Method and system for adapting a driving trajectory according to a sensor field of view

Abstract

The present invention relates to a method for adapting a driving trajectory according to a sensor field of view, in which, from the driving operation of a fleet of vehicles, each equipped with a sensor set for environmental monitoring, a trajectory of the other road user is provided from a set of trajectories for a respective detected intersection situation with another road user, from the respective measurement data of the respective sensor set, from which respective clusters of trajectories on a respective section of a road course are determined, and from the respective cluster determined a most probable trajectory (121) is formed, in which a respective starting point of the trajectories is determined and compared with respective road boundaries, and from a respective crossing below the road boundaries, the presence of a respective sensor field of view restriction is determined.and in which, when an autonomous driving system detects a sensor field of view restriction for the ego vehicle, the driving trajectory (221) is selected which provides a sensor field of view encompassing an area of ​​the respective road boundary. Furthermore, a system with which the procedure can be implemented is presented.

Description

[0001] The present invention relates to a method for adapting a driving trajectory according to a sensor field of view. Furthermore, a system is presented with which the method can be implemented.

[0002] For example, the field of view of a vehicle's sensors may be so severely restricted during a curve in the road that the sensors are unable to detect the rest of the road or other road users in the approaching area. When driving around the curve in the road, the time to collision (TTC) between the initial detection of an object and the potential collision may be very short.

[0003] Document DE 11 2020 002 546 T5 discusses a method and device for avoiding a collision between a vehicle and another road user. To this end, the vehicle is steered along a safety-priority route.

[0004] Document DE 11 2019 001 046 T5 discloses an information processing device comprising a detection unit and a calculation unit. The detection unit acquires learning data, including trajectory data related to a trajectory along which a mobile body has moved. Based on the acquired learning data, the calculation unit calculates a cost function related to the movement of the mobile body, wherein the calculation is performed using inverse reinforcement learning.

[0005] The publication DE 10 2016 216 335 B4 describes a method for generating second trajectories for a first route section from a number of initial trajectories for that route section. This method can also calculate an accident probability based on the initial trajectories if they approach each other too closely or even touch and / or intersect.

[0006] The document DE 10 2020 003 073 B3 relates to a method and a device for the automated driving of a vehicle. In the case of a curve ahead of the vehicle, a visibility restriction expected due to the curve is determined by at least one detection unit of an environmental sensor system directed in the direction of travel of the vehicle. If it is determined that the visibility of the at least one detection unit falls below a predetermined threshold, the vehicle automatically changes lanes to the outside lane of the curve, provided that an outside lane exists.

[0007] The publication DE 10 2021 108 071 A1 deals with the determination of safety information regarding the safety of a roadway arrangement for motor vehicles. A spatial distribution of a road user measure for the probability of occurrence of road users in different sections of the roadway arrangement is determined.

[0008] The publication DE 10 2022 204 041 A1 discloses a method for generating a hazard map dataset, wherein other road users are detected and their trajectories are determined using vehicle sensors. The trajectories are combined into spatially grouped trajectory sets, and a respective environment is divided into corresponding areas. For each area, an area-specific activity value determined by the number of trajectories running there is determined. The areas are then classified according to a predetermined scale depending on the respective activity value. Using the determined area-specific classifications of the areas, the hazard map dataset, in which the area-specific classifications are entered, is generated and provided.

[0009] The publication DE 10 2020 102 278 A1 discusses data acquisition using a server unit to provide navigation assistance to a driver in a motor vehicle using the server unit. A method for augmenting an object for displaying driving instructions in the using motor vehicle, as well as a driver information system comprising all components, are also provided.

[0010] The document DE 10 2021 209 507 A1 describes setting a detection unit of a vehicle during a travel journey, wherein the detection unit detects an environmental area in front of the vehicle during the travel journey, a current position of the vehicle is determined using a position determination unit, and a position of a route section in front of the vehicle is provided, and / or the position of the route section is detected using the detection unit. The method includes providing travel trajectories and / or movement behavior of multiple vehicles with respect to traveling along the route section, adjusting an orientation and / or alignment of the detection unit depending on the travel trajectories and / or movement behavior of the multiple vehicles, so that the route section in front of the vehicle is detected while traveling along the route section.

