Method for determining a bypass trajectory for a vehicle and for carrying out an at least semi-autonomous driving manoeuvre, electronic vehicle guidance system, etc.

Abstract

Aspects of the invention relate to methods for determining and using bypass trajectories (16) and / or approach trajectories (19) on the basis of a learned driving trajectory (13) and further information. The invention also relates to aspects of a vehicle guidance system (2), a vehicle (1), a computer program, and a control device (4).

Description

[0001] Method for determining a detour trajectory for a vehicle and for carrying out an at least semi-autonomous driving maneuver, electronic vehicle guidance system, etc.

[0002] One aspect of the invention relates to a method, in particular a computer-implemented method, for determining a travel trajectory for a vehicle for at least semi-autonomously avoiding an obstacle in an environment. Another aspect of the invention relates to a method for performing an at least semi-autonomous driving maneuver of a vehicle. Furthermore, one aspect of the invention relates to an electronic vehicle guidance system. Further aspects of the invention relate to a vehicle, a computer program, a computer-readable data carrier, and a control unit. Furthermore, one aspect of the invention relates to a method for determining an approach trajectory for a vehicle for approaching and merging into a learned driving trajectory.

[0003] Methods for maneuvering a vehicle along a given trajectory are known. For example, parking maneuvers can be repeated.

[0004] Conventional systems and methods only allow for minor deviations when following such a predetermined trajectory. This also means that obstacles encountered along the trajectory while following it will cause such a following maneuver, which is usually also performed fully autonomously, to be aborted.

[0005] It is an object of the present invention to provide a method in which the tracking of a known trajectory is made possible in an improved manner even under a wide variety of scenarios.

[0006] This task is solved by a method, an electronic vehicle guidance system, a vehicle, a computer program, a computer-readable data carrier, and a control unit.

[0007] One aspect of the invention relates to a method, in particular a computer-implemented method, for determining a vehicle's trajectory for at least semi-autonomously avoiding an obstacle in an environment. In particular, the method comprises the following steps:

[0008] - In particular, providing a learned driving trajectory, along which a vehicle has driven in the environment in a driving path, to a computing unit;

[0009] - In particular, providing obstacle information about an obstacle to the computing unit when subsequently following the driving trajectory in the environment;

[0010] - In particular, providing at least ground area information on ground areas of the environment that lie along the travel path outside the travel path to the computing unit;

[0011] - In particular, determining at least one detour trajectory branching off from the travel trajectory depending on the travel trajectory, the obstacle information and the ground area information with the computing unit.

[0012] Such a method now makes it possible to flexibly provide a known and learned driving trajectory that is to be used for a following maneuver. Therefore, the method now allows for driving maneuvers that do not necessarily have to be carried out entirely along the learned driving trajectory, without having to be aborted if obstacles occur. This is advantageous in that various scenarios that occur in reality, in which an obstacle may also be present along the learned driving trajectory during a following maneuver, can be managed, and the proposed method now also enables evasion of such an obstacle using the computer-implemented avoidance trajectory.This method now makes it particularly advantageous to intelligently determine at least one such avoidance trajectory based on specific information. This ensures that at least one such avoidance trajectory can be determined using this computing unit based on specific basic information. This particularly advantageous method enables the method to use the intelligence of the method to find a solution that allows the driver to utilize the travel trajectory and, with the avoidance trajectory, also to present a solution that allows an obstacle to be avoided, thus enabling the driver to follow the travel trajectory, particularly before and after this obstacle.In an intelligent way, this computer-implemented procedure thus presents a solution to be able to follow the learned driving trajectory in the best possible and comprehensive way and yet not have to interrupt the following when an obstacle occurs, but can drive past the obstacle in a needs-based and individually tailored manner.

[0013] A particularly advantageous feature of this method is that the specific information provided to this computing unit for determining the avoidance trajectory now also makes it possible to intelligently determine, if necessary, an avoidance trajectory that runs at least partially through surrounding areas that were not traveled during the travel to learn the travel trajectory. This ground area information is thus used in a particularly advantageous manner, so that a necessary avoidance trajectory can be determined in an emergency, allowing a safe and wider avoidance of the obstacle. This ground area information also makes it possible, if necessary, to use areas further away from the obstacle and the travel trajectory to allow the avoidance trajectory to run there.In particular, an assessment is made as to whether it is permissible to drive over a ground area with the vehicle and / or an assessment of the strength of the ground area for driving over with the vehicle.

[0014] This ground area information includes, in particular, information on the type of ground area or the surrounding subsurface and / or the current conditions of this ground area and / or the strength of this ground area. These specific criteria or parameters, which may be or may comprise ground area information, are not to be understood as exhaustive. In particular, ground area information also includes information that allows for an assessment of whether the area can be driven on in principle, particularly due to the weight of the vehicle, or whether damage to the ground area would occur in this regard and / or whether damage to the vehicle would occur in this regard if this ground area were driven on.In addition to or instead of this, an analysis can be performed, particularly by the computing unit, based on this ground area information to determine whether driving on such a ground area is permitted. For example, this may be permissible or impermissible due to regulatory restrictions and / or structural limitations. In this context, this may also be independent of whether the area would be generally passable, for example, due to the strength of the ground area.

[0015] This ground area information, particularly with regard to the position relative to the driving trajectory, enables a particularly advantageous specification of an information cluster in order to be able to determine a particularly needs-based detour trajectory.

[0016] In particular, ground areas are assessed which are up to greater than or equal to 1.00 m, in particular up to 5.00 m laterally away from the driving trajectory, in particular in order to be able to plan a detour trajectory in such widely spaced ground areas, in particular to be able to plan it depending on requirements.

[0017] In one embodiment, drivability information is determined using the computing unit depending on the ground area information. This drivability information determined internally in the computing unit is taken into account when determining the detour trajectory. Thus, in this embodiment, the ground area information provides basic information to enable a drivability assessment to be carried out using the computing unit. In addition to or instead of this, it is also possible for general drivability information to be provided to the computing unit independently of such internal determination of drivability information. This can, for example, be information that was recorded when driving through the environment during the learning of the driving trajectory. In addition to or instead of this, such provided drivability information can also be provided to the computing unit by another external unit.

