Method for determining a user-specific driving profile for automated driving of a vehicle

Abstract

Method for determining a user-specific driving profile (FP) for an automated drive of a vehicle (F1 to Fn), characterized in that: - several vehicle users (N1 to Nn) perform several automated drives, wherein the situation perception of the respective vehicle user (N1 to Nn) is determined location-related and, together with an associated driving parameter set ({kv, kd, kt,d, kt,v, ...}), which determines a driving behavior of the respective automated drive, is transmitted as user data (ND1 to NDn) to a vehicle-external server (1); - in the vehicle-external server (1) the driving parameter sets ({kv, kd, kt,d, kt,v, ...}) collected from all vehicle users (N1 to Nn) are stored.}) are clustered, whereby clusters are formed that represent different driving profiles (FP), - for each vehicle user (N1 to Nn) at least one driving profile (FP) preferred by the respective vehicle user (N1 to Nn) is identified by a statistical evaluation of their user data (ND1 to NDn), is assigned to the respective vehicle user (N1 to Nn) and is made available to the respective vehicle user (N1 to Nn) for retrieval.

Description

[0001] The invention relates to a method for determining a user-specific driving profile for automated driving of a vehicle.

[0002] As described in DE 10 2020 108 857 A1, a method for planning a target trajectory to be driven automatically by a vehicle is known from the prior art. The planning is based on determining a discrete set of candidates for the target trajectory and selecting one candidate from this set. The selection is based on predefined cost functions. If a change in the constraints to be observed and / or the driving tasks to be performed is detected, the selection is pre-controlled by adjusting the cost functions for individual trajectory segments of the candidates to the changed constraints and / or driving tasks. This allows lower costs to be assigned to trajectory segments that are better suited to complying with the changed constraints and / or performing the changed driving tasks than to other trajectory segments.

[0003] GB 2588639 A describes a method for automatically adapting a driving mode in a vehicle. The method includes receiving information related to the vehicle, a driver, and / or geographical details; determining the type of journey segment for an upcoming journey to navigate the vehicle based on this information; providing the information and the journey segment type as input to an initial trained learning model; and adapting the driving mode from one or more driving modes and values ​​of one or more vehicle parameters based on an output generated by the initial trained learning model.

[0004] The invention is based on the objective of providing a novel method for determining a user-specific driving profile for automated driving of a vehicle.

[0005] The problem is solved according to the invention by a method for determining a user-specific driving profile for automated driving of a vehicle with the features of claim 1.

[0006] Advantageous embodiments of the invention are the subject of the dependent claims.

[0007] In a method according to the invention for determining a user-specific driving profile for automated driving, in particular for highly automated or autonomous driving, several vehicle users perform several automated journeys. This can be done with the same vehicle or with several vehicles. The situational perception of each vehicle user is determined location-specifically and, together with an associated set of driving parameters that determine the driving behavior of the respective automated journey, is transmitted as user data to a server external to the vehicle, also referred to as a backend server or cloud. In particular, to determine the situational perception of each vehicle user location-specifically, the determined situational perception of each vehicle user is tagged with a location.On the vehicle-external server, the driving parameter sets collected from all vehicle users are clustered, forming clusters that represent different driving profiles. This clustering can also be referred to as classification, and the resulting clusters can accordingly be called classes. For each individual vehicle user, a statistical analysis of their user data identifies at least one driving profile preferred by that user—that is, one perceived as particularly pleasant—assigned to that user, and made available for retrieval.

[0008] The vehicle user is, for example, a driver, a passenger, or another vehicle occupant.

[0009] The solution according to the invention enables the automated creation of various driving profiles and their assignment to vehicle users. This eliminates the need for additional field studies with test subjects to determine different driving profiles.

[0010] Furthermore, the driving profiles created using the solution according to the invention are significantly more detailed than driving profiles that would be determined in field studies with test subjects, because with the solution according to the invention, data from an entire fleet of vehicle users of a respective manufacturer can be used, for example, to determine the driving profiles.

[0011] Furthermore, the solution according to the invention provides a larger amount of data for determining driving profiles than field studies. In such field studies, a preselection of data takes place to save costs. For example, critical situations are identified by speeds and braking decelerations. However, this results in a bias, i.e., a one-sidedness, because certain data are not evaluated. The vehicle user can determine this data much better because they are physically present and their perception of the situation is determined location-specifically.