[0011] Publication WO 2019 / 104112 A1 discusses methods for determining object motion and controlling autonomous vehicles. In one example embodiment, a computer system comprises processors and one or more tangible, non-transitory, computer-readable media collectively storing instructions that, when executed by the processors, cause the computer system to perform operations. The operations include acquiring data associated with a first object and one or more second objects in the environment of an autonomous vehicle. The operations include determining an interaction between the first object and the one or more second objects based at least in part on the data.The operations include determining one or more predicted trajectories of the first object within the surrounding environment based at least in part on the interaction between the first object and the one or more second objects. The operations include outputting data indicative of the one or more predicted trajectories of the first object.

[0012] The publication DE 10 2016 011 414 A1 relates to a method for warning a driver of a vehicle, in which a warning signal is transmitted to the vehicle, wherein the warning signal describes a hazard in the vehicle's surroundings, and the warning is issued to the driver depending on the warning signal. It is checked whether the hazard is within or outside the driver's current field of vision, and the warning is issued depending on whether the hazard is outside the current field of vision.

[0013] The document DE 10 2012 216 753 A1 discloses a method for assisting a driving maneuver of a vehicle, which method includes lateral control of the vehicle. The vehicle is guided along a trajectory from a current position of the vehicle to a target position, wherein at least one environment detection device detects a section of an area that will be traversed by the vehicle in the future and / or an area covered by the vehicle at the target position. The trajectory is calculated in such a way that the section of the area that will be traversed by the vehicle in the future and / or the area covered by the vehicle at the target position can be maximally detected by the environment detection device when the vehicle moves along the trajectory.

[0014] Against this background, it is an object of the present invention to propose a method for adapting a driving trajectory of an ego vehicle, which, when the sensor field of view is restricted, increases the time until an object with which there is a risk of collision is first detected. This is intended to gain time for collision-avoiding actions, such as braking and / or changing direction. Furthermore, a system is to be presented with which the method can be implemented.

[0015] To achieve the aforementioned object, a method for adapting a driving trajectory according to a sensor field of view is proposed, in which a respective trajectory of the other road user is provided from a set of trajectories from the respective measurement data of the respective sensor set during the driving operation of a fleet of vehicles, each equipped with a sensor set for environmental observation, for a respective intersection situation with another road user detected by the respective sensor set. From the set of trajectories, respective clusters of trajectories at a respective section of a road course are determined, and a respective trajectory bundle is formed from the trajectories of the respectively determined cluster. A most probable trajectory is formed from the respective trajectory bundle.A respective starting point of the trajectories of the respective trajectory bundle is determined and compared with the road boundaries of the respective section of the road. If the road boundaries are exceeded, the presence of a respective sensor field of view restriction is determined. When the ego vehicle approaches a respective section of a road with a detected sensor field of view restriction, the autonomous driving system selects the driving trajectory for the ego vehicle that provides a sensor field of view that covers an area of ​​the respective road boundary, compared to the sensor field of view restriction. The ego vehicle is controlled according to this driving trajectory.

[0016] The method according to the invention anticipates possible movements of other road users on a road with a confusing layout that could lead to a collision with the ego vehicle. It advantageously ensures that, by adjusting the driving trajectory, such road users are brought into the sensor field of view of the ego vehicle at an earlier point in time, so that detection is carried out by the vehicle's own sensors monitoring the surroundings and, in particular, the area in front of the vehicle. This implements anticipatory or predictive route planning for the ego vehicle, which enables timely braking or a timely change of direction to avoid a collision.

[0017] By observing the starting and end points of each trajectory, the invention allows conclusions to be drawn about its trajectory path, and a visible sensor field is derived from this, allowing the driving trajectory to be adjusted accordingly. For example, if the starting point of the trajectory for a road user approaching from the right lies between the road boundaries, the driving trajectory can be adjusted by drawing an arc toward the center of the road to expand the sensor field of view to a previously obscured area and enable early detection. Areas with an unrestricted sensor field of view are generally interrupted by areas with a restricted field of view and can therefore be clearly identified.

[0018] Typical sensors for environmental monitoring by a vehicle include radar, LiDAR and environmental cameras, especially cameras that are directed towards the area in front of the vehicle.

