Collision prevention method

Abstract

The invention relates to a method (M20) for collision prevention for a vehicle (100), comprising: providing (S13) a virtual vehicle model (150) which has different vehicle heights (250) at at least two points; providing (S14) at least one piece of obstacle information which is indicative of a position of an obstacle (220) and a clearance (230) below the obstacle (220); and determining the shortest distance (240) between the vehicle (100) and an obstacle (220) on the basis of the provided obstacle information and the provided vehicle model (150).

Description

[0001] 2023PF03425

[0002] 1

[0003] Collision avoidance procedures

[0004] The present invention relates to a collision avoidance method for a vehicle. The invention also relates to a parking method for a vehicle that employs the collision avoidance method. Furthermore, the invention relates to a computer program, a control device, and a vehicle.

[0005] Document US 2023 / 0 102 144 A1 discloses a virtual parking assistant. In the first step, a parking position for a specific vehicle is displayed using a superimposed realistic representation (so-called "augmented reality") of the vehicle in a garage, allowing the user to move the parking position. In the next step, a parking trajectory is calculated from the current position to the desired parking position. This includes a virtual collision check between a garage entrance and a cuboid-shaped vehicle model (so-called "bounding box"), which encompasses the vehicle including any attachments (e.g., bicycle rack, roof box). If the collision check indicates a potential collision, the user can review it again in the superimposed representation. This is therefore a very safe method, but it requires one or more user interventions.

[0006] It is particularly common in private garages for a low part of the vehicle (e.g., a hood) to be parked under an obstacle hanging from the garage ceiling (e.g., a bicycle, a rear carrier, or similar items), with which a higher part of the vehicle (e.g., the roof of the vehicle interior) would collide.

[0007] For driver-operated vehicles, DE 102022 131 873 A1 presents a solution within the framework of a procedure for creating a height-resolved obstacle profile by proposing not to warn a driver of a reflection detected above the height of a car hood. 2023PF03425

[0008] 2

[0009] Against this background, an object of the present invention is to provide a collision avoidance method that enables a low vehicle part to move under an obstacle without collision. It is desirable that the collision avoidance method be suitable for at least semi-autonomous low-speed maneuvering procedures.

[0010] Accordingly, a collision avoidance method for a vehicle is proposed. The method includes the step of providing a virtual vehicle model. The virtual vehicle model has different vehicle heights at at least two points. The method also includes the step of providing at least one piece of obstacle information, which is indicative of the position of a given obstacle and the clearance height below the obstacle. The method also includes the step of determining at least one shortest distance between the vehicle and an obstacle based on the provided obstacle information and the provided vehicle model.

[0011] By using a virtual vehicle model that has different vehicle heights at at least two points, the provided collision avoidance method can determine a shortest distance between the at least two vehicle heights, depending on the clearance height under the respective obstacle, which allows partial passage under the obstacle without collision.

[0012] This is therefore a virtual process, or a virtually executable process. It can also be described as a contactless collision avoidance method or a collision detection method.

[0013] The virtual vehicle model can be described as a virtual representation of a vehicle contour and / or as a description of a vehicle contour. Instead of "position," one can also say "area" or "section." It can also be said that the vehicle model represents at least different vehicle heights at at least two points.

[0014] 3

[0015] The term "vehicle height" is indicative at different locations. It does not refer to a maximum vehicle height, but rather to the vehicle's height at a specific point on the vehicle. Preferably, however, the point on the vehicle with the maximum height is included in the vehicle model as one of the points on the vehicle model.

[0016] Obstacle information can be described as a data set that can exist in a variety of forms. For example, if the method described in DE 10 2022 131 873 A1 is used, the obstacle information can be conveniently described in polar coordinates (vertical angle, horizontal angle, distance) from a vehicle-fixed reference point to an ultrasonic feature. The obstacle information can also be described in a Cartesian coordinate system. Alternatively, it can be described in a vehicle-related coordinate system, thus specifying a relative position. For example, when using SLAM techniques (simultaneous localization and mapping), it can be advantageous if the obstacle information is specified in relation to a map and / or other environmental features.

[0017] It is also possible that multiple pieces of obstacle information are provided for a complex obstacle. For example, in the case of a bicycle suspended from the ceiling, an ultrasound-based environmental image might depict both wheels.

[0018] An obstacle under which a clear height is available could, for example, be a parking space under a plane sloping towards the roadway, as in a so-called duplex garage, under a stair slope or under a roof slope.

[0019] The step of determining the shortest distance can also be described as: determining a shortest distance between the vehicle and one of the obstacles (if obstacle information is provided for multiple obstacles). It can also be said that the shortest distance between the vehicle and one of the obstacles is determined by the distance between the vehicle and the obstacle.

