Procedures for determining driver personalities

Abstract

Method (200), in particular a computer-implemented method, designed and set up for parameterizing driving functions based on personality components of drivers with different driver personalities, wherein the method comprises the steps: - one determination (210) of at least two personality components of different personality types for at least two drivers; - of determining (220) a value of at least one driving parameter of at least two drivers; and - a determination (230) of at least one parameterization depending on the personality components of the personality types and the expression of the at least one driving parameter of the at least two drivers; wherein the method is characterized in that the at least one parameterization is carried out by means of at least one inverse method on the personality components of the different personality types of the at least two drivers.

Description

[0001] The invention relates to a method designed and configured for parameterizing driving functions based on personality traits of drivers with different driving personalities. The invention relates to a method designed and configured to assign personality traits to drivers. The invention relates to a computer program product. The invention relates to a storage medium and a control unit. The invention relates to an electronic signal. The invention relates to a motor vehicle.

[0002] It is known from the state of the art to assign a driving profile to an individual driver in order to fulfill the driver's individual wishes, so that he can, for example, drive in a sporty manner or drive in an ECO driving mode to save fuel.

[0003] German patent DE 10 2018 202 146 A1 describes a driver assistance system that automatically uses situation-specific driving profiles to precisely fulfill or anticipate the driver's wishes. To do this, the system determines a candidate for the driving profile to be used based on various data sources. A selection is then made from these candidates according to a fixed ranking of the driving profiles.

[0004] German patent DE 10 2021 003 073 B3 describes a method for automatically selecting a driving mode. Here, too, a specific driver or driver type is first identified. Subsequently, a corresponding driver profile can be loaded, which influences the driving mode. The driving mode is then selected taking into account the driver profile and additional driving information.

[0005] DE 10 2010 014 076 A1 describes a method for adapting the driving behavior of a vehicle when the driver changes.

[0006] DE 10 2014 215 258 A1 describes a method and a device for automatically selecting driving modes.

[0007] DE 10 2021 207 781 A1 describes a method, a control device and a system for adapting an assisted or automated driving function of a vehicle.

[0008] DE 10 2018 130 622 A1 describes systems and procedures for adapting assistance systems.

[0009] DE 10 2019 118 184 A1 describes a system and a method for user-specific adaptation of vehicle parameters.

[0010] The applicant adheres to the safety standards required and applicable on the application date and priority date to provide the best possible protection for both drivers and other road users, and to ensure the highest possible level of road safety. However, the applicant is committed to continuously improving safety. Therefore, the aim is to further enhance safety for both the driver and other road users compared to the described procedures and driver assistance systems. The applicant also seeks to improve the ability to save energy during vehicle operation, for example, to increase range and further reduce CO2 emissions. This will also enhance the driving experience for drivers.

[0011] The object of the present invention is to increase safety, for example for a driver but also for other road users. It is the object of the invention to reduce energy consumption. It is the object of the invention to reduce CO2 emissions. Furthermore, it is the object of the invention to improve the driving experience.

[0012] The foregoing problem is solved by a method having the features of claim 1 and by a method having the features of claim 6. The problem is solved by a computer program product having the features of claim 11 and a storage medium and a control device having the features of claim 12. The problem is solved by an electronic signal having the features of claim 13. The problem is solved by a motor vehicle having the features of claim 14. Further features and details of the invention will become apparent from the dependent claims, the description, and the drawings. Features and details described in connection with the methods according to the invention naturally also apply in connection with the computer program product, the storage medium, the control device, the electronic signal, and the motor vehicle according to the invention.This also applies in reverse, so that with regard to the disclosure of the individual aspects of the invention, mutual reference is always made or can be made.

[0013] According to one aspect, the problem is solved by a method with the features of claim 1.

[0014] A method, in particular a computer-implemented method, is designed and configured for parameterizing driving functions based on the personality traits of drivers with different driving personalities. The method includes the step of determining at least two personality traits of different personality types for at least two drivers. The method includes the step of determining the level of at least one driving parameter for the at least two drivers. The method includes the step of determining at least one parameterization based on the personality traits of the personality types and the level of the at least one driving parameter for the at least two drivers. This at least one parameterization is performed using at least one inverse method on the personality traits of the different personality types in the at least two drivers.

[0015] The method can, in particular, be a computer-implemented method. This means that individual steps of the method are executed on a machine, especially a computing machine, i.e., a computer and / or a control unit. At least one step is executed on such a machine.

[0016] The procedure is designed and configured to perform parameterization of driving functions. These driving functions can be, in particular, automated driving functions. These are driving functions that are performed and / or adapted by a machine, such as the control unit of a motor vehicle, to adjust the vehicle's driving behavior. This can include, in particular, a driver assistance system whose operation can be adapted. In this context, parameterization means, in particular, that certain parameters have been determined which can be adjusted during driving and which can be set via driving functions by the machine, especially the control unit of the motor vehicle.

[0017] Personality traits can be assigned to drivers, and these traits can originate from different personality types. These personality types can be archetypes, which rarely or never exist in their purest form, but represent the basic driver types into which all drivers, or at least a majority of them, can be categorized according to their personality traits. Thus, each driver can be assigned a unique driver personality, which represents a blend of these archetypes. The following example can be used for clarification, drawing a mathematical analogy: The personality types, as archetypes, establish independent coordinates in a function space. Real drivers can then be positioned within this coordinate system.The aim is therefore to assign personality traits to drivers with different personalities in order to perform parameterization of as many drivers as possible, such as a fleet of vehicles. It is important that not every single driver has experienced all possible driving situations.

