Procedures for assessing the condition of a motor vehicle as well as motor vehicles

Abstract

The invention relates to a method for evaluating the condition of a motor vehicle (10), in which a sound signal is provided to an electronic computing unit (14) by means of at least one microphone (12) of the motor vehicle (10), wherein the sound signal characterizes sounds detected by means of the microphone (12) (V1), and the at least one sound signal is examined with regard to existing patterns by means of a trained model (V2), and depending on at least one pattern recognized by the model in the at least one sound signal, the condition of the motor vehicle (10) is evaluated (V4), wherein sound signals are only evaluated for situations in which at least one predetermined criterion is present (V3) which has an influence on background noise in the at least one sound signal.

Description

[0001] The present invention relates to a method for evaluating the condition of a motor vehicle and to a motor vehicle.

[0002] From DE 10 2017 200 372 A1, a method for diagnosing a motor vehicle is known. In this method, an acoustic signal is detected outside the vehicle cabin using at least one microphone, wherein the microphone is located outside the vehicle cabin and the acoustic signal is generated by a component of the vehicle. Furthermore, the method detects at least one state of the vehicle and classifies the acoustic signal. The classified acoustic signal is correlated with the at least one state of the vehicle. Finally, the method determines the cause of the acoustic signal based on the correlation of the classified acoustic signal with the at least one state of the vehicle.

[0003] Furthermore, DE 10 2022 127 924 B3 discloses a method for locating and / or classifying a problem in the chassis of a motor vehicle. In this method, a first acoustic signal from the chassis is recorded by a detection device in the motor vehicle, and a noise component is determined from this recorded acoustic signal by a computer in the motor vehicle. The computer then determines the vehicle parameter settings under which the noise component was identified. At least one of these vehicle parameter settings is then changed, and a second acoustic signal from the chassis is recorded under the changed parameter setting.The computer determines the change in the background noise between the first and second acoustic signals and locates and / or classifies the problem in the chassis area depending on the change in the background noise and the changed vehicle parameter setting.

[0004] Furthermore, DE 10 2007 051 261 A1 discloses a method for the acoustic assessment of a motor vehicle, in which acoustic signals from the vehicle are recorded using at least one microphone. The acoustic signals are preprocessed, and analytical procedures are performed on the preprocessed signals to determine characteristics. The characteristics are then classified into classes, and the classification results are compared with a reference classification.

[0005] The object of the present invention is to provide a solution by means of which the condition of a motor vehicle can be assessed with particular precision.

[0006] This problem is solved according to the invention by the subject matter of the independent claims. Further possible embodiments of the invention are disclosed in the dependent claims, the description, and the figures. Features, advantages, and possible embodiments set forth in the description for one of the subject matter of the independent claims are to be regarded, at least analogously, as features, advantages, and possible embodiments of the respective subject matter of the other independent claims, as well as of any possible combination of the subject matter of the independent claims, optionally in conjunction with one or more of the dependent claims.

[0007] The invention relates to a method for evaluating the condition of a motor vehicle, in particular a passenger car. In the method, a sound signal is provided to an electronic processing unit by means of at least one microphone in the motor vehicle. If the motor vehicle has several microphones, each microphone provides a sound signal to the electronic processing unit. The respective sound signals characterize sounds detected by the respective microphones. This means that at least one sound signal can be provided to the electronic processing unit by each microphone, with each sound signal characterizing a sound detected by the respective microphone. The electronic processing unit is, in particular, part of the motor vehicle.The respective sound signals, particularly in the form of voltage signals, can be provided by the microphones to the electronic processing unit. The method further provides that at least one sound signal is analyzed for previously existing patterns using a machine learning-trained model. This model can be executed by the electronic processing unit. Specifically, the model is trained using machine learning.

[0008] Alternatively or additionally, the model can be trained using rules, for example using fuzzy logic, or using a map-based approach.

[0009] Machine learning is a subfield of artificial intelligence and requires neural networks and deep learning. Machine learning develops, investigates, and uses statistical algorithms, also called learning algorithms. The trained model can, for example, be a neural network. Artificial neural networks are networks of artificial neurons that are interconnected and often organized in layers.

