Method for determining a navigation route for automated driving operation of a vehicle - Patents.com

Abstract

The present invention relates to a method for determining a navigation route (NR) for an automated driving operation of a vehicle, which comprises the steps of: - based on a digital attribute-based road map (AK), a node list is created which represents a number of geographic coordinates (KP1-KP7) of a route (R) determined by a navigation system; - the node list is transferred to a digital sensor-based road map (SK) which includes a representation of the environment detected by sensors; - based on the node list, precise lane-by-lane candidates are determined in the digital sensor-based road map (SK); - irrelevant candidate lane-by-lane candidates are identified and eliminated; - the route (R) is divided into a subgraph (T1) which includes only one possible lane and a subgraph (T2) which includes a number of possible lanes; - optimal lanes (oF) are identified; - combining the optimal lanes (oF) into a desired lane-by-lane precise navigation route (NR), thereby determining a lane-by-lane precise navigation route (NR) for an automated driving operation, in particular a highly automated or autonomous driving operation.

Description

[Technical field]

[0001] The present invention relates to a method for determining a navigation route for automated driving operation of a vehicle. [Background technology]

[0002] It is generally known from the prior art to assign a destination to a navigation system of a vehicle capable of autonomous operation, the navigation system determining a navigation route to the assigned destination, and, if possible, the vehicle traveling at least partially autonomously along the determined navigation route.

[0003] US Patent No. 10969232 describes the arrangement of standard definition (SD) and high definition (HD) road maps from various sources. In response to input of a destination, a route to the destination is defined and waypoints in the SD road map are generated from the defined route. A graph is created from the HD road map. The waypoints correspond to nodes and edges in the graph. One or more edges form a segment in the HD road map. A number of segments are identified to correspond to the route.

[0004] German patent application No. 102021006166.7 describes a method for data transfer between two digital road maps: the data transfer is performed between a digital sensor-based road map and a digital attribute-based road map, the sensor-based road map is divided into intersections and route segments, attributes for automated driving are assigned to the route segments, attributes for automated driving are assigned to the intersection segments, and it is checked whether the attribute values ​​for automatic crossing in the intersection areas can be made to comply with the turn restrictions in the attribute-based road map.

[0005] From US 2022 / 0032907 A1, an apparatus and a method for managing a vehicle traveling on a road along a route are known, in which the surrounding environment of the road along which the vehicle travels is detected by sensors arranged in the surrounding environment, map information relating to the road being traveled is detected, and based on the detected surrounding environment and the detected map information, information regarding the travel speed of the vehicle on the route or the passability of the route is determined and transmitted to the vehicle.

[0006] From US 2020 / 0200547 A1 a method is known for generating a driveable boundary of a road which is stored as a three-dimensional representation in a high-resolution three-dimensional map. In this method a two-dimensional projection of the road is created, which projection includes a number of points along the road. Each of these points is evaluated in terms of its driveability and based on this a driveable area and its boundaries are determined in the projection of the road. The determined driveable area and its boundaries are then converted into a three-dimensional representation and used to update the three-dimensional map. Summary of the Invention [Problem to be solved by the invention]

[0007] It is an object of the present invention to provide a method for determining a navigation route for automated driving operation of a vehicle which is improved over the prior art. [Means for solving the problem]

[0008] This object is achieved according to the invention by a method for determining a navigation route for automated driving operation of a vehicle having the features of claim 1.

[0009] Advantageous embodiments of the invention are the subject matter of the dependent claims.

[0010] The method for determining a navigation route for an automated driving operation of a vehicle according to the present invention aims to determine a precise lane-by-lane (lane-level precision) navigation route for an automated driving operation, in particular a highly automated or autonomous driving operation, as follows:

[0011] Based on the digital attribute-based road map, a node list is created that represents the geographic coordinates of the route determined by the navigation system. The node list is transferred to the digital sensor-based road map that includes the environment representation (description of the environment) detected by the sensors. Based on the node list, lane-by-lane precise driving section candidates are determined in the digital sensor-based road map. Irrelevant driving section candidates are identified and eliminated. The route is divided into a subgraph that includes only one possible lane and a subgraph that includes multiple possible lanes. In the subgraph that includes only one possible lane, the corresponding lane-by-lane precise driving section candidate exists in only one lane. In the subgraph that includes multiple possible lanes, the corresponding lane-by-lane precise driving section candidate exists in multiple lanes. Optimal lanes are identified. In the subgraph that includes only one possible lane, this lane is identified as the optimal lane. In the subgraph that includes multiple possible lanes, one of the multiple lanes is selected as the optimal lane. The optimal lanes are combined to form the desired lane-by-lane precise navigation route.