[0018] The procedure may include the step of determining at least two personality traits of different personality types for at least two drivers. The drivers may be selected from a majority of drivers. The majority of drivers may correspond to a set of drivers (in the mathematical sense).

[0019] In some embodiments, a subset of drivers may also be selected. In this context, selection should not be interpreted as requiring drivers to meet a specific set of criteria. The drivers can therefore be randomly selected. The number of drivers (greater than or equal to two) can also be random. This can be based, in particular, on which drivers are the first to provide the relevant information to determine the personality traits of the different personality types.

[0020] In embodiments, it may be provided that the drivers are compiled according to a catalog of criteria, for example in a specific or to be determined composition of personality traits or a specific or to be determined coverage of all personality types, also in approximately a specific distribution among individual drivers in the personality traits.

[0021] In some implementations, additional, abstracted information about each person's driving personality can be available. This information can, in particular, represent personality traits of personality types, allowing a driver to be assigned a driving personality as a mixture of archetypes. This makes it possible, in particular, to move beyond using only pre-defined driver profiles based on archetypes (in the form of personality types) and instead incorporate new, mixed types into the process. As described elsewhere, this can improve safety, reduce energy consumption during journeys, lower CO2 emissions, and enhance the driving experience.

[0022] The procedure can include the step of determining the value of at least one driving parameter for at least two drivers. As already explained, these drivers can be selected from the majority of drivers. The explanations regarding selection can also be applied accordingly here. Alternatively, a subset of drivers can be randomly selected, particularly if they are the first to provide information and data for the procedure during at least one trip. To obtain a sufficiently large dataset, records of trips taken by drivers from this subset can be provided, for example, via vehicle bus data.

[0023] In embodiments, the recordings (e.g., vehicle bus data) can contain information about driving behavior, particularly with regard to at least one driving parameter, from many journeys by different drivers.

[0024] In some embodiments, at least two drivers can be selected; in particular, a subset of drivers can be selected from the majority of drivers. The drivers can have completed at least one drive (per driver), as described elsewhere. In some embodiments, this can be a standardized test drive.

[0025] In some embodiments, these journeys may be random, particularly in real-world driving situations. Random journeys are understood here to mean that the journeys are not deliberately undertaken to feed the system with information about the driving behavior of one or more drivers. In other words, specific routes or traffic situations are not deliberately selected to gather this information. Rather, in some embodiments, information about the drivers' driving behavior, represented in particular by at least one driving parameter, can be fed into the system in everyday situations.

[0026] The expression of a driving parameter can be, in particular, the strength, type, or rate with which a corresponding driving parameter is executed by a driver during at least one journey. In other words, one could also speak of a characteristic with which a parameter is implemented by one or more drivers during at least one journey.

[0027] The procedure can include the step of determining at least one parameterization based on the proportions of personality types and the expression of at least one driving parameter. The explanations regarding the drivers and the expressions are as described elsewhere. Through parameterization, at least one expression of a driving parameter can be assigned to each personality type. Conversely, a driver with personality traits of these personality types—that is, a mixed type of archetypes—can be assigned at least one expression for a driving parameter according to a specific individual or group-based distribution.Based on this, adjustments can be made to vehicle electronics, control systems, brake and / or throttle response, injection systems, battery management, traction machine control, driver assistance systems and / or other settings.

[0028] In summary, and to put it another way: The driving styles (represented by a set of driving parameters) of many drivers can differ to varying degrees. These different driving styles can be taken into account when parameterizing driving functions, especially automated driving functions, using the method described. This goes beyond approaches where parameters for individual drivers can be generated using inverse methods. Either the driver's free driving can be analyzed, or their interventions in existing automated / assisted systems. However, these methods usually require very large datasets. In cases where the training data does not include all scenarios the driver might encounter later, problems with parameterization can arise in such unknown situations.Therefore, significantly more data is required. This amount of data cannot be provided under all circumstances. This can lead to safety issues if each driver is assigned completely individual parameters. Methods like the one described here allow for the maintenance of safe parameter ranges. This makes it possible to increase safety when operating motor vehicles with driving functions, especially safe driving functions. Furthermore, the identified safe parameter ranges may allow for the training and application of driver-specific driving functions, enabling appropriate adjustments to at least one driving function. This also makes it possible to determine the vehicle's driving behavior in relation to a mixed personality type, even for previously unknown drivers.The described parameterization allows, for example, the adjustment of energy consumption during a journey or the improvement of safety during a journey. The method described here can form a basis for the method described elsewhere, in which the generated parameterization can be used to enable the classification, grouping, clustering, and / or assignment of unknown drivers into mixed personality types based on personality traits, particularly based on their driving style (i.e., driving behavior).

[0029] One approach is to determine the at least two personality traits of different personality types based on a selection from a personality test, a self-identification, or a driving behavior analysis. This allows drivers to be assigned the personality traits of the various personality types. For each driver, additional abstracted information about their "driving personality" can be available. The driving personality corresponds to the personality traits of the personality types already described. The driving behavior of each driver in a larger fleet can thus be analyzed and classified. Subsequently, similar sets of parameters can be used for similar driver types (corresponding to a similarity in the described personality traits of the personality types).