[0020] 4

[0021] Obstacles are determined. It is also possible that the respective distance to several obstacles is determined.

[0022] By determining at least the shortest distance between the vehicle and the obstacle, the result of the procedure can be directly incorporated into motion planning.

[0023] The vehicle may be configured for manual and at least semi-autonomous parking and / or maneuvering. Therefore, the specified distance may be communicated to a driver, for example, as a visual and / or audible signal. It may also be communicated to another process or control unit as a measurable signal. Alternatively, the specified distance may be returned as a return value from a method or function to a calling method or function. In this case, the distance could be measurable, for example, as a distance traveled by the vehicle and / or as a distance maintained from the obstacle. The parking procedure described later serves as an example of the latter case.

[0024] The vehicle model is preferably indicative of the vehicle's contour or body shape. The following describes some simplifications of the vehicle contour to enable resource-efficient collision avoidance. The vehicle model can be formulated, for example, as a surface model, an edge model, a node model, and / or a volume model. These models are often convertible into one another. They can also exist simultaneously; consider, for example, a cube whose shape defines its edges, nodes, and faces. In the following variations, the elements "volume," "surface," "edge," and "node" refer to boundary elements and / or elements that delimit the vehicle model.

[0025] It is possible that the provided vehicle model has at least one inclined surface. An inclined surface is a surface that is neither horizontal nor vertical. One can also say that an inclined surface lies between a horizontal and a 2023PF03425

[0026] 5

[0027] The surface runs vertically. For example, a windshield, a hood, or the front of a vehicle can be described using an inclined surface. An inclined surface is easy to define, for example, using a normal to the surface (i.e., a vector) or using three points on the surface, such as three vertices. The inclined surface is preferably not curved. The inclined surface is preferably planar. The term "inclined" refers to a state in which the vehicle model is oriented relative to a horizontal surface.

[0028] The vehicle model may have at least seven surfaces. This proposal assumes one surface represents the vehicle floor, even if this is not modeled or measurable. In other words, the vehicle model may have at least six surfaces that do not represent the vehicle floor and / or are not the underside of the model. For example, a delivery van could be described with six surfaces representing the front, windshield, roof, rear, and two sides. Similarly, a station wagon and / or SUV could be described by having one surface each representing the front, hood, windshield, roof, rear, and one side. The surfaces are preferably flat or not curved to simplify the description.The areas counted here are external surfaces of the vehicle that are at least partially located on the outside and / or partially define the overall boundaries of the vehicle model.

[0029] The vehicle model can have up to 40 surfaces. Preferably, the vehicle model has up to 20, more preferably up to 16, and even more preferably up to 12 surfaces. Separately defined but adjacent surfaces should preferably not be counted as distinct surfaces. The fewer surfaces the vehicle model has, the lower the computational effort required to determine the shortest distance.

[0030] The vehicle model may have at least two inclined outer edges. An inclined outer edge runs neither horizontally nor vertically. This variant has the same advantages as the vehicle model above with at least one inclined surface. It may be 2023PF03425

[0031] 6

[0032] that the vehicle model has at least 15 outer edges. This variant has the same advantages as the vehicle model above with at least 6 surfaces that do not describe a vehicle floor / underside. The vehicle model may have up to 60 outer edges. Preferably, the vehicle has up to 42 outer edges, more preferably up to 28 outer edges, even more preferably up to 22 outer edges, and most preferably even only 12 outer edges.

[0033] The vehicle model may have at least 8 nodes. A node can be defined as a corner. The vehicle may have up to 40 nodes, preferably up to 32 nodes, more preferably up to 20 nodes, more preferably up to 24 nodes, and more preferably up to 16 nodes. The fewer nodes the model has, the less computational effort is required to determine the distance.

[0034] The vehicle model may have at least two primitive solids of different sizes and / or orientations, which may be separate or abutting, connected, and / or overlapping. These primitive solids are selected from the following list: cuboids (including cubes), four- to seven-sided pyramids, six- to twelve-sided truncated pyramids, prisms with triangular to hexagonal bases, cylinders, spheres, cones, and truncated cones. A "triangular prism" describes a volume formed by translating a triangle, thus having six vertices or nodes. Describing the vehicle model using such primitive solids is very easy to visualize and can be described with minimal effort. It can be efficient for the vehicle model to have up to six primitive solids.