[0012] With the method according to the invention, using a route determined by a navigation system, which is in particular edge-based and / or node-based, an optimal, lane-precise, i.e. lane-defined and as best possible automatically drivable driving path is determined from the point of view of a digital sensor-based road map in the form of a lane-precise navigation route which the vehicle can optimally follow in an automated driving operation. This driving path, i.e. lane-precise navigation route in the digital sensor-based road map, corresponds as best possible to a driving path, i.e. route, from a digital attribute-based road map. Using this driving path, i.e. lane-precise navigation route, the vehicle can be driven automatically at least along the automatically drivable driving sections of the navigation route, i.e. along the sections of the navigation route for which lane-precise driving section candidates have been determined, can initiate a timely handover of vehicle steering to or from the driver, and can enable manual driving, in particular based on driving recommendations by the navigation system, for non-automatically drivable driving sections of the navigation route, i.e. sections of the navigation route for which lane-precise driving section candidates have not been determined.

[0013] In one possible embodiment, the automated driving operation is blocked (prevented, prohibited) in sections of the navigation route where precise lane-by-lane candidate driving section has not been determined. In one possible embodiment, the automated driving operation is permitted (released) in sections of the navigation route where precise lane-by-lane candidate driving section has been determined. In one possible embodiment, highly automated or autonomous driving operation is permitted only in permitted sections. In this way, it is ensured that the automated driving operation, in particular the highly automated or autonomous driving operation, is performed only in sections of the navigation route suitable for it.

[0014] In one possible embodiment, in highly automated or autonomous driving operations, a takeover request is issued to the driver of the vehicle before reaching the blocked section, allowing the driver to take over vehicle control in a timely manner, thereby enabling uninterrupted driving operations and avoiding unsafe situations.

[0015] In one possible embodiment, if the determined (found) lane-by-lane precise navigation route includes one or more blocked sections, an alternative lane-by-lane precise navigation route is searched for that has no sections or has fewer sections where the automated driving operation is blocked, thereby finding a navigation route that has no sections or at least the fewest sections where the automated driving operation is blocked, thereby enabling the automated driving operation to be traveled from the starting point to the destination, or at least the longest possible section.

[0016] In one possible embodiment, for each subgraph containing a number of possible lanes, the optimal lane is identified using a cost function. In particular, in this case, each lane of each subgraph containing a number of possible lanes is assigned a cost, the magnitude of which depends in particular on the suitability of the respective lane for the automated driving operation, and in particular on the degree of the automated driving operation, i.e. in particular on the SAE-Level, i.e. on the automation level of the driving operation. The lower the SAE-Level, the higher the assigned cost. Alternatively or additionally, the necessary lane changes can also be included in this evaluation. The optimal lane is therefore identified by minimizing these costs. Instead of this cost function, it is also possible to identify the optimal lane using the inverse approach, i.e. a bonus function. In this case, the assignment of the bonus is performed inversely to the assignment of the costs. In the example of the SAE-Level, the higher the SAE-Level, the higher the assigned bonus. Alternatively or additionally, the necessary lane changes can also be included in this evaluation. The optimal lane is then identified by maximizing the bonus.

[0017] For example, the evaluation can be weighted using a cost or a bonus. Alternatively, for example, machine learning can be performed using training data and then the evaluation can be performed using, for example, a neural network.

[0018] In one possible embodiment, based on the lane-by-lane precise navigation route, lane change recommendations are generated for performing automatic or manual lane changes to follow the lane-by-lane precise navigation route.

[0019] Preferably, lane change recommendations for performing automatic lane changes are generated during highly automated or autonomous driving operations and output to a device for performing highly automated or autonomous driving operations.

[0020] Lane change recommendations for performing manual lane changes are generated and issued to a vehicle driver, particularly for vehicles not designed to perform automatic lane changes. The recommendations may therefore be issued to a driver during highly automated or autonomous driving operations or manual driving operations. Issuing recommendations during manual driving operations is effective because it guides the driver to an optimal lane of a lane-by-lane accurate navigation route in preparation for a later highly automated or autonomous driving operation.

