Method and device for operating a driverless ego vehicle
The method and device for operating an ego vehicle prioritize human safety by minimizing personal injury and death over material damage in collision scenarios, adapting maneuvers to avoid or reduce collision consequences based on the presence of humans and drivable lanes.
Patent Information
- Application Number
- DE102021003571
- Authority / Receiving Office
- DE · DE
- Patent Type
- Patents
- Current Assignee / Owner
- Filing Date
- 2021-07-12
- Publication Date
- 2025-07-10
- Estimated Expiration
- 2041-07-12
AI Technical Summary
Existing methods for operating autonomously driving vehicles do not adequately prioritize the safety and well-being of human occupants in collision scenarios, particularly when determining strategies to avoid or minimize collision consequences.
A method and device for operating an ego vehicle that performs a situation analysis to detect potential collisions with manned or unmanned vehicles, prioritizing strategies that minimize human injury or death over material damage by considering the cost of personal versus object damage, and implementing maneuvers such as lateral avoidance or emergency braking based on the presence of humans and the drivability of adjacent lanes.
The ego vehicle behaves like a responsible human driver by minimizing personal injury and death while accepting potential material damage, ensuring that lives and physical integrity of humans are prioritized over object damage, even if it means causing damage to itself.
Smart Images

Figure 00000013_0000 
Figure 00000014_0000 
Figure 00000014_0001
Abstract
Description
The invention relates to a method for operating a self-propelled ego vehicle according to the features of the preamble of claim 1 and to a device for operating a self-propelled ego vehicle.As described in DE 10 2017 011 831 A1, a method and a device for operating an autonomously driving vehicle are known from the prior art. The vehicle can be operated in a first operating mode provided for manned operation and in a second operating mode provided for unmanned operation, wherein a situation analysis for detecting a risk of collision with unprotected road users is carried out in both operating modes, and wherein a strategy for avoiding or reducing collision sequences is determined in both operating modes when detecting a risk of collision with at least one unprotected road user as a function of the respective operating mode.DE 11 2017 006 845 T5 describes that one or more sensors associated with a vehicle recognize a roadway state of a first roadway and an avoidance maneuver is performed by the vehicle to avoid the recognized roadway state. In response to one or more processors determining that the avoidance maneuver was successful, a record of the successful maneuver and roadway condition is stored in a database. After storing the record in the database, one or more computers associated with one or more vehicles are trained to perform the avoidance maneuver in response to determining that the one or more vehicles are experiencing the roadway condition that the vehicle has encountered.DE 10 2017 209 533 A1 describes a lane change assistance system for a motor vehicle, which is configured to automatically control the motor vehicle in response to a lane change request of a driver of the motor vehicle in the context of an automated lane change maneuver with at least automated lateral guidance starting from an initial lane to a target lane, comprising an environment sensor system for determining environmental information relating to the environment of the motor vehicle, a control device coupled to the environment sensor system, and a guidance device controlled by the control device for at least automated lateral guidance of the motor vehicle, wherein the control device is designed such that the control device determines a first trajectory for driving in the initial lane or for driving back to the initial lane and a second trajectory for driving to the target lane when the control device determines on the basis of the environmental information, wherein a collision risk with a first object exists when driving along the first trajectory and a collision risk with a second object when driving along the second trajectory, and the control device compares effects of a possibly occurring collision with the first object when driving along the first trajectory and effects of a possibly occurring collision with the second object when driving along the second trajectory with one another.DE 10 2011 111 895 A1 describes an apparatus for preventing a collision, which is provided in a vehicle, having a detection device which is designed to detect the environment of the vehicle and provide it in the form of measurement data; having a transmission / reception device which is designed to transmit the detected measurement data and record measurement data from further vehicles; having an evaluation device which is designed to evaluate the detected and recorded measurement data with respect to an imminent collision; having a control device which has an interface via which the control device controls the vehicle, via which further vehicles in the environment of the vehicle can be controlled and via which the control device can be controlled by a control device of a vehicle in the environment of the vehicle, in such a way that the imminent collision is prevented or an accident severity of the imminent collision is reduced.The object of the invention is to specify a method for operating an ego vehicle which runs in a non-driver manner, which method is improved compared to the prior art, and a device for operating an ego vehicle which runs in a non-driver manner, which device is improved compared to the prior art, for operating an ego vehicle which runs in a non-driver manner.The object is achieved according to the invention by a method for operating a self-propelled ego vehicle having the features of claim 1 and a device for operating a self-propelled ego vehicle having the features of claim 6.Advantageous embodiments of the invention are the subject matter of the dependent claims.In a method for operating a vehicle driving in a non-driver's way, referred to below as an ego vehicle, according to the invention, a situation analysis is carried out in the ego vehicle for detecting an overtaking maneuver carried out by a second vehicle, in which the second vehicle is likely to no longer avoid a collision with the ego vehicle and / or with a third vehicle, wherein, when this overtaking maneuver carried out by the second vehicle is detected as a function of whether it has been detected in the situation analysis that at least one of the vehicles likely to collide with one another is manned, a strategy for avoiding or reducing collision consequences is determined in the ego vehicle. In the situation analysis, a manning of the respective vehicle involved in the anticipated collision is therefore also detected, i.e. it is at least attempted to detect whether the respective vehicle is manned or unmanned.A device according to the invention for operating the self-propelled ego vehicle is designed and configured to carry out this method.The device comprises in particular a computing unit which is designed and configured to carry out a situation analysis for identifying a passing maneuver