[0035] Modern vehicles have many convex or curved exterior surfaces. Therefore, the virtual vehicle model may tangentially circumscribe the actual vehicle. This means that a projection of the virtual vehicle model lies tangentially on the outside of a projection of the actual vehicle. One advantageous application is, for example, obtaining intermediate results for other processes that can add a safety margin in a later phase. It is possible that the virtual vehicle model 2023PF03425

[0036] 7

[0037] The vehicle is described with a safety margin. This means that a projection of the virtual vehicle model describes or surrounds a projection of the actual vehicle at a distance. A homogeneous safety margin can be used in all directions. The safety margin can be a minimum value. It is also possible that a margin to one side of the vehicle is larger than the respective safety margin to other vehicle surfaces (especially to a hood), for example, to allow a driver to exit the vehicle.

[0038] The virtual vehicle model may be provided or described by a virtual base model and a user-defined allowance. The base model describes the actual vehicle tangentially or with a basic allowance (distance between the projections). The user-defined allowance is an allowance specified by the user (distance between the projections). A user may be able to adjust the allowance for each side of the vehicle to account for, for example, a roof load or a rear carrier. This approach proposes a way to make the vehicle model adaptable to a user's preferences using simple and transparent means.

[0039] The virtual vehicle model may only be indicative of a part of the vehicle. For example, modeling the rear of the vehicle may be unnecessary for forward parking.

[0040] The virtual vehicle model may describe the actual vehicle, including at least one permanent add-on part. This permanent add-on part could be selected from the following list, for example: an antenna, a side mirror, a trailer hitch, roof rails, a folding, extendable, retractable, and / or extendable vehicle element (e.g., a pop-up headlight, a folding side mirror, an extendable spoiler, or even a folding roof), and an aerodynamic component (e.g., a spoiler). Describing add-on parts is particularly advantageous with small safety margins or a safety margin-free implementation.

[0041] 8

[0042] For example, to maneuver / park a vehicle fully autonomously. For collision avoidance applications, it may be unnecessary to classify wheels as permanent attachments.

[0043] It is possible that at least one obstacle information is detected based on at least one detected sensor signal, whereby the at least one sensor signal is detected by an optical sensor (image / camera, radar, lidar) and / or by an ultrasonic sensor and / or obtained via sensor fusion. The obstacle information can be read from a map or manually programmed. If the obstacle information is obtained by the aforementioned vehicle sensors and / or via sensor fusion based on identical or different types of the aforementioned vehicle sensors, this leads to a more precise determination of the position, especially the relative position, of the respective obstacle.

[0044] The method may include: capturing a roadway based on at least one detected sensor signal, where the distance preferably corresponds to the length of the roadway between a (virtual) projection of the obstacle onto the roadway and a (virtual) projection of the vehicle model at the respective clearance height (i.e., a contour line) onto the roadway. Thus, in the case of curved or angled roadways, the distance to be traveled by the wheels can be precisely determined to facilitate route planning. Curved or angled roadways occur, for example, in sloping garage entrances or angled parking spaces, such as duplex parking spaces.

[0045] The method may include: providing a planned trajectory, wherein the distance preferably corresponds to a length of the trajectory between the obstacle and the vehicle model at the respective clearance height. The trajectory may be straight and / or (in the vertical direction) planar and / or (in the vertical direction) kinked and / or (in the vertical direction) curved and / or (in the horizontal direction). The trajectory may be single- or multi-section. It may be that instead of the trajectory2023PF03425

[0046] 9

[0047] A planned trajectory is provided and used to determine the distance. By using the planned trajectory as the basis for determining the shortest distance to an obstacle, obstacles not located along the route can be ignored. This improves the quality of distance information, especially in semi-autonomous driving procedures (parking, maneuvering). It is also possible that an obstacle is not located along a direct path between a current position and a target position, but is located along the planned trajectory. The significance of this obstacle can be recognized earlier by considering the trajectory. The planned trajectory is, for example, a trained trajectory.

[0048] An obstacle may be discarded / ignored if its relative position lies outside a driving lane. An obstacle may also be discarded / ignored if it is further away from a trajectory than a specified threshold. Finally, an obstacle may be discarded / ignored if its clearance exceeds the maximum vehicle height of the vehicle model. These steps reduce the amount of obstacle information used to determine the shortest distance, thus speeding up the calculation.

[0049] The distance can be determined longitudinally and / or in the direction of steering. The steering direction can be determined, for example, from the steering angles of the (steered and / or unsteered) wheels and the wheelbase. The resulting shortest distance between the vehicle and the obstacle is particularly important for manual driving.