[0021] In this way, the driving experience of highly automated or autonomous driving can be improved by pre-planning a lane change to an optimal lane and automatically executing it or recommending it to the driver, thereby minimizing unexpected interruptions of the highly automated or autonomous driving operation because the lane section ahead is not suitable for this driving operation.

[0022] The features described for the method for determining a navigation route for automated driving operation of a vehicle may, individually or in combination with each other, be part of a method for performing driving operation of a vehicle, in particular for performing highly automated or autonomous driving operation of a vehicle.

[0023] In the following, an embodiment of the invention will be explained in detail with reference to the drawings. [Brief description of the drawings]

[0024] [Figure 1] 1 illustrates a schematic diagram of a digital sensor-based road map. [Diagram 2] 1 illustrates a schematic diagram of a digital attribute-based road map. [Diagram 3] 3 shows a schematic representation of the digital sensor-based road map of FIG. 1 including geographic coordinates transferred from the digital attribute-based road map according to FIG. 2; [Figure 4] The digital sensor-based road map of FIG. 1 is shown diagrammatically together with possible driving trajectories resulting from the determined relevant candidate driving segments. [Diagram 5] Schematic showing partitioning into subgraphs and determining precise lane-by-lane navigation routes. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0025] In all the drawings, the same reference numerals are used to designate corresponding parts.

[0026] With reference to FIGS. 1 to 5, a method for accurately determining a navigation route for automatic driving operation (automated driving operation) of a vehicle, particularly for highly automated or autonomous driving operation, on a lane-by-lane basis will now be described.

[0027] In the prior art, autonomous driving is further developed, in particular by using a data-driven approach. Vehicles are equipped with a wide range of sensors that inspect the vehicle's surroundings from various aspects to provide surrounding environment data. Here, the sensor technologies used include, for example, radar, lidar, camera and / or ultrasonic and / or other sensor technologies suitable for detecting the surrounding environment. Using these surrounding environment data, a model of the environment is created, in particular by using machine learning techniques. The autonomous vehicle is introduced into this model, hypotheses are made by predicting future movements, from which possible courses of action for autonomous vehicle control are determined and executed.

[0028] To plausibly check the position estimates of an autonomous vehicle, to plausibly check the modelling and ultimately to plausibly check possible courses of action, the prior art not only relies on current sensor data of the vehicle itself, but also uses historical sensor data for safety reasons.

[0029] As illustrated in Fig. 1, these sensor data from past journeys, ideally of several different vehicles, are abstracted, in particular using machine learning techniques, and stored in a digital sensor-based road map SK, which shows the exact lane-by-lane traveled path segments recorded in the past and an image of the surrounding environment as detected (discovered) in terms of the stored sensor data. This is therefore a sensor-based road map SK that has been learned by the vehicle using sensor technology.

[0030] On the other hand, a navigation system requires a digital attribute-based road map AK, as exemplified in Fig. 2. This attribute-based road map AK is used for a purpose entirely different from that of the sensor-based road map SK. The attribute-based road map AK is primarily used to calculate a route from the current position of the vehicle to a specified destination. Therefore, the attribute-based road map AK has fundamentally different characteristics from the sensor-based road map SK.

[0031] The attribute-based road map AK is quite abstract. Otherwise, the route cannot be calculated efficiently by the navigation system built into the vehicle. In addition, the attribute-based road map AK consists of at least edges, also called links or road segments, nodes, also called junctions or intersections between links, and relations. Relations are relationships and conditions between edges, such as right and left turn restrictions AV.

[0032] Attribute-based road maps AK have many attributes that cannot or are difficult to automatically detect by sensors, such as road names, road types, certain points, street addresses, or specific restrictions at complex intersections (e.g., vehicle-based and / or time-based turn restrictions AV).

[0033] In particular, the attribute-based road map AK and the sensor-based road map SK differ in three main respects, which are explained below.

[0034] Unlike attribute-based road maps AK, which depict the road network according to a predefined set of rules optimized to enable efficient navigation functions, sensor-based road maps SK depict the surrounding environment in the way it is observed and measured by sensors.

[0035] As shown in Fig. 2, the attribute-based road map AK shows road-level geometry, i.e. edge-based geometry. In contrast, the sensor-based road map SK shows lane-level geometry, i.e. lane-based geometry, as shown in Fig. 1, where the sensor-based road map SK is shown with lanes. As can be seen by comparing Fig. 1 and Fig. 2, these differences are especially evident at complex intersections.