[0030] A personality test can be a standardized psychological test, such as the MDSI. The Multi-dimensional Driving Style Inventory (MDSI) is a personality test that assigns individual drivers a distribution of eight different personality types based on a questionnaire: reckless, anxious, risky, angry, high-velocity, distress-reducing, patient, and careful. For example, a driver might be 10% reckless, 50% angry, and 40% anxious. This also allows for the identification of archetypes.These profiles would correspond to approximately 100% of one personality type, but would exhibit no personality traits in the other personality types. Therefore, driving profiles that focus solely on sportiness or ECO mode do not reflect real drivers. Rather, such profiles deviate significantly from those of real drivers, at least for a portion of the driver population. Adjusting driving parameters to the mixed types described here allows for a closer approximation of real drivers' behavior, enabling the setting of driving parameters for a journey to be aligned with the current and / or ongoing needs of real drivers. This allows for particularly energy-efficient driving in one situation, while a more sporty program can be used in another.This approach is not limited to individual driving situations but can also be applied to the driver as a whole, since they can be assigned a corresponding personality profile that best represents them in most situations. Alternatively or additionally, from a sample of drivers, particularly as a subset of a larger group of drivers, such as a subpopulation of a fleet, at least one, and possibly several, mixed-type profiles can be created. This allows new drivers to be assigned a mixed type without having to study them in detail. Alternatively or additionally, it may also be possible that not every driver needs to have experienced every conceivable situation, which can pose a challenge with purely individualized profiles.

[0031] Alternatively or in addition to a personality test, drivers can also assign personality traits to themselves. They can perform a self-identification. However, this can lead to misjudgments or wishful thinking on the part of the driver(s) and may contradict their actual driving personality. Self-attribution can be carried out, in particular, using a vehicle's on-board computer, especially before starting a journey. A personality test, especially a standardized one, offers a certain degree of basic objectivity. However, it is potentially more complex to administer.

[0032] Alternatively or in addition to the personality test and / or self-identification, a driving behavior analysis can also be performed. This can be based on prior parameterization and thus includes, in particular, so-called legacy data. As already explained above, this allows the corresponding procedure to be carried out repetitively and / or iteratively in order to feed further datasets into an algorithm.

[0033] Based on one aspect, at least one driving parameter can be selected from a description of the driving's curvature, sinusoidal nature, or the activation patterns of the accelerator, brake, and / or steering inputs. This allows driving behavior to be parameterized, enabling the assignment of personality traits to a driving behavior, as described above. Conversely, corresponding driving styles and behaviors can also be assigned to a driving behavior, which can then be subdivided (and sorted) into corresponding subcategories and sub-behaviors.

[0034] Certain metrics of individual drivers' trajectories, such as descriptions of curvature, sinusoidal shape, and the activation patterns of accelerator, brake, and / or steering inputs, can correlate with the individual personality types (also referred to as personality attributes) of the MDSI. Thus, based on a sufficiently large dataset, drivers can be categorized into MDSI personality types solely based on their behavior in the vehicle (all of which can be read via bus data). However, this approach no longer maps archetypes to driver profiles; instead, it primarily identifies and maps mixed profiles.

[0035] Curvature can refer specifically to the curvature of a vehicle when cornering. However, it can also describe a deviation from a straight line on a straight stretch of road, for example, when the driver meanders around a center line and deviates from it. The latter can also be described as a sinusoidal shape. The intensity, frequency, and / or duration of throttle, brake, and / or steering inputs can vary between drivers, thus characterizing individual drivers.

[0036] Time series data, such as vehicle bus signals, trajectories and driver inputs, can also be transmitted to an algorithm that allows parameterization to be performed, as explained elsewhere.

[0037] In other words, and to summarize, information about a driver's journey, or about the journeys of different drivers, can be fed into an algorithm. This algorithm is particularly capable of outputting a description of the driver's personality traits. As explained in detail elsewhere, a driver's personality can correspond to a mixed type of personality traits from different personality types.

[0038] Based on one aspect, at least one parameterization can be performed using an inverse method based on the personality traits of the different personality types in at least two drivers. This allows for a direct parameterization to represent mixed types with varying degrees of expression of the personality traits.

[0039] The inverse method is a group of procedures that allow, particularly through the use of an algorithm, inferences to be drawn about a given situation. Inverse methods can be used to identify the personality traits of individual drivers, specifically in terms of personality types.

[0040] As an example of an inverse method, "Inverse Optimal Control" can be used. Inverse Optimal Control deals with solving and calculating the (unknown) parameters of an objective function in an optimal control problem such that given state and control trajectories are optimal. In other words, this method can be used to derive a so-called cost function from human behavior. This can be used, for example, for driver assistance functions to describe how a system (here, a driver assistance system) should behave. Human behavior is primarily determined by the values ​​of the driving parameters, as described elsewhere.The system here corresponds in particular to a motor vehicle and / or its control device and / or the corresponding components of a motor vehicle and / or a driver assistance system, in particular as controlled by a control device.

[0041] From one perspective, determining the at least one parameterization can be based on a correlation analysis between personality types, the personality traits of the personality types in the at least two drivers, and / or the expression of the at least one driving parameter in the at least two drivers. The correlations between driving parameters, as representations of driver input, and the personalities can be used to directly calculate the personality traits. The term "personality trait" can refer specifically to the previously explained personality traits of personality types in a driver. The term "personality trait" can be used synonymously. The correlations can be used, particularly directly, like coefficients. This allows the use of formulas for which these parameters can be found, which can be determined via the correlations.