[0019] The respective section of the road under consideration is defined by the presence of trajectories from intersection situations, whereby an intersection situation exists precisely when, during a period of time between an object detection in the vehicle's area in front and a passing of the vehicle, there is an intersection point between a traveled trajectory and the trajectory of the other road user.

[0020] In one embodiment of the method according to the invention, the autonomous driving system reduces the driving speed when traveling through the driving trajectory compared to traveling through without a detected sensor field of view restriction. By reducing the driving speed, the time at which a collision object is first detected is advantageously postponed, thereby gaining time to decelerate the ego vehicle up to the point of emergency braking.

[0021] In a further embodiment of the method according to the invention, the driving trajectory is selected by which the sensor field of view of the ego vehicle is aligned with the most probable trajectory of the respective section of a road course with a detected sensor field of view restriction.

[0022] In yet another embodiment of the method according to the invention, for the respective section of a road with a detected sensor field of view restriction, the presence of a road area requiring particularly cautious driving, particularly near a kindergarten, school, or care facility, is determined from map material. If such a section is present, the autonomous driving system controls a particularly slow trajectory.

[0023] In a further embodiment of the method according to the invention, if there is a small number of trajectories in a respective trajectory bundle with starting points outside the respective road boundary, a driving trajectory is selected by which the sensor field of view of the ego vehicle is aligned to a region of respective starting points of the small number of trajectories.

[0024] Furthermore, a system for adapting a driving trajectory according to a sensor field of view is claimed, which system comprises a data processing unit, a fleet of vehicles, each equipped with a sensor set for environmental monitoring, and among them an ego vehicle with a data storage device and an autonomous driving system. From a driving operation of the fleet, a respective trajectory of the other road user from a trajectory set is provided for a respective intersection situation with another road user detected by the respective sensor set of a respective vehicle from respective measurement data of the respective sensor set. The data processing unit is configured to • to determine from the set of trajectories the respective accumulations of trajectories at a respective section of a road, • to form a respective trajectory bundle from the trajectories of the respective determined accumulation, • to form a most probable trajectory from the respective trajectory bundle, • to determine a respective starting point of the trajectories of the respective trajectory bundle and to compare it with the respective road boundaries of the respective section of the road and • to determine whether there is a restriction in the sensor field of view when the road limits are exceeded and to store this restriction in the data unit of the ego vehicle together with GPS values ​​of the respective section of the road.

[0025] The autonomous driving system is configured to • to compare a current location with GPS values ​​for sections of a road with a detected sensor field of view restriction during driving, • when the ego vehicle approaches a respective section of a road with a detected sensor field of view restriction, to select a driving trajectory for the ego vehicle which is suitable for providing a sensor field of view covering an area of ​​the respective road boundary in relation to the sensor field of view restriction, and • to control the ego vehicle according to this driving trajectory.

[0026] The data processing unit can, for example, be a central server on which all measurement results from sensors in a vehicle fleet are stored and analyzed. Transmission of measurement data from the sensors can, for example, occur wirelessly or during an inspection of the respective vehicle. The processed data, e.g., relating to sections of a road with a detected sensor field of view restriction, can be transferred to the data unit provided to the ego vehicle ex works, for example, with continuous data updates being possible wirelessly. Since the ego vehicle is also part of the vehicle fleet, measurement data on intersection situations measured by the ego vehicle's sensor set can also be continuously transmitted to the data processing unit.

[0027] In one embodiment of the system according to the invention, the autonomous driving system is configured to reduce a driving speed when driving through the driving trajectory compared to driving through without a detected sensor field of view restriction.

[0028] In a further embodiment of the system according to the invention, the autonomous driving system is configured to select the driving trajectory which is suitable for aligning the sensor field of view of the ego vehicle to the most probable trajectory of the respective section of a road course with a detected sensor field of view restriction.

[0029] In yet another embodiment of the system according to the invention, the autonomous driving system is configured to determine from map material the presence of a road area with a particularly cautious driving style, in particular near a kindergarten, a school or a care facility, for the respective section of a road course with a detected sensor field of view restriction, and to control a particularly slow driving through the driving trajectory if such a presence exists.

