Method for supporting vehicle guidance based on travel lane and boundary estimation, assistance system and motor vehicle

Abstract

The invention relates to a method and an assistance system (5) for supporting the longitudinal and transverse guidance of a motor vehicle (2) on a road (1). The invention also relates to a correspondingly designed motor vehicle (2). In the method, a course (10) of a boundary (3), which delimits a travelable area of the road, and a travel lane (9) of the motor vehicle (2) are estimated on the basis of the acquired environment data. At least one distance (11) from the travel lane (9) to the estimated limit (10) is determined along the travel lane (9), said distance being given transversely to the travel direction. A data set (9, 11) is then generated and provided from the travel lane (9) and the determined distance (11), the travel lane (9) being specified in the data set (9, 11), and additionally a travelable area along the travel lane (9) being specified by the distance (11).

Description

Technical Field

[0001] The invention relates to a method and an assistance system for supporting the longitudinal and transverse guidance of a motor vehicle. The invention also relates to a correspondingly configured motor vehicle. Background Art

[0002] For various current and future assistance functions in motor vehicles, it is useful to determine the path that the respective motor vehicle may travel. This can be described as a so-called driving corridor or driving gallery of the motor vehicle. For example, such a driving corridor can be used to identify objects located in the respective environment. Such objects can then be used as target objects for controlling, for example, adaptive cruise control or other assistance functions and advanced or fully automated driving functions.

[0003] For example, DE 102018 129 079A1 describes a method for generating a driving corridor for a vehicle along a roadway. First, an initial corridor seeding is generated without taking into account all other vehicles on the roadway. Then, based on the identified objects, a first target limit of the corridor is generated for one side of the roadway. Furthermore, a second target limit is generated on the opposite second side of the roadway. Based on these limits, the first corridor seeding is adapted, thereby generating a driving corridor for the vehicle.

[0004] For example, DE 102009 009 047 A1 describes a method for object detection. In this method, a range image is acquired with the aid of a sensor system and a depth map of the environment is determined therefrom. A free space limit line is identified in the range image, which defines the extent of the obstacle-free area. Outside and along the free space limit line, the depth map is segmented by forming equal-width segments from pixels at the same or similar distance from a plane. The height of each segment is estimated as a part of an object located outside the obstacle-free area, so that each segment is characterized by the two-dimensional position of a base point and its height.

[0005] In principle, a robust and accurate recognition and representation of the corresponding environment can be used, for example, for corresponding predictions of driving corridors or trajectory planning, for example for emergency maneuvers. However, this remains a major challenge. For example, the map data that is often used for this purpose often has the problem that it is not up to date, i.e. it cannot adapt quickly enough to dynamic environmental changes, such as construction or blockades. The segmentation of the camera images of the environment that is often used is not always reliable and can be affected by many factors, such as environmental influences (such as heavy rain, glaring sunlight, snow, etc.). Another problem is that control devices or embedded systems with very limited computing power are usually used in motor vehicles to predict driving corridors and / or assistance functions based thereon, but the corresponding data and information must be processed and interpreted in real time or as quickly as possible during the operation of the motor vehicle.

[0006] An example of a method for establishing an environment model for a vehicle is described in DE 102014212 487A1. In this case, a roadway is acquired based on objects and / or free space limits and / or lane limits, which indicates a freely drivable area around the vehicle. The roadway includes at least one roadway segment, which in turn includes at least one roadway segment limit. The distance associated with the vehicle is determined. The roadway can also be provided to a driver assistance system. This method for establishing an environment model can reduce the cost required to establish a new driver assistance system and make it easier to integrate new sensor systems. Summary of the invention

[0007] The object of the present invention is to achieve a particularly robust, efficient and effective assisted or at least partially automated guidance of a vehicle.

[0008] According to the invention, this object is achieved by the subject matter according to the invention. Possible improvements and further developments of the invention are disclosed in the description and the drawings.

[0009] The method according to the invention can be used to support the longitudinal and transverse guidance of a motor vehicle on or along a road. In the method steps of the method according to the invention, environmental data are acquired which reflect or characterize the corresponding environment located ahead, in particular in the direction of travel of the motor vehicle. Acquiring environmental data can mean or include capturing environmental data by means of an environmental sensor system, in particular by means of a plurality of different environmental sensors, receiving environmental data via a data connection or interface and / or retrieving environmental data from a data memory.

