Method for detecting chassis defects of a vehicle

Abstract

The invention relates to a method for detecting chassis defects of a vehicle. The method is characterized in that - chassis noise (FG) and chassis parameters (FP) are determined during different operating conditions of the vehicle, - the determined chassis noises (FG) and chassis parameters (FP) are fed to an artificial neural network (1), which is used in a training (Tr) to generate reference data (RD) with different operating states with a defect-free chassis - recorded tone sequences (T') and sound levels (P') of chassis noise (FG'), - determined chassis parameters (FP') and - recorded noise (SG') was trained, - by means of the artificial neural network (1), a target-actual comparison is carried out between the determined chassis noises (FG) and chassis parameters (FP) and the learned chassis noises (FG') and chassis parameters (FP') and disturbing noises (SG) are continuously compensated using the reference data (RD) and - a chassis defect is detected by means of the artificial neural network (1) if a result of the target / actual comparison exceeds a predetermined threshold value. The invention further relates to a method for operating a vehicle.

Description

[0001] The invention relates to a method for detecting chassis defects of a vehicle according to the preamble of claim 1.

[0002] The invention further relates to a method for operating a vehicle.

[0003] From DE 10 2020 211 876 A1 a method for detecting a defect in a chassis of a vehicle is known with the following method steps: - Detecting an acoustic vibration and an acceleration using a first sensor system of the vehicle, - Detecting an event that exerts a mechanical influence on the vehicle by means of a second sensor arranged on the vehicle, - Assigning time information to the event, which describes a point in time at which the event exerts the mechanical influence on the vehicle, and - Detecting a defect based on the detected acoustic vibration and acceleration recorded at the time described by the time information and the detected event.

[0004] The first sensor system comprises an acoustic sensor for detecting the acoustic vibration in the form of sound or structure-borne sound. The first sensor system and the second sensor system are coupled to a control unit that executes the method. The control unit has signal processing units, each of which is a neural network or part of such a neural network.

[0005] Furthermore, DE 10 2019 210 931 A1 discloses a method and a device for detecting and verifying a faulty vehicle component. Detection is based on image data from a sensor system for environmental detection, and defects in the vehicle's chassis are detected using an artificial neural network. The artificial neural network is trained with reference image data, which is either configured as artificial image sensor data with predefined anomalies or as image sensor data with and without anomalies.

[0006] DE 10 2017 112 322 A1 describes a method and a device for autonomously monitoring the integrity of a vehicle. Vehicle vibration data and vehicle noise data are recorded, with the noise data comprising sound pressure level measurements taken inside and outside the vehicle using a microphone.

[0007] For further state-of-the-art information, reference can also be made to DE 10 2021 123 020 A1, which describes the detection of chassis defects using artificial neural networks. Noises on board a vehicle are recorded, and patterns in the recorded noises are recognized in order to assign the recognized patterns to predetermined classifications. To train artificial neural networks, a large number of noises with associated operating states are collected from a large number of vehicles.

[0008] Furthermore, reference is made to US 2022 / 0 097 717 A1, which describes the use of a neural network to estimate the operating noise of a vehicle component driven by the rotary motion of a drive, e.g., an automatic transmission. The value for an operating noise index is the difference between the sound pressure level of the operating noise and the sound pressure level of the vehicle's background noise.

[0009] The object of the present invention is to provide a novel method for detecting chassis defects of a vehicle and a novel method for operating a vehicle.

[0010] The object is achieved according to the invention by a method for detecting chassis defects, which has the features specified in claim 1, and by a method for operating a vehicle, which has the features specified in claim 9.

[0011] Advantageous embodiments of the invention are the subject of the subclaims.

[0012] In a method for detecting chassis defects of a vehicle, chassis noises caused by the chassis during operation of the vehicle are recorded by means of at least one acoustic sensor and fed to at least one artificial neural network for evaluation.

