Method and control unit for operating an automatic longitudinal and / or lateral guidance function of a vehicle

Abstract

This document describes a control unit for an autonomous and / or semi-autonomous ego vehicle (202). The control unit is configured to determine a course (811) of an ego lane (502) in which the ego vehicle (202) is driving, referred to as a behavior ego lane, based on driving trajectories (214, 216, 218, 220) of a plurality of vehicles (204, 206, 208, 210) within an environment of the ego vehicle (202). Furthermore, the control unit is configured to operate an automatic longitudinal and / or lateral guidance function of the ego vehicle (202) in accordance with the course (811) of the behavior ego lane (502).

Description

Technical Field

[0001] This document relates to vehicles, particularly autonomous or semi-autonomous vehicles. In particular, this document aims to adapt driving strategies for autonomous or semi-autonomous vehicles based on the collective behavior of other vehicles in the same environment. Background Technology

[0002] Autonomous and / or semi-autonomous vehicles are guided by a system that detects the geometry and / or route of road lanes. The geometry and / or route of road lanes can be determined using sensors capable of identifying lane boundaries, for example, by detecting lane markings that delineate lane boundaries. Alternatively, map data can be used to determine the geometry and / or route of road lanes. Lanes detected based on lane markings and / or map data may be referred to herein as sensed lanes.

[0003] Map data may be unavailable in unmapped areas and is generally unreliable in built-up areas. Additionally, sensors (e.g., camera sensors) may fail to detect lane geometry in adverse weather conditions or may fail to detect lane geometry on roads without lane markings or with faded or incorrect lane markings.

[0004] Driving strategies and / or driving functions of autonomous and / or semi-autonomous vehicles may be affected if the geometry and / or route of road lanes (i.e., sensing lanes) cannot be reliably detected (e.g., due to adverse weather conditions and / or due to the lack of lane markings or unreliable lane markings). The technical problem addressed in this document is to demonstrate reliable and safe driving functions for autonomous and / or semi-autonomous vehicles without using map data and / or sensor data regarding lane markings. This technical problem is solved by the independent claims. Preferred examples are defined in the dependent claims. Summary of the Invention

[0005] According to one aspect, a control unit for autonomous and / or semi-autonomous self-driving vehicles (particularly two-rail vehicles, such as cars, trucks, buses, etc.) is described. The control unit is configured to determine the route of its own lane, within which the self-driving vehicle is traveling, based on the travel trajectories of multiple vehicles within its environment; this self-driving lane is referred to as a behavioral ego lane. The behavioral ego lane can be determined without considering map data and / or sensor data regarding lane markings on the road over which the vehicle is traveling. In particular, the behavioral ego lane can be determined (only) based on the collective behavior of the vehicles within the self-driving vehicle's environment. The vehicle's behavior can be described using its travel trajectories.

[0006] The control unit can be configured to determine trajectory data regarding the travel trajectories of multiple vehicles within its (direct) environment. The vehicle itself may include one or more environmental sensors (such as cameras, radar sensors, lidar sensors, etc.). Sensor data from the one or more environmental sensors can be used to determine an environmental model of the vehicle's environment, where the environmental model can indicate objects (particularly other vehicles) within the vehicle's environment. Furthermore, the environmental model can indicate the travel trajectories of other vehicles (where the travel trajectories of the vehicles can indicate the position of the vehicles over time).

[0007] Furthermore, the control unit can be configured to cluster driving trajectories within each segment of a road segment sequence to determine a road shape sequence for the corresponding segment sequence on which its own vehicles are traveling. In other words, the road can be segmented into a sequence of segments. For each segment, different driving trajectories can be compared to identify the “dominant direction” of multiple driving trajectories within that segment. Specifically, it can be determined in which direction most vehicles are traveling within each segment of different segments. It can be assumed that the direction of most driving trajectories within a segment corresponds to the direction, route, or shape of the road within that segment. Therefore, a shape sequence describing the overall shape or route of the road can be determined. In particular, the road route along the segment sequence can correspond to the shape sequence, or can be determined based on the shape sequence (e.g., by filtering the shape sequence and / or by fitting an analytical function to the shape sequence).

[0008] Furthermore, the control unit can be configured to determine the number of lanes on the road. To this end, the control unit can be configured to assign multiple vehicles to one or more distinct adjacent lanes on the road. The number of lanes and / or the location of the lanes can be determined so that different vehicles can be assigned to different lanes in a definite manner. One or more distinct adjacent (traffic) lanes also include lanes for the traffic itself.

[0009] The route of its own lane can be determined in a precise manner based on the shape sequence of the corresponding segment sequence for the road, especially based on the route of the road.

[0010] Furthermore, a confidence metric can be determined for the route identified for the vehicle's own lane. The confidence metric indicates the degree of confidence that the route of the vehicle's own lane corresponds to the actual route of the lane on the road on which the vehicle is traveling. Alternatively or additionally, the confidence metric can indicate the confidence and / or definiteness with which the vehicle's own lane has been identified. Typically, the confidence and / or confidence metric increases with the number of vehicles in the vehicle's environment.

