Acoustic inspection of a motor vehicle

Abstract

Method (200) for valuing a motor vehicle (105), wherein the method (200) comprises the following steps: - Detection (210) of an operating noise during the execution of a predetermined function of a motor vehicle (105); - Determine (230) a disturbance factor associated with the operating noise; and - Training (235) an artificial neural network (135) to determine the disturbance factor of a provided functional noise, characterized in that a rating scale is specified or selected via an interaction device (140) (225) and the functional noise is rated with respect to the rating scale, wherein the rating scale - a first scale comprising a degree of wear or defect of an element of the motor vehicle involved in generating the operating noise (105), and - a second scale that expresses a person's subjective disturbance in the area of ​​the motor vehicle (105) due to the operating noise.

Description

[0001] The present invention relates to the acoustic testing of a motor vehicle. In particular, the invention relates to the evaluation of the operating noise of a function of the motor vehicle.

[0002] When a motor vehicle performs a function, a noise may be generated that is perceived by a user. The user may find the noise pleasant or unpleasant, loud or quiet. Furthermore, they may get the impression that the noise indicates a defect in a component of the motor vehicle involved in the function.

[0003] When developing a component for a motor vehicle, the operating noise generated during use is typically taken into account. This operating noise can be evaluated by a trained acoustics expert. To avoid biased assessments, several subjective evaluations are usually conducted and compared. The component can then be further developed based on these evaluations. Acoustic evaluation can require significant personnel and time resources, potentially delaying the component's development.

[0004] DE 10 2019 215 666 A1 relates to a method according to the preamble of claim 1.

[0005] One of the problems underlying the invention is to provide an improved technique for evaluating the operating noise of a motor vehicle. The invention solves this problem by means of the subject matter of the independent claims. Dependent claims describe preferred embodiments.

[0006] According to a first aspect, a first procedure for evaluating a motor vehicle includes steps of capturing a functional noise during the execution of a predetermined function of a motor vehicle; determining a disturbance factor associated with the functional noise; and training an artificial neural network (ANN) to determine the disturbance factor of a provided functional noise.

[0007] According to the invention, it is provided that a rating scale is specified or selected via an interaction device and the operating noise is evaluated with respect to the rating scale, wherein the rating scale - a first scale that includes a degree of wear or defect of an element of the motor vehicle involved in generating the operating noise, and - a second scale that expresses a person's subjective disturbance in the area of ​​the motor vehicle due to the operating noise.

[0008] The function of a motor vehicle can encompass virtually any function that can be meaningfully performed during normal use of the vehicle and that can produce a noise. Examples of such functions include opening or closing a door, hatch, or convertible top; operating a drivetrain, particularly a drive engine; using a ventilation system (HVAC: heating, ventilation, air conditioning); driving on different surfaces or at different speeds; or operating a control element, such as a gear selector or switch.

[0009] The KNN (Knowledge Neighborhood Network) can be trained to recognize a functional noise associated with a predetermined function of the vehicle and to assess how much this noise might disturb a person in the vicinity of the vehicle. This training can be performed using a large amount of data, in which disturbances from various devices or people have been evaluated. Data distortion due to bias on the part of a decision-maker can be avoided by simultaneously considering a large amount of trained data.

[0010] It is further preferred that operating sounds from different motor vehicles be used for training. The vehicles should be similar to each other to a predetermined degree to ensure comparability of the operating sounds. For example, the vehicles could be similar models but of different ages and / or mileages. The vehicles' equipment may also differ. The function can always be the same or a comparable one.

[0011] The trained KNN (cognitive-neighborhood) system can be used to enable the automatic evaluation of a functional noise in a motor vehicle. The repertoire of recognizable functional noises can be large and can be successively expanded through newly learned functional noises. Training can involve "deep learning," in which an intermediate layer of a multi-layered KNN is successively adapted to learned content. Training can also allow for the transfer determination of a functional noise that was not specifically learned, but for which one or more similar functional noises are known.

[0012] The noise factor can be determined in terms of the type and volume of the operating noise. The type of operating noise or the underlying function can be provided as a further input parameter. Alternatively, a k-nearest neighbors (KNN) algorithm can be trained to evaluate operating noises of only a predetermined function.