[0030] In a further embodiment of the system according to the invention, the autonomous driving system is configured to select a driving trajectory, if there is a small number of trajectories in a respective trajectory bundle with starting points outside the respective road boundary, which driving trajectory is suitable for aligning the sensor field of view of the ego vehicle to a region of respective starting points of the small number of trajectories.

[0031] Further advantages and embodiments of the invention will become apparent from the description and the accompanying drawings.

[0032] It is understood that the features mentioned above and those to be explained below can be used not only in the combination specified in each case, but also in other combinations or on their own, without departing from the scope of the present invention.

[0033] The characters are described coherently and comprehensively. Fig. 1 schematically shows a formation of a most probable trajectory in an embodiment of the method according to the invention. Fig. 2 schematically shows driving trajectories in the embodiment of the method according to the invention. Fig. 3 schematically shows a trajectory bundle in a further embodiment of the method according to the invention.

[0034] In Fig. 1 schematically shows the formation of a most probable trajectory 121 in one embodiment of the method according to the invention. In the representation 110, for a curved road 101, several trajectories 103 from detected intersection situations of respective sensor sets in vehicles of a fleet are shown. As an example of such a vehicle of the fleet, the ego vehicle 102 is shown, which is driving towards the curve in the direction of travel (arrow). During a data analysis of measurement data from vehicles of the fleet, all of these trajectories 103 are recognized as a cluster for a section of a road 101 and, as shown in representation 220, are combined into a most probable trajectory 121, e.g., by averaging. A trajectory 122 that is accessible to a full sensor field of view, for example, and which was averaged only from the two foremost trajectories 103, is rejected as improbable.

[0035] In Fig. 2 schematically shows driving trajectories 211, 221 in the embodiment of the method according to the invention. In the representation 210, the ego vehicle 102 follows a conventionally selected standard trajectory 211 for cornering and, halfway through cornering, is at a position 212 from which the sensor field of view cannot detect a right-hand road edge or a right-hand road boundary, even though a starting point of the most probable trajectory 121 lies there and, for example, a road user wishes to cross the road from there, as can be deduced from the fleet data. In contrast, a driving trajectory 221 intelligently selected with the aid of the method according to the invention enables a sensor field of view that advantageously detects the right-hand road edge from a position of the ego vehicle 222 when halfway through cornering. In the representation 230, the standard trajectory 211 is compared with the driving trajectory 221 according to the invention.The course of the driving trajectory 221 is shown to be curved more towards the centre of the road, which advantageously enables earlier observation of the right-hand side of the road.

[0036] In Fig. Figure 3 schematically shows a trajectory bundle with trajectories 303, 304, in which only a small number of trajectories—illustrated with only a single trajectory 304—have a starting point outside the right-hand road boundary. According to the invention, even in such a case, a driving trajectory 221 is selected through which the sensor field of view of the ego vehicle 102 can detect an area 305 of the starting point of trajectory 304. List of reference symbols 101 road curve 102 Ego Vehicle 103 Trajectory of a road user 110 trajectories of road users 120 Bundling of trajectories 121 Most likely trajectory 122 Unlikely trajectory 210 Curve approach with conventional standard trajectory 211 Conventional standard trajectory 212 Vehicle in curve 220 cornering with intelligently selected driving trajectory 221 Inventive travel trajectory 222 Ego vehicle in curve on inventive driving trajectory 230 Comparison of driving trajectories 300 Initial situation with few trajectories 303 trajectories with starting point on road 304 Trajectory with starting point at the roadside 305 Detection range for sensor field of view

