Method and system for determining a free space accessible to a vehicle

Abstract

The invention relates to a method for determining a free space (F) that can be driven through by an ego vehicle (2) by means of a plurality of free space segments (S0 - S3), based on which a check of the free space that can be driven through by the ego vehicle (2) without collision is carried out.

Description

The invention relates to the field of vehicle assistance systems. In particular, the invention relates to a method and a system for determining a free space that can be traveled by an ego vehicle.Systems for planning a free space which can be traveled by a vehicle without a collision are known in principle. In particular, systems are known which, based on an estimated trajectory of the ego vehicle and an estimated trajectory of a third-party vehicle, referred to below as a third-party vehicle trajectory, iteratively determine a free space to be traveled on without collision in each case for the entire prediction period.The problem here is that the calculation of the free space that can be traveled on without collision is very computationally expensive, since the iterative calculation of the free space is very complex on the basis of the plurality of objects to be taken into account and possibly likewise moving.Proceeding from this, it is an object of the invention to specify a method for determining a free space which can be traveled on by an ego vehicle, which enables a reliable calculation of the free space which can be traveled on without collision and which protects against computing resources.The object is achieved by a method having the features of independent claim 1. Preferred embodiments are the subject matter of the dependent claims. A system for determining a free space to be traveled by the ego vehicle is the subject matter of independent patent claim 10.According to a first aspect, a method for determining a free space which can be traveled on by an ego vehicle without collision is disclosed. The method comprises the following steps:First, the ego trajectory of the ego vehicle and a third-vehicle trajectory of at least one third-party vehicle are estimated in a prediction period. The ego trajectory predicts the movement of the ego vehicle, the third-vehicle trajectory predicts the movement of the third vehicle in the prediction period.A first free space is then defined. This first free space forms a basic framework for the free space to be determined, also referred to below as second free space or final free space, which is forwarded to a trajectory planner for planning a collision-free trajectory for the ego vehicle. The first free space is preferably defined without taking into account third-party objects, in particular without taking into account predicted positions of moving objects, in particular third-party vehicles.Subsequently, based on the estimated ego trajectory, the future positions of the ego vehicle are ascertained at different points in time in the prediction period. In other words, the prediction period is discretized over time and as a result a plurality of points in time in the prediction period is defined. Based on the estimated ego trajectory, the positions that the ego vehicle assumes at these points in time are determined. The points in time can be selected in particular equidistantly in the prediction period.Based on the determined future positions of the ego vehicle, the first free space is then divided into free space segments. The successive free space segments preferably adjoin one another directly. The free space segments preferably have a length running in the direction of travel which is greater than the vehicle length.Based on the third-vehicle trajectory, the future positions of the third-vehicle are estimated at the different points in time in the prediction period. The points in time correspond to those points in time which are used to determine the positions of the ego vehicle.Subsequently, for the respective points in time in the prediction period, the future position of the third vehicle at the respective point in time is separately compared with the free space segment in which the ego vehicle is located at this point in time. In particular, it is determined whether the third vehicle approaches the respective free space segment closer than a safety threshold value or is at least partially located in the respective free space segment.At least one free space segment is then restricted if the estimated future position of the third vehicle at the respective point in time indicates that the third vehicle is approaching the free space segment in which the ego vehicle is located at this point in time, closer than a safety threshold value or that the third vehicle at least partially enters this free space segment. "Limiting" is understood here in particular to mean reducing the area of the free space segment by modifying the lateral delimitation of the free space segment, such that the width of the free space segment running transversely to the direction of travel is reduced. This results in at least partially modified free space segments.Finally, a second, final free space is formed by joining the free space segments, wherein at least one of the free