[0012] The solution according to the invention also enables scalability, since the approach according to the invention can be easily applied automatically to many fleet data sets and different computing units.

[0013] The solution according to the invention also enables iterative improvement, since AI algorithms (AI = Artificial Intelligence) used to determine driving profiles are further trained with new data, i.e., new driving parameter set data and new data on location-related situation perception. This increases their performance and ultimately reduces errors to zero.

[0014] Furthermore, the solution according to the invention enables a reduction in costs, firstly due to the fact that field studies are no longer required and secondly due to the ever-decreasing development effort for complex algorithms as time progresses and the amount of data increases.

[0015] In one possible embodiment, several driving profiles preferred by the vehicle user are identified for at least one or more of the vehicle users, or for all vehicle users, through statistical analysis of their user data. These profiles are then assigned to the respective vehicle user and made available for retrieval, with a ranking of these preferred driving profiles being created according to their preference. This allows each vehicle user to select a desired driving profile from several that are generally suitable for them, one that corresponds to their current needs.For example, the respective vehicle user can then choose between a more comfortable and a more sporty driving profile, depending, for example, on whether he wants to reach a destination quickly or whether he has plenty of time to reach the destination and wants to perform other activities during the journey for which a comfortable driving profile is more suitable.

[0016] In particular, it is envisaged that for a subsequent automated drive, i.e., after at least one user-specific driving profile has been determined for the respective vehicle user, and following identification of the vehicle user, at least one driving profile preferred by the vehicle user, or the multiple driving profiles preferred by the vehicle user, will be retrieved from the vehicle-external server. The retrieved driving profile, or one of the retrieved driving profiles, will be selected and used as the basis for controlling the automated drive. The procedure then encompasses not only the determination of a user-specific driving profile for an automated drive of a vehicle, but also its use for carrying out the automated drive of the vehicle. It is therefore also a procedure for operating an automated driving vehicle.

[0017] In one possible embodiment of the method, the selection is automatic, whereby if only one driving profile is retrieved, that single driving profile is selected, and if multiple driving profiles are retrieved, the most preferred driving profile is advantageously selected. Alternatively, in another possible embodiment of the method, the selection of the retrieved driving profile or one of the multiple retrieved driving profiles is made by a selection action by the vehicle user, in particular a manual or, for example, acoustic one. An acoustic selection action is understood to mean, in particular, a voice command from the vehicle user. In this way, the vehicle's driving behavior is adapted to the driving profile selected automatically or by the vehicle user. When selecting, the vehicle user can choose the driving profile they currently desire from among several available driving profiles.When automatically selecting from several driving profiles, the driving profile currently preferred by the driver is highly likely to be chosen, as the most preferred driving profile of the vehicle user is automatically selected.

[0018] In one possible embodiment of the method, trajectory control is performed during each automated drive according to a planned trajectory, whereby the set of driving parameters that determines the driving behavior is used as the basis for trajectory planning during each automated drive. This applies in particular to the subsequent automated drive, i.e., after the at least one user-specific driving profile has been determined for the respective vehicle user, but advantageously also to the respective automated drive in which the situational perception of the respective vehicle user is determined location-specifically and, together with the associated set of driving parameters that determines the driving behavior of the respective automated drive, is transmitted as user data to the vehicle-external server.

[0019] In one possible embodiment of the method, clustering is location-specific for areas where unpleasant driving situations have been frequently reported. The use of a user-specific driving profile is particularly advantageous for such areas, as it helps to avoid the unpleasant driving situation experienced by the respective vehicle user. In other areas where no or only rare unpleasant driving situations have been reported, a universally applicable driving profile can be used, for example.

[0020] In one possible embodiment of the method, additional information about the situations present during the respective automated drive, such as rain, fog, snow, normal weather conditions, time of day, lane width, road classification, and / or other situations, is collected and transmitted to the vehicle-external server along with the user data. The clustering is then performed individually for the different situations. The at least one driving profile preferred by the respective vehicle user is then a situation-dependent preferred driving profile. This enables further improvement of the driving behavior through the use of a driving profile adapted to these specific situations, thus avoiding any unpleasant driving experiences resulting from currently unfavorable conditions.