[0010] In another method step of the method according to the invention, the trend of at least one limit or the trend of the corresponding limit feature is predicted, i.e. estimated, based at least on the acquired environmental data, thereby limiting the drivable area of ​​the road in the corresponding environment, and predicting, i.e. estimating the driving path that the motor vehicle is expected to drive through. The driving path extends from the current position of the motor vehicle in its driving direction away from the motor vehicle along the road or through the corresponding surrounding environment. In order to estimate the trend of the limit, for example, guardrails, grass, lanes or lane markings and / or the like can be detected or identified in the acquired environmental data. Therefore, the trend of the limit can be directly obtained from the environmental data, i.e. at least without considering the inevitable uncertainties contained therein. The corresponding predetermined estimation model can be used to finally estimate the trend of the limit and / or the driving path. This can be, for example, an algorithmic model, a simulation model, a training model based on machine learning, such as an artificial neural network, etc. In order to estimate the trend of the limit and / or the driving path, in addition to the environmental data, other data can also be acquired and used, such as current, immediately preceding and / or historical data on the operation or status of the motor vehicle, such as its steering angle, steering angle change rate, yaw rate, etc., map data, navigation data, movement data of other traffic participants in the corresponding environment, etc. Ultimately, this allows for a particularly precise, robust and reliable estimation of the course of the limits and / or the driving corridor.

[0011] In a further method step of the method according to the invention, at least one distance of the estimated course from the driving corridor to a limit is determined along the estimated driving corridor, which distance is given transversely to the driving direction of the motor vehicle, i.e. transversely to the longitudinal extension of the motor vehicle. For example, the distance can be determined at a plurality of or all points of the driving corridor. In particular, the distance can be determined for both sides or in two directions, i.e. to the left and to the right when viewed in the driving direction along the driving corridor. Depending on the identified or estimated limit, distances to a plurality of different limits can also be determined here.

[0012] In a further method step of the method according to the invention, a combined or integrated data set is generated and provided from the estimated course of the driving corridor and at least one determined distance (i.e., determined distance or distance value). In this case, it is also taken into account at what location, i.e., at what spatial position, the corresponding distance is given. In the data set generated and provided in this way, the estimated driving corridor and, in addition, at least by at least one distance, at least implicitly or explicitly, specify the area along the driving corridor that can be driven by the motor vehicle.

[0013] The driving corridor estimation has been expanded to include the drivable area. By combining or integrating in a data set, in which the distance and the driving corridor can be specified or included in a direct link to each other, a particularly efficient encoding and availability of this data can be achieved, for example compared to two separate data sets for the driving corridor and the drivable area, or compared to the subsequent determination of the driving corridor from a data set in which only the drivable area was initially specified, or vice versa. The data set proposed here, i.e. the combination from the driving corridor estimation and the limit estimation, can represent a particularly compact representation of the corresponding data or the corresponding environment. This makes it particularly easy for subsequent algorithms or auxiliary functions or auxiliary systems to remove and process these data from the data set. By this, during the operation of the corresponding motor vehicle, the corresponding auxiliary function can be executed or used particularly effectively and effortlessly, that is, even with corresponding limited hardware or computing resources, especially in real time.

[0014] For example, compared with a complete pixel-by-pixel segmented 3D environment model, the proposed representation of the drivable area by the distance from the driving corridor to its inner edge is particularly effective, because in addition to the driving corridor that is required for many auxiliary functions, only very little additional data needs to be saved. Therefore, for example, embedded systems can be used, which usually have relatively small computing power and relatively small data storage. Another advantage of the present invention is that by representing or describing the drivable area from the perspective of the driving corridor, that is, relative to the driving corridor, the drivable area that is actually only related to the longitudinal and lateral guidance of the motor vehicle is automatically specified or stored. For example, other areas that are drivable in principle but separated from the driving corridor by (possibly non-drivable) boundaries and are therefore not related to the motor vehicle can be kept out of consideration. For example, the search space for trajectory planning (whether along the driving corridor or for maneuvers to leave the driving corridor) can be correspondingly limited or constrained. As a result, the corresponding trajectory planning can also be performed in a particularly efficient, effective and low-cost manner.