space segments is modified.The method has the technical advantage that by dividing the free space extending over the entire prediction period into a plurality of free space segments, which are selected on the basis of the position of the ego vehicle and separate the check for the free space segments whether a third vehicle exactly enters this free space segment at the respective point in time at which the ego vehicle is located in the free space segment, the computational complexity is decisively reduced. In addition, by taking into account the movement of the ego vehicle when creating the free space segments, it is possible to distinguish whether or not a predicted movement of a third-party vehicle has relevance for free space planning.According to one exemplary embodiment, the first free space is designed in the form of a hose and runs along the estimated ego trajectory of the ego vehicle. As a result, the first free space forms a corridor in which the ego vehicle can fundamentally move in the prediction period and which can be used as a basis for determining whether this first free space is to be restricted on the basis of third-party vehicles or other obstacles.According to one exemplary embodiment, the free space segments each comprise the entire width of the first free space and each a partial length of the first free space. The partial length extends in the direction of travel of the ego vehicle, the width of the first free space is measured transversely to this direction of travel. In other words, the segmentation of the first free space thus takes place only in the direction of travel of the ego vehicle, but not transversely to the direction of travel of the ego vehicle. As a result, the first free space can be divided into free space segments, which can be assigned to the different points in time in the prediction period and to different positions of the ego vehicle in the prediction period.According to one exemplary embodiment, the free space segments are each formed centered with respect to the position which the ego vehicle has at a discrete point in time in the prediction period. The free space segment preferably has a length which is greater than the length of the ego vehicle. It is thus possible, based on the respective point in time, to determine a free space segment in which the ego vehicle is located at this point in time. Likewise, an assignment of a third-party vehicle position to a free space segment is possible via the respective point in time, in which the ego vehicle is located at this point in time. The determination of the free space to be traveled on without collision can thus be carried out in each case on the basis of the relevant regions of the first free space.According to one exemplary embodiment, the restriction of at least one free space segment comprises an at least partial reduction in the width of the free space segment, i.e. the free space segment is laterally restricted in such a way that no overlap with the third vehicle or a safety zone assigned to this third vehicle takes place.According to one exemplary embodiment, a safety zone is formed around the third vehicle. The free space segment is restricted or tailored if the estimated future position of the third-party vehicle indicates at a certain point in time an at least partial intrusion of the safety zone of the third-party vehicle into the free space segment in which the ego vehicle is located at this point in time. The safety zone can create a safety buffer which has the effect that the second free space, which is produced by joining the free space segments together, is collision-free with high certainty.According to one exemplary embodiment, the size of the safety zone is selected as a function of the quality of the third-party vehicle trajectory.In particular, if the quality of the third-vehicle trajectory is low, i.e. the third-vehicle trajectory is subject to a relatively large uncertainty, the safety zone can be selected to be greater than in the case of a high quality of the third-vehicle trajectory. As a result, the safety during assisted or autonomous driving processes can be significantly increased.According to one exemplary embodiment, the length of the free space segment is selected as a function of the quality of the ego trajectory and / or the quality of the third-party vehicle trajectory. In particular, if the quality of the ego trajectory and / or of the third-vehicle trajectory is low, the length of the free space segment can be selected to be greater than in the case of an ego trajectory and / or of the third-vehicle trajectory having a higher quality. This prevents that no or only inadequate restrictions on the free space occur in the case of inaccurate estimates of the position of the ego vehicle and / or of the third vehicle on account of too fine a discretisation of the free space.According to one exemplary embodiment, the length of the free space segment is selected as a function of the speed of the ego vehicle. In particular, the length of the free space segments can be increased with increasing speed of the ego vehicle. This increases the safety of the planned free space.According to a further aspect, a system for determining a free space to be traveled by