[0013] The level of disturbance can be determined based on the relative loudness of one or more predetermined partial noises encompassed by the functional noise. The functional noise can comprise a superposition and / or sequence of partial noises, which may have different qualities. The function may involve the operation of several vehicle components, and a partial noise may originate from one of these components or from a combination of several. For example, a first partial noise might be a squeak, and a second a rumble. The partial noises can be determined as having varying degrees of disturbance, with the less disturbing partial noise potentially masking the more disturbing one, so that the functional noise as a whole is less disturbing than the more disturbing of the two partial noises. The relative loudness of the partial noises can be considered for this determination.The KNN can be trained to assess different partial sounds simultaneously without having to perform the otherwise usual signal-technical separation of the partial sounds.

[0014] The level of disturbance can indicate whether the operating noise is acceptable or not. Determining the acceptability of an operating noise associated with a component's function can be crucial during the development of a motor vehicle or a component thereof. If the operating noise is unacceptable, the component must be appropriately modified. This may require a different technical approach to the component's function.

[0015] In another embodiment, it can be determined on a completed motor vehicle whether an operating noise must be considered anomalous or falls within a normal range of acceptable operating noises. A customer who makes a complaint about the motor vehicle due to an emitted operating noise can obtain an objective assessment using the KNN (Knowledge, Neighbors, Nearest) method.

[0016] The function is preferably implemented in such a way that any noise superimposed on the functional noise can be disregarded. For this purpose, the motor vehicle or the function in question can be operated under controlled conditions. Such conditions can be met, for example, on a test bench where driving the motor vehicle can be simulated without the vehicle leaving the test bench. Predetermined conditions for the function or the propagation of the functional noise can be repeatedly induced on the test bench.

[0017] The controlled conditions may include the requirement that the other noise, which does not originate from the execution of the predetermined function, is quiet enough not to influence the evaluation of the functional noise. In another embodiment, the other noise may be known and deliberately disregarded in the determination. In yet another embodiment, the other noise has characteristics that make it easily distinguishable from the functional noise being evaluated, allowing it to be disregarded. For example, the other noise may occupy a different frequency range or fluctuate with a different rhythm than the functional noise.

[0018] According to a further aspect of the present invention, a second method for evaluating a motor vehicle comprises steps of detecting a functional noise during the execution of a predetermined function of a motor vehicle; determining a disturbance factor associated with the functional noise by means of an artificial neural network trained to determine a disturbance factor of a provided functional noise; and providing the determined disturbance factor.

[0019] According to the invention, it is provided that a rating scale is specified or selected via an interaction device and the operating noise is evaluated with respect to the rating scale, wherein the rating scale - a first scale that includes a degree of wear or defect of an element of the motor vehicle involved in generating the operating noise, and - a second scale that expresses a person's subjective disturbance in the area of ​​the motor vehicle due to the operating noise, includes.

[0020] The KNN (Knowledge-nearest Neighbors) can be trained, in particular, using a first method described herein. Functional sounds learned in this way can advantageously be used to evaluate further functional sounds.

[0021] In a further preferred embodiment, the presence of a disturbance factor can be used to determine whether a defect exists in a component of the motor vehicle. This component is typically involved in generating the operating noise. In particular, a defect resulting from wear, imprecise assembly, or an exceeded tolerance of a component can be detected more effectively in this way. The function may be unaffected or only minimally impaired by the defect.

[0022] For example, it can be determined whether the noise indicates a worn bearing, an inaccurate fit, or imprecise guidance of a moving part. A lack of lubrication or maintenance, or damage due to aging, can also be identified based on the operating noise. For instance, the operating noise of a windshield wiper can help determine whether a wiper blade is worn or brittle.

[0023] Furthermore, it is preferable to determine the severity of the defect. The severity can be determined using k-nearest neighbors (KNN) analysis or a traditional technique based on a specific disturbance factor. The determined severity can facilitate a decision as to whether a particular defect requires specific action or whether the vehicle can continue to be operated unchanged. A developing defect can be monitored. The point in time when the severity of the defect exceeds a predetermined threshold can be predicted.

