A vehicle unified path planning method based on position relationship of entrance and exit lanes of intersections

Abstract

The application discloses a vehicle unified path planning method based on position relationship between entrance and exit roads of an intersection, and is used for describing a path of a vehicle through the intersection. The application describes the position relationship between the entrance road and the exit road through which the vehicle passes by establishing a coordinate system based on the entrance road of the intersection; uses a third-order Bezier curve to describe the path of the vehicle through the intersection, establishes a correlation between control parameters of the Bezier curve fitting path and position relationship parameters of the entrance road and the exit road of the intersection, and thus constructs the vehicle unified path planning method based on the position relationship between the entrance road and the exit road of the intersection. Compared with the existing model, the application can directly determine the vehicle path through the physical environment characteristics of the intersection, is not limited to a single turning direction and a geometric shape of the intersection, and can support accurate prediction of the motion of the artificial driving vehicle and the path planning of the automatic driving vehicle.

Description

Technical Field

[0001] The present invention relates to the field of traffic safety, and in particular to a unified vehicle path planning method based on the positional relationship between entrances and exits of an intersection. Background Art

[0002] Intersections are crucial nodes in urban road networks, serving as a hub and distributor for the multiple traffic flows converging there. Because vehicles change direction at intersections, their paths merge, collide, and intersect, leading to numerous conflict points and frequent traffic accidents. Statistics in my country indicate that traffic accidents at intersections account for approximately 30% of all traffic accidents. Unlike ordinary roads, intersections do not have lane divisions. Vehicle paths through intersections are not restricted by lanes, allowing drivers to take a wide variety of different routes through the intersection. Therefore, vehicle paths are a key factor in determining safe driving behavior at intersections.

[0003] Many researchers have studied vehicle path modeling within intersections, most focusing on left- and right-turn paths and the factors influencing them. On one hand, researchers such as Alhajyaseen Wael KM et al., Ma Zian et al., Zhou Donghao et al., and Abdeljaber Osama et al. modeled the paths and trajectories of left-turning vehicles in right-hand traffic, or right-turning vehicles in left-hand traffic, to prevent collisions between turning vehicles and oncoming vehicles. On the other hand, researchers such as Asano Miho et al., Alhajyaseen Wael KM et al., Dias Charitha et al., and Chen Peng et al. studied the paths and trajectories of left-turning vehicles in left-hand traffic, or right-turning vehicles in right-hand traffic, to prevent collisions between turning vehicles and pedestrians and non-motorized vehicles. These studies have identified factors influencing vehicle paths or trajectories at intersections, including intersection layout characteristics (e.g., intersection angle, turning radius, number of exit lanes), vehicle motion conditions (e.g., speed, vehicle type, turning angle), and traffic flow conditions (e.g., volume). These studies usually model the paths or trajectories of vehicles turning in different directions. Although the established models can more accurately describe the paths or trajectories of each different turning direction, these studies are limited to turning vehicles in a single direction and lack universality in practical applications.

[0004] Some researchers have used different curves or curve combinations to model turning paths for vehicles at intersections. For example, Asano Miho et al. and Alhajyaseen Wael KM et al. applied an approximate method based on Euler spirals to model the trajectory of individual left-turning vehicles, where each trajectory was fitted by an Euler spiral and a circular curve. Subsequently, Alhajyaseen Wael KM et al. used a combination of Euler spirals and circular curves to estimate the paths of left-turning vehicles and constructed parameter distributions for the Euler spirals and circular curves as functions of intersection geometry, vehicle speed, and vehicle type. However, this curve combination results in each component of the entire path being explained by a different variable, increasing the complexity of practical applications. Furthermore, some researchers have used spline curves and Bezier curves to plan the paths of autonomous vehicles. These studies typically focus on vehicles turning at regular intersections. However, due to urban geography, irregular intersections are extremely common on real roads, such as diagonal intersections, roundabouts, T-intersections, and Y-intersections. Currently, research on modeling vehicle paths at irregular intersections remains relatively lacking. The Bezier curve is determined by movable control points. The curve can move regularly as the positions of the control points change, which has the advantage of smoothing the path.