[0050] The shortest distance may be determined in several angular sectors relative to a direction of travel (forward or reverse) and / or to a longitudinal direction of the vehicle. This allows a human driver, for example, to reliably identify the obstacle. 2023PF03425

[0051] 10

[0052] According to a further aspect of the invention, a method for parking a vehicle is proposed. The proposed method comprises the steps of: providing a target position for the vehicle, executing the steps of the collision avoidance method according to one of the preceding variants / embodiments, and generating a control signal for a steering and / or drive device of the vehicle depending on at least the shortest distance. The embodiments and features described for the proposed collision avoidance method apply accordingly to the proposed parking method.

[0053] The proposed method can, for example, be implemented as a function of a parking assistance system. The parking assistance system is specifically designed for semi-autonomous or fully autonomous driving of the vehicle. Semi-autonomous driving means, for example, that the parking assistance system controls a steering device and / or an automatic transmission. Fully autonomous driving means, for example, that the parking assistance system also controls a drive system and a braking system.

[0054] According to a further aspect of the invention, a method for maneuvering a vehicle is proposed. The proposed method comprises the steps of: executing the steps of the collision avoidance method according to one of the preceding variants / embodiments, and generating a trajectory for the vehicle depending on at least the shortest distance. The embodiments and features described for the proposed collision avoidance method apply accordingly to the proposed maneuvering method.

[0055] Furthermore, a computer program product is proposed which includes instructions that, when executed by a computer, cause it to perform one of the methods described above. The computer program product thus has the features and advantages of the respective method. 2023PF03425

[0056] 11

[0057] A computer program product, such as a computer program tool, can be provided or delivered from a server on a network, for example, as a storage medium such as a memory card, USB stick, CD-ROM, DVD, or as a downloadable file. This can be done, for example, in a wireless communication network by transmitting the corresponding file containing the computer program product or tool.

[0058] According to a further aspect of the invention, a control device for a vehicle is proposed. The proposed control device is configured to execute the aforementioned collision avoidance, parking, and / or maneuvering procedure according to one of the presented variants or options. The embodiments and features described for the proposed method apply accordingly to the proposed control device. The control device may, for example, have an interface for reading obstacle information, a memory for providing the virtual vehicle model, and a processing unit for determining the shortest distance to an obstacle. Preferably, the control device has an interface for outputting a sensor signal indicative of the shortest distance.

[0059] According to a further aspect of the invention, a vehicle is proposed which has a proposed control device. The embodiments and advantages described for the proposed method and / or the proposed control device thus apply accordingly to the proposed vehicle.

[0060] The vehicle is, for example, a passenger car or a truck. The vehicle preferably includes a number of sensor units designed to detect the vehicle's driving status and its surroundings. Examples of such sensor units include imaging devices such as a camera, radar (radio detection and ranging), or lidar (light detection and ranging), ultrasonic sensors, positioning sensors, wheel angle sensors, and / or wheel speed sensors. Each sensor unit is designed to output a 2023PF03425

[0061] 12

[0062] Sensor signals are set up, for example to the parking assistance system or driving assistance system, which performs semi-autonomous or fully autonomous driving depending on the detected sensor signals.

[0063] A vehicle's level of automation, for example, corresponds to an automation level according to the SAE classification system. The SAE classification system was published in 2014 by SAE International, a standards organization for motor vehicles, as J3016, "Taxonomy and Definitions for Terms Related to On-Road Motor Vehicle Automated Driving Systems." It is based on six different levels of automation and considers the degree of system intervention required and the driver's attention required. The SAE automation levels range from Level 0, which corresponds to a fully manual system, through driver assistance systems in Levels 1 and 2, to semi-autonomous (Levels 3 and 4) and fully autonomous (Level 5) systems, where no driver is required.An autonomous vehicle (also known as a driverless car, self-driving car, and robotic car) is a vehicle that is able to perceive its surroundings and navigate without human input, and it corresponds to SAE automation level 5.

[0064] Other possible implementations of the invention also include combinations of features or embodiments described previously or subsequently with regard to the exemplary embodiments, even if not explicitly mentioned. In such cases, the person skilled in the art will also add individual aspects as improvements or additions to the respective basic form of the invention.

[0065] Further advantageous embodiments and aspects of the invention are the subject of the dependent claims and the exemplary embodiments of the invention described below. The invention is further explained below with reference to preferred embodiments and the accompanying figures. 2023PF03425

[0066] 13

[0067] Fig. 1 schematically shows a top view of a proposed vehicle having a proposed control device configured to carry out the proposed method;

[0068] Fig. 2 schematically shows a comparison between a delivery van and a virtual vehicle model;

[0069] Fig. 3 shows a schematic comparison between a sedan and a virtual vehicle model;

[0070] Fig. 4 schematically shows a comparison between a sedan and another virtual vehicle model;

[0071] Fig. 5 schematically shows a comparison between a sedan and another virtual vehicle model;

[0072] Fig. 5 schematically shows a comparison between a station wagon and a virtual vehicle model;

[0073] Fig. 7 schematically shows in three views a comparison between a sedan and another virtual vehicle model;

[0074] Fig. 8 illustrates a driving situation in which several obstacles of different heights in an area in front of a vehicle of different heights are compared to determine the shortest distance to a collision; and

[0075] Fig. 9 shows a flowchart of the proposed procedure. 2023PF03425

[0076] 14

[0077] In the figures, identical or functionally equivalent elements have been given the same reference symbols, unless otherwise indicated.

