parameter value of the longitudinal movement distance in the (i+1)th iteration based on the parameter values ​​of the yaw angle, the longitudinal movement distance in the (i)th iteration, and the lateral movement distance in the (i)th iteration; substituting the parameter value of the longitudinal movement distance in the (i+1)th iteration into the lane polynomial equation to obtain the parameter value of the lateral movement distance in the (i+1)th iteration; if the lateral movement distance in the (i+1)th iteration... If the difference between the updated parameter value and the parameter value updated in the i-th iteration of the lateral movement distance is less than zero, then the parameter value of the longitudinal movement distance is determined to be the parameter value updated in the (i+1)-th iteration of the longitudinal movement distance. If the difference between the parameter value updated in the (i+1)-th iteration of the lateral movement distance and the parameter value updated in the i-th iteration of the lateral movement distance is greater than or equal to zero, then i is set to i+1, and the process is repeated from the step of determining the parameter value updated in the (i+1)-th iteration of the longitudinal movement distance based on the yaw angle, the parameter value updated in the i-th iteration of the longitudinal movement distance, and the parameter value updated in the i-th iteration of the lateral movement distance. Here, i is an integer greater than or equal to 0, the parameter value updated in the 0th iteration of the longitudinal movement distance is equal to the initial value of the longitudinal movement distance, and the parameter value updated in the 0th iteration of the lateral movement distance is equal to the initial value of the lateral movement distance.

[0010] Secondly, the lane prediction system described in the embodiments of this application includes: a lane prediction module, used to execute the lane prediction method as described above.

[0011] Optionally, the lane prediction system further includes: a lane perception module connected to the lane prediction module, used for: fitting a lane polynomial equation; and sending the lane polynomial equation to the lane prediction module.

[0012] Optionally, the lane prediction system further includes: a chassis module connected to the lane prediction module, used to: send the vehicle's motion state data to the lane prediction module.

[0013] Thirdly, a vehicle according to an embodiment of this application includes the lane prediction system described above.

[0014] In summary, the technical solution provided in this application predicts the longitudinal movement distance of the vehicle based on the vehicle's motion state data and the lane line polynomial equation; based on the longitudinal movement distance, the lane line coefficients in the lane line polynomial equation are processed by time-series recursion to obtain a coefficient prediction sequence; the predicted values ​​of the lane line coefficients in the coefficient prediction sequence within multiple prediction periods are smoothed to obtain the target values ​​of the lane line coefficients; and the lane line polynomial equation is updated based on the target values ​​of the lane line coefficients to obtain the predicted lane lines. This application's embodiment predicts the parameter values ​​of the lane line coefficients within multiple future prediction periods based on the vehicle's real-time motion state data and the lane line polynomial equation fitted by the vehicle's actual perception. This allows for multiple intelligent updates of the lane line coefficient parameter values ​​within a longer perception fitting update period, making the changes in the lane line coefficients more continuous and smooth. This avoids control jitter and command jumps caused by the planning control's calculation cycle being shorter than the perception fitting update cycle, ensuring the smoothness of the vehicle control process and effectively improving the vehicle control performance. Furthermore, this embodiment uses multiple predicted values ​​for comprehensive calculation instead of a single prediction, avoiding the random errors of a single prediction. While preserving the true geometric trend of the lane lines, it averages perceived jitter and abnormal jumps in the time domain, effectively improving the accuracy of the lane line coefficient and achieving real-time accurate prediction of lane lines. In addition, this embodiment implements forward prediction of the lane line coefficient instead of backward smoothing, eliminating the inherent phase lag effect of traditional low-pass filtering, ensuring the timeliness of vehicle control, and further improving the vehicle control effect. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 This is a flowchart of a lane line prediction method provided in an embodiment of this application; Figure 2 This is a flowchart of another lane prediction method provided in the embodiments of this application; Figure 3 This is a schematic diagram of a lane prediction system provided in an embodiment of this application; Figure 4 This is a schematic diagram of another lane prediction system provided in an embodiment of this application; Figure 5 This is a schematic diagram of a vehicle provided in an embodiment of this application. Detailed Implementation

[0017] 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 a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the protection scope of this application.

[0018] Please see Figure 1 , Figure 1 This is a flowchart of a lane line prediction method provided in an embodiment of this application. Figure 1 As shown, the lane line prediction method may include the following steps S100 to S400.

[0019] Step S100: Based on the vehicle's motion state data and the lane line polynomial equation, predict the vehicle's longitudinal movement distance; Step S200: Based on the longitudinal movement distance, perform time-series recursive processing on the lane line coefficients in the lane line polynomial equation to obtain the coefficient prediction sequence; wherein, the coefficient prediction sequence includes the predicted values ​​of the lane line coefficients in multiple prediction periods; Step S300: Smooth the predicted values ​​of the lane line coefficient over multiple prediction periods to obtain the target value of the lane line coefficient; Step S400: Update the lane line polynomial equation based on the target value of the lane line coefficient to obtain the predicted lane line.

[0020] Motion state data is used to indicate the driving status of a vehicle. This application does not limit the specific type of motion state data; in practical applications, it can be flexibly set according to requirements. For example, motion state data may include, but is not limited to, at least one of the following: yaw rate (unit: degrees / second) and vehicle speed (unit: meters / second).

[0021] The lane line polynomial equation is used to indicate the lane line fitted by vehicle perception, and can be obtained by a perception fusion algorithm based on the road image corresponding to the vehicle. This application does not limit the specific implementation form of the lane line polynomial equation; in practical applications, it can be flexibly set according to requirements. For example, the lane line polynomial equation can be implemented as a cubic polynomial equation or a fifth-order polynomial equation, etc. Furthermore, the lane line predicted in this application embodiment (i.e., the predicted lane line) can be the lane line corresponding to the lane where the vehicle is located, including but not limited to any of the following: solid line, dashed line, solid-dashed line, continuous straight lane line, discontinuous straight lane line, discontinuous dashed lane line, double solid line, etc.