carried out by a second vehicle, in which the second vehicle is expected to no longer be able to avoid a collision with the ego vehicle and / or with a third vehicle, and to determine a strategy for avoiding or reducing collision sequences when this passing maneuver carried out by the second vehicle is identified as a function of whether it has been identified in the situation analysis that at least one of the expected vehicles colliding with one another is manned.This strategy is preferably determined in cases in which it has been detected during the situation analysis that a collision due to full braking of the ego vehicle or due to a safe avoidance of the ego vehicle to an existing adjacent lane is likely not to be able to be avoided. A safe avoidance to a neighboring lane is to be understood here as meaning that neither the ego vehicle nor other road users are endangered by the avoidance to the neighboring lane.The solution described advantageously achieves the effect that the ego vehicle driving in a non-driver's way behaves like a responsible human driver of the vehicle, i.e. the life, the physical integrity and a minimization of a severity of injury to humans and thus the avoidance or minimization of personal damage to an avoidance or minimization of damage to the person, in particular also damage to the person of the ego vehicle and, for example, a load on the ego vehicle, being prioritized. The solution described thus also accepts damage to the ego vehicle and / or to the load thereof if the life and the health of humans can be saved as a result or at least the severity of injury can be reduced, i.e. if personal damage can be avoided or at least minimized as a result. The strategy can also include, for example, an action by which damage to the person, in particular on the ego vehicle and the load thereof, is caused, in order to thereby avoid or at least minimize damage to persons, in particular also with respect to occupants in the other vehicle involved or in the plurality of other vehicles involved and also with respect to persons outside the vehicles. The method thus advantageously achieves the effect that the self-propelled ego vehicle behaves in such a way that the lives and the physical integrity of humans are saved or at least their severity of injury is minimized.When determining the strategy for avoiding or reducing the collision consequences, damage to persons and damage to things are thus advantageously taken into account, wherein, for example, damage to the environment is also taken into account, expediently as damage to things. In this case, personal damage is weighted more heavily, in particular weighted much more heavily than person damage, so that when determining the strategy for avoiding or reducing collision consequences, the avoidance or at least minimization of personal damage is prioritized over the avoidance or minimization of person damage, in particular is prioritized considerably, for example with a weighting that is many times higher or higher by at least one order of magnitude or several orders of magnitude. In this case, in the determination of the strategy for avoiding or reducing the collision sequences, all currently detected road users are advantageously taken into account, i.e. road users, in particular persons, outside the ego vehicle, the second vehicle and the third vehicle, in particular persons in an area which is taken into account for avoiding the ego vehicle, thus in particular on a roadway edge area, in particular laterally next to the lane of the ego vehicle, in particular to an area next to the roadway of the vehicle.In one possible specific embodiment, if it has not been recognized whether the respective vehicle is manned, it is assumed that it is manned. This ensures that even if it is not safe, but there is the possibility that the respective vehicle is manned, the lives, the physical integrity and the minimization of the severity of injury of humans are prioritized over damage to things.According to the invention, an estimation of the costs of the collision sequences is carried out for the determination of the strategy for avoiding or reducing the collision sequences. As already mentioned above, costs are taken into account in this case by damage to persons and damage to things, wherein, for example, damage to the environment is also taken into account as damage to things. For example, this cost consideration already results in the above-described higher, in particular very much higher, weighting of the damage to persons compared to the damage to things, since damage to persons, if all the costs associated therewith are taken into account, are usually very much higher than damage to things. For example, treatment costs and damage replacement can be used as damage to persons, damage replacement for pain, permanent disability, loss of service and disadministration for a rear bliquor. In one possible embodiment, for example, flat values for damage to persons and damage to things can be specified as costs, wherein, as described above, damage to persons are weighted higher, in particular much higher, than damage to things.If the estimated costs exceed a predefined threshold value, an avoidance trajectory for the ego vehicle is determined according to the invention as a strategy for avoiding or reducing the collision sequences. If the estimated costs do not exceed the predefined threshold value, or if no avoidance trajectory can be determined, as a strategy for avoiding or reducing the collision sequences, for example, it is determined that the ego vehicle is maximally decelerated, i.e. a standard emergency braking is determined and carried out as a strategy, wherein the ego vehicle is advantageously kept in its lane.In one possible specific embodiment, if it is detected in the situation analysis that the second vehicle passes the third vehicle and approaches the ego vehicle in its lane, so that a collision of the ego vehicle with the second vehicle is imminent, as a strategy for avoiding or reducing the collision sequences, it is ascertained that the ego vehicle is avoiding the second vehicle, in particular is avoiding laterally, in particular is avoiding the roadway edge region, in particular next to the lane or next to the roadway of the ego vehicle, in order to avoid the collision of the ego vehicle with the second vehicle. The corresponding avoidance trajectory is thus determined.However, this is advantageously determined as a strategy for avoiding or reducing the collision sequences only if, as a result of this avoidance of the ego vehicle, no persons outside these three vehicles, in particular no persons on the roadway edge region, are endangered, i.e. in particular only if no persons are located on the roadway edge region. Furthermore, if it is detected that the second vehicle is unmanned and if, in addition, the ego vehicle is unmanned, this is advantageously carried out only if it is detected that the roadway edge region can be traveled in order to avoid the ego vehicle. Otherwise, as a strategy for avoiding or reducing the collision sequences, it is determined that the ego vehicle remains in its lane and is maximally decelerated, i.e. carries out the standard emergency braking.In one possible specific embodiment, if it is detected in the situation analysis that the second vehicle passes the