[0078] Fig. 1 shows a schematic bird's-eye view of a vehicle 100. The vehicle 100 is, for example, a car located in an environment 140. The car 100 has a parking assistance system 110, which is designed, for example, as a control device. In addition, several environmental sensor devices 120, 130 are arranged on the car 100, including, for example, optical sensors 120 and ultrasonic sensors 130. The optical sensors 120 include, for example, visual cameras, radar, and / or lidar. The optical sensors 120 can each capture an image of a respective area from the environment 140 of the car 100 and output it as an optical sensor signal. The ultrasonic sensors 130 are configured to detect the distance to objects located in the environment 140 and to output a corresponding sensor signal.By means of the sensor signals detected by the sensors 120, 130, the parking assistance system 110 is able to drive the car 100 semi-autonomously or even fully autonomously. In addition to the optical sensors 120 and ultrasonic sensors 130 shown in Fig. 1, the vehicle 100 may be equipped with various other sensor devices 120, 130. Examples of these are a microphone, an accelerometer, an antenna with a coupled receiver for receiving electromagnetically transmitted data signals, and the like.

[0079] Figures 2 to 7 illustrate exemplary virtual vehicle models 150a to 150f.

[0080] Figure 2 shows the example of a typical delivery van, designated as vehicle 100. Because delivery vans are often taller than passenger cars, maneuvering in parking garages involves a particularly high risk of getting stuck on an obstacle near the garage ceiling. In this example, the virtual vehicle model 150a has the shape of a cuboid with a chamfer. The shape of the virtual vehicle model 150 can also be described as a prism with a pentagonal base. Depending on whether or not one considers a base surface of the virtual vehicle model, one can say that the virtual 2023PF03425

[0081] 15

[0082] Vehicle model 150a has 6 or 7 outer surfaces. Furthermore, this vehicle model 150a has 10 nodes 190, which are connected by 14 outer edges 180.

[0083] Depending on the implementation, the virtual vehicle model 150a can be defined as a volume (height, width, length, angle, and chamfer size), or it can be defined by the location of the nodes 190, the location of the edges 180, or the location of the surfaces 170. It is preferred to define the location of the nodes 190 and to define the edges 180 and / or the surfaces 170 in relation to the nodes 190.

[0084] Figure 3 shows an example where the vehicle 100 is a sedan. Sedans have longer and flatter hoods compared to vans, and they often have a trunk lid that is lower than the roof over a passenger compartment. Figure 3 illustrates an example of a virtual vehicle model 150b that has 10 surfaces 170. The surfaces 170 can, for example, be defined by nodes 190. The vehicle model 150b can also, for example, be a polyhedron whose vertices are defined by the nodes 190.

[0085] Figure 4 shows another example of the limousine from Figure 3. Here, the virtual vehicle model 150c is composed of several primitive bodies 160. The primitive bodies 160 are cuboids of different sizes. This virtual vehicle model 150c can be described as an arrangement of three adjacent cuboids, where, from front to back, a first cuboid with a first height represents the engine compartment, a second cuboid with a roof height represents the cabin, and a third cuboid with a third height represents the trunk, with the first and third heights each being lower than the second height. The virtual vehicle model 150c can also be described as an arrangement of several upright cuboids of different heights one behind the other. An upright cuboid is understood here to be a cuboid whose side faces are assumed to be perpendicular to a roadway of the vehicle 100.It is mentioned that the front of the middle cuboid is an area that partially limits the vehicle model 150c because it is partially obscured by the front cuboid. 2023PF03425.

[0086] 16

[0087] Figure 5 shows another example of the sedan. Here, the virtual vehicle model 150d is also composed of several primitive solids 160. These are a cuboid, an irregular truncated pyramid, and a prism. The cuboid has the height of the trunk. The cuboid extends from behind the vehicle 100 to in front of a windshield. The truncated pyramid is positioned in front of the cuboid. The truncated pyramid has a forward-sloping top surface. The truncated pyramid has an upright rectangular front face that is narrower (smaller in width) than the cuboid, so that the sides of the truncated pyramid converge from back to front. The prism, for example, has trapezoidal bases that are aligned with the sides of the cuboid. For other body shapes, an irregular quadrilateral might be a better fit than a trapezoid.