[0042] For various drivers who have undergone the MDSI personality test, the method described here can be used to assess the influence of the personality traits identified in the MDSI on the drivers' driving trajectories. Correlation analyses, such as Spearman and / or Pearson correlations, can be performed for this purpose. Different aspects of driver behavior correlate with the identified personality traits and can be identified using the methods described here and used for parameterization, as described elsewhere.

[0043] From one perspective, a classification machine learning model can be trained to assign drivers to at least one driving parameter by adding their personality types and personality traits. Both the input and output of the system (here defined as the entirety of operators and parameters representing personality types and personality traits) can be known for a very large number of drivers. Therefore, machine learning approaches such as neural networks can be used. In particular, standard supervised feed-forward networks can be employed. One such example is a fully connected multilayer perceptron (FC-MLP). An FC-MLP is specifically defined as a fully connected artificial neural feed-forward network with at least three layers (input, output, and at least one hidden layer).The same time series data descriptions used in correlation analysis can be employed. Alternatively or additionally, further preprocessing of the time series data can be performed. Alternatively or additionally, a Long Short-Term Memory (LSTM) network can be used. An LSTM network is specifically a type of Recurrent Neural Network (RNN) designed for processing sequential data, such as time series. This allows for the utilization of the temporal context of the time series data.

[0044] Driver behavior can be categorized into different clusters based on driver personality (as mixed types, as described elsewhere). This allows for the use of unsupervised and / or semi-supervised approaches. The existing dataset can be subjected to clustering, enabling the assignment of labels to the personality traits of personality types and / or mixed types. These labels, based on driver personality, then serve to label the clusters. A new driver to be classified can be automatically assigned to one of these clusters, as explained below.

[0045] According to an independent assessment, a procedure, specifically a computer-implemented method, is designed and configured to assign personality traits of personality types to drivers. This is done to adjust at least one driving parameter. The parameterization is performed using at least one inverse method based on the personality traits of different personality types in at least two drivers. The procedure includes the step of determining the specific value of at least one driving parameter in at least one driver.

[0046] This involves determining, in one step, the personality traits of at least two personality types in at least one driver. This determination is based in particular on at least one driving parameter.

[0047] The procedure includes a step of adapting at least one driving function, in particular based on the determined personality components of at least two personality types of the driver.

[0048] As described elsewhere, individual drivers can be assigned to certain personality traits (personality components of personality types). The parameterization method for driving functions based on the personality components of drivers with different personalities, described elsewhere, allows for the analysis of correlations between various parameters, particularly for automated driving functions. Using previously explained inverse methods, parameters can be determined for individual drivers whose personality traits are known. These parameters can then be assigned to different driver personality traits (synonymous with personality traits, which represent the composition of personality components from different personality types) via correlation analyses.This makes it possible to find and group drivers with similar characteristics from a very large pool of drivers (e.g., a fleet). Instead of always optimizing parameters for just one driver, all recorded scenarios from drivers with (strongly) similar behavior can be used. Appropriate thresholds or deviation standards can be defined to establish similarity. This results in a much larger pool of scenarios for parameter optimization.

[0049] The parameters found and the corresponding, correlating driver personalities can then be used to release a secure parameter set, which can fluctuate around defined and previously secured values.

[0050] To illustrate this, consider a driving function with an adaptable parameter set. The main parameters (those most correlated with personality profiles, which are the combination of personality traits from different personality types) are designated as a1, a2, a3, and a4. A matrix can be created to show how these parameters can be adapted based on the driver's personality. For example, for a driver who is 30% "careful," a3 might be reduced by 10% and a4 by 5%. Increases are also possible. The individual factors can depend on the strength of the personality trait within a mixed profile of personality traits from different personality types.

[0051] The driver in question could be at least one other driver, particularly a third driver. The driver could therefore be a "new driver," meaning a driver who is not already included in an existing database of personality traits from personality types.

[0052] This determination can be based on at least one driving parameter. Since driving behavior (exhibiting the driving parameters as described elsewhere) can be categorized into different clusters depending on driver personality, unsupervised and / or semi-supervised approaches, as described elsewhere, can be used. In this case, the existing dataset would be subjected to clustering. The clusters could then be labeled using the existing labels indicating driver personality. A new driver to be classified can then be automatically assigned to one of these clusters.

[0053] According to one perspective, determining the level of at least one driving parameter can be based on the driver's driving behavior during a journey. This allows a driver to be assigned a personality profile, particularly based on their driving style during active operation. This profile is primarily based on a blend of personality traits from different personality types. As described elsewhere, this can involve at least two personality types. Reference is made here to the previously stated explanations, and repetition of text passages is avoided for the sake of conciseness and readability.

[0054] Alternatively or additionally, at least one parameter of an automated driving function can be adjusted. This adjustment can be selected from at least one maximum lateral acceleration, at least one braking point, or at least one acceleration point. This allows the method to adjust at least one parameter to meet or approximate the driver's expectations and / or preferences by selecting one of the aforementioned parameters or similar parameters. In particular, it is possible to imitate or approximate driver behavior in semi-automated driving mode.

[0055] From one perspective, determining the personality traits of at least two personality types can be based on a correlation analysis of the expression of at least one driving parameter. Alternatively or additionally, determining the personality traits of at least two personality types can be based on an inverse method. This allows a driver to be assigned a personality profile, particularly based on their driving style during operation, which is primarily based on a mixed type of personality traits from different personality types. This can involve at least two personality types, as already described elsewhere. Reference is made here to the previously stated explanations, and repetition of text passages is avoided for the sake of conciseness and readability.