an ego vehicle is disclosed. The system comprises a computing unit configured to perform the following steps:estimating the ego trajectory of the ego vehicle and a third-vehicle trajectory of at least one third-vehicle for a prediction period;defining a first free space;determining the future positions of the ego vehicle at different points in time in the prediction period based on the estimated ego trajectory;dividing the first free space into free space segments based on the determined future positions of the ego vehicle;estimating the future positions of the third vehicle at the different points in time in the prediction period based on the third vehicle trajectory;separately for the points in time in the prediction period, in each case, comparing the future position of the third vehicle at the respective point in time with the free space segment in which the ego vehicle is located at this point in time;restricting at least one free space segment if the estimated future position of the third vehicle at the respective point in time indicates an approach of the third vehicle to the free space segment in which the ego vehicle is located at this point in time closer than a safety threshold value or an at least partial intrusion of the third vehicle at the respective point in time into the free space segment in which the ego vehicle is located at this point in time;forming a second free space by joining the free space segments together.The terms "approximately", "substantially" or "about" mean, for the purposes of the invention, deviations from the exact value in each case by + / - 10%, preferably by + / - 5%, and / or deviations in the form of changes which are insignificant for the function.Developments, advantages and possible applications of the invention also result from the following description of exemplary embodiments and from the figures. All features described and / or graphically depicted are fundamentally the subject matter of the invention, either alone or in any combination, independently of their summary in the claims or their reference back. The contents of the claims are also made part of the description.The invention is explained in more detail below on the basis of the figures using exemplary embodiments. The following are shown: FIG. 1 shows, by way of example, a schematic illustration of a system for determining a free space which can be traveled on without collision; FIG. 2 shows, by way of example and schematically, a plan view illustration of a driving situation of an ego vehicle in which free space planning based on a plurality of free space segments is used; and FIG. 3 is a block diagram illustrating the method steps of a method for determining a free space that can be traveled on without a collision.FIG. 1 shows, by way of example and schematically, a system 1 for determining a free space F which can be traveled by an ego vehicle 2. The system 1 is preferably provided in the ego vehicle 2. The system 1 comprises a computing unit 4. this computing unit 4 receives information from a sensor system 5 of the ego vehicle 2. the sensor system 5 is designed to detect the environment of the ego vehicle 2, so that the information provided by the sensor system 5 contains environmental information. The sensor system 5 can comprise, for example, one or more radar sensors, ultrasonic sensors, one or more lidar sensors and / or one or more cameras.The arithmetic unit 4 is configured to ascertain the free space F that the ego vehicle 2 can travel in a collision-free manner on the basis of the environmental information. The free space F can be defined, for example, by a left and a right boundary line. The left and right boundary line can be formed, for example, by a line segment connecting a plurality of support points.The free space F determined by the computer unit 4 can be forwarded to a trajectory planner 6, which, on the basis thereof, determines the ego trajectory traveled by the ego vehicle 2.FIG. 2 schematically illustrates a traffic situation in which the free space that can be traveled on by the ego vehicle 2 without collision can be used.In the exemplary embodiment shown, the ego vehicle 2 travels in the direction of travel FR on a roadway that has a plurality of lanes for this direction of travel. The ego vehicle 2 travels in the right lane, for example. An obstacle H, in the example shown a stationary vehicle, is located in this lane. This obstacle H must be traveled around by the ego vehicle 2 by the ego vehicle 2 performing a lane change. The obstacle H thus restricts the free space F which can be traveled on without collision.In addition, the free space F that can be traveled on without collision is influenced by a third-party vehicle 3 that is located on the left lane, moves at a higher speed, for example, and changes lanes to the middle lane.The free space F which can be traveled on by the ego vehicle 2 without collision is determined by defining a first free space which is defined on the basis of the expected movement of the ego vehicle in the prediction period, segmenting this first free space into a plurality of free space segments S 0-S 3 following one another in the direction of travel FR, and partially restricting the first free space in the