[0022] In step S100, the longitudinal movement distance of the vehicle is predicted based on the motion state data and the lane line polynomial equation. The longitudinal movement distance refers to the displacement component along the vehicle's longitudinal axis. Typically, the vehicle coordinate system has its origin at the rear axle center, with the positive X-axis pointing forward and the Y-axis perpendicular to the X-axis. The longitudinal movement distance is the displacement component along the X-axis, and the lateral movement distance is the displacement component along the Y-axis. Therefore, the longitudinal movement distance is related to the vehicle's travel distance and its yaw angle.

[0023] Based on this, in some embodiments, the above step S100 may include: predicting the yaw angle of the vehicle based on the yaw rate in the motion state data; predicting the vehicle's travel distance based on the vehicle speed in the motion state data; if the yaw angle is zero, determining that the longitudinal movement distance is equal to the travel distance; if the yaw angle is non-zero, determining the longitudinal movement distance based on the yaw angle, the travel distance, and the lane line polynomial equation.

[0024] Among them, motion state data is the data of the vehicle in the current calculation cycle of planning and control. Therefore, the yaw angle, driving distance and longitudinal movement distance predicted based on the motion state data all correspond to the current calculation cycle.

[0025] The yaw angle can be calculated from the yaw rate in the motion state data. In some embodiments, the product of the yaw rate and the cycle duration of the planning control can be used as the yaw angle. For example, if the cycle duration of the planning control is 20 milliseconds and the yaw rate is in degrees per second, then the yaw angle (in degrees) can be calculated using the following formula 1.

[0026] Formula 1: yaw = 0.02 × yawrate.

[0027] The travel distance can be calculated based on the vehicle speed in the motion state data. In some embodiments, the travel distance can be obtained by multiplying the vehicle speed by the cycle duration of the planning control. For example, if the cycle duration of the planning control is 20 milliseconds and the vehicle speed is in meters per second, the travel distance (in meters) can be calculated using the following formula 2.

[0028] Formula 2: distance = 0.02 × speed.

[0029] The yaw angle can be used to determine whether a vehicle is traveling in a straight line, thus determining the longitudinal distance traveled. If the yaw angle is zero, it indicates that the vehicle is traveling in a straight line, and the longitudinal distance traveled is equal to the distance traveled. If the yaw angle is non-zero, it indicates that the vehicle is not traveling in a straight line, and the longitudinal distance traveled needs to be further determined by combining the yaw angle, the distance traveled, and the lane line polynomial equation.

[0030] In some embodiments, determining the longitudinal movement distance based on the yaw angle, the driving distance, and the lane line polynomial equation may include: determining an initial value for the longitudinal movement distance based on the yaw angle, the driving distance, and the constants in the lane line polynomial equation; substituting the initial value of the longitudinal movement distance into the lane line polynomial equation to obtain an initial value for the lateral movement distance; and iteratively updating the longitudinal movement distance based on the yaw angle, the lane line polynomial equation, the initial value of the longitudinal movement distance, and the initial value of the lateral movement distance.

[0031] This application embodiment can perform perception fitting on the lane lines on both the left and right sides of the lane where the vehicle is located. Taking the lane line polynomial equation as a cubic polynomial equation as an example, the lane line polynomial equation can be as shown in Formula 3 below.

[0032] Formula 3: Y = coeff0 + coeff1 × X +coeff2 × X^2 + coeff3 × X^3.

[0033] Where X represents the distance in front of the vehicle, and Y represents the lateral distance from the lane line at point X in front of the vehicle to the center of the vehicle's rear axle; coeff0 is a constant in the lane line polynomial equation, which represents the distance between the center of the vehicle's rear axle and the corresponding side lane line; coeff1 is the coefficient of the first term in the lane line polynomial equation, which represents the tangent of the angle between the vehicle and the corresponding side lane line; coeff2 is the coefficient of the second term in the lane line polynomial equation, which represents the curvature of the corresponding side lane line; and coeff3 is the coefficient of the third term in the lane line polynomial equation, which represents the rate of change of curvature of the corresponding side lane line. Therefore, lane line coefficients include constants and multi-term coefficients (such as first-term, second-term, and third-term coefficients).

[0034] The initial value of the longitudinal movement distance can be determined based on the yaw angle, the travel distance, and the constants in the lane line polynomial equation. For example, combining Equations 1 to 3 above, the initial value of the longitudinal movement distance X_Length can be calculated using Equation 4 below.

[0035] Formula 4: X_Length = ( distance / (-yaw) - coeff0 ) × tan(-yaw).

[0036] Then, the initial value of the longitudinal movement distance is substituted into the lane line polynomial equation to obtain the initial value of the lateral movement distance. For example, combining Equations 3 and 4 above, the initial value of the lateral movement distance Y_Length can be calculated by Equation 5 below.

[0037] Formula 5: Y_Length = coeff0 + coeff1 × X_Length + coeff2 × X_Length^2 +coeff3 × X_Length^3.

[0038] Since there is a cycle difference between the update cycle of perception fitting and the calculation cycle of planning and control, in order to improve the accuracy of longitudinal movement distance, this embodiment of the application performs iterative updates on the longitudinal movement distance for multiple prediction cycles based on the yaw angle, the lane polynomial equation, the initial value of the longitudinal movement distance, and the initial value of the lateral movement distance, so as to verify whether the corresponding lateral movement distance approaches the lateral movement distance obtained by the lane polynomial equation.