[0056] One aspect that can be considered is the grouping of drivers with similar characteristics of at least one driving parameter, in order to adapt the at least one driving function based on the determined personality traits of at least two personality types. This makes it possible to base the parameterization of driving functions on the determined personality traits of at least two personality types and to enable the adaptation of at least one driving function.

[0057] An adjustment of the driving function can be achieved by modifying a parameter of the driving function. This parameter can be designed, as described elsewhere, to influence a driving parameter. This could involve, for example, the setting of a braking system, a traction motor, a control device, and / or an injection system. It could also involve adjusting the power supply from a high-voltage battery or redistributing power in a hybrid vehicle.

[0058] Alternatively or additionally, drivers with similar personality traits can be grouped to adjust at least one driving function based on the expected similarities in at least one driving parameter. This allows drivers to be grouped (also referred to as clustering) based on their assigned driving personalities (as mixed types, as explained elsewhere), since it can be expected that new drivers with a corresponding driving personality will also benefit from appropriate parameter selection and adjustments to a driving function. This enables situation-adapted parameter settings for a driving function, allowing, for example, the integration of a sportier driving style with an ECO mode (fuel-saving mode) and potentially alternating between the two.Safety can also be enhanced if the parameters of the driving function can be adjusted according to the driver's personality, which is a blend of personality traits from different personality types. For example, a driver with a higher "angry" personality trait can drive more sportily by adjusting at least one driving function to achieve higher acceleration. Braking behavior can also be adjusted, thereby increasing safety. Conversely, for other driver personalities with a higher "careful" personality trait, acceleration can be reduced, thus saving energy.

[0059] One aspect of this is that in a trained classification machine learning model, particularly one trained using a method as described elsewhere, a driver can be assigned to a cluster of drivers with similar proportions of personality types. This training of machine learning models may have been carried out according to the previously explained methods. Specifically, the trained neural networks, as described in detail elsewhere (and omitted here for the sake of conciseness and readability), are fed at least one driving parameter, allowing the driver to be assigned a label relating to their personality. For further information on this driver personality, please refer back to the previous explanations.

[0060] From an independent perspective, a computer program product can be designed and configured to be executed on a machine. When the computer program product is executed on the machine, one of the procedures described in detail elsewhere can be performed. The computer program product is, in particular, machine-readable code and / or an electrical signal configured to be read by a machine in order to transmit instructions to the machine, such as to perform a procedure of the type described elsewhere. A computer program product can, in particular, be designed as machine-readable code, especially as an algorithm capable of performing a corresponding inverse method and / or a correlation analysis.Alternatively or additionally, the machine-readable code can also execute instructions to adjust at least one driving function. Thus, a computer program product can be described by the features, properties, and advantages as described and explained for the procedures. The reverse is also true. For the sake of conciseness and readability, a repetition of all these features, properties, and advantages is omitted here.

[0061] From an independent perspective, a storage medium can incorporate a computer program product as described. Alternatively or additionally, a control device can incorporate a computer program product as described. The computer program product is designed and configured to perform one of the described procedures. Thus, the storage medium and / or the control device can be described by the features, properties, and advantages as described and explained for the procedures and / or the computer program product. The reverse is also true. For the sake of conciseness and readability, a repetition of all these features, properties, and advantages is omitted here.

[0062] From an independent perspective, an electronic signal can be configured with a parameterization determined according to a procedure as described. This electronic signal can hold and / or transmit a data set to enable the assignment of driver personalities to a driving parameter, thereby adapting a driving function. In particular, the electronic signal can contain a collection of mixed personality traits from the set of personality types. This makes it possible to perform this parameterization in a cloud-based process and / or to exchange and / or implement it in a collective network of vehicles within a fleet.

[0063] From an independent perspective, a motor vehicle may have a storage medium, particularly one already described elsewhere. Alternatively or additionally, the motor vehicle may have a control unit, particularly as described elsewhere. The motor vehicle may be configured and equipped to receive and / or transmit an electronic signal, particularly as described elsewhere. Alternatively or additionally, the motor vehicle may be configured and equipped to execute a computer program, particularly as described elsewhere, in such a way as to perform at least one of the procedures, particularly as described elsewhere, in particular to adapt at least one driving function based on personality traits of a driver's personality type.This allows the advantages to be realized, as already described with regard to the processes, the computer program product, the storage medium, the control unit, and / or the electronic signal. The motor vehicle can be described by the characteristics and properties of all these processes, devices, and / or systems. The reverse is also true. For the sake of readability and conciseness, a repetition of all these characteristics is omitted here.

[0064] The methods, devices, and systems described here and elsewhere can achieve a corresponding effect not only at the individual level of the driver or the vehicle. Rather, the various methods listed here also allow for the exploitation of savings potential at the fleet level, particularly with regard to energy consumption during fleet operation. In particular, the overall ecological footprint of the fleet can be optimized, for example, because certain drivers with a particular driving style tend to drive faster and / or accelerate more quickly on average. However, this can be offset by adjusting at least one driving function for other drivers who do not exhibit such driving behavior.