[0005] Therefore, the present invention describes the positional relationship between the entrance and exit lanes of the intersection, uses the Bezier curve to fit the vehicle path, and establishes a correlation between the positional relationship parameters of the entrance and exit lanes and the Bezier curve control parameters, thereby constructing a unified vehicle driving path planning method that is not limited to a single turning direction and regular intersections. Summary of the Invention

[0006] To address the shortcomings of existing models, the present invention aims to provide a vehicle path planning method that can uniformly describe the path of a vehicle through an intersection and is applicable to different driving directions and intersection geometries. The method specifically includes the following steps:

[0007] Step 1: Based on the positional relationship between the entrance and exit roads of the intersection, a coordinate system based on the entrance road of the intersection is established;

[0008] Step 2, based on the intersection entrance road-based coordinate system established in step 1, describing the positional relationship between the intersection entrance road and the exit road;

[0009] Step 3: Based on the Bezier curve, describe the path of the vehicle passing through the intersection;

[0010] Step 4: Based on the positional relationship between the entrance and exit lanes of the intersection obtained in step 2 and the path curve of the vehicle passing through the intersection obtained in step 3, a unified vehicle path planning method based on the positional relationship between the entrance and exit lanes of the intersection is established;

[0011] Furthermore, the coordinate system based on the intersection entrance road described in step 1 is based on the intersection entrance road, the midpoint of the stop line of the intersection entrance road is the origin of the coordinate system, the straight line where the stop line is located is the X-axis of the coordinate system, and the perpendicular line of the stop line is the Y-axis of the coordinate system;

[0012] Furthermore, the positional relationship between the entrance and exit roads of the intersection described in step 2 is represented by parameters w, h, and α, where w and h represent the horizontal and vertical distances between the midpoint of the entrance road and the midpoint of the exit road, respectively, and α represents the complementary angle of the angle between the straight line containing the entrance and exit roads, that is, the angle between the straight line containing the exit road and the Y axis.

[0013] Furthermore, the path of the vehicle passing through the intersection in step 3 is described by a third-order Bezier curve, which is composed of control points P o 、P d , P1 and P2 control, where P o Indicates the starting point of the vehicle path, with coordinates (0,0); P d Indicates the end point of the vehicle path, with coordinates (w, h); P1 controls the shape of the vehicle's path after entering the intersection, located on the extension line of the intersection entrance center, with coordinates (0, L1); P2 controls the shape of the vehicle's path when it is about to leave the intersection, located on the extension line of the intersection exit center, with coordinates (w-L2sinα, h-L2cosα); L1 represents the control point P o The distance between L2 and P1, L2 represents the control point P d The distance between P2 and the intersection entrance and exit lanes is calculated. For each set of intersection entrance and exit lanes, the optimal control parameters L1 and L2 of a set of vehicle paths can be obtained through the actual vehicle trajectory data, including the following steps:

[0014] Step 31: extract the world coordinates of the starting point and the end point of the vehicle path. The starting point coordinate is expressed as (x orig,world ,y orig,world ), the end point coordinate is expressed as (x dest,world ,y dest,world );

[0015] Step 32, extracting the vehicle trajectory coordinates of the continuous time series within the starting point and end point coordinate ranges described in step 31, which can be expressed as (x world ,y world ), the coordinates of the trajectory points satisfy:

[0016]

[0017] Step 33, based on the intersection entrance road-based coordinate system established in step 1, measuring the counterclockwise rotation angle γ relative to the world coordinate system;

[0018] Step 34, based on the world coordinates (x orig,world ,y orig,world ), the world coordinates (x world ,y world ), and the coordinate system rotation angle γ obtained in step 33, the world coordinates of the vehicle trajectory point are converted to the coordinates (x, y) based on the intersection entrance road:

[0019]

[0020] Step 35: For each set of vehicle path data for the entrance and exit lanes of the intersection, the Bezier curve control parameters L1 and L2 are optimized using a differential evolution algorithm to solve an optimal Bezier curve that minimizes the distance between the vehicle path and the Bezier curve.

[0021] Furthermore, the unified vehicle path planning method based on the position relationship between the entrance and exit lanes of the intersection described in step 4 can be described by the following expression:

[0022] L1=f1(w,h,α);

[0023] L2=f2(w,h,α);

[0024] When a set of intersection entrance and exit position relationship parameters is given, an optimal Bezier curve controlled by the Bezier curve fitting path control parameters L1 and L2 can be drawn to describe the path of the vehicle through the intersection; the expression of the Bezier curve is as follows:

[0025]

[0026] Where: t is the curve parameter between 0 and 1, n is the number of control points (including the start and end points), Pi is the i-th control point;

