Method for bypassing an obstacle

Abstract

A method for avoiding an obstacle, in which an obstacle (5) located in the lane of a vehicle (1) is detected and an evasive trajectory is determined, characterized in that the evasive trajectory is determined as a function of at least one lateral boundary line on the obstacle side, the lateral distance of the lateral boundary line relative to the original boundary is continuously recalculated as a function of the current distance of the vehicle (1) from the obstacle (5) and the lateral distance is reduced as the distance of the vehicle (1) from the obstacle (5) decreases, wherein the evasive trajectory is displayed to a driver of the vehicle for avoiding the obstacle along the evasive trajectory and / or controllable units of the vehicle are controlled autonomously or semi-autonomously for avoiding the obstacle along the evasive trajectory.

Description

[0001] The invention relates to a method for avoiding an obstacle, in which an obstacle located in the lane of a vehicle is detected and an evasive trajectory is determined. State of the art

[0002] German patent DE 10 2006 042 666 A1 describes a method for preventing a vehicle from colliding with an obstacle. Using environmental sensors, objects on the road are detected, and an evasive trajectory is determined to bypass the obstacle, taking into account the current state of the vehicle. If the driver's evasive reaction is insufficient to avoid a collision, an autonomous intervention in the vehicle is carried out in conjunction with the driver's reaction. This intervention can affect the vehicle's braking system, steering system, and drive system.

[0003] In DE 10 2008 016 377 A1, a method for avoiding an obstacle is described in which an obstacle located in the lane of a vehicle is detected and an evasive trajectory is determined, which is continuously recalculated depending on the current distance of the vehicle from the obstacle.

[0004] The chosen evasive trajectory is determined from a family of possible evasive trajectories, taking into account an optimization function where the integral of the square of the curvature curve takes a minimum. Disclosure of the invention

[0005] The invention is based on the objective of determining an evasive trajectory for safely circumventing an obstacle.

[0006] This problem is solved according to the invention by the features of claim 1. The dependent claims specify advantageous further developments.

[0007] Using the method according to the invention, an obstacle located in a vehicle's lane can be safely avoided, thus preventing a collision. For this purpose, an evasive trajectory is determined along which the vehicle moves, or is intended to move, to circumvent the obstacle. Both embodiments are possible in which the evasive trajectory is determined automatically, but the obstacle is circumvented along the trajectory by driver intervention, and embodiments with autonomous or semi-autonomous circumvention of the obstacle through automatically performed interventions in controllable components of the vehicle, in particular the steering system and, if applicable, the drive and braking system. If the vehicle moves along the evasive trajectory through driver intervention, the evasive trajectory represents a recommendation that the driver should follow by adjusting the steering angle.

[0008] The obstacle can be a stationary obstacle or a moving obstacle, such as a third vehicle.

[0009] In the method according to the invention, the evasive trajectory is determined as a function of a lateral boundary line on the obstacle side. The lateral boundary line represents one boundary of an evasive corridor within which the vehicle moves to bypass the obstacle. The opposite boundary or line of the evasive corridor is formed, for example, by the center line of the roadway or by the opposite edge of the roadway. The actual evasive trajectory lies within the evasive corridor bounded by the lateral boundary lines. Within the evasive corridor, the evasive trajectory lies at a defined distance from the lateral boundary line on the obstacle side, for example, parallel to it or with its instantaneous center of rotation on a bisector between the lateral boundary lines.

[0010] The boundary line on the side of the obstacle is determined relative to the original boundary, which is formed by the adjacent lane without the obstacle, according to a calculation formula with a lateral distance that is continuously recalculated as a function of the vehicle's current distance from the obstacle. The lateral distance – viewed perpendicular to the longitudinal direction of the road or carriageway – of the newly determined boundary line defining the evasive corridor relative to the original boundary line is thus continuously recalculated, with each recalculation taking into account the vehicle's current distance from the obstacle, which continuously decreases due to the vehicle's movement.As the vehicle's distance to the obstacle's crossing point decreases, the trajectory corridor narrows towards the obstacle. Due to this reduction in distance, and taking this distance into account when determining the evasive trajectory, a reduced change in curvature is achieved from the vehicle's current position to the target position at the crossing point, viewed longitudinally. This reduced change in curvature results from the fact that, as long as the vehicle's view-ahead corridor is not breached, the trajectory does not continue parallel to the obstacle. Otherwise, a large change in curvature would be required in the area parallel to the obstacle to compensate for the initial curvature. The curvature is calculated as the second derivative of, for example, a third-degree polynomial.