[0010] The trained model is used to examine whether the respective noise signals exhibit patterns, particularly repetitive patterns. The procedure thus allows verification of whether a sequence occurs in the noise signal under investigation that appears multiple times consecutively, especially at regular intervals. For example, the pattern of the noise signal may be such that a specific sequence occurs at a fixed interval, such as once per rotor rotation of a vehicle's drive motor or 4.6 times per rotor rotation. Furthermore, the procedure includes an assessment of the vehicle's condition based on at least one pattern detected in the noise signals by the model.In particular, if a period of the sequence occurring several times at regular intervals can be correlated with the rotation of a rotor of the vehicle's drive motor, it can be assumed that the noise characterized by the respective sound signal is caused by the vehicle itself and not by another road user or an object in the vehicle's vicinity. Based on the detected pattern, a fault in the vehicle can be inferred, such as a non-round bearing seat in the engine housing of the drive motor or an eccentrically running gear in the vehicle's drive train.Depending on the type of pattern detected and thus the detected, regularly repeated sequence of the noise signal, as well as, if applicable, the relationship between the regularity of the occurrence of the sequence and the rotor rotation, it is possible to determine what kind of fault the motor vehicle is experiencing.

[0011] The procedure further stipulates that sound signals are only evaluated in situations where at least one predefined criterion is present that influences background noise in the at least one sound signal. If it is determined that at least one predefined criterion is not present, then all sound signals are not evaluated. For example, if the vehicle is traveling at 200 km / h on the motorway, external noise from the vehicle is so loud that noises generated by the vehicle itself cannot necessarily be reliably and clearly recorded in the respective sound signals for the model by the respective microphones. By evaluating the sound signals only when at least one predefined criterion is present, reliable pattern recognition by the model in the at least one sound signal can be ensured.At least one criterion is chosen such that, when it is determined that at least one criterion is met, ambient noise from the vehicle's surroundings is particularly quiet and therefore as inconspicuous as possible in the respective recorded sound signals. This allows the model to identify specific patterns in the sound signals particularly easily and reliably.

[0012] In a possible further development of the invention, it is provided that the criterion is deemed to be met if it is determined that the speed of the motor vehicle is below 150 kilometers per hour, in particular below 100 kilometers per hour, and / or there is no sound-reflecting wall within a predetermined distance from the motor vehicle, and / or the motor vehicle is outside a tunnel, and / or the motor vehicle is at a distance from a bridge, and / or no rain is falling on the motor vehicle. The respective criteria mentioned above influence the respective background noise when the respective sounds are detected by the at least one microphone of the motor vehicle.If the relevant criteria are met, this indicates that the vehicle is most likely located in an environment where ambient noise is particularly quiet, or where there is very little ambient noise not directly caused by the vehicle. The slower the vehicle travels, the less noise is captured by the microphones due to wind noise and the noise caused by the vehicle's tires rolling on the road. If a sound-reflecting wall is located within the specified distance of the vehicle, then sound reflections from this wall can cause significant background noise to be recorded by the microphones.In this scenario, the wall can reflect noises generated by the vehicle or by other road users or objects towards the vehicle. This can cause the sound signals provided by at least one microphone to be heavily dominated by background noises reflected from the wall, making it difficult to recognize patterns in the sound signals using the model. If the vehicle is inside a tunnel, the tunnel walls can reflect sound, potentially leading to the aforementioned disadvantages. If the vehicle is on a bridge, the bridge can vibrate due to the vehicle's or other vehicles' movements, causing periodic background noises in the sounds recorded by the at least one microphone.This can cause patterns in the respective sound signal to be altered by interference or made difficult or even impossible for the model to recognize by being masked. If rain is dripping onto the vehicle, the individual drops can cause loud background noise in the sounds recorded by the at least one microphone. Similarly, a wet road surface can lead to increased background noise in the sounds recorded by the at least one microphone. Adherence to these criteria ensures that the proportion of background noise in the sounds captured by the at least one microphone is particularly low, and consequently, the respective patterns can be recognized particularly easily and precisely by the model.

[0013] In a possible further development of the invention, data for assessing whether at least one predetermined criterion is met is acquired by means of at least one speed sensor of the motor vehicle and / or by means of at least one environmental perception device of the motor vehicle. The speed sensor of the motor vehicle can be used to determine the speed at which the motor vehicle is moving. Consequently, it can be determined particularly easily and precisely whether the speed of the motor vehicle is below 150 km / h. The at least one environmental perception device can, for example, comprise a camera and / or a radar and / or a LiDAR.The environmental perception device makes it particularly easy and reliable to determine whether there is a sound-reflecting wall within a specified distance of the vehicle, and / or whether the vehicle is inside or outside a tunnel, and / or whether the vehicle is on, under, or at a distance from a bridge. Alternatively or additionally, the environmental perception device can determine whether rain is dripping onto the vehicle.