[0027] Compared with the existing technology, the beneficial effects of the present invention are: establishing a coordinate system based on the entrance road of the intersection, which can describe the positional relationship between the entrance road and the exit road passed by the vehicle through the intersection; using a third-order Bezier curve to describe the vehicle path, and on this basis constructing a unified intersection vehicle path planning method that is not limited to a single turning direction and intersection geometry, which can determine the vehicle path through the physical environment characteristics of the intersection, which is conducive to the automatic driving vehicle to accurately predict the movement of the manually driven vehicle. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Figure 1 Schematic diagram of a coordinate system based on an intersection entrance road in an embodiment of the present invention;

[0029] Figure 2Schematic diagram of the positional relationship between the entrance and exit roads of an intersection in an embodiment of the present invention;

[0030] Figure 3 A schematic diagram of a vehicle path at an intersection based on a Bezier curve in an embodiment of the present invention;

[0031] Figure 4 This is a schematic diagram of an intersection drawn using the simulation software UC-Win / Road in an embodiment of the present invention;

[0032] Figure 5 Schematic diagram of the driving simulator experiment process in an embodiment of the present invention;

[0033] Figure 6 This is a descriptive statistical feature table of 96 sets of positional relationship parameters between intersection entrance and exit lanes and corresponding Bezier curve control parameters of vehicle paths in an embodiment of the present invention;

[0034] Figure 7 Schematic diagram of vehicle motion video data acquired by drone aerial photography in an embodiment of the present invention;

[0035] Figure 8 25 sets of parameter values ​​of actual paths and their errors with the fitted paths in an embodiment of the present invention;

[0036] Figure 9 Schematic diagram comparing the actual path and the fitted path in an embodiment of the present invention. DETAILED DESCRIPTION

[0037] The present invention is described in detail below with reference to the accompanying drawings, attached tables and examples.

[0038] The present embodiment provides a unified vehicle path planning method based on the position relationship between the entrance and exit lanes of an intersection, comprising the following steps:

[0039] Step 1: Based on the position relationship between the entrance and exit of the intersection, establish Figure 1 The coordinate system shown is based on the intersection entrance road;

[0040] Step 2, based on the intersection entrance road coordinate system established in step 1, is described as follows Figure 2 The positional relationship between the entrance and exit roads of the intersection shown;

[0041] Step 3, based on Bezier curve, is described as follows Figure 3 The paths of vehicles shown through the intersection;

[0042] Step 4: Based on the positional relationship between the entrance and exit lanes of the intersection obtained in step 2 and the path curve of the vehicle passing through the intersection obtained in step 3, a unified vehicle path planning method based on the positional relationship between the entrance and exit lanes of the intersection is established;

[0043] In one embodiment, the coordinate system based on the intersection entrance road described in step 1 is based on the intersection entrance road, the midpoint of the stop line of the intersection entrance road is the origin of the coordinate system, the straight line where the stop line is located is the X-axis of the coordinate system, and the perpendicular line of the stop line is the Y-axis of the coordinate system;

[0044] In one embodiment, the positional relationship between the entrance and exit roads of the intersection described in step 2 is represented by w, h, and α, where w and h represent the horizontal and vertical distances between the midpoint of the entrance road exit and the midpoint of the exit road entrance, respectively, and α represents the supplementary angle of the angle between the straight line on which the entrance and exit roads are located, that is, the angle between the straight line on which the exit road is located and the Y axis. In this embodiment, 96 groups of intersection entrance and exit roads were drawn using the simulation software UC-Win / Road, as shown in the schematic diagram. Figure 4 As shown;

[0045] In one embodiment, the path of the vehicle passing through the intersection in step 3 is described by a third-order Bezier curve, which is composed of control points P o 、P d , P1 and P2 control, where P o Indicates the starting point of the vehicle path, with coordinates (0,0); P d Indicates the end point of the vehicle path, with coordinates (w, h); P1 controls the shape of the vehicle's path after entering the intersection, located on the extension line of the intersection entrance center, with coordinates (0, L1); P2 controls the shape of the vehicle's path when it is about to leave the intersection, located on the extension line of the intersection exit center, with coordinates (w-L2sinα, h-L2cosα); L1 represents the control point P o The distance between L2 and P1, L2 represents the control point P d The distance between P2 and P3; In this embodiment, 20 test drivers were recruited, and Figure 5 The driving simulator test shown in the figure obtained 20 actual vehicle trajectory data corresponding to each set of intersection entrance and exit lanes, and used the differential evolution algorithm to obtain a set of optimal control parameters L1 and L2 for the vehicle paths. The steps included the following:

[0046] Step 31: Extract the world coordinates of the starting point and the end point of the vehicle path. The starting point coordinate is expressed as (xorig, world, y orig,world ), the end point coordinate is expressed as (x dest,world ,y dest,world );

[0047] Step 32, extracting the vehicle trajectory coordinates of the continuous time series within the starting point and end point coordinate ranges described in step 31, which can be expressed as (x world ,y world ), the coordinates of the trajectory points satisfy:

[0048]

[0049] Step 33, based on the intersection entrance road-based coordinate system established in step 1, measuring the counterclockwise rotation angle γ relative to the world coordinate system;

[0050] Step 34, based on the world coordinates of the vehicle path starting point (xorig,world,yorig,world) obtained in step 31, the world coordinates of the vehicle trajectory point (x world ,y world ), and the coordinate system rotation angle γ obtained in step 33, the world coordinates of the vehicle trajectory point are converted to the coordinates (x, y) based on the intersection entrance road:

[0051]

[0052] In step 35, for each set of vehicle path data for the entrance and exit lanes of the intersection, the Bezier curve control parameters L1 and L2 described in step 3 are optimized using a differential evolution algorithm to find an optimal Bezier curve that minimizes the distance between the vehicle path and the Bezier curve. The parameter values ​​of the differential evolution algorithm are given below:

[0053] The population size is 20 and the maximum number of iterations is 30;

[0054] In addition, the range of L1 is [0, hw / tanα], and the range of L2 is [0, w / sinα];

[0055] In this embodiment, since the distances between the 20 vehicle paths within each set of intersection entrance and exit lanes and the Bezier curves are difficult to calculate, the optimization objective of minimizing the distances between the vehicle paths and the corresponding Bezier curves is transformed into minimizing the sum of the distances between the vehicle trajectory points and the corresponding Bezier curves. Furthermore, since it is difficult to calculate the distances between points and Bezier curves, each Bezier curve is divided into 500 sub-points, and the distance from a trajectory point to the Bezier curve is approximated as the distance between the trajectory point and its nearest sub-point. Thus, minimizing the sum of the distances between each trajectory point and its nearest sub-point is used as the optimization objective of the differential evolution algorithm.

[0056] In this embodiment, 96 sets of intersection entrance and exit position relationship parameters w, h and α, as well as the descriptive statistical characteristics of the Bezier curve control parameters L1 and L2 corresponding to the vehicle path are obtained. Figure 6 As shown;

[0057] Furthermore, in the unified vehicle path planning method based on the position relationship between the entrance and exit lanes of the intersection described in step 4, in this embodiment, the expressions of the Bezier curve control parameters and the position relationship parameters of the entrance and exit lanes and their goodness of fit are obtained by the stepwise regression fitting method as follows:

[0058] L1=0.407h+0.523,R 2 =0.517;

[0059] L2=0.709(|w|+h)-7.052,R 2 =0.623;

[0060] The above expression is converted into the calculation formula of L1 and L2 as follows:

[0061] L1=0.407h+0.523;

[0062] L2=0.709|w|+0.302h-7.575;

[0063] When a set of intersection entrance and exit position relationship parameters is given, an optimal Bezier curve controlled by the Bezier curve fitting path control parameters L1 and L2 can be drawn. The expression of the Bezier curve is as follows:

[0064]

[0065] Where: t is a curve parameter between 0 and 1; n is the number of control points (including the start and end points), in this embodiment, n = 4; P i is the i-th control point;

[0066] In one embodiment, a drone is used to capture video data of vehicles passing through intersections of different shapes, as shown in the schematic diagram. Figure 7 As shown in the figure, the actual vehicle trajectory data is extracted through image processing methods, and the intersection dimensions are measured using a rangefinder. The actual vehicle trajectory data is used to verify the accuracy of the planning method. Several sampling points are selected at equal intervals on each actual trajectory, and the distance from each sampling point to the corresponding Bezier curve fitting path is calculated, which is considered as the error of each sampling point. The root mean square error (RMSE) of the sampling point errors is calculated to represent the deviation between the actual path and the fitting path. The error for each pair of paths is calculated as follows:

[0067]

[0068] Where: T represents the number of sampling points, Δrt represents the error of the t-th sampling point;

[0069] In this embodiment, the errors between 25 sets of actual paths and fitted paths are calculated, as shown in Figure 8 As shown in the figure, the average root mean square error of the 25 cases is 0.4613m. Among them, the path diagrams of one left turn, one staggered straight-ahead, and one right turn are selected respectively, as shown in the figure. Figure 9 The present invention can directly obtain the path of a vehicle passing through an intersection through the physical environment characteristics of the intersection, and can be uniformly applied to intersections of different shapes and sizes and vehicles with different turning directions.