[0011] Furthermore, a shortened trajectory to the area in front of the crossing point automatically intensifies the intervention, as the target point is reached sooner at the end of the planned trajectory. It also ensures that the corridor shortened to the crossing point, or its lateral boundary, does not intersect the blocked area.

[0012] The trajectory can be implemented as a third-degree polynomial that determines the lateral distance between the lateral boundary line of the avoidance corridor and the lateral boundary of the original lane as a function of the longitudinal distance between the vehicle and the obstacle. The coefficients of the third-degree polynomial represent the lateral offset, the yaw angle between the corridor boundary line and the vehicle's longitudinal axis, the curvature, and the rate of change of curvature. These coefficients, which are continuously recalculated to recalculate the avoidance trajectory, are determined from vehicle boundary conditions relating to the vehicle's position at the starting point of each calculation and at the point of passage, as well as the orientation of the vehicle's longitudinal axis relative to the lane.

[0013] The corridor length, or the foresight for trajectory planning, can be continuously shortened as the distance to the obstacle crossing point decreases, with the aim of curving the planned trajectory more sharply away from the obstacle. Shortening the permissible movement range in the longitudinal direction of the corridor thus allows for a focus on the relevant immediate area in front of the relevant obstacle crossing point, thereby generating the necessary lateral offset.

[0014] Advantageously, at each calculation time, several points on the current lateral boundary line are determined, from which the avoidance trajectory is calculated in order to narrow the avoidance corridor and thus generate a collision-free clear area.

[0015] According to the invention, the lateral clearance is reduced as the distance of the vehicle from the obstacle decreases, for example according to a linear relationship. According to a further advantageous embodiment, a weighting factor is introduced to determine the lateral clearance as a function of the current distance between the vehicle and the obstacle. This weighting factor increases as the distance to the obstacle decreases and is, for example, varied between 0 (initial position) and 1 (passage point at the obstacle) depending on the distance.

[0016] To determine the current lateral distance of the side boundary line relative to the original boundary as a function of the current vehicle distance from the obstacle, it is advantageous to consider the distance-dependent weighting factor, which reaches a maximum upon reaching the obstacle – the passing point. The coefficients of the polynomial can be calculated from the lateral distance at various points.

[0017] According to a further advantageous embodiment, the endpoint of the various, continuously recalculated evasive trajectories remains the same; this is, in particular, the point at which the vehicle passes the obstacle, which is advantageously located axially at the beginning of the obstacle. The distance of the vehicle to the point at which the obstacle passes decreases continuously. The point at which the obstacle passes can be detected in the vehicle by means of environmental sensors that monitor a forward-looking corridor located in front of the vehicle. Advantageously, the forward-looking corridor has a minimum axial extent relative to the vehicle's longitudinal axis, for example, 16 m, whereby the considered endpoint of the evasive corridor must not fall below this minimum axial extent.If the distance between the vehicle and the passing point at the obstacle falls below the minimum extent of the view-through corridor, the endpoint is set to the minimum extent of the view-through corridor, which can therefore also be located axially behind the passing point.

[0018] The process takes place in a control unit within the vehicle. The vehicle may be equipped with a steering system featuring an electric servo motor for automatic evasive maneuvers, which can be controlled via signals from the control unit. The steering system may also include a superimposed steering gear for actively controlling and translating a steering angle to the vehicle's steerable wheels, allowing the superimposed steering gear to add a superimposed steering angle to the steering angle specified by the driver.

[0019] Further advantages and practical designs can be found in the additional requirements, the figure description, and the drawings. These show: Fig. 1 In schematic representation a vehicle on a roadway with an obstacle in front of it, projecting laterally into the roadway, Fig. 2 the vehicle in two positions at different times as it approaches the obstacle along a trajectory.