[0024] According to a further aspect of the present invention, a device for evaluating a motor vehicle comprises a microphone for capturing an operating noise during the execution of a predetermined function of the motor vehicle; an artificial neural network configured to determine a disturbance factor for a provided operating noise; and an output device for providing the determined disturbance factor. The ANN can be trained by a first method described herein.

[0025] According to the invention, the device for specifying or selecting a rating scale is equipped via an interaction device for evaluating the operating noise with respect to the rating scale, wherein the rating scale - a first scale that includes a degree of wear or defect of an element of the motor vehicle involved in generating the operating noise, and - a second scale that expresses a person's subjective disturbance in the area of ​​the motor vehicle due to the operating noise, includes.

[0026] The device can enable an evaluation of functional noise that is significantly faster than evaluation by a trained person. In one embodiment, the evaluation can be performed within a short, predetermined time, which may be on the order of a few seconds or even less than a second. Real-time evaluation of the functional noise is also possible. This can facilitate the manual adjustment of elements involved in the emission of the functional noise. The elements can be modified or adjusted under continuous acoustic monitoring and evaluation. The accuracy or reliability of the evaluation can be high due to the use of machine learning techniques.When evaluating a functional noise, it is also possible to specify the degree of certainty with which a predetermined noise or partial noise was detected, so that the statement made can be better verified.

[0027] According to yet another aspect of the invention, a motor vehicle comprises at least one permanently mounted microphone for a device described herein. Preferably, the motor vehicle comprises several microphones mounted at various predetermined locations within the vehicle. This allows for standardized sampling of functional sounds on the vehicle, thus facilitating easier comparison with functional sounds on another vehicle under similar conditions. Some of the microphones may be provided solely for testing or development purposes and not be installed in a production vehicle. Other microphones may be installed as standard equipment in the vehicle, making them readily available for acoustic evaluation of a function.

[0028] Examples of locations where a microphone might be placed include the interior, the engine compartment, or the trunk. A microphone can also be mounted on the exterior of the vehicle, for example, on the roof or underbody.

[0029] The first method, the second method, the apparatus, and the motor vehicle realize different aspects of the same invention presented herein. Features or advantages of the methods can be transferred to each other or to the apparatus, or vice versa.

[0030] The invention will now be described in more detail with reference to the attached drawings, in which: Fig. 1. a system; and Fig. 2. A flowchart of a process is illustrated.

[0031] Fig. Figure 1 shows a system 100 comprising a motor vehicle 105 and a device 110 for evaluating the operating noise of the motor vehicle 105. For recording the operating noise, a first microphone 115 is mounted in an interior area of ​​the motor vehicle 105, and a second microphone 120 is mounted in an engine compartment. A third microphone 125 is also provided outside the motor vehicle 105, positioned, for example, in a predetermined direction and at a predetermined distance from the motor vehicle 105. In further embodiments, additional or different microphones 115-125 may be provided, which can be mounted at predetermined locations on or relative to the motor vehicle 105.

[0032] A functional noise occurs when a predetermined function of the motor vehicle 105 is performed. The function typically involves a mechanical movement of one or more elements of the motor vehicle 105, with the functional noise occurring as an accompanying noise. The functional noise, or a component thereof, can also be intentionally produced, for example, by emitting a predetermined noise through a loudspeaker.

[0033] An example function might involve opening or closing the convertible top of a motor vehicle 105 and take approximately 4 to 8 seconds. During this process, various partial noises can be produced, which may overlap or follow one another. For example, operating the convertible top could include creaking, hissing, whirring, squeaking, grinding, or banging noises. The device 110 can help to assign a disturbance factor to the operating noise, indicating how strong or in what way the noise is desired or undesired, or disturbing or not disturbing.

[0034] In another example, a satisfying clicking sound might be desirable when closing a vehicle door of motor vehicle 105, but some satisfying clicking sounds may be perceived by a user as too loud, inappropriate for the intended function, or dissonant. Furthermore, the clicking sound may be perceived as intrusive or disturbing by a person not operating motor vehicle 105. Device 110 enables a repeatable and objective assessment of the clicking sound and its potential for disturbance.