ego vehicle and the third vehicle is approaching in oncoming traffic, so that a collision of the third vehicle with the second vehicle is imminent, it is ascertained as a strategy for avoiding or reducing the collision sequences that the ego vehicle avoids the second vehicle, in particular escapes laterally, in particular avoids on the roadway edge region, in particular next to the lane of the ego vehicle, in order thereby to enable the second vehicle to avoid and thus to avoid the collision of the third vehicle with the second vehicle. The corresponding avoidance trajectory is thus determined.However, this is advantageously determined as a strategy for avoiding or reducing the collision sequences only if, as a result of this avoidance of the ego vehicle, no persons outside these three vehicles, in particular no persons on the roadway edge region, are endangered, i.e. in particular only if no persons are located on the roadway edge region. Furthermore, if it is detected that the second vehicle and the third vehicle are unmanned, this is advantageously carried out only if it is detected that the roadway edge region can be traveled in order to avoid the ego vehicle.According to the invention, as a strategy for avoiding or reducing the collision sequences, it is determined that the ego vehicle is avoiding to a roadway edge region ifthe second vehicle approaches the ego vehicle in the lane of the ego vehicle and is manned, the ego vehicle is unmanned and no persons are located in the edge region of the roadway, orthe second vehicle approaches the ego vehicle in the lane of the ego vehicle, the roadway edge region is drivable and no persons are located in the roadway edge region, orthe second vehicle passes over the ego vehicle, the second vehicle and / or the third vehicle is manned, the ego vehicle is unmanned and no persons are located on the roadway edge region, orthe second vehicle passes over the ego vehicle, the roadway edge region is drivable and no persons are located on the roadway edge region.This means that, in the mentioned cases, the corresponding avoidance trajectory for the ego vehicle is determined as a strategy for avoiding or reducing the collision sequences.For example, as a strategy for avoiding or reducing the collision sequences, it is determined that the ego vehicle remains in its lane and decelerates to the maximum extent if the second vehicle approaches the ego vehicle in the lane of the ego vehicle and is unmanned, the ego vehicle is unmanned andat least one person is located on the roadway edge region, and / orthe roadway edge region is not drivable.This means that the standard emergency braking of the ego vehicle is determined in the mentioned cases as a strategy for avoiding or reducing the collision sequences, wherein the ego vehicle is advantageously kept in its lane.What is achieved as a result is that damage to persons is primarily avoided or at least reduced, and moreover damage to persons is advantageously also avoided or at least reduced, or if no persons are in risk, damage to persons is avoided or at least reduced.Whether or not the roadway edge region is trafficable can be ascertained, for example, on the basis of corresponding map data of a digital map and / or by means of a surroundings detection sensor system of the ego vehicle and / or on the basis of information which is transmitted to the ego vehicle, for example, by a vehicle-to-vehicle communication and / or vehicle-infrastructure communication and / or by a communication of the ego vehicle with at least one external device, in particular an information server.Whether or not the second vehicle and / or the third vehicle is manned can be determined, for example, by means of the environment detection sensor system of the ego vehicle and / or on the basis of information which is transmitted to the ego vehicle, for example, by a vehicle-to-vehicle communication and / or vehicle-infrastructure communication and / or by a communication of the ego vehicle with at least one external device, in particular an information server.Whether the ego vehicle is manned if the ego vehicle has the basic possibility for this can be determined, for example, by means of an occupant detection sensor system of the ego vehicle, for example by means of a seat occupancy detection and / or by means of at least one interior camera of the ego vehicle.The term "manned" is to be understood in particular as meaning that at least one human is present, i.e. is present on or in the vehicle referred to as manned. This can be a driver / vehicle driver of the vehicle or another occupant, for example a passenger. If the vehicle is unmanned, there is therefore no human on or in the vehicle.Exemplary embodiments of the invention are explained in more detail below with reference to drawings.The following are shown: FIG. 1 schematically shows an ego vehicle in a plan view from above, FIG. 2 schematically shows a traffic situation with the ego vehicle from FIG. 1 and a second vehicle passing a third vehicle and coming in the opposite direction to the ego vehicle, FIG. 3 schematically shows a traffic situation with the ego vehicle from FIG. 1 and a second vehicle passing the ego vehicle in case of oncoming traffic, FIG. 4 schematically shows a traffic situation with the ego vehicle from FIG. 1, a second vehicle passing the ego vehicle in oncoming traffic, and a person in a roadway boundary region adjoining a lane of the ego vehicle, FIG. 5 schematically shows a further traffic situation with the ego vehicle from FIG. 1 and a second vehicle passing a third vehicle and arriving at the ego vehicle, and FIG. 6 schematically shows a device for operating an ego vehicle driving in a non-driver's way.Corresponding parts are provided with the same reference numerals in all figures.With reference to FIGS. 1 to 6, a method and a device 1 for operating a vehicle driving in a non-driver's way, referred to below as ego vehicle Ego, are described below. The ego vehicle Ego is thus a vehicle that is autonomous and is driving without the presence of a driver or vehicle driver in the ego vehicle Ego, in particular a road vehicle.The ego vehicle Ego is illustrated by way of example in FIG. 1. FIGS. 2 to 5 show, by way of example, different traffic situations with the ego vehicle Ego driving in a non-driver fashion, and FIG. 6 shows, by way of example, a schematic illustration of the device 1.In the examples shown, the ego vehicle Ego is always not only free of drivers, but also unmanned, i.e. no passengers are present as occupants in the ego vehicle Ego. In the examples shown, the ego vehicle Ego is designed as a truck, more precisely as a semi-trailer having a semi-trailer tractor and a semi-trailer. In other embodiments, the ego vehicle Ego can also be designed, for example, as another truck, as an omnibus, as a passenger car, as a transporter or in another way. For example, it can then also be provided that although the ego vehicle Ego is driving in a non-driver's way, occupants are present as passengers in the ego vehicle Ego, i.e. then not unmanned.FIG. 1 shows the ego vehicle Ego by way of example in a schematic plan view illustration. To carry out the driver-less driving mode, the ego vehicle Ego comprises a surroundings detection sensor system 2 having at least one surroundings detection sensor 2.1, 2.2, 2.3 or