[0021] In one possible embodiment of the method, the situational perception of the respective vehicle user is determined automatically, for example, by monitoring the vehicle user with a camera and / or by the vehicle user providing feedback via a user input, i.e., through voluntary feedback from the vehicle user. The user input can be manual or acoustic, i.e., voice input. Monitoring the vehicle user enables the determination of the situational perception automatically, without disturbing the vehicle user or distracting them from other activities. Furthermore, this method of determining the situational perception does not require active participation from the vehicle user.The alternative or additional determination of the situational perception through the situational perception statement of the respective vehicle user enables an even more precise determination of the situational perception based on an active statement about this by the vehicle user, thereby avoiding possible interpretation errors through vehicle user observation.

[0022] Exemplary embodiments of the invention are explained in more detail below with reference to drawings.

[0023] This shows: Fig. 1. Schematically, a procedure for determining a user-specific driving profile for automated driving. Fig. 2 schematic driving parameter sets and two driving profiles, and Fig. 3 schematic driving parameter sets and driving profiles for different situations.

[0024] Corresponding parts are marked with the same reference symbols in all figures.

[0025] Fig. Figure 1 shows a schematic representation of a method for determining a user-specific driving profile for automated driving, in particular for highly automated or autonomous driving, of a vehicle F1 to Fn. In this method, several vehicle users N1 to Nn perform several automated journeys. This can be done with the same vehicle F1 to Fn or, as in the example shown, with several vehicles F1 to Fn. The situational perception of each vehicle user N1 to Nn is determined location-specifically and, together with an associated set of driving parameters {kv,kd,kt,d,kt,v,...}, which determines the driving behavior of the respective automated journey, is transmitted as user data ND1 to NDn to a vehicle-external server 1, also referred to as a backend server.In particular, to determine the situational perception of the respective vehicle user N1 to Nn in relation to their location, the determined situational perception of the respective vehicle user N1 to Nn is stamped with a location. The situational perception of the respective vehicle user N1 to Nn is determined, for example, automatically by monitoring the vehicle user with a camera, especially with an interior camera in a passenger compartment of the respective vehicle F1 to Fn, and / or by providing feedback from the respective vehicle user N1 to Nn via a user input, i.e., via feedback from the respective vehicle user N1 to Nn, which is voluntary. The user input can be, for example, manual or acoustic.

[0026] In the vehicle-external server 1, the driving parameter sets {kv,,kt,d,kt, v,...} collected from all vehicle users N1 to Nn are clustered, forming clusters that represent different driving profiles FP. This clustering is location-specific, for example, for areas where a high number of unpleasant driving situations have been reported.

[0027] For each individual vehicle user N1 to Nn, at least one driving profile FP preferred by the respective vehicle user N1 to Nn, i.e., perceived as particularly pleasant, is identified through a statistical evaluation of their user data ND1 to NDn, assigned to the respective vehicle user N1 to Nn and made available to the respective vehicle user N1 to Nn for retrieval.

[0028] If several preferred driving profiles FP are identified for the respective vehicle user N1 to Nn, it is advantageous to also create a ranking of these preferred driving profiles FP according to preference.

[0029] For a subsequent automated drive, i.e., after at least one user-specific driving profile FP has been determined for the respective vehicle user N1 to Nn, the at least one driving profile FP preferred by the vehicle user N1 to Nn, or the multiple driving profiles FP preferred by the vehicle user N1 to Nn, are retrieved from the vehicle-external server 1 after identification of the vehicle user N1 to Nn. The retrieved driving profile FP, or one of the retrieved driving profiles FP, is selected and used as the basis for controlling the automated drive. The procedure thus includes not only the determination of a user-specific driving profile FP for an automated drive of a vehicle F1 to Fn, but also its use for carrying out the automated drive of the vehicle F1 to Fn. It is therefore also a procedure for operating an automated vehicle F1 to Fn.

[0030] The selection of the driving profile FP can, for example, be automatic. If only one driving profile FP is requested, that single driving profile FP is selected, and if multiple driving profiles FP are requested, the most preferred driving profile FP is advantageously selected. Alternatively, the selection of the requested driving profile FP, or one of the multiple requested driving profiles FP, can be made by a selection action by the vehicle user N1 to Nn.