[0035] The device 110 preferably comprises a control device 130 configured for implementing machine learning methods. For this purpose, the control device 130 preferably includes a KNN 135. Optionally, an interaction device 140 is provided, which allows an operator to input a parameter. The interaction can also allow the display or modification of a parameter.

[0036] Optionally, an output device 145 is provided, by means of which a specific disturbance factor of a functional noise can be output. In the illustrated embodiment, the output device 145 comprises, for example, three lights similar to a traffic light, so that a first light indicates that a detected functional noise is acceptable, a second light indicates that it is barely or no longer acceptable, and a third light indicates that it is unacceptable. Other configurations or output elements are also possible. The output device 145 can be located in the vicinity of the motor vehicle 105, so that it can be read by a person in the vicinity of the motor vehicle 105.

[0037] The motor vehicle 105 and the device 110 are preferably located in an acoustically controlled environment, for example in a hall or on a test stand. This ensures comparability of operating noises for different versions of the function.

[0038] It is proposed that the KNN 135 be trained in a first phase to recognize a disturbance factor of a functional noise that occurs during the execution of a predetermined function of the motor vehicle 105. The function can be specified or selected using the interaction device 140. The training can be based on a large number of functional noises that are associated with the same function of the motor vehicle 105. The functional noises can be generated on different motor vehicles 105. The conditions under which the function is executed can be the same or differ slightly from each other in a predetermined way.

[0039] Different conditions during the generation of a functional noise can be caused by another function being executed simultaneously. For example, a functional noise may vary depending on the vehicle's speed. Background noise caused by another function or environmental influences can partially mask or override the functional noise.

[0040] The attribution of a disturbance factor to a functional noise can be carried out by a person at this stage. The person can indicate the disturbance factor in relation to a predetermined scale or question. For example, the disturbance factor can indicate how loud, unpleasant, or inappropriate the functional noise is considered to be. The scale can also indicate how clearly the functional noise indicates a defect in a component involved in producing the functional noise, or how severe the defect is.

[0041] In a second phase, the operating noise of the motor vehicle 105 can be evaluated using the KNN 135, and a specific disturbance factor can be provided by means of the output device 145. Preferably, the underlying function is also specified here by means of the interaction device 140.

[0042] Fig.Figure 2 shows a flowchart of a procedure 200 for evaluating a motor vehicle 105. The procedure 200 can essentially be carried out using a device 110 and requires a previously trained KNN 135.

[0043] In step 205, it can be determined which function of the motor vehicle 105 is to be evaluated. The function can be specified or selected using the interaction device 140. In step 210, the function can be executed, and any resulting operating noise can be recorded.

[0044] The recorded functional noise can then be evaluated using the KNN 135 algorithm with respect to a disturbance factor on a predetermined scale. In step 215, various partial noises encompassed by the functional noise can be identified and / or evaluated. In step 220, the acoustic interaction of these partial noises can be determined. This interaction can take into account the temporal sequence or relative loudness of the partial noises. It can then be determined whether the partial noises influence each other, for example, by mutually attenuating or amplifying each other.

[0045] In step 225, a rating scale can be selected. The rating scale can be specified or selected via the interaction device 140. In a subsequent step 230, the operating noise can be rated with respect to the scale. A disturbance factor can be determined, which generally indicates where the rating of the operating noise lies on the selected scale. A first example scale could include the degree of wear or defect of a component of the motor vehicle 105 involved in generating the operating noise. A second example scale could express a person's subjective disturbance caused by the operating noise in the vicinity of the motor vehicle 105. Other scales are also possible.

[0046] In step 235, the specific rating can be provided, for example, by means of the output device 145. An indication of the scale used can also be provided. The scale can include discrete or continuous values, and the rating can be output in any appropriate manner, for example, graphically, as a numerical value, or as a symbolic value.

[0047] It should be noted that procedure 200, with slight modifications, can also be used to train a KNN 135 for a specific evaluation. In step 230, instead of providing an evaluation, an evaluation assigned to the investigated functional noise can be recorded. This evaluation can be created in any way, for example, manually by a person. In step 235, the KNN 135 can be trained with respect to the functional noise and the evaluation. A scale to be used for this purpose can be predefined, for example, by selecting or entering it using the interaction device 140.