respective free space segment S 0-S 3 if an obstacle H, a third vehicle 3, etc. is located in the free space segment S 0-S 3. In other words, the free space planning is not carried out for a total prediction period, which may be 0 s to 5 s, for example, but the first free space, which specifies the spatial region in which the ego vehicle 2 will move during the prediction period, is segmented. Based on these free space segments S 0-S 3, it is then checked whether an obstacle, a third vehicle or the like is located in the respective free space segment S 0-S 3 at the time at which the ego vehicle 2 is located in this free space segment S 0-S 3, in order to focus the free space planning on relevant regions and thereby limit the necessary computing resources.The method will be described in more detail below based on FIG. 2.First, the ego trajectory ET of the ego vehicle 2 is estimated in a prediction period. This ego trajectory ET is preferably a rough estimate that can result in information from past prediction periods. Likewise, a third-vehicle trajectory DT of a third-vehicle 3 is ascertained by ego vehicle 2, this third-vehicle 3 having been detected, for example, by sensor system 5 of ego vehicle 2. During the ascertainment of the third-vehicle trajectory DT, quality information can also be provided which is a measure of quality of the estimation of the third-vehicle trajectory DT. Thus, the quality of the third-vehicle trajectory DT can be higher, for example, when the third-vehicle 3 has already been detected by the sensor system 5 of the ego vehicle 2 for a relatively long period of time than when the third-vehicle 3 has only been detected recently.A first free space is then defined. This is preferably carried out on the basis of the estimated ego trajectory ET of the ego vehicle 2. the first free space can be designed, for example, in the form of a hose and have a length and a width, wherein the length extends in the direction of travel and at least substantially corresponds to the route which the ego vehicle 2 covers in the prediction period. The width of the first free space extends transversely, in particular perpendicularly, to the direction of travel FR of the ego vehicle 2. the first free space thus specifies a corridor in which the ego vehicle 2 can move fundamentally in the prediction period without taking into account objects which restrict the free space.Subsequently, based on the ego trajectory ET of the ego vehicle 2 and discrete times t0- t3in the prediction period, it is determined at which positions the ego vehicle 2 will be located in the future. This position determination can relate exclusively to the position of the ego vehicle 2 in the direction of travel FR, i.e. the transverse position of the ego vehicle 2 does not have to be taken into account for this purpose.FIG. 2 shows the positions of the ego vehicle 2 at the times t 0-t 3. Based on these positions, the first free space can be segmented, wherein the free space segments S 0-S 3 are formed by the segmentation. The free space segments S 0-S 3 extend in the direction of travel FR over a partial length of the first free space. Transversely to the direction of travel, these are limited by the boundaries of the first free space, as is indicated by the bold lines in FIG. 2. The free space segments S 0-S 3 are each assigned to a time t 0-t 3 in the prediction period, for example the free space segment S 0 is assigned to the time t 0, the free space segment S 1 is assigned to the time t 1, etc. Each free space segment S 0-S 3 is arranged centered around the position of the ego vehicle 2 at the respective time, as seen in the direction of travel FR. The free space segment S 0-S 3 has a length extending in the direction of travel FR that is greater than the length of the ego vehicle 2.Subsequently, one or more free space segments S 0-S 3 are restricted on the basis of the objects or obstacles H, which fall into the respective free space segments S 0-S 3 at the times t 0-t 3. In other words, an object or an obstacle H is used to restrict a free space segment S 0-S 3 only if this object or obstacle H likewise falls at least partially into this free space segment S 0-S 3 at the time t 0-t 3 at which the ego vehicle 2 is located in the respective free space segment S 0-S 3. A safety zone S can be formed around an object or obstacle H. This safety zone S can be larger than the base area of the object or obstacle H, for example, and thus form a safety buffer, in order to take account of position estimates of the object or obstacle H, which are subject to uncertainties, for example. In FIG. 2, the safety zone S around the obstacle H and the third-party vehicle 3 are indicated with a rectangular contour around the obstacle H and the third-party vehicle 3.As explained above, the future movement of the third vehicle 3 in the prediction period is described by a third vehicle trajectory DT. Based on this third vehicle trajectory DT, it is possible to determine at which position the third vehicle 3 is located at the points in time t 0-t 3.If the third vehicle 3, an obstacle H or another object, also penetrates into this free space segment S 0-S 3 or lies entirely therein at the time t 0-t 3 at which the ego vehicle 2 is located