[0088] Figure 6 shows a modification of the virtual vehicle model 150e from Figure 5. This virtual vehicle model 150f is particularly suitable, for example, for station wagons, vans based on passenger cars (also known as high-roof station wagons), and / or SUVs. In the virtual vehicle model 150f, the prism extends over the entire top surface of the cuboid. Otherwise, the description for Figure 5 applies.

[0089] Figure 7 shows an even more complex virtual vehicle model 150g, consisting of 4 primitive bodies with a total of 15 external surfaces (or 16 external surfaces if the underside is included), 26 edges, and 20 nodes. For clarity, Figure 7 has three partial views: a top view (a), a side view (b), and a front view (c).

[0090] The virtual vehicle model 150g can also be described or defined in several ways. For example, the virtual vehicle model 150g can be provided as a polyhedron with 20 nodes (190) and 26 edges (180). The polyhedron can be defined, for example, by its nodes (190). 2023PF03425

[0091] 17

[0092] Furthermore, the virtual vehicle model 150g can be described as an arrangement consisting of a cuboid, a prism, and two irregular truncated pyramids, i.e., four primitive solids 160. The cuboid is upright and extends over a lower area of ​​the vehicle's interior. The front of the cuboid is simultaneously a back face of a truncated pyramid, which contains the engine compartment and tapers towards the front to a smaller, upright rectangle (as in vehicle models 150e and 150f). The back of the cuboid is simultaneously a front face of the prism. The prism is a trapezoidal prism whose bases are the top and bottom faces, so that one back face of the prism is narrower than the front face. Finally, the top of the cuboid is a bottom face of the second truncated pyramid. The truncated pyramid has a rectangular top face, which is located on or above the roof of vehicle 100 and is smaller than the bottom face of the truncated pyramid.The side faces of the truncated pyramid are inclined inwards, so that the top of the truncated pyramid is located inside the bottom of the truncated pyramid when viewed from above.

[0093] As can be seen from the top view in Fig. 7a) and even more clearly from the front view in Fig. 7c), the virtual vehicle model 150g is so wide that (in a virtual overlay of the vehicle model 150g over the actual vehicle 100) the side mirrors 260 are completely within the virtual vehicle model 150g. The side mirrors 260 are an example of a permanent add-on part found on almost every vehicle 100. By setting up the virtual vehicle model 150g to describe the permanent add-on parts, the virtual vehicle model 150g can be designed to more closely resemble the actual vehicle 100 in areas other than the sides of the vehicle.

[0094] This measure therefore results in an increase in the area of ​​the surroundings 140 of vehicle 100 that is recognized as passable.

[0095] As can be seen in Figures 2 to 7, the virtual vehicle models 150 ag each have a safety margin 200, which can be seen as the gap between the respective virtual vehicle model 150 and the projection of the vehicle 100. The safety margin 200 serves to compensate for possible measurement errors when determining the position of the 2023PF03425

[0096] 18

[0097] This allowance compensates for obstacles when determining distance and / or braking the vehicle. It is, for example, 15 cm. The safety margin of 200 can be described as the minimum distance between vehicle 100 and vehicle model 150.

[0098] As can be seen in Figures 2 to 6 and 8, the virtual vehicle model 150 may have ground clearance, meaning that the virtual vehicle model 150 does not extend down to the road surface. As can be seen in Figure 7, it is also possible that the virtual vehicle model 150 does extend down to the road surface.

[0099] Next, using the driving scene in Fig. 8 and the flowchart in Fig. 9, a proposed method M10 for parking the vehicle 100 is described, which includes the steps of a proposed method M20 for collision avoidance for the vehicle 100.

[0100] In step S11, a target position 210 is provided in the vicinity 140 of the vehicle 100. The target position 210 is located, for example, in a garage. In the area of ​​the target position 210, obstacles 220a-d hang from above, so that there is a clear height 230a-d under each obstacle 220a-d.

[0101] In step S12, a travel path for the vehicle is determined. The travel path is a virtual projection of the vehicle 100 onto the road surface at every position of the vehicle 100 from its current position to the target position 210. The travel path can be described as an area to be swept by the vehicle 100 during a planned journey. The travel path is preferably calculated with a lateral allowance. In simplified terms, the travel path can also be calculated as a region of a predefined width on both sides of a trajectory to the target position 210.