[0018] Not only in this context, providing means receiving information by the processing unit, for example. It is also possible that providing means the processing unit itself retrieves the relevant information, for example, from one or more memories. Thus, this information is only provided to the processing unit in such units, such as memories.

[0019] In one embodiment, the ground area information is compared with reference ground area information by the computing unit. Depending on the comparison, it is determined whether a ground area is classified as passable. In one embodiment, the computing unit thus performs a classification. This can, for example, be such that a ground area is classified as passable or impassable. A third class can, for example, be specified such that it is unknown whether this ground area is passable or not. Such a comparison with reference ground area information in particular allows for a diverse analysis for different ground areas. In this regard, the above-mentioned classification is also made possible very reliably and precisely.

[0020] In one embodiment, navigable soil types and / or navigable unevenness and / or impassable zones within a soil area are specified as reference soil area information. Additionally or instead of this, impassable soil types and / or impassable zones and / or impassable unevenness can be specified as reference soil area information. This also enables a very detailed analysis, and very different soil areas can therefore still be analyzed very precisely, and in particular, classified.

[0021] Examples of soil types include asphalt, paving, gravel, sand, or pebbles. Grass or lawns can also be considered soil types in this context. A drivable unevenness can be, for example, a slight depression or elevation. These could be small holes or gullies, or low, humped traffic calming elements.

[0022] Impassable soil types, particularly those that do not rise above a passable area in terms of their level, can be, for example, water areas such as a pool or a pond. Impassable soil types can also be, for example, deep mud. Impassable zones, especially those that are not permitted or approved for driving, can be, for example, a flowerbed or an enclosed lawn area or a front garden. Impassable uneven surfaces can also be provided, whereby these can be understood as elevations or depressions that, for example, have a level change of at least 20 cm compared to the adjacent surface. It is therefore possible to drive over slight elevations and / or depressions in order to be able to navigate the bypass trajectory.On the other hand, if the avoidance trajectory is planned accordingly, it avoids having to drive over undesirably high or undesirably steeply sunken areas. This also advantageously prevents damage to the vehicle and / or this ground area.

[0023] In one embodiment, if a ground area is classified as passable, the avoidance trajectory is planned using the computing unit, at least for some areas of this ground area. This assessment of a ground area makes it possible to plan the avoidance trajectory particularly conveniently. This means that the passability of the ground area also allows for a wider detour around the obstacle if necessary. This can also improve the safe avoidance of the obstacle.

[0024] Preferably, a ground area is considered as a basis in terms of its size and / or geometry and / or position so that it borders the driving path. In particular, it is also possible to consider a ground area so that it borders the driving trajectory. This avoids the occurrence of unconsidered areas that would otherwise arise between the driving path and the ground area. This avoids areas that might need to be assessed and that are relevant for planning the detour trajectory.

[0025] In one embodiment, if a ground area is classified as unknown for driving and / or cannot yet be detected by a detection unit from the current position of the vehicle, but is not classified as not impassable, the computing unit considers substitute information to assess this ground area to determine its passability. Especially when a ground area cannot be clearly classified as passable or impassable, a further analysis using the proposed method is to be provided in this state of unknown. This is advantageous in that a ground area initially classified as unknown is not discarded.This advantageous embodiment also makes it possible to conduct a further analysis of such an initially unknown classified ground area in order to, based on this further analysis, possibly arrive at a more precise result as to whether the ground area is drivable or not. Thus, such a more in-depth and multi-stage analysis of a ground area can also achieve the following: if it can be assessed as drivable through a cascaded analysis, a more specific bypass trajectory and / or multiple bypass trajectories can be planned, which then also run through this ground area.

[0026] As replacement information, past

[0027] Ground area information is provided to the computing unit. This past

[0028] Ground area information is, in particular, information that was recorded while driving through the surroundings to learn the driving trajectory, in particular with at least one detection unit of the vehicle. Even if, in this context, the ground area is initially classified as unknown when determining the avoidance trajectory, it can still be concluded from this known past ground area information whether this ground area is passable or not. For example, such a classification as unknown can be made if, while following the driving trajectory and upon encountering such an obstacle, this ground area cannot currently be recorded with the detection unit and / or can only be recorded inaccurately.Based on this specific substitute information, it can then be determined, at least with an assessment probability, whether this ground area, which was identified as passable in the past, can also currently be classified as passable or impassable. In particular, in the advantageous exemplary embodiment, a safety buffer or a discount on the accuracy of the statement is applied when using substitute information. For example, this result can be checked on the basis of the substitute information using other plausibility checks. In particular, but not only here, processes such as reducing the speed of the vehicle when then traveling along the detour trajectory and / or taking into account current weather information and / or time of day information and / or season information can also be incorporated.This will then make it possible to make the relevant statements on trafficability, which will then also be made on the basis of the substitute information, more precise and, if necessary, make the assessment result more meaningful.

[0029] In one embodiment, the computing unit plans a detour trajectory even in a ground area classified as unknown if, based at least on the substitute information, navigability can be assumed. In particular, the additional advantageous procedures mentioned above can then be used in this context. It is also possible in this context for several alternative detour trajectories to be planned, in particular even suggested, in such a scenario. This also allows a selection to be made, particularly with regard to which of the then several planned detour trajectories offers the highest level of safety.

[0030] In one embodiment, if a local unknown zone with regard to drivability is detected in a drivable ground area, a check is carried out, in particular by the computing unit, to determine whether the avoidance trajectory through the unknown zone is planned or therefore must be continuously planned. If this is the case, it can be assessed based on provided comparison information whether the unknown zone can be assessed as drivable or not. Comparison information in this context can be information that was recorded when the driving trajectory was learned. For example, such a local unknown zone can manifest itself in the possible occurrence of a small unevenness. This could be, for example, a small rill extending across the roadway. This example is not intended to be exhaustive but merely to explain how such situations are to be understood.If in such a case it is known in the driving path that such an element, for example a gutter, is located there, which was driven over during learning and thus recognized as being drivable, and if this gutter extends laterally beyond the driving path and thus into the adjacent ground area, and is not recognized there but classified as a local unknown zone, it can be concluded on the basis of the information on how this configuration runs and is oriented that this element, in this case the gutter, can also be driven over in this ground area if it could be driven over in the driving path.