[0020] In Fig. Figure 1 is a schematic representation of a motor vehicle 1 on a roadway 2 which has two lanes. The vehicle 1, traveling in the right lane, is laterally bounded on one side by the right lane boundary 3 and on the other by the center line 4. An obstacle 5 protrudes laterally into the lane of the vehicle 1, covering the right lane boundary 3 and lying partly inside and partly outside the right lane.

[0021] Vehicle 1 is equipped with environmental sensors, such as radar sensors, which scan the area in front of vehicle 1 on the roadway 2. Using these environmental sensors, a forward-looking corridor 6 is scanned in front of vehicle 1, within which obstacles on the roadway can be detected. As vehicle 1 approaches obstacle 5, the obstacle enters the forward-looking corridor 6 and can thus be detected by vehicle 1. To avoid a collision between vehicle 1 and obstacle 5, vehicle 1 must move along an evasive trajectory that bypasses obstacle 5. This evasive trajectory is defined in Fig. 2 shown.

[0022] In order to effectively bypass obstacle 5 with the least possible curvature of the evasive trajectory, the evasive trajectory T is continuously recalculated as a function of the distance between vehicle 1 and obstacle 5 as the vehicle approaches obstacle 5. Fig. Figure 2 shows vehicle 1 as an example at two different times t1 and t2; accordingly, the trajectories T and various points y are shown at the two times t1 and t2.

[0023] The avoidance trajectory T is recalculated at regular intervals, in the exemplary embodiment according to Fig.Figure 2 shows the evasive trajectory T for time t1 with a solid line and for time t2 with a dashed line. The evasive trajectory T runs from the current vehicle coordinate system to a passing point P at the obstacle, which axially designates the beginning of the obstacle 5 at its corner protruding into the lane, with an additional lateral safety margin.

[0024] Each trajectory T is calculated from the current lateral boundary line y, which is derived from a third-degree polynomial according to t=c0+c1⋅x+c2⋅x2+c3⋅x3 is determined, where y denotes the lateral displacement and x the longitudinal extent along the direction of travel. c0 represents the lateral displacement at x=0. C1 the yaw angle difference, c2 the curvature at x=0 and c3 the change in curvature.

[0025] The lateral boundary line y(t1) or y(t2) on the obstacle side forms the basis for determining the trajectories T(t1) and T(t2), which are calculated from a geometric relationship between the lateral boundary lines. For example, the trajectory T can be calculated from a lateral displacement of the lateral boundary line y, where the lateral displacement takes into account the transverse distance between the boundary line y and the center line 4.

[0026] In each determination of the lateral boundary line y at the different times, three points are taken into account, with the first point being y c0 axially at the height of the vehicle's own coordinate system, the second point y c1 axially at half the distance x H / 2 between the vehicle and the passing point P and the third point y c2The obstacle 5 lies axially at the height of the crossing point P. At time t1, the obstacle 5 is detected by the vehicle 1's environmental sensors. The distance between the vehicle 1 and the obstacle 5 at the height of the crossing point P is x. H (t1). The three points y c0 (t1), y c1 (t1) and y c2 (t1) lie on the lateral boundary line y(t1); opposite the original roadway boundary 3 with the associated points y c0,org (t1), y c1,org (t1) and y c2,org (t1) the two points y c1 (t1) and y c2 (t1) a transverse distance, whereas the first point y c0 (t1) with the point y c0,org (t1) coincides with the original lane boundary 3.

[0027] At time t2, the three points y c0 (t2), y c1 (t2) and y c2 (t2) on the lateral boundary line y(t2) and point opposite the original roadway boundary 3, on which the assigned points yc0,org (t2), y c1,org (t2) and y c2,org (t2) a lateral distance. Since all avoidance trajectories y are oriented towards the crossing point P of obstacle 5, provided a minimum distance of the view corridor is not undercut, the points y c2 (t1) and y c2 (t2) together.

