Method for planning a target path for a motor vehicle

Abstract

The invention relates to a method for planning a target path for the at least highly automated guidance of a motor vehicle within a parking space from a starting position located within the parking space to a target position located within the parking space, comprising the following steps: Receiving a digital map of the parking lot, wherein the digital map specifies at least one critical area of the parking lot, Planning the target path from the starting position to the target position based on the digital map, wherein the target path is planned using a BiRRT* algorithm, according to which a first search tree is iteratively constructed from the starting position to the target position based on the digital map and according to which a second search tree is iteratively constructed from the target position to the starting position, wherein the target path is planned based on the first and second search trees, wherein, for the iterative construction of the search trees, samples of motor vehicle poses are drawn from the digital map, based on which the search trees are iteratively constructed, wherein it is specified that, in the context of drawing the samples, the at least one critical area is prioritized so that a sample is drawn from the at least one critical area with a higher probability than from non-critical areas. The invention relates to a device, a computer program and a machine-readable storage medium.

Description

[0001] The invention relates to a method for planning a target path for at least highly automated driving of a motor vehicle within a parking space, a device, a computer program and a machine-readable storage medium. State of the art

[0002] The patent specification EP 3 323 697 B1 discloses a method for planning a trajectory for autonomous parking of a motor vehicle in a parking area with multiple parking spaces.

[0003] The published patent application DE 10 2018 008 685 A1 discloses a method for training an artificial neural network.

[0004] The published patent application CN 115 560 771 A discloses a method for planning a path.

[0005] The published patent application CN 115 408 779 A discloses a method for testing an algorithm for a parking assistant. Disclosure of the invention

[0006] The object underlying the invention is to provide a concept for efficiently planning a target path for at least highly automated guidance of a motor vehicle within a parking space from a starting position located within the parking space to a target position located within the parking space.

[0007] This object is achieved by means of the respective subject matter of the independent claims. Advantageous embodiments of the invention are the subject matter of the respective dependent subclaims.

[0008] According to a first aspect, a method is provided for planning a target path for at least highly automated driving of a motor vehicle within a parking space from a starting position located within the parking space to a target position located within the parking space, comprising the following steps: Receiving a digital map of the parking lot, wherein the digital map specifies at least one critical area of the parking lot, Planning the target path from the starting position to the target position based on the digital map, wherein the target path is planned using a BiRRT* algorithm, according to which a first search tree is iteratively constructed from the starting position to the target position based on the digital map and according to which a second search tree is iteratively constructed from the target position to the starting position, wherein the target path is planned based on the first and second search trees, wherein, for the iterative construction of the search trees, samples of motor vehicle poses are drawn from the digital map, based on which the search trees are iteratively constructed, wherein it is specified that, in the context of drawing the samples, the at least one critical area is prioritized so that a sample is drawn from the at least one critical area with a higher probability than from non-critical areas.

[0009] According to a second aspect, a device is provided which is configured to carry out all the steps of the method according to the first aspect.

[0010] According to a third aspect, a computer program is provided, comprising instructions which, when the computer program is executed by a computer, for example by the device according to the second aspect, cause the computer to carry out a method according to the first aspect.

[0011] According to a fourth aspect, a machine-readable storage medium is provided on which the computer program according to the third aspect is stored.

[0012] The invention is based on and incorporates the finding that the above object is achieved by using a digital map of the parking space to plan the target path from the starting position to the target position of the parking space. This digital map specifies at least one critical area of the parking space. This at least one critical area is taken into account by the BiRRT* algorithm in such a way that, during sampling, i.e., the drawing of samples, the at least one critical area is sampled with higher priority than non-critical areas. In other words, it is more likely that a sample will be drawn from the at least one critical area than a sample from non-critical areas.

[0013] The digital map thus specifies one or more critical areas. The rest of the digital map is thus made up of only non-critical areas. In other words, the parking lot is completely divided into one or more critical areas and one or more non-critical areas. Samples are more likely to be taken from a critical area than from the entire set of non-critical areas.