[0011] The control unit is also configured to operate automatic longitudinal and / or lateral guidance functions of the vehicle, particularly vehicle following functions (such as ACC) and / or active lane keeping functions, based on the route taken within its own lane. As a result, reliable and robust operation of the autonomous and / or semi-autonomous vehicle can be achieved, particularly without considering map data and / or lane markings on the road on which the vehicle is traveling. Alternatively or additionally, for example, by comparing the determined route taken within its own lane with sources of lane sensing (such as data derived from map data and / or sensor data indicating lane markings), the route taken within its own lane can be used to verify whether the map data and / or (detected) lane markings can be trusted.

[0012] Longitudinal and / or lateral guidance functions may include a vehicle following function, configured to perform longitudinal and / or lateral guidance of the vehicle based on a target vehicle traveling in front of it within its own lane. The vehicle following function may be configured to maintain the vehicle at a defined distance from the target vehicle, and / or adjust the vehicle's speed based on the target vehicle's speed, and / or perform lateral movement of the vehicle based on the target vehicle's trajectory.

[0013] The control unit can be configured to determine the trajectory of a target vehicle, which is referred to herein as the target trajectory. The target trajectory can be determined based on location and / or trajectory data (as described above). The target trajectory can be compared with the route of the vehicle's own lane, and the vehicle following function can be operated based on the comparison between the target trajectory and the route of the vehicle's own lane, thereby increasing the reliability and safety of the vehicle's target following function.

[0014] Specifically, the control unit can be configured to determine the values ​​of one or more deviation parameters, which indicate the degree of deviation of the target trajectory from the route of its own lane. Example deviation parameters include: the lateral offset between the target trajectory and the route of its own lane, and / or the angle between the target trajectory and the route of its own lane (at any point along the relevant longitudinal region). Based on the values ​​of one or more deviation parameters, the vehicle following function can be operated in a particularly robust and safe manner.

[0015] The control unit can be configured to determine whether the target trajectory deviates from the route of the behavior's own lane by more than a deviation threshold. The deviation threshold can depend on a confidence metric regarding the behavior's own lane, specifically such that if the confidence metric decreases, the deviation threshold increases (thus allowing for greater deviation), and / or such that if the confidence metric increases, the deviation threshold decreases.

[0016] If the deviation parameter between the target trajectory and the collective behavior's own lane (i.e., the route of the behavior's own lane) exceeds a given deviation threshold, the vehicle following function can be automatically interrupted (at least for the target vehicle). As a result, the robustness and safety of the vehicle following function can be increased.

[0017] The control unit can be configured to determine lane sensor data, which indicates lane markings on the road over which the vehicle is traveling (e.g., using a camera pointing at the road surface). Furthermore, the control unit can be configured to determine a route for sensing its own lane based on the lane sensor data. The sensing of its own lane can be based on lane markings and, possibly, map data. On the other hand, sensing its own lane can be determined independently of the behavior of other vehicles within its environment (specifically, independently of trajectory data regarding the trajectories of multiple vehicles within its environment).

[0018] The vehicle can be (automatically) guided laterally based on the route it senses within its own lane and the route it acts within its own lane, thus providing lane-keeping functionality. By taking into account both sensing and acting within its own lane, the robustness and safety of the lane-keeping function can be increased.

[0019] The control unit can be configured to compare the route sensed in its own lane with the route acted in its own lane. Specifically, the values ​​of one or more deviation parameters can be determined. Example deviation parameters include: lateral offset between the route sensed in its own lane and the route acted in its own lane, and / or angle between the route sensed in its own lane and the route acted in its own lane. The lane-keeping function can be operated based on the comparison between the route sensed in its own lane and the route acted in its own lane.

[0020] For example, comparing the route of the sensed lane with the route of the acting lane can be useful at construction sites where lane markings on the road can be confusing. Specifically, at construction sites, the original lane markings on the road can be overridden by temporary lane markings that are effective for the duration of the construction. The reliability of active lane keeping functionality can be improved by considering the determined route of the acting lane (in addition to considering the route of the sensed lane).

[0021] Specifically, the control unit can be configured to determine whether the route of the sensing lane deviates from the route of the behavior of the lane by more than a deviation threshold. The deviation threshold may depend on a confidence metric of the behavior of the lane, specifically such that if the confidence metric decreases, the deviation threshold increases, and / or if the confidence metric increases, the deviation threshold decreases.

[0022] If it is determined that the lane being sensed deviates from the lane being acted upon by more than a deviation threshold, lane keeping functionality based on (e.g., solely based on) the sensed lane can be interrupted, and / or lane keeping functionality can be performed based on the lane being acted upon. As a result, the robustness and / or safety of the lane keeping function can be increased.