[0015] For example, the driving path can be specified as a polygon, which contains, in addition to the spatial x and y coordinates, additional information or data, such as the orientation or direction of the driving path perpendicular thereto and the width. For each point or section of the driving path or polygon, the distance to at least one of the limits therein can then be stored in the data set. Likewise, the course of the corresponding limit can be specified or stored in the data set, for example as another polygon. The drivable area along the driving path can then be read or extracted from the data set or from the polygon for the driving path extended with the corresponding data or information.

[0016] For estimating the course of the driving corridor or the limit, for example, a known information filtering mechanism can be used, which uses an information matrix (i.e., the inverse of the covariance matrix) and an information vector or information state vector linked to the estimation vector via the information matrix. As a result, the course of the driving corridor and the limit can be estimated in one working step or operation, i.e., also particularly quickly, efficiently and with low expenditure.

[0017] In a possible embodiment of the present invention, a plurality of various, especially different data are combined with each other, especially fused, for estimating the limits and driving channels. In particular, data from a plurality of different types of sensors can be combined here. Other types of data that do not need to come from traditional sensors can also be used, such as trajectory or group movement data, map data, etc. indicating the movement of other vehicles in the corresponding environment. For example, several hypotheses for the road trend can be generated from various data types, and thus several hypotheses are generated at least indirectly for the limit trend and / or directly for the limit trend. These hypotheses can be combined with each other. For this purpose, the hypothesis can be weighted, averaged, etc., for example. These hypotheses can be polygons from different sources, for example, based on different data types or combinations of data types. This can, for example, involve road limits, road markings, the history of other traffic participants, group trajectories, etc. By using a plurality of different data types proposed here, more accurate, robust and reliable estimates can be achieved. For example, problems or effects involving or hindering a single sensor or a single data type may not affect other sensors or data types. Therefore, under many different conditions, there is always a particularly high probability that at least one data type can be obtained, from which the trend of the limit can be obtained or correctly estimated.

[0018] In another possible embodiment of the present invention, a distinction is made between a hard limit that cannot be driven over and a soft limit that can be driven over. Thus, for example, the distance can be determined consistently only to the hard limit or only to the soft limit, or a first distance to the soft limit and a second distance to the hard limit can be determined. These distances can all be specified in the data set. Therefore, the drivable area can or will be defined in a corresponding hierarchical manner. For example, for normal operation without errors and obstacles, in which the motor vehicle is only guided to the nearest soft limit in the corresponding direction, this drivable area can be defined by the distance to the soft limit. On the other hand, for emergency operation, in which driving over the soft limit can be allowed and the motor vehicle can be guided to the nearest hard limit in the corresponding direction (for example, to avoid a collision with another traffic participant), the drivable area can or will be defined by the nearest hard limit. Therefore, it is possible to achieve vehicle guidance adapted to the situation in a particularly simple and effective manner.

[0019] In a possible improvement of the present invention, only the distance from the driving channel to the closest hard limit on both sides of the driving channel and / or only the distance from the driving channel to the detectable or estimated soft limit farthest from both sides of the driving channel is determined and specified in the data set. Therefore, the distance specified in the data set can be the distance to the soft limit hypothesis farthest from or farthest from the driving channel in two lateral directions and / or the distance to the hard limit hypothesis closest to. In other words, if multiple soft limits and / or multiple hard limits are detected at different distances from the driving channel, the closer limit among the multiple soft limits or the farther limit among the multiple hard limits can be rejected. This means that the corresponding data processing work can be saved. By limiting the limits used or considered here, the entire relevant drivable area can only be specified or limited in a particularly simple and effective manner. If multiple distances to different types of limits are specified in the data set, they can be indicated or marked accordingly. This can also realize corresponding adaptive functions based on the data set in a particularly simple, efficient and low-cost manner, such as trajectory planning.

[0020] In a possible development of the invention, lane markings or road markings are used as soft limits, i.e. are taken into account or classified accordingly. From such soft limits, it is possible to clearly and reliably derive the information that the corresponding area, for example the area between a pair of such road markings or the area around the corresponding vehicle trajectory, is actually drivable. At the same time, such soft limits allow for particularly simple and efficient trajectory planning.