[0013] According to the invention, the method is characterized in that - the chassis noise and chassis parameters are determined during different operating conditions of the vehicle, - the determined chassis noises and chassis parameters are fed into the artificial neural network, which is used in a training to generate reference data with different operating conditions with a defect-free chassis - recorded sound sequences and sound levels of chassis noise, - determined chassis parameters and - recorded background noises were trained, - using the artificial neural network, a target-actual comparison is carried out between the determined chassis noises and chassis parameters and the learned chassis noises and chassis parameters, and the disturbing noises are continuously compensated using the reference data, and - a chassis defect is detected by means of the artificial neural network if the result of the target-actual comparison exceeds a specified threshold.

[0014] As the degree of vehicle automation increases, responsibility is transferred from the driver to the vehicle or the vehicle manufacturer. This means that the vehicle must be able to recognize in a timely manner when a vehicle condition no longer ensures safe participation in road traffic. While the driver is generally still responsible for the condition of the vehicle, premature wear or defects in chassis components can occur, for example due to contact with a curb or driving over potholes, which cannot be detected during a regular or cyclical vehicle inspection or general inspection. If a driver activates an automated driving function in their vehicle, for example according to SAE Level 3 or higher, there is a risk that the vehicle with a damaged chassis will take over the driving task.Likewise, chassis damage can occur during automated driving. In highly automated vehicles, for example, those operating at SAE Level 5, there is a risk of undetected chassis defects because the vehicles are not driven by human drivers. However, there is also a risk that a human driver may fail to detect chassis defects, including chassis wear.

[0015] The present method particularly advantageously enables the automated, early detection of chassis defects, i.e., chassis damage and / or chassis wear, even outside of regular or cyclical vehicle inspections in a workshop. This allows accidents resulting from safety-relevant chassis defects and the associated personal injury and property damage to be avoided, or at least the risk of their occurrence can be reduced. Compensating for background noise using reference data, which is stored, for example, in a reference curve, enables particularly accurate detection of chassis defects. The method also enables remote diagnosis of the chassis, for example, in the event of a vehicle user's complaints, and an assessment and notification to the vehicle user as to whether and to what extent the vehicle is drivable in the presence of a chassis defect.Furthermore, the method is simple and can be implemented with little hardware effort, since sensors already present in or on the vehicle, for example a microphone for detecting chassis noise and background noise, can be used.

[0016] According to one possible embodiment of the method, the type and / or severity of the chassis defect and / or at least one component causing the chassis defect is / are determined based on the target-actual comparison. This enables a tailored selection of measures to reduce the risk of impacts on the vehicle's operation. For this purpose, one possible embodiment of the method provides for different chassis defects to be learned during training of the artificial neural network, with defined tone sequences and sound levels of chassis noise being assigned to each chassis defect depending on the chassis parameters occurring.

[0017] According to a further possible embodiment of the method, the detection of chassis defects is carried out while the vehicle is in operation and / or in an automated test mode performed at the request of a vehicle user or during a workshop visit while the vehicle is stationary. Performing detection while the vehicle is in operation enables continuous or regular automated checking of the chassis for defects that do not impair the vehicle's ability to perform its driving task or the comfort of a vehicle user. Performing detection in test mode enables the chassis to be checked for defects under defined and reproducible conditions. This enables particularly precise detection of chassis defects and very good comparability with results already determined in previous checks.

[0018] According to another possible embodiment of the method, in the test mode, when the vehicle is stationary, steering actuators and / or chassis actuators are activated to actively change the chassis parameters, or a request is issued to a vehicle user to perform at least one action that influences the chassis. This enables a reproducible excitation of the chassis to generate chassis movements and thus the testing of the chassis for defects under defined and reproducible conditions.

[0019] According to a further possible embodiment of the method, the noises are at least - Noise from a vehicle drive and / or - Noise from a vehicle brake and / or - Rolling noise from the vehicle’s wheels and / or - Wind noise and / or - Ambient noise is taken into account. By considering these noises as background noise, influences on the results during the acoustic detection of chassis defects can be largely avoided.

[0020] According to a further possible embodiment of the method, at least - Suspension travel and / or rebound travel of the chassis and / or - Vibration frequencies of chassis components and / or - a loading condition of the vehicle and / or - Proper movements of the chassis and / or - a rack force of a steering system of the vehicle and / or - a steering angle of a vehicle's steering and / or - a steering angle velocity of the vehicle's steering system is taken into account. By taking these chassis parameters into account, the vehicle's various operating states can be represented at least almost completely. The steering system can comprise front-axle steering and / or rear-axle steering.