[0036] Unlike the attribute-based road map AK, which shows turn-restricted AVs, the sensor-based road map SK models turn options through the presence of dedicated turn lanes.

[0037] In the sensor-based road map SK in Figure 1, all turn and driving options AF at intersections are shown. In the attribute-based road map AK in Figure 2, turn restrictions AV at intersections are shown with dashed arrows. The difference in abstraction level is clear.

[0038] The sensor-based road map SK contains information for automated driving operation, while the attribute-based road map AK enables route calculation and route guidance for example in a navigation system.

[0039] Accordingly, German patent application no. 102021006166.7 describes a matching technique that allows information from a sensor-based road map to be mapped onto a much more abstract attribute-based road map, in particular on the basis of figures 1 to 5 and the associated description of the drawings. German patent application no. 102021006166.7, in particular its figures 1 to 5 and the description of the drawings, are incorporated herein by reference.

[0040] The method described below enables in a particularly optimal manner the mapping of a route R from a start point SP to a destination point ZP determined by a navigation system on the basis of a digital attribute-based road map AK, which is particularly suitable for performing automated driving operations, in particular highly automated or autonomous driving operations, onto a digital sensor-based road map SK, thereby enabling the vehicle to travel the route R in a particularly automated driving operation, in particular in a highly automated or autonomous manner.

[0041] The overall system in which this mapping is implemented consists in particular of a vehicle equipped with an apparatus for performing automated driving operations, in particular highly automated or autonomous driving operations, based on a digital sensor-based road map SK, and a navigation system with a digital attribute-based road map AK. The navigation system can be implemented in the vehicle using a local database or can be provided as a back-end based online solution, i.e. with a data connection between the vehicle and a server external to the vehicle, or both embodiments can be combined.

[0042] In the method described here, particularly in step I, a node list is created based on the digital attribute-based road map AK, which represents the geographic coordinates KP1-KP7 of the route R determined by the navigation system. These geographic coordinates KP1-KP7 are also called nodes.

[0043] First, a route R is calculated in the navigation system. This route R or a desired, in particular a predicted driving path is advantageously transferred to a device for performing automated driving operations, in particular highly automated or autonomous driving operations, in order to enable the vehicle to follow the route R or the driving path with the highest possible degree of automation (SAE level). In the following, the method is described using a route R determined by the navigation system, although a driving path can also be used instead of the route R.

[0044] In particular, the route R determined by the navigation system is initially available within the navigation system as a directed list (sequence) of edges and / or nodes in the map graph of the attribute-based road map AK. This list typically consists of alphanumeric or numeric IDs, i.e. identification codes to the edges / nodes in the map graph. All edges and / or nodes are completely included.

[0045] In particular, in sub-step I.1, a mapping is made to a list containing only nodes, since the geometric course of the route R is completely and unambiguously contained in the complete node list, which is the information available to make a mapping to the sensor-based road map SK.

[0046] If the route list of a route R determined by the navigation system consists of edges and nodes, in particular, all edges shall be removed from the route list, so that the route list contains a complete list of only the nodes along the route R.

[0047] If the route list of a route R determined by the navigation system consists of edges only, then in particular, the respective start and end nodes of each edge shall be explored in the map graph. The complete list of edges is replaced by the complete list of nodes along the route R.

[0048] In particular, in sub-step I.2, the node list along the route R is looked up using the IDs mentioned above, but these IDs are unknown to the sensor-based road map SK, so the node list is mapped to a common exchange format known to both.

[0049] The IDs of the complete list of nodes along the route R are transposed into a dataset (data record) in a known exchange format for both road maps AK, SK alike. For example, the geographic coordinates may be in WGS84 format and may optionally be enriched with additional data, for example altitude information and road names or numbers.

[0050] Furthermore, in the method described herein, particularly in step II, the node list is transferred to a digital sensor-based road map SK which includes a representation (description) of the environment (surroundings) detected by the sensors, particularly in a common exchange format.

[0051] For example, the node list of the route R is transferred in a common exchange format to the computation unit, which accesses the sensor-based road map SK and performs a mapping, ie, matching, to the sensor-based road map SK.

[0052] FIG. 2 shows a complete node list of a route R including nodes, ie, multiple geographic coordinates KP1-KP7, in an attribute-based road map AK.