[0014] During the iterative construction of the search trees, a sample is drawn for each iteration. According to the previously known procedure, the sample is drawn evenly across the entire parking lot. According to the concept described here, however, the sample is not drawn evenly across the entire parking lot, but rather the one or more critical areas are prioritized so that the probability distribution is weighted more heavily toward the one or more critical areas. Based on such a probability distribution, the BiRRT* algorithm draws a sample with a higher probability from a critical area than from a non-critical area.This results in the technical advantage, for example, that areas of the parking lot that are difficult for a motor vehicle to drive through can be taken into account more efficiently and better when planning the target path than with uniform sampling, i.e. taking a uniformly distributed sample.

[0015] Purely uniform sampling for target path planning may therefore not be ideal. For example, on ramps or tight curves, it may be more difficult to determine valid target path sections using the BiRRT* algorithm, whereas it may be easier on a straight road. Therefore, it is advantageous to focus the target path planning iterations on the more difficult regions, the critical areas. In the case of BiRRT*, this focus can be achieved by sampling new target poses from these critical areas more frequently, i.e., by taking samples from these critical areas more frequently than for non-critical areas. This increases the chance of finding possible solutions for these areas and thus the chance of finding a path from start to finish in a shorter time.

[0016] This results in the technical advantage in particular that a concept is provided for the efficient planning of a target path for the at least highly automated guidance of a motor vehicle within a parking space from a starting position located within the parking space to a target position located within the parking space.

[0017] The term "at least highly automated driving" encompasses highly automated driving, fully automated driving, and autonomous driving. This means, for example, that the motor vehicle drives in a highly automated, fully automated, or autonomous manner.

[0018] Highly automated driving corresponds to automation level 3 according to the definition of the Federal Highway Research Institute (BASt). Fully automated driving corresponds to automation level 4 according to the BASt definition. Autonomous driving corresponds to automation level 5 according to SAE (J3016), where SAE stands for "Society of Automotive Engineers."

[0019] The planned target path is thus suitable so that the motor vehicle can be guided or driven from the starting position to the target position based on it, at least in a highly automated manner.

[0020] For example, the vehicle performs an AVP process within the parking space. AVP stands for "Automated Valet Parking," which translates into German as "automatic parking process."

[0021] An AVP process includes, for example, at least highly automated driving of the motor vehicle from a drop zone, also called a drop-off location, to a parking location, and at least highly automated driving of the motor vehicle from a parking location to a pick-up location, also called a pickup zone. At a drop-off location, a driver of the motor vehicle drops off the motor vehicle for an AVP process. At a pick-up location, the motor vehicle is picked up after the AVP process has ended.

[0022] A parking space, as defined in this description, can also be referred to as a parking area and serves as a parking space for motor vehicles. The parking space thus forms, in particular, a contiguous area that includes several parking spaces (in the case of a parking space on private property) or parking bays (in the case of a parking space on public property). According to one embodiment, the parking space can be enclosed by a parking garage. In particular, the parking space is enclosed by a garage.

[0023] The abbreviation “BiRRT*” stands for “Bidirectional Rapidly Exploring Random Trees”.

[0024] A pose of the motor vehicle, i.e. a motor vehicle pose, is defined, for example, by x, y, φ and a speed.

[0025] The phrase “at least one” means “one or more”.

[0026] When the singular is used for the critical area, the plural should always be read along with it and vice versa.

[0027] In one embodiment of the method, it is provided that before drawing a sample, a random decision is made as to whether a sample is to be drawn from the at least one critical area or from the entire map, wherein, if the sample is to be drawn from the at least one critical area, a sample is drawn from the at least one critical area, wherein, if the sample is to be drawn from the entire map, the sample is drawn from the entire map.

[0028] This provides, for example, the technical advantage that the sample can be drawn efficiently.

[0029] Computational and memory efficiency can be advantageously ensured by prioritizing certain areas. Furthermore, the execution time required to find an initial or optimal solution to the path planning problem can be advantageously reduced.

[0030] In one embodiment of the method, it is provided that the two possibilities regarding the decision from which the sample is to be drawn are each weighted with a predetermined probability, in particular with a weight of 50% each.

[0031] This provides, for example, the technical advantage that the sample can be drawn efficiently.

[0032] In one embodiment of the method, if there are several critical areas and a decision is made that a sample is to be drawn from a critical area, one of the several critical areas is randomly selected from which the sample is to be drawn, wherein the sample is drawn from the one randomly selected critical area.

[0033] This provides, for example, the technical advantage that the sample can be drawn efficiently.

[0034] In one embodiment of the method, it is provided that the at least one critical region has a quadrilateral shape, in particular a rectangular shape.