[0021] According to a further possible embodiment of the method, the disturbing noises and / or chassis parameters and / or the chassis noises are determined at least as a function of - a speed of the vehicle and / or - a longitudinal acceleration of the vehicle and / or - lateral acceleration of the vehicle and / or - an acceleration of the vehicle in the vertical direction, - a brake pressure of a brake of the vehicle and / or - Data describing the states of a vehicle's drivetrain is determined. By taking these parameters into account, the representation of the vehicle's various operating states can be further improved.

[0022] In a method according to the invention for operating a vehicle, the aforementioned method is used to determine whether a chassis defect is present. If a chassis defect is present, the following measures are / will be taken: - a message concerning the chassis defect is displayed inside the vehicle and / or on a terminal device and / or - a driving instruction is issued within the vehicle and / or on a device and / or - a control behavior of a vehicle's driving dynamics control system is adjusted and / or - a maximum driving speed of the vehicle is limited and / or - an automated driving function is deactivated and / or - during automated driving of the vehicle, the driving speed of the vehicle is reduced and / or - information concerning the chassis defect is transmitted to a central processing unit and / or control center external to the vehicle and / or - during automated driving, the vehicle is guided to a specified position.

[0023] Due to the automated detection of chassis faults, the method particularly advantageously enables the prompt implementation of measures upon detection of chassis faults to alert a vehicle user or vehicle owner to the presence of a chassis defect and / or to implement measures relating to vehicle operation in order to prevent accidents resulting from safety-relevant chassis faults and the associated personal injury and property damage, or at least to reduce the risk of such accidents occurring. For example, this can prevent automated driving functions from being activated when the vehicle is not roadworthy. The method also enables remote diagnosis of the chassis, for example, in the event of a vehicle user raising complaints, and an assessment and notification to the vehicle user as to whether and to what extent the vehicle is roadworthy in the presence of a chassis defect.Furthermore, by taking this action, a vehicle user can be conveniently informed of further action in the event of a detected chassis defect, such as a route to the nearest workshop, calling a towing service, etc. The described remote diagnosis is also possible, for example, for several or all vehicles in a fleet, such as fully automated driverless taxis.

[0024] According to one possible design of the procedure, the measure and / or a characteristic of the measure is / are selected depending on the type and / or severity of the landing gear defect. This enables a targeted adaptation of the type of measure and / or its characteristic, i.e., the degree and / or severity and / or extent of the measure, in order to select the best possible approach for dealing with or rectifying the landing gear defect depending on the specific landing gear defect.

[0025] Embodiments of the invention are explained in more detail below with reference to a drawing.

[0026] It shows: Fig. 1 schematically shows a process flow of a method for detecting defects in a vehicle chassis.

[0027] In the only Fig.1 schematically shows a sequence of a possible embodiment of a method according to the invention for detecting chassis defects of a vehicle.

[0028] Depending on the extent of the damage to the chassis, chassis defects on a vehicle generate chassis noises (FG), such as creaking during steering maneuvers due to a worn tie rod end or rattling and clattering noises during one-sided compression due to worn coupling rod joints. In order to detect chassis defects, including chassis wear, in a timely manner during or before the start of a journey, the method described below is used to acoustically detect chassis defects by recording and evaluating chassis noises (FG).

[0029] In a method step S1, chassis noises FG are recorded during various operating states of the vehicle using at least one acoustic sensor, for example, a microphone. The chassis noises FG include, for example, noises resulting from bearing play in the chassis bearings, mechanical friction of chassis components, and / or movements of chassis components.

[0030] Simultaneously with the chassis noise FG, background noise SG caused by vehicle operation and / or its environment is recorded. Background noise SG includes, for example, at least the noise of the vehicle's drive system, the noise of the vehicle's brakes, the rolling noise of the vehicle's wheels, wind noise, and / or ambient noise.