[0014] In a further possible embodiment of the invention, it is provided that the route guidance and / or the lateral guidance and / or the longitudinal guidance of the motor vehicle are adapted to bring the motor vehicle into a situation in which the at least one predetermined criterion is met. A set route, along which the motor vehicle is guided manually and / or by means of a driver assistance function, can thus be adapted to bring the motor vehicle into at least one situation in which the at least one predetermined criterion is met. By adapting the lateral guidance and / or the longitudinal guidance of the motor vehicle, a lane change to another lane can be implemented or a target course, in particular with a lateral deviation, can be set.It can thus be ensured that during the journey of the motor vehicle, the situation arises at least once in which at least one predefined criterion is met, and consequently at least one sound signal provided by at least one microphone of the motor vehicle can be evaluated and thus examined for the presence of a pattern.

[0015] In a further possible embodiment of the invention, it is provided that an action of the vehicle is triggered depending on the result of the assessment of the vehicle's condition. In particular, the action can be designed to react to a fault of the vehicle detected based on a determined pattern. This enables particularly safe operation of the vehicle, even if a fault is present, since at least one action can react to the detected fault.

[0016] In this context, it is specifically intended that the following actions will be taken: a warning will be issued in the passenger compartment of the vehicle, particularly via the vehicle itself or via a mobile electronic device located in the passenger compartment; and / or a request to visit a repair shop will be issued in the passenger compartment; and / or a stop procedure for the vehicle will be initiated; and / or a spare parts order will be automatically triggered. Issuing the request to visit a repair shop will encourage the driver to take the vehicle to a repair shop. Consequently, the vehicle can be inspected at the repair shop, specifically for the fault identified based on the determined pattern. Furthermore, this identified fault can be rectified at the repair shop, thus enabling particularly safe operation of the vehicle.Initiating the vehicle's stopping process brings it to a halt, for example, onto a shoulder or in a parking lot. Consequently, the vehicle leaves the flow of traffic, thus eliminating any potential hazard to itself or other road users. During the spare parts ordering process, a specific part is ordered that corresponds to the fault identified based on the determined pattern. This automatically triggered spare parts order allows for a particularly rapid repair of the vehicle, thereby resolving the fault very quickly.

[0017] In a further possible embodiment of the invention, the model is trained using one or more acoustic phenomena detected on a test bench. This test bench can, in particular, be a so-called end-of-line test bench. The end-of-line test bench enables comprehensive functional testing of the motor vehicle at the end of a manufacturing process. The model can thus be trained using noise signals recorded when the motor vehicle is positioned on the end-of-line test bench and / or using acoustic phenomena detected on the end-of-line test bench. The model can therefore be trained using insights or data obtained on an end-of-line test bench when examining at least one motor vehicle, and in particular several motor vehicles, which, after training, enables the model to reliably recognize patterns in noise signals provided by microphones.

[0018] In a further possible embodiment of the invention, the at least one recognized pattern and the vehicle evaluation based on this pattern are provided to a higher-level computing unit. This makes it possible to use the recognized pattern and the vehicle evaluation based on this pattern for training a further model. Furthermore, depending on the recognized pattern received by the higher-level computing unit and the vehicle evaluation based on this pattern, a test specification and / or a test method at the end-of-line test bench can be adapted. This allows for the further development and improvement of end-of-line vehicle testing using data from vehicles in the field.

[0019] The invention further relates to a motor vehicle with at least one microphone configured to provide a sound signal to an electronic processing unit, wherein the sound signal characterizes the sounds detected by the microphone. The motor vehicle further comprises a model, trained in particular by machine learning, which is configured to examine the sound signals for existing patterns. Alternatively or additionally, the model can be trained using rule-based methods, for example, fuzzy logic, or using map-based methods.Furthermore, the motor vehicle includes the electronic computing unit, which is designed to evaluate the state of the motor vehicle based on at least one pattern recognized by the model in the at least one sound signal and to trigger the evaluation of the motor vehicle based on sound signals which have been detected in situations in which at least one predetermined criterion is not present which has an influence on background noise in the at least one sound signal.In other words, the electronic processing unit is configured to trigger the evaluation of specific noise signals by the model, or the assessment of the motor vehicle based on the noise signals, only when the electronic processing unit has determined that the at least one noise signal is associated with a situation in which the at least one predefined criterion influencing background noise in at least one noise signal is present. Furthermore, the electronic processing unit can be configured to use the trained model to examine the at least one noise signal for existing patterns. The motor vehicle is thus configured to carry out the procedure as already described in connection with the inventive method for assessing the condition of a motor vehicle.