[0023] According to another aspect, a motor vehicle (particularly a car, truck, or bus) is described, which includes the control unit described in this document.

[0024] According to another aspect, a method for operating automatic longitudinal and / or lateral guidance functions of a self-contained vehicle is described. The method includes: determining the route of a self-contained lane in which the self-contained vehicle is traveling, referred to as a behavior self-contained lane, based on the travel trajectories of multiple vehicles within the environment of the self-contained vehicle. Furthermore, the method includes: operating the automatic longitudinal and / or lateral guidance functions of the self-contained vehicle based on the route of the behavior self-contained lane.

[0025] According to another aspect, a software program is described. This software program is adapted to execute on a processor, and when executed on a processor, it is adapted to perform the method steps outlined in this document.

[0026] According to another aspect, a storage medium is described. This storage medium may include a software program adapted to execute on a processor, and when executed on the processor, the software program is adapted to perform the steps outlined in this document.

[0027] According to another aspect, a computer program product is described. This computer program may include executable instructions that, when executed on a computer, perform the method steps outlined in this document.

[0028] It should be noted that the methods and systems, including their preferred embodiments as outlined in this patent application, can be used alone or in combination with other methods and systems disclosed in this document. Furthermore, all aspects of the methods and systems outlined in this patent application can be combined arbitrarily. In particular, the features of the claims can be combined with each other in any manner. Attached Figure Description

[0029] The invention will now be described by way of example with reference to the accompanying drawings, in which:

[0030] Figure 1 An exemplary method for determining one or more lanes of a road is shown;

[0031] Figure 2 An exemplary traffic scenario on a road is shown;

[0032] Figure 3 Exemplary segments and trajectory clusters of a road traffic scenario are shown;

[0033] Figure 4 An exemplary multi-curve fitting function for estimating the shape of a road is shown;

[0034] Figure 5 A first example of lane configuration is shown;

[0035] Figure 6 A second example of lane configuration is shown;

[0036] Figure 7 A third example of lane configuration is shown;

[0037] Figure 8a An example of the target following function is shown;

[0038] Figure 8b An example of the lane-keeping function is shown; and

[0039] Figure 9 A flowchart is shown as an example method for operating a vehicle.

[0040] The illustrations in the accompanying drawings are schematic. It should be noted that in different drawings, similar or identical elements are given the same reference numerals or are given reference numerals that differ from the corresponding reference numerals only in the first digit. Detailed Implementation

[0041] Figure 1An example method 100 is illustrated for determining one or more lanes of a road in the environment of a vehicle (referred to herein as an ego vehicle). Lanes determined using method 100 are referred to herein as behavioral lanes because they are determined based on the collective behavior of vehicles within the ego vehicle's environment. Method 100 can be executed on a computer or electronic control unit of the ego vehicle. The ego vehicle's environment can be defined by sensor-detectable areas of the vehicle. Method 100 may include receiving and / or determining 102 multiple objects (e.g., other vehicles) in the ego vehicle's environment. For example, method 100 may receive and / or determine 102 multiple objects from an object detection component of the ego vehicle. For example, method 100 may receive and / or determine 102 multiple objects from an environmental model of the ego vehicle. An environmental model of the ego vehicle can be generated by fusing sensor data from one or more environmental sensors of the ego vehicle. Example environmental sensors for the ego vehicle are cameras, radar sensors, lidar sensors, ultrasonic sensors, etc.

[0042] Method 100 may further include receiving and / or determining 104 multiple trajectories for multiple objects (particularly vehicles) within the vehicle's own environment. Multiple trajectories can be received from the vehicle's trajectory determination component. For example, multiple trajectories can be determined by reconstructing a motion profile of the objects within the vehicle's own environment using temporary position and mileage data of the objects. In other words, multiple trajectories may include observations about the past motion of one or more of the multiple objects. Multiple trajectories can be received from the vehicle's own environment model.

[0043] Furthermore, method 100 may include estimating and / or determining the shape of road 106 based on multiple trajectories for multiple objects. When estimating the shape of road 106 based on multiple trajectories, method 100 may determine the number of road segments; cluster multiple trajectories in each of the one or more segments according to one or more shape similarity measures; determine trajectory clusters of the clustered multiple trajectories in each of the one or more segments, wherein the selected trajectory clusters may include a majority of trajectories that have the same or similar shape in a particular segment of the one or more segments; and estimate the overall shape of the road based on the determined trajectory clusters in each of the one or more segments.

[0044] Figures 2 to 4 An exemplary estimation of road shape in the environment of a vehicle is shown. Specifically, Figure 2An exemplary traffic scenario 200 of a road that can be detected by vehicle 202 (also referred to herein as itself) is illustrated. Vehicle 202 can identify and / or detect multiple vehicles in the environment of vehicle 202, including vehicles 204, 206, 208, 210, and 212. Additionally, the multiple vehicles may include itself. Vehicle 202 can also identify multiple trajectories corresponding to vehicles 202 through 212, including trajectories 214, 216, 218, 220, 222, and 224. Trajectory 214 may relate to vehicle 204, trajectory 216 may relate to vehicle 206, trajectory 218 may relate to vehicle 208, trajectory 220 may relate to vehicle 210, trajectory 222 may relate to vehicle 212, and trajectory 224 may relate to vehicle 202.