in the respective free space segment S 0-S 3 (for example, the obstacle H and the third vehicle 3 into the free space segment S 2 at the time t 2), this free space segment S 0-S 3 is correspondingly restricted or cut to size, such that the object or obstacle H, including its possibly present safety zone S, lies outside the respective free space segment S 0-S 3.As can be seen in FIG. 2, neither the obstacle H nor the third-party vehicle 3 is located in the free space segments S 0 and S 1. These thus remain unrestricted.At the time t 2, at which the ego vehicle 2 is located in the free space segment S 2, both the obstacle H and the third vehicle 3 are located in the free space segment S 2. Therefore, the boundaries of the free space segment S 2 which restrict the free space segment S 2 in its width are changed in such a way that the third vehicle 3 and the obstacle H, including their possibly present safety zone S, lie outside the free space segment S 2. This is indicated in FIG. 2 by the angular shape of the bold lateral boundaries of the free space segment S 2. The same applies in an analogous manner also to the free space segment S 3 and the time t 3.After the modification of one or more free space segments, a second free space is formed by assembling the free space segments S 0-S 3. This second free space forms the free space which can be traveled on without collision in the prediction period and which can be used by the trajectory planner 6 to calculate collision-free trajectories.The length of the free space segments S 0-S 3 measured in the direction of travel FR of the ego vehicle 2 can be selected as a function of the quality of the estimate of the own trajectory ET and / or as a function of the quality of the third-party trajectory DT. In particular, if the quality of the own trajectory ET and / or of the third-vehicle trajectory DT is low, the length of the free space segments S 0-S 3 can be selected to be greater than in the case of a high quality of the own trajectory ET and / or of the third-vehicle trajectory DT, in order to increase the safety in the determination of the free space which can be traveled on without collision.Alternatively or additionally, the length of the free space segments S 0-S 3 measured in the direction of travel FR of the ego vehicle 2 may depend on the speed of the ego vehicle 2. In particular, the length of the free space segments S 0-S 3 can be increased with increasing speed of the ego vehicle 2.As explained above, a safety zone S can be set up around a third-party vehicle 3, which is used in the restriction of the free space segments when the third-party vehicle 3 enters a free space segment S 0-S 3 with its safety zone S. The size of the safety zone S, in particular the distance of the edge of the safety zone from the outer line of the vehicle body of the third vehicle 3, can be determined as a function of the quality of the third vehicle trajectory DT. In particular, if the prediction of the third-vehicle trajectory DT is uncertain and thus the quality of the third-vehicle trajectory DT is low, the safety zone S can be selected to be greater than in the case of a high quality, in order to increase the safety of the determination of the free space to be traveled on without collision.FIG. 3 shows a schematic block diagram illustrating the steps of the method for determining a free space to be traveled by an ego vehicle.First, the ego trajectory of the ego vehicle and a third-vehicle trajectory of at least one third-vehicle is estimated for a prediction period (S 10).In addition, a first clearance is set (S 11).Subsequently, based on the estimated ego trajectory, the future positions of the ego vehicle are ascertained at different points in time in the prediction period (S 12).Based on the determined future positions of the ego vehicle, the first free space is divided into free space segments (S 13).Based on the third-vehicle trajectory, the future positions of the third-vehicle at the different times in the prediction period are estimated (S 14).Subsequently, in each case for the points in time in the prediction period, the future position of the third vehicle at the respective point in time is separately compared with the free space segment in which the ego vehicle is located at this point in time (S 15).At least one free space segment is then restricted or tailored if the estimated future position of the third vehicle at the respective point in time indicates an approach of the third vehicle to the respective free space segment in which the ego vehicle is located at this point in time, closer than a safety threshold value (the safety threshold value can be predefined by the safety zone, for example), or an at least partial intrusion of the third vehicle into this free space segment (S 16).Finally, a second clearance is formed by joining the clearance segments (S 17).The invention has been described above with reference to exemplary embodiments. It is to be understood that numerous changes and modifications are possible without thereby departing from the scope of protection defined by the claims.List of reference characters1 System 2 Ego vehicle 3 Third vehicle 4 Computing unit 5 Sensor system 6 Trajectory planner DT Third vehicle trajectory ET Ego trajectory F Free space H Obstacle S Safety zone S 0-S 3 Free space segment t 0-t 3 Time