[0028] The transverse distance y c0 , y c1 and y c2 for each time period according to yc0=yc0,org+coeff0(xH)⋅dyP yc1=yc1,org+coeff1(xH)⋅dyP yc2=yc2,org+coeff2(xH)⋅dyP from the points y c0,org , y c1,org and y c2,org as well as a weighting coefficient coeff0, coeff1 or coeff2 and the transverse distance dy P The weighting coefficients coeff0, coeff1, and coeff2 of the passing point P to the original lane boundary 3 are determined. Hof vehicle 1 from obstacle 5. The weighting coefficients are not constant; they vary depending on the distance between a minimum and a maximum, for example in the range between 0 and 1, with the weighting factor reaching its maximum at the point of passage P when the obstacle is reached.

[0029] Given knowledge of the transverse distance between the points y c0 , y c1 and y c2 can according to c0=f0(yc0,yc1,yc2,xH) c1=f1(yc0,yc1,yc2,xH) c2=f2(yc0,yc1,yc2,xH) c3=f3(yc0,yc1,yc2,xH) the coefficients c0, c1, c2 and c3 of the third-degree polynomial from functions f0, f1, f2 and f3, taking into further consideration the distance x H The distance between the vehicle and the obstacle is calculated. The functions f0 to f3 are given and can be implemented as linear or non-linear functions, for example, as cubic curves.

[0030] Upon reaching the passing point P at the obstacle 5, the vehicle 1 is advantageously positioned laterally midway between the passing point and the center line 4. The obstacle 5 is passed along a straight line, after which the vehicle can return to its original lateral position on the roadway along a trajectory.

[0031] The evasive trajectories T are either displayed to the driver so that they can maneuver around the obstacle 5 along the trajectories by making appropriate steering adjustments. It is also possible that, for example, through the activation of driver assistance systems, the vehicle automatically passes the obstacle 5 along the evasive trajectories.

Claims

[1] Method for avoiding an obstacle, wherein an obstacle (5) located in the lane of a vehicle (1) is detected and an evasive trajectory is determined, characterized by , that the evasive trajectory is determined as a function of at least one lateral boundary line on the obstacle side, the lateral distance of the lateral boundary line relative to the original boundary is continuously recalculated depending on the current distance of the vehicle (1) from the obstacle (5) and the lateral distance is reduced as the distance of the vehicle (1) from the obstacle (5) decreases, wherein the evasive trajectory is displayed to a driver of the vehicle for driving around the obstacle along the evasive trajectory and / or controllable units of the vehicle are controlled autonomously or semi-autonomously for driving around the obstacle along the evasive trajectory. [2] Method according to claim 1, characterized by, that the lateral distance depends linearly on the current distance of the vehicle (1) from the obstacle (5). [3] Method according to any one of claims 1 to 2, characterized by , that to determine the lateral distance a weighting factor is calculated as a function of the current distance of the vehicle (1) from the obstacle (5). [4] Method according to claim 3, characterized by , that the weighting factor increases with decreasing distance to the obstacle (5). [5] Method according to claim 4, characterized by , that the weighting factor reaches a maximum upon reaching the obstacle (5). [6] Method according to any one of claims 3 to 5, characterized by , that the weighting factor retains the value it has when passing the obstacle (5) as it does when reaching the obstacle (5). [7] Method according to any one of claims 1 to 6, characterized by, that the crossing point of the evasive trajectory lies on the median bisector between the center of the roadway and the lateral boundary line of the currently determined collision-free corridor. [8] Method according to any one of claims 1 to 7, characterized by , that at each calculation time several points on the current lateral boundary line are determined. [9] Method according to any one of claims 1 to 8, characterized by that the crossing point of all continuously recalculated evasive trajectories remains the same. [10] Method according to claim 9, characterized by , that the passing point is located axially at the beginning of the obstacle (5). [11] Method according to any one of claims 1 to 10, characterized by , that the corridor length or the forecast for trajectory planning is shortened with decreasing distance to the crossing point at the obstacle. [12] Control device for carrying out the method according to any one of claims 1 to 11. [13] Vehicle with a control unit according to claim 12. [14] Vehicle according to claim 13, characterized by , that a steering system with an electric servo motor can be controlled to bypass the obstacle (5) on the evasive trajectory. [15] Vehicle according to claim 13 or 14, characterized by , that an active steering system with superimposed steering gear can be controlled to bypass the obstacle (5) on the evasive trajectory.

Images