[0045] Figure 3 Presented Figure 2 Exemplary segment clustering 300 of multiple trajectories in an exemplary traffic scenario 200. Specifically, segment clustering 300 includes exemplary segments 302 and 304, or roads (along the longitudinal direction of a road). Each segment 302, 304 includes at least a portion of one or more of multiple trajectories 214 to 224. For example, segment 302 includes a portion of each of the multiple trajectories, and segment 304 also includes a portion of each of the multiple trajectories. Partial clustering of trajectories within a specific segment (e.g., segment 302 or segment 304) is performed using one or more similarity measures to form one or more trajectory shapes within that specific segment. Figure 3 As shown, segment 302 comprises a single cluster because all the trajectories of segment 302 have similar or identical shapes. Furthermore, segment 304 comprises three clusters: a first cluster comprising trajectory 216, a second cluster comprising trajectories 214 and 222, and a third cluster comprising trajectories 218, 220, and 224.

[0046] Within each segment, a trajectory cluster comprising the majority of trajectories can be identified (e.g., the cluster containing the highest number of trajectories compared to other clusters). For example, segment 302 comprises a single trajectory cluster because all trajectories have similar or identical shapes. Therefore, the single trajectory cluster of segment 302 comprises the majority of trajectories. Segment 304 has three clusters. The third trajectory cluster of segment 304 comprises three trajectories and therefore comprises the majority of trajectories in segment 304.

[0047] Different methods can be used to determine whether a pair of trajectories of a segment are similar or identical. For example, the lateral distance between two trajectories can be calculated within a unique region defined by a particular segment. Trajectories are considered similar or identical when the distance is below a predefined threshold. Additionally or alternatively, the span of lateral deviation, the longitudinal distance before reaching a specific lateral bifurcation, the azimuth deviation, and / or time-warped measures can be used to determine whether two or more trajectories are similar or identical to each other.

[0048] Furthermore, when determining a (majority) cluster of trajectories from multiple trajectories in each of one or more segments, a consistency check can be performed. If a particular segment has a preceding segment, for example, segment 304 has a preceding segment 302, and the trajectory cluster of the particular segment including the majority trajectories deviates from or is inconsistent with the determined trajectory cluster of the preceding segment, then the trajectory cluster of the particular segment that does not include the majority trajectories can be selected. For example, when one or more trajectories in the trajectory cluster of the majority trajectories of the preceding segment are discontinuous, but are not interrupted by one or more trajectories in the trajectory cluster of the particular segment including the majority trajectories, the trajectory cluster of the particular segment including the majority trajectories may be inconsistent with the determined trajectory cluster of the preceding segment. In this case, different trajectory clusters can be determined for the particular segment, particularly clusters in which the trajectories show a higher degree of continuity relative to the selected trajectory clusters in adjacent segments.

[0049] Figure 4 This demonstrates the use of a multi-curve fitting process for... Figure 2 An exemplary estimation of road shape for an exemplary traffic scenario 400. The road shape estimation is based on trajectories included in the determined trajectory clusters for all segments. For example, Figure 4 An exemplary multi-curve fitting function 402 is shown for those portions of trajectories 216, 218, and 220 included in the determined trajectory cluster. Preferably, the multi-curve fitting function 402 describes a curve that simultaneously best fits all trajectories included in the determined trajectory cluster. The shape of the curve of the multi-curve fitting function can be used to estimate the shape of a road. For example, an approximate shape curve can be obtained by using a regression method to simultaneously fit an approximative clothoid function to the trajectory—with a specific lateral offset parameter for each trajectory.

[0050] Furthermore, method 100 may include determining one or more lanes of road 108 (referred to herein as behavioral lanes) using the estimated road shape and multiple objects and / or multiple trajectories of multiple objects. To determine one or more lanes of road 208, method 100 may generate candidate lanes for each of the multiple trajectories, wherein the shape of the candidate lane is similar to or identical to the estimated road shape, and wherein, at least in segments where the shape of the trajectory is similar to or identical to the estimated road shape, the trajectory of the multiple trajectories is at the center of the candidate lane or within a predetermined range around the center of the candidate lane. Furthermore, method 100 may determine one or more lane sets from the generated candidate lanes, lane sets also referred to hereinafter as lane configurations, wherein the lane sets include only lanes that are distinct from each other from the generated candidate lanes; and method 100 may determine a score value for each of the determined one or more lane configurations, wherein the score value increases when the trajectory of a particular lane in the lane set is at the center of the particular lane or within a predefined range around the center of the particular lane. Finally, method 100 may determine one or more lanes of the road based on the lane set with the maximum score value. One or more (collective behavioral) lanes of the road may have the same shape as the estimated road shape in the vicinity of their own vehicle 202.