[0021] In a possible improvement of the present invention, structural facilities, in particular guardrails, walls, such as soundproof walls or fences, ditches and vegetation areas, are used as hard boundaries, that is, they are taken into account or graded accordingly. Such hard boundaries can define corresponding boundary conditions, for example for trajectory planning, in order to avoid, for example, a collision of a motor vehicle with such a hard boundary or a loss of control of the motor vehicle when attempting to drive on such a hard boundary. Therefore, specifically identifying such hard boundaries can improve the safety when guiding the vehicle. At the same time, for example, for emergency or evasive actions, the search space used for trajectory planning can also be strictly limited, which can correspondingly achieve faster, more efficient and more reliable trajectory planning. For example, possible drivable space and corresponding safety potential are not abandoned, which may occur if the drivable space is only strictly limited by soft boundaries. On the other hand, if areas outside the hard boundaries that cannot be driven over are strictly excluded in terms of drivability or trajectory planning, additional data processing work can be saved and the corresponding safety risks can be reduced.

[0022] In a possible improvement of the invention, hard limits are further distinguished into physical limits that cannot be driven over on the one hand and rule-based limits that are not allowed to be driven over on the other hand only according to predetermined rules. If and in the case of such hard limits in the corresponding environment, the distances from the driving path to the physical limit and from the driving path to the rule-based limit are separately specified in the data set. In the current sense, a physical limit can be a physical obstacle that makes it impossible for a motor vehicle to cross it. Such a physical limit can be, for example, a protective barrier, a soundproof wall, a vegetation belt with corresponding dense trees, etc. On the contrary, from a purely physical technical (i.e. practical) point of view, a rule-based limit can in principle be driven over, but for example, at least in normal operation, this may not be allowed. Such a rule-based limit can be, for example, a vegetation belt with only corresponding low or flat vegetation, a grass field, a relatively shallow ditch, a curb, etc. In any case, an attempt to drive over a physical limit can be excluded. However, in certain predetermined cases, such as the maximum escalation level of an emergency operation, etc., an attempt to drive over a rule-based limit can be allowed or considered. Accordingly, the hard limits or corresponding distances in the data set can be illustrated, marked or graded with signs. Therefore, it is also possible to specify or define the drivable area in a hierarchical manner for the corresponding different situations. For example, the area between the rule-based limit and the nearest but far-away physical limit can be classified or specified as a non-drivable area for normal operation or for a lower upgrade level of emergency operation, but can be classified or specified as a drivable area or as a conditionally drivable area for emergency operation or the highest upgrade level of emergency operation. In this sense, the conditionally drivable area can be drivable according to the following conditions: for example, no collision occurs thereby, such as a collision with a pedestrian or a cyclist or with an oncoming vehicle, etc., and at the same time, even in normal operation, such a collision can be avoided in the drivable area by avoiding in the conditionally drivable area (especially only thereby). In normal operation or in a lower upgrade level of emergency operation, the search space for trajectory planning can be correspondingly limited, which can help reduce the data processing work and therefore help particularly fast and low-consumption trajectory planning. At the same time, by utilizing all possibilities according to the specific situation, an improvement in overall safety can be achieved.

[0023] In order to identify the respective boundaries, ie their type or kind, for example, a corresponding object identification can be used.

[0024] In another possible embodiment of the invention, in a section along the driving corridor, in which the limit cannot be detected or estimated, the distance is set to zero. In other words, in such a section, the drivable area is limited to the corresponding side or boundary line of the driving corridor. As a result, a particularly high level of safety can be achieved in vehicle guidance, because the vehicle guidance does not have to rely on assumptions about the limit trends or purely speculative extrapolations, etc. Through the embodiment of the invention proposed here, it is possible to guide the vehicle only in the area that is actually identified as drivable in a particularly simple and low-cost manner. This is particularly feasible without a large amount of additional data processing or modeling expenditure.