[0035] This provides, for example, the technical advantage of efficiently defining the critical region. In particular, the provision of a quadrilateral shape, especially a rectangular shape, has the technical advantage of enabling a computationally efficient execution of the BiRRT* algorithm.

[0036] In one embodiment of the method, it is provided that the motor vehicle pose is determined by pose coordinates, which are drawn in a uniformly distributed manner from respectively predetermined corresponding pose coordinate intervals during the drawing of the sample.

[0037] This provides, for example, the technical advantage that the sample can be drawn efficiently.

[0038] For example, a motor vehicle pose is defined by its orientation and position. The position is defined by an x-coordinate and a y-coordinate of a Cartesian coordinate system. A motor vehicle pose is defined by x, y, φ (orientation), and a velocity. For example, a pose is defined relative to a parking space coordinate system.

[0039] A pose coordinate interval is an interval comprising several pose coordinates.

[0040] For example, the device is programmed to execute the computer program.

[0041] The method is carried out, for example, by means of the device.

[0042] Device features result analogously from corresponding process features and vice versa. This means that the technical functionalities of the process result from the corresponding technical functionalities of the device and vice versa.

[0043] For example, the method is a computer-implemented method.

[0044] For example, the method includes outputting target path signals representing the planned target path.

[0045] The device comprises, for example, an input configured to receive the digital map of the parking space. The device comprises, for example, a processor configured to plan the desired path. The device comprises, for example, an output configured to output the desired path signals. The processor comprises, for example, one or more processors.

[0046] The device is, for example, a computer. The device is, for example, implemented in a cloud infrastructure.

[0047] The embodiments and exemplary embodiments described here can be combined with one another in any way, even if this is not explicitly described.

[0048] The invention is explained in more detail below using preferred embodiments. These show: Fig. 1 a flowchart of a method according to the first aspect, Fig. 2 a device according to the second aspect, Fig. 3 a machine-readable storage medium according to the fourth aspect, Fig. 4 a parking space and Fig. 5 the parking lot of the Fig. 4, where no critical area around the ramp is specified, which is preferentially considered by the BiRRT* algorithm with regard to sampling.

[0049] In the following, the same reference symbols may be used for the same features.

[0050] Fig. 1 shows a flowchart of a method for planning a target path for at least highly automated driving of a motor vehicle within a parking space from a starting position located within the parking space to a target position located within the parking space, comprising the following steps: Receiving 101 a digital map of the parking lot, wherein the digital map specifies at least one critical area of the parking lot, Planning 103 the target path from the starting position to the target position based on the digital map, wherein the target path is planned 105 using a BiRRT* algorithm, according to which a first search tree is iteratively constructed 107 from the starting position to the target position based on the digital map and according to which a second search tree is iteratively constructed 109 from the target position to the starting position, wherein the target path is planned based on the first and second search trees, wherein, for the iterative construction of the search trees, samples of motor vehicle poses are drawn 111 from the digital map, based on which the search trees are iteratively constructed, wherein it is specified 113 that, in the course of drawing the samples, the at least one critical area is prioritized so that a sample is drawn from the at least one critical area with a higher probability than from non-critical areas.

[0051] Fig. 2 shows a device 201 which is configured to carry out all steps of the method according to the first aspect.

[0052] Fig. 3 shows a machine-readable storage medium 301 on which a computer program 303 is stored. The computer program 303 includes instructions that, when executed by a computer, cause the computer program 303 to execute a method according to the first aspect.

[0053] Fig. 4 shows a parking space 401.

[0054] An x, y coordinate system 403 is placed on the parking lot 401. The coordinate system 403 includes an x-axis 405 and a y-axis 407. The abscissa of the coordinate system 403 is the x-axis 405. The ordinate of the coordinate system 403 is the y-axis 407.

[0055] A starting position 409 is located within parking space 401. A destination position 411 is located within parking space 401. A motor vehicle is to be guided from starting position 409 to destination position 411 in a manner that is at least highly automated. According to the concept described here, a target path for the motor vehicle is planned from starting position 409 to destination position 411.

[0056] The parking lot 401 includes a ramp 413. A critical area 415 is defined, which includes the ramp 413.