[0065] Further advantages, features, and details of the invention will become apparent from the following description, in which exemplary embodiments of the invention are described in detail with reference to the drawings. The features mentioned in the claims and in the description can be essential to the invention individually or in any combination. The drawings schematically show: Fig. 1A a schematic acceleration situation; Fig. 1B a schematic braking situation; Fig. 1C a schematic cornering situation; Fig. 2. An example matrix with driving parameters for different personality traits of personality types; Fig. 3. An overview of exemplary driver personalities as mixed profiles of personality traits from different personality types; Fig. 4 a schematic representation of an embodiment of a method; and Fig. 5 a schematic representation of an embodiment of a method.

[0066] Fig. Figure 1A shows a schematic acceleration situation 1 of a motor vehicle 30, which can be controlled by a control device 10, by a computer program product 20. This computer program product 20 is designed and configured to perform one of the procedures 200, 300 as described elsewhere. For the sake of simplicity, no drivers are shown here. However, different drivers can differ in their driving parameters, such as acceleration behavior 8a, 8b. In particular, a first driver can exhibit a first acceleration behavior 8a, whereby the first driver accelerates more slowly in similar or identical situations. A second driver can exhibit a second acceleration behavior 8b, whereby the second driver accelerates more quickly in similar or identical situations.

[0067] Fig. Figure 1B shows a schematic braking situation 2 of a motor vehicle 30, which can be controlled by a control unit 10, by a computer program product 20. This computer program product 20 is designed and configured to perform one of the procedures 200, 300, as described elsewhere. For the sake of simplicity, no drivers are shown here. However, different drivers can differ in their driving parameters, such as braking behavior 7a, 7b. In particular, a first driver can exhibit a first braking behavior 7a, whereby the first driver brakes more cautiously in similar or identical situations. A second driver can exhibit a second braking behavior 7b, whereby the second driver brakes more strongly in similar or identical situations.

[0068] Fig. Figure 1C shows a schematic cornering situation 3. Drivers (not shown) can control a motor vehicle on a roadway bounded by shoulders 4 and divided into lanes by a median strip 6. A first driver can demonstrate a first cornering behavior 5a, staying within their assigned lane. This allows for higher cornering acceleration, necessitating appropriate braking and / or acceleration when entering and exiting the curve. A second driver can demonstrate a second cornering behavior 5b in the same or a similar situation. In this case, the second driver follows an ideal line to minimize cornering acceleration. It is explicitly emphasized that drivers must always adapt their driving to the situation and adhere to the legal regulations.

[0069] These and other differences in driving behavior are reflected in driving parameters, such as different acceleration and / or braking inputs. Furthermore, different drivers can also select different steering responses. These examples are merely illustrative; many other possible driving parameters can be described.

[0070] Fig. Figure 2 shows an example matrix 100, which schematically depicts four identified driving parameters a1, a2, a3, and a4, along with their corresponding correlated driver personalities. These can be used to create a validated parameter set that can fluctuate by defined and previously validated values. For clarity, a driving function with a correspondingly adaptable parameter set can be assumed. The main parameters (and those most correlated with the driver personalities) are a1, a2, a3, and a4, purely for illustrative purposes. A matrix can now be created to show how the driving parameters can be adapted based on the driver personality, for example, for a driver whose personality trait has been determined to be 30% (from "careful" 135). In this case, a3 can be reduced by 10% and a4 by 5%. Specifically, the driving parameter a2 remains unchanged, while the driving parameter a1 is increased by 5%.The individual factors depend in particular on the strength of the personality trait. The same approach can be taken with the other personality traits listed here as examples: highvelocity 110, risky 120, and patient 140. In this case, a highvelocity 110 personality trait of 10% can lead to a 10% reduction in the first driving parameter a1, while a2 and a3 remain unchanged, and a4 can be increased by 5%. Similarly, a risky 120 personality trait of 20% can lead to a 5% reduction in the first driving parameter a1, while a2 can remain unchanged, a3 can be increased by 10%, and a4 can be increased by 5%. The personality component of patient 140 can be determined to be 40% and can lead to an increase of the first driving parameter a1 by 5%, a2 can be increased by 5%, a3 can be reduced by 5% and a4 can be reduced by 5%.For a driver with a corresponding driving personality, a driving profile can be set in which, for example, aspects of their personality are reflected in the setting of a driving function. In other embodiments, all settings can be taken into account. In a simple embodiment, the overall driving function can be set by summing the individual driving parameter adjustments. Thus, for a corresponding driver personality, driving parameter a1 would be reduced by 5%, driving parameter a2 increased by 5%, driving parameter a3 reduced by 5%, and driving parameter a4 would remain unchanged (at a basic setting). Driving parameter a1 could be a brake setting, driving parameter a2 a transmission setting, driving parameter a3 an acceleration setting, and driving parameter a4 a steering setting.These allocations and the stated percentages are purely examples for better understanding.

[0071] Alternatively or additionally, at least one parameter of an automated driving function can be adjusted. This parameter can be selected from at least a maximum lateral acceleration, at least a braking point, or at least an acceleration point. This allows the parameters of an automated driving function, as expected by the driver, to mimic or support the driver's driving behavior. It is possible that, in semi-autonomous driving mode, a driver's driving style can be imitated or approximated.