[0031] In one embodiment, the detour trajectory is planned as branching off from the driving trajectory. Thus, in one embodiment, it is possible for the computer-implemented method to determine a detour trajectory that is then used when the vehicle has already moved a certain distance while maneuvering along the learned driving trajectory. This specific planning of a detour trajectory thus enables a skillful and intelligent deviation from the driving trajectory in an intermediate region of the driving trajectory in order to avoid an obstacle. In addition to or instead of this, it is also possible for the detour trajectory to be planned as merging into the driving trajectory.This also makes it possible, for example, if an obstacle is located at the very beginning of the learned driving trajectory, for this driving trajectory not to be used for following the route at its starting point, but rather to be bypassed adjacent to it at the beginning by the bypass trajectory in order to be able to bypass the obstacle at the beginning of the driving trajectory. In such a constellation, a continuous junction into the driving trajectory via the bypass trajectory is then particularly advantageously enabled. In another exemplary embodiment, it is also possible for the bypass trajectory to be planned both as a branch off from the driving trajectory and as a rejoining of the driving trajectory.

[0032] In one embodiment, it is advantageous if the avoidance trajectory can be planned with a lateral distance from the travel trajectory that is greater than 1.00 m, in particular with a maximum lateral distance from the travel trajectory that is between 1.00 m and 3.00 m. This is another very advantageous embodiment, since such a tolerance zone can provide a relatively large distance by which the vehicle can deviate from the travel trajectory when avoiding an obstacle. Especially in such constellations, the above-mentioned detection, testing, and analysis of ground areas is advantageous.Since, at such large distances, which may be necessary to avoid an obstacle, ground areas may also be used that are significantly removed, at least laterally, from the areas covered by the driving path when learning the driving trajectory, diverse and flexible avoidance trajectories can be determined in a particularly advantageous manner. This advantageous embodiment in particular also generates very far-reaching evasive driving options as avoidance trajectories, so that a wide variety of large obstacles and / or obstacles positioned within the driving path can be avoided very reliably and in a very needs-based manner, at least semi-autonomously, and in particular fully autonomously, using the avoidance trajectory.Such large distances from driving trajectories are impossible even with conventional systems. As already explained at the beginning, this leads to the fact that only small obstacles or obstacles located only at specific locations can be avoided to a minimum. Otherwise, corresponding driving maneuvers must be aborted when following a learned trajectory. This is avoided with the proposed method, and in particular with this advantageous embodiment. The analysis of ground areas and their drivability is particularly advantageous.

[0033] In one embodiment, weather information, in particular at least current, and / or in particular at least current time of day information is provided to the computing unit. In one embodiment, this information is taken into account when assessing the drivability of a ground area and / or when determining the avoidance trajectory. This allows an even more detailed and precise assessment to be made even for ground areas with a specific ground area type. For example, if the ground area is made of earth or grass and therefore no permanently solid surface such as asphalt, concrete or pavement is present, this additional information can be used to assess its drivability even more precisely. For example, if this ground type is softened by prolonged and intense rain, the assessment of drivability may be different than in permanently fine and dry weather.This also applies to information regarding snowfall, ice, or the like. In this regard, one exemplary embodiment also allows for a weighting when assessing the navigability of a ground area based on the aforementioned parameters, so that a ground area fundamentally characterized by the soil type and / or unevenness and / or a specific zone exhibits different navigability suitability levels depending on different weather information and / or time of day information.

[0034] A further aspect of the invention relates to a method for performing an at least semi-autonomous, in particular fully autonomous, driving maneuver of a vehicle. In particular, this method comprises the following steps:

[0035] - Providing at least one detour trajectory as determined by a method according to the above-mentioned aspect or an advantageous embodiment thereof, in particular to an electronic vehicle guidance system;

[0036] - Avoiding the obstacle based on the avoidance trajectory, in particular controlled by the electronic vehicle guidance system, in particular as a fully autonomous driving maneuver. In particular, during this semi-autonomous driving maneuver, the vehicle is moved along the learned driving trajectory, in particular up to the section specified by the avoidance trajectory.

[0037] A further independent aspect of the invention relates to a method for carrying out an at least semi-autonomous, in particular fully autonomous, driving maneuver of a vehicle, in particular comprising the following steps:

[0038] - In particular, at least semi-autonomous driving of the vehicle along a provided, learned driving trajectory along which a vehicle has driven in a driving path;

[0039] - In particular, detecting an obstacle in the driving path with at least one detection unit of the vehicle and generating obstacle information;

[0040] - In particular, detecting ground areas of the environment that lie outside the driving path when following the learned driving trajectory along the driving path with at least one detection unit of the vehicle and generating ground area information therefor, and / or providing at least past ground area information at least of ground areas that lie outside the driving path when driving to learn the driving trajectory along the driving path;

[0041] - In particular, determining at least one detour trajectory branching off from the travel trajectory depending at least on the travel trajectory, the obstacle information and the ground area information and / or the past ground area information;

[0042] - In particular, bypassing the obstacle along the avoidance trajectory.

[0043] In this method, the driving maneuver, i.e., the at least partial retracing of the learned driving trajectory, is performed. If an obstacle occurs along the learned driving trajectory and / or within the driving path during this retracing, at least one possible avoidance trajectory is determined during this retracing. If such an avoidance trajectory is determined and assessed as suitable, the obstacle is avoided at least semi-autonomously, in particular fully autonomously, along the avoidance trajectory.