[0020] Further features of the invention may become apparent from the following description of the figures and from the drawings. The features and combinations of features mentioned above in the description, as well as the features and combinations of features shown below in the description of the figures and / or in the figures themselves, can be used not only in the combinations specified, but also in other combinations or individually, without departing from the scope of the invention.

[0021] The drawing shows in: Fig. 1. A schematic perspective view of a motor vehicle; and Fig. 2. A procedural scheme for a procedure for evaluating the condition of a motor vehicle.

[0022] In Fig. Figure 1 shows a schematic perspective view of a motor vehicle 10. The motor vehicle 10 includes at least one microphone 12. In this case, the motor vehicle 10 includes several microphones. Fig. 1 microphones 12, each schematically marked with a box. The respective microphones 12 can be directed with their detection range into the surroundings of the motor vehicle 10 or into an interior of the motor vehicle 10, in particular into the passenger compartment of the motor vehicle 10. The respective microphones 12 are configured to provide respective sound signals for an electronic processing unit 14 of the motor vehicle 10, wherein the respective sound signals characterize respective sounds detected by the microphones 12. The processing unit 14 is in Fig. 1 is also only schematically marked with a box. The motor vehicle 10 further comprises a machine learning-trained model, which may be stored on a memory device of the motor vehicle 10. The electronic processing unit 14 is specifically configured to examine the noise signals for existing patterns using the trained model. The electronic processing unit 14 is further configured to evaluate the motor vehicle 10 depending on at least one recognized pattern and at least one noise signal. In particular, the motor vehicle 10 is configured to carry out a procedure for evaluating a state of the motor vehicle 10.

[0023] A procedural scheme belonging to the procedure is contained in Fig. Figure 2 shows that the method comprises five process steps, V1 to V5. In the first process step, V1, sound signals are provided to the electronic processing unit 14 by means of the microphones 12. The respective sound signals from the respective microphones 12 characterize the sounds captured by the respective microphones 12. In the second process step, V2, the sound signals are examined for existing patterns using the trained model. In particular, the model is trained on acoustic phenomena detected at a test bench, especially an end-of-line test bench.

[0024] In the third procedural step V3, it is examined whether at least one predefined criterion influencing background noise in at least one noise signal is present. If it is determined that at least one predefined criterion is present, the procedure follows the arrow from the third procedural step V3 to the fourth procedural step V4. In particular, it is determined that the criterion is present if it is established that the speed of the motor vehicle 10 is below 150 kilometers per hour and / or there is no sound-reflecting wall within a predefined distance from the motor vehicle 10 and / or the motor vehicle 10 is outside a tunnel and / or the motor vehicle 10 is at a distance from a bridge and / or no rain is dripping onto the motor vehicle 10.Data for assessing whether at least one predefined criterion is met can be obtained by means of at least one speed sensor of the motor vehicle 10 and / or by means of at least one environmental perception device 18 of the motor vehicle 10. The speed sensor 16 and the environmental perception device 18 are in . Fig. 1 each schematically marked with a box.

[0025] If it is determined that the current route guidance does not lead the vehicle 10 to a situation where at least one predefined criterion is met, then the route guidance of the vehicle 10 can be adjusted to bring the vehicle 10 into a situation where at least one predefined criterion is met. Thus, the route along which the vehicle 10 travels can be adjusted to ensure that, while driving along the adjusted route, the vehicle 10 is in a situation where at least one predefined criterion is met at least once. Alternatively or additionally, lateral and / or longitudinal guidance of the vehicle 10 can be adjusted to bring the vehicle 10 into a situation where at least one predefined criterion is met.

[0026] In the fourth process step V4, the electronic processing unit 14 evaluates the condition of the motor vehicle 10 based on at least one pattern detected in the noise signals by the model. In the fifth process step V5, an action of the motor vehicle 10 is triggered based on the result of the evaluation of its condition. Thus, if the fourth process step V4 determines that a fault in the motor vehicle 10 exists with a probability above a certain threshold, then at least one action of the motor vehicle 10 can be triggered in the fifth process step V5.Specifically, the following actions are taken: a warning is issued in the passenger compartment of vehicle 10, and / or a request to visit a repair shop is issued in the passenger compartment of vehicle 10, and / or a stop procedure for vehicle 10 is triggered, and / or a spare parts order for vehicle 10 is automatically initiated. Alternatively or additionally, it is possible that the at least one recognized pattern and the evaluation of vehicle 10 based on this pattern are made available to a higher-level computing system, in particular a server.