advantageously, as illustrated here, a plurality of surroundings detection sensors 2.1, 2.2, 2.3. The respective environment detection sensor 2.1, 2.2, 2.3 is designed, for example, as a lidar sensor, radar sensor or as a camera, for example a stereo camera. As described and illustrated in FIG. 1, a plurality of environment detection sensors 2.1, 2.2, 2.3 are advantageously provided. In this case, a plurality of identical environment detection sensors 2.1, 2.2, 2.3 or combinations of one or more environment detection sensors 2.1, 2.2, 2.3 of one type with one or more environment detection sensors 2.1, 2.2, 2.3 of at least one other type can be provided.FIG. 1 also schematically shows detection ranges E 1, E 2, E 3 of environment detection sensors 2.1, 2.2, 2.3. In the example shown, these extend to an area in front of the ego vehicle Ego and to an area laterally next to the ego vehicle Ego.Furthermore, the ego vehicle Ego has a computing unit 3. The computing unit 3 is a component of the device 1 shown in FIG. 6 For example, the environment detection sensor system 2 is also a component of this device 1.The self-propelled ego vehicle Ego is advantageously located in a respectively present traffic infrastructure by means of its environment detection sensor system 2 and on the basis of map data of a digital map LK, and its driving behavior is advantageously matched to other road users detected and measured in particular by means of its environment detection sensor system 2.During this sensorless driving operation of the ego vehicle Ego, as shown by way of example in FIGS. 2 to 5 on the basis of various traffic situations, on traffic routes with oncoming traffic, dangerous overtaking maneuvers by other vehicles F 2, F 3 may occur, in the examples shown here, by a second vehicle F 2, wherein the ego vehicle Ego is taken over by the second vehicle F 2, as shown in FIGS. 3 and 4, or a third vehicle F 3 arriving at the ego vehicle Ego is taken over by the second vehicle F 2, as shown in FIGS. 2 and 5.In order for the ego vehicle Ego to react correctly in such traffic situations, the method described here for operating the ego vehicle Ego driving in a non-driver mode provides for a situation analysis to be carried out in the ego vehicle Ego for detecting an overtaking maneuver carried out by a second vehicle F 2, in which the second vehicle F 2 is likely to no longer be able to avoid a collision with the ego vehicle Ego and / or with a third vehicle F 3. If this is the case, i.e. if this passing maneuver carried out by the second vehicle F 2 is detected, a strategy for avoiding or reducing collision sequences is determined in the ego vehicle Ego. In principle, the strategy provides that the ego vehicle Ego initiates full braking if it can hereby avoid a collision, or that it changes to a possibly present adjacent lane if the change is possible without risk and a collision can thereby be avoided. The method described here is concerned with traffic situations in which a collision cannot be avoided in this way. In such cases, the strategy and thus its determination is dependent on whether it has been recognized during the situation analysis that at least one of the vehicles Ego, F 2, F 3 that are likely to collide with one another has been manned.The method achieves the effect that the ego vehicle Ego moving in a non-driver's way behaves as a responsible human driver of the vehicle, i.e. the life, the physical integrity and a minimization of a severity of injury to humans and thus the avoidance or minimization of personal damage to an avoidance or minimization of object damage, in particular also object damage to the ego vehicle Ego and, for example, a load on the ego vehicle Ego. Thus, the method also accepts damage to the ego vehicle Ego and / or to the load thereof if the life and health of humans can be saved as a result or at least the severity of injury can be reduced, i.e. if personal damage can be avoided or at least minimized as a result. The strategy can also include, for example, an action by which damage to the person, in particular to the ego vehicle ego and its load, is caused, in order to thereby avoid or at least minimize damage to persons. The method thus advantageously achieves the effect that the ego vehicle Ego moving in a non-driver manner behaves in such a way that the lives and the physical integrity of people are saved or at least their severity of injury is minimized.When determining the strategy for avoiding or reducing the collision consequences, damage to persons and damage to things are thus advantageously taken into account, wherein, for example, damage to the environment is also taken into account, expediently as damage to things. In this case, personal damage is weighted more heavily, in particular weighted much more heavily than person damage, so that when determining the strategy for avoiding or reducing collision consequences, the avoidance or at least minimization of personal damage is prioritized over the avoidance or minimization of person damage, in particular is prioritized considerably, for example with a weighting that is many times higher or higher by at least one order of magnitude or several orders of magnitude. In this case, in the determination of the strategy for avoiding or reducing the collision sequences, all currently detected road users are advantageously taken into account, i.e. road users, in particular persons P, outside the ego vehicle Ego, the second vehicle F 2 and the third vehicle F 3, in particular persons P in a region which is taken into account for avoiding the ego vehicle Ego, thus in particular on a roadway edge region FB, in particular laterally next to the lane of the ego vehicle Ego, in particular laterally or next to the roadway of the ego vehicle Ego.For example, if it has not been recognized whether the respective vehicle F 2, F 3 is manned, it may be provided that it is assumed that it is manned. This ensures that even if it is not safe, but there is the possibility that the respective vehicle F 2, F 3 is manned, the lives, the physical integrity and the minimization of the severity of injury of humans are prioritized over damage to the things.For example, for the determination of the strategy for avoiding or reducing the collision sequences, a determination of costs of the collision sequences is carried out. If the costs determined exceed a predefined threshold value, an avoidance trajectory AT is determined, for example, as a strategy for avoiding or reducing the collision sequences. If the determined costs do not exceed the predefined threshold value, or if no avoidance trajectory AT can be determined, as a strategy for avoiding or reducing the collision sequences, it is determined, for example, that the ego vehicle Ego is maximally decelerated, i.e. carries out a standard emergency braking NB, and in the process advantageously remains in its lane.For example, if it is detected in the situation analysis that the second vehicle F 2 passes the third vehicle F 3 and approaches the ego vehicle Ego in its lane, so that a collision of the ego vehicle Ego with the second vehicle F 2 is imminent, it is determined as a strategy for avoiding or reducing the collision sequences that the ego vehicle Ego avoids the second vehicle F 2, in particular escapes laterally, in particular avoids the roadway edge region