[0044] Advantageous embodiments of the above-mentioned first aspect of an independent method are to be regarded as advantageous embodiments of this method for carrying out an at least semi-autonomous driving maneuver.

[0045] In particular, these steps, as provided for determining the avoidance trajectory, can then also be carried out during the follow-up drive, in particular by the electronic vehicle guidance system, preferably by a computing unit of the vehicle guidance system.

[0046] In particular, the method for performing the driving maneuver provides that a control unit of the vehicle is configured to generate control signals for at least one functional unit of a vehicle, depending on information, in particular the at least one avoidance trajectory and / or the at least one approach trajectory, in order to drive through the sub-section. In particular, these control signals are generated by the control unit. The computing unit can be a component of the vehicle. However, it can also be arranged externally. For example, it can be arranged in a data center or in another vehicle. The driving maneuver is, in particular, a shunting maneuver. In particular, this means that the speed of the vehicle over the entire distance of the driving maneuver is less than or equal to 40 km / h, in particular less than or equal to 30 km / h.

[0047] A functional unit can be a braking system of the vehicle, a steering system of the vehicle, a vehicle assistance system, a vehicle drive system, etc.

[0048] A further aspect of the invention relates to an electronic vehicle guidance system with at least one computing unit and at least one detection unit. The vehicle guidance system is designed to carry out a method according to the above-mentioned aspect or an advantageous embodiment thereof. In particular, this method is carried out with the electronic vehicle guidance system.

[0049] A detection unit can be, for example, a camera, a lidar sensor, a radar sensor, an ultrasonic sensor, or the like. The detection unit can be a vehicle-mounted detection unit.

[0050] A further aspect of the invention relates to a vehicle with such an electronic vehicle guidance system. The vehicle may be a motor vehicle. It may, for example, be a passenger car or a truck.

[0051] A further aspect of the invention relates to a computer program or a computer program product, comprising instructions which, when the program is executed by a computer, such as a computing unit, cause the computer to carry out the method according to an above-mentioned aspect or an advantageous embodiment, in particular in at least partial steps thereof.

[0052] A further aspect of the invention relates to a control unit for a vehicle, which is configured to generate control signals for at least one functional unit of a vehicle, depending on information generated by a method according to the above-mentioned aspect or an advantageous embodiment thereof, in order to carry out the method at least along the avoidance trajectory, in particular along the learned travel trajectory and the avoidance trajectory. A further independent aspect of the invention relates to a method, in particular a computer-implemented method, for determining an approach trajectory for a vehicle to approach and merge into a learned travel trajectory. In particular, this method comprises the following steps:

[0053] In particular, providing a learned driving trajectory, along which a vehicle has driven in the environment in a driving path, to a computing unit;

[0054] In particular, providing current position information of the vehicle away from the travel trajectory to the computing unit;

[0055] In particular, providing ground area information on ground areas of the environment that lie in the area of ​​the current position of the vehicle and the travel trajectory to the computing unit;

[0056] In particular, determining at least one approach trajectory leading into the travel trajectory depending at least on the travel trajectory, the ground area information and the position information with the computing unit.

[0057] This method now makes it possible, when following a learned driving trajectory is desired, to initially approach the driving trajectory semi-autonomously, in particular fully autonomously, in a simple manner and now also through the flexible planning of an approach trajectory. Since this method now provides diverse and more flexible options for determining such an approach trajectory, a vehicle can thus also be guided into the learned driving trajectory in a more needs-based and situation-adapted manner and can essentially only be brought onto this learned driving trajectory in order to then be able to follow, in particular begin, along this learned driving trajectory at least semi-autonomously, in particular fully autonomously.

[0058] Embodiments of the above-mentioned aspect relating to the avoidance trajectory, in particular its determination by the computing unit, are also to be regarded as advantageous embodiments of the method for determining the approach trajectory according to the computer-implemented method. Thus, the above-mentioned steps, as mentioned in the computer-implemented method for determining the avoidance trajectory, can also be used as a basis for determining the approach trajectory or also apply thereto. The relevant aspects, as explained for the avoidance trajectory, also apply to the approach trajectory within a corresponding framework.

[0059] A further aspect of the invention then also relates to a method for carrying out an at least semi-autonomous driving maneuver of a vehicle, in particular comprising the following steps:

[0060] - Providing at least one approach trajectory as determined by a method according to the above-mentioned aspect or an advantageous embodiment thereof;

[0061] - Approaching and merging the vehicle into the learned driving trajectory based on the approach trajectory.

[0062] For this aspect of determining the approach trajectory, a method for performing an at least semi-autonomous driving maneuver of a vehicle is also provided, as explained above and is performed when actually following the learned driving trajectory. Furthermore, for this aspect of the approach trajectory, the aspects relating to the electronic vehicle guidance system, the vehicle, the computer program, the computer-readable data storage device, and the control unit are also provided accordingly. The explanations presented above also apply in this context to the aspect of the approach trajectory.

[0063] In one embodiment, it is possible for both the determination of an approach trajectory and the determination of an avoidance trajectory to be carried out in a common process, in particular a computer-implemented method, and / or a method for performing an at least semi-autonomous driving maneuver. Thus, it is also possible for such a follow-up drive to initially approach the learned driving trajectory using such a specific approach trajectory, then merge into it, then drive along the driving trajectory, and then, if an obstacle occurs, avoid this obstacle based on a specific avoidance trajectory.

[0064] A further aspect of the invention relates to a method, in particular a computer-implemented method, for determining an additional trajectory, such as a bypass trajectory for bypassing an obstacle or an approach trajectory for a vehicle to approach and enter a learned driving trajectory, in addition to an existing, in particular learned driving trajectory, comprising the following steps:

[0065] - Providing a learned driving trajectory, along which a vehicle has driven in an environment in a driving path, to a computing unit;

[0066] - Providing current position information of the vehicle away from the travel trajectory to the computing unit and / or providing obstacle information about an obstacle along the travel path when subsequently following the travel trajectory in the environment to the computing unit;

[0067] - Check whether the current position of the vehicle is outside the driving path and / or check whether bypassing the obstacle requires the vehicle to be moved outside the driving path, using the computing unit;

[0068] - if the check shows that the position of the vehicle is outside the travel path and / or avoiding the obstacle requires that the vehicle must be moved outside the travel path, providing ground area information to the computing unit at least on ground areas of the environment which are outside the travel path, in particular in the area of ​​the current position of the vehicle and / or in the area of ​​the obstacle;

[0069] - Determining at least one additional trajectory which flows into the travel trajectory or branches off from the travel trajectory depending at least on the ground area information, the travel trajectory and / or the travel path, and the position information and / or the obstacle information with the computing unit.