[0027] It is possible that the third process step V3 is performed before the second process step V2 or before the first process step V1, and that the second process step V2 following the third process step V3, or the first process step V1 following the third process step V3, is only performed if it has been determined during the third process step V3 that at least one predefined criterion has been met. The sequence of process steps V1, V2, and V3 can therefore be: V1-V2-V3, V3-V1-V2, or V1-V3-V2.

[0028] The method enables the particularly early detection of the onset of wear on the motor vehicle 10. The method can effectively prevent delayed access to the motor vehicle 10 when damage occurs, as well as the occurrence of irreparable damage.

[0029] The microphones 12 allow for the evaluation of drive noise, chassis noise, and / or driving noise of the vehicle 10. Changes in noise can then be detected by the electronic processing unit 14 and evaluated using the model via AI pattern recognition. Depending on the risk assessment, a stop, a warning, a workshop repair, and / or a spare parts order can then be triggered. The procedure can pay particular attention to the vehicle 10's speed and its surroundings. At a speed of 200 kilometers per hour on a highway, external noise may be so loud that wear noises from the vehicle 10 may be particularly difficult to perceive.In particular, situations can be sought using at least one speed sensor, for example an electronic stability program, and / or using at least one environmental perception device 18, which detects tunnels, reflective walls and / or bridges, in which the respective wear noises of the motor vehicle 10 can be determined particularly easily and accurately in the respective noise signals. If there is a suspicion of wear of the motor vehicle 10, but the environmental noises of the motor vehicle 10 are particularly loud, then the route guidance of the motor vehicle 10 can be adjusted such that a route of the motor vehicle 10 is planned in such a way that the procedure can be used based on at least one situation occurring in which the at least one predefined criterion is present.In this context, the speed of the vehicle 10 on the respective route and the surroundings of the vehicle 10 can be taken into account. Alternatively or additionally, information on rain conditions or a wet road surface can also be used.

[0030] Within the framework of this process, the AI ​​pattern recognition system, and thus the model itself, can be trained using acoustic phenomena from an end-of-line test bench. This allows the model to recognize, quantify, and evaluate these acoustic phenomena in the vehicle 10 based on the noise signals provided by the microphones 12. This enables the implementation of a control loop for reject optimization at the end-of-line test bench. Furthermore, depending on the recognized acoustic phenomenon, components of the vehicle 10 can be replaced as needed to improve acoustic comfort. Alternatively, measures for acoustic optimization of the vehicle 10, or of other manufactured vehicles, can be initiated based on at least one pattern recognized in the vehicle 10.Based on data provided to the higher-level computing unit by the vehicle 10, in particular the at least one recognized pattern and the evaluation of the vehicle 10 based on this pattern, an automotive manufacturer can, in addition to assessing phenomena in the vehicle 10, gain insight into a tolerance level or implement a sustainable improvement of secondary acoustic measures such as damping and insulation in the vehicle 10, for example with regard to Shore hardness, engine mounting, or bulkhead insulation. Mutual cancellation of tolerance levels in the vehicle 10 is possible and increases the acceptance rate for the customer.

[0031] For example, a 4.6th order of noise in a B-bearing can be observed at 390 Hz and 450 Hz, depending on the gear ratio, at speeds of approximately 85 km / h and 120 km / h of the vehicle 10. This phenomenon is caused by a non-roundness of a bearing seat in the engine housing of the vehicle 10. The acoustics can be assessed using limit drives in a few vehicles 10. An immediate measure for acoustic optimization, particularly for reducing noise by three to five decibels, is to replace the cutting inserts of a housing machine. The aforementioned interval can be correlated with a field situation, and the scrap rate of vehicles 10 can be significantly reduced at given maintenance intervals, since vehicles 10 in which a defect has been detected can be recalled to a workshop at an early stage.The described phenomenon of the 4.6th engine order occurs 4.6 times per rotor revolution of the motor vehicle 10. A phenomenon of the first engine order, which thus occurs once per rotor revolution, can be a humming or rattling sound generated by a splined connection of a gear pair on the rotor shaft and the transmission input shaft. This acoustic phenomenon always occurs when the gearing runs eccentrically and is restricted too much by a centering seat.