[0053] Figure 3 shows the complete node list of route R mapped to the sensor-based road map SK in a common exchange format. Due to the generalized characteristics of the attribute-based road map AK, the geographic coordinates KP1-KP7 may deviate significantly from the actual lane-by-lane precise driving path of the sensor-based road map SK and therefore from the possible lane-by-lane precise navigation route NR.

[0054] Furthermore, in the method described herein, particularly in step III, accurate driving section candidates are determined on a lane-by-lane basis in the digital sensor-based road map SK based on the node list.

[0055] Due to the large deviations in the geometrical representation between the two road maps AK, SK, which are evident from a comparison of Figures 2 and 3, tolerance ranges TB are provided, particularly for the nodes, i.e. the geographic coordinates KP1-KP7, imported from the attribute-based road map AK into the sensor-based road map SK. Each tolerance range TB can be, for example, larger in the lateral direction, i.e. in the direction perpendicular to the vehicle's driving direction (transverse direction), than in the driving direction.

[0056] All lane-level accurate driving sections within the respective tolerance range TB are the driving section candidates to be used, i.e. lane-level accurate driving section candidates. If there are no lane-level accurate driving sections in the sensor-based road map SK for some consecutive nodes, i.e. geographic coordinates KP1-KP7, within the tolerance range TB, it can be assumed that there is a small deviation between the two road maps AK, SK. In this case, some of these individual nodes, i.e. geographic coordinates KP1-KP7, can be skipped. If there are no lane-level accurate driving sections in the sensor-based map SK for a sufficient number of consecutive nodes, i.e. geographic coordinates KP1-KP7, within the tolerance range TB, it can be assumed that no automated driving operation is possible within this section along the nodes, i.e. geographic coordinates KP1-KP7. Further steps in this case are described below in step V.

[0057] Furthermore, in the method described herein, particularly in step IV, irrelevant or invalid travel segment candidates are identified and eliminated or discarded (deleted), thereby leaving relevant travel segment candidates RF.

[0058] In particular, in sub-step IV.1, candidate segments that have a direction of travel opposite to the path of the route R are discarded, i.e. excluded. Candidate segments that move further away from the route R, in particular those that eventually leave the route R, are discarded, i.e. excluded. Candidate segments that join the route R from outside, i.e. from the other side of the route R, are discarded, i.e. excluded. Candidate segments that are not part of the route R but for example just pass nearby are also discarded, i.e. excluded.

[0059] In particular, in sub-step IV.1, dead paths are eliminated. Candidate driving sections that correspond to a driving course, i.e. a route R, but whose further course makes it impossible to follow the route R, are eliminated retroactively until a point is reached that allows a lane change to an alternative driving section that follows the route R further. This is, for example, the next point reached retroactively, which has a dashed road marking at which a lane change is possible. Similarly, in the reverse, i.e. in the opposite procedure (mirroring), this also applies to candidate driving sections that run in parallel based on the history, but can only be reached by changing lanes along the route R from a certain point in the direction of travel. In the illustrated example, a lane change from the left lane in the driving direction along the route R to the right lane in the driving direction along the route R is only possible up to the third geographic coordinate KP3. Thus, from the third geographic coordinate KP3 onwards, the driving section in the left lane in the driving direction along the route R is not a relevant candidate driving section RF.

[0060] As shown in FIG. 4, the remaining relevant driving section candidates RF allow one driving trajectory FT or multiple driving trajectories FT (illustrated by dashed lines) of the vehicle in either the case of directly driving in the right lane, as illustrated by the dashed lines, or the case of first driving in the left lane and then changing lanes to the right lane, if permitted.

[0061] Furthermore, in the method described herein, and particularly in step V, the route R is partitioned (segmented) into a subgraph T1 containing only one possible lane and a subgraph T2 containing multiple possible lanes. This is illustrated in Figure 5 for the entire route R from a start point SP to a destination point ZP, where the different segments S are indicated by boxes.

[0062] As shown, there may be a segment So without a possible subgraph. A desired lane-by-lane precise navigation route NR for automated driving operation, particularly highly automated or autonomous driving operation, has gaps in such a segment So. Automated driving is not possible in this segment So. In a section of the navigation route in such a segment So where there is no possible subgraph and a lane-by-lane precise driving section candidate could not be determined, automated driving operation, for example with SAE levels 3, 4 and 5, is blocked, i.e. at least highly automated and autonomous driving operation is not allowed (released). Therefore, for highly automated or autonomous driving operation, it is necessary to hand over the guidance of the vehicle to the driver of the vehicle before reaching the blocked section. Therefore, a handover request is issued to the driver of the vehicle before reaching the blocked section.