[0070] Advantageous embodiments of the above-mentioned aspects of an independent method are to be regarded as advantageous embodiments for this method for determining an additional trajectory.

[0071] Embodiments of the invention are explained in more detail below with reference to schematic drawings. They show:

[0072] Fig. 1 is a schematic representation of an environmental scenario with an embodiment of a vehicle according to the invention and an embodiment of an electronic vehicle guidance system according to the invention;

[0073] Fig. 2 is a schematic plan view of an embodiment for maneuvering a vehicle; Fig. 3 is a schematic plan view of another embodiment of a scenario for driving a vehicle;

[0074] Fig. 4 is a plan view of yet another embodiment of a scenario for driving a vehicle;

[0075] Fig. 5 is a schematic plan view of yet another embodiment for driving a vehicle; and

[0076] Fig. 6 is a schematic plan view of an embodiment for driving a vehicle.

[0077] In the figures, identical or functionally identical elements are provided with the same reference symbols.

[0078] Fig. 1 shows a schematic representation of a vehicle 1. The vehicle 1 has an electronic vehicle guidance system 2. The electronic vehicle guidance system 2 preferably has at least one computing unit 3. In particular, the vehicle 1 has a control unit 4. The control unit 4 can be a component of the electronic vehicle guidance system 2. It is possible for the computing unit 3 to be a component of the control unit 4. However, the computing unit 3 can also be arranged separately. The computing unit 3 can be arranged in the vehicle or externally thereto, for example in a data center or in another vehicle. In addition, the electronic vehicle guidance system 1 has a detection unit 5 for detecting the surroundings of the vehicle 1. The detection unit 5 can have at least one camera and / or at least one ultrasonic sensor and / or at least one radar sensor and / or at least one lidar sensor.The detection unit 5 may comprise only one such component. However, multiple components may also be provided. The detection unit 5 may also be designed as an environment detection system.

[0079] In particular, a so-called all-round vision system can also be provided in this context.

[0080] A computer program can be a component of the computing unit 3 or the control unit 4. This program can be designed to include commands that, when executed by a computer, in particular the computing unit 3, cause the computer to perform a method for determining a bypass trajectory and / or an approach trajectory for a vehicle 1 for at least semi-autonomously bypassing an obstacle and / or for approaching and merging with a learned driving trajectory.

[0081] Furthermore, Fig. 1 shows an environmental situation 6 in an environment 7. In particular, a building 8 with a parking area 9, for example a garage or a carport or the like, is shown here.

[0082] Furthermore, a roadway 10 is shown in the environment 7 according to the depicted environmental scenario 6. This roadway 10 is intended for driving with the vehicle 1. In the exemplary embodiment, this roadway 10 leads to the infrastructure facility, here the parking area 9.

[0083] It is provided that the vehicle 1 can drive at least semi-autonomously, in particular fully autonomously, at least to the parking area 9. For this purpose, one exemplary embodiment provides for the vehicle 1 to drive a driving trajectory on the roadway 10 during a learning drive. This driving trajectory is recorded. It is possible for the driving trajectory to be recorded, for example, using odometry data from the vehicle 1. In addition to or instead of this, however, other localization methods can also be used. For example, the detection of the surroundings 8 with the at least one detection unit 5 can also be provided here, and this information can be used for localization. In addition, however, a SLAM method, for example, can also be used.

[0084] In one embodiment, floor areas 11 and 12 are also recorded during this learning run.

[0085] Generally speaking, floor areas 11 and 12 are areas that represent the subsurface. They are therefore particularly characterized by ground-side surface areas or floor surfaces. In contrast, objects can also be present in the environment 8, which can also be detected accordingly. In this context, objects can be temporarily static or dynamic, as well as permanently stationary objects. They are arranged on the floor surfaces of the floor areas. In this context, objects are particularly raised elements that protrude upwards from the floor surface, in particular protrude at least 30 cm upwards, for example.

[0086] Fig. 2 shows a simplified scenario in a schematic plan view. Fig. 2 shows a situation in which a learned driving trajectory 13 already exists. In addition, a driving lane 14 is also shown here. The vehicle 1 has moved in the driving lane 4, in particular during the learning run for learning the driving trajectory 13. The driving lane 14 is therefore not just a thin line, as is characterized in particular by the driving trajectory 13, but rather a surface strip characterized by a width perpendicular to this driving trajectory 13. In Fig. 2, only a lateral fictitious boundary 14a of the driving lane 14 is shown by way of example. In this regard, the driving lane 14 can have a width that corresponds to a width of the vehicle 1. It is also possible for the driving lane 14 to have a width that is slightly greater than the width of the vehicle 1.

[0087] In this context, it is provided in one embodiment that the surface area of ​​the driving tube 14 is formed as a central floor area, which is classified as being drivable with the vehicle 1.

[0088] In the traffic scenario shown in Fig. 2, in which the vehicle 1 performs a follow-up drive, which is an example of an at least semi-autonomous driving maneuver of the vehicle 1, the present example shows that at least one obstacle 15 is present. This obstacle 15 is positioned such that the vehicle 1 cannot drive past the obstacle 15 solely along the learned driving trajectory 13. Therefore, in such an embodiment, it is provided that a detour trajectory for detouring around the obstacle 15 is determined.