[0031] In each automated journey, trajectory control is carried out according to a planned trajectory, whereby the driving parameter set {kv,,kt,d,kt,v,...}, which determines the driving behavior, is used as the basis for trajectory planning in each automated journey.

[0032] In one possible embodiment of the method, additional information about the situations present during the respective automated drive, such as rain, fog, snow, normal weather conditions, time of day, lane width, road classification, and / or other situations, is determined and transmitted together with the user data ND1 to NDn to the vehicle-external server 1, whereby the clustering is then carried out individually for the different situations. The at least one driving profile FP preferred by the respective vehicle user N1 to Nn is then a situation-dependent preferred driving profile FP.

[0033] In summary, vehicles F1 to Fn drive in an automated, in particular highly automated or autonomous, manner, collecting various data which are then transmitted to the vehicle-external server 1 and processed there. Vehicles F1 to Fn then receive user-specific data from the vehicle-external server 1, according to their vehicle user N1 to Nn, which is used to influence the driving behavior of the respective vehicle F1 to Fn.

[0034] Further details of the procedure are described below.

[0035] In particular, during automated, especially highly automated or autonomous, driving, each vehicle F1 to Fn continuously performs trajectory planning to execute the respective automated journey. The result of the trajectory planning is an optimal target trajectory. This optimal target trajectory forms the basis for trajectory control. For example, it specifies not only a desired path that the respective vehicle F1 to Fn should follow, but also a desired dynamic, in particular speed and acceleration, with which this path should be traversed.

[0036] Trajectory planning is based, for example, on the current driving state of the respective vehicle F1 to Fn, in particular with regard to speed and acceleration; on target requirements regarding comfort and safety, such as a desired speed, desired limits of longitudinal and lateral acceleration, desired distances to obstacles and desired distances to a lane center; on weighting values ​​that indicate which of the target requirements should be prioritized over the other target requirements and to what extent; and on a recorded environmental situation, such as the geographical vehicle position, lane layout, lane position of the vehicle F1 to Fn, orientation of the vehicle F1 to Fn relative to the lane, obstacles in the vehicle environment, and available driving space.

[0037] The respective vehicle F1 to Fn includes, for example, means for recording the situational awareness of the vehicle user N1 to Nn. This determines whether the vehicle user N1 to Nn perceives the current situation as more pleasant or more unpleasant. These means include, for example, the camera, in particular the interior camera, for monitoring the vehicle user, such as observing the eyes and gaze direction of the vehicle user N1 to Nn and any secondary activities the vehicle user N1 to Nn may be engaged in. For example, wide-open eyes focused on the road or gripping the steering wheel are indicators of a situation perceived as unpleasant, while a detected distraction of the vehicle user N1 to Nn, such as closed eyes, a gaze wandering from the road, or secondary activities, such as using a smartphone, are indicators of a situation perceived as pleasant.Alternatively or additionally, the means for recording the situational perception can also include a control unit via which the vehicle user N1 to Nn directly communicates his situational perception via an operating input.

[0038] Each vehicle, F1 to Fn, continuously transmits its perceived situation and situation data describing the current situation to the external server 1. This data is the user data ND1 to NDn. The situation data includes the parameters used for trajectory planning, i.e., a respective set of driving parameters {kv,,kt,d,kt, v,...}.

[0039] The vehicle-external server 1 thus receives from a large number of vehicles F1 to Fn the situation data of the respective vehicle F1 to Fn and the situation perception of the respective vehicle user N1 to Nn in the form of the respective user data ND1 to NDn.

[0040] In the example shown, the received user data ND1 to NDn are fed into a first block B1 on the vehicle-external server 1, where they are clustered. Each of the received user data ND1 to NDn is then assigned to one or more clusters.

[0041] In the example shown, the clustered data CDs created in the first block B1 are fed to a second block B2 and a third block B3 in the vehicle-external server 1.

[0042] In the second block B2, several driving profiles FP are created from the clustered data CD.

[0043] The generated driving profiles FP are fed into the third block B3.

[0044] In the third block, B3, the driving profiles FP and the clustered data CD are used to assign the driving profiles FP to the vehicle users N1 to Nn, and thus, in the example shown, also to the vehicles F1 to Fn, since in the example each vehicle user N1 to Nn only uses their vehicle F1 to Fn. One or more of the driving profiles FP can be assigned to each vehicle user N1 to Nn, and thus, in the example, to each vehicle F1 to Fn.