[0039] In some embodiments, the iterative update of the longitudinal movement distance based on the yaw angle, the lane polynomial equation, the initial value of the longitudinal movement distance, and the initial value of the lateral movement distance may include: determining the parameter value of the longitudinal movement distance in the (i+1)th iteration based on the parameter values ​​of the yaw angle, the longitudinal movement distance in the i-th iteration, and the parameter value of the lateral movement distance in the i-th iteration; substituting the parameter value of the longitudinal movement distance in the (i+1)th iteration into the lane polynomial equation to obtain the parameter value of the lateral movement distance in the (i+1)th iteration; if the lateral movement distance is updated in the (i+1)th iteration... If the difference between the new parameter value and the parameter value updated in the ith iteration is less than zero, then the parameter value for the longitudinal movement distance is determined to be the parameter value updated in the (i+1)th iteration. If the difference between the parameter value updated in the (i+1)th iteration and the parameter value updated in the ith iteration is greater than or equal to zero, then let i equal i+1, and start executing again from the step of determining the parameter value updated in the (i+1)th iteration based on the yaw angle, the parameter value updated in the ith iteration for the longitudinal movement distance, and the parameter value updated in the ith iteration for the lateral movement distance.

[0040] Where i is an integer greater than or equal to 0. This application embodiment performs iterative updates on the longitudinal and lateral movement distances over multiple prediction cycles, but the specific number of iterations is not limited. In practical applications, it can be arbitrarily and flexibly set by combining the update cycle of perception fitting and the calculation cycle of planning control. For example, if the update cycle of perception fitting is 100 milliseconds and the calculation cycle of planning control is 20 milliseconds, then iterative updates can be performed within 5 prediction cycles. It should be understood that the parameter value of the longitudinal movement distance in the 0th iteration update is equal to the initial value of the longitudinal movement distance, and the parameter value of the lateral movement distance in the 0th iteration update is equal to the initial value of the lateral movement distance.

[0041] Based on the parameter values ​​of the yaw angle, the longitudinal movement distance updated in the i-th iteration, and the parameter value of the lateral movement distance updated in the i-th iteration, the parameter value of the longitudinal movement distance updated in the (i+1)-th iteration can be determined. For example, if the yaw angle is yaw, the parameter value of the longitudinal movement distance updated in the i-th iteration is X_Length, and the parameter value of the lateral movement distance updated in the i-th iteration is Y_Length, then the parameter value of the longitudinal movement distance updated in the (i+1)-th iteration can be calculated using the following formula 6.

[0042] Formula 6: X_Length_new = X_Length + Y_Length × tan(yaw).

[0043] Substituting the parameter value of the longitudinal movement distance updated in the (i+1)th iteration into the lane line polynomial equation, the parameter value of the lateral movement distance updated in the (i+1)th iteration can be obtained. For example, combining with Equation 6 above, the parameter value Y_Length_new of the lateral movement distance updated in the (i+1)th iteration can be calculated by Equation 7 below.

[0044] Formula 7: Y_Length_new = coeff0 + coeff1 × X_Length_new + coeff2 × X_Length_new ^2 + coeff3 × X_Length_new ^3.

[0045] Then, the lateral movement distance can be verified by calculating the difference between the parameter value of the lateral movement distance updated in the (i+1)th iteration and the parameter value of the lateral movement distance updated in the ith iteration, and determining whether the difference is less than zero. For example, combining Equations 5 and 7 above, this difference Y_Length_Judge can be calculated by Equation 8 below.

[0046] Formula 8: Y_Length_Judge = Y_Length_new - Y_Length.

[0047] If the difference Y_Length_Judge between the parameter value of the lateral movement distance updated in the (i+1)th iteration and the parameter value of the lateral movement distance updated in the ith iteration is less than zero, then the longitudinal movement distance and the lateral movement distance will no longer be iteratively updated. Instead, the parameter value of the longitudinal movement distance will be determined to be the parameter value of the longitudinal movement distance updated in the (i+1)th iteration, so that the parameter value of the longitudinal movement distance updated in the (i+1)th iteration can be used in subsequent calculation steps.

[0048] If the difference Y_Length_Judge between the parameter value of the lateral movement distance updated in the (i+1)th iteration and the parameter value of the lateral movement distance updated in the ith iteration is greater than or equal to zero, then further iterations of the longitudinal and lateral movement distances are needed. In this case, we can set i to i+1 and start again from the step of determining the parameter value of the longitudinal movement distance updated in the (i+1)th iteration based on the yaw angle, the parameter value of the longitudinal movement distance updated in the ith iteration, and the parameter value of the lateral movement distance updated in the ith iteration, until the difference Y_Length_Judge between the parameter value of the lateral movement distance updated in the (i+1)th iteration and the parameter value of the lateral movement distance updated in the ith iteration is less than zero. It should be noted that there may be a limit to the number of iterations for the iterative updates of the longitudinal and lateral movement distances. For example, in the iterative update of 5 prediction cycles described in the above embodiment, if the updated i is greater than or equal to the number of iterations after setting i to i+1, then the iterative update will stop. Instead, the parameter value of the longitudinal movement distance in the last iteration update will be used as the parameter value of the final output of the longitudinal movement distance, so that the parameter value of the longitudinal movement distance in the last iteration update can be used in subsequent calculation steps.

[0049] In step S200, the lane line coefficients in the lane line polynomial equation are processed using a time-series recursion based on the longitudinal movement distance to obtain a coefficient prediction sequence. This coefficient prediction sequence includes the predicted values ​​of the lane line coefficients over multiple prediction periods; that is, the predicted values ​​of the lane line coefficients over multiple prediction periods can be updated based on the longitudinal movement distance and written into the coefficient prediction sequence. Since the lane line coefficients may include multiple coefficients, such as the aforementioned constants, linear coefficients, quadratic coefficients, and cubic coefficients, each coefficient can be constructed into its own coefficient prediction sequence for easy differentiation.