[0089] It is possible that, in one embodiment, the obstacle 15 is detected, for example, with the detection unit 5. This allows obstacle information to be generated in this embodiment that was detected by the detection unit 5 itself. In particular, this occurs during the tracking drive.

[0090] Furthermore, it is provided that the surroundings 8 are detected, in particular with at least the detection unit 5. In addition to the surface area of ​​the travel path 14, ground areas located laterally thereto, for example, the ground area 11, are also detected. Such a ground area directly borders the travel trajectory 13 and / or the travel path 14 laterally.

[0091] In one embodiment, it is provided that a detour trajectory for avoiding the obstacle 15 is determined by a computer-implemented method, in particular with the computing unit 3. For this purpose, the following steps are carried out in particular:

[0092] - Providing the learned driving trajectory 13, along which the vehicle 1 has driven in the environment 7 in the driving path 14, to the computing unit 3;

[0093] - Providing obstacle information about the at least one obstacle 15 during the subsequent following of the travel trajectory 13 in the environment 7 to the computing unit 3;

[0094] - Providing at least ground area information on ground areas 11, 12 of the environment 7, which along the travel path 14 are located at least partially outside the travel path 14, to the computing unit 3;

[0095] - Determining at least one detour trajectory 16 branching off from the travel trajectory 13 depending on the travel trajectory 13, the obstacle information and the ground area information with the computing unit 3.

[0096] It is also possible that, to determine such a detour trajectory 16, as shown here as an example in Fig. 2, ground area information acquired and generated during the follow-up drive is no longer used as a basis, but rather, in addition to or instead of the detour, historical ground area information is used as a basis. This information can be that acquired for ground areas 11 and / or 12 during the learning drive for learning the driving trajectory 3.

[0097] When determining the detour trajectory 16, which in this context also constitutes planning of the detour trajectory 16, reference ground area information can be provided to the computing unit 3. In one embodiment, it is possible for currently acquired ground area information and / or past ground area information provided to the computing unit 3 to be compared with such reference ground area information. Depending on this comparison, a ground area 11 and / or 12 can be classified. A classification into a drivable ground area or an impassable ground area or an unknown ground area is possible here. In the embodiment in Fig. 2, a situation is shown in which the vehicle

[0098] I is already partially executing this follow-up journey along the learned travel trajectory 13. It is therefore intended here that the detour trajectory 16 is planned to branch off from the travel trajectory 13 and then again merge into the travel trajectory 13. This is intended in the present case because the obstacle 15 is positioned such that the learned travel trajectory 13 is used, or can be used, both before and after the obstacle 15 for the driving maneuver.

[0099] In the embodiment shown in Fig. 2, the ground area information is available in such a comprehensive manner and the ground area 11 provided here is dimensioned and recorded in such a way that the entire avoidance trajectory 16 can be planned therein. If this ground area 11 is thus classified as drivable, this avoidance trajectory 16 can be planned entirely within this ground area 11. As is also provided in this context, the avoidance trajectory 16 extends laterally outside the driving path 14. In particular, it thus protrudes at least partially into the ground area

[0100] II. In particular, it is provided that the bypass trajectory 16 can generally be planned with a maximum lateral distance a from the travel trajectory 13 that is greater than 1.00 m, in particular with a maximum lateral distance a from the travel trajectory that is between 1.00 m and 3.00 m. This also makes it possible for the vehicle 1 to be moved away from the travel trajectory 13, for example, by at least half a vehicle width, preferably also by an entire vehicle width, in this lateral direction.

[0101] Fig. 3 shows a schematic plan view corresponding to Fig. 2. In contrast to Fig. 2, it is provided here that ground area information has been recorded currently and in this regard only from the ground area 11. Here, too, it is provided, by way of example, that the ground area 11 is assessed as passable. In the scenario presented here, it is therefore provided that an area 17 which lies in the lateral direction to the travel trajectory 13 between the travel trajectory 13 and the ground area 11 does not have current ground area information. However, since this intermediate area 17 lies in the travel path 14, such a scenario can be based on ground area information known from the past. In particular, it is known here, for example due to localization and / or other position information, that this intermediate area 17 is an area of ​​the travel path 14.Since the driving lane 14 was characterized as passable during the learning run and also saved, it can be concluded in this scenario in Fig. 3 during this follow-up run that this area 17 is passable for the vehicle 1, particularly at least with regard to its strength. This is true even if this area cannot currently be directly detected during the follow-up run and / or no current ground area information is available in this regard.

[0102] Here, too, the bypass trajectory 16 can then be completely planned accordingly.

[0103] Fig. 4 shows a schematic plan view of a further exemplary embodiment. In this exemplary embodiment, the environment 8 is such that, during this follow-up drive, the ground area 11 in the current position can only be detected over an area that is insufficient for planning the entire avoidance trajectory 16. Therefore, in the exemplary embodiment according to Fig. 4, in order to avoid the obstacle 15, ground area information from a ground area 18 must also preferably be taken into account, which in one exemplary embodiment cannot yet be detected during the follow-up drive starting from the position of the vehicle 1, in particular cannot be detected with the detection unit 5. In such an exemplary embodiment, it is then possible for ground area information from the ground area 18 that was detected, for example, during the learning drive of the vehicle 1 to be preferably provided and used.Thus, in such an embodiment, if this ground area information of the ground area 18 characterizes the ground area 18 as passable during the assessment, it is also possible to plan a complete detour trajectory 16. This then runs, as shown by way of example in Fig. 4, through the ground area 11 assessed as passable and the ground area 18 assessed as passable using this substitute information. In particular, it is therefore provided that in this current situation, as shown in Fig. 4, the ground area 18 is initially classified as unknown with regard to passability. This multi-stage assessment scenario for the ground area 18, if substitute information is available here, then also enables it to be classified as passable or impassable.It is possible not only in this context that additional current information, such as current weather information and / or current time of day information, is taken into account in order to increase the meaningfulness of this assessment result.

[0104] However, such additional information can be taken into account not only in such a scenario, in particular for the assessment, in particular for the better assessment of such an unknown soil area 18, but also for the assessment of a currently detectable soil area 11 and / or 12.