[0031] To record the chassis noise FG and background noise SG, tone sequences T and sound levels P are recorded.

[0032] Furthermore, simultaneously with the recording of the chassis noises FG and background noises SG, a recording of chassis parameters FP is carried out that depend on a respective operating state of the vehicle present during the recording or that describe this operating state. The chassis parameters FP recorded include at least the suspension travel and / or rebound travel of the chassis, vibration frequencies of chassis components, a vehicle load state, the chassis's own movements, a rack force of a vehicle's steering system, a steering angle of a vehicle's steering system, and / or a steering angular velocity of a vehicle's steering system. The steering system can comprise a front-axle steering system and / or a rear-axle steering system. The vehicle's own sensors are used for the recording.

[0033] Furthermore, the background noise SG, chassis parameters FP, and / or chassis noise FG are determined as a function of vehicle parameters CP. The vehicle parameters CP include, for example, a vehicle speed, a longitudinal acceleration of the vehicle, a lateral acceleration of the vehicle, a vertical acceleration of the vehicle, a brake pressure of a vehicle brake, and / or data describing the states of a vehicle drive.

[0034] The detection of chassis defects, i.e. the recording of chassis noises FG, noises SG and chassis parameters FP as well as the evaluation of these, is carried out during driving operation FB of the vehicle, in which movements of the vehicle and resulting chassis movements take place, i.e. during normal driving of the vehicle.

[0035] After the chassis noise FG, the background noise SG, and the chassis parameters FP have been recorded or determined, these are processed in a further process step S2 during vehicle operation FB using a signal processing system. For this signal processing, the chassis noise FG, the background noise SG, and the chassis parameters FP are fed into a trained artificial neural network 1.

[0036] The artificial neural network 1 is or was before use in the driving operation FB of the vehicle in a training Tr to generate reference data RD, for example a reference curve, with in different operating states with a defect-free chassis - recorded sound sequences T' and sound levels P' of chassis noise FG', - determined chassis parameters FP' and - recorded background noises SG'. The training Tr thus records the noises caused by the landing gear in various operating states when the landing gear does not exhibit any landing gear defects.

[0037] As disturbing noises SG', for example, at least noises from a vehicle's drive, noises from a vehicle's brakes, rolling noises from the vehicle's wheels, wind noise and / or ambient noise are recorded.

[0038] To record the chassis noise FG' and background noise SG', tone sequences T' and sound levels P' are recorded.

[0039] To assign the recorded chassis noises FG' to the various operating states, for example, compression and rebound at different vertical speeds, steering angle changes, cornering, driving over speed bumps, etc., the recorded chassis noises FG' are assigned to the simultaneously determined chassis parameters FP'. As chassis parameters FP', at least the suspension travel and / or rebound travel of the chassis, vibration frequencies of chassis components, a vehicle load state, inherent movements of the chassis, a rack force of a vehicle's steering system, a steering angle of a vehicle's steering system, and / or a steering angular velocity of a vehicle's steering system are recorded. On-board sensors are used for recording. Furthermore, for assignment to the operating states, the background noises SG', chassis parameters FP', and / or chassis noises FG' are determined depending on vehicle parameters CP'.In particular, the noise SG' can also be assigned to the vehicle's operating states. The vehicle parameters CP' include, for example, a vehicle speed, a longitudinal acceleration of the vehicle, a lateral acceleration of the vehicle, a vertical acceleration of the vehicle, a brake pressure of a vehicle's brake, and / or data describing the states of a vehicle's drive.

[0040] The training Tr of the artificial neural network 1 can be performed, for example, during the development of the vehicle or chassis, or during a learning routine after a defect-free chassis inspection. The learning routine can be started, for example, after a specialist workshop has inspected or repaired the chassis and ensured that the chassis is in the correct condition, in order to compensate for any changes in disturbance variables SG', such as a change in rolling noise due to a different tire configuration.

[0041] In the signal processing carried out by means of the trained artificial neural network 1 in method step S2, the disturbing noises SG detected during driving operation FB are continuously compensated by means of the network 1 in a compensation K based on the reference data RD.