[0050] The temporal recursive processing refers to the forward recursive prediction of the parameter values ​​of the lane line coefficients. Based on this, in some embodiments, step S200 may include: substituting the longitudinal movement distance into the lane line polynomial equation in the k-th prediction period to obtain the predicted value of the constant in the lane line polynomial equation in the (k+1)-th prediction period; and determining the predicted value of the polynomial coefficients in the lane line polynomial equation in the (k+1)-th prediction period based on the longitudinal movement distance and the derivative of the lane line polynomial equation in the k-th prediction period. Here, the lane line coefficients include constants and polynomial coefficients; k is an integer greater than or equal to 0. Furthermore, the lane line polynomial equation in the 0th prediction period is the lane line polynomial equation fitted by the actual vehicle perception; therefore, the predicted value of the constant in the 0th prediction period is equal to the parameter value of the constant in the lane line polynomial equation fitted by the actual perception, and the predicted value of the polynomial coefficients in the 0th prediction period is equal to the parameter value of the polynomial coefficients in the lane line polynomial equation fitted by the actual perception.

[0051] This application does not limit the number of loops for the time-series recursive processing of lane line coefficients. In practical applications, the number of loops can be arbitrarily and flexibly set in combination with the update cycle of perception fitting and the calculation cycle of planning control. For example, if the update cycle of perception fitting is 100 milliseconds and the calculation cycle of planning control is 20 milliseconds, then iterative updates can be performed within 5, 8, 10, or 15 prediction cycles to cover the cycle difference between the update cycle of perception fitting and the calculation cycle of planning control.

[0052] For example, consider iterative updates of lane line coefficients over 15 prediction periods. If the lane line polynomial equation actually perceived by the vehicle is as shown in Formula 3 above, then the lane line coefficients can include a constant coeff0, a linear coefficient coeff1, a quadratic coefficient coeff2, and a cubic coefficient coeff3. For the constant coeff0, a coefficient prediction sequence coeff0_update[0~14] can be constructed; for the linear coefficient coeff1, a coefficient prediction sequence coeff1_update[0~14] can be constructed; for the quadratic coefficient coeff2, a coefficient prediction sequence coeff2_update[0~14] can be constructed; and for the cubic coefficient coeff3, a coefficient prediction sequence coeff3_update[0~14] can be constructed. Taking the calculation of the lane line coefficient in the (k+1)th prediction period as an example, the predicted values ​​of the constant coeff0, the linear coefficient coeff1, the quadratic coefficient coeff2, and the cubic coefficient coeff3 in the kth prediction period are coeff0_update[k], coeff1_update[k], coeff2_update[k], and coeff3_update[k], respectively. Based on this, the predicted value of the constant coeff0 in the (k+1)th prediction period, coeff0_update[k+1], can be calculated by the following formula 9. The predicted value of the linear term coefficient coeff1 in the (k+1)th prediction period, coeff1_update[k+1], can be calculated using Equation 10 below, i.e., by taking the derivative of Equation 9 below; the predicted value of the quadratic term coefficient coeff2 in the (k+1)th prediction period, coeff2_update[k+1], can be calculated using Equation 11 below, i.e., by taking the derivative of Equation 10 below; the predicted value of the cubic term coefficient coeff3 in the (k+1)th prediction period, coeff3_update[k+1], can be calculated using Equation 12 below, assuming that the rate of curvature change remains constant over a short period. Here, X_Length_final in Equations 9 to 11 represents the final longitudinal movement distance output in step S100 above.

[0053] Formula 9: coeff0_update[k+1] = coeff0_update[k]+ coeff1_update[k] × X_Length_final + coeff2_update[k] × X_Length_final^2 + coeff3_update[k] × X_Length_final^3.

[0054] Formula 10: coeff1_update[k+1] = 3 × X_Length_final ^2 × coeff3_update[k]+ 2 ×

[0055] Formula 11: coeff2_update [k+1] = 6 × X_Length_final × coeff3_update[k]+coeff2_update[k].

[0056] Formula 12: coeff3_update[k+1] = coeff3_update[k].

[0057] In step S300, the predicted values ​​of lane line coefficients in the coefficient prediction sequence over multiple prediction periods are smoothed to obtain the target values ​​of the lane line coefficients. A smoothing coefficient can be introduced during the smoothing process to fuse the predicted values ​​of the lane line coefficients over multiple prediction periods based on the smoothing coefficient. The smoothing coefficient can be corrected based on the yaw angle to distinguish between straight-line driving and non-straight-line driving.

[0058] Based on this, in some embodiments, step S300 may include: determining a smoothing coefficient sequence based on the vehicle's yaw angle, the smoothing coefficient sequence including correction values ​​of the smoothing coefficient over multiple prediction periods; performing statistical processing on the correction values ​​of the smoothing coefficient over multiple prediction periods to obtain a cumulative value of the smoothing coefficient; performing weighted summation processing on the predicted values ​​of the lane line coefficient over multiple prediction periods based on the correction values ​​of the smoothing coefficient over multiple prediction periods to obtain a fused value of the lane line coefficient; and determining a target value of the lane line coefficient based on the fused value of the lane line coefficient and the cumulative value of the smoothing coefficient.

[0059] This application does not limit the method of assigning values ​​to the smoothing coefficient, and it can be flexibly set according to actual needs in practical applications. In some embodiments, the correction values ​​of the smoothing coefficient in multiple prediction periods may be the same or different; the smoothing coefficients corresponding to different coefficients in the lane line coefficients may be the same or different.

