Method and device for operating an infotainment system of a vehicle

Abstract

Method for operating an infotainment device (20) of a vehicle, which has a selection module (22) for media pieces, which is designed to select media pieces from a set of predetermined media pieces for signaling by the infotainment device (20) depending on predetermined rules, in which - at least one physiological parameter is determined depending on at least one biosignal of a vehicle user recorded with a predetermined sensor, which is recorded during and / or after the signalling of a selected piece of media, - for each of the physiological parameters determined, at least one characteristic feature is determined depending on the respective predetermined physiological parameter, wherein the at least one characteristic feature is determined depending on a time-based signal curve of the physiological parameter and / or a frequency-based signal curve of the physiological parameter, - depending on the characteristic features determined, an emotion rating from a predefined emotion rating range or from a set of predefined emotion rating classes is assigned to the selected media piece, - the determined characteristic features are compared with respective example instances of a set of given example instances, - depending on the comparison, the emotion rating is assigned to the signalled media piece, and - in which the emotion evaluation comprises a valence value or a valence class which characterises how pleasant the signalling of the media piece is perceived by the vehicle user, wherein the valence value or the valence class are available after a classification and the emotion evaluation is projectable onto a valence scale.

Description

[0001] The invention relates to a method and a device, as well as a computer program, for operating an infotainment system of a vehicle. Furthermore, the invention relates to a computer program product.

[0002] Infotainment systems are regularly used in motor vehicles to reproduce predefined infotainment data. Infotainment data is primarily presented acoustically and / or visually. The vast array of music tracks, such as MP3 songs and podcasts, in many media databases available today (music collections on a hard drive and / or in an online music store) awakens the desire in some users for automated music selection. Therefore, recommendation systems are used, particularly in the music sector, to provide personalized program preparation for the user.However, such recommendation systems for music tracks and / or podcasts usually comprise a large number of operating steps and are therefore not practical, or only with limitations, in the automotive environment, for example, in which the operator, who is primarily concerned with the driving task, can devote only a fraction of his or her overall attention to operating a music playback system.

[0003] DE 2007 008 815 A1 discloses a method for acoustically outputting system information in a vehicle, in which attributes are assigned to the music audible in the vehicle interior and the acoustic output of the system information is dependent on the attributes.

[0004] DE 103 43 683 A1 discloses a system for providing information depending on the factors that stress the driver when driving a motor vehicle. The system comprises a) a device for recording objective stress factors acting on the driver, b) a device for recording driver activities, c) a device for recording the driver's individual characteristics that influence driving the vehicle, d) an information processing unit for generating an information profile depending on the stress factors acting on the driver, the driver activities, and the individual characteristics of the driver that influence driving the vehicle, and e) an output device for outputting the information determined by the information profile.

[0005] DE 10 2005 058 227 A1 discloses an emotion-based software robot for automobiles in which the emotion of a driver and the behavior caused by such an emotion are presumed and expected when input data such as states, commands and behaviors of a driver, automobile situations, environmental situations of the automobile, etc., are recognized, based on offline learned results regarding a change in the emotions of each driver, and each vehicle information is prioritized so that services provided by a telematics system, etc. can be implemented proportionately to correspond to the mood of the driver.

[0006] DE 10 2010 036 666 A1 discloses a method for intelligent music selection in vehicles based on user preferences and driving conditions.

[0007] XIE, Zhibing; GUAN, Ling: Multimodal information fusion of audio emotion recognition based on kernel entropy component analysis. In: 2012 IEEE International Symposium on Multimedia, 10-12 Dec. 2012, 8 pp. - ISBN 978-1-4673-4370-1, discloses a method for multimodal information fusion in audio emotion recognition based on kernel entropy component analysis.

[0008] The object underlying the invention is to provide a method, a device and a computer program for operating an infotainment device as well as a computer program product that enable a simple and efficient improvement and / or adaptation and / or personalization of the provision of media pieces by the infotainment device.

[0009] The problem is solved by the features of the independent patent claims. Advantageous developments of the invention are characterized in the subclaims.

[0010] According to a first and second aspect, the invention is characterized by a method and a corresponding device for operating an infotainment system of a vehicle. The infotainment system comprises a selection module for media pieces, which is designed to select media pieces from a set of predefined media pieces for signaling by the infotainment system depending on predefined rules. At least one physiological parameter is determined depending on at least one biosignal of a vehicle user detected with a predefined sensor, which is detected during and / or after the signaling of a selected media piece. For the determined physiological parameter(s), at least one characteristic feature is determined depending on the respective predefined physiological parameter.Depending on the characteristic features determined, the selected media piece is assigned an emotion rating from a predefined emotion rating value range or from a set of predefined emotion rating classes.