[0042] Subsequently, in a further method step S3, during driving operation FB of the vehicle, an acoustic evaluation of the detected chassis noises FG, filtered for the background noises SG in method step S2, and the chassis parameters FP assigned to the chassis noises FG are carried out by means of the artificial neural network 1. A target-actual comparison is performed between the determined chassis noises FG and chassis parameters FP and the learned chassis noises FG' and chassis parameters FP'. A chassis defect is detected by means of the artificial neural network 1 if a result of the target-actual comparison exceeds a predetermined threshold value. In addition, one possible embodiment provides that the type and / or severity of the chassis defect and / or at least one component causing the chassis defect is / are determined based on the target-actual comparison.

[0043] In a further method step S4, during vehicle operation FB, a chassis status, i.e., whether or not a chassis defect is present, is output to other vehicle systems. These systems execute measures that can prevent accidents resulting from safety-relevant chassis faults and the associated personal injury and property damage, or at least reduce the risk of their occurrence. The measure and / or a characteristic of the measure can be selected depending on the type and / or severity of the chassis defect.

[0044] For example, if a chassis defect is present, in a method step S5, an output unit is used to output an indication relating to the chassis defect within the vehicle and / or on a terminal device, for example a mobile phone of a vehicle user.

[0045] For example, the vehicle user can be informed that there is a chassis defect. The vehicle user can also be informed about the type and severity of the chassis defect.

[0046] Furthermore, a driving instruction can be displayed within the vehicle and / or on the device, informing the vehicle user, for example, that they should reduce the vehicle's speed and / or visit a repair shop. This can also include a suggested route to a suitable nearby repair shop or an automated call for a towing service.

[0047] In addition or alternatively to the measures carried out in process step S5, measures relating to the driving operation FB of the vehicle are carried out in a process step S6.

[0048] In this case, for example, the control behavior of a vehicle's driving dynamics control system can be adapted by, for example, reducing control thresholds in the driving dynamics control system in order to increase the driving stability of the vehicle in critical situations when there is a chassis defect, for example a worn tie rod end during dynamic driving.

[0049] The maximum possible driving speed of the vehicle can also be limited both during control of the vehicle by a vehicle user and during automated, in particular highly automated or autonomous, control of the vehicle.

[0050] Furthermore, it is possible that an automated driving function of the vehicle is deactivated, the driving speed of the vehicle is automatically reduced during the automated driving mode FB of the vehicle and / or the vehicle is guided to a predetermined position during the automated driving mode FB.

[0051] For all measures carried out in process step S6 that influence the driving operation FB of the vehicle, additional information describing the corresponding measure can be output to the vehicle user in process step S5.

[0052] In addition to issuing instructions and information to the vehicle user in process step S5 and implementing measures affecting the vehicle's driving operation (FB) in process step S6, the information relating to the chassis defect can also be transmitted to an external central processing unit and / or control center in process step S4. The central processing unit or control center can then execute and / or initiate appropriate measures to prevent accidents resulting from safety-relevant chassis faults and the associated personal injury and property damage, or at least reduce the risk of their occurrence.For example, in the event of a chassis defect in a fully autonomous vehicle, such as a driverless so-called robotaxis, the control center can automatically guide the vehicle to a defined location, such as a workshop or assembly point or the control center, and / or stop the vehicle from operating.

[0053] The previously described procedure for detecting and reporting chassis defects can be carried out during a test mode PM when the vehicle is stationary, as an alternative to being carried out during normal driving operation FB of the vehicle.

[0054] The PM test mode is activated and executed automatically, for example cyclically at predefined intervals, at the request of the vehicle user or while the vehicle is in a workshop. In the PM test mode, when the vehicle is stationary, steering actuators and / or chassis actuators are controlled to actively change the chassis parameters or a request is issued to a vehicle user to perform at least one action that influences the chassis. For example, while the PM test mode is executed, a central processing unit, on which method steps S1 to S6 are also carried out, controls the steering actuators and / or chassis actuators of an active chassis to perform defined movements on the chassis. The control is carried out in such a way that possible chassis defects can be quickly detected. For example, an oscillating control of a rack and pinion steering is carried out for this purpose.The recording, processing and evaluation of the chassis noises FG, background noises SG and chassis parameters FP is carried out in the test mode PM analogously to the previously described recording, processing and evaluation carried out during the driving operation FB of the vehicle.