FB, in particular adjacent to the lane, in particular adjacent to the roadway of the ego vehicle Ego, in order to avoid the collision of the ego vehicle Ego with the second vehicle F 2, as shown in FIG. 2. The corresponding avoidance trajectory AT is thus determined.However, this is advantageously determined as a strategy for avoiding or reducing the collision sequences only if, as a result of this avoidance of the ego vehicle Ego, no persons P outside these three vehicles Ego, F 2, F 3, in particular no persons P on the roadway edge region FB, are endangered, i.e. in particular only if no persons P are located on the roadway edge region FB. Furthermore, if it is detected that the second vehicle F 2 is unmanned, and if the ego vehicle Ego is unmanned, this is advantageously carried out only if it is detected that the roadway edge region FB can be traveled in order to avoid the ego vehicle Ego. Otherwise, as a strategy for avoiding or reducing the collision sequences, it is determined that the ego vehicle Ego remains in its lane and is maximally decelerated, i.e. carries out the standard emergency braking NB.For example, when it is detected in the situation analysis that the second vehicle F 2 passes the ego vehicle Ego and the third vehicle F 3 is approaching in oncoming traffic, so that a collision of the third vehicle F 3 with the second vehicle F 2 is imminent, it is determined as a strategy for avoiding or reducing the collision sequences that the ego vehicle Ego avoids the second vehicle F 2, in particular escapes laterally, in particular avoids onto the roadway edge region FB, in particular next to the lane of the ego vehicle Ego, in order thereby to enable the second vehicle F 2 to avoid and thus avoid the collision of the third vehicle F 3 with the second vehicle F 2, as shown in FIG. 3. The corresponding avoidance trajectory AT is thus determined.However, this is advantageously determined as a strategy for avoiding or reducing the collision sequences only if, as a result of this avoidance of the ego vehicle Ego, no persons P outside these three vehicles Ego, F 2, F 3, in particular no persons P on the roadway edge region FB, are endangered, i.e. in particular only if no persons P are located on the roadway edge region FB. Furthermore, if it is detected that the second vehicle F 2 and the third vehicle F 3 are unmanned, this is advantageously carried out only if it is detected that the roadway edge region FB can be traveled in order to avoid the ego vehicle Ego.The method is explained below with reference to the traffic situations according to FIGS. 2 to 5.In the traffic situation according to FIG. 2, the second vehicle F 2 passes over the third vehicle F 3 and arrives at the ego vehicle Ego in its lane. In the ego vehicle Ego, the situation analysis is carried out to identify this passing maneuver carried out by the second vehicle F 2, in which the second vehicle F 2 is expected to no longer be able to avoid the collision with the ego vehicle Ego, and this passing maneuver carried out by the second vehicle F 2 is thus identified. It is recognized in the situation analysis that there is no possibility that the second vehicle F 2 can end the passing maneuver and, even by maximally delaying the second vehicle F 2 and the ego vehicle Ego, the collision of the second vehicle F 2 with the ego vehicle Ego can no longer be avoided. Furthermore, it is recognized in the situation analysis that the second vehicle F 2 is manned, at least with very high probability, or it is assumed that the second vehicle F 2 is manned. There is thus the great risk of the second vehicle F 2 colliding with the ego vehicle Ego with a very high death-following probability for occupants in the second vehicle F 2. The ego vehicle Ego is unmanned, as already mentioned.As a function of the situation analysis having been detected or assuming that the second vehicle F 2 is manned, the strategy for avoiding or reducing the collision sequences is then determined in the ego vehicle Ego. In the example shown, this strategy is avoidance trajectory AT for ego vehicle Ego via roadway edge region FB laterally next to the lane of ego vehicle Ego, in order to thereby avoid oncoming second vehicle F 2. That is to say, as a strategy for avoiding or reducing the collision sequences, it is determined that the ego vehicle Ego is avoiding to the roadway edge region FB, in particular since there are no persons P on the roadway edge region FB that could be endangered by this avoidance of the ego vehicle Ego. When determining this strategy for avoiding or reducing the collision sequences, i.e. when determining this avoidance trajectory AT, it is immaterial whether or not the roadway edge region FB can be traveled on, since the ego vehicle Ego is unmanned. The possibility of the occurrence of object damage to the ego vehicle Ego and / or to its load is thus accepted, in order thereby to avoid personal damage to the occupants of the second vehicle F 2.In the traffic situation according to FIG. 3, the second vehicle F 2 passes over the ego vehicle Ego and thus comes against the third vehicle F 3 in oncoming traffic in its lane. In the ego vehicle Ego, the situation analysis is carried out to identify this passing maneuver carried out by the second vehicle F 2, in which the second vehicle F 2 is expected to no longer be able to avoid the collision with the third vehicle F 3, and this passing maneuver carried out by the second vehicle F 2 is thus identified. It is recognized in the situation analysis that there is no possibility that the second vehicle F 2 can end the passing maneuver, that the ego vehicle Ego can no longer delay sufficiently to allow the second vehicle F 2 to cut in in good time in front of the ego vehicle Ego, and that the second vehicle F 2 can no longer delay sufficiently to cut in behind the ego vehicle Ego. Furthermore, it is recognized in the situation analysis that the second vehicle F 2 and / or the third vehicle F 3 is manned, at least with very high probability, or it is assumed that the second vehicle F 2 and / or the third vehicle F 3 is manned. There is thus the great risk of the second vehicle F 2 colliding with the third vehicle F 3 with a very high death-following probability for occupants in the second vehicle F 2 and / or in the third vehicle F 3. The ego vehicle Ego is unmanned, as already mentioned.As a function of the situation analysis having been detected or assuming that the second vehicle F 2 and / or the third vehicle F 3 is manned, the strategy for avoiding or reducing the collision sequences is then determined in the ego vehicle Ego. In the example shown, this strategy is the avoidance trajectory AT for the ego vehicle Ego via the roadway edge region FB laterally next to the lane of the ego vehicle Ego, in order thereby to enable the second vehicle F 2 to avoid. That is to say, as a strategy for avoiding or reducing the collision sequences, it is determined that the ego vehicle Ego is avoiding to the roadway edge region FB, in particular since there are no persons P on the roadway edge region FB that could be endangered by this avoidance of the ego vehicle Ego. The second vehicle F 2 can thereby yield to the third vehicle F 3 in the previous lane of the ego vehicle Ego. When determining this strategy for avoiding or reducing the collision sequences, i.e. when