[0045] The driving profiles FP assigned to the vehicle users N1 to Nn, and thus in the example shown to the vehicles F1 to Fn, are allocated to the respective vehicle user N1 to Nn, more precisely to the vehicle F1 to Fn used by the respective vehicle user N1 to Nn.

[0046] If a vehicle F1 to Fn is assigned only one of the driving profiles FP, then this single driving profile FP is selected automatically. Otherwise, if the vehicle F1 to Fn is assigned multiple driving profiles FP, these are offered to the vehicle user N1 to Nn for selection, or the selection is made automatically.

[0047] The selected driving profile FP is used in the respective vehicle F1 to Fn to improve, for example, the determination of an optimal target trajectory.

[0048] The number of clusters, and therefore driving profiles (FP), is fixed to a predetermined value. For example, four driving profiles (FP) are distinguished. One driving profile (FP) is a sleep driving profile, in which the vehicle user (N1 to Nn) can relax or sleep. This sleep driving profile is characterized, for example, by predetermined minimum vehicle accelerations (F1 to Fn). Another driving profile (FP) is a defensive driving profile (FP), which is characterized in particular by the fact that few overtaking maneuvers are performed, especially compared to a normal driving profile (FP) and a time-optimized driving profile (FP). The normal driving profile (FP) is characterized, for example, by average driving behavior.The time-optimized driving profile FP is characterized, for example, by the fact that many overtaking maneuvers are carried out, especially in comparison to the defensive and normal driving profile FP, and that the maximum possible speed is driven.

[0049] Alternatively, it could be planned, for example, that a meaningful number of clusters, and thus a corresponding number of driving profiles FP, are calculated from the user data ND1 to NDn transmitted to the vehicle-external server 1. The calculation of the number of clusters can be performed, for example, using the elbow method of the Within-Cluster Sum of Squared Errors (WSS) of various k in k-means, as described, for example, in https: / / medium.com / analytics-vidhya / how-to-determine-the-optimal-k-fork-means-708505d204eb in conjunction with https: / / de.wikipedia.org / wiki / K-Means-Algorithmus. The unknown / variable driving parameters k serve as the basis for the clustering. v , kd , k t,d , k t,v , ... of the driving parameter sets {k v , k d , k t,d , k t,v , ...} selected, which are usually used as a quality criterion for optimizing a trajectory or selecting a trajectory from a family of trajectories, as described and cited below in formulas (3.8) on p. 40 and (3.18) on p. 43 in conjunction with formula (3.4a) on p. 37 in Schucker, Jeremias, “Trajektorienplanung und Fahrzeugführung für hochautomatisiertes Fahren auf der Autobahn.” (2020), https: / / tuprints.ulb.tu-darmstadt.de / 12686 / 1 / 2020-07-28_Schucker_Jeremias.pdf:

[0050] "By planning a target speed, the final costs of (3.4a) also change, and the following cost function is obtained for calculating the family of trajectories: Jv(Fx,s)=kte,vte,v+kv(v(te,v)−vtarget(te,v))2+12∫t0te,vFx,s2(t)dt.

[0051] With k te,v and k vThe final costs are weighted and t e,v is the length of the maneuver." Jd(δS)=kte,dte,d+kd(d(te,d)−dtarget)2+12∫t0te,dδS2(t)dt.

[0052] This includes k te,d and k d the weighting factors and t e,d the maneuver digesters." (End quote)

[0053] The quality functions J v , J d are optimized to find an ideal trajectory. This is done using weighting factors {k v , k d , k t , k t,v , ...} determines how, for example, the duration of the maneuver and the acceleration are to be weighted. A defensive driver, for instance, accepts a longer duration for an overtaking maneuver in order to drive with as little lateral acceleration as possible. Thus, the ratio k d / k t , smaller for him than for an aggressive vehicle user.