Claims

[1] Method for adapting a driving trajectory according to a sensor field of view, in which a respective trajectory (103, 303, 304) of the other road user is provided from a set of trajectories from a driving operation of a fleet of vehicles, each of which is equipped with a sensor set for environmental observation, to a respective intersection situation with another road user detected by the respective sensor set, from respective measurement data of the respective sensor set, in which respective clusters of the trajectories (103, 303, 304) at a respective section of a road course are determined from the set of trajectories and a respective trajectory bundle is formed from the trajectories (103, 303, 304) of the respectively determined cluster, in which a most probable trajectory (121) is formed from the respective trajectory bundle, in which a respective starting point of the trajectories (103, 303,304) of the respective trajectory bundle is determined and compared with the respective road boundaries of the respective section of the road and if the road boundaries are exceeded, the presence of a respective sensor field of view restriction is determined, and in which an autonomous driving system, • when the ego vehicle (102) approaches a respective section of a road with a determined sensor field of view restriction for the ego vehicle (102), the driving trajectory (221) is selected which, compared to the sensor field of view restriction, provides a sensor field of view covering an area of ​​the respective road boundary, and • the ego vehicle (102) is controlled according to this travel trajectory (221). [2] Method according to claim 1, wherein the autonomous driving system reduces a driving speed when driving through the driving trajectory (221) compared to driving through without a detected sensor field of view restriction. [3] Method according to one of the preceding claims, in which the driving trajectory (221) is selected by which the sensor field of view of the ego vehicle is aligned with the most probable trajectory (121) of the respective section of a road course with a detected sensor field of view restriction. [4] Method according to one of the preceding claims, in which, for the respective section of a road course with a detected sensor field of view restriction, the presence of a road area with a particularly cautious driving style, in particular at a kindergarten or a school or a care facility, is determined from map material, and in the case of such a presence, the autonomous driving system controls a particularly slow driving through the driving trajectory (221). [5] Method according to one of the preceding claims, in which, when there is a small number of trajectories (304) in a respective trajectory bundle with starting points outside the respective road boundary, a driving trajectory (221) is selected by which the sensor field of view of the ego vehicle (102) is aligned to a region (305) of respective starting points of the small number of trajectories (304). [6] System for adapting a driving trajectory according to a sensor field of view, which comprises a data processing unit, a fleet of vehicles, each equipped with a sensor set for environmental observation, and among them an ego vehicle (102) with a data memory and an autonomous driving system, wherein from a driving operation of the fleet to a respective intersection situation with another road user detected by the respective sensor set of a respective vehicle, a respective trajectory (103, 303, 304) of the other road user of a trajectory set is provided from respective measurement data of the respective sensor set, wherein the data processing unit is configured to • to determine from the set of trajectories the respective clusters of the trajectories (103, 303, 304) at a respective section of a road course, • to form a respective trajectory bundle from the trajectories (103, 303, 304) of the respectively determined accumulation, • to form a most probable trajectory (121) from the respective trajectory bundle, • to determine a respective starting point of the trajectories (103, 303, 304) of the respective trajectory bundle and to compare it with the respective road boundaries of the respective section of the road course and • to determine the presence of a respective sensor field of view restriction when the road boundaries are undershot and to store this together with GPS values ​​of the respective section of the road in the data unit of the ego vehicle (102), and the autonomous driving system is configured to • to compare a current location with GPS values ​​for sections of a road with a detected sensor field of view restriction during driving, • when the ego vehicle (102) approaches a respective section of a road with a determined sensor field of view restriction for the ego vehicle (102), to select a driving trajectory (221) which is suitable for providing a sensor field of view covering an area of ​​the respective road boundary in relation to the sensor field of view restriction, and • to control the ego vehicle (102) according to this travel trajectory (221). [7] The system of claim 6, wherein the autonomous driving system is configured to reduce a driving speed when traveling through the travel trajectory (221) compared to traveling through without a detected sensor field of view restriction. [8] System according to one of claims 6 or 7, wherein the autonomous driving system is configured to select the driving trajectory (221) which is suitable for aligning the sensor field of view of the ego vehicle (102) with the most probable trajectory (121) of the respective section of a road with a detected sensor field of view restriction. [9] System according to one of claims 6 to 8, wherein the autonomous driving system is configured to determine from map material the presence of a road area with a particularly cautious driving style, in particular at a kindergarten or a school or a care facility, for the respective section of a road course with a detected sensor field of view restriction, and in the event of such a presence to control a particularly slow driving through the driving trajectory (221). [10] System according to one of claims 6 to 9, wherein the autonomous driving system is configured to select a driving trajectory (221) when there is a small number of trajectories (304) in a respective trajectory bundle with starting points outside the respective road boundary, which driving trajectory is suitable for aligning the sensor field of view of the ego vehicle (102) to a region of respective starting points of the small number of trajectories (304).

Images