[0051] Figures 5 to 7 The illustration depicts an exemplary process for determining one or more lanes of Road 108 using the estimated road shape, multiple objects, and multiple trajectories of the multiple objects. Specifically, Figure 5 The first example of lane configuration 500 is shown. Figure 6 A second example of lane configuration 600 is shown, and Figure 7 A third example of lane configuration 700 is shown. Generally, lane configurations can satisfy a predefined set of constraints. Preferably, each lane in the lane configuration must have the same shape as the estimated road shape. Furthermore, preferably, two adjacent lanes in the lane configuration must be spaced apart by a predetermined minimum distance, for example, 2 meters.

[0052] To determine one or more lanes of a road, a trajectory from a plurality of trajectories that is at least partially similar to or identical to the estimated road shape is identified. For each trajectory that is at least partially similar to or identical to the estimated road shape, a candidate lane is generated. The candidate lane can be any possible lane of the road. The lane configuration includes one or more candidate lanes. If a candidate lane satisfies the constraints of the lane configuration as described above, the candidate lane can be considered through the lane configuration. Preferably, at least in the portion of the trajectory that is similar to or identical to the estimated road shape, the candidate lane is centered on one of the trajectories.

[0053] like Figure 5As shown, lane configuration 500 may include candidate lane 502 and candidate lane 504. Candidate lane 502 may be centered on a combination of trajectory 214 and trajectory 216. Candidate lane 504 may be centered on trajectory 218. Candidate lane 502 and candidate lane 504 share boundaries with each other. Figure 6 Lane configuration 600 is depicted, including a single candidate lane 602. The single candidate lane is centered on trajectory 220. Figure 7 Lane configuration 700 is presented, which includes three candidate lanes: candidate lane 702 centered on the combination of trajectory 214 and trajectory 216, candidate lane 704 centered on trajectory 218, and candidate lane 706 centered on trajectory 220.

[0054] Additionally, method 100 may include determining a score value for each of the determined one or more lane sets or lane configurations. The score value of a lane configuration may increase when a trajectory is centered on or within a predefined range around the center of a particular lane (i.e., a particular candidate lane). In other words, the score value defines the number of trajectories that can be covered by a particular lane set or lane configuration. Lane configuration 500 may have a score value of 3 because candidate lanes 502 and 504 of lane configuration 500 cover three trajectories. Lane configuration 600 may have a score value of 1 because candidate lane 602 covers only one trajectory. Lane configuration 700 may have a score value of 4 because candidate lanes 702, 704, and 706 cover four trajectories. However, candidate lanes 702, 704, and 706 of lane configuration 700 overlap, thus not satisfying the lane configuration constraints. Therefore, lane configuration 500 has the maximum score value. Since lane configuration 500 provides the maximum score value, method 100 may determine the lanes of the road based on lane configuration 500.

[0055] Advantageously, the road lanes can be determined using object trajectories describing the collective behavior of objects near the vehicle 202. This allows the vehicle to continue operating using (semi-)autonomous driving features or functions when lane detection systems based on lane markings are no longer available. Furthermore, this method allows (semi-)autonomous driving systems to evaluate the quality of data provided by marking-based lane detection systems. Additionally, this method can effectively support driver assistance systems. For example, if a majority of objects are observed to be curving to the right, the driver assistance system can use this information to infer the presence and geometry of lanes when no further information about the road lanes is available.

[0056] Method 100, as described above, can determine one or more (behavioral) lanes of a road. The determined lanes can be used by various driver assistance systems to provide more robust and readily available driver assistance systems in complex, high-volume traffic scenarios. For example, (behavioral) lanes 502, 504 of the road, determined from the collective behavior of vehicles surrounding the vehicle 202, can be used to directly control the vehicle 202 for active lane-keeping functionality. The vehicle 202 can mimic group behavior in active lane-keeping functionality. Furthermore, lanes 502, 504 of the road, determined from collective behavior, can be used to assign objects to the determined lanes or to sort objects using the determined lanes. This can be used to determine which objects on the road are in the same lane as the vehicle 202. Therefore, method 100 can be used to determine which vehicles on the road are associated with an adaptive cruise control (ACC) system. More specifically, method 100 can be used to determine which vehicles on the road should be considered relevant when controlling the longitudinal dynamics characteristics of their own vehicle 202 (e.g., acceleration, deceleration, and braking). Furthermore, the identified lanes on the road can be used to classify how other objects are being manipulated, e.g., to what extent other objects are changing lanes, canceling lane changes, or leaving the road. The vehicle 202 can use this information to perform its own maneuvers, e.g., accelerating and / or decelerating to perform a lane change, and / or accelerating and / or decelerating to cancel a lane change.