[0102] For example, the driving tube and / or trajectory may be provided by an external means or another method, so that steps S11 and S12 may coincide. 2023PF03425

[0103] 19

[0104] In step S13, the virtual vehicle model 150 is provided. In this example, this is again the virtual vehicle model 150c of Fig. 4, composed of three cuboids (i.e., primitive bodies 160). The cuboids have vehicle heights 250a-c as shown in Fig. 8. It is also possible, for example, that the virtual vehicle model 150 is provided as a basic model with user-defined safety margins 200.

[0105] In step S14, at least one obstacle information is provided. In the present example, an image of the environment 140 is acquired by sensors 120 and 130. Depending on sensor signals from sensors 120 and 130 of the vehicle 100, the relative position to the vehicle 100 and the clearance height 230a-d are then determined for each of the obstacles 220a-d.

[0106] In step S15, the amount of obstacle information is reduced. Based on the driving path provided in S12 and the obstacle information provided in S14, all obstacles whose relative position does not overlap with the driving path are discarded. For example, each ultrasonic feature can be assigned a width (e.g., a horizontal extent from the vehicle's perspective 100) depending on a camera image, which is then checked for overlap with the driving path. As a result, only obstacles 220 located on / in the driving path are considered.

[0107] In step S16, the amount of obstacle information is reduced again. Obstacle information for obstacles 220 whose clearance height 230 is greater than a maximum vehicle height 250b is discarded. Because the clearance height 230d is greater than the vehicle height 250b, obstacle 220d is no longer considered.

[0108] A subsequent step S17 is executed for each of the remaining obstacles 220a-c. In this step, a distance between the vehicle 100 and the respective obstacle 220a-c is determined depending on the virtual vehicle model 150c. 2023PF03425

[0109] 20

[0110] Here, for each of the obstacles 220a-c, it is first selected by comparing the clearance height 230a-c with the vehicle heights 250a-c which of the primitive bodies 160 the respective obstacle 220a-c could collide with.

[0111] Then the possible distances 240a-f between the primitive bodies 160 and the obstacles 220a-c are calculated. The distances 240 are assumed to be parallel to the roadway, for example, by calculating the distances 240 between a respective virtual projection of the obstacles 220 onto a roadway plane and a respective virtual projection of the outer surfaces 170 of the cuboids 160.

[0112] For inclined outer surfaces 170, for example, a virtual contour line of the respective surface can be considered at the height of the clear height 230 of the respective obstacle. Then, the distance between the virtual obstacle projection and a virtual projection of the contour line onto the road surface can be determined.

[0113] If the virtual vehicle model 150 is an edge model, then, for example, the respective virtual intersection points of the respective clearance height 230 with the edges 180 can be virtually projected onto the road surface and connected by virtual straight lines. The distance between the respective virtual straight line and the virtual obstacle projection can then be determined.

[0114] If the virtual vehicle model 150 is a node model (a point cloud), virtual intersection points between any two nodes 190 located above and below the clearance height 230 can be determined, these can then be virtually projected onto the road surface and the distance determined.

[0115] The step of (virtually) projecting onto the road surface can be performed by calculating a coordinate in the longitudinal direction of the vehicle and a coordinate in the transverse direction of the vehicle. 2023PF03425

[0116] 21

[0117] If vehicle 100 is directly in front of the target position 210, the distance is a length in the longitudinal direction of the vehicle. In other cases, for example, the length of the trajectory between the obstacle 220 and the virtual vehicle model 150 can be determined as the distance.

[0118] As a result of step S17, the five distances 240a-e are determined in magnitude in the example in Fig. 8.

[0119] In step S18, the shortest distance between the vehicle 100 and an obstacle 220 is determined. This can be referred to as the critical distance. For this purpose, the shortest of the five distances 240a-e determined in S17 is calculated. In this case, the shortest distance between one of the obstacles 220a-d and the vehicle 100 is the distance 240e between the second-lowest obstacle 220b and the front of the vehicle. This distance is significantly longer than the distances between the front of the vehicle and the nearest obstacle 220d or the nearest obstacle 220c that is deep enough for the vehicle to pass through.

[0120] In step S19, a control signal for a steering and / or drive unit is determined based on the shortest distance 240e calculated in S18. This signal enables collision-free travel to the target position 210. The control signal is thus generated so that the vehicle 100 gets as close as possible to the target position 210 without colliding with any of the obstacles 220. The control signal is then output to the steering and / or drive unit.