[0032] The model can be configured to recognize previously unknown acoustic phenomena and thus patterns. For example, the data transmitted from the vehicle 10 to the higher-level computer can be used to determine how often a phenomenon occurs during operation in different vehicles 10 and how often the respective vehicles 10 in which this phenomenon has been observed have been taken to a workshop. Depending on the result of this evaluation, a limit value for the acoustic phenomenon can be adjusted, particularly in the end-of-line test bench, and / or a measure for acoustic optimization can be implemented in further produced vehicles 10 and / or in the at least one vehicle 10 that is supplying the data. The model is specifically trained to recognize which engine order the acoustic phenomenon, and thus the pattern, belongs to.Furthermore, the model can be trained to recognize the cause of the detected acoustic phenomenon or pattern. Specifically, it is intended that only patterns with levels above a predefined threshold in the respective noise signals are examined. Thus, only acoustic phenomena above a predefined sound level are investigated by the model for the presence of a pattern.

[0033] Overall, the invention demonstrates how a novel acoustic monitoring of a vehicle's condition can be carried out using AI pattern recognition. Reference symbol list 10 motor vehicle 12 microphones 14 electronic computing unit 16 Speed ​​sensor 18 Environmental perception device V1 to V5 respective process steps 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 10 2017 200 372 A1

[0002] DE 10 2022 127 924 B3

[0003] DE 10 2007 051 261 A1

[0004]

Claims

[1] Method for assessing the condition of a motor vehicle (10), in which - a sound signal is provided to an electronic computing unit (14) by means of at least one microphone (12) of the motor vehicle (10), wherein the sound signal characterizes sounds detected by means of the microphone (12) (V1), and - by means of a trained model that examines at least one noise signal with regard to existing patterns (V2), and - depending on at least one pattern detected by the model in which at least one noise signal the condition of the motor vehicle (10) is assessed (V4), - where noise signals are only evaluated for situations in which at least one predefined criterion (V3) is present, which has an influence on background noise in the at least one noise signal. [2] Method according to claim 1, wherein it is determined that the criterion (V3) is met if it is determined that - the speed of the motor vehicle (10) is below 150 km / h, and / or - there is no sound-reflecting wall within a specified distance from the motor vehicle (10), and / or - the motor vehicle (10) is outside a tunnel, and / or - the motor vehicle (10) is located at a distance from a bridge, and / or - a drop of rain does not fall on the motor vehicle (10). [3] Method according to claim 1 or 2, wherein data for assessing whether the at least one predetermined criterion is met are determined by means of at least one speed sensor (16) of the motor vehicle (10) and / or by means of at least one environmental perception device (18) of the motor vehicle (10) (V3). [4] Method according to one of the preceding claims, wherein a route guidance and / or a lateral guidance and / or a longitudinal guidance of the motor vehicle (10) are adapted to bring the motor vehicle (10) into a situation in which the at least one predetermined criterion is met. [5] Method according to one of the preceding claims, wherein depending on a result of the assessment of the condition of the motor vehicle (10) an action of the motor vehicle (10) is triggered (V5). [6] Method according to claim 5, wherein one action (V5) - a warning is issued in a passenger compartment of the motor vehicle (10), and / or - a request to visit a repair shop is issued in the passenger compartment of the motor vehicle (10), and / or - a stopping process of the motor vehicle (10) is triggered, and / or - a spare parts order is automatically triggered. [7] Method according to one of the preceding claims, wherein the model is trained on the basis of one or more acoustic phenomena detected on a test rig. [8] Method according to any of the preceding claims, wherein the at least one detected pattern and the evaluation of the motor vehicle (10) based on this pattern are made available to a higher-level computing device. [9] motor vehicle (10), with - at least one microphone (12) which is configured to provide a sound signal to an electronic computing unit (14) (V1), wherein the sound signal characterizes sounds detected by means of the microphone (12), - a trained model configured to examine at least one sound signal for existing patterns (V2), - the electronic computing unit (14), which is configured to evaluate the state of the motor vehicle (10) depending on at least one pattern recognized by the model in the at least one sound signal and to trigger the evaluation of the motor vehicle (10) on the basis of sound signals which have been recorded in situations in which at least one predetermined criterion which has an influence on background noise in the at least one sound signal is not present.

Images