[0054] For example, a respective set of driving parameters {k} can be retrieved from the vehicle-external server 1. v , k d , k t,d , k t,v , ...} are transmitted to various vehicle users N1 to Nn, more precisely to their respective vehicles F1 to Fn, and are evaluated by them, for example, by means of direct user feedback, i.e., by indicating their perception of the situation, for example via a display query, or by means of an indirect determination of their perception of the situation, in particular through vehicle user observation. Alternatively or additionally, it may be provided, for example, that each vehicle user N1 to Nn can select one or more parameter values ​​from the driving parameter set {k v , k d , k t, k t,v , ...}, for example, a minimum distance to a vehicle in front, is set automatically.

[0055] Fig. Figure 2 schematically shows an example with different sets of driving parameters represented here as points {k v , k d , k t,d , k r,v , ...}, which are transmitted from the vehicle-external server 1 to various vehicle users N1 to Nn, more precisely to their respective vehicles F1 to Fn, and thus tested by them. In this process, the vehicle users N1 to Nn are observed or can, as described, actively provide feedback to evaluate a set of driving parameters {k v , k d , k t, k t,v , ...} to evaluate as pleasant or unpleasant. The driving parameter sets {k v , k d , k t , k t,v, ...} different vehicle users N1 to Nn are collected and clustered using k-means and WSS. This results, for example, in two driving profiles FP: a very defensive driving profile, shown on the left, which is perceived as pleasant by the first vehicle user N1, and a dynamic driving profile FP on the right, which is perceived as unpleasant by the first vehicle user N1, but is preferred, for example, by a second vehicle user N2.

[0056] The respective vehicle user N1 to Nn can directly access the driving parameter sets {k} via a query, especially after autonomous driving, or via vehicle user monitoring, which, for example, detects that they are asleep or tense or have their hands close to the steering wheel, or directly in vehicle settings. v , k d , k t,d , k t,v...} evaluate and, for example, adjust. Driver takeover by the vehicle user N1 to Nn, for example by braking, accelerating, and / or steering, can also be included in the evaluation. If, for example, the autonomous driving system of the vehicle F1 to Fn does not merge into a gap during an overtaking maneuver and the vehicle user N1 to Nn therefore ultimately takes over the steering and merges manually, the minimum distance to the vehicle in front can be minimized via the driving parameter kd in the performance function.

[0057] The selection of a specific driving profile FP can be made by the vehicle user N1 to Nn themselves. For example, they can choose a defensive driving profile FP for night driving and a fast driving profile FP for their commute to work to arrive at the workplace as quickly as possible.

[0058] For example, parameter tuning for trajectory evaluation can be implemented by subtly modifying the set driving parameters kv, kd, kt, d, kt, v,... as in an optimization function. Active or passive feedback, as described above, determines whether this parameter change was beneficial or detrimental, i.e., whether it was perceived as pleasant or unpleasant by the respective vehicle user N1 to Nn. Another possible embodiment could include a selection of driving parameter sets based on time of day, weather, or road conditions. For instance, a more defensive driving profile FP could be automatically selected in snow / fog / rain, at night, and / or on rural roads. However, these situations are perceived differently as "dangerous," so these driving profiles FP are also individually assigned to the respective vehicle user N1 to Nn.

[0059] Fig.Figure 3 compares the change in perceived situation in fog, rain, or snow in the left diagram and in a normal situation in the right diagram. This results in different driving profiles FP, which are perceived as comfortable by the vehicle user F1 to Fn. Thus, the driving profile FP in the left diagram, i.e., in fog, rain, or snow, is positioned further to the left and is therefore significantly more defensive than the driving profile FP in the right diagram in a normal situation, i.e., the driving parameters k v , k d , k t,d , k t,v , ... of the driving profile FP in the left diagram, in fog, rain or snow, show smaller values ​​than in the right diagram in normal conditions.

[0060] Vehicle user N1 to Nn finds a more defensive driving profile FP, where, for example, the driving parameters kv and kd are smaller, more comfortable in fog, rain, or snow. They prefer to drive at a lower speed and with a greater distance to the vehicle in front, accepting a longer travel and maneuvering time as a consequence. Another vehicle user N1 to Nn, who, for example, lives in a ski resort, feels safer and would prefer a different set of driving parameters FP with the same kv and kd. The ideal set of driving parameters FP can be determined, for example, using a k-means clustering algorithm. Other classification algorithms are also possible, such as Decision Tree, SVM, Random Forest, or similar algorithms.