[0060] For example, the initial value of the smoothing coefficient is smooth, and the range of the initial value can be [20, 100], such as the initial value smooth can be 40; the yaw angle is yaw, and the correction value of the smoothing coefficient smooth_coeffs can be calculated by the following formula 13.

[0061] Formula 13: smooth_coeffs = max (0.0, 1 – yaw × smooth) × 0.9.

[0062] Taking the iterative update of lane line coefficients over 15 prediction periods as an example, a smoothing coefficient sequence smoothcoeff [0~14] can be constructed. The correction value of the smoothing coefficient can be the same across multiple prediction periods, so smoothcoeff [0~14] can all be assigned the value of smooth_coeffs calculated by formula 13. In practical applications, the correction value of the smoothing coefficient can also be different across multiple prediction periods. For example, a decay factor can be introduced to reflect time-varying decay.

[0063] Since the lane line polynomial equation is a smooth curve, during smoothing, the same smoothing factor can be set for each coefficient in the lane line coefficients to avoid deformation of the final smoothed lane line due to different smoothing factors. Based on this, the smoothing coefficient sequences corresponding to the above constant coeff0, the linear coefficient coeff1, the quadratic coefficient coeff2, and the cubic coefficient coeff3 are coeff0_scale[0~14], coeff1_scale[0~14], coeff2_scale[0~14], and coeff3_scale[0~14], respectively. They can be the same, and the smoothing coefficient sequences corresponding to each coefficient are assigned values ​​as shown in Formulas 14 to 17 below.

[0064] Formula 14: coeff0_scale[0~14] = smoothcoeff [0~14].

[0065] Formula 15: coeff1_scale[0~14] = smoothcoeff [0~14].

[0066] Formula 16: coeff2_scale[0~14] = smoothcoeff [0~14].

[0067] Formula 17: coeff3_scale[0~14] = smoothcoeff [0~14].

[0068] Then, the correction values ​​of the smoothing coefficients over multiple prediction periods are statistically processed to obtain the cumulative value of the smoothing coefficients. For example, based on the smoothing coefficient sequences corresponding to the above coefficients, the correction values ​​of the smoothing coefficients over multiple prediction periods in the smoothing coefficient sequence can be summed to obtain the cumulative value of the smoothing coefficients. Combining formulas 14 to 17 above, the cumulative value of the smoothing coefficients corresponding to each coefficient is calculated using formulas 18 to 21 below.

[0069] Formula 18: coeff0_scale_final = Σ_{k=0}^{14} smoothcoeff[k].

[0070] Formula 19: coeff1_scale_final = Σ_{k=0}^{14} smoothcoeff[k].

[0071] Formula 20: coeff2_scale_final = Σ_{k=0}^{14} smoothcoeff[k].

[0072] Formula 21: coeff3_scale_final = Σ_{k=0}^{14} smoothcoeff[k].

[0073] Furthermore, based on the correction values ​​of the smoothing coefficient over multiple prediction periods, the predicted values ​​of the lane line coefficient over multiple prediction periods are weighted and summed to obtain the fusion value of the lane line coefficient. That is, the fusion value of the lane line coefficient is obtained by summing the products of the predicted values ​​of the lane line coefficient and the correction values ​​of the smoothing coefficient over multiple prediction periods. For example, the lane line coefficient includes the aforementioned constant coeff0, the linear coefficient coeff1, the quadratic coefficient coeff2, and the cubic coefficient coeff3, and combined with the aforementioned coefficient prediction sequence: coeff0_update[0~14], coeff1_update[0~14], coeff2_update[0~14], coeff3_update[0~14], and the smoothing coefficient sequence coeff0_scale[0~14], coeff1_scale[0~14], coeff2_scale[0~14], coeff3_scale[0~14], the fusion value of each coefficient in the lane line coefficient can be calculated by the following formulas 22 to 25.

[0074] Formula 22: coeff0_update_final = coeff0_update[0~14] × coeff0_scale[0~14].

[0075] Formula 23: coeff1_update_final = coeff1_update[0~14] × coeff1_scale[0~14].

[0076] Formula 24: coeff2_update_final = coeff2_update[0~14] × coeff2_scale[0~14].

[0077] Formula 25: coeff3_update_final = coeff3_update[0~14] × coeff3_scale[0~14].

[0078] Based on the fusion value of the lane line coefficients and the cumulative value of the smoothing coefficients, the target value of the lane line coefficients can be finally determined. For example, the target value of the lane line coefficients can be obtained by dividing the fusion value of the lane line coefficients by the cumulative value of the smoothing coefficients. For example, the lane line coefficients include the aforementioned constant coeff0, the linear coefficient coeff1, the quadratic coefficient coeff2, and the cubic coefficient coeff3, and combined with the cumulative values ​​of the smoothing coefficients coeff0_scale_final, coeff1_scale_final, coeff2_scale_final, and coeff3_scale_final calculated by formulas 18 to 21, and the fusion values ​​of the lane line coefficients coeff0_update_final, coeff0_update_final, coeff0_update_final, and coeff0_update_final calculated by formulas 22 to 25, the target value of each coefficient in the lane line coefficients can be calculated by formulas 26 to 29 below.

[0079] Equation 26: coeff0_smooth = coeff0_update_final / coeff0_scale_final.

[0080] Equation 27: coeff1_smooth = coeff1_update_final / coeff1_scale_final.

[0081] Equation 28: coeff2_smooth = coeff2_update_final / coeff2_scale_final.

[0082] Equation 29: coeff3_smooth = coeff3_update_final / coeff3_scale_final.