[0025] Another aspect of the present invention is an assistance system for a motor vehicle. The assistance system according to the present invention has an interface for acquiring environmental data characterizing the corresponding environment located ahead in the direction of travel, a processor device coupled thereto, and a computer-readable data memory coupled to the processor device, the processor device being, for example, a microchip, a microcontroller or a microprocessor, etc. The assistance system according to the present invention is configured to execute the method according to the present invention, in particular automatically. For this purpose, a corresponding operating program or computer program can be stored in the data memory, which is encoded or implemented in conjunction with the method steps, processes or measures described in the method according to the present invention or corresponding control instructions. The operating program or computer program can then be executed by the processor device in order to implement the corresponding method or realize its implementation. The assistance system according to the present invention can be designed, for example, as a control device, in particular an embedded system.

[0026] Another aspect of the invention is a motor vehicle having an environment sensor system for acquiring environment data characterizing the corresponding environment located ahead in the direction of travel and an assistance system according to the invention. The environment sensor system can be part of the assistance system or can be coupled to the assistance system, for example via the onboard electrical system of the motor vehicle. The motor vehicle according to the invention can in particular be a motor vehicle mentioned in conjunction with the method according to the invention and / or in conjunction with the assistance system according to the invention, or correspond thereto.

[0027] Further features of the invention can be seen in the drawings and the description of the drawings. The features and combinations of features mentioned above in the description of the drawings and the features and combinations of features shown below in the description of the drawings and / or alone in the drawings can be used not only in the combination specified in each case but also in other combinations or alone without departing from the scope of the invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] The attached drawings are Figure 1 An example and schematic overview is shown to illustrate possibilities for supporting vehicle guidance. DETAILED DESCRIPTION

[0029] Figure 1 An example and schematic partial overview is shown to illustrate a method for supporting the longitudinal and lateral guidance of a vehicle. Specifically, a section of a road 1 is shown here, which includes a curve. In particular, this can be a graphical representation of a corresponding data set. For the purpose of illustration, a motor vehicle 2 is shown here moving along the road 1. In principle, the motor vehicle 2 can move on the entire road 1, that is, within the lane edge 3. Therefore, the lane edge 3 (shown here on the outside) can be a hard limit of the drivable area or space that cannot be driven over, such as a crash barrier, a soundproof wall, etc. However, in practice, there can also be soft limits that can be driven over, such as lane markings 13 or regulations for maintaining a distance from the lane edge 3, thereby limiting the area or space that is actually to be driven through (at least outside of an emergency). In addition, for technical assistance functions or at least partially automated vehicle guidance, the technical identification or determination of the actual drivable area (possibly graded according to soft and hard limits) is necessary or at least useful. For this purpose, the motor vehicle 2 is equipped with an environmental sensor system 4 and an assistance system 5. Environmental sensor system 4 can be used to record environmental data that reproduce or characterize the environment, in particular the environment ahead in the direction of travel of motor vehicle 2. These environmental data can then be acquired by assistance system 5 via interface 6 and processed using processor 7 and data memory 8.

[0030] The environmental sensor system 4 can include a plurality of different sensors and can also acquire additional data via the interface 6, which can then also be processed accordingly by the auxiliary system 5. The auxiliary system 5 can generate a driving channel estimate 9 and a limit estimate 10 from the acquired data. The driving channel estimate 9 predicts the driving channel of the motor vehicle 2 along the road 1. The limit estimate 10 estimates the trend of one or more limits of the drivable area around the driving channel. In the example shown here, the limit estimate 10 can be, for example, along a lane or lane marking 13. Similarly, further limit estimates can be performed for the trend of the lane edge 3. Here, a distinction can be made between soft limits that can be driven over and hard limits that cannot be driven over. Hard limits can strictly limit the drivable area because it cannot be driven over or over. On the other hand, soft or gentle limits can be, for example, such a mark: it may, that is, under certain predetermined conditions, be driven over or over, or the driving over or over is possible or allowed.

[0031] Along the driving corridor estimate 9, a distance 11 to one or more limit estimates 10 given in the local transverse direction can be determined for each point or section. If and in the case of such different limits or such different limits can be identified or estimated, the distance 11 to the soft limit and the distance to a hard limit that is closest, in particular in the corresponding direction, can be determined respectively. This is explained here by way of example for several locations along the driving corridor estimate 9. For the sake of clarity, only some of the distances 11 are shown here, and only exemplary selections thereof are explicitly marked. The distance 11 can be, in particular, from the driving corridor estimate 9 (for example, its edge or center line) to the hard limit that is closest on the corresponding side or in the corresponding direction and / or to the soft limit that is farthest from each other or farthest from the driving corridor estimate 9, which is closer to the driving corridor estimate 9 than the closest hard limit. An undefined section 12 is also shown here in an exemplary manner, in which a limit estimate 10 cannot be generated at least on one side of the driving corridor estimate 9. Therefore, in the undefined section 12, the distance 11 in the corresponding direction can be set to zero.