[0072] Fig. Figure 3 shows corresponding examples of driver personalities as determined in an MDSI personality test (80), in comparison to a self-identification (90), in which drivers assess themselves. Significant differences can become apparent, as illustrated in the four example diagrams. In addition to the previously explained personality types high-velocity (110), risky (120), careful (135), and patient (140), the personality types angry (115), anxious (125), reckless (130), and distress-reduction (145) are also identified. In the driver personality shown in the lower right, the self-identification (90) consists of a mixture of 30% high-velocity (110) and 30% 135. This means the diagram lies across the axes and is therefore difficult to display. These are also examples intended to aid understanding.These driver personalities, representing a blend of personality traits of the indicated strength of the various personality types, serve as a basis for carrying out the procedure described below. A corresponding matrix 100 can be generated and maintained as an electronic signal after parameterization has been performed according to one of the procedures 200 described below. Such an electronic signal can be made available to a cloud computing facility, a control unit 20, or another machine for processing. It can also be transmitted to another motor vehicle 30.

[0073] Fig. Figure 4 schematically shows an embodiment of a method 200. This method is in particular designed and configured as a computer-implemented method 200 for parameterizing driving functions based on personality traits of drivers with different driver personalities.

[0074] Procedure 200 includes, in particular, a step 210 of determining at least two personality traits of different personality types for at least two drivers. This procedure can involve not just two drivers, but a large number of drivers, who in turn represent a subset of a larger set of drivers. This set of drivers can represent a fleet, which may consist, for example, of all motor vehicles 30 equipped with a corresponding control device 10.

[0075] The procedure includes, in particular, a step of determining 220 a characteristic of at least one driving parameter of at least two drivers.

[0076] The procedure includes in particular a step of determining at least one parameterization depending on the personality components of the personality types and the expression of at least one driving parameter of the at least two drivers.

[0077] The procedure includes, in particular, a step involving the transmission of parameterization (240), in which the corresponding assignments of personality components, the resulting mixed types, and / or driving functions and their adjustments for driving parameters are transmitted, for example, to a control unit (10) of another vehicle. Alternatively or additionally, transmission to a cloud for storage and / or further distribution is also possible.

[0078] Determining the at least two personality traits of different personality types can be based on a selection from a personality test or a self-identification, as is the case with regard to the Fig. 3 has already been described. Alternatively or additionally, it can be based on a driving behavior analysis, as is also the case with regard to the Fig. 5 is described.

[0079] At least one driving parameter can be selected from a description of the driving curves, the sinusoidal nature of the driving, the activation patterns of the accelerator, brake and / or steering inputs. Corresponding driving parameters are shown in exemplary and schematic form in the Fig. Listed 1A to 1C.

[0080] At least one parameterization can be carried out using at least one inverse method on the personality components of the different personality types in the at least two drivers.

[0081] Determining the at least one parameterization (230) can be based on a correlation analysis between the personality types, the personality proportions of the personality types in the at least two drivers, and / or the expression of the at least one driving parameter in the at least two drivers. In particular, a matrix (100), such as the one described in the Fig. The parameters shown in Figure 2 serve as a basis for parameterizing the driving function. This allows for adjustments to the driving function parameters, thereby increasing safety during journeys. Alternatively or additionally, the fleet's energy consumption can be reduced, often accompanied by a reduction in greenhouse gas emissions.

[0082] A classification machine learning model can be trained to assign drivers to at least one driving parameter by adding their personality types and personality traits. Both the input and output of the system (here defined as the entirety of operators and parameters representing personality types and personality traits) can be known for a very large number of drivers. Therefore, machine learning approaches such as neural networks can be used. In particular, standard supervised feed-forward networks can be employed. One such example is a fully connected multilayer perceptron (FC-MLP). An FC-MLP is specifically defined as a fully connected artificial neural feed-forward network with at least three layers (input, output, and at least one hidden layer).The same time series data descriptions used in correlation analysis can be employed. Alternatively or additionally, further preprocessing of the time series data can be performed. Alternatively or additionally, a Long Short-Term Memory (LSTM) network can be used. An LSTM network is specifically a type of Recurrent Neural Network (RNN) that can be designed for processing sequential data, such as time series. This allows the temporal context of the time series data to be utilized.

[0083] Driver behavior can be categorized into different clusters based on driver personality (as mixed types, as described elsewhere). This allows for the use of unsupervised and / or semi-supervised approaches. The existing dataset can be subjected to clustering, enabling the assignment of labels to the personality traits of personality types and / or mixed types. These labels, based on driver personality, then serve to label the clusters. A new driver to be classified can be automatically assigned to one of these clusters, as explained below.

[0084] Fig.Figure 5 shows a schematic representation of an embodiment of a method 300, which can be designed and configured, in particular, as a computer-implemented method 300 to assign personality traits of personality types to drivers. This is done, in particular, to adjust at least one driving parameter.

[0085] Procedure 300 can include a step of determining 310 a value of at least one driving parameter for at least one driver. This driver can, in particular, be at least one other driver, especially a third driver. This can also be a subset of drivers that, in particular, has not been used as part of a training subset for training machine learning approaches. Furthermore, this cannot, in particular, be the part of the subset that has already been used for a correlation analysis or for an inverse method to realize an initial parameterization. Thus, a driving function can also be adapted for new drivers.

[0086] The procedure 300 can include a step of determining 320 the personality components of at least two personality types in the at least one driver, based on the at least one driving parameter.

[0087] The procedure 300 may include a step of adapting 330 at least one driving function, based on the determined personality components of at least two personality types of the driver.

[0088] Determining the value of at least one driving parameter (310) can be based on the driver's driving behavior during a single trip. Information on the driving behavior of a single driver can also be collected across multiple trips. This applies equally to other drivers. This allows for iterative adjustment of the parameterization, as the information obtained can be fed back into a process (200). Furthermore, determining the value of at least one driving parameter (310) based on one or more trips allows a driver's driving behavior to contribute to its grouping and, in particular, to form its basis.