[0083] In step S400, the lane line polynomial equation is updated according to the target value of the lane line coefficients to obtain the predicted lane line. That is, the parameter values ​​of the lane line coefficients in the lane line polynomial equation fitted by the actual perception of the vehicle are replaced with the target values ​​of the lane line coefficients, thereby obtaining the final updated lane line polynomial equation, which is used to indicate the predicted lane line. For example, combining Equation 3 and Equations 26 to 29 above, the final updated lane line polynomial equation can be as shown in Equation 30 below.

[0084] Formula 30: Y = coeff0_smooth + coeff1_smooth × X + coeff2_smooth × X^2+ coeff3_smooth × X^3.

[0085] In summary, the technical solution provided in this application predicts the longitudinal movement distance of the vehicle based on the vehicle's motion state data and the lane line polynomial equation; based on the longitudinal movement distance, the lane line coefficients in the lane line polynomial equation are processed by time-series recursion to obtain a coefficient prediction sequence; the predicted values ​​of the lane line coefficients in the coefficient prediction sequence within multiple prediction periods are smoothed to obtain the target values ​​of the lane line coefficients; and the lane line polynomial equation is updated based on the target values ​​of the lane line coefficients to obtain the predicted lane lines. This application's embodiment predicts the parameter values ​​of the lane line coefficients within multiple future prediction periods based on the vehicle's real-time motion state data and the lane line polynomial equation fitted by the vehicle's actual perception. This allows for multiple intelligent updates of the lane line coefficient parameter values ​​within a longer perception fitting update period, making the changes in the lane line coefficients more continuous and smooth. This avoids control jitter and command jumps caused by the planning control's calculation cycle being shorter than the perception fitting update cycle, ensuring the smoothness of the vehicle control process and effectively improving the vehicle control performance. Furthermore, this embodiment uses multiple predicted values ​​for comprehensive calculation instead of a single prediction, avoiding the random errors of a single prediction. While preserving the true geometric trend of the lane lines, it averages perceived jitter and abnormal jumps in the time domain, effectively improving the accuracy of the lane line coefficient and achieving real-time accurate prediction of lane lines. In addition, this embodiment implements forward prediction of the lane line coefficient instead of backward smoothing, eliminating the inherent phase lag effect of traditional low-pass filtering, ensuring the timeliness of vehicle control, and further improving the vehicle control effect.

[0086] The lane prediction method provided in this application will be described below with an example.

[0087] Steering Lane Keeping Assist (SLKA) is an intelligent driving assistance system defined as follows: when the vehicle's onboard sensors detect that the vehicle has unintentionally deviated from its lane, the lane keeping system will automatically correct the vehicle's driving trajectory to prevent the vehicle from leaving its lane and causing danger. The lane keeping system is mainly applied in the following scenarios: (1) when the vehicle deviates from its lane and is about to cross the lane line, where the lane line refers to the lane line of the vehicle's lane. The lane line types include solid lines, dashed lines, and solid-dashed lines. Common lane lines include continuous straight lane lines, straight intermittent lane lines, dashed intermittent lane lines, double solid lines, etc.; (2) when the vehicle deviates from its lane, the road type can include straight roads, left and right curves, straight roads leading into curves, curves leading into straight roads, etc.

[0088] In related technologies, lane keeping systems detect lane line scenarios during the vehicle correction process triggered by the function. However, they rely solely on the perceived and fitted lane line information without any smoothing. If lane lines change abruptly in scenarios such as ramps, this may lead to premature and erroneous function triggering or unclear lane lines, resulting in discontinuous lateral control of the vehicle. Furthermore, these lane keeping systems often use smoothing methods such as low-pass filtering for the perceived lane lines, which can cause a significant delay between the smoothed lane line information and the actual lane lines. This can lead to delayed function activation or delayed control in cases of high lateral speed, resulting in situations such as vehicle crossing the lane lines, vehicle overtaking the lane lines, or uneven control.

[0089] In view of this, the embodiments of this application predict the lateral and longitudinal movement distance of the vehicle based on the vehicle's current speed and yaw rate, calculate the change of the lane line coefficient within the prediction time based on the lateral and longitudinal movement distance, set the lane line smoothing coefficient based on the vehicle's yaw angle, and smooth the predicted lane line coefficient change.

[0090] Please see Figure 2 , Figure 2 This is a flowchart of another lane prediction method provided in an embodiment of this application. For example... Figure 2 As shown, the lane line prediction method may include the following steps S201 to S210.

[0091] Step S201: Acquire real-time motion state data of the vehicle. This motion state data may include, for example, yaw rate and vehicle speed.

[0092] Step S202: Obtain the lane line polynomial equation fitted by the vehicle's actual perception. This lane line polynomial equation can be, for example, a cubic polynomial equation, as shown in Formula 3 above. During the vehicle's actual perception fitting, lane line confidence, total lane line length, and lane line following status can also be output. The lane line confidence ranges from 0 to 100, indicating whether the current lane line is blurry or clear; generally, a higher value indicates a clearer lane line. The total lane line length refers to the distance between the center of the vehicle's rear axle and the actual end point of the lane line captured by the camera. The lane line following status indicates whether the lane line is real or a predicted fit.

[0093] Step S203: Predict the longitudinal movement distance of the vehicle based on the motion state data and the lane line polynomial equation. Specifically, the yaw angle of the vehicle can be predicted based on the yaw rate in the motion state data, as shown in Formula 1 above; the vehicle speed can be predicted based on the vehicle speed in the motion state data, as shown in Formula 2 above; if the yaw angle is zero, the longitudinal movement distance is determined to be equal to the travel distance; if the yaw angle is non-zero, the longitudinal movement distance is determined based on the yaw angle, travel distance, and lane line polynomial equation, and the longitudinal movement distance is iteratively updated, as shown in the above embodiment.