[0057] like Figure 8a As shown, a vehicle 202 can be operated according to a target vehicle 802 traveling in front of it within the same (behavioral) lane 502. The target vehicle 802 can be used to control the longitudinal dynamics of the vehicle 202 (e.g., in the context of an ACC system). If the target vehicle 802 performs a lane change, the vehicle 202 should be released from following the target vehicle 802 to avoid a potential collision between the vehicle 202 and a vehicle in the next lane 504.

[0058] As described above, the control unit of the vehicle 202, and particularly the vehicle 202, can be configured to determine one or more (behavioral) lanes 502, 504 of the road based on the collective behavior of vehicles 204, 206, 208, 210 traveling on the road. Specifically, the trajectories 214, 216, 218, 220 of vehicles 204, 206, 208, 210 (as in...) can be determined. Figure 2(As described in the context). Furthermore, the trajectories 214, 216, 218, and 220 of vehicles 204, 206, 208, and 210 can be clustered for different road segments 302 and 304 to determine the overall shape of the road for the segment sequences 302 and 304 (as described in...). Figure 3 (As described in the context). The shape of the road can be described using curve 402, which is fitted to the shape of segment sequences 302, 304 (e.g., ...). Figure 4 (As shown in the diagram). Using curve 402, which describes the collective shape of the road, and using the positions and / or trajectories of different vehicles 204, 206, 208, 210, one or more (behavioral) lanes 502, 504 of the road can be identified (as shown in the diagram). Figures 5 to 7 (as described in the context).

[0059] One or more lanes 502, 504 of a road, which have been determined (only) by the collective behavior of vehicles 204, 206, 208, 210 traveling on the road, may be referred to herein as behavioral lanes. On the other hand, one or more lanes of a road, which have been determined (only) by map data and / or sensor data indicating lane markings on the road, may be referred to herein as sensing lanes.

[0060] Analysis of the collective behavior of vehicles 204, 206, 208, and 210 within the environment of their own vehicle 202 provides a set of models (e.g., a set of curves) describing the geometry and / or routes 811 of the behavioral lanes 502 and 503, as inferred from the collective behavior of traffic participants. Furthermore, a confidence metric can be determined, providing an indication of the degree to which models of one or more behavioral lanes 502 and 504 can be trusted. Examples of confidence metrics include the number of trajectories 214, 216, 218, and 220 contributing to the collective shape 402 (different segments 302 and 304), and / or the number of vehicles 204, 206, 208, and 210 that can be semantically assigned to travel in each lane 502 and 504. Specifically, models for the behavioral lane 502 within which the own vehicle 202 and the target vehicle 802 are traveling can be determined. In addition, a confidence metric for the behavior of lane 502 can be determined.

[0061] To determine whether a target vehicle 802 is changing lanes, a motion profile or trajectory 812 of the target vehicle 802 (which may correspond to the trajectory of its own vehicle 202 following the target vehicle 802) can be compared with a model and / or route 811 of one or more driving lanes 502, 504. Specifically, a lateral offset 821 between the motion profile 812 of the target vehicle 802 and the route 811 of its own driving lane 502 can be determined. Alternatively or additionally, an angle 822 between the motion profile 812 of the target vehicle 802 and the route 811 of its own driving lane 502 can be determined. Therefore, the values ​​of one or more deviation parameters 821, 822 describing the deviation of the motion profile 812 from its own driving lane 502 can be determined.

[0062] Furthermore, the values ​​of one or more deviation parameters 821, 822 can be used to determine whether the deviation of the motion profile 812 from its own lane 502 exceeds a tolerable deviation threshold or a deviation zone (corridor) around its own lane 502. Then, it can be determined that the target object 802 is changing lanes and / or the target vehicle 802 can no longer be used by its own vehicle 202 as a target object for performing object following.

[0063] To reduce the number of false detections, a threshold can be determined based on one or more confidence metrics used for one or more behavior lanes 502, 504. Specifically, the confidence threshold can depend on the number of trajectories 214, 216, 218, 220 that contribute to determining the shape 811 of the behavior lane 502 (wherein the deviation threshold typically increases as the number of trajectories decreases). Alternatively or additionally, the deviation threshold can depend on the number of vehicles 204, 206, 208, 210 falling within the behavior lane 502 (wherein the deviation threshold typically increases as the number of vehicles decreases).

[0064] Therefore, a possible confidence measure for behavior within lane 502 is the number of objects 204, 206, 208, and 210 that contribute to the road shape 811 within specific segments 302 and 304. This confidence measure can be called "shape weight." It should be noted that although object trajectories 214, 216, 218, and 220 may have the same shape, they may actually be located within different lanes 502 and 503. The confidence measure can consider whether object trajectories 214, 216, 218, and 220 are located within their own lane 502.

[0065] Another confidence metric used for behavior of lane 502 could be the actual number of trajectories that contribute to a particular lane (specifically, lane 502 itself). This confidence metric is less than or equal to the shape weight (which is typically lane agnostic).