[0061] In the solution described here, driving profiles FP are determined without test drivers and with the help of feedback from vehicle users N1 to Nn. This results in an individual driving profile FP with a corresponding individual set of driving parameters {k} for each vehicle user N1 to Nn. v , k d , k t ,, k t,v , ...}, which can also be adapted to the environmental conditions. This results in greater acceptance and increased comfort. QUOTES INCLUDED IN THE DESCRIPTION

[0000] This list of documents cited by the applicant was automatically generated and is included solely for the reader's convenience. The list is not part of the German patent or utility model application. The DPMA accepts no liability for any errors or omissions. Cited patent literature

[0000] DE 102020108857 A1

[0002] GB 2588639 A

[0003]

Claims

[1] Method for determining a user-specific driving profile (FP) for automated driving of a vehicle (F1 to Fn), characterized by , that - several vehicle users (N1 to Nn) perform several automated journeys, whereby a situational perception of the respective vehicle user (N1 to Nn) is determined location-specifically and together with an associated driving parameter set ({k v , k d , k t,d , k t,v , ...}), which determines the driving behavior of the respective automated journey, as user data (ND1 to NDn) is transmitted to a vehicle-external server (1), - in the vehicle-external server (1) the driving parameter sets collected from all vehicle users (N1 to Nn) ({k v , k d , k t,d , k t,v , ...}) are clustered, forming clusters that represent different driving profiles (FP), - for each vehicle user (N1 to Nn) at least one preferred driving profile (FP) is identified by a statistical evaluation of their user data (ND1 to NDn), assigned to the respective vehicle user (N1 to Nn) and made available to the respective vehicle user (N1 to Nn) for retrieval. [2] Method according to claim 1, characterized by , that in order to determine the situational perception of the respective vehicle user (N1 to Nn) in relation to the location, the determined situational perception of the respective vehicle user (N1 to Nn) is stamped with a location stamp. [3] Method according to any one of the preceding claims, characterized by , that the driving profile (FP) preferred by the respective vehicle user (N1 to Nn) is identified as a driving profile (FP) that is perceived as pleasant by the respective vehicle user (N1 to Nn). [4] Method according to any one of the preceding claims, characterized by, that for at least one of the vehicle users (N1 to Nn) several driving profiles (FP) preferred by the vehicle user (N1 to Nn) are identified through the statistical evaluation of his user data (ND1 to NDn), are assigned to the vehicle user (N1 to Nn) and are made available to the vehicle user (N1 to Nn) for retrieval, whereby a ranking of these preferred driving profiles (FP) is created according to the preference. [5] Method according to any one of the preceding claims, characterized by, that for a subsequent automated journey after identification of the vehicle user (N1 to Nn) the at least one driving profile (FP) preferred by the vehicle user (N1 to Nn) or the several driving profiles (FP) preferred by the vehicle user (N1 to Nn) is / are retrieved from the vehicle-external server (1), wherein the retrieved driving profile (FP) or one of the retrieved several driving profiles (FP) is selected and used as the basis for controlling the automated journey. [6] Method according to claim 5, characterized by , that the selection is made automatically, whereby if several driving profiles (FP) are retrieved, the most preferred driving profile (FP) is selected, or that the selection is made via a selection action by the vehicle user (N1 to Nn). [7] Method according to any one of the preceding claims, characterized by, that during the respective automated journey, trajectory control is carried out according to a planned trajectory, whereby the driving parameter set ({k v , k d , k t,d , k t,v , ...}), which determines the driving behavior, is used as the basis for trajectory planning in the respective automated drive. [8] Method according to any one of the preceding claims, characterized by , that the clustering is location-specific for areas where a high number of unpleasant situations have been reported. [9] Method according to any of the preceding claims, characterized by, that additional information about situations present during the respective automated driving is determined and transmitted together with the user data (ND1 to NDn) to the vehicle-external server (1), wherein the clustering is carried out individually for the different situations and wherein the at least one driving profile (FP) preferred by the respective vehicle user (N1 to Nn) is a situation-dependent preferred driving profile (FP). [10] Method according to any one of the preceding claims, characterized by , that the situation perception of the respective vehicle user (N1 to Nn) is automatically determined by vehicle user observation with a camera and / or by a situation perception indication of the respective vehicle user (N1 to Nn) via an operating input.

Images