[0011] Advantageously, the recorded biosignals can be used to quantify the vehicle user's emotional reactions. Emotion-specific patterns are extracted and classified from the biosignals.

[0012] At least one characteristic feature is determined based on a time-based signal profile of the physiological parameter. This allows for the determination of time-dependent characteristic properties of the biosignal, thereby significantly reducing the amount of data to be processed and simultaneously increasing the performance of the classification algorithm.

[0013] Alternatively or additionally, at least one characteristic feature is determined based on a frequency-based signal profile of the physiological parameter. This allows frequency-dependent characteristic properties of the biosignal to be determined, thereby significantly reducing the amount of data to be processed and simultaneously increasing the performance of the classification algorithm.

[0014] Biosignals encompass autonomous, energetically and materially measurable physical quantities generated by living organisms. One or more sensors are used to record the biosignals, for example, an electrocardiogram sensor and / or an electromyogram sensor and / or a respiratory sensor and / or a skin conductance sensor. The sensors record the biosignals, and transducers convert them into electrical quantities that can then be processed automatically. Functional measurements derived from the biosignals, such as heart rate or respiratory rate, are referred to as indicators, parameters, or characteristic values.

[0015] The identified characteristic features are compared with respective example instances from a set of predefined example instances, and based on the comparison, the emotion rating is assigned to the signaled media piece. The example instances can also be referred to as example feature vectors. This has the advantage that instance-based learning algorithms can be used. Instance-based learning is based on a very simple idea: To predict the class of a new, unknown object, the object to be classified is compared with the predefined example instances, the instance most similar to the object to be classified is determined, and the objective function and / or function value stored for this instance is assigned to the unknown object.The advantage is that no information contained in a training set is abstracted away, and thus the information is not lost. The training set, in particular, includes the set of example instances.

[0016] The emotion evaluation is preferably forwarded to the selection module via a predefined interface. Depending on the reaction to the current media piece, the playback of the selected media piece can be interpreted as desired or undesired. The vehicle user is not required to provide explicit feedback. The vehicle user is therefore not distracted from other tasks, in particular from controlling the vehicle, by the task of providing feedback. In particular, the emotional reaction can be used as a quantitative input for the selection module, and the selection module can be adapted. In particular, a classification of the media pieces by the selection module can be adapted. The biosignals advantageously enable the emotional reactions to be quantified much more precisely than binary quantification. An initial parameterization of the selection system can be dispensed with.

[0017] Different classification methods from the field of machine learning can be used to assign the emotion rating to the selected media piece based on the determined characteristic features.

[0018] A media piece can include, for example, a piece of music, a podcast, a Twitter message, a point-of-interest piece, traffic information, and / or an email, etc. Alternatively, the term "infotainment piece" can also be used.

[0019] In a further advantageous embodiment of the first and second aspects, the emotion rating is assigned to the selected media piece based on a predefined instance-based classification algorithm. For example, a k-nearest neighbor algorithm or a K* algorithm, or an eigenvalue determination, can be used.

[0020] In a further advantageous embodiment of the first and second aspects, a distance measure between a feature vector and the respective example instance is determined for at least some of the example instances of the set, wherein the distance measure is determined according to a predetermined rule and the feature vector comprises at least some of the determined characteristic features. Depending on the determined distance measures, the rating is assigned to the media piece. The distance measure can comprise a Euclidean distance and / or a weighted Euclidean distance and / or a Minkowski distance, and so on. This has the advantage that a similarity between the feature vector and the respective example instances can be quantitatively assessed very easily.

[0021] The emotion rating includes a valence value or valence class, which characterizes how pleasant the signaling of the media piece is perceived by the vehicle user. The valence value or valence class is available after classification, and the emotion rating can be projected onto a valence scale. This has the advantage that the emotion rating can be very easily evaluated and further processed.

[0022] In a further advantageous embodiment of the first and second aspects, the distance measure comprises an entropy-based distance measure. This advantageously enables a more reliable assignment of the emotion rating to the detected patterns in the biosignals. It is assumed that the emotion and / or the change in emotion during listening to the media pieces is attributable to the respective media piece. Since the valence value of the emotion is available after classification, it can be deduced whether the driver likes the media content currently being played.

[0023] According to a third aspect, the invention is characterized by a computer program for operating an infotainment device, wherein the computer program is designed to carry out the method for operating an infotainment device or an advantageous embodiment of the method on a data processing device.