[0094] Step S204: Based on the longitudinal movement distance, perform time-series recursive processing on the lane line coefficients in the lane line polynomial equation to obtain the coefficient prediction sequence. The coefficient prediction sequence corresponding to each coefficient in the lane line coefficients can be shown in Formulas 9 to 12 above.

[0095] Step S205: Determine the smoothing coefficient sequence based on the vehicle's yaw angle. The smoothing coefficient sequence corresponding to each coefficient in the lane line coefficients can be shown in Formulas 13 to 17 above.

[0096] Step S206: Statistically process the correction values ​​of the smoothing coefficients over multiple prediction periods to obtain the cumulative value of the smoothing coefficients. The cumulative value of the smoothing coefficients corresponding to each coefficient in the lane line coefficients can be shown in Formulas 18 to 21 above.

[0097] Step S207: Based on the correction values ​​of the smoothing coefficient over multiple prediction periods, perform a weighted summation of the predicted values ​​of the lane line coefficient over multiple prediction periods to obtain the fused value of the lane line coefficient. The fused value of each coefficient in the lane line coefficient can be shown in Formulas 22 to 25 above.

[0098] Step S208: Determine the target value of the lane line coefficient based on the fusion value of the lane line coefficient and the cumulative value of the smoothing coefficient. The target values ​​for each coefficient in the lane line coefficient can be shown in formulas 26 to 29 above.

[0099] Step S209: Determine whether the total length of the lane lines is greater than a preset threshold. The preset threshold can be the minimum allowable lane length. If the total length of the lane lines is greater than the preset threshold, proceed to step S210; otherwise, proceed to step S201.

[0100] Step S210: Update the lane line polynomial equation based on the target value of the lane line coefficient to obtain the predicted lane line. The parameter values ​​of the lane line coefficient in Formula 3 above can be replaced with the target values ​​of the lane line coefficient calculated by Formulas 26 to 29 above. The final updated lane line polynomial equation can be as shown in Formula 30 above.

[0101] related Figure 2 For further explanation of the steps shown and their beneficial effects, please refer to the above embodiments; they will not be repeated here.

[0102] Based on the same inventive concept, this application also provides a lane prediction system for implementing the lane prediction method described above. The solution provided by this lane prediction system is similar to the implementation described in the above method; therefore, the specific limitations of the one or more lane prediction system embodiments provided below can be found in the limitations of the lane prediction method described above, and will not be repeated here.

[0103] Please see Figure 3 , Figure 3 This is a schematic diagram of a lane prediction system provided in an embodiment of this application. Figure 3 As shown, the lane prediction system 300 may include a lane prediction module 310. The lane prediction module 310 can be used to implement the lane prediction method described in any of the above embodiments.

[0104] Please see Figure 4 , Figure 4 This is a schematic diagram of another lane prediction system provided in an embodiment of this application. For example... Figure 4 As shown, the lane prediction system 300 may include a lane prediction module 310. The lane prediction module 310 can be used to implement the lane prediction method described in any of the above embodiments.

[0105] In some embodiments, such as Figure 4 As shown, the lane prediction system 300 may further include a lane perception module 320. The lane perception module 320 is connected to the lane prediction module 310, and is used to: fit a lane polynomial equation; and send the lane polynomial equation to the lane prediction module 310. In some embodiments, the lane perception module 320 may include a data acquisition unit and a perception fitting unit. The data acquisition unit may be used to acquire road images of the road where the vehicle is located; for example, the data acquisition unit may be implemented as a camera to acquire road images ahead of the vehicle. The perception fitting unit may perceive and fit a lane polynomial equation based on the road images acquired by the data acquisition unit.

[0106] In some embodiments, such as Figure 4As shown, the lane prediction system 300 may further include a chassis module 330. The chassis module 330 is connected to the lane prediction module 310, and is used to: acquire vehicle motion state data; and send the motion state data to the lane prediction module 310. In some embodiments, the chassis module 330 may include sensors, cameras, gyroscopes, etc., which are not limited in this application embodiment.

[0107] In some embodiments, such as Figure 4 As shown, the lane prediction system 300 may further include a planning and control module 340. The planning and control module 340 is connected to the lane prediction module 310 and can perform planning and control of the vehicle based on the lane polynomial equation finally output by the lane prediction module 310, such as realizing the lane keeping function and planning the path, etc. The embodiments of this application do not limit this.

[0108] Each module or unit in the aforementioned lane prediction system can be implemented entirely or partially through software, hardware, or a combination thereof. These modules or units can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the operations corresponding to each module.

[0109] Accordingly, this application also provides an electronic device. This electronic device includes a memory, a processor, and a computer program or instructions stored in the memory and executable on the processor. When the processor executes the computer program or instructions, it implements the steps of the lane prediction method described above. Since the lane prediction method has been described in detail above, it will not be repeated here.

[0110] Accordingly, embodiments of this application also provide a computer-readable storage medium storing a computer program or instructions thereon, which, when executed by a processor, implements the steps of the lane prediction method described above. Since the lane prediction method has been described in detail above, it will not be repeated here.

[0111] Accordingly, embodiments of this application also provide a vehicle 10, such as Figure 5 As shown. The vehicle 10 may include the aforementioned electronic equipment or the aforementioned lane prediction system. The vehicle has all the beneficial effects of the aforementioned electronic equipment and lane prediction system, which will not be elaborated here.

[0112] In the above embodiments, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0113] Those skilled in the art will understand that the accompanying drawings are merely schematic diagrams of exemplary embodiments and may not be to scale. The modules or processes shown in the drawings are not necessarily essential for implementing this application and therefore should not be used to limit the scope of protection of this application.