[0070] The above embodiments are only used to illustrate the present invention, and the various steps of the method can be changed. Any equivalent transformations and improvements based on the technical solution of the present invention should not be excluded from the scope of protection of the present invention.

Claims

1. A unified vehicle path planning method based on the position relationship between the entrance and exit lanes of an intersection, characterized in that: The steps include: Step 1: Based on the positional relationship between the entrance and exit roads of the intersection, a coordinate system based on the entrance road of the intersection is established; Step 2, based on the intersection entrance road-based coordinate system established in step 1, describing the positional relationship between the intersection entrance road and the exit road; Step 3: Based on the Bezier curve, describe the path of the vehicle passing through the intersection; Step 4: Based on the positional relationship between the entrance and exit lanes of the intersection obtained in step 2 and the path curve of the vehicle passing through the intersection obtained in step 3, a unified vehicle path planning method based on the positional relationship between the entrance and exit lanes of the intersection is established; The coordinate system based on the intersection entrance road described in step 1 is based on the intersection entrance road, the midpoint of the stop line of the intersection entrance road is the origin of the coordinate system, the straight line where the stop line is located is the X-axis of the coordinate system, and the perpendicular line of the stop line is the Y-axis of the coordinate system; The positional relationship between the entrance and exit lanes of the intersection described in step 2 is represented by parameters w, h, and α, where w and h represent the horizontal and vertical distances between the midpoints of the entrance and exit lanes, respectively, and α represents the complementary angle between the straight lines containing the entrance and exit lanes, i.e., the angle between the straight line containing the exit lane and the Y-axis. The path of the above-mentioned vehicle through the intersection is described by a third-order Bezier curve, which is controlled by control points Po, Pd, P1 and P2, where Po represents the starting point of the vehicle path with coordinates (0,0); Pd represents the end point of the vehicle path with coordinates (w,h); P1 controls the shape of the vehicle's path after entering the intersection and is located on the extension line of the center of the intersection entrance road with coordinates (0,L1); P2 controls the shape of the vehicle's path as it is about to leave the intersection and is located on the extension line of the center of the intersection exit road with coordinates (w-L2sinα,h-L2cosα); L1 represents the distance between control points Po and P1, and L2 represents the distance between control points Pd and P2; corresponding to each set of intersection entrance and exit roads, a set of optimal control parameters L1 and L2 for the vehicle path can be obtained through the actual vehicle trajectory data, including the following steps: Step 5: Extract the world coordinates of the starting point and the end point of the vehicle path. The starting point coordinate is expressed as (x orig,world ,y orig,world ), the end point coordinate is expressed as (x dest,world ,y dest,world ); Step 6: Extract the vehicle trajectory coordinates of the continuous time series within the starting point and end point coordinate range described in step 5, which can be expressed as (x world ,y world ), the coordinates of the trajectory points satisfy: Step 7, based on the intersection entrance road-based coordinate system established in step 1, measuring the counterclockwise rotation angle γ relative to the world coordinate system; Step 8: Based on the world coordinates (x orig,world ,y orig,world ), the world coordinates of the vehicle trajectory point obtained in step 6 (x world ,y world ), and the coordinate system rotation angle γ obtained in step 7, the world coordinates of the vehicle trajectory point are converted to the coordinates (x, y) based on the intersection entrance road: Step 9: For each set of vehicle path data at the entrance and exit of the intersection, the differential evolution algorithm is used to optimize the Bezier curve control parameters L1 and L2, thereby solving an optimal Bezier curve so that the distance between the vehicle path and the Bezier curve is minimized.

2. The unified vehicle path planning method based on the position relationship between the entrance and exit lanes of an intersection according to claim 1 is characterized in that: The unified vehicle path planning method based on the position relationship between the entrance and exit lanes of the intersection described in step 4 can be described by the following expression: L1=f1(w,h,α); L2=f2(w,h,α); When a set of intersection entrance and exit position relationship parameters is given, an optimal Bezier curve controlled by the Bezier curve fitting path control parameters L1 and L2 can be drawn to describe the path of the vehicle through the intersection; the expression of the Bezier curve is as follows: Where: t is the curve parameter between 0 and 1, n is the number of control points (including the start and end points), P i is the i-th control point.

Images