Claims

[1] Method for detecting chassis defects of a vehicle, wherein chassis noises (FG) caused by the chassis are detected by means of at least one acoustic sensor during operation of the vehicle and are fed to at least one artificial neural network (1) for evaluation, characterized by , that - the chassis noise (FG) and chassis parameters (FP) are determined during different operating conditions of the vehicle, - the determined chassis noises (FG) and chassis parameters (FP) are fed to the artificial neural network (1), which is used in a training (Tr) to generate reference data (RD) with different operating states with a defect-free chassis - recorded tone sequences (T') and sound levels (P') of chassis noise (FG'), - determined chassis parameters (FP') and - recorded noise (SG') was trained, - by means of the artificial neural network (1), a target-actual comparison is carried out between the determined chassis noises (FG) and chassis parameters (FP) and the learned chassis noises (FG') and chassis parameters (FP') and disturbing noises (SG) are continuously compensated using the reference data (RD) and - a chassis defect is detected by means of the artificial neural network (1) if a result of the target-actual comparison exceeds a predetermined threshold value. [2] Method according to claim 1, characterized by that the type and / or severity of the chassis defect and / or at least one component causing the chassis defect is / will be determined based on the target / actual comparison. [3] Method according to claim 1 or 2, characterized by that the detection of chassis defects is carried out during driving operation (FB) of the vehicle and / or in a test mode (PM) while the vehicle is stationary. [4] Method according to claim 3, characterized by that the test mode (PM) is carried out automatically, at the request of a vehicle user or during a workshop visit. [5] Method according to claim 3 or 4, characterized by that in the test mode (PM) when the vehicle is stationary, steering actuators and / or chassis actuators are controlled to actively change the chassis parameters (FP) or a request is issued to a vehicle user to carry out at least one action influencing the chassis. [6] Method according to one of the preceding claims, characterized by that as noise (SG, SG') at least - Noise from a vehicle drive and / or - Noise from a vehicle brake and / or - Rolling noise from the vehicle’s wheels and / or - Wind noise and / or - Ambient noise is taken into account. [7] Method according to one of the preceding claims, characterized by that the chassis parameters (FP, FP') are at least - Suspension travel and / or rebound travel of the chassis and / or - Vibration frequencies of chassis components and / or - a loading condition of the vehicle and / or - Proper movements of the chassis and / or - a rack force of a steering system of the vehicle and / or - a steering angle of a steering system of the vehicle and / or - a steering angle speed of a vehicle's steering system is / will be taken into account. [8] Method according to one of the preceding claims, characterized by that the disturbing noises (SG, SG') and / or chassis parameters (FP, FP') and / or the chassis noises (FG, FG') are at least dependent - a speed of the vehicle and / or - a longitudinal acceleration of the vehicle and / or - lateral acceleration of the vehicle and / or - an acceleration of the vehicle in the vertical direction, - a brake pressure of a brake of the vehicle and / or - Data describing the states of a vehicle's drive system can be determined. [9] Method for operating a vehicle, wherein a method according to one of the preceding claims is used to determine whether a chassis defect is present, and in the case of a chassis defect, as a measure - a message concerning the chassis defect is displayed within the vehicle and / or on a terminal device and / or - a driving instruction is issued within the vehicle and / or on a terminal device and / or - a control behavior of a vehicle's driving dynamics control system is adjusted and / or - a maximum driving speed of the vehicle is limited and / or - an automated driving function is deactivated and / or - during automated driving (FB) of the vehicle, the driving speed of the vehicle is reduced and / or - information concerning the chassis defect is transmitted to a central processing unit and / or control center external to the vehicle and / or - during automated driving (FB) the vehicle is guided to a specified position. [10] Method according to claim 9, wherein the measure and / or a form of the measure is / are selected depending on a type and / or severity of the chassis defect.

Images