[0114] The lane prediction method, lane prediction system, and vehicle provided in the embodiments of this application have been described in detail above, and specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the technical solutions and core ideas of this application. Those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application.

Claims

1. A lane line prediction method, characterized in that, The method includes: Based on the vehicle's motion state data and the lane line polynomial equation, the longitudinal movement distance of the vehicle is predicted. Based on the longitudinal movement distance, the lane line coefficients in the lane line polynomial equation are subjected to time-series recursive processing to obtain a coefficient prediction sequence; wherein, the coefficient prediction sequence includes the predicted values ​​of the lane line coefficients in multiple prediction periods; The predicted values ​​of the lane line coefficient over multiple prediction periods are smoothed to obtain the target value of the lane line coefficient. The lane line polynomial equation is updated based on the target value of the lane line coefficient to obtain the predicted lane line.

2. The method according to claim 1, characterized in that, The step of performing time-series recursive processing on the lane line coefficients in the lane line polynomial equation based on the longitudinal movement distance to obtain a coefficient prediction sequence includes: Substituting the longitudinal movement distance into the lane polynomial equation in the k-th prediction period, the predicted value of the constant in the lane polynomial equation in the (k+1)-th prediction period is obtained. Based on the longitudinal movement distance and the derivative of the lane polynomial equation in the k-th prediction period, the predicted value of the polynomial coefficients in the lane polynomial equation in the (k+1)-th prediction period is determined. Wherein, the lane line coefficient includes the constant and the polynomial coefficient; and k is an integer greater than or equal to 0.

3. The method according to claim 1, characterized in that, The step of smoothing the predicted values ​​of the lane line coefficient over multiple prediction periods to obtain the target value of the lane line coefficient includes: Based on the yaw angle of the vehicle, a smoothing coefficient sequence is determined; wherein, the smoothing coefficient sequence includes correction values ​​of the smoothing coefficients within multiple prediction periods; The correction values ​​of the smoothing coefficient over multiple prediction periods are statistically processed to obtain the cumulative value of the smoothing coefficient; Based on the correction values ​​of the smoothing coefficient over multiple prediction periods, the predicted values ​​of the lane line coefficient over multiple prediction periods are weighted and summed to obtain the fused value of the lane line coefficient. The target value of the lane line coefficient is determined based on the fusion value of the lane line coefficient and the cumulative value of the smoothing coefficient.

4. The method according to any one of claims 1 to 3, characterized in that, The step of predicting the longitudinal movement distance of the vehicle based on the vehicle's motion state data and the lane line polynomial equation includes: Based on the yaw rate in the motion state data, predict the yaw angle of the vehicle; Based on the vehicle speed in the motion state data, predict the vehicle's travel distance; If the yaw angle is zero, then the longitudinal movement distance is determined to be equal to the travel distance; If the yaw angle is non-zero, the longitudinal movement distance is determined based on the yaw angle, the travel distance, and the lane line polynomial equation.

5. The method according to claim 4, characterized in that, Determining the longitudinal movement distance based on the yaw angle, the travel distance, and the lane line polynomial equation includes: The initial value of the longitudinal movement distance is determined based on the yaw angle, the travel distance, and the constants in the lane line polynomial equation. Substituting the initial value of the longitudinal movement distance into the lane polynomial equation, we obtain the initial value of the lateral movement distance. The longitudinal movement distance is iteratively updated based on the yaw angle, the lane line polynomial equation, the initial value of the longitudinal movement distance, and the initial value of the lateral movement distance.

6. The method according to claim 5, characterized in that, The step of iteratively updating the longitudinal movement distance based on the yaw angle, the lane line polynomial equation, the initial value of the longitudinal movement distance, and the initial value of the lateral movement distance includes: Based on the yaw angle, the parameter values ​​of the longitudinal movement distance updated in the i-th iteration, and the parameter values ​​of the lateral movement distance updated in the i-th iteration, the parameter value of the longitudinal movement distance updated in the (i+1)-th iteration is determined; where i is an integer greater than or equal to 0, the parameter value of the longitudinal movement distance updated in the 0-th iteration is equal to the initial value of the longitudinal movement distance, and the parameter value of the lateral movement distance updated in the 0-th iteration is equal to the initial value of the lateral movement distance; Substituting the parameter value of the longitudinal movement distance updated in the (i+1)th iteration into the lane polynomial equation, we obtain the parameter value of the lateral movement distance updated in the (i+1)th iteration. If the difference between the parameter value of the lateral movement distance updated in the (i+1)th iteration and the parameter value of the lateral movement distance updated in the ith iteration is less than zero, then the parameter value of the longitudinal movement distance is determined to be the parameter value of the longitudinal movement distance updated in the (i+1)th iteration. If the difference between the parameter value of the lateral movement distance updated in the (i+1)th iteration and the parameter value of the lateral movement distance updated in the i-th iteration is greater than or equal to zero, then let i equal i+1, and start executing again from the step of determining the parameter value of the longitudinal movement distance updated in the (i+1)th iteration based on the yaw angle, the parameter value of the longitudinal movement distance updated in the i-th iteration, and the parameter value of the lateral movement distance updated in the i-th iteration.

7. A lane line prediction system, characterized in that, The system includes: A lane prediction module for performing the lane prediction method as described in any one of claims 1 to 6.

8. The system according to claim 7, characterized in that, The system also includes: The lane line perception module, connected to the lane line prediction module, is used to: fit a lane line polynomial equation; and send the lane line polynomial equation to the lane line prediction module.

9. The system according to claim 7, characterized in that, The system also includes: The chassis module, connected to the lane prediction module, is used to send the vehicle's motion state data to the lane prediction module.

10. A vehicle, characterized in that, The vehicle includes the lane prediction system as described in any one of claims 7 to 9.

Images