determining this avoidance trajectory AT, it is immaterial whether or not the roadway edge region FB can be traveled on, since the ego vehicle Ego is unmanned. The possibility of the occurrence of object damage to the ego vehicle Ego and / or to its load is thus accepted, in order thereby to avoid personal damage to the occupants of the second vehicle F 2.If it is detected in the traffic situation according to FIG. 3 in the situation analysis that the second vehicle F 2 and the third vehicle F 3 are unmanned and that the roadway edge region FB laterally next to the lane of the ego vehicle Ego can be traveled, the strategy for avoiding or reducing the collision sequences in the ego vehicle Ego is advantageously likewise determined in the form of the avoidance trajectory AT for the ego vehicle Ego via the roadway edge region FB laterally next to the lane of the ego vehicle Ego, in order thereby to enable the second vehicle F 2 to avoid. That is to say, as a strategy for avoiding or reducing the collision sequences, it is determined that the ego vehicle Ego is avoiding to the roadway edge region FB, in particular since there are no persons P on the roadway edge region FB that could be endangered by this avoidance of the ego vehicle Ego. The second vehicle F 2 can thereby yield to the third vehicle F 3 in the previous lane of the ego vehicle Ego. As a result, damage to the object with respect to the second and third vehicles F 2, F 3 is avoided, and damage to the object for the subject vehicle Ego is not to be feared on account of the trafficability of the roadway edge region FB.The traffic situation according to FIG. 4 is similar to the traffic situation according to FIG. 3, i.e. the second vehicle F 2 passes over the ego vehicle Ego and thus arrives at the third vehicle F 3 in oncoming traffic in its lane. In the ego vehicle Ego, the situation analysis is carried out to identify this passing maneuver carried out by the second vehicle F 2, in which the second vehicle F 2 is expected to no longer be able to avoid the collision with the third vehicle F 3, and this passing maneuver carried out by the second vehicle F 2 is thus identified. It is recognized in the situation analysis that there is no possibility that the second vehicle F 2 can end the passing maneuver, that the ego vehicle Ego can no longer delay sufficiently to allow the second vehicle F 2 to cut in in good time in front of the ego vehicle Ego, and that the second vehicle F 2 can no longer delay sufficiently to cut in behind the ego vehicle Ego. Furthermore, in the situation analysis, person P is now recognized on roadway edge area FB laterally next to the lane of ego vehicle Ego. Furthermore, it is recognized in the situation analysis that the second vehicle F 2 and / or the third vehicle F 3 is manned, at least with very high probability, or it is assumed that the second vehicle F 2 and / or the third vehicle F 3 is manned.The strategy for avoiding or reducing the collision sequences is then determined in the ego vehicle Ego. The avoidance trajectory AT over the roadway edge region FB laterally next to the lane of the ego vehicle Ego is excluded in this case, since this would lead to the person P being at risk on the roadway edge region FB. It is assumed here that, even if it is detected in the situation analysis that the second vehicle F 2 and / or the third vehicle F 3 is manned, these occupants are better protected by the respective vehicle F 2, F 3, in particular by its passive occupant protection devices, and by energy absorption possibilities of the second vehicle F 2 and of the third vehicle F 3 than the unprotected person P on the roadway edge region FB in the event of a collision with the ego vehicle Ego. As a strategy for avoiding or reducing the collision sequences, it is therefore determined that the ego vehicle Ego remains in its lane, i.e. does not travel an avoidance trajectory AT, and therefore does not avoid the roadway edge region FB laterally next to the lane of the ego vehicle Ego.The traffic situation according to FIG. 5 is similar to the traffic situation according to FIG. 2, i.e. the second vehicle F 2 passes over the third vehicle F 3 and arrives at the ego vehicle Ego in its lane. In the ego vehicle Ego, the situation analysis is carried out to identify this passing maneuver carried out by the second vehicle F 2, in which the second vehicle F 2 is expected to no longer be able to avoid the collision with the ego vehicle Ego, and this passing maneuver carried out by the second vehicle F 2 is thus identified. It is recognized in the situation analysis that there is no possibility that the second vehicle F 2 can end the passing maneuver and, even by maximally delaying the second vehicle F 2 and the ego vehicle Ego, the collision of the second vehicle F 2 with the ego vehicle Ego can no longer be avoided. Furthermore, however, it is now recognized in the situation analysis that the second vehicle F 2 is unmanned. As already mentioned, the ego vehicle Ego is also unmanned. This is also detected in the situation analysis. There is thus the great risk of the second vehicle F 2 colliding with the ego vehicle Ego, but without a risk of death, since there are no occupants present both in the second vehicle F 2 and in the ego vehicle Ego. Furthermore, it is recognized in the situation analysis that roadway edge region FB laterally next to the lane of ego vehicle Ego is not drivable.As a function of the situation analysis having detected that the second vehicle F 2 and the ego vehicle Ego are unmanned, the strategy for avoiding or reducing the collision sequences is then determined in the ego vehicle Ego. In the example shown, this strategy is the stay of the ego vehicle Ego in its lane and a maximum deceleration of the ego vehicle Ego. This minimizes damage to the object caused by the collision of the ego vehicle Ego with the second vehicle F 2. Damage to persons does not occur because the vehicles Ego, F2 colliding with one another are unmanned. In contrast, when the roadway edge region FB that cannot be traveled on is avoided, higher damage would have to be expected, in particular for the ego vehicle Ego and / or its load.If it is detected in this traffic situation illustrated in FIG. 5 that the roadway edge region FB can be traveled laterally next to the lane of the ego vehicle Ego, the strategy for avoiding or reducing the collision sequences is likewise determined in the ego vehicle Ego on the basis of the situation analysis having detected that the second vehicle F 2 and the ego vehicle Ego are unmanned. This strategy is now the avoidance trajectory AT for the ego vehicle Ego over the roadway edge region FB laterally next to the lane of the ego vehicle Ego, in order thereby to avoid the oncoming second vehicle F 2. That is to say, as a strategy for avoiding or reducing the collision sequences, it is determined that the ego vehicle Ego is avoiding to the roadway edge region FB now identified as being drivable, in particular since there are no persons P on the roadway edge region FB that could be endangered by this avoidance of the ego vehicle Ego. In this way, the collision of the second vehicle F 2 with the ego vehicle Ego and the damage resulting therefrom can be avoided, and due to the trafficability of the roadway edge region FB laterally next to the lane of the ego