[0063] In the illustrated example, a fully automated or autonomous driving operation is not possible along the entire route R, because when driving in a fully automated or autonomous manner, the vehicle needs to reach the destination, i.e. the destination point ZP of the route R, without the driver, which would not be possible if the driving task needs to be handed over to the driver during that time. When the vehicle drives in a fully automated or autonomous manner, the route R is discarded and an alternative route R is searched for that does not include a segment So without a possible subgraph.

[0064] In the partitioning (segmentation) step, as described above, in particular in step V, a graph containing all possible trajectories along the route R is partitioned into a subgraph T1 with only one possible lane, i.e. in which the relevant lane-by-lane precise driving section candidate RF exists in only one lane, and a subgraph T2 with multiple possible lanes, i.e. in which the relevant lane-by-lane precise driving section candidate RF exists in multiple lanes.

[0065] Sections of the route R that cannot be mapped onto the sensor-based road map SK, i.e. where automated driving operation is not possible, are treated separately, as described above in step III. As already mentioned, automated driving operation is not possible here, so driving is handed over to the driver before reaching the section in question. After the end of such a section, automated driving can be restored, i.e. automated driving operation can be resumed. Alternatively, a method for performing automated driving operation without a sensor-based road map SK can be used, in which automated driving operation is possible without handing over to the driver in those segments So without a possible subgraph.

[0066] Furthermore, in the method described herein, particularly in step VI, optimal lanes oF are identified. In a subgraph T1 that includes only one possible lane, the only possible lane is the optimal lane oF, since there is no alternative for performing an automated driving operation. The vehicle must stay in this lane. In a subgraph T2 that includes multiple possible lanes, one of the lanes must be selected as the optimal lane oF for creating a lane-by-lane accurate navigation route NR. The selection takes into account (determines) which of the possible lanes allows driving at the highest possible automation level, i.e. the highest possible SAE level, with the fewest possible lane changes, the fewest possible and / or fewest possible hindrances requiring the fewest possible and / or fewest possible speed adjustments.

[0067] Thus, the selection of the optimal lane oF in the subgraph T2 comprising multiple possible lanes is made in particular by considering (determining) the results of the evaluation, for example with a cost function. The optimal, i.e. lane with the smallest cost, is selected as the optimal lane oF and is thus used to enable the vehicle to drive autonomously along the subgraph T2. For example, lane changes along the subgraph T2 are also taken into account if the corresponding data from the sensor-based road map SK allow lane changes. The cost function may include one or more or all of the following parameters: These parameters are included in the cost function, for example, proportionally to the length along the route R (depending on the proportion of the length):

[0068] Costs decrease as the level of automation (SAE level) of autonomous driving increases. The level of automation is an attribute of the sensor-based road map SK, and individual road sections can have different automation levels. -Costs go down as the proportion of autonomous driving operations increases. Costs increase with the number of handovers required between the driver and the autonomous vehicle. Conversely, costs decrease as the autonomous vehicle segments become longer. -Costs increase with the number of lane changes performed. Costs increase depending on the presence of current traffic delays, which can be determined from precise lane-by-lane traffic information, for example. The costs are reduced depending on the actual lane-by-lane driving speed, which can be determined in particular by the sensor system.

[0069] The individual cost parameters can, for example, in particular be weighted relative to one another depending on the system design and translated (transformed) into a common overall cost function which takes all partial parameters (sub-parameters) into account accordingly.

[0070] Instead of this cost model of evaluation, in which the adverse costs are summed up and minimized as a whole, a bonus value can also be calculated and maximized on the contrary. At points where the above-mentioned costs fall, the bonus value increases, and at points where the above-mentioned costs rise, the bonus value decreases. Here too, for example, the individual bonus parameters can be weighted relative to one another, in particular depending on the system design, and transferred (transformed) accordingly into a common overall bonus function that takes all partial parameters into account. The difference is that the cost function minimizes the costs, whereas the bonus function maximizes the bonus value.

[0071] Instead of a weighted evaluation of the individual arguments, in particular the abovementioned parameters, machine learning can be carried out, for example on the basis of training data, in which case the evaluation is carried out, for example, by means of a neural network.