[0048] Reference sign 100 System 105 motor vehicles 110 Device 115 first microphone 120 second microphone 125 third microphone 130 Control device 135 artificial neural network, ANN 140 Interaction facility 145 Dispensing device 200 procedures 205 Determine function 210 Execute function and record operating noise 215 Determine partial noise Determine 220 volume levels 225 Classify operating noise 230 Evaluate operating noise / Record evaluation 235 Provide rating / Learn rating

Claims

[1] Method (200) for valuing a motor vehicle (105), wherein the method (200) comprises the following steps: - Detection (210) of an operating noise during the execution of a predetermined function of a motor vehicle (105); - Determine (230) a disturbance factor associated with the operating noise; and - Training (235) an artificial neural network (135) to determine the interference factor of a provided functional noise, characterized by , that a rating scale is specified or selected via an interaction device (140) (225) and the operating noise is rated with respect to the rating scale, wherein the rating scale - a first scale comprising a degree of wear or defect of an element of the motor vehicle involved in generating the operating noise (105), and - a second scale that expresses a person's subjective disturbance in the area of ​​the motor vehicle (105) due to the operating noise. [2] Method (200) according to claim 1, wherein, when the specified disturbance factor is provided, an indication of the scale used is given. [3] Method (200) according to claim 1 or 2, wherein the disturbance factor - regarding the type and volume of the operating noise and / or - is determined on the basis of the relative loudness of one or more predetermined partial noises encompassed by the operating noise (230). [4] Method (200) according to one of the preceding claims, wherein the disturbance factor indicates whether the operating noise is acceptable or not. [5] Method (200) according to one of the preceding claims, wherein the function is carried out (210) in such a way that a noise superimposed on the functional noise can be neglected. [6] Method (200) for valuing a motor vehicle (105), wherein the method (200) comprises the following steps: - Detection (210) of an operating noise during the execution of a predetermined function of a motor vehicle (105); - Determining (230) a disturbance factor associated with the functional noise using an artificial neural network (135) trained to determine a disturbance factor of a provided functional noise; and - Providing (235) the specified disturbance factor, characterized by , that a rating scale is specified or selected via an interaction device (140) (225) and the operating noise is rated with respect to the rating scale, wherein the rating scale - a first scale comprising a degree of wear or defect of an element of the motor vehicle involved in generating the operating noise (105), and - a second scale that expresses a person's subjective disturbance in the area of ​​the motor vehicle (105) due to the operating noise, includes. [7] Method (200) according to claim 6, wherein when the specified disturbance factor is provided, an indication of the scale used is given. [8] Method (200) according to claim 6 or 7, wherein - based on the disturbance factor, it is determined whether a defect exists in an element of the motor vehicle (105), - determines the severity of the defect and / or - a time when the severity of the defect exceeds a predetermined threshold is predicted. [9] Device (110) for evaluating a motor vehicle (105), wherein the device (110) comprises the following elements: - a microphone (115-125) for recording an operating noise during the execution of a predetermined function of a motor vehicle (105); - an artificial neural network (135) designed to determine a disturbance factor for a provided functional noise; and - an output device (145) for providing the specified disturbance factor, characterized by , that the device (110) is configured to specify or select (225) a rating scale via an interaction device (140) for evaluating the operating noise with respect to the rating scale, wherein the rating scale - a first scale comprising a degree of wear or defect of an element of the motor vehicle involved in generating the operating noise (105), and - a second scale that expresses a person's subjective disturbance in the area of ​​the motor vehicle (105) due to the operating noise, includes. [10] Device (110) according to claim 9, wherein the output device (145) is configured to provide an indication of the scale used when the specified disturbance factor is provided. [11] motor vehicle (105), comprising - at least one permanently attached microphone (115, 120) for a device (110) according to claim 9 or 10, - a device (110) according to claim 9 or 10 and / or - a device (110) for carrying out a method according to any one of claims 1 to 8.

Images