vehicle Ego, no damage is to be feared either due to this avoidance of the ego vehicle Ego.If, in this traffic situation illustrated in FIG. 5, it is recognized that the roadway edge region FB can be traveled laterally next to the lane of the ego vehicle Ego, however, if a person P is recognized thereon, then the stay of the ego vehicle Ego in his lane and a maximum deceleration of the ego vehicle Ego are expediently determined as a strategy for avoiding or reducing the collision consequences, since the person P would be endangered by the avoidance to the lateral roadway edge region FB. Here, too, this would expediently also be determined as a strategy if the second vehicle F 2 were manned, since it can be assumed that the occupants in the second vehicle F 2 are better protected than the unprotected person P on the roadway edge region FB.Whether or not the roadway edge region FB can be traveled on can be determined, for example, on the basis of corresponding map data of the digital map LK and / or by means of the environment detection sensor system 2 of the ego vehicle Ego and / or on the basis of information which is transmitted to the ego vehicle Ego, for example, by a vehicle-to-vehicle communication and / or vehicle-infrastructure communication and / or by a communication of the ego vehicle Ego with at least one external device, in particular an information server.Whether or not the second vehicle F 2 and / or the third vehicle F 3 is manned can be determined, for example, by means of the environment detection sensor system 2 of the ego vehicle Ego and / or on the basis of information which is transmitted to the ego vehicle Ego, for example, by a vehicle-to-vehicle communication and / or vehicle-infrastructure communication and / or by a communication of the ego vehicle Ego with at least one external device, in particular an information server.Whether the ego vehicle Ego is manned if the ego vehicle Ego has the basic possibility for this can be determined, for example, by means of an occupant detection sensor system of the ego vehicle Ego, for example by means of a seat occupancy detection and / or by means of at least one interior camera of the ego vehicle Ego.In order to decide for the strategy for avoiding or minimizing the collision sequences whether the avoidance trajectory AT should be planned and the ego vehicle Ego should travel this avoidance trajectory AT, i.e. whether the ego vehicle Ego should avoid laterally, in particular to the roadway edge region FB laterally next to the lane of the ego vehicle Ego, or whether the ego vehicle Ego should only delay to the maximum, i.e. should perform the standard emergency braking NB, a cost determination is advantageously carried out and checked whether the determined costs are higher than a predefined threshold value. For cost determination, as described above, personal damage and object damage are taken into account. For the damage to persons, it is taken into account, for example, whether persons involved are most likely to die or are injured, for example. For example, a most likely injury severity is also taken into account. For example, the damage to the person concerned is taken into account as damage to the person concerned, the charge of the person concerned and, for example, additionally damage to the environment and damage to the other road users involved, in particular to the second vehicle F 2 and third vehicle F 3. For example, it is also taken into account whether damage to hazardous material could occur.If the determined costs do not exceed the predefined threshold value, the ego vehicle Ego is maximally decelerated as a strategy for avoiding or minimizing the collision sequences, i.e. the standard emergency braking NB is determined and carried out as a strategy.If the determined costs exceed the predefined threshold value, a check is made as a strategy for avoiding or minimizing the collision sequences as to whether the avoidance trajectory AT is possible, and if it is possible, this avoidance trajectory AT is used, i.e. the avoidance of the ego vehicle Ego is determined as a strategy for avoiding or minimizing the collision sequences, in particular laterally, in particular across the roadway edge region FB laterally next to the lane of the ego vehicle Ego. If this is not possible, i.e. if no avoidance trajectory AT can be determined, the ego vehicle Ego is also maximally decelerated here as a strategy for avoiding or minimizing the collision sequences, i.e. the standard emergency braking NB is determined and carried out as a strategy.FIG. 6 shows the device 1 for operating the self-propelled ego vehicle Ego. It is designed and configured to carry out the described method. The ego vehicle Ego includes this device 1.The device 1 comprises the aforementioned computing unit 3, the digital map LK, a sensor processing unit 4 and a data fusion unit 5. A current position of the ego vehicle Ego is determined by means of a position determination unit 6 of the ego vehicle Ego and the digital map LK, for example by means of a global navigation satellite system. Data fusion 5 of the sensor information processed by sensor processing 4 and the current position of ego vehicle Ego is then carried out. Both the merged data and the current position of the ego vehicle Ego are fed to a behavior and planning module 7 of the computing unit 3.The described situation analysis and a planning of the described strategy for avoiding or reducing the collision sequences take place therein in a corresponding situation analysis and planning module 8. In an extreme scenario module 9, the described traffic situations in which a passing maneuver, in which the second vehicle F 2 is expected to no longer avoid a collision with the ego vehicle Ego and / or with the third vehicle F 3, are recognized and analyzed by the second vehicle F 2 in order to determine the strategy for avoiding or reducing the collision sequences.In a first step S 1, it is checked whether such a traffic situation, i.e. a passing maneuver carried out by the second vehicle F 2, in which the second vehicle F 2 is expected to no longer be able to avoid a collision with the ego vehicle Ego and / or with the third vehicle F 3, has been detected. If n, no trajectory adaptation kTA takes place, i.e. no strategy for avoiding or reducing collision sequences is determined. A trajectory T determined in another way for carrying out the keyless driving operation of the ego vehicle Ego is then expediently retained.If yes, i.e. if in the first step S 1 a passing maneuver carried out by the second vehicle F 2, in which the second vehicle F 2 is expected to no longer be able to avoid a collision with the ego vehicle Ego and / or with the third vehicle F 3, has been detected, then in a second step S 2 a determination of the costs of the collision sequences is carried out and a check is made as to whether the determined costs are higher than a predefined threshold value. For cost determination, as described above, personal damage and object damage are taken into account. For the damage to persons, it is taken into account, for example, whether persons involved are most likely to die or are injured, for example. For example, a most likely injury severity is also taken into account. For example, the damage to the person concerned is taken into account as damage to the