[0121] Although the present invention has been described using exemplary embodiments, it can be modified in many ways. 2023PF03425

[0122] 22

[0123] REFERENCE MARK LIST

[0124] 100 vehicles

[0125] 110 Control device

[0126] 120 optical sensor

[0127] 130 Ultrasonic sensor

[0128] 140 surroundings

[0129] 150 virtual vehicle models

[0130] 160 primitive bodies

[0131] 170 area

[0132] 180 edge

[0133] 190 knots

[0134] 200 surcharge

[0135] 210 Target position

[0136] 220 Obstacle

[0137] 230 clear height

[0138] 240 distance

[0139] 250 vehicle height

[0140] 260 permanent attachment

[0141] M10 Procedure for parking a vehicle

[0142] M20 Collision avoidance procedure for a vehicle

[0143] 511 Providing a target position for the vehicle

[0144] 512 Providing a trajectory and / or driving path for the vehicle 513 Providing a virtual vehicle model

[0145] 514 Provide at least one obstacle information

[0146] 515 Reducing obstacle information depending on the driving lane 516 Ignoring obstacles whose clearance is greater than a maximum vehicle height

[0147] 517 Determining the distance between the vehicle and each obstacle depending on the virtual vehicle model 2023PF03425

[0148] 23

[0149] S18 Determining the shortest distance between the vehicle and an obstacle S19 Determining a control signal to a steering and / or drive device

Claims

2023PF03425 24 PATENT CLAIMS 1. Collision avoidance method (M20) for a vehicle (100), comprising: Providing (S13) a virtual vehicle model (150) which has different vehicle heights (250) at at least two locations; Providing (S14) at least one obstacle information indicative of the position of a given obstacle (220) and the clearance height (230) below the obstacle (220); and Determining a shortest distance (240) between the vehicle (100) and an obstacle (220) based on the provided obstacle information and vehicle model (150).

2. Method according to claim 1, characterized in that the provided vehicle model (150) has at least one inclined outer surface (170) and / or at least 7 outer surfaces (170).

3. Method according to claim 1 or 2, characterized in that the vehicle model (150) has at least two inclined outer edges (180) and / or at least 15 outer edges (180).

4. A method according to any of the preceding claims, characterized in that the vehicle model (150) has at least two primitive bodies (160) of different sizes and / or orientations, which are separated from each other or abutting each other, connected to each other and / or overlapping each other, wherein the primitive bodies (160) are preferably selected from: cuboids including cubes, four- to seven-sided pyramids, six- to twelve-sided truncated pyramids, prisms with three- to six-sided bases, cylinders, spheres, cones and truncated cones. 2023PF03425 25 5. Method according to one of the preceding claims, characterized in that the virtual vehicle model (150) describes the actual vehicle (100) with a safety margin (200).

6. Method according to one of the preceding claims, characterized in that the virtual vehicle model (150) is provided by a virtual base model which describes the actual vehicle (100) and at least one surcharge specified by a user.

7. Method according to one of the preceding claims, characterized in that the virtual vehicle model (150) describes the actual vehicle (100) including at least one permanent attachment part (260), wherein the at least one permanent attachment part (260) is preferably selected from an antenna, a side mirror, a trailer hitch, a roof rail, a fold-out and / or extendable vehicle element and an aerodynamic component.

8. Method according to one of the preceding claims, characterized in that the obstacle information is detected depending on at least one detected sensor signal, wherein the at least one sensor signal is obtained from an optical sensor (120) and / or from an ultrasonic sensor (130) and / or by means of sensor fusion.

9. Method according to one of the preceding claims, characterized in that the method comprises: detecting a roadway course depending on at least one detected sensor signal, wherein the distance (240) preferably corresponds to a length of the roadway course between a projection of the obstacle (220) onto the roadway course and a projection of the vehicle model (150) at the level of the respective clearance height (230) onto the roadway course.

10. Method according to one of the preceding claims, characterized in that the method comprises: providing (S12) a planned trajectory, wherein the distance2023PF03425 26 (240) preferably corresponds to a length of the trajectory between the obstacle (220) and the vehicle model (150) at the level of the respective clearance height (230).

11. Method according to one of the preceding claims, characterized in that the distance (240) is determined in the longitudinal direction and / or in the steering direction of the vehicle (100).

12. Method (M10) for parking a vehicle (100), comprising: Providing (S11) a target position (210) for the vehicle (100); Performing the steps of the method (M20) according to one of the preceding claims; and Determining (S19) a control signal to a steering and / or drive device of the vehicle (100) depending on at least the shortest distance (240).

13. Computer program product comprising instructions which, when the program is executed by a computer, cause it to execute the method (M10, M20) according to any one of claims 1-12.

14. Control device (110) for a vehicle (100) which is configured to perform the method (M10, M20) according to one of claims 1 - 12.

15. Vehicle (100) with a control device according to claim 14.

Images