[0105] Fig. 5 shows a schematic representation of a further scenario in which a follow-up drive of the vehicle 1 takes place on the basis of the learned driving trajectory 13. In this exemplary embodiment, it is provided that a ground area 11 can be currently detected, in particular, for example, at least with the detection unit 5. The ground area 11 can also be classified as drivable here, except for a local unknown zone 11a in this ground area 11. With regard to the unknown zone 11a, no current ground area information is available in the exemplary embodiment. In such a scenario, in which the unknown zone 11a also extends into the driving path 14, it is possible that substitute information is again taken into account here.For example, such a thin and / or strip-like and / or otherwise locally geometrically formed unknown zone 11a can be assigned substitute information originating from the past. For example, this area may have been crossed during learning of the travel trajectory 13, particularly in the area of ​​the travel path 14. In this regard, it may be known, for example, that a shipping channel is present here. This was recognized as being navigable in the area of ​​the travel path 14. If this unknown zone 11a extends positionally and / or geometrically in the scenario according to Fig. 5 at a corresponding point in the travel path 4 and extends, as shown in Fig.5, laterally out of the navigation channel 14 into the ground region 11, it can be concluded in such a scenario that this navigation channel extends from the navigation channel 14 into the ground region 11 and that it can therefore also be crossed in this unknown zone 11a. Therefore, in such a scenario, the unknown zone 11a can be classified as navigable. It should be mentioned at this point that the exemplary explanation of such an unknown zone 11a based on such a navigation channel is by no means to be understood as conclusive. A navigation channel is to be understood merely as an example and symbolic for such an element.For a variety of other aspects that can be assigned to such an unknown zone 11a and, based on the scenario explained, occur in particular in the driving path 14 due to the corresponding local and / or geometric specification and then extend beyond it, the information of this element in the driving path 14 can be used to infer corresponding information outside the driving path 14. In such an embodiment, if the unknown zone 11a can be classified as passable, a bypass trajectory 16 can be planned that also extends into or crosses this unknown zone 11a.

[0106] In the examples shown in Fig. 2 to Fig. 5, it is then possible to avoid the obstacle 15. In particular, this is done based on this planned avoidance trajectory 16.

[0107] In all the exemplary embodiments explained so far, as well as in the further exemplary embodiments yet to be explained, it should be noted that, in addition to the previous scenarios, the detection of objects in the environment 6 is also possible, in particular at least during the tracking of the vehicle 1. In this context, objects are different environmental elements than ground areas. In this context, objects are elements extending upwards from the ground surface of a ground area, in particular elements extending upwards by at least 30 cm.

[0108] In other embodiments, it is also possible for the bypass trajectory 16 not to be defined as a branching and joining trajectory to the travel trajectory 13. It can also be defined solely as a branching bypass trajectory 16 or solely as a joining trajectory.

[0109] Fig. 6 shows a schematic top view of a scenario in which a learned driving trajectory 13 is also shown. To carry out the retracing of the driving trajectory 13, an approach trajectory 19 is provided here. The illustration in Fig. 6 shows a situation in which the vehicle 1 is still positioned at a distance from the driving trajectory 13 before the start of the retracing. In order to be able to carry out the retracing of the driving trajectory 13, the vehicle 1 must move towards the driving trajectory 13. This is achieved by the approach trajectory 19, which is planned such that it merges into the driving trajectory 13. In order to be able to determine a suitable approach trajectory 19, the following, in particular computer-implemented, method is carried out.

[0110] Providing the learned driving trajectory 13, along which the vehicle 1 has driven in the environment 7 in the driving path 14, to a computing unit 3;

[0111] Providing current position information of the vehicle 1 away from the travel trajectory 13 to the computing unit 3; Providing ground area information about ground areas 11, 12 of the environment 7 that lie in the area of ​​the current position of the vehicle 1 and the travel trajectory 13 to the computing unit 3;

[0112] Determining at least one approach trajectory 19 leading into the travel trajectory 13 depending at least on the travel trajectory 13, the ground area information and the position information with the computing unit 3.

[0113] In the example shown, the ground area 11 is present. Vehicle 1 is also located within this area. It is also possible that the ground area 11 extends in such a way that the vehicle 1, in its current position, is not yet or not yet completely positioned within the ground area 11. If this ground area 11 is assessed as passable according to the analysis presented, the approach trajectory 19 can be planned.

[0114] In this context, it is preferably possible for such planning of an approach trajectory 19 to be possible even when the vehicle 1 is at a lateral distance a from the travel trajectory 13, which can be up to 3.00 m. In particular, this is thus possible when the distance a is greater than 1.00 m. This means that the method can also be used to determine an approach trajectory 19 when the vehicle 1, in its current position, is still relatively far laterally from the travel trajectory 13. Especially in such constellations, it is then particularly advantageous to carry out corresponding analyses based on the ground regions 11, 12 adjacent to the travel trajectory 13 in order to be able to assess the drivability of such ground regions 11, 12.This also makes it possible to plan more diverse and / or longer and, above all, more extensive approach trajectories 19 in order to be able to move the vehicle 1 at least semi-autonomously, in particular fully autonomously, toward the travel trajectory 13. This enables scenarios that occur before a follow-up drive in order to subsequently be able to carry out the follow-up drive along the travel trajectory 13.

[0115] It is also possible for the embodiment according to Fig. 6 to be combined with an embodiment according to Fig. 2 to Fig. 5. This then also makes it possible for the vehicle 1 to initially approach the travel trajectory 13 using such a specifically determined approach trajectory 19 and then, when subsequently following the learned travel trajectory 13, to determine an avoidance trajectory 16, as explained above, if a corresponding obstacle 15 occurs. In another embodiment, it is also possible for the scenario in Fig.

[0116] 6, an obstacle 15 is present in the area toward the travel trajectory 13, so that the follow-up trajectory 19 can also be planned such that this obstacle can be avoided on the way to the travel trajectory 13. This then also enables a combined approach and avoidance trajectory. This can also be planned accordingly.