person concerned, the charge of the person concerned and, for example, additionally damage to the environment and damage to the other road users involved, in particular to the second vehicle F 2 and third vehicle F 3. For example, it is also taken into account whether damage to hazardous material could occur.If this check as to whether the determined costs are higher than the predefined threshold value is answered with no n, then the maximum deceleration of the ego vehicle Ego is determined as a strategy for avoiding or reducing the collision sequences, i.e. the standard emergency braking NB of the ego vehicle Ego in order to reduce the collision sequences.If this check as to whether the determined costs are higher than the predefined threshold value is answered with yes, then a third step S 3 checks whether the avoidance trajectory AT for the ego vehicle Ego is possible as a strategy for avoiding or reducing the collision sequences, i.e. the avoidance of the ego vehicle Ego, in particular to the roadway edge region FB laterally next to the lane of the ego vehicle Ego. If n, i.e. if this is not possible, then the maximum deceleration of the ego vehicle Ego is determined as a strategy for avoiding or reducing the collision sequences, i.e. the standard emergency braking NB of the ego vehicle Ego, in order to reduce the collision sequences. If j, i.e. if the avoidance trajectory AT is possible, then the determined avoidance trajectory AT is applied, i.e. used for the sensorless driving operation of the ego vehicle Ego.The respective result of the extreme scenario module 9 is transmitted to a trajectory generator 10 of the behavior and planning module 7 and used for trajectory generation. The generated trajectory T is then transmitted to an actuator system 11 of the ego vehicle Ego and used there for carrying out the driver-less driving operation. The actuator system 11, also referred to as actuator system, comprises in particular a steering device, a braking device and a drive train of the ego vehicle Ego.The computing unit 3, in particular its behavior and planning module 7, advantageously further comprises a data record 12, in which all relevant data of the described method and of the described procedure are recorded in the apparatus 1. This is important for subsequent legal and police awareness and analysis. In particular, if the strategy for avoiding or reducing collision consequences could avoid damage to the second and third vehicles F 2, F 3 and their occupants, but if damage was caused thereby to the ego vehicle Ego, this data record 12 is very important for proving who is originally responsible for these damage and therefore has to replace it.
Claims
Method for operating an ego vehicle (Ego) driving in a non-driver manner, wherein a situation analysis for detecting an overtaking maneuver carried out by a second vehicle (F2), in which the second vehicle (F2) is expected to no longer be able to avoid a collision with the ego vehicle (Ego) and / or with a third vehicle (F3), is carried out in the ego vehicle (Ego), wherein a strategy for avoiding or reducing collision sequences is determined in the ego vehicle (Ego) when this overtaking maneuver carried out by the second vehicle (F2) is detected as a function of whether it has been detected in the situation analysis that at least one of the expectedly colliding vehicles (Ego, F2, F3) has been manned, wherein an estimation of the costs of the collision sequences is carried out for the ascertainment of the strategy for avoiding or reducing the collision sequences, wherein an avoidance trajectory (AT) is ascertained as a strategy for avoiding or reducing the collision sequences, if the estimated costs exceed a predefined threshold value, - it is ascertained that the ego vehicle (Ego) is maximally decelerated, if the estimated costs do not exceed a predefined threshold value or if no avoidance trajectory (AT) can be ascertained, wherein it is ascertained as a strategy for avoiding or reducing the collision sequences that the ego vehicle (Ego) avoids on a roadway edge region (FB), - if the second vehicle (F2) approaches the ego vehicle (Ego) on the lane of the ego vehicle (Ego) and is manned, the ego vehicle (ego) is unmanned and no persons (P) are located on the roadway edge area (FB), or - if the second vehicle (F2) approaches the ego vehicle (ego) on the lane of the ego vehicle (ego), the roadway edge area (FB) is trafficable and no persons (P) are located on the roadway edge area (FB), or - if the second vehicle (F2) passes the ego vehicle (ego), the second vehicle (F2) and / or the third vehicle (F3) is manned, the ego vehicle (ego) is unmanned and no persons (P) are located on the roadway edge area (FB), or if the second vehicle (F2) passes over the ego vehicle (Ego), the roadway edge region (FB) is drivable and no persons (P) are located on the roadway edge region (FB).Method according to Claim 1, characterized in that the strategy is determined in cases in which it has been detected during the situation analysis that a collision can probably not be avoided by full braking of the ego vehicle (Ego) or by safe avoidance of the ego vehicle (Ego) to an existing adjacent lane.Method according to one of the preceding claims, characterized in that, if it has not been detected whether the respective vehicle (Ego, F2, F3) is manned, it is assumed that it is manned.Method according to one of the preceding claims, characterized in that person damage and object damage are taken into account in the determination of the strategy for avoiding or reducing the collision consequences, person damage being weighted more highly than object damage.Method according to one of the preceding claims, characterized in that the strategy for avoiding or reducing the collision sequences is determined to be that the ego vehicle (Ego) remains in its lane and delays it to the maximum if the second vehicle (F2) approaches the ego vehicle (Ego) in the lane of the ego vehicle (Ego) and is unmanned, the ego vehicle (Ego) is unmanned and - at least one person (P) is located on the edge region (FB), and / or - the edge region (FB) is not drivable.Device (1) for operating a self-propelled ego vehicle (Ego) designed and configured to carry out a method according to one of the preceding claims.Device (1) according to Claim 6, comprising a computing unit (3) which is designed and configured to carry out a situation analysis for identifying a passing maneuver carried out by a second vehicle (F2), in which the second vehicle (F2) is expected to no longer be able to avoid a collision with the ego vehicle (Ego) and / or with a third vehicle (F3), and to determine a strategy for avoiding or reducing collision consequences when this passing maneuver carried out by the second vehicle (F2) is identified as a function of whether the situation analysis has identified that at least one of the vehicles (Ego, F2, F3) which are expected to collide with one another has been manned.
Citation Information
Patent Citations
Device and method for preventing a vehicle collision, vehicle
DE102011111895A1
Method and device for operating an autonomously driving vehicle
DE102017011831A1
Lane change assistance system and lane change assistance method with increased safety for the driver and other road users
DE102017209533A1
TRAINING A SELF-DRIVING VEHICLE
DE112017006845T5