[0066] Depending on the use case, different combinations of these two confidence values ​​can be used (because different functions may have different tolerances in terms of availability, false positives, etc.). For example, for a function in which lane keeping is deactivated when the followed object changes lanes, considering only the shape weight reset confidence metric may be sufficient, as the relatively small number of false positives is tolerable (because the function will stop). However, if the function involves actively following the behavior itself in lane 502, it may be necessary to assign several objects to the behavior itself in lane 502 to increase the safety of the function.

[0067] The trajectory 812 of the target vehicle 802 and the route 811 of its own lane 502 may not be compared in absolute coordinates, but can be compared without considering possible constant lateral offsets between trajectory 812 and route 811. Specifically, trajectory 812 can be laterally moved to contact route 811 of its own lane 502 at a specified point. Furthermore, it can be verified whether trajectory 812 and route 811 deviate from each other. By ignoring the lateral offset between trajectory 812 and route 811, different driver behaviors (such as driving in the middle of the lane or near one side of the lane) can be considered within the same method.

[0068] Upon detecting that the target vehicle 802 is changing (behavioral) lanes 502 and 504, object tracking may be terminated and / or alternative vehicles for object tracking may be identified. Alternatively or additionally, the route 811 of behavioral lane 502 may be used to perform active lane keeping for its own vehicle 202.

[0069] In another scenario (in) Figure 8b As shown in the diagram, the vehicle 202 can perform automatic lane keeping, for example, based on lane markings on the road on which the vehicle 202 is traveling. Automatic lane keeping may be undesirable if there are obstacles in lane 502 on which the vehicle 202 is traveling and / or if traffic participants collectively decide not to follow lanes 852, 854 (as defined by markings) (e.g., this may occur in high-density traffic situations).

[0070] Lane keeping can be performed using sensing lanes 852 and 854 (which have been determined based on lane markings on the road). The control unit of the vehicle 202 can be configured to detect deviations between the vehicle's own lane 502 and the sensed own lane 852. This can be achieved by determining the values ​​of one or more deviation parameters 821 and 822 (such as lateral offset 821 or angle 822), which describe the deviation between the vehicle's own lane 502 and the sensed own lane 852.

[0071] Whether the deviation between the self-behaving lane 502 and the sensed lane 852 exceeds a deviation threshold can be determined based on the values ​​of one or more deviation parameters 821, 822. The deviation threshold may depend on one or more confidence metrics used for the self-behaving lane 502. If a deviation exceeding the deviation threshold is detected, automatic lane keeping can be terminated and / or automatic lane keeping can switch from the sensed lane 852 to the self-behaving lane 502. Alternatively or additionally, object following can be automatically activated (to perform object following instead of lane keeping).

[0072] Figure 9 A flowchart of an example method 900 for operating an automated longitudinal and / or lateral guidance function for a self-contained vehicle 202 is shown. Method 900 can be operated by a control unit of the self-contained vehicle 202. Method 900 includes determining, based on the travel trajectories 214, 216, 218, 220 of multiple vehicles 204, 206, 208, 210 within the environment of the self-contained vehicle 202, 901 the route 811 of a self-contained lane 502 in which the self-contained vehicle 202 is traveling, referred to as a behavioral self-contained lane. The route 811 of the self-contained lane 502 can be determined in a manner as outlined in this document (e.g., using method 100).

[0073] Furthermore, method 900 includes operating automatic longitudinal and / or lateral guidance functions (particularly lane keeping or vehicle following functions) of the vehicle itself 202 based on the route 811 of the behavior of its own lane 502. Robust and reliable automatic longitudinal and / or lateral guidance functions can be provided by taking into account the behavior of the own lane 502, which has been determined based on the collective behavior of vehicles 204, 206, 208, 210 within the environment of the vehicle itself 202.

[0074] It should be noted that the term "comprising" does not exclude other elements or steps, and the use of the articles "a" or "an" does not exclude multiple. Furthermore, elements described in association with different embodiments may be combined. It should also be noted that reference numerals in the claims should not be construed as limiting the scope of the claims. It should be noted that the specification and drawings merely illustrate the principles of the proposed methods and systems. Those skilled in the art will be able to implement various arrangements, which, although not explicitly described or shown herein, embody the principles of the invention and are included within the spirit and scope of the invention. Furthermore, all examples and embodiments outlined in this document are primarily and explicitly intended for illustrative purposes only to aid the reader in understanding the principles of the proposed methods and systems. Moreover, all statements provided herein regarding the principles, aspects, and embodiments of the invention and their specific examples are intended to cover their equivalents.