[0072] Furthermore, in the method described herein, in particular in step VII, the determined lane-by-lane precise navigation route NR is formed by combining the identified optimal lanes oF. That is, a reconstructed, autonomously drivable, navigable optimal driving path oF is formed in the sensor-based road map SK. The lane-by-lane precise navigation route NR is shown in FIG. 5 by a dashed line. The lanes of the subgraph T1 containing only one possible lane, i.e. the optimal lane oF, are combined with the subgraph containing the respective optimal lanes oF identified as described above in the subgraph T2 containing multiple possible lanes to form a global graph. If the global graph has one or more gaps, in the illustrated example the sixth and ninth segments S in the driving direction, then autonomous driving operation based on the data of the sensor-based road map SK is not possible. In that case, a handover to the driver must be made or an alternative route R must be requested in the navigation system. That is, an alternative lane-by-lane precise navigation route NR can be searched for which there are no sections where autonomous driving operation is blocked or which have fewer sections where autonomous driving operation is blocked.

[0073] Since the lane-by-lane precise navigation route NR is composed of a set of optimal lanes oF, it is possible to generate a lane change recommendation that will lead the vehicle to the respective optimal lane oF when following the lane-by-lane precise navigation route NR. The lane recommendation can be a control instruction that is output to a device for performing an automated or highly automated driving operation in an automated or highly automated driving operation to initiate the execution of an automated lane change. The lane recommendation can also be an execution instruction issued to the driver to instruct the driver to manually execute a lane change. This is particularly useful for vehicles that are not designed to execute an automated lane change. In that case, the issuance of a lane change recommendation to the driver can be performed in a highly automated or autonomous driving operation, but also in a manual driving operation. In that case, the issuance of a lane change recommendation in a manual driving operation is effective since the driver is initially directed to the optimal lane oF in preparation for a subsequent highly automated or autonomous driving operation that may be performed. [Prior art documents] [Patent documents]

[0074] [Patent Document 1] U.S. Pat. No. 1,096,9232 [Patent Document 2] German Patent Application No. 102021006166.7 [Patent Document 3] US Patent Application Publication No. 2022 / 0032907 [Patent Document 4] US Patent Application Publication No. 2020 / 0200547

Claims

1. 1. A method for determining a navigation route (NR) for automated driving operation of a vehicle, comprising: - a node list is created, which represents a number of geographic coordinates (KP1-KP7) of a route (R) determined by the navigation system, based on a digital attribute-based road map (AK); - the node list is transferred to a digital sensor-based road map (SK) containing a representation of the environment detected by the sensors; - based on the node list, accurate lane-by-lane candidate driving sections are determined in the digital sensor-based road map (SK); - irrelevant candidate driving segments are identified and eliminated; - said route (R) is partitioned into a subgraph (T1) containing only one possible lane and a subgraph (T2) containing several possible lanes; - Optimal lanes (oF) are identified, - The optimal lanes (oF) are combined to form a desired lane-by-lane accurate navigation route (NR); thereby determining an accurate lane-by-lane navigation route (NR) for the automated driving operation, in particular for a highly automated or autonomous driving operation.

2. In a section of the navigation route (NR) where the accurate driving section candidate in lane unit is not determined, an automatic driving operation is blocked. The method of claim 1.

3. In the section of the navigation route (NR) where the accurate driving section candidate in lane unit is determined, an automatic driving operation is permitted. The method according to claim 1 or 2.

4. The highly automated or autonomous driving operation is permitted only in permitted sections. The method according to claim 3.

5. In the highly automated or autonomous driving operation, a handover request is issued to the driver of the vehicle before reaching a blocked section. The method according to any one of claims 2 to 4.

6. When the determined lane-by-lane precise navigation route (NR) includes one or more blocked sections, an alternative lane-by-lane precise navigation route (NR) without blocked sections or with fewer blocked sections is searched for. The method according to any one of claims 2 to 5.

7. In each of the subgraphs (T2) including a plurality of possible lanes, the optimal lane (oF) is identified using a cost function. The method according to any one of claims 1 to 6.

8. Based on the lane-by-lane precise navigation route (NR), a lane change recommendation is generated for performing a manual or automatic lane change to follow the lane-by-lane precise navigation route (NR). The method according to any one of claims 1 to 7.

Images