[0057] According to the concept described here, the target path is planned using a BiRRT* algorithm. Two search trees are constructed: a first search tree 417 and a second search tree 419. The first search tree 417 is constructed from the starting position 409 to the target position 411. The second search tree 419 is constructed in the opposite direction, from the target position 411 to the starting position 409.

[0058] During the construction of the search trees 417, 419, samples of motor vehicle poses are taken, whereby these samples are drawn with a higher probability from the critical area 415 than from the rest of the parking space 401. Thus, more samples are drawn from the area 415 than from the rest of the parking space 401. This results in the technical advantage, for example, that more possible solutions for a desired path are found within the area 415, which can increase computing time and solution quality.

[0059] By way of example, several potential vehicle poses for the first search tree 417 are provided with the reference numeral 421. By way of example, several potential vehicle poses for the second search tree 419 are provided with the reference numeral 423.

[0060] Fig. 5 shows parking lot 401 of the Fig.4, wherein no critical region 415 is defined around the ramp 413, which is given priority by the BiRRT* algorithm with regard to sampling. Accordingly, the samples are drawn evenly distributed and do not prioritize the critical region 415 encompassing the ramp 413. QUOTES CONTAINED IN THE DESCRIPTION

[0000] This list of documents submitted by the applicant was generated automatically and is included solely for the convenience of the reader. This list is not part of the German patent or utility model application. The DPMA assumes no liability for any errors or omissions. Cited patent literature

[0000] EP 3 323 697 B1

[0002] DE 10 2018 008 685 A1

[0003] CN 115 560 771 A

[0004] CN 115 408 779 A

[0005]

Claims

[1] Method for planning a target path for the at least highly automated guidance of a motor vehicle within a parking space (401) from a starting position (409) located within the parking space (401) to a target position (411) located within the parking space (401), comprising the following steps: Receiving (101) a digital map of the parking space (401), wherein the digital map specifies at least one critical area (415) of the parking space (401), Planning (103) the target path from the starting position (409) to the target position (411) based on the digital map, wherein the target path is planned (105) using a BiRRT* algorithm, according to which a first search tree (417) from the starting position (409) to the target position (411) is iteratively constructed (107) based on the digital map and according to which a second search tree (419) from the target position (411) to the starting position (409) is iteratively constructed (109), wherein the target path is planned based on the first and second search trees (419), wherein, for the iterative construction of the search trees, samples of motor vehicle poses (421, 423) are drawn (111) from the digital map, based on which the search trees are iteratively constructed, wherein it is specified (113) that, in the course of drawing the samples, the at least one critical area (415) is prioritized, so that a sample is drawn from the at least one critical area (415) with a higher probability than from non-critical areas. [2] Method according to claim 1, wherein before drawing a sample it is randomly decided whether a sample should be drawn from the at least one critical area (415) or from the entire map, wherein if the sample is to be drawn from the at least one critical area (415), a sample is drawn from the at least one critical area (415), wherein if the sample is to be drawn from the entire map, the sample is drawn from the entire map. [3] Method according to claim 2, wherein the two possibilities concerning the decision from which the sample is to be drawn are each weighted with a predetermined probability, in particular each weighted with 50%. [4] Method according to claim 2 or 3, wherein, if there are several critical areas (415) and a decision is made that a sample is to be taken from a critical area (415), one of the several critical areas (415) is randomly selected from which the sample is to be taken, the sample being taken from the one randomly selected critical area (415). [5] Method according to one of the preceding claims, wherein the at least one critical region (415) has a quadrilateral shape, in particular a rectangular shape. [6] Method according to one of the preceding claims, wherein the at least one critical area (415) defines an area of the parking space (401) within which a maneuvering space for the motor vehicle is smaller than in the immediate vicinity of the at least one critical area (415). [7] Method according to claim 6, wherein the at least one critical area (415) comprises a ramp (413) of the parking lot (401) and / or a curve. [8] Method according to one of the preceding claims, wherein the motor vehicle pose (421, 423) is determined by pose coordinates which are each drawn in a uniformly distributed manner from respectively predetermined corresponding pose coordinate intervals as part of drawing the sample. [9] Device (201) which is arranged to carry out all the steps of the method according to one of the preceding claims. [10] Computer program (303) comprising instructions which, when the computer program (303) is executed by a computer, cause the computer to carry out a method according to one of claims 1 to 8. [11] Machine-readable storage medium (301) on which the computer program (303) according to claim 10 is stored.

Images