Claims

Patent claims 1 . Method, in particular a computer-implemented method, for determining a bypass trajectory (16) for a vehicle (1) for at least semi-autonomously bypassing an obstacle (15) in an environment (7), comprising the following steps: Providing a learned driving trajectory (13), along which a vehicle (1) has driven in the environment (7) in a driving path (14), to a computing unit (3); Providing obstacle information about an obstacle (15) to the computing unit (3) when subsequently following the travel trajectory (13) in the environment (7); Providing at least ground area information on ground areas (11, 12) of the environment (7) that lie along the travel path (14) outside the travel path (14) to the computing unit (3); Determining at least one detour trajectory (16) branching off from the travel trajectory (13) as a function of the travel trajectory (13), the obstacle information and the ground area information with the computing unit (3).

2. Method according to claim 1, wherein, depending on the ground area information, navigability information is determined with the computing unit (3) and this internally determined navigability information is taken into account when determining the bypass trajectory (16).

3. Method according to claim 1 or 2, wherein the ground area information is compared with reference ground area information with the computing unit (3) and depending on the comparison it is determined whether a ground area (11, 12) is classified as passable.

4. The method according to claim 3, wherein drivable soil types and / or drivable unevennesses and / or drivable zones are specified as reference soil area information, and / or impassable soil types and / or impassable zones and / or impassable unevennesses are specified as reference soil area information.

5. The method according to claim 3 or 4, wherein when a ground area (11, 12) is classified as passable, the bypass trajectory (16) is planned at least in part on this ground area (11, 12) with the computing unit (3).

6. Method according to one of the preceding claims 3 to 5, wherein when a ground area (11, 12) is classified as unknown for driving and / or cannot yet be detected by a detection unit (5) from the current position of the vehicle (1), but is not classified as not impassable, substitute information for assessing this ground area (11, 12) is taken into account by the computing unit (3) for determining the drivability.

7. The method according to claim 6, wherein past ground area information is provided to the computing unit (3) as replacement information, which was recorded when driving through the environment (7) for learning the driving trajectory (13), in particular with at least one recording unit (5) of the vehicle (3).

8. The method according to claim 7, wherein a detour trajectory (16) is also planned in a ground area (11, 12) which has been classified as unknown if drivability can be assumed on the basis of the substitute information.

9. Method according to one of the preceding claims, wherein when a local unknown zone (11a) is detected in a drivable ground area (11, 12), it is checked with regard to drivability whether the bypass trajectory (16) through the unknown zone (11a) is or must be planned, and if this is the case, it is assessed on the basis of provided comparison information whether the unknown zone (11a) can be assessed as drivable or not.

10. Method according to one of the preceding claims, wherein the bypass trajectory (16) is planned as branching off from the travel trajectory (13) and / or as leading into the travel trajectory (13).

11. Method according to one of the preceding claims, wherein the bypass trajectory (16) can be planned with a lateral distance (a) to the travel trajectory (13) that is greater than 1.00 m, in particular with a maximum lateral distance (a) to the travel trajectory (13) that is between 1.00 m and 3.00 m.

12. Method according to one of the preceding claims, wherein weather information and / or time of day information is provided to the computing unit (3), this information is taken into account when assessing the drivability of a ground area (11, 12) and / or when determining the detour trajectory (16).

13. A method for carrying out an at least semi-autonomous driving maneuver of a vehicle (1), comprising the following steps: Providing at least one detour trajectory (16) as determined by a method according to one of the preceding claims; Avoiding the obstacle (15) based on the avoidance trajectory (16).

14. A method for carrying out an at least semi-autonomous driving maneuver of a vehicle (1), comprising the following steps: at least semi-autonomous driving of the vehicle (1) along a provided, learned driving trajectory (13) along which a vehicle (1) has driven in a driving path (14); Detecting an obstacle (15) in the travel path (14) with at least one detection unit (5) of the vehicle (1) and generating obstacle information; Detecting ground regions (11, 12) of the surroundings (7) that lie outside the travel path (14) when traveling along the learned travel trajectory (13) along the travel path (14) with at least one detection unit (5) of the vehicle (1) and generating ground region information therefor, and / or providing at least past ground region information of at least ground regions (11, 12) that lie outside the travel path (14) when traveling to learn the travel trajectory (13) along the travel path (14); Determining at least one detour trajectory (16) branching off from the travel trajectory (13) depending at least on the travel trajectory (13), the obstacle information and the ground area information and / or the past ground area information; Avoid the obstacle () along the avoidance trajectory ().

15. Electronic vehicle guidance system (2) with at least one computing unit (3) and with at least one detection unit (5), wherein the vehicle guidance system (2) is designed to carry out a method according to one of the preceding claims.

16. Vehicle (1) with a vehicle guidance system (2) according to claim 14.

17. A computer program comprising instructions which, when executed by a computer, cause the computer to carry out the method according to one of the preceding claims 1 to 12.

18. A computer-readable data carrier on which the computer program according to claim 17 is stored.

19. Control device (4) for a vehicle (1), which is designed to generate control signals for at least one functional unit of a vehicle (1) depending on information as generated by a method according to one of the preceding claims 1 to 12, in order to carry out the driving at least along the detour trajectory (16).

20. Method, in particular a computer-implemented method, for determining an approach trajectory (19) for a vehicle (1) for approaching and entering a learned driving trajectory (13), comprising the following steps: Providing a learned driving trajectory (13), along which a vehicle (1) has driven in the environment (7) in a driving path (14), to a computing unit (3); Providing current position information of the vehicle (1) away from the travel trajectory (13) to the computing unit (3); Providing ground area information on ground areas (11, 12) of the environment (7) that lie in the area of the current position of the vehicle (1) and the travel trajectory (13) to the computing unit (3); Determining at least one approach trajectory (19) leading into the travel trajectory (13) depending at least on the travel trajectory (13), the ground area information and the position information with the computing unit (3).

Images