[0024] According to a fourth aspect, the invention is characterized by a computer program product comprising executable program code, wherein the program code, when executed by a data processing device, executes the method for operating an infotainment device or an advantageous embodiment of the method.

[0025] Embodiments of the invention are explained below with reference to the schematic drawings.

[0026] They show: Fig. 1 an exemplary block circuit system of an infotainment system, Fig. 2 an exemplary flow chart for a program for operating an infotainment device and Fig. 3 an example of an emotion model.

[0027] Elements of the same construction or function are provided with the same reference symbols throughout the figures.

[0028] Fig. 1 shows an infotainment system 10. The infotainment system 10 has an infotainment device 20. The infotainment device 20 includes, for example, a media item database 24 and / or a receiving unit for receiving media items from predefined vehicle-external databases and / or from predefined vehicle-external servers. The media item database 24 is configured, for example, to store a plurality of music pieces and / or podcasts and / or electronic messages.

[0029] The infotainment device 20 has, for example, a control unit 26 for controlling the output and / or for evaluating operating instructions.

[0030] For example, the infotainment device 20 is assigned an acoustic and / or optical output unit 29, 28. Alternatively, the infotainment device 20 can include the acoustic and / or optical output unit 29, 28. The infotainment device 20 is configured to output the media pieces in a predetermined manner using the acoustic and / or optical output unit 29, 28.

[0031] The infotainment device 20 further comprises a selection module 22. The selection module 22 can also be referred to as a recommendation module. The selection module 22 is designed, for example, for the personalized, automated selection of media pieces signaled by the infotainment device 20. The selection module 22 selects media pieces from a set of predefined media pieces, for example, from the media piece database 24, depending on predefined rules.

[0032] The selection module 22 is configured, for example, to analyze and classify media pieces, in particular music pieces, that are actively selected by the vehicle user. The selection module 22 can use different machine learning algorithms and / or different input variables to learn user preferences for the media piece selection. In particular, the selection module 22 can use different input variables for the respective media piece selection, for example, recorded operating variables of the vehicle and / or environmental variables of the vehicle and / or biosignals of the vehicle user.

[0033] For this purpose, the selection module 22 can, for example, be designed to classify the media pieces according to predetermined moods and to determine a current mood of the driver depending on detected biosignals of a vehicle user and to select a media piece depending on the determined mood, which is then signaled by the infotainment device 20.

[0034] The selection module 22 is preferably designed to select the media pieces depending on one or more emotion ratings, which are each provided in response to the signaling of previously selected media pieces at a predetermined interface of the selection module 22.

[0035] A detection device 15 is assigned to the infotainment system 10. Alternatively, the infotainment system 10 may include the detection device 15. The detection device 15 comprises one or more predetermined sensors for detecting one or more biosignals.

[0036] The detection device 15 is designed, for example, to detect electrocardiogram signals and / or electroencephalogram signals, blood pressure signals, respiration signals, body temperature signals, sweat secretion signals, and / or electrical conductivity signals of the skin of a vehicle user. The detection of the biosignals is preferably non-invasive.

[0037] Furthermore, the infotainment system 10 has an evaluation device 30, which can also be referred to as a device for operating an infotainment device 20.

[0038] The evaluation device 30 is configured to execute a program for operating the infotainment device 20. For this purpose, the evaluation device 30 comprises, for example, a computing unit and a program memory. The evaluation device 30 is signal-coupled to the detection device 15 and to the infotainment device 20, in particular to the selection module 22.

[0039] Fig. 2 shows an example flow chart of the program.

[0040] The program is started in step S10. The program is started, for example, when the selection module 22 of the infotainment system 20 is activated. It can be provided that the selection module 22 is activated upon activation of the infotainment system 20 and / or after a predetermined operation of the infotainment system 20 by the vehicle user.

[0041] In a step S12, at least one physiological parameter is determined based on at least one detected biosignal of the vehicle user, which is detected during and / or after the signaling of a selected media piece and provided by the detection device 15. Preferably, at least one physiological parameter is determined for each detected biosignal.

[0042] For example, in step S12, a heart rate and / or a blood pressure and / or a respiratory rate and / or a body temperature and / or an electrical conductivity of the skin of the vehicle user is determined.

[0043] Step S12 comprises, for example, a normalization and / or scaling of the biosignals and / or an elimination and / or at least reduction of interference components.

[0044] In a step S14, at least one characteristic feature is determined for each of the determined physiological parameters, depending on the respective predefined physiological parameter. In step S14, a feature vector is preferably determined for the currently selected and signaled media piece. In this case, both characteristic features are preferably provided that are determined depending on a frequency-based signal profile of the physiological parameter, characteristic features are provided that are determined depending on a time-based signal profile, and characteristic combination features that characterize both time-dependent and frequency-dependent properties.