[0089] Determining the personality traits of at least two personality types can be based on a correlation analysis of the expression of at least one driving parameter. Alternatively or additionally, determining the personality traits of at least two personality types can be based on an inverse method.

[0090] The procedure 300 can include the step of grouping 340 drivers with similar characteristics of at least one driving parameter in order to adapt the at least one driving function based on the determined personality components of at least two personality types.

[0091] Alternatively or additionally, the procedure 300 may include the step of grouping 340 drivers with similar personality traits in order to adapt the at least one driving function based on the expected similar characteristics of the at least one driving parameter.

[0092] A classification machine learning model can be trained, in particular according to a method 200 as described elsewhere. This allows a driver to be assigned to a cluster of drivers with similar proportions of personality types. The classification machine learning model was trained specifically as described elsewhere. In particular, the driving behavior of one or more drivers is fed into a trained model, allowing a label to be attached to each driver. This enables the driver to be assigned a driver personality (as a mixed type with personality traits from the various personality types) and thus subjected to clustering. This also allows new drivers to be included and represented in the process, enabling the driving functions to be adjusted accordingly.

[0093] The preceding explanation of the embodiments describes the present invention solely by way of examples. Naturally, individual features of the embodiments can be freely combined with one another, provided this is technically feasible, without departing from the scope of the present invention.

Claims

[1] Method (200), in particular a computer-implemented method, designed and set up for parameterizing driving functions based on personality components of drivers with different driver personalities, the method comprising the steps: - one determination (210) of at least two personality components of different personality types for at least two drivers; - of determining (220) a value of at least one driving parameter of at least two drivers; and - a determination (230) of at least one parameterization depending on the personality components of the personality types and the expression of at least one driving parameter of the at least two drivers; wherein the procedure characterized by The requirement is that at least one parameterization is carried out using at least one inverse method on the personality components of the different personality types in the at least two drivers. [2] Method (200) according to claim 1, characterized by , that the determination (210) of at least two personality components of different personality types is based on a selection from a personality test (80), a self-identification (90) or a driving behavior analysis. [3] Method (200) according to one of claims 1 or 2, characterized by , that at least one driving parameter is selected from a description of the curvature of the driving, the sinusoidal nature of the driving, the activation profiles of accelerator, brake and / or steering inputs. [4] Method (200) according to any of the preceding claims, characterized by , that the determination (230) of the at least one parameterization is based on a correlation analysis between the personality types, the personality components of the personality types in the at least two drivers and / or the expression of the at least one driving parameter of the at least two drivers. [5] Method (200) according to any of the preceding claims, characterized by , that a classification machine learning model is trained to assign at least one driving parameter by adding the personality types and personality components of the drivers. [6] Method (300), in particular a computer-implemented method, trained and set up to assign personality components of personality types to drivers in such a way as to adapt at least one parameter of a driving function, characterized by , that at least one parameterization has been carried out using at least one inverse method on the personality components of the different personality types of at least two drivers, the procedure comprising the following steps: - of determining (310) a characteristic of at least one driving parameter in at least one driver, in particular as at least one further, in particular third driver; - a determination (320) of the personality components of at least two personality types in the at least one driver, based on the at least one driving parameter; and - an adaptation (330) of at least one driving function based on the determined personality components of at least two personality types of the driver. [7] Method (300) according to claim 6, characterized by , that the determination (310) of the value of at least one driving parameter is based on the driver's driving behavior during a journey; and / or wherein at least one parameter of an automated driving function is adapted, selected from at least one maximum lateral acceleration, at least one braking point or at least one acceleration point. [8] Method (300) according to one of the preceding claims 6 or 7, characterized by, that the determination (320) of the personality components of at least two personality types is based on a correlation analysis of the expression of at least one driving parameter and / or that the determination (320) of the personality components of at least two personality types is based on an inverse method. [9] Method (300) according to any one of claims 6 to 8 above, characterized by the step: - a grouping (340) of drivers with similar characteristics of the at least one driving parameter in order to adapt the at least one driving function based on the determined personality components of at least two personality types; and / or - a grouping (340) of drivers with similar personality traits to perform the adaptation of the at least one driving function based on the expected similar expressions of the at least one driving parameter. [10] Method (300) according to any one of claims 6 to 9 above, characterized by , that in a trained classification machine learning model, in particular trained according to a method (200) according to claim 6, a driver is assigned to a cluster of drivers with similar proportions of personality types. [11] Computer program product (20), designed and configured to be executed on a machine, wherein, when executing the computer program product (20) on the machine, a method (200, 300) according to one of the preceding claims is carried out. [12] Storage medium and / or control device (10) comprising a computer program product (20) according to claim 11, designed and configured to perform a method (200, 300) according to any one of claims 1 to 10. [13] Electronic signal comprising a parameterization determined by a method (200, 300) according to any one of claims 1 to 5. [14] Motor vehicle (30), in particular comprising a storage medium and / or a control device (10) according to claim 12, wherein the motor vehicle is configured and designed to receive and / or transmit an electronic signal, in particular according to claim 14; and / or wherein the motor vehicle is configured and designed to execute a computer program product (20), in particular according to claim 11, such that it is used to carry out a method according to any one of claims 1 to 10, in particular to adapt at least one driving function based on personality traits of a driver's personality type.

Images