Claims

Method for determining a free space (F) to be traveled by an ego vehicle (2), the method comprising the following steps: - estimating the ego trajectory (ET) of the ego vehicle (2) and a third-vehicle trajectory (DT) of at least one third vehicle (3) for a prediction period (S10); - defining a first free space (S11); - determining the future positions of the ego vehicle (2) at different times (t1 - t3) in the prediction period based on the estimated ego trajectory (ET) (S12); - dividing the first free space into free space segments (S0 - S3) based on the determined future positions of the ego vehicle (2) (S13); estimating the future positions of the third vehicle (3) at the different points in time in the prediction period based on the third vehicle trajectory (DT) (S 14); separately for the points in time (t0- t3) in the prediction period, respectively, comparing the future position of the third vehicle (3) at the respective point in time (t0- t3) with the free space segment (S0- S3) in which the ego vehicle (2) is located at this point in time (t0- t3) (S15); restricting at least one free space segment (S0 - S3) if the estimated future position of the third vehicle (3) at the respective time (t0 - t3) indicates (S16) an approach of the third vehicle (3) to the free space segment (S0 - S3) in which the ego vehicle (2) is located at this time closer than a safety threshold value or an at least partial penetration of the third vehicle (3) at the respective time (t0 - t3) into the free space segment (S0 - S3) in which the ego vehicle (2) is located at this time; forming a second free space by joining the free space segments (S0 - S3) (S17).Method according to Claim 1, characterized in that the first free space is of hose-like design and runs along the estimated ego trajectory (ET).Method according to Claim 1 or 2, characterized in that the free space segments (S0 - S3) each comprise the entire width of the first free space and each a partial length of the first free space.Method according to one of the preceding claims, characterized in that the free space segments (S0 - S3) are each formed centered with respect to the position which the ego vehicle (2) has at a discrete point in time (t0 - t3) in the prediction period.Method according to one of the preceding claims, characterized in that the restricting of at least one free space segment (S0 - S3) comprises an at least partial reduction in the width of the free space segment (S0 - S3).Method according to one of the preceding claims, characterized in that a safety zone (S) is formed around the third vehicle (3), and in that at least one free space segment (S0 - S3) is restricted if the estimated future position of the third vehicle (3) indicates an at least partial penetration of the safety zone (S) into the respective free space segment (S0 - S3).Method according to Claim 6, characterized in that the size of the safety zone (S) is selected as a function of the quality of the third-party vehicle trajectory (DT).Method according to one of the preceding claims, characterized in that the length of the free space segment (S0 - S3) is selected as a function of the quality of the ego trajectory (ET) and / or the quality of the third-party vehicle trajectory (DT).Method according to one of the preceding claims, characterized in that the length of the free space segment (S0 - S3) is selected as a function of the speed of the ego vehicle (2).System for determining a free space (F) which can be traveled by an ego vehicle (2), wherein the system (1) has a computing unit (4) which is configured to carry out the following steps: - estimating the ego trajectory of the ego vehicle (2) and a third-vehicle trajectory (DT) of at least one third-party vehicle (3) for a prediction period; - defining a first free space; - determining the future positions of the ego vehicle (2) at different times (t0 - t3) in the prediction period on the basis of the estimated ego trajectory (ET); - dividing the first free space into free space segments (S0 - S3) on the basis of the determined future positions of the ego vehicle (2); estimating the future positions of the third vehicle (3) at the different points in time (t0 - t3) in the prediction period based on the third vehicle trajectory (DT); separately for the points in time (t0 - t3) in the prediction period, respectively, comparing the future position of the third vehicle (3) at the respective point in time (t0 - t3) with the free space segment (S0 - S3) in which the ego vehicle (2) is located at this point in time (t0 - t3); restricting at least one free space segment (S0 - S3) if the estimated future position of the third vehicle (3) at the respective time (t0 - t3) indicates an approach of the third vehicle (3) to the free space segment (S0 - S3) in which the ego vehicle (2) is located at this time closer than a safety threshold value or an at least partial penetration of the third vehicle (3) at the respective time (t0 - t3) into the free space segment (S0 - S3) in which the ego vehicle (2) is located at this time; forming a second free space by joining the free space segments (S0 - S3).

Images