Claims

1. A method (900) for operating an automated longitudinal and / or lateral guidance function of its own vehicle (202), wherein the method (900) comprises: - Based on the driving trajectories (214, 216, 218, 220) of multiple vehicles (204, 206, 208, 210) in the environment of the self-vehicle (202), the route (811) of the self-vehicle lane (502) is determined, the self-vehicle is driving in the self-vehicle lane, the self-vehicle lane is referred to as the behavior self-vehicle lane; as well as - Operate the automatic longitudinal and / or lateral guidance function of the vehicle itself (202) according to the route (811) of the vehicle itself (502); The longitudinal and / or lateral guidance function includes a vehicle following function, which includes: performing longitudinal and / or lateral guidance of the vehicle (202) based on a target vehicle (802) traveling in front of the vehicle (202) within the vehicle's own lane (502); and The method further includes: - Determine the trajectory (812) of the target vehicle (802), the trajectory of the target vehicle is referred to as the target trajectory (812); - Compare the target trajectory (812) with the route (811) of the behavior's own lane (502); and - The vehicle following function is operated based on the comparison between the target trajectory (812) and the route (811) of the behavior's own lane (502); The method further includes: - Determine the trajectory data (214, 216, 218, 220) of the multiple vehicles (204, 206, 208, 210); - Divide the road on which the vehicle (202) is traveling into a sequence of segments (302, 304) consisting of multiple segments. - Based on one or more shape similarity measures, cluster the travel trajectories (214, 216, 218, 220) of the plurality of vehicles (204, 206, 208, 210) within each segment of the segment sequence (302, 304) to determine the shape of the road in the corresponding segment for each segment of the segment sequence (302, 304), wherein the determined corresponding shapes of the plurality of segments of the road constitute the shape sequence of the road; and - Based on the shape sequence of the road, determine the route (811) of the lane (502).

2. The method according to claim 1, wherein the method further comprises: - Determine the values ​​of one or more deviation parameters (821, 822), the values ​​of the one or more deviation parameters indicating the degree of deviation of the target trajectory (812) from the route (811) of the behavior's own lane (502); as well as - Operate the vehicle following function according to the values ​​of the one or more deviation parameters (821, 822).

3. The method of claim 2, wherein the one or more deviation parameters (821, 822) comprise: - The lateral offset (821) between the target trajectory (812) and the route (811) of the behavior's own lane (502); and / or - Angle (822) between the target trajectory (812) and the route (811) of the lane (502) of the behavior itself.

4. The method according to any one of claims 1 to 3, wherein the method further comprises: - Determine whether the target trajectory (812) deviates from the route (811) of the behavior's own lane (502) by more than a deviation threshold; as well as - If it is determined that the target trajectory (812) deviates from the route (811) of the behavior's own lane (502) by more than the deviation threshold, the vehicle following function with respect to the target vehicle (802) is interrupted.

5. The method according to claim 4, wherein the method further comprises: - Determine a confidence metric, which indicates: - The confidence level of the route (811) of the behavior itself lane (502) corresponding to the actual route of the lane of the road, the vehicle itself (202) is traveling on the road; and / or - Confidence and / or definiteness, the lane of the behavior itself has been determined with said confidence and / or definiteness (502). as well as - The deviation threshold is determined based on the confidence metric.

6. The method of claim 5, wherein the deviation threshold is determined based on the confidence metric such that if the confidence metric decreases, the deviation threshold increases, and / or if the confidence metric increases, the deviation threshold decreases.

7. The method according to any one of claims 1 to 3, wherein the method further comprises: - Determine lane sensor data indicating lane markings of the road lanes on which the vehicle (202) is traveling; - Based on the lane sensor data, determine the route for sensing its own lane (852); as well as - Based on the route of the sensed own lane (852) and the route (811) of the behavior of the own lane (502), the vehicle (202) is guided laterally to provide lane keeping function.

8. The method according to claim 7, wherein the method further comprises: - Compare the route of the sensing self-lane (852) with the route (811) of the behavior self-lane (502); as well as - The lane keeping function is operated based on the comparison between the route of the sensed lane (852) and the route (811) of the behavior lane (502).

9. The method of claim 8, wherein the method further comprises: - Determine whether the route of the sensing self-lane (852) deviates from the route (811) of the behavior self-lane (502) by more than a deviation threshold; as well as - If it is determined that the route of the sensing self-lane (852) deviates from the route (811) of the behavior self-lane (502) by more than the deviation threshold, the lane keeping function is interrupted based on the sensing self-lane (852), and / or the lane keeping function is performed based on the behavior self-lane (502).

10. The method according to any one of claims 1 to 3, wherein in order to determine the route of the lane itself based on the shape sequence of the road, the method further comprises: - Based on the shape sequence, determine the route (402) of the road along the segment sequence (302, 304). - Assign the plurality of vehicles (204, 206, 208, 210) to one or more distinct adjacent lanes (502, 504) of the road; wherein the one or more distinct adjacent lanes (502, 504) include the vehicle's own lane (502); and - Based on the route (402) of the road, determine the route (811) of the lane (502) of the road itself.

11. A control unit for an autonomous and / or semi-autonomous self-driving vehicle (202), wherein the control unit is configured to perform the method according to any one of claims 1 to 10.

Images