[0032] The driving corridor estimate 9 and the determined distance 11 can define an area in which the motor vehicle 2 can be driven. In the undefined section 12, the edge of the drivable area can coincide with the corresponding edge of the driving corridor estimate 9. The driving corridor estimate 9 and the distance 11 or the drivable area defined thereby are summarized in a single data set. This data set can then be used, for example, for further assistance functions of the motor vehicle 2. For this purpose, the data set can be stored, for example, in a data memory 8 and / or provided or transmitted via an interface 6.

[0033] Overall, the described example shows how an estimate or prediction of the driving corridor can be used to determine the drivable area and the associated limits for lateral guidance of the vehicle and how the corresponding data can be used in a particularly simple and efficient manner.

[0034] Reference Numbers List

[0035] 1 Street

[0036] 2 Motor Vehicles

[0037] 3 Lane Edge

[0038] 4 Environmental Sensors

[0039] 5. Auxiliary systems

[0040] 6 Interface

[0041] 7 Processor

[0042] 8 Data storage

[0043] 9 Driving channel estimation

[0044] 10 Boundary Estimation

[0045] 11 Distance

[0046] 12 Undefined section

[0047] 13 Lane markings.

Claims

1. A method for supporting the longitudinal and lateral guidance of a motor vehicle (2) on a road, wherein Acquiring environmental data characterizing a corresponding environment located ahead in the direction of travel of the motor vehicle (2), Based on the environmental data, a course (10) of a boundary (3, 13) is estimated, which limits a drivable area of ​​the road in the corresponding environment, and a driving path (9) that the motor vehicle (2) is expected to travel through, Along the driving path (9), at least one distance (11) of the estimated course (10) from the driving path (9) to the limit (3, 13) is determined, the distance being given transversely to the driving direction, and A data set is generated and provided from the driving corridor (9) and the determined distance (11), wherein the driving corridor (9) is specified in the data set and the drivable area along the driving corridor (9) is additionally specified by the distance (11).

2. The method according to claim 1, It is characterized in that A plurality of different data types are combined with one another for the estimation (10) of the limits (3, 13) and the driving corridor (9).

3. The method according to any one of the preceding claims, It is characterized in that A distinction is made between a hard limit (3) which cannot be driven over and a soft limit (13) which can be driven over.

4. The method according to claim 3, It is characterized in that Only the distance (11) from the driving corridor (9) to the estimated hard limits (3) closest to the sides of the driving corridor (9) and / or only the distance (11) from the driving corridor (9) to the estimated soft limits (10, 13) farthest from the sides of the driving corridor (9) are determined and specified in the data set.

5. The method according to claim 3 or 4, It is characterized in that Use road markings as soft boundaries (13).

6. The method according to any one of claims 3 to 5, It is characterized in that Use structures, particularly guardrails, walls, ditches and vegetated areas, as hard boundaries (3).

7. The method according to any one of claims 3 to 6, It is characterized in that The hard limit (3) is further differentiated into a physical limit that cannot be crossed by driving and a rule-based limit that is not allowed to be crossed by driving only according to a predetermined rule, and the distance (11) from the driving passage (9) to the physical limit and the distance from the driving passage to the rule-based limit are separately specified in the data set.

8. The method according to any one of the preceding claims, It is characterized in that In a section (12) along the driving path (9) in which the limit (10) cannot be estimated, the distance (11) is set to zero.

9. An assistance system (5) for a motor vehicle (2), comprising an interface (6) for acquiring environmental data characterizing a corresponding environment located ahead in the direction of travel, a processor device (7) and a computer-readable data memory (8) coupled to the processor device, wherein the assistance system (5) is configured to execute a method according to any one of the preceding claims.

10. A motor vehicle (2) having an environment sensor system (4) for acquiring environment data characterizing the corresponding environment located ahead in the direction of travel, and an assistance system (5) according to claim 9.

Images