[0045] In a step S16, depending on the determined characteristic features, an emotion rating from a predefined emotion rating value range or from a set of predefined emotion rating classes is assigned to the selected media piece.

[0046] For this purpose, for example, the determined characteristic features that form a feature vector are compared with respective example instances of a set of predefined example instances. The emotion rating is assigned to the selected media piece depending on a predefined instance-based classification algorithm. For each example instance of the set, a distance measure between the feature vector and the respective example instance is determined. A k-nearest neighbor algorithm, for example, is used as the classification algorithm, and the respective distance measure represents, for example, the Euclidean distance between the feature vector and the respective example instance. Depending on the determined distance measures, the emotion rating is assigned to the media piece. Alternatively or additionally, an entropy-based distance measure is used, for example.

[0047] The emotion rating is provided, for example, at a predefined interface for forwarding to the selection module 22.

[0048] The program is run again, for example, for each piece of media selected by the selection module 22. The program is terminated in step S18. The program is terminated, for example, by deactivating the selection module 22.

[0049] Fig. Figure 3 shows an example of a two-dimensional emotion model. A valence is assigned to a first axis. The valence characterizes how positive or negative an emotion is. In other words, the valence characterizes how pleasant something is perceived by a person. A second axis is assigned to an activity. The activity characterizes an arousal triggered by an event. The combination of both values ​​results in the emotion. Fig.The emotion model shown in Figure 3 includes eight basic emotions: anger, fear, surprise, joy, anticipation, acceptance, sadness, disgust.

[0050] Basic emotions are emotions that cannot be further traced back to other emotions or emotions from which all other - more complex - emotions are composed.

[0051] The emotion rating includes, in particular, a valence value that characterizes how pleasant the signaling of the media piece is perceived by the vehicle user. The valence value can thus be used by the selection module 22 as feedback from the vehicle user to better adapt the selection of media pieces to the vehicle user's preferences. List of reference symbols 10 Infotainment system 15 Recording device 20 Infotainment system 22 Selection module 24 Media Item Database 26 Control unit 28 optical output unit 29 acoustic output unit 30 Evaluation device S10 ...S18 program steps

Claims

[1] Method for operating an infotainment device (20) of a vehicle, which has a selection module (22) for media pieces, which is designed to select media pieces from a set of predetermined media pieces for signaling by the infotainment device (20) depending on predetermined rules, in which - at least one physiological parameter is determined depending on at least one biosignal of a vehicle user recorded with a predetermined sensor, which is recorded during and / or after the signalling of a selected piece of media, - for each of the physiological parameters determined, at least one characteristic feature is determined depending on the respective predetermined physiological parameter, wherein the at least one characteristic feature is determined depending on a time-based signal curve of the physiological parameter and / or a frequency-based signal curve of the physiological parameter, - depending on the characteristic features determined, an emotion rating from a predefined emotion rating range or from a set of predefined emotion rating classes is assigned to the selected media piece, - the determined characteristic features are compared with respective example instances of a set of given example instances, - depending on the comparison, the emotion rating is assigned to the signaled media piece, and - in which the emotion evaluation comprises a valence value or a valence class which characterises how pleasant the signalling of the media piece is perceived by the vehicle user, wherein the valence value or the valence class are available after a classification and the emotion evaluation is projectable onto a valence scale. [2] Method according to claim 1, wherein the emotion rating is assigned to the selected media piece depending on a predetermined instance-based classification algorithm. [3] Method according to claim 1 or 2, in which - for at least some of the example instances of the set, a distance measure between a feature vector and the respective example instance is determined, wherein the distance measure is determined according to a predetermined rule and the feature vector comprises at least some of the determined characteristic features and - depending on the determined distance measurements, the rating is assigned to the media piece. [4] Method according to one of the preceding claims, wherein the distance measure comprises an entropy-based distance measure. [5] Device for operating an infotainment device (20) of a vehicle, wherein the infotainment device (20) has a selection module (22) for media pieces, which is designed to select media pieces from a set of predetermined media pieces for signaling by the infotainment device (20) depending on predetermined rules and the device is designed to carry out a method according to one of claims 1 to 4. [6] Computer program for operating an infotainment device (20) of a vehicle, wherein the computer program is designed to carry out a method according to one of claims 1 to 4 when executed on a data processing device. [7] A computer program product comprising executable program code, wherein the program code, when executed by a data processing device, carries out the method according to any one of claims 1 to 4.

Images